1 ORGANIC INORGANIC HYBRID PHOTOVOLTAIC CELLS By RENJIA ZHOU A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY U NIVERSITY OF FLORIDA 2012
2 2012 Renjia Zhou
3 To Xiaoc hang Miao
4 ACKNOWLEDGMENTS The success of my PhD study is only possible because of the contribution from many people in my life. First, I want to ac knowledge my advisor, Prof. Jiangeng Xue, for guiding and supporting my PhD study. His knowledge academic attitude and high research standard are a great example for me to follow in the present study and in the future career. I also would like to acknowl edge Prof. Paul H. Holloway, Prof. Franky So, Prof. Scott Perry, Prof. Charles Cao, and Prof. Simon Phillpot for kindly serving as my Ph D committee members. I would not be able to complete this work without the help from many brilliant members in Dr. Xue in this Gator community and beyond. Dr. Ying Zheng, Dr. Williams Hammond, and Dr. Jason Myers gave me initial training in solar cell fabrication and characterization, equipment usage, and lab maintenance. Dr. Jihua Yang and Dr. Aiwei Tang were the pioneer members in hybrid solar cell project and t heir study gave me much valuable help for my latter work. The much insightful discussion and experimental help from Dr. Wei Zhao, Dr. Sang Hyun Eom, Dr. Yixing Yang, Dr. Edward Wrze sniewski, Weiran Cao, John Mudrick, Nathan Shewmon, Zhifeng Li, and Matthew Rippe are also gratefully acknowledged Besides, I also would like to a cknowledge many collaborators in the SETP project: Dr. Lei Qian synthesized the ZnO NPs, Dr. Romain Stadler a nd Dr. Dongping Xie synthesized the functional oligomers, Marc Plaisant provided the ternary nanoparticles, and Prof. Paul Holloway, Prof. John Reynolds and Prof. Kirk Schanze gave much valuable advice to complete this project. Prof. Wei You at University of North Carolina provided the polymer PBnDT. Assistance in the technical measurement s from Dr. Gerald Bourne, Dr. Kerry Siebein, Zhuo Chen and Eric Lambers is gratefully acknowledged. I also want to
5 acknowledge Dr. Song Chen, Dr. Changhua Liu, Rui Qing, and many other friends for enjoying the life at the Gator Nation. The financial support from Department of Energy, National Science Foundation, and Florida Energy System Consortium is also acknowledged. I owe the most to my family. I thank my parents my parents in law and my uncle for their love and support throughout my whole academic journey; and my brother and sister for taking care of our parents and sharing the life Finally, I want to thank my wife, Xiaochang, for love, care, and creating a new lif e together And t his dissertation is dedicated to her.
6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ .......... 10 LIST OF FIGURES ................................ ................................ ................................ ........ 11 LIST OF ABBREVIATIONS ................................ ................................ ........................... 17 ABSTRACT ................................ ................................ ................................ ................... 22 CHAPTER 1 INTRODUCTION OF PHOTOVOLTAIC TECHNOLOGY ................................ ....... 24 1.1 Introduction ................................ ................................ ................................ ....... 24 1.2 Status of Photovoltaic Technologies ................................ ................................ 25 1.3 Photovoltaic M aterials and Devices ................................ ................................ .. 27 1.3.1 c Si Photovoltaic Cells ................................ ................................ ............. 27 1.3.2 Thin Film Photovoltaic Cells ................................ ................................ .... 31 1.3.3 Third Generation Photovoltaic Cells ................................ ........................ 33 184.108.40.206 Dye sensitized solar cells ................................ ............................... 33 220.127.116.11 Organi c solar cells ................................ ................................ .......... 35 1.4 Organic inorganic Hybrid Photovoltaic Cells ................................ ..................... 37 1.5 Characterization of Photovoltaic Cells ................................ .............................. 38 1.5.1 Current Voltage Measurement and Photovoltaic Parameters ............... 38 1.5.2 Solar Spectrum ................................ ................................ ........................ 40 1.5.3 Quantum Efficiency ................................ ................................ ................. 41 1.6 Overview of This Dissertation ................................ ................................ ........... 43 2 INTRODUCTION TO COLLOIDAL NANOCRYSTALS ................................ ........... 46 2.1 Introduction ................................ ................................ ................................ ....... 46 2.2 Electronic Structure of Nanocrystals ................................ ................................ 47 2.3 Co lloidal Nanocrystals Growth ................................ ................................ .......... 49 2.3.1 Synthesis of Quantum Dots ................................ ................................ ..... 49 2.3.2 Synthesis of Anisotropic Nanocrystals ................................ ..................... 54 2.4 Surface Chemistry ................................ ................................ ............................ 5 5 2.5 Optical and Electronic Properties ................................ ................................ ...... 57 2.5.1 Optical Properties ................................ ................................ .................... 57 2.5.2 Electronic Properties ................................ ................................ ............... 60 2.6 Application in Photonic Energy Conversion Devices ................................ ......... 63 2.6.1 Quantum Dot Photovoltaic Cells ................................ .............................. 63 2.6.2 Quantum Dot Light Emitting Diodes ................................ ........................ 65
7 3 INTR ODUCTION TO ORGANIC ELECTRONIC MATERIALS AND DEVICES ...... 68 3.1 Introduction ................................ ................................ ................................ ....... 68 3.2 Electronic Structure and Properties of Organic S emiconductors ...................... 71 3.2.1 Atomic Orbital Hybridization and Bonding ................................ ............... 71 3.2.2 Excitons in Organic Solids ................................ ................................ ....... 73 18.104.22.168 Exciton types ................................ ................................ .................. 73 22.214.171.124 Exciton Properties ................................ ................................ .......... 74 126.96.36.199 Exciton Motion ................................ ................................ ............... 77 3.3 Charge Transport in Organic Solids ................................ ................................ .. 79 3.4 Organic Photovoltaic Cells ................................ ................................ ................ 81 3.4.1 Principle of Organic Photovoltaic Cells ................................ .................... 81 3.4.2 Progress of Organic Photovoltaic Cells ................................ ................... 85 188.8.131.52 Conjugated polymers ................................ ................................ ..... 86 184.108.40.206 Morphology ................................ ................................ .................... 87 220.127.116.11 Device architecture ................................ ................................ ........ 88 4 EFFEC T OF COLLOIDAL NANOCRYSTALS ON HYBRID PHOTOVOLTAIC CELLS ................................ ................................ ................................ .................... 90 4.1 Introduction ................................ ................................ ................................ ....... 90 4.2 Colloidal Nanocrystals Synthesis and Process ing ................................ ............ 91 4.2.1 Spherical Nanoparticles Synthesis ................................ .......................... 91 4.2.2 Nanorods Synthesis ................................ ................................ ................ 91 4.2.3 Nanocrystal Processing ................................ ................................ ........... 92 4.3 Hybrid Film Characterization and PV Cell Fabrication ................................ ...... 93 4.3.1 Nanocrystals and Hybrid Film Characterization ................................ ....... 93 4.3.2 Hybrid PV Cell Fabrication ................................ ................................ ...... 93 4.4 Nanocrystal Characterization and Properties ................................ .................... 94 4.4.1 Nanocrystal Size ................................ ................................ ..................... 94 4.4.2 Nanocrystal Shape and Composition ................................ ...................... 95 4.5 Effect of Device Aging on Performance ................................ ............................ 99 4.6 Effect of Nanocrystal Size on Device Performance ................................ ........ 102 4.7 Effect of Nanocrystal Sha pe on Device Performance ................................ ..... 106 4.8 Other Colloidal Nanocrystals ................................ ................................ .......... 107 4.9 Summary ................................ ................................ ................................ ........ 107 5 SOLUTION PROCESSED MULTI FUNCTIONAL ZINC OXIDE NANOPARTICLE CATHODE INTERLAYER ................................ ........................ 109 5.1 Introduction ................................ ................................ ................................ ..... 109 5.2 Synthesis and Characterization of ZnO Nanoparticles ................................ .... 110 5.3 Effect of a ZnO NP Layer on Device Performance ................................ .......... 111 5.4 Role of the Z nO NP Layer ................................ ................................ ............... 114 5.5 Effect of the ZnO NP Layer on Device Environmental Stability ....................... 118 5.6 Summary ................................ ................................ ................................ ........ 119
8 6 EXTENDING SPECTRAL RESPONSE AND ENHANCING PHOTOVOLTAGE BY CONJUGATED POLYMERS ................................ ................................ .......... 121 6.1 Introduction ................................ ................................ ................................ ..... 121 6.2 PCPDTBT:CdSe Hybrid Processing on Device Performance ......................... 124 6.2.1 Phase Separation of PCPDTBT:CdSe Hybrid Thin Films ...................... 124 6.2.2 Effect of Processing Solvents on Device Performance .......................... 127 6.2.3 Effect of Annealing Temperature on Device Performance ..................... 129 6.3 Effect of the ZnO Layer on the PCPDTBT Device Performance ..................... 131 6.3.1 Device Efficiency ................................ ................................ ................... 131 6.3.2 Role of the ZnO NP Lay er ................................ ................................ ..... 135 6.3.3 Device Stability ................................ ................................ ...................... 136 6.4 Enhancing Photovoltage Using a Deep HOMO Polymer ................................ 138 6.5 Summary ................................ ................................ ................................ ........ 140 7 ENGINEERING POLYMER N ANOCRYSTAL INTERFACE BY CHEMICAL TREATMENT ................................ ................................ ................................ ........ 142 7.1 Introductio n ................................ ................................ ................................ ..... 142 7.2 Synthesis and Processing of CdSe Nanorods ................................ ................ 143 7.3 CdSe Nanorods and Polymer:CdSe Hybrid Film Characterization ................. 144 7.3.1 TEM Measurement ................................ ................................ ................ 144 7.3.2 FTIR Measurement ................................ ................................ ................ 145 7.3.3 NMR Measu rement ................................ ................................ ............... 145 7.3.4 XPS Measurement ................................ ................................ ................ 146 7.3.5 AFM Measurement ................................ ................................ ................ 146 7 .4 Effect of EDT Treatment on Device Performance ................................ ........... 146 7.5 Effect of EDT Treatment on Nanorods and Hybrid Films ................................ 154 7.5.1 UV Vis A bsorption ................................ ................................ ................. 154 7.5.2 Hybrid Film Morphology ................................ ................................ ......... 155 7.5.3 Chemical Treatment on Surface Chemistry of CdSe Nanocrystals ....... 157 7.5.4 Charge Transport of Hybrid Films ................................ .......................... 163 7.6 Nature of the EDT Treatment on Device Performance ................................ ... 163 7.7 Surface Chemistry of CdSe Nanorods ................................ ............................ 165 7.8 Summary ................................ ................................ ................................ ........ 167 8 GRAFTING CONJUGATED OLIGOMERS TO COLLOID AL NANOCRYSTALS .. 169 8.1 Introduction ................................ ................................ ................................ ..... 169 8.2 Design and Properties of Functional Oligomers ................................ .............. 170 8.3 Oligomer grafted Nanocrystal Hybrids ................................ ............................ 173 8.4 Oligomer grafted nanocrystal Hybrid Photovoltaic Cells ................................ 175 8.5 Summary ................................ ................................ ................................ ........ 178 9 SOLUTION PROCESSED METAL OXIDE ANODE INTERLAYER ...................... 180 9.1 Introduction ................................ ................................ ................................ ..... 180
9 9.2 Synthesis and Characterization of Tungsten Oxide Precursor and Thin Film 181 9.2.1 Synthesis of Tungsten Oxide Precursor ................................ ................ 181 9.2.2 Characterization of Tungsten Oxide Thin Film ................................ ...... 182 9.3 Solution processed Tungsten Oxide Thin Film ................................ ............... 183 9.4 PV Devices with Tungsten Oxide Anode Interlayer ................................ ......... 188 9.4.1 Solution processed Polymer Solar Cells ................................ ............... 188 9.4.2 Vacuum deposited Small molecule Solar Cells ................................ ..... 192 9.5 Summary ................................ ................................ ................................ ........ 194 10 CONCLUSIONS AND FUTURE WORK ................................ ............................... 196 10.1 Conclusions ................................ ................................ ................................ .. 196 10.1.1 Organic inorganic Hybrid Materials ................................ ..................... 196 10.1.2 Organic inorganic Interface ................................ ................................ 198 10.1.3 Semiconductor Metal Interface ................................ ......................... 200 10.2 Future Work ................................ ................................ ................................ .. 2 01 10.2.1 Organic inorganic Hybrid Materials ................................ ..................... 202 10.2.1.1 Colloidal nanocrystals ................................ ................................ 202 10.2.1.2 Organic semiconductors ................................ ............................ 203 10.2.1.3 Organic inorganic hybrids ................................ .......................... 204 10.2.2 Multi junction Organic inorganic Hybrid PV Cells ................................ 204 APPENDIX: LIST OF PUBLICATIONS AND CONFERENCE PRESENTATIONS ...... 207 LIST OF REFERENCES ................................ ................................ ............................. 211 BI OGRAPHICAL SKETCH ................................ ................................ .......................... 223
10 LIST OF TABLES Table page 1 1 Spectral mismatch factor ( M) for organic based photovoltaic cells based on various active layer s. ................................ ................................ .......................... 41 5 1 Summary of photovoltaic performance under 1 sun AM 1.5 G illumination for P3HT:CdSe hybrid photovoltaic cells. ................................ .............................. 114 6 1 Summary of performance parameter of PCPDTBT:CdSe hybrid solar cells processed by different solvent mixtures and related surface roughness and domain size of the active layer. ................................ ................................ ........ 131 7 1 Summary of photovoltaic performance under 1 sun AM 1.5 G illumination for polymer:colloidal nanocrystal hybrid photovoltaic cells. ................................ .... 151 9 1 Summary of photovoltaic performance under 1 sun AM 1.5 G illu mination for P3HT:CdSe hybrid photovoltaic cells. ................................ .............................. 191
11 LIST OF FIGURES Figure page 1 1 (color) Statistics of energy consumption in the world.. ................................ ........ 25 1 2 (color) The best research efficiency of the single junction PV cells made of various semiconducting materials.. ................................ ................................ ..... 27 1 3 (color) Schematic illustration of the device structures of crystallin e Si cell and CIGS cell ................................ ................................ ................................ ............ 28 1 4 (color) Schematic illustration of device structures of dye sensitized sol ar cell using either liqui d or solid electrolyte ................................ ................................ .. 34 1 5 (color) Schematic illustration of roll to roll processes for potential third generation PV cell manufacturing.. ................................ ................................ ..... 36 1 6 Equivalent circuit and typical J V curves of a PV cell. ................................ ...... 39 1 7 (color) The standard AM 1.5 G solar spectrum and simulated illumination spectrum from Xe Arc bulb. ................................ ................................ ................ 41 1 8 (color) The typic al external quantum efficiency of a bilayer CuPc/C60 solar cell. ................................ ................................ ................................ ..................... 43 2 1 (color) The comparison of electronic en ergy states of a semiconductor at different length scales. ................................ ................................ ........................ 48 2 2 (color) Energy levels of some common inorganic semiconductors .................... 50 2 3 Colloidally synthesized C dSe nanostructures ................................ .................... 53 2 4 (color) Schematic illustration of a colloidal nanocrystal containing both an inorganic core and an organic ligand shell. ................................ ........................ 55 2 5 (color) Optical properties of colloidal CdSe nanocrystals with different sizes. .... 58 2 6 (color) Device structures of Schottky type an d depleted heterojunction quantum dot photovoltaic cells. ................................ ................................ .......... 64 2 7 (color) Qu antum dot light emitting diodes ................................ ........................... 66 3 1 (color) Organi c molecules with different structural complexities. ........................ 69 3 2 Molecular structures of some common small molecules and conjugated polymers for organic based optoelectronic devices. ................................ ........... 69
12 3 3 (color) Optoelectronic devices made using conjugated organics as active /responsive materials. ................................ ................................ ......................... 70 3 4 (color) Electronic configuration o f a carbon atom in ground and hybridized states. ................................ ................................ ................................ ................. 71 3 5 (color) Schematic illustration of three types of excitons in a solid ....................... 74 3 6 (color) Jablonski energy diagram.. ................................ ................................ ...... 76 3 7 (color) Consecutive steps for photocurrent generation from incident light in a bulk heterojunction organic based photovoltaic cells. ................................ ........ 82 3 8 TEM image of a polym er:CdSe nanocrystal hybrid film ................................ ...... 83 3 9 (color) Energy level diagram of the possible main processes of exciton diss ociat ion in an organic based PV cell ................................ ........................... 84 4 1 TEM images of CdSe NPs with different sizes ................................ ................... 95 4 2 (color) UV Vis absorption spectr a of CdSe nanoparticles with different sizes. ... 96 4 3 TEM images of composition varied semiconductor nanorods ............................ 97 4 4 (color) UV V is absorption spectra of composition tunable semiconductor nanorods. ................................ ................................ ................................ ........... 98 4 5 (color) XRD patterns of composition tunable Cd chalcogenide nanorods. ......... 99 4 6 (color) Aging effect of hybrid PV cells based on colloidal nanocrystals.. .......... 100 4 7 (color) Nanocrystal size effect ................................ ................................ .......... 103 4 8 (color) EQE of P3HT:CdSe hybrid PV devices based on different nanoparticle sizes. ................................ ................................ ............................ 104 4 9 (color) J V characteristics of electron only devices based on different CdSe NP s izes.. ................................ ................................ ................................ ......... 105 4 10 (color) Nanocrystal shape effect. ................................ ................................ ...... 106 4 11 (color) Hybrid PV devices based on CdS nanocrystals and P3HT. .................. 108 5 1 TEM image and XRD pattern of ZnO nanoparticles synthesized by wet chemistry method. ................................ ................................ ............................ 110 5 2 (color) Device structure and schema t ic energy level diagram of P3HT: CdSe hybrid PV cells with a ZnO NP layer. ................................ ................................ 111
13 5 3 (color) Typical J V characteristics of P3HT:CdSe (5 nm) hybrid PV cells without and with a ZnO NP layer und er 1 sun AM 1.5G illumination and their corresponding EQE ................................ ................................ .......................... 112 5 4 (color) Typical J V characteristics of P3HT:CdSe hybrid PV cells without and with a ZnO NP layer and using CdSe nanoparticle s (6.8 nm) under 1 sun AM 1.5G illumination. ................................ ................................ ....................... 113 5 5 (color) Typical J V characteristics of P3HT:CdSe (5 nm) hybrid PV cells without and with a ZnO NP layer under the dark condition. .............................. 115 5 6 (color) Tapping AFM morphology of P 3HT:CdSe hybrid film without and with a ZnO NP layer ................................ ................................ ................................ 116 5 7 (color) Optical intensity prof iles of P3HT:CdSe hybrid photovoltaic cells.. ........ 117 5 8 (color) EQE and the corresponding phase change of P3HT:C dSe hybrid PV cells without and with a ZnO NP layer as a function of applied bias. ................ 118 5 9 (color) Typical J V characteristics of a P3HT:CdSe hybrid PV cell tested after fabrication and storage at the air for over two months under 1 sun AM 1.5G illumination. ................................ ................................ .............................. 119 6 1 (color) Chemical structures and s chematic energy level diagram of conjugated polymers. ................................ ................................ ....................... 122 6 2 (color) UV Vis absorption spectra of c onjugated polymers and CdSe nanocrystals in dissociated states. ................................ ................................ ... 123 6 3 (color) Tapping mode AFM topological and phase images of PCPDTBT:CdSe NP hybrids processed from different solvent mixtur es.. ........ 124 6 4 TEM images of PCPDTBT:CdSe NP hybrids with different film thickness ........ 127 6 5 (color) J V characteristics and EQE o f hybrid PCPDTBT:CdSe NP PV cells with the active layers proce ssed from different solvent mixtures. ..................... 128 6 6 (color) Plots of the dependence of J sc FF, p and R rms on the processing solvent mixtures for hybrid PCPDTBT:CdSe NP PV devices. .......................... 130 6 7 (color) J V characteristics and EQE of hybrid PCPDTBT:CdSe NP PV devices with the active layers t hermally treated at different temperatures. ...... 132 6 8 (color) J V characteristics of hybrid PV cells based on 6.8 nm CdSe particles and different conjugated polymers under 1 sun AM 1.5G illumi nation. ................................ ................................ ................................ ....... 133
14 6 9 (color) Effect of a ZnO NP layer on the performance of hybrid PV cells using a low gap polymer as donor material tested under variable illumination intensity. ................................ ................................ ................................ ........... 134 6 10 (color) Effect of the ZnO NP layer on the external and internal quantum efficiencies of PCPDTBT:CdSe hybrid PV cells. ................................ ............... 135 6 11 (color) O ptical intensity profiles and the calculated short circuit current density of the PCPDTBT:CdSe hybrid PV cells ................................ ................ 137 6 12 (color) Evolution of the photovoltaic performance parameters of an unenc apsulated PCPDTBT:CdSe hybrid PV cell with a ZnO NP layer upon exposure to the ambient condition. ................................ ................................ ... 137 6 13 (color) J V characteristics under 1 sun AM 1.5G ill umination and EQE of hybrid PV d evices based on CdSe nanorods and medium gap polymers. ....... 139 6 14 (color) J V characteristic of a hybrid PV device based on CdSe nanorods and PB n DT. ................................ ................................ ................................ ....... 140 7 1 TEM images of CdSe nanorods with different aspect ratio. .............................. 144 7 2 (color) Air exposure time dependent J V characteristics of PCPDTBT:CdSe NR hybrid PV cells. ................................ ................................ ........................... 148 7 3 (color) Performance enhancement in polymer:nanocrystals hybrid photovoltaic cells upon the EDT treatment. ................................ ...................... 150 7 4 (color) Ef fect of the EDT treatment on EQE of hybrid polymer:CdSe NR PV devices. ................................ ................................ ................................ ............ 151 7 5 (color) J V characteristics and EQE of P3HT:CdSe NR PV devices ba sed on shorter nanorods ................................ ................................ ........................ 152 7 6 (color) J V characteristics and EQE of P3HT:CdSe NR PV devices with various chemical treatment. ................................ ................................ .............. 154 7 7 (color) UV Vis absorption spectra o f polymer:CdSe hybrid films. ..................... 155 7 8 (color) Surface m orphology of PCPDTBT:CdSe hybrid films. ........................... 156 7 9 TEM images of P3HT:CdSe hy brid films without and with the EDT treatment. 157 7 10 (color) FTIR spectra of CdSe nanorods upon various treatments.. ................... 159 7 11 (color) 31 P NMR spectra of ligands exchanged by pyridine and ligands exchanged by EDT from nanorods. ................................ ................................ .. 160
15 7 12 (color) XPS high resolution spectra of CdSe nanorods upon various treatment. ................................ ................................ ................................ ......... 162 7 13 (color) J V characteristics of electron only and hole only devices .................. 164 7 14 (color) Schematic illustration of surface che mistry of colloidally synthesized CdSe nanorods. ................................ ................................ ................................ 166 7 15 (color) Proposed cleavage mechanism of alkylphosphonic acid molecules from CdSe nanocrystals upon the EDT treatment. ................................ ........... 167 8 1 (color) Schematic drawing of colloidal nanocrystals grafted by conjugated oligomers through strong chemical in teraction ................................ ................. 170 8 2 (color) C hemical structures of phosphonic acid functionalized oligomers. ....... 171 8 3 (color) UV Vis absorption spectra of functional oligomers and 6 nm CdSe particles. ................................ ................................ ................................ ........... 172 8 4 (color) Schematic energy level diagram of conjugated oligomers and CdSe nanocrystals. ................................ ................................ ................................ .... 173 8 5 (color) TGA of as synthesized and oligomer grafted CdSe nanopar ticles. ....... 174 8 6 (color) Evolution of the fluorescence of OPE E and OPE A upon addition of CdSe NPs into the solution, and evolution of the peak fluorescence int ensities for the ester and acid forms of OPE, T6 and T4BTD upon incremental addition of CdSe NCs. ................................ ................................ ... 175 8 7 TEM images of as synthesized and oligomers grafted CdSe nanocrystals. ..... 176 8 8 (color) J V characteristics and EQE of hybrid PV cells based on oligomers grafted nanocrystals. ................................ ................................ ........................ 177 8 9 (color) Tapping mode AFM topological image of T4BTD A:CdSe nan oparticle hybrid. ................................ ................................ ................................ ............... 17 8 9 1 (color) TGA of ammonium tungsten oxide bulk materials. ................................ 183 9 2 (color) XPS of tungsten oxide precursor s annealed at different temperature. .. 185 9 3 (color) XRD patterns of tungsten oxide precursors annealed at different temperature. ................................ ................................ ................................ ..... 186 9 4 (color) Tapping mode AFM topological images of (a) ITO with a 3 nm WO 3 (b) ITO with a 10 nm WO 3 (c) bare ITO substrate, and (d) ITO with a 20 nm PEDOT:PSS layer. ................................ ................................ ........................... 187 9 5 (color) Tr ansmittance of various interlayers deposited upon glass substrates. 188
16 9 6 (color) J V characteristics of P3HT:PCBM solar cells using different anode interfacial layer and their correspondi ng EQE. ................................ ................. 190 9 7 (color) Effect of annealing temperature of WO 3 interlayers on the performance of P3HT:PCBM solar cells. ................................ .......................... 192 9 8 (color) Effect of a WO 3 interlayer on the bilayer CuPc/C60 solar cells. ............ 193 9 9 (color) Effect of a WO 3 interlayer on the mixed p lanar SubPc:C60/C60 solar cells ................................ ................................ ................................ ................. 194 10 1 (color) The plot of power conversion efficiency ( p ) vs year in bulk heterojunction organic inorganic hybrid photovoltaic cells.. .............................. 197 10 2 (color) J V characteristics of a tandem cell c ontaining a PCPDTBT:CdSe hybrid cell and a small molec ule cell using ZnO NPs as inter connecting layer and the corresponding subcells. ................................ .............................. 205
17 LIST OF ABBREVIATION S Ac Acetate AFM Atomic force microscopy AM Air mass AR Aspect ratio BDT Benzenedithiol BHJ Bulk heterojunction BR Bi molecular recombination BOS Balance of system CB Conduction band, chlorobenzene (chapter 3) CELIV Carrier extraction by linearly increasing voltage CF Chloroform CIGS Copper indium gallium diselenide CIS Copper indium selenide CPVs Concentrated photovoltai cs CRZ Charge recombination zone CS Charge separation CT Charge transfer C V Capacitance voltage CV Cyclic voltammetry DA Donor acceptor DAP E Secondary phosphine chalcogenides Dialkylphosphine chalcogenides o DCB o dichlorobenzene DIO 1,8 D iiodooctane DMSO Dimethyl sulfoxide
18 DNA Deoxyribonucleic acid DOS Density of states DSSC Dye sensitized solar cells EDT Ethanedithiol EFG Edge defined film fed growth EQE External quantum efficiency ETA Extremely thin semiconductor absorber FETs Field effect transisto rs FF Fill factor FWHM Full width at half maximum FTIR Fourier transform infrared spectroscopy GPC G el permeation chromatography GR Germinate recombination HTMs H ole transporting materials HOMO Highest occupied molecular orbital ICBA I ndene C60 bisadduct I QE Internal quantum efficiency ITO Indium tin oxide LCOE Levelised cost of electricity LED Light emitting diodes LUMO Lowest unoccupied molecular orbital MEG Multi exciton generation MCCs Molecular chalcogenide complexes MEH PPV Poly( 2 methoxy 5 (2' ethylh exyloxy) p phenylene vinylene ) MO Molecular orbital
19 NC Nanocrystals NIR Near infrared NMR Nuclear magnetic resonance spectroscopy NP Nanoparticles J V Current density voltage OA Oleic acid ODE 1 Octadecene ODT 1, 8 Octanedithiol OPVs Organic photovolta ic cells P3HT Poly(3 hexyl thiophene) PA Phosphonic acid PC61BM C61 butyric acid methyl ester PCE, or p Power conversion efficiency PCDTBT Poly((9 (1 octylnonyl) 9H carbazole 2,7 diyl) 2,5 thiophenediyl 2,1,3 benzothiadiazole 4,7 diyl 2,5 thiophenediyl) PCPDTBT Poly (2,1,3 benzothiadiazole 4,7 diyl( 4,4 bis(2 ethylhexyl) 4H cyclopenta(2,1 b:3,4 b') dithioph ene 2,6 diyl)) PEDOT:PSS Poly(3,4 ethylenedioxythiophene):poly(styrenesulfonate) PSCs Polymer solar cells PV P hotovoltaic PPA P P (di n tetradecyl) dihydrogen pyrophosphonic acid QDs Quantum dots SCLC Space charge limited current SEM Scanning electron m icroscopy SILAR Successive ion layer adsorption and reaction SM Small molecule
20 SSSC Solid state sensitized solar cells TBP Tributylphosphine TBP E Tributylphosphine chalcogenides TDPA Tetradecylphosphonic acid TEA Triethylamine TGA Thermogravimetric analys is TEM Transmission electron microscopy TMAH T etrame thylammonium hydroxide TMOs Transition metal oxides TMS E Bis(trimethylsilyl) chalcogenides TOF Time of flight TOP Trioctylphosphine TOPO Trioctylphosphine oxide TOPS Trioctylphosphine sulfide TOPSe Trioc tylphosphine selenide VB Valence band XRD X ray diffraction XPS X ray photoelectron spectroscopy UPS Ultraviolet photoelectron spectroscopy J sc Short circuit current density, mA/cm 2 n Electron concentration A Absorption efficiency CT Charge transfer / exciton dissociation efficiency ED Exciton diffusion efficiency EQE EQE External quantum efficiency
21 n i Intrinsic carrier concentration IQE IQE Internal quantum efficiency p PCE Power conversion efficien cy P 0 I llumination intensity p Hole concentration R rms Root mean square roughness V oc Open circuit voltage, V V CB V olume of chlorobenzene
22 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Ful fillment of the Requirements for the Degree of Doctor of Philosophy ORGANIC INORGANIC HYBRID PHOTOVOLTAIC CELLS By Renjia Zhou December 2012 Chair: Jiangeng Xue Major: Materials Science and Engineering Organic inorganic hybrid materials that can potent ially combine the low cost and processing versatility of organic materials with high electronic performance and stability of inorganic semiconductors have been extensively used as active layer in photovoltaic cells, light emitting diodes, and photodetector s This dissertation endeavors to better understand operation mechanism as well as improve the performance of organic inorganic hybrid photovoltaic cells using blends of conjugated polymers/oligomers and colloidal inorganic nanocrystals First we study th e effect of colloidal ly synthesized CdSe nanocrystals on the performanc e of hybrid photovoltaic cells. The device using p oly(3 hexylthiophene) and CdSe show s nanocrystal size dependent performance and a maximum p ower conversion efficiency ~ 2.0% that attrib ut es to improved charge transport and organic inorganic hybrid morphology as well The study has been extended to improve device performance by including a solution processed ZnO nanoparticle layer between the hybrid active layer and the cathode as a resu lt of combinational optical, electrical, and morphological effects Further enhancement in device performance has been achieved by selecting conjugated polymers with low energy gap and energy level s better aligned
23 with CdSe nanocrystals The device using a low gap polymer shows spectral response up to ~ 850 nm and a power conversion efficiency 3.5%; and the device using a polymer with deep highest occupied molecular orbitals leads to an open circuit voltage as high as 0.9 V. O rganic inorganic interface that govern s exciton dissociation (charge transfer) and charge transport in hybrid photovoltaic cells has been particularly emphasized First, we introduce chemical treatment to engineer the organic inorganic interface, which leads to a 30 90% enhancement in de vice performance an d a record high power conversion efficiency 5% in bulk heterojunction hybrid photovoltaic cells. Moreover we also design ed conjugated oligomers with functional group s to directly i nterface colloidal nanocrystals for hybrid photovoltaic cells. Finally, tungsten oxide thin film was prepared by depositing precursor solution into supporting substrates and then thermally annealed at a mild temperature. Organic photovoltaic cells using these tungsten oxide thin film s as anode interlayer have s hown performance comparable to those using traditional organic interlayer.
24 CHAPTER 1 INTRODUCTION OF P HOTOVOLTAIC TECHNOLO GY 1.1 Introduction Searching for clean, reliable, renewable, and affordable energy has become a critical state strategy for most of the countries in the 21st century. Fossil fuels including coal, oil and natural gas are non renewable, highly carbon emitted, and region limited. The proven reserves are not able to meet the demand of the rapid growing in global economy and population in t he next one century. Nuclear energy, though it is clean, could lead to a series of safety and subsequent environmental problems. Wind energy, hydropower, and geothermal energy, are clean and renewable if utilized properly, but they are location limited. Ha rvesting energy directly from the Sun is the only approach to meet the requirement of both sustainable growth in human society and maintenance of a clean and healthy environment. However, even renewable energy technology has shown significant progress in t he past several decades, fossil energy has still dominated our energy consumption style nowadays, as which possesses >85% of all consumption energy (Figure 1 1). 1 Photovoltaics or solar electricity, one of the renewable energy produced directly from the Sun, have shown rapid growth since the first demonstration of modern PV cells in silicon in the 1950s. 2 The electricity generated by photovoltaics is compatible to a range of applications including on and off grid, is not constraint to climate and geographic lo cation, and has relatively low maintenance and operation cost. A PV system contains PV cells (or modules) and auxiliary components called balance of system (BOS) including inverter, batteries, mechanical structure, etc. Nowadays, the first generation and s econd generation PV modules have already went into market, and
25 Figure 1 1. (color) Statistics of energy consumption in the world. Fossil energy (oil, coal, consumption, which is pr edicted to decrease to 75% in 2035 as the fraction of renewable energy grows. Other renewables include wind, solar cells, solar thermal, etc. Adapted from World Energy Outlook 2011, International Energy Agency. According to the data from International Rene wable Energy Agency, the thin film PV module prices had fallen below $1/watt to $0.84 0.93 /W, and the prices of the crystalline Si modules varied in the range of $1.02/W $1.24/W at the beginning of 2012. 1 Even excluding the huge price drop due to the market factor (excess supply than demand), these prices still have a gap to the cost of the Sunshot initiative created by US Department of Energy in 2010, which suggested the cost of solar modules could be reduced to $0.50/W and the levelised cost of elect ricity (LCOE) to $1/W by 2020. 1 Thus, together with the diffuse nature of the sunlight, further improving this technology is still a must to optimize the energy consumption in our only earth. 1.2 Status of Photovoltaic Technologies The modern PV technology started in crystalline Si based on pn junctions in the 1950s, and subsequently other materials such as GaAs, CdTe, Cu 2 S/CdS etc. had also
26 been developed for PV application in a fast manner. 2 The power conversion efficiencies of the PV cells based on these inorganic semiconductors had reached to or above 10% at the first twenty years (Figure 1 2), which together with the driving by the energy crisis and pursuit of renewab le energy by many countries led to the emerging, progress, and mature of the PV industry in the 1980s. PV module production was firstly realized in Si pn junction, which has been well known as first generation PV technology. Though several other PV technol ogies have been developed for module production, crystalline Si modules (c Si) still have the dominant share in PV market nowadays. The early attempt to scale up the thin film PV technologies from laboratory centimeter square to module is not successful; w hile n owadays thin film PV modules based on amorphous Si Si), CdTe, copper indium selenide (CIS) and copper indium gallium diselenide (CIGS) that are known as second generation PV technology have become the major competitor of the c Si modules. The primary ef fort in PV industry is to increase the efficiency of the module and simultaneously reduce the cost of materials and production. The efficiency of single junction solar cells is constraint to the Shokley Queisser limit derived by considering the loss in spe ctral coverage, exciton/charge recombination and black body radiation. 1 For example, Si solar cells have the Shokley Queisser limit of 32.7%. Black body radiation accounts for 7% when the device operated at room temperature, and this temperature depend ent loss is unavoidable. S pectrum loss and recombination can be engineered in terms of materials and device architecture to break this limit and maximize the efficiency. For example, staking cell in series with each subcell absorbing a specific wavelength rang e of photons is a practical approach to maximize absorption and
27 Figure 1 2. (color) The best research efficiency of the single junction PV cells made of various semiconducting materials. The data is taken from efficiency chart by NREL, and only the cham pion efficiency at that year was used. ultimately enhance efficiency. In fact, though multi junction GaAs based cells have shown very high efficiency (>30%), the high manufacture cost makes it only accessible to the application in aerospace and military in dustry. In addition, reducing exciton and charge recombination requires more optimal morphology or crystal quality of semiconductors and the better contact between the semiconductor and metal electrode. The semiconductor materials, fabrication, operation m echanism, and characterization of the PV cells will be surveyed in the following parts to evaluate the advantage and disadvantage of the different types of PV technologies. 1.3 Photovoltaic Materials and Devices 1.3.1 c Si Photovoltaic Cells Silicon, an ab undant element in the earth, plays the most important role in the modern technology revolution. Si has an indirect bandgap of 1.1 eV and can be manufactured in a very large quantity at low cost and with ultra high purity. The production of c Si solar cell s mainly includes the Si ingot/wafer production, PV cell
28 Figure 1 3. (color) Schematic illustration of the device structures of crystalline Si cell (a) and CIGS cell (b). fabrication, and module assembly. In general, c Si can be classified as single cry stalline Si (sc Si), multicrystalline Si (mc Si), and edge defined film fed growth (EFG) ribbon Si. 3 The highest cell efficiencies of sc Si and mc Si are 25% and 20% (Figure 1 2), 4 respectively, however, the efficiency of their commercial modules is typically 14 19%. Figure 1 3a shows the device structure of a crystalline Si solar cell. S lim metal grid at the illumination side forms one contact to the diode and allow s sunlight to touch the semiconductor for harvesting. An antireflective layer is typically coated to reduce the absorption loss due to reflection in Si surface. The semiconduct or diode contains an n type and a p type semiconductor, which touches each other to form a metallurgical junction or pn junction that is the basic for an inorganic solar cell. In principle, when the semiconductor active layer absorbs a solar photon with en ergy higher than its band gap, an electron in valence band (VB) will be excited to empty conduction band (CB), resulting in the generation of a weakly bounded electron
29 hole pair called exciton. The photogenerated exciton subsequently dissociates into free electron and hole in an ultrafast manner (~10 12 10 15 s), which will diffuse or drift under electrical field to the opposite collecting electrodes. Charge recombination could occur during the transport process. The collection of these photogenerated cha rge carriers leads to the photocurrent in the device and also photovoltage built up across the two electrodes. The absorption coefficient as a reflection of the photogeneration in a semiconductor with desired thickness for direct transition can be describe d by (1 1) where A is a constant, h is Plank constant, is frequency, E g is the band gap of a semiconductor. F or an indirect semiconductor, the absorption also in volves the photon absorption or emission with required momentum. Thus, the absorption coefficient can be described as (1 2) Thus, for Si solar cells, the indirect band gap requires very thick active layer (hundreds of micrometer s) to harvest most of the sunlight. 2 Recombination is an important electronic process occurred in solar cells and the very process leading to loss in photocurrent. The mechanism includes recomb ination through traps in the forbidden gap, radiative recombination, and Auger recombination. The net recombination rate through a single level trap within the forbidden gap can be expressed as (1 3)
30 w he re p, n, and n i are the concentration of hole, electron, and intrinsic carriers, respectively; and carrier lifetime is given by section, is the thermal velocity of the carriers, and N T is the concentration of the traps. Note that the recombination rate is only dependent on the minority carrier, the recombination rate for a p type semicon ductor in low injection condition can be simplified as The net recombination rate due to radiative processes is given by (1 4) where B is a constant, and for a p type semicon ductor at low injection condition, the equation can be re written as where 2 The net recombination rate due to Auger process is given by (1 5) Thus, the total recombina tion rate in a semiconductor is the sum of these individual rates. The carrier transport in conventional inorganic semiconductors includes drift and diffusion processes. The drift velocity is proportional to the electric field and related to carrier mobili ty. The driving force for diffusion is the difference in carrier concentration and in thermal equilibrium, and the diffusion coefficient (D) can be expressed by Einstein relationship (Chapter 3). In solar cells, lattice and ionized impurity scattering are the main scattering mechanism in carrier transport. By solving the minority carrier diffusion equation together with appropriate boundary conditions, the current voltage characteristics for an inorganic solar cell can be derived as
31 (1 6) w here I o1 and I o2 are the saturated dark current due to recombination in the quasi neutral region and in the space charge region, respectively. 2 1.3.2 Thin Film Photovoltaic Cells The high cost in both materials processing and device fabrication of c Si cells stimulates to develop other PV technologies. Thin film solar cells then gradually move to the commercial stage for their advantages such as low material usage (<1% thick of c Si cell), low manufacturing cost (potential wet chemistry processing), and compatible with a variety of light weight and even flexible supporting substrates. Nowadays three types of thin film Si), CdTe, and CIGS have been commercially developed with high efficiency (Figure 1 2). Here we exemplify the fabrication and principle of the CIGS cells. Figure 1 3b shows the device structure of a typical CIGS cell. The back contact Mo layer that is very sensitive to water is deposited by sputtering at high vacuum. The CIGS layer is typically deposited by multi source co evaporation and two stage processes of precursor deposition followed by selenium annealing. A thi n CdS layer is deposited by chemical bath deposition to form the CIGS CdS junction. Due to the environmental safety concern, CdS has also been replaced by Cd free compounds such as ZnS, ZnO, ZnSe, etc However, the deposition limit of these Cd free compoun d s makes CdS still the priority choice, since it can protect the active layer and the junction during the subsequent depositions. A thin ZnO window layer with high resistance (1 100 e the diode quality. I t is argued that if the CdS layer is thick enough, the ZnO layer is not necessary
32 to improve device performance. Then a doped high conductivity ZnO layer are deposited by sputtering or chemical vapor deposition, followed by deposition of Ni/Al grid to enhance charge collection. 2 Though efficiency of laboratory cells has reached to 20% by optimization in materials processing and device fabrication, the underlying operation mechanism is still not fully understood. Photogeneration occurs a t the CIGS layer and the space charge region at the CIGS layer and also CdS layer leads to the charge separation. The photogeneration loss is typically due to the surface and front metal grid reflection, the absorption in the CdS layer and TCO layer, and t he incomplete absorption in CIGS layer. The recombination loss reflects in the open circuit voltage, which is also a function of the band gap of the CIGS. 2 Recombination typically occurs at the defects and imperfect structure in the CIGS layer, while the r ecombination at the CIGS CdS junction is trivial by proper doping. In addition, the grain boundaries in the CIGS layer could also affect the transport and recombination. The CdS layer creates the type inversion at the CIGS CdS interface and is fully deplet ed. The band alignment at the CIGS in creating the type inversion. Other than using high vacuum and high temperature deposition method for the CIGS layer, recently Mitzi and co workers d eveloped a precursor approach to deposit CIGS layer through solution processing. 5 6 These solution processed cel ls maintain the conventional CIGS structure and show efficiency 10 15%, which makes this technology more promising for future photovoltaics. Due to the high cost in indium, the alternative
33 absorbers with earth abundant element such as copper zinc tin selen ide (CZTSe) have also been developed with high efficiency. 6 7 1.3.3 Third Generation Photovoltaic Cells Third gene ration photovoltaic technologies are mostly at the laboratory research stage yet grow in a fast pace recently. The third generation PV technologies generally include concentrated photovoltaics (CPVs), dye sensitized solar cells (DSSC), organic solar cells (OPVs), and other novel and emerging solar cell concepts such as quantum dot solar cells (Chapter 2.6), intermediate band cells, etc. 3 The CPVs utilize the optical devices to concentrate the sun light into a small but high efficiency multi junction cell. The optical devices need to be oriented to the su n and cooling system is necessary to reduce the performance loss due to high operation temperature. Since it focuses on the optical management, rather than introducing new materials or new concept in PV technologies, we do not intend to discuss in detail a bout CPVs. Instead, we will particularly introduce DSSC and OPVs as examples of the development of third generation PV technologies. 18.104.22.168 Dye s ensitized s olar c ells Dye sensitized solar cells (DSSC) have received significant attention as major advanceme nt made by Grtzel and co workers in the early 1990s. 8 9 In a typical manner, it consists of a transparent conductive substrate such as fluorine doped SnO 2 a photoelectrode such as mesoporous TiO x thin film, a photosensitizer such as Ru complex, a redox electrolyte such as I /I 3 redox ions, and a counter electrode with high electrocatalytic activity such as Pt (Figure 1 4). When illuminated, Ru complex photosensitizer absorbs the incident photons and is excited from a ground state to an excited state, following by the injection of the
34 Figure 1 4 (color) Schematic illustration of device structures of dye s ensitized solar cell using either liquid (left) or solid electrolyte (right). The solid state DSSC may have a thin absorber layer and a hole transporting layer, here we simplify them as solid electrolyte layer. excited electron to TiO x CB. These electrons then transport to the TCO through diffusion in TiO x and ultimately reach to the counter electrode through the external loading. The oxidized photosensitize r then accepts electron from an I ion redox mediator to regenerate the ground state, and simultaneou sly the I is oxidized to I 3 The oxidized redox mediator is re reduced to I ion by diffusing to the counter electrode. Overall, the incident photon is converted to electron with retained chemical states of the involving materials. The energy gap of the photosensitizer determines the photocurrent, and the energy offset between the Fermi level of TiO x and redox potential of the mediator determines the output voltage. The energy offsets between the photosensitizer LUMO and TiO x CB and between the redox medi ator potential and the photosensitizer HOMO should be sufficient high (> 200 mV) for efficient charge injection / electron transfer reaction. Unlike conventional inorganic solar cells and organic solar cells as well, only electron is generated in photosens itizer and directly injected to TiO x which leads to the
35 absence of charge recombination process and of electric field for charge separation. 9 Besides, charge transport occurs in TiO x phase that is different to the photogeneration sites, which leads to the very efficient charge separation. Though t he efficiency of liquid DSSC has reached to > 10% even in a submodule level, 4 the durability and safety of this type of cells is a big concern since the liquid electrolyte may cause potential corrosion and leakage DSSC utilizing solid state electrolytes (SSSC) then have received much attention (Figure 1 4). 10 Organic, inorganic, and organic inorganic hybrid semiconducting and hole transporting materials (HTMs) have been developed as solid state electrolyte and even as light absorber I nstead of using organic dye or solid state HTMs as lig ht absorber, an extremely thin semiconductor absorber (ETA, < 50 nm) is typically coated upon the TiO x The mechanism in SSSC is not fully proven yet it may share some elements of DSSC. Recently SSSCs with efficiency 10% using newly developed p type direct bandgap semiconductor CsSnI 3 as HTMs and the dye N719 as absorber or spiro OMeTAD as p type hole conductor and organic inorganic hybrid perovskite ( CH 3 NH 3 PbI 2 Cl ) as ETA have been reported, 11 12 suggesting the promising commercialization of this type of cells in the near future. 22.214.171.124 Organic s olar c ells Organic photovoltaic cells that utilize organic small mole cule or polymer to harvest solar photons have received considerate attention since the first bilayer structured cell with efficiency >1% invented by Tang in the 1980s. 13 The laboratory efficiency has reached to ~10% for organic PV cells based on either vacuum deposited small molecules or solution processed polymers (Figure 1 2). 4 The difference in operation principle compared to i norganic solar cells mainly lies in the difference in
36 Figure 1 5. (color) Schematic illustration of roll to roll processes for potential third generation PV cell manufacturing Adapted from h ttp://www.nitto.com. fundamental physical properties between o rganic and inorganic semiconductors. OPVs require a donor acceptor heterojunction for charge transfer and separation and the separated charges transport individually at the donor and acceptor components to the respective collecting electrodes. 13 14 The operation principles and research progress will be detailed in Chapter 3. Compared to PV technologies based on inorganic semiconductors, the advantages of the OPVs are self evident: organic materials are cheap, abundant, tailorable, processable, flexible, and light weight; OPVs can be manufactured using roll to roll processing in a variety of substrates including flexible substrates (Fig ure 1 5), be semi transparent, and be ultrathin (100 200 nm in organic active layer). 15 The flexible and light weight nature enables to install the OPVs in various places including roof, window, bus station, bag, or even the back of electronic devices, etc. How ever, the challenge of OPV is also very straightforward: the relatively low efficiency, particularly in module and the stability of both organic materials and devices.
37 1.4 Organic inorganic Hybrid Photovoltaic Cells The relatively low mobility and dielectr ic constant in organic semiconductors are the main reasons for the loss in photocurrent and photovoltage in organic photovoltaic cells. On the contrary, i norganic semiconductors have high electronic performance due to their high mobility and dielectric con stant and high environmental stability as well. Thus, combining the advantages of high electronic performance and stability of inorganic semiconductors with processing flexibility and high absorption coefficient of organic semiconductors is the right pursu it for the new generation of mesoscale materials in optoelectronic application. Organic inorganic hybrid photovoltaic cells, thanking to the advent of the solution processable colloidal nanocrystals, have then emerged as an alternative to all organic solar cells in the beginning of this century. 16 The first polymer nanocrystal hybrid solar cell with bulk heterojunction structure was demonstrated by Greenham et al. in 1996 with efficiency < 0.1%. 17 In 200 2 Alivisatos and co workers demonstrated polymer nanorod PV cells with p 1.7% under AM 1.5G 1 sun illumination. 16 This conceptual publication together with rapid progress in nanocrystal synthesis has then fueled much research interest in photovoltaic application of polyme r:nanocrystal hybrid materials. Though the progress is in a slow and zigzag pace, the p has reached to a level of 3 5% in this community by tailoring colloidal nanocrystals and conjugated polymers, engineering p olymer na nocrystals interface, and optimizing device architecture. 18 The operation mechanism is similar to polymer solar cells and based on the organic inorganic (polymer n anocrystal) donor acceptor junction. Conjugated polymers serve as electron donor s and hole transporting materials, and colloidal nanocrystals serve as electron acceptor s and transporter. Photogeneration occurs both in conjugated
38 polymers and colloidal nano crystals once the incident photons have energy higher than their energy gaps; and charge transfer happens at a polymer nanocrystal interface. The separated electron and hole then transport in nanocrystal phase and polymer phase, respectively, and then ar e collected by the corresponding electrodes. Nonetheless, compared to all organic solar cells, the involvement of nanocrystals creates an organic inorganic interface that has been recognized as main challenge in the advancement of this technology. This org anic inorganic interface includes the chemical, morphological, and electronic interfaces, and thus far all of these are still not well understood, which will be part icularly surveyed in this dissertation 1.5 Characterization of Photovoltaic Cells 1.5.1 Cu rrent Voltage Measurement and Photovoltaic Parameters Current voltage ( I V ) measurement is typically used to determine the diode characteristics and critical performance parameters of a PV cell. The test cell in the dark or under illumination is biased with variable voltage load and the current is sensed by an Agilent semiconductor parameter analyzer. The simulated AM 1.5 G solar illumination is provided by using a Xe arc lamp. The intensity of solar illumination is calibrated using a standard Si PV tes t cell equipped with a series of neutral density filters. A KG1 filter has been used to match the solar spectrum. The spectral mismatch between the simulated and standard solar spectrum for particular materials system has been corrected, as discussed in th e following part. As the equivalent circuit shown in Figure 1 6a, a PV cell can be modeled as an ideal diode with a parallel current source due to photocurrent ( J ph ), a parallel shunt resistance ( R sh ) due to the leakage current, and a series resistance ( R s ) due to finite conductivity of semiconductors and the contact resistance between semiconductors and
39 electrodes. 2 The series resistance and shunt resistance of the PV cells can be obtained by f itting the J V curves as shown in Figure 1 6b through the Schottky equation 19 (1 7) where J s is the saturation current density of the ideal diode, n is the ideality factor, k is T is the absolute temperature. The PV performance parameters including short circuit current density ( J sc ), open ci rcuit voltage ( V oc ) and fill factor (FF) can be extracted directly from the J V characteristic under illumination. Power conversion efficiency ( p ) that is the ratio of the maximum electrical power output ( P m,out ) to the incident optical power density ( P 0 ) can be calculated by the expression (1 8) where J sc is the short circuit current density and V oc is the open circuit voltage, both of which can be directly obtained from J V characteristics; FF is the fill factor, which is the squareness of the J V characteristics. Figure 1 6. Equivalent circuit (a) and typical J V curves (b) of a PV cell. Her e R s is the series resistance, R sh is the shunt resistance, J ph is the photocurrent.
40 For an ideal PV cell with R s = 0 and R sh the incident power intensity, namely, J sc / P 0 = constant; and the V oc can be derived with the following expression: (1 9) 1 .5. 2 Solar Spectrum The emission of radiation from the sun is similar to the black body radiation, and the light reaching the Earth surface can be approximated as parallel streams of solar photons due to the far distance between the Sun and the Earth. Since the atmosphere absorption can affect the spectral content and intensity of the solar radiation reaching the Earth surface, air mass (AM) has been introduced to define the actual spectral content and intensity. The AM 1.5 G spectrum normalized to a power density of 100 mW/cm 2 (1 sun) h as been widely used as a standard for comparing solar cell performance (Figure 1 7). Here G (global) means the measured spectrum includes the diffuse component yet D (direct) does not. In fact, the spectrum of the simulated AM 1.5G solar illumination from Xe arc lamp even after filtering is not well matched with the reference spectrum (Figure 1 7 ). The absorption coefficient and range of organic semiconductors are also vastly different from Si materials. Hence, in order to accurately determine the illuminat ion intensity, a spectral mismatch factor has been introduced and expressed as 2 (1 10) where E R is the reference spectral intensity, E S is the simulated sour ce intensity, S R is the spectral responsivity of Si cell, S T is the spectral responsivity of test PV cell that is
41 Figure 1 7 (color) The standard AM 1.5 G solar spectrum and simulated illumination spectrum from Xe Arc bulb. calculated from EQE by Each function must be integrated over the photoresponse wavelength of the test cell. By adjusting the solar simulator spectral irradiance, I R,R can be equal to I S,R Hence, the real I sc ( I R,T ) of the test c ell under the reference cell can be expr essed as = ( I R,R = I S,R ) (1 11) Though there is some difference in reference and simulated spectra, the M in organic based solar cell is close to unity for most of the material systems. The M of some material systems for organic based solar cells in this study has been listed in Table 1 1. 1 .5. 3 Quantum Efficiency Quantum efficiency measurement is another characterization technique to Table 1 1. Spectral mismatch factor ( M) for organic based photovoltaic cells based on various active layers. Active layer P3HT:PCBM P3HT:CdSe PCPDTBT:CdSe CuPc/C60 SubPc/C60 M 0.98 0.99 0.98 0.99 0.98
42 assess performance of a PV cell. The photoresponse range and intensity of the cell can be revealed through QE study. The light illuminated from halogen lamp passes through monochromator to give a low intensity light (1 2 ) with desired wavelength. The monochromatic light has been mechanically chopped at 400 Hz before shining upon a calibrated Newport 818 UV Si detector for overall intensity measurement (I 0 ) or a test device for photocurrent measurement (I d ). A S tanford Research Systems 830DSP lock in amplifier with synchronous frequency of the mechanical chopper, together with a Keithley 428 current amplifie r to reduce noise/signal ratio, has been used to monitor the photocurrent of the device. The test devices were also illuminated by a separate constant white light bias from a halogen lamp with an intensity of ~ 0.5 1 sun to create similar optical and ele ctric fields as I V measurement. Hence, the EQE can be calculated by (1 12) where h is Plank constant, q is the elementary charge, I d is the photocurrent of the test device at a given wavelength, I 0 is the intensity of the Si detector at a given wavelength, R D is the responsivity of the Si detector at a given wavelength. The short circuit current density of the test cell under reference 1 sun (= 100mW/cm 2 ) AM 1.5G illumination can also be calcu lated using EQE through the expression (1 13)
43 where is the reference AM 1.5G power intensity. The difference between the J sc obtained from J V measurement and the J sc calculated from EQE should be very small if the simula tor was calibrated correctly, typically within 5% in error (Figure 1 8). The Si detector was used to measure the intensity of light before and after passing the test device. The absorbance of the device was calculated by (1 14) w here I r ( ) is the reflective light measured by the detector Figure 1 8 (color) The typical external quantum efficiency ( EQE ) of a bilayer CuPc/C60 solar cell. The integrated J sc from EQE is 4.0 mA/cm 2 which is almost the same as the J sc (3.9 mA/cm 2 ) determin ed from J V characteristics under 1 sun AM 1.5G illumination. 1 .6 Overview of This Dissertation This dissertation focuses on photovoltaic cells using blends of organic semiconductors and co lloidal nanocrystals. We intend to provide a better understanding on fundamental physical properties of organic inorganic hybrid thin films and to demonstrate their unique feasibility in photovoltaic application. The first three chapters
44 provide the fundamental context for the co ming work present in this dissertation C hapter 4 to Chapter 9 can be generally divided into three parts: Chapter 4 to Chapter 6 focuses on the fundamental properties of organic inorganic hybrid materials and their influence on the performance of hybrid PV cells; Chapter 7 and Chapter 8 strengthe n on organic inorganic interface engineering to enhance the hybrid PV cell performance; Chapter 9 studies the preparation of metal oxides as anode interfacial layer for organic based PV cells. In part I first, C hapter 4 demonstrates the s ynthesis, processing, and properties of colloidal CdSe nanocrystals and their effect on the performance of organic inorganic hybrid PV cells. Chapter 5 intends to enhance the performance of hybrid PV cells by including a solution processed ZnO nanoparticle layer between the active layer and the cathode. Chapter 5 endeavors to enhance the photocurrent and photovoltage of hybrid PV cells in the perspective of conjugated polymers selection and processing. The part II intends to address the fun damental issue of organic inorganic interface in hybrid PV cells. Chapter 7 focuses on engineering the polymer nanocrystal interface by chemical treatment to enhance the performance of hybrid PV cells and to understand the fundamental chemical and physical mechanisms of the organic inorganic interface. Chapter 8 further attempts to engineer the organic inorganic interface by directly grafting conjugated oligomers with functional group to colloidal nanocrystals. The part III including Chapte r 9 introduces the low temperature solution processing of metal oxides as anode interfacial layer for organic based PV cells. Organic based PV cells using these solution processed metal oxides have been
45 demonstrated with comparable performance to those cel ls based on traditional organic interlayer. Finally, Chapter 10 concludes this work and provides the future research directions for organic inorganic hybrid photovoltaic cells.
46 CHAPTER 2 INTRODUCTION TO COLL OIDAL NANOCRYSTALS 2 .1 Introduction Crystalline semiconductor solids are the fun damental building blocks for mode rn electronics and optics. F urther advance s in the semiconductor industry will require the downsizing of these solids to a nanometer scale regime The fundamental properties of these solids become size dependent when one of their dimensions reaches this scale. 20 24 For example, the band gap of CdSe, a prototypical semiconducting solid, can be varied from 1.7 eV to 3.0 eV by simply tailoring its physical size yet maintaining the chemical composition. 25 26 The size dependent properties are the main stimulus for the wide spread study, both in academia and in industry, of these nanoscale solids (nanocrystals) in the past two decades. 21 23 27 Colloidal nanocrystals are referred to solution grown, organic monol ayer stabilized, electronically isolated, and nanometer scale inorganic particles. 18 27 These nanocrystals are co mpose d of tens to thousands of atoms and bridge the gap between molecules and bulk crystals in both physical dimension as well as physical/chemical properties. Compared to their bulk counterparts, nanocrystals exhibit two main characteristics: i ). the intrinsic properties of nanocrystals are subjected to the quantum confinement effect; ii ). the surface properties of nanocrystals become more and more significant as the fraction of surface atoms increased with size decreasing and with the involvement of the stabilizing organic monolayer. 21 26 Thus, this opens a great opportuni ty to tune the physical properties of the colloidal nanocrystals by simply adjusting their composition, size, shape, and surface ligands; 26 31 the size dependent properties combined with solution processability also make nanocrystals a promising
47 candidate as building blocks for a variety of optoelectronic devices and as fluorophores in biomedical application s 32 35 In particular, this dissertation will focus on the synthesis and processing of colloidal nanocrystals for photovoltaic application. In this chapter, though it is not intended to cover all the basic knowledge and technological breakthroughs, aims to provide a context for the work that will be presented concerning colloidal nanocrystals. The electronic structure and the synthesis of colloidal nanocrysta ls will be first introduced, followed by surveying the surface chemistry due to its critical importance of governing the physical properties and in a variety of application. The optical and electronic properties and the application in optical energy conversion devices will be also discussed. 2 .2 Electronic Structure of Nanocrystals Molecules are formed by chemically bonding atoms together. When two atoms approach each other to form covalent bond, according to molecular orbitals theory, the atomic orbitals interact to form bonding molecular orbitals (MO) and anti bondi ng MO. The bonding electrons will preferentially reside on the lower energy level bonding MO and leave higher energy level anti bonding MO empty (Figure 2 1), which minimizes the free energy of the new molecule. If these molecules assemble to form an infin ite solid, the significant interaction of the MOs leads to the splitting of the energy levels; accordingly, the bonding and anti bonding MOs of the molecule lead to the formation of two continuous energy levels valence band (VB) and conduction band (CB) of the solid. An energy gap or band gap (E g ) has simultaneously deve loped to separate the VB and CB. T he VB will be occ upied by the covalent electrons, and the CB will be unoccupied (Figure 2 1). For a bulk semiconductor, for example, Si, the absorption of a visible photon (energy higher than E g ) will excite a valence electron to the conduction band,
48 resulting in the formation of an electron hole pair or exciton. Due to the high dielectric in bulk inorganic semiconductors can be easily dissociated into two free polarons. Once generated, these free polarons together with their precursor excitons are highly delocalized among the crystalline solids. However, the previous continuous energy levels of bulk solids, according to quantu m mechanics, will be discret ized for nanocrystals (Figure 2 1), and the density of electronic states (DOS) and energy gap correspondingly varied as the size varies. 36 If the nanocrystal size is comparable to or less than its exciton Bohr radius (r B ) that is related to the effective electron and hole masses and die lectric coefficient, the photogenerated excitons will be spatially confined, rather than delocalized. 36 This quantum confinement will more likely lead to exciton recombination for nanocrystals as a relaxation manner for the excited electrons, rather than exciton dissociation in bulk semiconductors. This argument h as been well demonstrated in the state of the art quantum dots (QDs), which show emission (radiative recombination process with an emissive photon) yield approaching unity. 26 37 Also, the phenomenon that the emission Figure 2 1. (color) The comparison of electronic energy states of a semiconductor at different length scales
49 sizes decrease indicates that the quantum confinement becomes more and more significant and energy level of the nanocrystals is more discret iz ed. According to the particle in a box model, the band gap shift of nanocrystals can be approximately predicted b y (2 1) where h R the nanocrystal radius, e the elementary charge, the dielectric coefficient, and m e and m h the effective electron and hole masses, respectively. 38 Thus, the band gap of nanocrystals can be calculated by (2 2) Since the Coulomb interaction term (decrease as R 1 ) is relatively small compared to the quantum localization term (increase as R 2 ), thus th e band gap of the nanocrystals always i ncreases as the size decreases. 2 .3 Colloidal Nanocrystals Growth 2 .3.1 Synthesis of Quantum Dots Quantum dots refer to inorganic nanocrystals that confine excitons in all three spatial dimensions and exhibit size dep endent absorption and emission properties. 21 31 38 Quantum dots have been studied since the pioneering of nano technology and are the typical starting morphology for investigating the properties of new nanoscale materials. A number of inorganic bulk materials can be tailored with nanometer dimensions that lead to new chemical and physical properties. For example, as shown in Figure 2 2, the
50 energy level and band gap of bulk materials can be engineered to follow the scaling rule. 39 40 The formation of colloidal nanocrystals (including quantum dots) is generally governed by nucleation and growth in a system containing precursors, organic surfactants, solvents a nd organic impurities originating from the source chemicals. 27 31 41 The precursors are chemically transformed to active species upon heating to a desired temperature, which leads to nucleation after sufficient accumulatio n of these active species. F urther growth is controlled by both the abundance of the acti ve species and the organic surfactants. Most of the precursors are metal organic complexes that can be decomposed at 50 350 C 42 43 and the surface energy of solids in nanoscale could be significantly lower than in the bulk, it is feasible to colloidally generate nanocrystals at relatively low temperature at which the organic surfactants and solvents are stable 27 44 Thermal decomposition was the first strategy for colloidal synthesis of nanocrystals. 27 31 41 42 45 To synthesize CdE (E = S, Se, and Te) nanocrystals, cadmium alkyls (dimethylcadmium) and bis(trimethylsilyl)chalcogenides were used as Figure 2 2. (color) Energy level s of some common inorga nic semiconduc tors. Due to the low band gaps of Pb chalcogenides, their energy levels shown here correspond to the quantum dots with size 2 3 nm, which is also the dimension of interest for optoelectronic application
51 Cd and chalcogen sources, respectively; and alkylpho sphine oxides (trioctylphosphine oxide, technical TOPO) were selected as both solvent and surface capping ligand. 31 In a similar manner, other metal alkyls such as diethylzinc and dibenzylmercury have been correspondingly chosen to produce other II VI nanocrystals. I t is worthwhile to mention that these metal alkyls are not decomposed into reactive specie directly, but rather, reacted with a trace of alkyl phosphonic acid (PA, from technical TOPO) to form metal PA complex as metal source precursors. 28 Inspired by this observation, Peng et al. employed CdO together with phosphonic acids as starting chemicals to replace those toxic pyrophoric, hygro scopic, expensive, and unstable metal alkyls for producing high quality CdE nanocrystals. 42 Since then other metal oxides and metal acetates (Ac) such as PbO, MnO, ZnO, CdCl 2 MnCl 2 PbCl 2 ZnCl 2 and Zn(Ac) 2 have also been gradually developed for nanocrystal synthesis. 41 44 46 Bis(trimethylsilyl) chalcogenides (TMS E) and trialkylphosphine chalc ogeneides such as trio ctylphosphine selenide (TOPSe) and tributylphosphine selenide (TBPSe) are the typical ch oices as anion precursors. 31 The recent work by Evans et al. revealed that the secondary phosphine chalcogenides (DAP E, dialkylphosphine chalcogenides), rather than tertiary phosphine chalcogenides (TBP E and TOP E) are the reactive species for the nuclei formation. 47 The secondary phosphines exist as an impurity of the commercial TBP and TOP, which have been identi fied as vehicles for delivering chalcogenides (such as selenium) to the secondary phosphines during the reaction. This observation also paves a way to increase the nanocrystal yield by intentional ly adding a required amount of secondary phosphine into tert iary phosphine fo r preparing anionic precursors. In addition, since the reactivity of the anionic precursors such as
52 TMS E, DAP E TOP E, and TBP E varies due to the difference in the P=E bond strength, synthesis of ternary or quaternary nanocrystals with u niform chemical reactivity. 48 Apart from the cationic and anionic components, organic ligands are another essential chemical compo nent of colloidal nanocrystals. These organic ligands initially serve to stabilize the nanocrystals and may form a covalent, ionic, or dative bonding with the inorganic component. Carboxylic acids, alkylphosphosphate, alkylphosphine oxide, alkylthiols, and alkyla mines are typical surface capping ligands. 27 31 During crystal growth, these ligands play mainly two roles: bef ore nucleation, as mentioned, reacting with metal precursors to form reactive and decomposable precursors; after nanocrystal surface. For example, the alkylphosphonic acid (PA), a highly populated ligand on CdSe nanorods, first reacts with CdO to form Cd PA complex as a cationic precursor ; and then during nanocrystal growth specifically bind s to some crystal planes that suppress their further growth and direct the nanocryst al growth on other exposed planes. 28 It should be mentioned tha t during the nanocrystal growth there is a dynamic equilibrium between ligand adsorpt ion and ligand desorption that is highly dependent on many factors such as nanoparticle ligand binding strength, ligand solvent interaction, steric hindrance, temperature, etc. 29 The termination of nanocrystal growth has been realized by rapidly cooling the reaction mixture below a temperature threshold or diluting the monomer concentration. However, in many cases, Ostwald ripening could occur once the monomer becomes
53 depleted or its concentration decreases to a threshold, leading to the dissolution of the smaller nanocrystals and the further growth of larger nanocrystals. 2 7 The control of monomer concentration, reaction temperature, and reaction time are rather empi rical at this moment, and vary from one system to another. In spite of that, the size of nanocrystal can still be facilely controlled for almost all developed nanocrystals under the concept of size distribution focusing growth. 27 H igh monomer concentration will lead to th e growth of smaller particles with a faster rate than that of larger ones, and consequently result in focusing of nanoparticles into monodispersity. 30 43 Though the coordinating solvents such as TOPO have exhibited great successes in the synthesis of CdSe nanocrystals, many studies have found that it is difficult to be extended to other nanocrystal systems with h igh quality. 49 Non coordinat ing solvents such as 1 octadecene (ODE) then have been utilized to address this issue. This synthesis strategy can be realized by utilizing elemental anionic precursors. The key advantage between coordinating and non coordinating approaches is that the mon omer reactivity can be tuned by controlling the ligand concentration in the non coordinating solvent. 49 Thus, the nucleation and growth can be better controlled for h igh quality nanocrystal synthese s such as CdS, PbS, PbSe, etc. 50 52 Figure 2 3 Colloidally synthesized CdSe nanostructures. Adapted with permission from Ref. 26 Copyright 2010 American Chemical Society.
54 2 .3.2 Synthesis of Anisotropic Nanocrystals The morphology of nanocrystal s can be further manipulated by controlling the monomer concentration (Figure 2 3) 26 28 30 The fast nucleati on rate and then relatively low monomer concentration will result in the growth of quasi spherical nanocrystals (quantum dots), since the growth rate for every crystal facet is slow and relatively the same due to low monomer concentration. However, anisotr opic nanocrystal growth requires much higher monomer concentration with a kinetic growth mechanism. Since the crystal growth rate exponentially depends on the surface energy of crystal facets; at a high monomer concentration and in a kinetic growth regime, the facets with higher energy will grow much faster than those with lower energy, resulting in the anisotropic crystals such as rods, tetrapods, disks, etc. 28 53 54 In term of CdSe nanocrystals, the high incoming flux of monomers will diffuse to the nuclei sphere that will be mainly consumed by the c axis of the CdSe nanocrystals. The selective adhesion of surfactants on crystal facets is also an important approach to control the anisotropic growth of nanocrystals. The mechanism is that the surface energy of the certain facets will be lowered when surfactants sel ectively adhere to them and then the growth rate of these facets will be decreased or even prohibited. Alkylphosphonic acids, which usually exist as impurities in technical grade TOPO, have firstly and so far largely been chosen as surfactants for growing anisotropic CdSe nanocrystals. 28 These PA molecules preferentially bind to the Cd rich (01 0) and (11 0) facets, which are o ften the side facets of nanorods. 55 The strength for selective adhesion of PA molecules is also related to the length of alkyl chain, which could direct the growth of anisotropic nanostructures.
55 2 .4 Surface Chemistry Nanocrystals have a very high surface to volume atom ratio, over 50% of the atoms are locate d on the surface for a 2 nm quantum dot. 41 Such a high surface to volume atom ratio leads to a strong domination of the properties of colloidal nanocrystals and also a significant influence on the application. In general, a weak coord inating ligand (or non coordinating ligand) and a strong coordinating ligand will both be used during synthesis. 42 The weak or non coordinating ligand typically serves as solvent and controls the reactivity of the monomer; while the strong coordinating ligand reacts with metal precursors to form metal ligand complex and participates in nanocrystal growth by selecti vely binding to crystal planes. The surface ligands in nanocrystal surface can be generally classified as L type ligands and X type ligands. 18 56 58 The L type ligands (weak or non coordinating ligands) are neutral molecules binding to nanocrystal surface by donating a lone pair of electrons an d forming a dative bond. TOPO, TOP, TOPSe, and alkylamine are typical L type Figure 2 4. (color) Schematic illustration of a colloidal nanocrystal containing both an inorganic core and an organic ligand shell. A layer of cation is also absorbed to the i norganic core surface. The organic shell layer contains both neutral L type ligands and charged X type ligands.
56 ligands. T he X type ligands (strong coordinating ligands) are negatively charged molecules that could ionically or covalently bond to the cation rich nanocrystal surface. The typical X type ligands include alkylphosphonic acids such as TDPA and alkylcarboxylic acid such as oleic acid and stearic acid. The qualitative and quantitative identifications of these surface ligands can be performed by Fou rier transform infrared spectroscopy (FTIR), nuclear magnetic resonance spectroscopy (NMR), X ray photoelectron spectroscopy (XPS), etc. 58 Due to the dynam ic equi librium nature of nanocrystal ligand interaction, the surface chemistry of colloidal nanocrystals can be readily tailored through ligand exchange and removal. The exchange or removal of ligands from the nanocrystal surface is related to the bindin g strength between nanocrystal and ligands, the steric hindrance of the ligands, and the ligand solvent interaction. 18 Post synthes e s purification (dispersing in good solvent and precipitating with poor solvent) is usually emp loyed to remove the unbound ligands or dissociated li gands in the incubated solution. I t also liberate s some weakly bonded ligands from the nanocrystal surface either in the form of free ligands or metal ligand complex es Ligand exchange with short chain molecules such as butylamine or pyridine is the other typical protocol to remove the L type ligands. 59 Nonetheless, the ionic or cov alent bonding nature of the X type ligands makes them much more difficult to remove by standard purification methods or exchanged with amine based molecules. The alternative approach to repel these X type ligands is by using small molecule s with strong nuc leophilic group s such as EDT, etc. 60 Finally, when exposed to the air, nanocrystal surface is also very vulnerable to
57 oxid ization, resulting in liberation of metal ligand complex es or pure ligands and simultaneous formation of insulating metal oxides. 2 .5 Optical and Electronic Properties 2 .5.1 Optical Properties Unlike their bulk crystal and small molecule counterpart s colloidal nanocrystals exhibit unique optical properties governed by the chemical composition, size, structure, and surface properties 20 25 38 44 A bsorption of a photon with energy higher than the band gap energy leads to the excitation of a valence electron to the CB. This electronic transition is subjected to the selection rule: the wave vector K should be conserved before and afte r transition. The K is automatically conserved for the direct semiconductor for this transition, while a phonon is required for the indirect semiconductor. Note that the K values have also been quantized f or the finite nanocrystal size. The optical absorpt ion of colloidal nanocrystals can be readily tailored by varying size without changing chemical composition. CdSe nanocrystals, for example, exhibits absorption across the entire visible range by varying the size from 7 nm to 4 nm (Figure 2 5 ). 25 61 Due to the discrete energy levels of the quantum dots, the absorption edges corresponding to the heavy hole, light hol e and spit orbital transitions for bulk CdSe disappeared; instead, a number of absorption features appeared (Figure 2 1 ) which are due to the nS h 1S e and nP h 1P e transitions. 62 63 For instance, the first two absorption edges for 5 nm CdSe nanocrystals can be assigned to the 1S 2/3 1S e and 2S 3/2 1S e transitions, the third to the 1P 3/2 1P e transition, and ot hers to the higher energy level transitions (Figure 2 5 ). In addition the quantum confinement effect which leads to the drastic overlapping of wave functions of the charge carriers also results in
58 a significant increase in the absorption coefficient (or oscillator strength) as the nanocrystal size decreases. 38 In a practical viewpoint, such high absorption coefficient makes nanocrystals potential candidates as building block for photovoltaic devices with ultra thin light absorbing layer s 64 Multi exciton generation (MEG) or carrier mult iplication is a phenomenon that generates two or more excitons by absorbing one photon with energy at least twice the band gap of the nano crystal 65 67 The excite d electron may release its excess kinetic energy to excite another electron from HOMO level in a collision like Auger process, resulting in generation of a bi exciton or multi exciton and an internal quantum efficiency (IQE) over 100%. This phenomenon has been observed in some nanocrystal systems such as PbS and PbSe, with over 100% external quantum efficiency (EQE) being observed recently for a PbSe quantum dots solar cell. 68 Figure 2 5 (color) Optical properties of colloidal CdSe nanocrystals with different sizes. The inset indicates the absorption and emission processes of colloidal nanocrystals interacting with incident photons.
59 Emission, reverse process of absorption, is also dependent on nanocrystal size and surface property. As shown in Figure 2 5 the excited electron in the higher energy level could relax to the lowest energy level and then recombine with the hole in the ground state; the excess energy f or this recombination will be released either as heat for non radiative recombination or as a photon for radiative recombination. The electron relaxation process i s well demonstrated by the fact that the emission spectrum is very narrow (typical full width with hal f maximum, FWHM, ~20 30 nm) and becomes longer compared to the absorption spectrum (Figure 2 5 ). The recombination process is rather complicated and varies in different categories of nanocrystals and surface properties. Some nanocrystals can captu re one type of charge carrier in the shallow traps at the nanocrystal surface and the recombination process is achieved by capture of the other oppositely charged carrier; whereas most nanocrystals that have much higher energy level s confine electron and h ol e within the finite nanocrystal thus the emission process happens when the wave function s of electron and hole overlap to a desired degree. 26 62 The size dependent emission is typically r eflected in the emission spectrum at the band edge; while in some cases, t he other emission bands appear at higher energy levels resulting from intraband recombination due t o the surface defects. Surface passivation with inorganic shells has been introduced to reduce surface defects or traps that could potentially damage the quantum yield of quantum dots. Apart from this, an inorganic shell with wider energy gap also further confines the exciton within the nanocrystal thus increasing the possibility of radiative recombination. This concept has been demonstrated in a variety of nanocrystal systems through successive ion layer adsorption and reaction (SILAR) 69 a s well as non epitaxial growth 70 The shell
60 layer can be deposited layer by layer and the overall nanocrystals maintain monodispersity. Type II quantum dots with staggered VBs and CBs between the core and shell materials, firstly demonstrated by Bawendi and co wor kers, provide another approach to engineer the w ave functions between electron and holes. 71 For example, CdTe CdSe core shell quantum dots having a VB maximum in CdTe while a CB minimum in CdSe could separately confine the hole in the VB of CdTe and electron in the CB of CdSe. C arrier recombination occurs across the interface between the core and the shell. Recently Nie and co worker demonstrated by tuning the lattice strain between core an d shell materials with type II band alignment, the core shell nanocrystals not only preserve high photoluminescence yield but also can tune the emission spectrum. 72 Hence, the emissive wavelength is feasible to be tuned by varying the thicknesses of the core and shell layers, and even emissive photon energy lower than the lowest bulk bandgap of core and shell materials can be achieved. 2 .5.2 Electronic Properties The electronic properties of colloidal nanocrystals are t o a large extent, molecule like and dominated by the surface properties. The charge transport properties (carrier mobility and type) of colloidal nanocrystals in solid thin film state can be measured by time of flight (TOF), carrier extraction by linearly increasing voltage (CELIV), field effect transistors (FETs), and space charge limited current (SCLC), etc. 44 Though there is some deviation between different meth ods, in general, it follows the similar order of magnitude and trends. The carrier density and lifetime of QD thin film s can be determined by cap acitance voltage (C V) analysis and transient photovoltaic measurement respectively
61 Colloidal nanocrystal s ha ve been regarded as electronically atom s because the organic monolayers physically isolate the nanocrystals and separate electronic interaction with the adjacent counterparts. 36 An effective approach to accomplish the electronic exchange between nanocrystals is by replacing the long chain or ganic monolayers with short chain small molecules. Exchanging the TOPO capped CdSe nanocrystals with pyridine, for example, leads to two to three orders of magnitude increase in electron mobility, reaching 10 4 10 6 cm 2 V 1 s 1 depending on the nanocrys tal size. 73 Treating nanocrystals with bidentate molecules such as ethanedithiol (EDT) or benzenedithiol (BDT) further shortens the contact distance between nanocrystals and lead s to the enhancement in carrier mobility to ~10 4 10 6 cm 2 V 1 s 1 for the demonstrated nanocrystals such as CdSe, PbS, PbSe, etc. 74 75 Talapin e t al. treated the PbSe nanoparticle thin films with hydrazine to further decrease the interparticle spacing and increase the electronic coupling. 76 They found that the electron and hole mobilities of the PbSe nanoparticle thin film s can be as high as 0.9 and 0.2 cm 2 V 1 s 1 respectively; and the transport of the carrier type can be switched by varying the treating conditions. Furthermore, passivating with inorganic molecular chalcogenide complexes (MCCs) brings the transport property of colloidal nanocrystals to a level similar to amorphous Si, with e ~16 cm 2 V 1 s 1 obtained for C dSe quantum dots passivated with In 2 Se 4 2 77 78 The charge transport among colloidal nanocrystals has been considered mainly through hopping and tunneling. The hopping behavior can be described by the early models proposed by Mott, Efros and Shklovskii. 79 82 The Mott h opping model primarily describes the conduction in a disordered system, and illustrates that the hopping
62 ij ) between charge localized site (E i ) and a nearby site (E j ) and the spatial distan ce (r ij ). When the hopping between the nearest neighboring sites dominates at low temperature, deviated from Arrhenius behavior, the temperature dependent conductivity can be expressed as (2 3) w here is the localization length and the dielectric c onstant. F or depicting the hopping conductivity other than the nearest sites, the express ion can be changed to T 1/4 dependence. The Efros Shklovskii hopping model assumes that the quantum localization len gth is much smaller than the spatial distance between hopping sites, thus the temperature dependence of conductivity is given by (2 4) I ij is equal to or smaller than the the Efros Shklovskii mechanism can be changed to the Mott model. This Coulomb gap is around twice the energy required to remove one charge from the particle, and can be approximated as e 2 for a quantum dot with a radius r The coupling energy be tween the two neighboring particles that determines the tunneling probability can be approximated as (2 5) w here h is the tunneling rate, m* the effective mas s of the charge carrier s and are the height and width of the tunneling barrier, respectively. Apparently, the coupling energy and tunneling rate increase
63 several systems of colloidal QDs that dec reasing the interparticle distance sharply increases the electrical conductivity. As mentioned, exchanging long alkyl chain molecule on the nanocrystal surface with short chain molecule s or small molecule s is the typical means to decrease the interparticle the electrical transport. 2 .6 Application in Photonic Energy Conversion Devices 2 .6.1 Quantum Dot Photovoltaic Cells Q uantum dot photovoltaic (PV) cells have rapidly emerged as a potential new technology recently. T he advantages inherent from colloidal quantum dots for PV application include the compatibility with a variety of low cost wet chemistry processing techniques and flexible substrates, readily customized absorption to match the solar spectrum, potential ly e fficient MEG, etc. PbS and PbSe have large bulk Bohr radii that offers great opportunity to tune the absorption c overage by varying their sizes. The QD PV cells were pioneered based on PbS and PbSe QDs as light harvesting materials and with a structure of QDs sandwiched by two electrodes, as shown in Figure 2 6 a 64 65 83 The QDs were built up layer by layer with a thickness of several hundred nanometer s to 1 micrometer upon a transp arent ITO electrode that forms O hmic contact with the QDs for Schottky solar cells; and each QD layer was treated with EDT or BDT to remove the bulky organic ligands. A low work function metal electrode (Al, Ag, Ca, etc.) was then vacuum deposited on the top of QDs to complete device fabrication. 64 Upon illumination, the p type light absorbing QD layer generates excitons and subsequently dissociates into free electron and hole under the assistance of buil t in potential. The p type QDs have a Fermi level close to that of the transparent electrode
64 Figure 2 6 (color) Device struct ures of Schottky type (a) and depleted heterojunction (b) quantum dot photovoltaic cell s Adapted with permission from American Chemical Society. and band bending is not necessary for the extraction of hole from QD layer; whi le the low work function metal has > 0.5 V difference in work function that produces a band bending in the QD metal interface. Such band bending favors the electron extraction and also provides an energetic barrier for hole transport. The asymmetric electrodes provide a built in potential that could sweep out the charge carriers in the entire depletion regions. In order to fully extract the photogenerated cha rge carriers, high electron and hole mobilities are essential, particularly for the minority carrier. For p type QDs, the depletion region is typically ranging from 100 150 nm, thus the mino rity carrier electrons generated in the neutral region and close to the ITO side need to overcome the long traveling distance (nearly the whole QD layer thickness) to reach the metal electrode. Despite high EQE that has been achieved for QD PV cells in the visible region, the EQE in the near infrared region is still lo w (< 20%), which is mainly due to 4 cm 1 ) in this region. 64 Increasing the QD layer thickness to ~ 1m could lead to > 90% absorption in the near infrared while the charge carrier (minority carrier) is not likely to escape such a thick film before recombination
65 occurs To address this absorpt ion extraction compromise requires higher carrier mobility, better design of device architecture, and more efficient removal/passivation of QD defects. 84 To date, by tailoring the QD size and composition and engineering device architec ture, p of 3 5% has been reported for these Schottky type QD PV cells by several independent groups. 64 68 74 75 84 85 Sargent and co workers recently introduced depleted heterojunction concept for QD PV cells where QDs were built up on the top of TiO x pre coated upon a fluorine doped tin oxide (FTO) substrate (Figure 2 6b) 86 The s e depleted heterojunction PV cells have the advantages that the minority carrier s can separate more efficiently since charge generation occurs at the illuminated electron accepting TiO x side. Together with improved atomic passivation of colloidal QDs, p ~6% has been achieved for depleted heterojunction QD PV cells recently. 87 Moreover, MEG has also been realized in PbSe QD solar cells with dual EDT and hydrazin e treatment during QD layer building up, resulting in significant contribution to photocurrent generation with peak EQE ~114%. 68 However, in order to transform this technology to real commercialization, further improvement in p is a must, which again relies on the better understanding of both device physics and chemistry/material sci ence of colloidal quantum dots. 2 .6. 2 Quantum Dot Light Emitting Diodes Solution processed electroluminescent quantum dot light emitting diodes (Q D LEDs) with multi layer structure are another example of optoelectronic application of colloidal nanocrystals (Figure 2 7) The high color purity and photophysical stability of quantum dots together with their compatib ility with roll to roll processing ar e the potential advantages to expl ore this technology. QD LEDs have a typical structure of transparent electrode (ITO) / hole transporting layer (HTL) / QD emissive layer / electron
66 Figure 2 7 (color) Quantum dot light emitting diodes. (a) A typical qu antum dot LED with light extraction at the hole injection side; (b) pictures of red, green, and transporting layer (ETL) / reflective cathode (Al) (Figure 2 7) The basic operation mechanism is the injected electron and hole from corres ponding electrodes recombine at the QD layer with photon emission. The efficiency of a QD LED is directly related to the PL quantum efficiency of quantum dots. The organic monolayer protected quantum dots are vulnerabl e to external stimuli that lead to the instability of QDs and thus the unreliability of quantum efficiency. Coating with an inorganic shell layer upon the quantum dots is the typical approach to engineer QDs. The type I core shell structur e such as CdSe/CdS and CdSe/ZnS not only enhances the physical stability but also increases the radiative recombination efficiency by confining charge carriers within the core. Though they are more prone to non radiative recombination due to the separate l ocation of electron and hole, the type II core shell QDs such as CdTe/CdSe have also been intensively studied to tune the emission wavelength from visible to near infrared regime. The high efficiency QD LEDs are typically fabricated using core shell QDs nowadays.
67 The HTL and ETL have been introduced to balance the hole and electron injection and increase the LED efficiency. The typical HTL materials include bis(4 butylphenyl) bis(phenyl)benzidine) (PolyTPD), PEDOT:PSS, poly( 9,9 dioctyl fluorene co N (4 butylphenyl) diphenylamine) (TFB) NiO, MoO 3 WO 3 etc.; and Alq 3 ZnO, TiO x etc. have been usually used as ETL materials in QD LEDs. The energy level, mobility, carrier concentrations (inorganic), and processing conditions of HTL and ETL materials are the parameters affecting the LED performance. Also blending or interfacing QDs with hole transporting polymers as hole transporting and emissive layer has also been investigated. E xploring HTL s with high work function to directl y inject hole s into QDs is still a challenge. In addition, though QD LEDs with a solution processed ZnO or TiO x layer ha ve shown much improvement in efficiency and lifetime, the accompanied Auger process also leads to loss of a large amount of charge carri ers. QD LEDs have witnessed great progress in the past decade, with EQE ~1% for blue, ~10% for green, and ~20% for red at brightness ~100 1000 cd/m 2 achieved. Further advance as a commercial interest requires not only the increase in device performance b y finely tailoring quantum dots and HTL and ETL materials and optimizing device architecture but also the development of economical processing in current QD LEDs.
68 CHAPTER 3 INTRODUCTION TO ORGA NIC ELECTRONIC MATERIALS AND DEVICE S 3 .1 Introduction This ch apter introduces a class of molecular solids that hold together through weak van der Waals interaction. These molecular solids are typical organic materials primarily composed of carbon and hydrogen atoms. The atoms that form the molecules are bound ed toge ther through strong covalent interaction, while the intermolecular interaction is mainly van der Waals interaction (molecular interaction) The fact that the organic solids have strong intramolecular interaction but weak intermolecular interaction result s in drastic difference in the optical electronic and mechanical properties 88 compared to the inorganic counterparts and also offer great opportunity to broaden their application to a wide range through economical wet chemistry processing methods. In general, organic materials can be classified as discreted small molecules, polymers, and biological molecules (Figure 3 1 a) The polymers and biological molecules are macromolecules with molecular weight in the range of 10 000 g/mol to several millions Da. P olymers have molecular weight and structure polydispersities and strictly speaking, are not chemically pure; biological molecules have the most complex structures but are chemically exclusive (Figure 3 1a). This dissertation focuses on organic semiconducting materials that have electrical conductivity in between metals and insulators Both small molecules and natural/ synth etic conjugated polymers can be organic semiconductors; though some biological molecules such as deoxyribonucleic acid (DNA) have ever been studied in optoelectronic devices, these materials are neither classified as the organic semiconductors nor the stud y of interest here.
69 Figure 3 1. (color) Organic molecules with different structural complexities. The weak intermolecular interaction and good affinity with organic solvents make organic materials dissolvable in solution that makes them compatible with s olution processing methods for thin film deposition. Furthermore, the weak intermolecular interaction also results in good mechanical ductility of organic materials that can be flexibly shaped to follow the supporting substrates. In addition, the modern sy nthetic chemistry enables the development of thousands of new small molecules and polymers Figure 3 2. Molecular structures of some common small molecules and conjugated polymers for organic based optoelectronic devices.
70 with customized properties every day, as shown in Figure 3 2. Hence, all of these intrinsic and extrinsic properties offer great potential to manufacture large area, cheap and flexible organic optoelectronic dev ices. In fact, organic semiconductors have fueled much interest both in acade mia and in industry since the pioneer work by Tang who first demonstrated a simple bilayer heterojunction organic solar cell with p exceeding 1% in 1986 13 and successively a bilayer heterojunction organic LED exhibiting 1% quantum efficiency in 1987 together with VanSlyke. 89 After that, Friend and co workers demonstrated the first LED made using conjugated polymers in 1990. 90 Today, organic LED (OLED) has already went into market as displays in cell phone, jumbo size TV, etc.; and small area state of the art organic photovoltaic cells (OPVs) fabricated in laboratory exhibit p ~10%, 4 approaching the commercial needs (Figure 3 3). Certainly, the application of organic semiconductors is not limited to OLED and OPVs, many other devices s uch as field effect transistors (FETs), 91 photodetectors, 14 and sensors have also been developed using both small molecules and polymers (Figure 3 3). The study of organic semiconductors can be traced back to 1960s and a vast amount of documentary literatures on this field ha ve been published from then on. 88 92 Figure 3 3. (color) Optoelectronic devices made using conjugated organics as active /responsive materials.
71 Figure 3 4. (color) Electronic configuration of a carbon atom in ground and hybridized states. Thus, this chapter will only go through some basic knowledge of organic electronic mat erials and devices that provides as a context for the coming work. In particular, photovoltaic cells based on organics and organic inorganic hybrids will be strengthened, which is the theme of this dissertation 3 .2 Electronic Structure and Properties of Organic Semiconductors 3 2. 1 Atomic Orbital Hybridization and Bonding The bonding nature of carbon atom governs the properties of the organics. The electronic configuration of the c arbon atom at the ground state is 1s 2 2s 2 2p 2 There are four valence electrons in the outer electronic level, with two electrons paired in the 2s orbital and the other two unpaired in the 2p orbital. The outer 2s and three 2p orbitals may mix to create a se t of equivalent degenerate orbitals and relocate the four electrons unpaired. This orbital mixing can be occurred to 2s orbital with one, two, or all three 2p orbitals to create sp, sp 2 or sp 3 hybridization, respectively (Figure 3 4). Before hybridization only the unpaired two electrons in the p orbitals could share with other atoms to form covalent bonds; and hybridization makes all the valence electrons available for bonding. Though a slight increase in the total energy of the carbon atom is
72 required af ter hybridization, this energy increase can be compensated by the energy released during bond formation. As mentioned earlier in chapter 2 when two atoms approach each other, the atomic orbitals interact to form bonding MOs and anti bonding MO s. The bonds formed between s, p, and sp n bond, and the bonds formed by overlapping two p z bond. Benzene, for example, has 2s and two 2p orbitals hybridized, resulting in three sp 2 hybridization orbitals and one unaltered 2p z orbital. The bond in benze ne is perpendicular to the plane containing all the carbon atoms and creates a delocalized electron density above and below the carbon atom plane; on the contrary, the sp 2 orbitals generates localized electron density in the ring plane and between carbon a bond) bond) between the carbon atoms in benzene is the basis of the conjugated organics that exhibit favorable electronic property. bond is st bond for the larger orbital overlapping. bon Despite the bonding among atoms within an organic molecule is covalent (E cov =2 4 e V), the intermolecular bonding is typically weak van der Waals interaction (E vdW =10 3 10 2 eV). Thus, the electronic structure and property of organic solids are, to a large extent, molecule like Nonetheless, since the energy of the van der Waals interact ion varies with distance (r) given as following: (3 1)
73 the molecular orientation and packing could significantly influence the electronic properties in solid state. When the molecular orbitals c onsiderably overlap as two molecules approach each other, the repulsive component of dispersion force becomes significant. This repulsion is also distance dependent; and the potential energy rising from both this contribution and intermolecular attraction is typically described by Lennard Jones potential: 93 (3 2) w This weak intermo lecular interaction endows organic molecules/solids with unique properties. Apart from good affinity with organic solvents and mechanical ductility, organic semiconductors typically have much lower charge carrier mobility (10 6 1 cm 2 V 1 s 1 ) than inorga nic counterparts. 88 Furthermore, excitons generated in organic solids are strongly bounded in the nearby molecules, rather than rapidly dissociated and deloc alized within the whole solid. 3 .2.2 Excitons in Organic Solids 3 .2.2.1 Exciton types Similar to the inorganic semiconductors, the interaction of organic solids with light leads to promoting an electron from a HOMO to an empty LUMO and simultane ously leaving a positive hole in the HOMO. The photogenerated electron and hole are not free particles in the respective energy level, but bounded as a neutral electron hole pair or called exciton. 88 The excitons can be generally classified to three types based on the difference in binding energy to the original molecule: Frenkel, charge tr ansfer, and Wannier Mott (Figure 3 4). 88 As shown in Figure 3 5, Wannier Mott exciton is usually
74 present in inorgani c semiconductors for the high dielectric constant and strong atomic interaction. Its Coulombic binding energy is a few meV and Bohr radius on the order of 10 nm. Due to the low dielectric constant and weak intermolecular interaction, the exciton generated in organic solids is typically Frenkel. The facts that the binding energy can be as high as 0.1 1 eV 94 95 but the Bohr radius is only on the order of 1 nm make Frenkel exciton highly localized on the molecule. Charge transfer exciton delocalizes to a few molecules, which occurs in solids with long range order or as an intermediate process during exciton dissoc iation in either donor acceptor inte rface or organic based devices. 3 .2.2.2 Exciton Properties Excitons can be formed by optical excitation or free electron and hole recombination in organic solids. Th e binding energy of excitons is mainly contributed from the Coulombic, electron lattice and electron electron interactions. This binding energy should be overcome if excitons want to dissociate into free charge carriers. Figure 3 5. (color) Schematic illustration of three types of excitons in a solid. (a). a Frenkel exciton, localized within a single molecule; (b). a charge transfer exciton, slightly delocalized to two or a few adjacent molecules; (c). a Wannier Mott exciton, highly delocalized with radius much greater than the lattice constant a L
75 Other th an binding energy, exciton diffusion length and lifetime are other important parameters in or ganic based electronic devices. The behavior of the exciton (generation and recombination) involves a series of optical transitions that also determines the absorp tion and emission properties of the organic solids. Though the transitions between energy states contain several energetic processes such as electronic, vibrational and rotational processes, only the electronic transition has the energies the same as those of near infrared, visible and ultra violet photons that are of interest for optoelectronic device application. Hence, electronic transition will be only considered for the study of exciton behavior in organic solids here. Jablonski energy diagram shown i n in Figure 3 6, summarizes a series of optical and electronic transitions that exciton can undergo in organic semiconductors. The generated exciton by optical absorption has a singlet type due to spin conservation and a series of radiative and non radiati ve transition will follow. The possibility of these transitions, according to the selection rule, is dependent on the time scale. The exciton can relax to dissipate energy such as undergoing vibrational relaxation or internal conversion with conserved spin state ; and can also further absorb photonic energy that may lead to transitions to higher energy levels such as from S 1 to S 2 etc. Exciton may also undergo intersystem crossing when the vibrational levels of the two excited states overlap resulting in t he change of spin state. Exciton generated by recombination of a free electron with a free hole has lower energy state than the free electron due to its binding energy. S ince the spin state of electron is not restricted by that of hole during this recombin ation process, the generated exciton could be either singlet or triplet (the theoretical permutation ratio of
76 singlet to triplet is 1 to 3). The singlet exciton can relax to the singlet ground state accompanying with emission of a photon or fluorescence; a nd the triplet exciton can also relax to the singlet ground state in a molecule with allowed intersystem crossing, resulting in phosphorescence. Since phosphorescence involves a spin orbital coupling process, its time scale (0.1 1 ms) is much longer than t hat of fluorescence (1 100 ps). B oth of fluorescence and phosphorescence have much longer lifetime than that of vibrational transition (0.1 ps). Hence, the exciton thermally relaxes to the lowest excited state before undergoing photon emission processes, l eading to red shift in emission spectrum in organic semiconductors. This shift can also be explained by the Frank Condon effect, which demonstrates photon emission requires an emission of a phonon due to the coordination coordinate shift between the ground state and the excited state, thus resulting in lowering in photon emission energy. Other than radiative processes, there are also a number of non radiative processes that an exciton in organic solids may carry on, as shown in Figure 3 6. Figure 3 6. (co lor) Jablonski energy diagram. The radiative transitions and non radiative transitions are indicated by squiggly arrows and straight arrows, respectively. Adapted from www.olympusmicro.com.
77 3 .2. 2.3 Exciton Motion The motion of the excitons in organic solid s can be described using the thermally activated hopping model. The exciton transport is important to understand the mechanism and optimization of organic based optoelectronic devices. Several mechanisms have been proposed to describe the exciton motion in cluding energy transfer and energy migration. First, cascade or trivial energy transfer is a process of reabsorption of a photon emitted from a donor molecule by an acceptor molecule. 88 This energy transfer is important at long distances and when there is strong overlap between the donor emission and acceptor absorption spectra. Furthermore, this process also leads to l engthen the apparent lifetime of the singlet state. The second similar to cascade energy transfer, also depends on the spectrum overlap between the donor emission and acceptor absorption. The energy received by the acceptor from the donor can be transferred to the degenerat e states and subsequently relaxed to vibronic states of low energy, which produces a dephasing of the excited state and also vibrational decay. In a very J acceptor electronic state and J is the strength of the interaction between the donor and the acceptor), the transfer rate by the assumption that the transfer of electronic excitation energy between a donor and an acceptor is mediated by dipole dipole interaction can be writte n as 88 (3 3)
78 w here F D is the normalized fluorescence emission spectrum, A ) is the normalized acceptor absorption cross section, n 0 is the refraction index of the solvent, c is the speed of the light, D is the natural lifetime of the donor, and integration is over all the frequencies. Hence, it is straightforward that the tra nsfer is not occurred unless there is overlap between F D and A Dexter extended the theory of resonance energy transfer to exciton states with electron exchange between donor and acceptor. This transfer distance is very short (~ 1nm), but it is of i mportance in the case of triplet triplet energy transfer. The transfer rate is given by 96 (3 4) where DA is the exchange energy interaction between molecules, F D (E) and F A (E) are the normalized phosphorescence spectrum of the donor and absorption spectrum of the acceptor molecule, respe ctively. Due to the non directional nature of these energy transfer processes, the exciton diffusion can be used to describe the exciton migration with time (3 5) where the concentration of the excitons and D is the diffusion coefficient. D iffusion length ( l ) a distance that an exciton can travel before relaxing back to the ground state is related to the diffusion coefficient and exciton lifetime and given by where Z=6, 4, and 2 for three, two, and one dimensions, respectively. The
79 diffusion length for organic semiconductors is typically 5 20 nm, as experimentally determined by methods such as luminescence quenching. 14 97 3 3 Charge T ransport in Organic Solids The charge transport of solids can be very different depending on the bonding nature and int ermolecular interaction. Inorganic crystals that h o ld together through strong covalent or ionic bonding lead to the strong overlapping of atomic orbitals in three dimensions and then the formation of transport band. The charge transport can be highly deloc alized in this band that is slightly affected by the present impurities, defects, and phonon scattering. However, molecular solids that h o ld together through weak van der Waals interaction exhibit much more complicated transport behavior C harge carrier mo bility in organic solids can be influenced by a number of internal and external factors such as molecular packing, impurities, molecular weight and distribution, disorder, temperature, pressure, electric field, etc. 98 100 The angle dependent charge carrier mobility of pentacene single crystal is an example of molecular packing. 88 In inorganic solids, charge transport is mainly governed by the electronic interaction, and the electron phonon interaction is very small that is considered as a perturbation. However, in molecular crystals, the e lectron phonon interactions are comparable to or even larger than the electronic interaction, which leads to the formation of polaron, quasiparticle of phonon dressed electronic charge. 99 Organic crystals with long range order at low temperature may exhibit band like transport behavior. A slightly delocalized transport energy band can be formed in organic solids with high structural regularity cm 2 V 1 s 1 at room temperature. 88 T his band like transport is subjected to the
80 environmental temperature as ( n > 1 ) 101 and t his temperature dependence attributes to the large electron phonon interaction in molecular solids. In fact, most of the organic solids are m orphologically amorphous due to their bonding nature. Hopping model is mostly popular for describing the transport behavior of injected charge carrier within organic solids. The typical motion of highly localized polaron in amorphous organic solids is char acterized by hopping from site to site and scattered as a result of interaction with the local environment. The mobility in term of electric field and temperature then can be expressed as 102 (3 6) where F is the electric field, is the activation energy for intermolecular hopping, and is a constant. The mobility of charge carrier in molecular solids can be determined by time of flight, field effect transistors, spa ce charge limited current, etc. 99 Time of flight is resultant mobility. FETs method has been favored for organic crystals, while mobility determined by FETs is highly sensitive to the organic electrod e and organic dielectric contact preparation the surface polarity dielectric topology etc. SCLC offers a convenient means to determine the mobility from the electrical characteristics of diodes with organic layer sandwiched by two electrodes. This metho d is particularly useful for ultrathin films (tens to hundreds of nanometers). The charge injection is determined by the choice of electrodes and only one type of charge carriers can be extracted for the chosen electrodes. Ideally for organic thin film wi th traps, J V curves may exhibit several regions: i ). a linear regime where the transport is injection limited, ii ). an
81 intermediate region, iii ). an SCLC region where J scales quadratically with applied voltage, iv ). a trap charge limited conduction (TCLC) region where J scales cubically with V and v ). a trap free SCLC regime. In the first SCLC regime, the J V 88 103 (3 7) where L is the film thickness, r is t he dielectric constant, and is the charge carrier mobility. 3 .4 Organic Photovoltaic Cells 3 .4 .1 Principle of Organic Photovoltaic Cells Unlike the conventional inorganic PV cells, the operation mechanism of organic based PV cells including all organic a nd organic inorganic hybrid solar cells are based on donor acceptor (DA) heterojunction formed between two different materials with staggering energy levels. 13 14 As shown in Figure 3 7, the incident photons can be absorbed by the donor and acceptor materials, leading to the generation of a tightly bounded exciton. 88 104 106 The photogenerated exciton then diffuses to the DA hetero interface with sufficient energy difference for dissociation. 104 However, the binding ene rgy rising from the electron hole mutual Coulomb attraction is sufficiently high (0.3 1 r : ~2 4), thus the energy offsets between the LUMOs of the donor and the acceptor or/and between the HOMOs of the donor and the acceptor should be higher than this binding energy for achieving exciton dissociation into free electron and hole (polarons) at the DA interface. The dissociated free polarons are then collected by the respective electrodes.
82 Figure 3 7 (color) Consecutive steps for photocurrent generation from incident light in a bulk heterojunction organic based photovoltaic cells. (Courtesy of J. Xue). The absorption efficiency ( A ) depends on the absorption coefficients of the donor and acceptor materials and their collective thickness. 106 The exciton diffusion efficiency ( ED ) depends on the exciton diffusion length and the degree of phase separation. 107 108 U niform and delicate phase separation is essential for efficient exciton diffusion in organic based PV cells. 108 109 Through careful control in donor and acceptor materials processing, phase separation with an ideal size ( around twice of the exciton diffusion length) is achievable in bulk heterojunction (BHJ) PV cells. 110 For exam ple, as shown in Figure 3 8, hybrid film based on polymer and CdSe nanocrystals (~ 6 7 nm) shows well separated binary phases with phase size ~10 nm, around the twice of the exciton diffusion length. 61 In contrast, exciton diffusion is the ma jor limiting factor in bilayer heterojunction PV cells 14 The exciton dissociation (charge transfer) process is not straightforward, but rather complicated with a series of competitive sub processes involved. 111 A number of emerging studies indicate that charge transfer state is of critical importance in both photocurrent generation and open circuit voltage ( V oc ) in organic solar cells 111 114 When
83 Figure 3 8. TEM image of a polymer:CdSe nanocrystal hybrid film. (a) the original image; (b) the binary enhanced ima ge as a guidance for the eye, the length of the dark area is mostly < 10 nm, about one nanocrystal size; and the length of the white area is about 10 nm. The dark ar ea is nanocrystal network and the white area is conjugated polymer matrix. exciton diffuses to the DA interface, the initial step may generate an interfacial charge transfer (CT) state (Figure 3 9). This initial hot CT state will thermalize with an increased electron hole separation distance. Germinate recombination (GR) may occur in the singlet state (S 0 ) at the relaxed CT state. 111 Besides, the relaxed CT state can undergo rapid spin mixing between its singlet and triplet states, which could result i n the transfer from the singlet state to the triplet state (T 1 ). Otherwise, the CT state can undergo full charge separation into free positive and negative polarons. According to the Onsager theory, 115 the efficiency of exciton dissociation is critically dependent upon the thermalization length and Coulomb capture radius. In the viewpoint of kinetics in exciton dissociation, the steps between thermalization and charge transfer and between GR and charge separation (CS) are regarded as the main competitive processes. Due to the ultrafast nature of charge transfer, 116 the typical charge transfer efficiency for bulk heterojunction PV cells based on conjugated polymers /small molecules and fullerene deri vatives is close to unity; while this efficiency is still under debate for BHJ
84 Figure 3 9. (color) Energy level diagram of the possible mai n processes of exciton dissociation in an organic based PV cell CT: charge transfer; CS: charge separation; GR: g erminate recombination; and BR: bimolecular recombination. polymer:nanocrystal hybrid PV cells due to the complex nanocrystal surface / interface. 18 However, the G R during exciton dissociation has been independently V oc 113 Thus, managing the GR is essential in achieving both unity CT and high V oc Charge carrier transport the final step of photocurrent generation, also critically relies on the active layer morphology (Figure 3 8) and the donor and acceptor mobilities. 14 The highly percolated network of the donor and acceptor m aterials has been regarded as the ideal morphology for efficient charge transport. 107 The phase separation in polymer:nanocrystal hybrid s as shown in Figure 3 8, is in a degree of ~10 nm. 61 Such a delicate separation is of advantage for efficient exciton diffusion; while it may be too fine to carrier transport due to the possible wavefunction overlapping of electron and hole This overlapping may lead to an increased bimolecular recombination, which, to a large extent, a ccounts for the loss of photocurrent in organic based PV cells. 14 Addressing this exciton diffusion charge transport compromise rising
85 from the donor acceptor phase separation is of great importance for high efficiency BHJ PV cells. Thus, the external quantum efficiency (EQE or EQE ) that is defined as the ratio of the number of collected electrons to the number of incident photons is the multiplication of the efficiencies of each process and expressed by EQE = A ED CT CC = A I QE (Figure 3 7). 3 4.2 Progress of Organic P hotovoltaic Cells Organic PV cells have witnessed drastic progress during the last decade with p reaching 10% recently. 4 Organic PV cells have been typically classified as small molecule (SM) and polymer PV cells. The SM PV cells utilize organic small molecules as donor and fullerene (C 60 or C 70 ) as acceptor. Though SM PV cells using solution processing have shown some significant progress recently, 117 the high crystallinity of small molec ules together with unfavored phase separation make high vacuum thermal evaporation technique still the primary choice. 14 Polymer solar cells (PSCs) are using co njugated polymers as donor and fullerene derivatives as acceptor and based on bulk heterojunction structure. 107 The PSCs are fabricated from solution by spin coatin g, dip coating, even roll to roll, etc. There are many similarities in the research of small molecule and polymer solar cells; here we emphasize the latter with a brief progress overview The ultrafast energy transfer from a conjugated polymer to C 61 butyr ic acid methyl ester (PC 61 BM) observed independently by Heeger and Yoshino and their coworkers in early 1990s is the basis of the development of BHJ polymer solar cells. 116 118 Though the planar structured PSCs was firstly demonstrated in 1993 119 this structure suffers from inefficient exciton dissociation and charge carrier transport, which
86 received little attention after introduction of bulk heterojunction concept. 107 BHJ PSCs were pioneered by Heeger and co workers by blending poly( 2 methoxy 5 (2' ethylhexyloxy) p phenylene viny lene ) (MEH PPV) and PC 61 BM. 107 BHJ PSCs have then shown dramatic progress by tailoring conjugated polymers and fullerene derivatives, optimizing the active layer mo rphology, and engineering the device structure. 3 .4 .2.1 Conjugated polymers Though MEH PPV was firstly developed for the BHJ polymer solar cells, poly(3 hexyl thiophene) (P3HT) has then shown to play a leading role due to its wider absorption coverage, hig her molecular regioregularity, and correspondingly higher hole mobility. 107 120 122 Partly due to the vast commercial availability, good air stability and facile processing, P3HT:PC 61 BM blend has become the standard system for testing new device architecture and studying device physics. 123 The p of P3HT:PCBM cells falls within the range of 2 5%. 123 However, the relative large energy gaps of PPV and PT d erivatives make PSCs fabricated using these materials only absorb small part of solar photons, leading to relatively low J sc (< 12 mA/cm 2 ). 123 To harvest a large portion of the near infrared (NIR) solar photons, a series of low gap polymers have been developed in the past sever al years. One of the common basic designing rules for the low gap polymers is to co polymerize an electron donating unit with an electron accepting unit. This copolymerization leads to the energy level overlapping of the electron donating unit molecule and electron accepting unit molecule, resulting in raising HOMO level lowering LUMO level and thus lowering the energy gap. 124 Poly (2,1,3 benzothiadiazole 4 ,7 diyl( 4,4 bis(2 ethylhexyl) 4H cyclopenta(2,1 b:3,4 b') dithiophene 2,6 diyl)) (PCPDTBT) is the first low gap copolymer developed based on this design rule for PSCs with p >3% and with significant harvesting of NIR
87 photons. 125 The unending endeavor in develo ping semiconducting copolymers by a number of research groups lead to the numerous reports of PSCs based on the blends of these copolym ers and PCBM with J sc in the range of 12 16 mA/cm 2 125 130 Apart from narrowing the bandgap, the mobility, intermolecular interaction, and molecular chain p acking of polymers can also significantly affect J sc Finally, it is worth mentioning that replacing C 60 with C 70 in PCBM typically results in 10% increasing in J sc due to b etter absorption of the latter. 131 For these state of the art cells based o n low gap polymers, though the J sc reaches 15 mA/cm 2 the V oc is still not high enough (< 0.7 V), partly because the energy level alignment between polymer donor and fullerene acceptor is not optimized. Leclerc and co workers then developed p oly((9 (1 octy lnonyl) 9H carbazole 2,7 diyl) 2,5 thiophenediyl 2,1,3 benzothiadiazole 4,7 diyl 2,5 thiophenediyl) (PCDTBT) with HOMO level of 5.5 eV, ~ 0.3 0.4 eV lower than P3HT and PCPDTBT. 132 Such deep HOMO can be of great benefit to increase the V oc of PSCs using model acceptor PCBM, since empirically the V oc follow the equation : V 8) High V oc (0.9V) was indeed obtained for PSCs based on blends of PCDTBT and PC 71 BM. 128 3 .4 .2.2 Morphology D elicate phase separation of dono r and acceptor blends is of critical importance for BHJ organic solar cells. Morphology control is usually achieved by thermal and solvent annealing for PSCs using conjugated polymers and fullerene derivatives. For example, P3HT:PCBM film casted from dichl orobenzene after slow solvent annealing
88 results in higher electron and hole mobilities, more balanced charge t ransport and better absorption. 121 Similarly, higher crystallinity and better phase sep aration have also been observed in P3HT:PCBM with thermal annealing. 133 134 Thus, the p of PSCs based on P3H T:PCBM with improved morphology after either solvent or thermal annealing reaches to 4 5%, around 1 2 times higher than control devices without any treatment. Although thermal and solvent a nnealing are very effective in P3HT:PCBM cells, these methods are n ot as effective as in PSCs based on low gap polymers such as PCPDTBT. 125 Bazan and co workers discovered that processing PCPDTBT:PCBM blends using dichlorob enzene mixing with some additives such as alkanedithiol could result in better morphology favored for exciton diffusion and charge carrier extraction, and thus the significant increase in J sc from 10 mA/cm 2 to 16 mA/cm 2 and in p from 3.5% to 5.5%. 135 This discovery has then been extensively extended to a variety of materials system with better morphol ogy. 3 .4 .2.3 Device a rchitecture The basic device structure of organic solar cells is the active layer sandwiched by a transparent electrode such as ITO and a reflecting electrode such as Al and Ag. PEDOT:PSS has typically used as anode interlayer to smooth the ITO surface and accordingly increase work function. Moreover, transition metal oxides such as MoO 3 WO 3 NiO, and V 2 O 5 have also been developed as anode interlayer with comparable device performance. 136 137 On the other hand, LiF, 138 TiO x 128 and ZnO, 139 either processed by vacuum deposition or solution casting have usually introduced between the active layer and the cathode to improve the performance of PSCs. The high transparency makes TiO x and ZnO promising to tune t he optical field distribution within the active layer.
89 Moreover the transparent and conductive metal oxides are also applicable as inter connecting layer for solution processed multi junction solar cells. Multi junction solar cells have the advantages of absorbing solar photons in each junction with selective wavelengths and simultaneously adding the voltages produced in each junction. Due to relative large bandgap of organic materials, making PSCs with multi junction co uld be of exceptional benefit. 140 The early example by Kim et al. demonstrated a tandem cell with PCPDTBT:PCBM as front cell (the cell close to transparent electrode) to absorb NIR photons and P3HT:PCBM as back cell (the cell close to the reflective electrode) to harvest visible photons. 141 The cell showed open circuit voltage addition of two subcells and p as high as 6.5%, even close to the efficiency addition of the t wo single junction PCPDTBT:PCBM cell (3.0%) and P3HT:PCBM cell ( 4.7% )
9 0 CHAPTER 4 EFFECT OF COLLOIDAL NANOCRYSTALS ON HYBR ID PHOTOVOLTAIC CELL S 4 .1 Introduction P rogress in synthetic chemistry o f inorganic materials enables us to tailor nanocrystals with variable size, shape, compo sition, and surface chemistry. 27 For example, the emissive property, which can be man ipulated by engineering nanocrystal size and shape, makes colloidal nanocrystal s rising star emitter s in light emitting diodes 32 and biological labeling/imaging. 34 35 Colloidal nanocrystals that not only ha ve been regarded to inherit the intrinsic properties (high ele ctron mobility and environmental stability) of their bulk counterparts solution processability were introduced into polymer photovoltaic cells as electron acceptor 16 The main roles of the colloidal nanocrystals in hybrid PV cells are photogeneration ( light absorption, or exciton generation), charge transfer by creating a donor acceptor junction with organic material and electron tra nsport. With these in mind, recently, it is a trend to study organic inorganic hybrid or quantum dot PV cells by tailoring colloidal nanocrystals with defined chemical, electronic, and morphological structures. 18 84 In this chapter, we attempt to understand the effect of chemical composition, size, shape, and even surface chemistry of colloidal nanocrystals on the performance of hybrid PV cells. The electronic structure of the colloidal nanocrystals will be also probed to relate with the performance of the hybrid PV cells.
91 4 .2 Colloidal N anocrystals Synthesis and P rocessing 4 .2.1 Spherical Nanoparticles S ynthesi s Spherical CdSe nanoparticles (or quantum dots) were synthesized according to the reported method with some modification 61 142 76 mg CdO (Alfa Aesar), 3 mL oleic acid (Sigma Aldrich ), and 3.0 g trioctylphosphine oxide ( Sigma Aldrich) were mixed and heated to 280 o C under N 2 flow. A 1.0 mL trioctylphosphine (Sigma Aldrich ) solution containing 78 mg Se (Sigma Ald rich) was injected at 280 o C. The reaction mixture was kept for 1 5 min at 27 0 290 o C to produce CdSe nanoparticles with average sizes of 4 7 nm. The reaction was then terminated by adding 5 mL tolu ene into the mixture. CdSe n anoparticles were purified b y precipitation with methanol and re dissolution in toluene for three cycles CdS quantum dots were synthesized as follows 143 CdO (154 mg), OA (3 mL), and ODE (45 mL) were mixed and heated to 250 o C under N 2 flow, at which temperature S/ODE (38 mg/6 mL) was injected. A liquots were taken for monitoring the nanocrystal size. After reaction, the solution was naturally cooled to R.T., and purified by methanol preci pitation and hexane dispersion. 4 .2.2 Nanorods S ynthesis CdS nanorods were synthesized according to the published method. 144 CdO (0.105 g), octadecylphosphonic acid (ODPA, 0.45 g), and TOPO (1.45 g) were mixed and heated to 320 o C under N 2 flow until a clear mixture solution obtained. The mixture was then cooled to 120 o C and pumped for 1 h at this temperature before re heating to 320 o C, at which point 1 mL TOP was injected. After the temperature returned to 320 o C, the S/TOP (120 mg/1.5 mL) mixture was injected and the react ion was kept at 310 320 o C for 85 min. The mixture was naturally cooled to room temperature, and 5 mL
92 toluene was added. CdS x Se 1 x nanorods were synthesized t he same as CdS nanorods, with the exception of S/Se/TOP injection (85 mg S/0.3 mL TBP/0.5 mL TOP ; 15 mg Se/0.15 mL T BP/0.5 mL TOP/ 0.3 mL toluene). CdSe nanorods were synthesized in a similar manner as CdS nanorods 145 CdO (0.205 g), TOPO (3.0 g), and TDPA (0.85 g) were mixed and heated to 320 o C under N 2 flow until a clear mixture solution obtained. The mixture was th en cooled to room temperature and aged for 24 h and then re heated to 320 o C, at which point Se/TBP/TOP/toluene (63 mg/0.3mL/1.0mL/0.3mL) mixture was injected The reaction was kept at 250 280 o C for 40 min and then was naturally cooled to room tempera ture. The width and aspect ratio of CdSe nanorods were tuned by varying the growth temperature. 4 .2.3 Nanocrystal P rocessing The purified CdSe nanoparticles (~80 100 mg) were mixed with 10 mL pyridine and sonicated for 10 min at room temperature. The nano particles / pyridine mixtures were then transferred to a three necked flask and stirred for another 24 h under N 2 flow at room temperature. After ligand exchange, the nanoparticles were precipitated from pyridine solution by hexane. The precipitated nanopa rticles were dried in a vacuum oven. P olymer:CdSe hybrid solution was prepared by dispersing 30 mg CdSe nanoparticles and 3 mg of conjugated polymers in chlorobenzene/pyridine (90/10, v/v) co solvent. CdS and CdSe nanorods were purified by precipitation wi th methanol and dispe rsing in toluene for 6 cycles. The purified nanorods were mixed with 5 10 mL pyridine and refluxed for 24 h. Then the nanorods were collected by hexane precipitation. In order to completely remove the exchanged alkyl chain ligands, t he
93 pyridine exchanged nanorods were re dissolved in toluene and then centrifuged to collect the nanorod solids. Finally, the nanorods were dried and re dispersed in chloroform before mixing with conjugated polymers with a concentration of 30 mg nanorods an d 3 mg polymer s for characterization and device fabrication. 4.3 Hybrid Film Characterization and PV Cell Fabrication 4.3.1 Nanocrystals and Hybrid Film C haracterization The purified CdSe nanoparticles and nanorods were characterized using a JEM 2010F tran smission electron microscope (TEM) and a JEOL 200CX TEM both with a 200 keV electron beam energy. UV Vis absorption was recorded on a Cary UV Vis absorption spectrometer. The X powder X ray diffractometer. The surface morphology of the hybrid thin films was characterized using a Veeco Nanoscope atomic force microscopy (AFM) operated in tapping mode. 4.3.2 Hybrid PV Cell Fabrication G lass substrates pre patterned with i ndium tin oxide as anode (ITO) (R s : 20 were cleaned by successive sonication with soap, deionized water, acetone, and isopropanol for 15 min, followed by exposure to UV ozone for 15 min. A poly(3,4 ethylenedioxythiophene): poly(styrenesulfonate) (PEDOT:PSS Clevious P, H.C.Stark ) layer was first spin coated on the ITO substrates and annealed in air at 150 o C for 15 min. P3HT:CdSe hybrid active layer w as spin coated in a N 2 glove box with H 2 O and O 2 both < 0.1 ppm and then annealed at 150 o C for 30 min. For devices based on nanorods, the act ive layers were processed from chloroform and annealed at 1 2 0 o C for 20 min. The devices were completed by vacuum deposition of an Al cathode through a shadow mask, resulting in each device with an active device area of approximately 4
94 mm 2 The PV devices were characterized by J V measurement under dark and 1 sun AM 1.5G illumination. The EQE measurement was performed as described in Chapter 1. 4.4 Nanocrystal C haracterization and Properties 4.4 .1 N anocrystal S ize CdSe nanoparticles with different size we re sy nthesized, as shown in Figure 4 1. The 4 nm particles are almost monodisperse and self assembled into hexagonal structure in carbon coated copper grid. When the sizes increased to ~5 nm and 6 nm, the nanoparticles become less monodisperse; while furth er increasing nan oparticle size by prolonging reaction time results in significant size defocusing and presence of both spherical particles and elongated rods (Figure 4 1d) The difference in size dispersity is likely due to monomer consumption. 43 The initial monomer concent ration and precursor injection temperature control the concentrati on of nuclei. The growth to a large extent, is also monomer concentration dependent. At high monomer concentration, the monomer diffusion to each nucleus is higher than the consumption rate resulting in monodispersity; however, as the monomer consumed, the monomer reaching each particle becomes unequal, which leads to the size polydispersity. Ost wald growth will appear as the monomer is further consumed to a threshold, which leads to the di ssolution of small particles (particles with high surface energy) and the growth of large particles, as shown in Figure 4 1d. UV Vis absorption spectra of the CdSe nanocrystals were shown in Figure 4 2. The first excitonic peak s of all four samples are app arent, while they become less pronounced as nanocrystal size increases another indication of size polydispersity. Due to the quantum confinement effect, a s the nanoparticle size decrease s the
95 Figure 4 1. TEM images of CdSe NPs with different sizes: (a ) 4.0 nm, (b) 5.0 nm, (c) 6.1 nm, and (d) 6.8 nm. The scale bars are 20 nm. absorption blue shifts and the corresponding band gap increases. 146 The estimated band gaps of these nanocrystals, based on the absorption onset, are in the range of 1.8 eV (6.8 nm) 2.0 eV (4.0 nm). The broadening in band gap as nanoparticle size decrease s also leads to the corresponding shift in the CB and VB of the nanocrystals. The magnitude of the change in CB and VB is related to the effective mass of electron and hole of the bulk materials, respectively. Both band gap and energy level changes should be carefully taken into account once the nanocrystals are used as the active materials in optoelectronic devices. 4.4 .2 Nanocrystal S hape and C omposition Nanocrystal shape and composition, another two important parameters determining the nanocrystal properties, can be tailored in a simultaneous manner. Using facet defining organic surfactants such as alkyl phosphonic acids, elongated nanorods,
96 Figure 4 2. (color) UV Vis absorption spectra of CdSe nanoparticles with different sizes. rather than quantum dots can be achieved 28 Figure 4 3 shows TEM images of CdS, CdS x Se 1 x and CdSe nanorods passivated by alkyl phosphonic acids. The width and aspect ratio of these nanorods can be readily controlled by var ying monomer concentration, reaction time and t emperature, and surfactant type For example, nanorods synthesized using longer alkyl chain s have shorter width and higher aspect ratio, which is likely due to the steric hindrance effect of the long alkyl cha in s resulting in less favored side growth. I ncreasing the grow th temperature can result in short and fat nanorods, suggesting growth favored along the side direction. 147 This may be due to the increasing kinetic energy for ligand attach ment and lig and detach ment and also the more uniform monomer distribution along the nanocrystal head and side. Finally, branched nanocrystals have frequently appeared in nanorod synthesis, which is likely due to the impurity of alkylphosphonic acids and the present o f zinc blende CdSe nanoparticles at the initial growth stage (Figure 4 3). 148 Though elongated nanocrystals may have dimension in one direction much larger than their Bohr radius, the quantum confine ment effect is more or less conserved
97 Figure 4 3. TEM images of composition varied semiconductor nanorods: (a) CdS, width:3.1 nm, A.R. ~30; (b) CdS x Se 1 x width: 3.2 nm, A. R. ~ 21; (c) CdSe, width:4.0, A. R. ~ 10. The scale bars for (a, b): 100 nm, and (c): 50 nm. compared to the spherical nanoparticles with size similar to the ir width. 146 149 The abso rption spectra of nanorods, shown in Figure 4 4, exhibit characteristic and size dependent absorption properties For example, CdS nanorods show a very sharp first excitonic absorption peak and pronounced second absorption peak with a band gap of ~ 2.6 eV. The band gap of the nanocrystals can also be readily tuned by changing the nanorod composition. For example, if S and Se precursors are co injected into Cd precursor solution, ternary nanorods CdS x Se 1 x can be produced. The recent study by Ruberu et al. d emonstrated that CdSe would grow first due to the higher reactivity of Se/TOP and the resulting nanocrystals are partly alloyed with non uniform rod like structure. 147 I n this case, the na norods are ve ry uniform though further study is needed for the composition distribution in these ternary nanorods. The nanorods also show a red shift in absorption co mpared to pure CdS nanorods, yet the excitoni c peaks become less pronounced, partly due to the non uni form distribution of CdSe and CdS components. In addition, pure CdSe nanorods show defined absorption spectrum with a band gap of 1.9 eV (Figure 4 4)
98 Figure 4 4. (color) UV Vis absorption spectra of composition tunable semiconductor nanorods. X ray diff raction is employed to further characterize thes e nanorods. As shown in Figure 4 5, all nanorods show characteristic peaks of wurtzite phase and peak broadening due to the finite dimension. 149 CdS x Se 1 x nanorods shows defined diffraction peaks locating in between those of CdS and CdSe nanorods, suggesting partial alloy of S the composition of the alloyed nanorods can be estimated as CdS 0.34 S e 0.66 The nanocrystal surface is also of critical importance for understanding its growth and properties and for optoelectronic devices and biomedical application However, it is not straightforward to quantitatively identify the nanocrystal surface due t o its complicated nature rising from synthesis and crystallization, the characterization techniques including optical and electrical measurements typically reveal the collective properties of the nanocrystals rather than individual particle 25 Since this dissertation focuses on the application of colloida l nanocrystals in photovoltaic cells, here we will not particularly discuss the nanocrystal surface yet will definitely connect nanocrystal surface when the nanocrystals become a dominant factor of device performance.
99 Figure 4 5 (color) XRD patterns of composition tunable Cd chalcogenide nanorods. 4 5 Effect of D evice Aging on P erformance We began this study in the P3HT:CdSe hybrid PV cells, partly because this system is more mature in processing and device fabrication. Though the previous study indicate d hybrid PV cells based on spherical nanoparticles were not efficient (< 1%), 16 the advancement in nanocrystal synthesis makes them worthwhile to re visit as acceptor in hybrid PV cells. An abnormal aging behavior has been observed for th e unencapsulated P3HT:CdSe hybrid PV cells when exposed to air. 61 As shown in Figure 4 6a, the initial test showed high dark current and small rectification ratio at =1/ 1 V (~60), while the dark current was decreased and rectification ratio increased to ~450 after exposing to the air for 0.5 h. Though the dark current was further decreased after 2 h exposure, the rectification ratio also decreased to ~18 due to the reducing in forward current. The J V characteristics under 1 sun AM 1.5G ill umination show corresponding behaviors. First, low J sc = 3.2 mA/cm 2 and V oc = 0.42 V were exhibited under initial test. After exposure to the air for 0.5 h, both J sc and V oc were drastically increased to 4.7 mA/cm 2
100 and 0.68 V, respectively, resulting in th e increase of p from initial 0.6% to 1.5%. The device performance can be maintained at this stage for ~ 0.5 1 h, and varied due to the difference in sample preparation. After 2 h exposure, however, the J sc was decreased to the initial level (3.0 mA/cm 2 ) and p was dropped to < 0.5% due to the drastic decrease in FF and subsequent increase in series resistance (from initial 3.0 2 to 202 2 ) This aging behavior has also been observed in hybrid PV cells based on CdSe nanorods. As shown in Figure 3 6 b, the dark curre nt decreased when exposed to air for a period of time and simultaneously the rectification ratio increased. Similarly, the device showed low J sc and V oc at the initial test, and then gradually increased. The Figure 4 6 (color) Aging ef fect of hybrid PV cells based on colloidal nanocrystals. J V characteristics of hybrid PV cells based on P3HT and CdSe quantum dots (a) or nanorods (b) under dark and 1 sun AM 1.5G solar illumination.
101 maximum device performance was observed after exposure to the air for 5 h and retained at this level for several hours. After 24 h, the device efficiency was dropped to less than half of the best value due to the large decrease in V oc The increase in V oc can be mainly explained by the decrease in dark curren t as exposed to the air, since the V oc is the voltage at which the dark current equals to the photocurrent and a reduced dark current requires a higher voltag e with equal photocurrent. However, t he change in J sc and FF is less understood at this moment. In order to rule out the potential influence of post fabrication relaxation of the hybrid films, we tested the devices that had been kept at a N 2 filled glove box for 12 h. These devices exhibited the similar initial aging behavior as exposed to the air as t he freshly prepared devices. Further study indicates that this aging behavior also occurs in hybrid PV devices using other conjugat ed polymers such as PCPDTBT ( Chapter 7 ), in the devices with a ZnO NP layer (Chapter 5 and Chapter 6) and in the devices wit h resin encapsulation as well. T his behavior has seldom appeared in all organic P3HT:PCBM PV cells. Together, we attribute this phenomenon to the interaction of CdSe n anocrystals with water / oxygen. Air induced photoluminescence enhancement was observed i n several colloidal nanocrystal systems. 150 155 The early study by Koberling et al. 151 showed the exposure of CdSe nanocrystals to the oxygen changes both the intensity and fluctuation of photoluminescence, which is likely due to the injection of electron s to the oxygen via an Auger process leaving a positively charged nanocrystal. The air induced fluorescence also occurred in CdSe/ZnS core shell nanocrystals. Muller et al. 153 proposed that the neutralization of charged non emissive nanocrystals by oxygen adsorption facilitated by
102 the presenc e of water is the mechanism of this air induced f luorescence enhancement. T he study by Cordero et al. 150 showed water molecule s are responsible for the increas ed luminescence intensity of CdSe nanocrystals. The initial adsorption of water molecules passivates t he CdSe nanocrystal surface and leads to the enhance ment of the luminescence, while the extended exposure to air leads to the formation of an oxide layer that consequently results in luminescence quenching The observation by Pechstedt et al. 155 further demonstrated that photoinduced fluorescence enhancement of CdSe/ZnS nanocrystals requires the involvement of water molecules while only the presence of oxygen leads to the generation of non emissive nanocrystals due t o the photoinduced electron transfer to the oxygen. Hence, these studies give much insight on revealing the abnormal aging effect of hybrid PV cells based on CdSe nanocrystals. The exposure of the devices to air leads to the adsorption of oxygen and water molecules to the CdSe nanocrystal surface. In a limited period of time, this adsorption passivates the dangling bond of the nanocrystal surface and reduces the density of surface defects. In other words, this reduces exciton and/or charge carrier recombina tion as reflected by the increase in J sc and V oc Further exposure could lead to the formation of an oxide layer in the nanocrystal surface, resulting in the transition of a semiconductor to a quasi insulator and thus the huge inc rease in nanocrystal resis tance, as reflected in the drastic decrease in J sc and/or FF o f hyb rid PV cells (Figure 4 6). 4.6 Effect of Nanocrystal Size on Device P erformance The optical and electronic properties of colloidal nanocrystals are size dependent due to the quantum confin ement effect, which motivates us to study the effect of nanocrystal size on the performance of hybrid PV cells. The J V characteristics of
103 hybrid PV cells based on CdSe nanocrystals with different sizes are shown in Figure 4 7a. The device based on 4 nm CdSe shows J sc = 1.5 mA/cm 2 and V oc = 0.81 V. As the nanoparticle size increases from 4 nm to ~ 7 nm, the J sc increases accordingly yet V oc decreases slig htly. The plot shown in Figure 4 7b indicates a quasi linear relationship between J sc and nanoparticle size. Since the size dependence of V oc is not as significant as J sc p follows the same size dependent trend as J sc (Figure 4 7b). The maximum efficiency of the hybrid PV cells using P3HT and ~ 7 nm spherical CdSe particles is 1.90.2%, which is, to our knowledge, one of the highest efficiency reported based on the P3HT:CdSe system. 61 We endeavored to further increase the nanoparticle size to achieve higher device performance. Unfortunately, this attempt was unsuccessful due to the difficult y in dispersing large particles (size > 8 nm) in organic co solvents. The wavelength dependent EQE of hybrid PV cells based on P3HT and CdSe Figure 4 7. (color) Nanocrystal size effect: J V characteristics (a) of hybrid PV devices based on CdSe nanopart icles and P3HT and the corresponding plots of short circuit current density ( J sc ) and power conversion efficiency ( p ) versus nanocrystal size (b)
104 nanoparticles with dif ferent size is shown in Figure 4 8. The device using 6.8 nm particles shows significan tly higher EQE than the 4.0 nm device, which is consistent with J V measurement Besides, the EQE of the 6.8 nm device is red shifted slightly compared to the 4.0 nm device due to its lower band gap. Furthermore, the 6.8 nm device shows the highest EQE of 42%, about three times higher than that of the 4.0 nm device. Figure 4 8. (color) EQE of P3HT:CdSe hybrid PV devices based on different nanoparticle s izes. Though larger nanoparticles slightly extend the absorption, the significant ly higher EQE of the device with larger nanoparticle s indicates that other electro nic processes are the determining factors. We then measured electron mobility of P3HT:CdSe hybrid films with different nanoparticle size following the SCLC method. The electron only device has a structure of P3HT:CdSe hybrid film sandwiched by Al. Hence, the electron mobility is attained by fitting the J V c haracteristics shown in Figure 4 (eq. 3 7, r = 6.5 and L = 100 nm ). Thus, the electron mobilities ( e ) of hybrid films with 4.0 nm and 6.8 nm CdSe particles are (62)
105 10 6 and (52) 10 5 cm 2 V 1 s 1 respectively. I mprovement in electron mobility in hybrid thin film can be of great benefit in th e extraction of photogenerated charge carrier. Plus, the improved electron mobility is more balanced with hole mobility (110 4 cm 2 V 1 s 1 ) of P3HT in hybrid system, which is a lso reflected in the increase in FF for the 6.8 nm device. Furthermore, nanoparti cles with larger size have less surface atoms and thus less surface defect density. These surface defects could act as recombination centers for photogenerated excitons and charge carriers, resulting in lower photocurrent for the device with small nanopart icles. Thus, the improved electron mobility together with less surface defect density lead to the enhanced J sc and p for the devices based on large nanoparticles. Finally, the fact that the slight decrease in V oc as size increases could attribute to the d ecrease in energy offset between polymer HOMO and nanocrystal CB due to the decrease in nanocrystal band gap. Figure 4 9. (color) J V characteristics of electron only devices with a structure of Al/P3HT:CdSe/Al based on different CdSe NP sizes. The sol id lines are fitted
106 4.7 Eff ect of Nanocrystal Shape on Device P erformance The early study by Hyunh et al. 16 showed that the nanocrystal shape has a pronounced influence on the performance of hybrid PV cells. N anocrystals with elongated shape may provide direct charge transport pathway and facilitate charge carrier collection. We prepared CdSe nanorods with different width and aspect ratio for hybrid PV cells. The J V characteristics of the hybrid PV cells usi ng different nanorods were shown in Figure 3 10a. As a comparison, PV cell using CdSe nanoparticles with size comparable to the nanorod width shows J sc = 4.2 mA/cm 2 V oc = 0.60 V, FF = 0.40, and p = 1.0%. The cell using CdSe nanorods with width of 6.4 nm and aspect ratio of 2 shows increase J sc =4.9 mA/cm 2 V oc = 0.74, and p =1.4%. The performance of hybrid Figure 4 10. (color) Nanocrystal shape effect: J V characteristics (a) and EQE (b) of P3HT:CdSe hybrid PV devices based on different NC shapes.
107 PV cell has been further increased with J sc = 5.9 mA/cm 2 V oc = 0.72, FF = 0.54, and p = 2.3% when nanorods with higher aspect ratio used. The wavelength dependent EQE further demonstrates the adva ntage of nanorods as acceptor s in hybrid PV cells (Figure 4 600 nm ( mainly corresponding to P3HT absorption ) and at > 600 nm ( mainly correspond ing to the nanocrystal absorption ) increase as nanorod aspect ratio increases, indicating more balanced charge carrier extraction and particularly better electron extraction as reflected in the higher EQE at the low wavelength range. 4.8 Other Colloidal N anocrystals CdSe nanocrystals have been mostly studied in hybrid PV devices 16 including in this dissertation while bulk CdSe has band gap ~1.7 eV that limits the harvesting of near infrared region and is also not environmentally benign. Other colloidal nanocrystals such as CdS 156 157 158 ZnO 159 Si 160 CuInS 161 PbS 83 162 and PbS e have also been developed for photovoltaic application including hybrid PV cells. Here we synthesized some CdS nanocrystals for hybrid PV cells. The J V characteristics of hybrid P3HT:CdS PV cells are shown in Figure 4 11. The cell based on 3 nm quantum dots shows very limited J sc and V oc resulting in an p < 0.1%. The cell using CdS nanorods shows improved J sc and V oc resulting in an p = 1.3%. Further engineering in size, shape, and surface chemistry is needed to improve the device performance based on CdS nanocrystals 4.9 Summary In this chapter we discuss the synthesis and characterization of colloidal nanocrystals for hybrid PV cells. The abnormal aging effect of hybrid PV cells using CdSe nanocrystals have been observed, which is attribute d to th e complicate d surface
108 Figure 4 11. (color) Hybrid PV devices based on CdS nanocrystals and P3HT. Both devices were EDT treated before Al deposition (See details in Chapter 7) chemistry, particularly the interaction between the CdSe nanocrystals and moi sture / oxygen. Hybrid PV cells also show nanocrystal size dependent performance that is mainly accounted by the increased electron mobility and reduced surface defect density. Furthermore, we demonstrated that nanocrystals with elongated structure could b e beneficial for hybrid PV cells, since the nanorods may provide direct charge transport pathway s for more efficient charge collection. The results based on both CdSe and CdS show that higher efficiency is indeed observed compared to their spherical counte rparts.
109 CHAPTER 5 SOLUTION PROCESSED MULTI FUNCTIONAL ZINC OXID E NANOPARTICLE CATHODE INTERLAYER 5 .1 Introduction Though photogeneration in bulk heterojunction organic inorganic hybrid solar cells occurs at the photoactive layers, metal (electrode) semic onductor (active layer) junction is also of great importance, which affect s the charge injection and collection. Operation of bulk heterojunction solar cells requires a pair of electrodes with asymmetric work function to facilitate electron and hole inject ion/collection. Nonetheless, this results in a common challenge that work function of electrode material is not sufficiently low or not high enough to align with the energy level s of the semiconducting active layer, consequently leading to the loss of phot ocurrent and voltage. Moreover, the direct contact between electrode and active layer may cause exciton recombination at the electrode active layer interface. 163 Introducing electrode interlayer to physically isolate the contact between el ectrode and active layer and simultaneously increase (for anode) or decrease (for cathode) work function is a feasible approach to address these issues. This concept was early practiced by depositing a PEDOT:PSS layer upon ITO substrate, which witnessed gr eat success in improving the performance of solution processed organic based solar cells. On the other hand, introducing a transparent and semiconducting metal oxide layer to separate the active layer and cathode has also been studied in organic based sola r cells. 139 164 166 This chapter focuses on the effect of a solution processed ZnO nanoparticle (NP) cathode interlayer on the performance of hybrid solar cells. The roles of the ZnO NP laye r will be also comprehensively surveyed
110 5 .2 Synthesis and C haracterization of ZnO N anoparticle s ZnO nanoparticles 167 168 were synthesized by dropwise addition of a stoichiometric amount of tetrame thylammonium hydroxide (TMAH) / ethanol solution into zinc acetate dihydrate / dimeth yl sulfoxide (DMSO) solution and stirred at room temperature at the air for 1h. After reaction, the nanoparticles were firstly collected b y precipitation in ethyl acetate, and then re dissolved in ethanol and precipitated by hexane for another twice. The p urified nanoparticles were dissolved in ethanol with a concentration of ~30 mg/mL. ZnO NPs were dropped on carbon coated copper grid for TEM measurement. For XRD study, dried ZnO NPs powder was used. The thickness of the ZnO NPs processed from ethanol and deposited upon Si substrates was determined by ellipsometry measurement. As shown in Figure 5 1a, ZnO nanoparticles are uniformly distributed and have an average size ~3 nm with narrow size distribution. The XRD pattern shows that the nanoparticles are cry stalline with a wurtzite structure and the peaks are being broadened due to their polycrysta llinity and small size (Figure 5 1b). Figure 5 1 TEM image (a) and XRD pattern (b) of ZnO nanoparticles synthesized by wet chemistry method
111 Figure 5 2 (color ) Device structure and schematic energy level diagram of P3HT : CdSe hybrid PV cells with a ZnO NP layer. 5 .3 Effect of a ZnO NP Layer on Device P erformance The device structure and schematic energy level diagram of hybrid solar cells have been shown in Fi gure 5 2. The fabrication of polymer:CdSe hybrid solar cells is th e same as described in Chapter 4 The P3HT:CdSe active layer was spin coated upon ITO substrates pre coated with a PEDOT:PSS layer from chlorobenzene/pyridine solution. A ~25 nm thick ZnO NP layer was then spin coated upon the active layer, following by annealing at 150 o C for 30 min at a glove box. Due to the simplicity in synthesis and processing, CdSe nanoparticles were chosen to exemplify the effect of a ZnO NP layer on the performance of hybrid solar cells. Even using nanoparticles, we demonstrate the power conversion efficiency is comparable to those using anisotropic nanocrystals such as nanorods and tetrapods. 16 169 171 As indicated in the energy level diagram, ZnO has a CB ~ 4.2 eV aligned with both CdSe nanoparticles and Al cathode that enables efficient charge extraction from the active layer and it has a very deep VB ~ 7.6 eV that may block back transport of
112 hole into Al cathode. The large optical gap makes ZnO NPs very transparent for the visible light travel/propagation and limits photogeneration at the ZnO NP layer. We first demonstrate the effect of a ZnO NP layer on hybrid solar cells using archetypal polymer P3HT and CdSe nanoparticles (5 nm), and the J V characteris tics have been shown in Figure 5 3a and the corresponding performance parameters listed in Table 5 1. The device without a ZnO layer shows J sc = 2.6 mA/cm 2 V oc = 0.75 V, FF = 0.42, and p = 0.8%; while the insertion of a ZnO NP l ayer significantly increases J sc Figure 5 3 (color) Typical J V characteristics of P3HT:CdSe (5 nm) hybrid PV cells without and with a ZnO NP layer under 1 sun AM 1.5G illumination (a) and their corresponding EQE (b).
113 (4.4 mA/cm 2 ) and FF(0.47), leading to a relative 87% enhancement in p The enhancement in J sc is also reflected in EQE, as shown in Figure 5 3b, in which the device with the ZnO l the P3HT absorption, suggesting photogenerated holes have been more efficiently collected. The performance o f hybrid PV cells based on spherical nanoparticles, as demonstrated in chapter 4 is dependent on the size of nanoparticles. 61 Here we further demonstrate the effect of the ZnO NP layer on hybrid PV cells using larger nan oparticles. As shown in Figure 5 4, the inclusion of a ZnO NP layer in hybrid PV Figure 5 4 (color) Typical J V characteristics of P3HT:CdSe hybrid PV cells without and with a ZnO NP layer and using CdSe nanoparticles (6.8 nm) with mixed spheres and rods under 1 sun AM 1 .5G illumination.
114 cells using 6.8 nm CdSe significantly increases the J sc and FF, leading to p as high as 2.4% with a relative 50% increase compared to the control device. Such efficiency is comparable to the state of the art hybrid solar cells b ased on P 3HT and CdSe nanorods / tetrapods. 16 167 168 The performance parameters in Table 5 1 indicate that the efficiency enhancement of hybrid PV cells with a ZnO layer is primarily from the increase in J sc and secondarily from FF, while the V oc has been slightly decreased. This V oc decrease is mainly due to the increase in dark current for the ZnO devices (Figure 5 5), and the diode characteristics are more or less maintained. The increase in J sc and FF is a consequence of combinatorial electronic, optical, and morphological effects. 5 .4 Role of the ZnO NP L ayer We investigate the roles of the ZnO NP layer on hybr id PV cells in the perspective of electronics, optics, and morphology. First, the ZnO NP layer physically eliminates the direct contact between the active layer and the Al, which prevents the Table 5 1. Summary of photovoltaic performance under 1 sun AM 1 .5 G illumination for P3HT:CdSe hybrid photovoltaic cells. Active layer ZnO J SC (mA/cm 2 ) V OC (V) FF P (%) Enhancement (%) P3HT:CdSe (5 nm) no 2.6 0.75 0.42 0.8 -yes 4.4 0.72 0.47 1.5 87 P3HT:CdSe (5.5 nm) no 4.2 0.75 0.40 1.0 -yes 5.8 0.61 0.50 1.8 80 P3HT:CdSe (6.8 nm) no 5.8 0.64 0.44 1.6 -yes 6.5 0.70 0.52 2.4 50
115 direct quenching of photogenerated excitons at the active layer Al interface and inside the active layer as Al penetrated during vacuum deposition. Second, coating a 2 5 nm thick ZnO layer upon P3HT:CdSe hybrid active layer reduces the root mean square surface roughness from 9.7 nm to 3.8 nm as shown in Figure 5 6. The smoother surface is anticipated to improve the contact between the active layer and the cathode and thu s enhance the collection of photogenerated charge carriers. Moreover, the transparent nature of ZnO enables to change the optical field distribution of the hybrid active layer, as early demonstrated in organic solar cells using either TiO x 164 or ZnO 139 The optical field profiles of the devices without and with a ZnO layer have been calculated based on a transfer matrix theory. The optical constants for the calculation are determined by ellipsometry m easurement. As shown in Figure 5 7, the optical intensity has been re positioned as a ZnO NP layer inclu ded, resulting in higher optical field distribution in the active layer. In particular, this effect is more Figure 5 5 (color) Typical J V characteristics of P3HT:CdSe (5 nm) hybrid PV cells without and with a ZnO NP layer under the dark condition.
116 Figure 5 6. (color) Tapping AFM morphology of P3HT:CdSe hybrid film without (a) and with a ZnO NP layer (b). de due to the optical interference, where n is the refractive index of the hybrid active layer. This optical effect has been reflected in t he EQE, which is much higher at evice with a ZnO layer (Figure 5 3b). Thus, this optical effect e nables to tune the absorption of the active layer to maximize the device photogeneration in both conjugated polymers and CdSe nanoparticles by varying the ZnO layer thickness. In addition, the optical effect slightly shifts the photogeneration zone closer to the cathode, which may be advantageous due to the relatively lower electron mobility (~10 5 10 6 cm 2 61 than the hole mobility (~10 4 10 5 cm 2 121 thus resultin g in more balanced charge collection as reflected in the slight increase in FF. T he deep VB (~7.6 eV) of the ZnO suggests the possible blocking of hole diffusion to the cathode (Figure 5 2). We verify this behavior by measuring the EQE of the devices as a function of applied bia s at given wavelengths (Figure 5 8). The EQE of the device without a ZnO layer collected at all the given wavelengths exhibits a transition point (a minimum EQE) at a voltage ~ 0.94 0.97V, accompanying with a sharp ~ 180 o phase cha nge. As observed in the CuPc:C60 mixed heterojunction solar
117 Figure 5 7. (color) Optical intensity profiles of P3HT:CdSe hybrid photovoltaic cells. The materials in each layer were labeled accordingly. cells 172 this transition behavior in EQE indicates the reversal of photocurrent direction and the photogenerated holes can freely drift into the cathode at high biases, which may also facilitate the hole diffusion from the active layer to the cathode. Nonetheless, the EQE of the device with a ZnO layer shows a monotonic decrease with the bias yet there is no a sharp phase change observed. As indicated by the CuPc/C60 planar heterojunction solar cells 172 this behavior suggests the photocurrent direction is unchanged even at high bias where the direction of the electric field is reversed (from the anode to the cathode) and causes the drift of the hole from the active layer to the
118 cathode. The unchanged photocurrent direction is only compensated by a higher magnitude of the hole diffusion from the active layer to the anode, which is achieved by the hole blocking characteristics of the ZnO NP layer. Figure 5 8. (color) EQE and the corresponding phase change of P3HT:CdSe hybrid PV cells without (a) and with (b) a ZnO NP layer as a function of applied bias. 5 .5 Effect of the ZnO NP Layer on Device E nvi ronmental S tability The chemical and morphological instabilities in organic based solar cells are a very challenge to move this technology into commercial realit y. 173 The presence of chemically varied surface and interface in colloidal nanocrystals agai nst the environment further deteriorates this concern. The study in ab normal aging effect in Chapter 4 exemplifies the influence of the environment on the performance of the hybrid solar cells. This is truly the case for the unencapsulated devices without a ZnO layer, which typically degrades after exposure to the air for hours. 61 Nonetheless, the inclusion of a ZnO NP layer, though the spherical nanoparticles may not pack in a closest fashion, does drastically improve the device stability wit hout any encapsulation when stored at an environment with humility 35 60% and temperature 21 24 o C. 167 171 A s shown in Figure 5 J sc and p retain ~70% of the original value. Such an enhancement in device stability indicates the
119 Figure 5 9. (color) Typical J V characteristics of a P3HT:CdSe hybrid PV cell tested after fabrication and storage at the air for over two months under 1 sun A M 1.5G illumination. additional advantages of the ZnO compared to low work function metals such as Ca and Mg and insulating compounds such as LiF. The stability enhancement is mainly attributed to the hindrance of the penetration of moisture and oxygen fr om cathode side, which reduces the possibility of insulating CdSe surface, etching the PEDOT:PSS active layer interface, and generating an insulating Al 2 O 3 layer at the active layer Al interface. In addition, the physical isolation of the active layer from Al cathode also prevent s the potential chemical reaction of Al with the conjugated polymers and retards the degradation of the polymer Al interface. 173 5 .6 Summary This chapter describes the introduction of a solution processed ZnO nanoparticle layer to hybrid photovoltaic cells, resulting in both significantly enhanced efficiency and envi ronmental stability. Hybrid solar cells based on P3HT and CdSe nanoparticles with a ZnO NP layer show ~50 90% increase in p which is primarily contributed from an
120 increase in J sc and secondly from FF. Though only spherical CdSe nanocrystals and P3HT were tested in this study, the concept of inserting a solution processed ZnO NP cathode interlayer to enhance device performance is applicable to the devices using anisotropic CdSe nanocrystals, other conjugated polymers and beyond. The roles of the ZnO NP lay er in hybrid solar cells have been systematically studied. The ZnO NP layer mainly serves to smoothen the hybrid active layer to provide better contact with Al cathode, adjust the optical field distribution inside the active layer for more optimal absorpti on, prevent the hole leakage from the active layer to the cathode, assist in extraction of photogenerated electrons, and prevent the direct quenching of photogenerated exciton at the active layer Al interface. Furthermore, the ZnO NP layer also prevents the penetration of moisture and oxygen and the leakage of Al atom or cluster into the active layer, which together leads to a dramatic improvement in device environmental stability with ~70% efficiency retained after exposure to the air for over 2 months.
121 CHAPTER 6 EXTENDING SPECTRAL RESPONSE AND ENHANCING PHOTOVOLTAGE BY CONJUGATED POLYMERS 6 .1 Introduction Modern synthetic chemistry enables chemists to tailor organic small molecules and polymers with desired chemical structures and physical properties, w hich broadens the application of organic materials to the fields originally occupied by inorganic materials. Drastic progress in organic solar cells during the past several years is a very examp le of technological improvement driven by the development of s ynthetic chemistry. Most of the newly developed polymers primarily for organic solar cells can be also applicable in organic inorganic hybrid photovoltaic cells, provided that the energy levels of the polymers are matched with those of inorganic nanocrysta ls. In fact, to date, organic materials for hybrid solar cells are mostly directly borrowed from organic solar cells, rather than specially tailored for interfacing with inorganic nanocrystals. 18 Although there is only ~ 30 vol. % of conjugated polymer in the active layer of an organic inorganic hybrid solar cell, the polymer contribute s > 80% photocurrent generat ion due to the high absorption coefficient (~10 5 cm 1 ). Moreover, the open circuit voltage of organic based PV cells is generally proportional to the energy offset of acceptor LUMO level and donor HOMO level. 174 Thus, the design rule s for organic donor materials in term of el ectronic structure are as follows: a low energy gap to harvest most of the solar photons, and an appropriate energy level alignment with nanocrystal acceptor to facilitate efficient charge transfer and simultaneously create the highest energy offset with a cceptor LUMO level. 174 One common strategy to synthesize low gap polymers is co polymerizing an electron rich unit with an electron deficient unit,
122 Figure 6 1. (color) Chemical structures and s chematic energy level diagram of conjugated polymers. such as PCPDTBT 125 with energy gap of 1.45 eV as shown in Figure 6 1 and Figure 6 2. Due to the inter chain charge transfer, the high LUMO level of the electron rich unit (acceptor) and the low HOMO level of the electron deficient unit (donor) defin e the ultimate energy gap. Introducing a strong electron withdrawing unit or group like fluorine atom s into backbone of conjugated polymers is a means to reduce both the HOMO and LUMO levels with energy gap retained such as PB n DT FTAZ. 175 As shown in Figure 6 1 and Figure 6 2, compared to P3HT, PB n DT FTAZ has almost the s ame energy gap but lower HOMO level (5.36 eV). Bulk heterojunction photovoltaic cells critically rely on the uniform and nanometer scale accurate phase separation of donor and acceptor materials for efficient exciton diffusion and charge carrier collection 16 135 Organic inorganic hybrid
123 Figure 6 2. (color) UV Vis absorption spectra of conjugated polymers and C dSe nanocrystals in dissociated states. materials have an organic phase (organic small molecule or conjugated polymer) and an organic inorganic hybrid phase (colloidal nanocrystals having an inorganic core and a mon olayer of organic surfactant). The miscib ility of the organic component and the colloidal nanocrystal compon ent in a particular system is a great challenge for their vast difference in affinity with organic solvents. Moreover, the processing conditions for colloidal nanocrystals are relatively si milar, while those of conjugated polymers could be a huge difference. In addition, the molecular packing 122 of the organic component and the phase separation 176 of the organic inorganic hybrid materials are also dependent on the post processing treatment, such a s solvent or thermal annealin g. In this chapter, we endeavor to enhance photocurrent and photovoltage of organic inorganic hybrid photovoltaic cells through selecting conjugated polymers with different energy gaps and energy levels. Besides, we exemplify the effect of conjugated
124 Fi gure 6 3 (color) Tapping mode AFM topological (a, c, e) and phase (b, d, f) images of PCPDTBT:CdSe NP hybrids processed from different chlorinated solvent:pyridine mixtures. The volume ratio of chlorinated solvent/pyridine is 90/10. polymers on processing conditions in organic inorganic hybrid morphology and their corresponding device performance. In addition, as suggested in Chapter 5 the effect of a ZnO NP layer on hybrid PV cells will be further strengthened using the low gap polymer PCPDTBT. 6 .2 PCPDT BT:CdSe Hybrid Processing on Device P erformance 6 2.1 Phase S eparation of PCPDTBT :CdSe Hybrid Thin F ilms Phase separation of donor and acceptor materials is of significant importance for bulk heterojunction hybrid solar cells. 107 135 PCPDTBT:CdSe hybrid thin films were prepared using chlorinated solvent/pyridine co solvents (v/v 90/10) with concentration of
125 3 mg P CPDTBT and 30 mg CdSe nanoparticles. The AFM topological and phase images of t hese films are shown in Figure 6 3. Though all films show sub 100 nm size domains that consist of both PCPDTBT and CdSe nanoparticles as evidenced from phase images, the film rou ghness and uniformity of the domain are very different. The film processed from chloroform ( CF ) : pyridine co solvent shows a relatively wide domain size distribution (30 90 nm) and a high surface roughness (root mean square, R rms 9.2 nm) due to domain clustering. The film processed from chlorobenzene ( CB ) : pyridine co solvent shows smaller and more uniform domain of 30 50 nm and much smoother surface with an R rms of 2.7 nm. The domain size has been increased to 60 100 nm when processed from o dich lorobenzene (DCB) : pyridine co solvent that is larger than those processed from both CF and CB; but the R rms is 6.9 nm, even smaller than the CF film. The vast difference in hybrid film morphology is not only attributed to the hybrid solvent affinity but also the physical properties of co solvents themselves, particularly the solvent evaporation rate. CF has a boiling point of 62 C and a high evaporation rate of 11.6 with reference of butyl acetate as 1, which leads to fast evaporation of chloroform durin g spin coating and results in forming large domains; while pyridine has a boiling point of 115 C, 53 C higher than chloroform, which may result in formation of the small domains. Thus, the wide domain size distribution and high surface roughness of CF pr ocessed film are mainly due to fast evaporation rate of chloroform and large boiling point gap be tween chloroform and pyridine. CB has a high boiling point of 131 C with only 16 C gap to pyridine and evaporation rate of 1.1, which could result in narrow domain size d istribution and thermodynamically controlled domain size growth. The
126 even evaporation of CB and pyridine also led to smooth surface. DCB has an even higher boiling point of 180 C and lower evaporation rate of < 1 and possesses of 90% in volume in DCB co solvents, which enables thermodynamically controlled growth of domain and also results in narrow domain size distribution. The highe r R rms compared to that processed from CB co solvent may be due to the larger boiling point gap between DCB and pyridine. We further studied the phase separation of CB processed PCPDTBT:CdSe hybrid films using TEM. The thickness of the hybrid films w as var ied in order to probe phase separation in a fine r le ngth scale. As shown in Figure 6 4a, hybrid film with 10 20 nm thickness shows uniform distribution of nanoparticles in the whole film and nanoparticles tend to connect each other. If the hybrid film thic kness increased to ~30 nm, no nanoparticle aggregation has been observed yet a small degree of nanoparticle percolated network appeared. Further increasing the film thickness to ~ 50 nm and then ~ 80 nm (the device thickness) also does not lead to nanopart icle aggregation, but rather, the obvious formation of percolated network of nanoparticles. This nanoparticle phase has size 6 15 nm (1 2 times of individual nanoparticle size) and is uniformly distributed throughout the entire film. Such a high level of d elicate phase separation together with large interfacial area between the polymer and nanocrystals ensure efficient e xciton diffusion and subsequent carrier transport. In comparison, the early work by Huynh et al. 16 showed a significant extent of nanoparticle aggregation and inhomogeneous distribution of nanoparticles for P3HT:CdSe hybrid film processed from CF, which leads to large phase separation of polymers (>50 nm), much higher than the exciton diffusion length. The difference in phase separation between this work and the
127 early publication s may be attribute d to the use of processing solvents. The evaporation rate of CF may be too fast to form uniform polymer and nanocrystal phases, while CB has much lower evaporation rate that ensures the formation of uniform and energetically favored polymer and nanocrystal phases. Figure 6 4 TEM images of PCPDTBT:CdSe NP hybrids with different film thickness processed from CB :pyridine mixtures. (a). 10 20 nm; (b). ~ 30 nm; (c). ~ 50 nm; (d). ~80 nm. The average size of CdSe NPs: 6.8 nm, and scale bars: 20 nm. 6 2.2 Effect of P rocessing S olvents on Device Performance PCPDTBT:CdSe hybrid PV cells were fabricated using different processing solvents. The thicknesses of the active layer and the ZnO NP layer are ~80 100 nm and 25 30 nm as determined by ellipsometry, respectively; and the CdSe nanoparticles are 6.1 0.3 nm. Figure 6 5a shows the J V characteristics of the hybrid PV cells under 1 sun AM 1.5G simulated illumination. The photovoltaic performance parameters, and
128 Figure 6 5 (color) J V characteristics (a) and EQE (b) of hybrid PCPDTBT:CdSe NP PV cells with the active layers processed fro m different chlorinated solve nt/ pyridine mixtures. domain size and surface roughness of the active layers are summarized in Table 6 1. The CF device shows a J sc = 3.6 mA/cm 2 and a V oc = 0.72 V; while the J sc has been increased to 6.2 mA/cm 2 using DCB as processing solvent yet V oc is r elatively unchanged. The J sc has been further increased to 8.1 mA/cm 2 when CB used as the solvent, accompanying with a slight increase in V oc = 0.78 V. The EQE spectra as a function of incident light wavelength were collected and shown in Figure 6 5b. All three devices show photovoltaic response from 350 nm to >800 nm, but the values and spectra of EQE are quite different. Both the CB and DCB
129 devices have much higher EQE than the corresponds to the nanocrystal absorption, suggesting that a more uniform phase separation leads to better charge extraction through the nanocrystal network and polymer matrix. The corresponding enhancement in FF (from 36 % of CF device to 46% of CB and DCB devices) also confirms the benefit of increased phase separation. The EQE for the absorption) is considerably low, indicating the inefficient exciton dis sociation resulting from too large domain size or insufficient phase separation. On the contrary, t he CB processed film has the most uniform and delicate phase separation that leads to the much higher EQE at balanced morphology for e fficient exciton dissociation and charge carrier extraction. We also used mixed chlorinated solvents to process the PCPDTBT:CdSe hybrid films for device f abrication. As shown in Figure 6 6 and Table 6 1, J sc increases as the volume concentration of CB ( V CB ) increases in both solvent systems. The R rms values and domain size s of these films monotonically decrease as the V CB increases in both solvent systems. Furthermore, p follows the similar trends as J sc ; and FF is relatively unchanged for CB:DCB system but has a sharp increase as the V CB increases from 0% to 20% for CF:CB system. Hence, as expected yet empirical, hybrid films with the smoothest surface and the smallest and most uniform domain lead to the highest photocurrent, FF V oc and consequently the best p in hybrid PV cells. 6 .2.3 Effect of Annealing T emperature on Device Performance Post deposition thermal treatment is the other strategy to manipulate phase sep aration of donor and acceptor materials, which has been well demonstrated as a means to improve the performance of both organic and hybrid PV cells using P3HT as
130 Figure 6 6 (color) Plots of the dependence of J sc FF, p and R rms on the processing solv ent mix tures for hybrid PCPDTBT:CdSe PV devices. donor material. 61 134 However, the study by Heeger and co work ers showed that thermal annealing does not enhance the performance of polymer solar cells using the amor phous low gap polymer PCPDTBT. 135 Here we show that appropriate thermal annealing is still of great benefit for PCPDTBT:CdSe hybrid PV cells proces sed from CB system. 177 As shown in Figure 6 7, though device without t hermal annealing shows the lowest dark current, it also exhibits the lowest photocurrent and FF The J sc increases 2 3 times as the annealing temperature increases; while a decrease in J sc has been observed once annealing temperature increased to 180 o C. T he thermal annealing may lead to the further segregation of the nanocrystal and polymer components and thus
131 Table 6 1. Summary of performance parameter of PCPDTBT:CdSe hybrid solar cells processed by different solvent mixtures and related surface roughnes s and domain size of the active layer. Solvent mixture* J sc (mA/cm 2 ) V oc (V) FF (%) p (%) R rms (nm) Domain (nm) CF 3.6 0.716 36.1 0.9 9.2 62 21 CF:CB (80/20) 4.6 0.741 41.4 1.4 6.2 42 13 CF:CB (50/50) 4.9 0.724 43.1 1.5 6.9 39 11 CF:CB (20/80) 5.4 0.734 42.3 1.7 6.1 39 11 CB 8.1 0.783 45.7 2.9 2.7 39 8 CB:DCB (50/50) 7.7 0.731 44.5 2.5 4.7 46 12 CB:DCB (20/80) 6.9 0.726 43.5 2.2 6.1 52 13 DCB 6.2 0.724 45.7 2.1 6.9 74 14 Volume percent more defined nanocrystal and polyme r phases. Furthermore, the remaining solvent residues such as pyridine that can weakly bound nanocrystal surface can be removed upon annealing, leading to reduc ing the recombination centers for photogenerated excitons and charge carriers. Finally, the furt her removal of pyridine ligands together with the thermal activation may lead to more intimate connect between nanocrystals, resulting in better electron transport as indicated in the corresponding enhancement of FF 6 .3 Effect of the ZnO Layer on the PCPD TBT Device Performance 6 .3.1 Device Efficiency The study in chapter 5 unambiguously indicates both the efficiency and air stability of hybrid PV cells can be enhanced by including a solution processed ZnO NP layer between the active layer and the cathode. Here we extend to study this ZnO effect
132 Figure 6 7 (color) J V characteristics (a, inset: dark current) and EQE (b) of hybrid PCPDTBT:CdSe NP PV devices with the active layers thermally treated at different temperatures. The active layers were processe d from CB:pyridine mixture solvents. in hybrid PV cells using conjugated polymers rather than the archetypical P3HT and also demonstrate the compatibility of low gap polymers in hybrid PV cells for harvesting near infrared photons. Figure 6 8 shows the J V characteristics of hybrid PV cells using 6.8 nm CdSe particles but different conjugated polymers under simulated 1 sun AM 1.5G illumination. The PCPDTBT de vice without a ZnO layer shows J sc = 7.2 mA/cm 2 and p = 2.7%, both of which are higher than the P 3HT device with a ZnO layer, indicating the compatibility of the low gap polymer in hybrid PV cells. 167 The PCPDTBT
133 devic e with a ZnO NP layer shows J sc = 9.2 mA/cm 2 V oc = 0.77 V, FF = 0.49, and p = 3.5%, representing ~30% enhancement compared to the device without a ZnO layer. Note that the efficiency of the device with a ZnO layer is comparable to those of the devices using CdSe nanorods and tetrapods and PCPDTBT 178 Figure 6 8. (color) J V characteristics of hybrid PV cells based on 6.8 nm CdSe particles and different conjugated polymers under 1 sun AM 1.5G illumination. The PCPDTBT devices were also tested under variable illuminatio n intensity ( P 0 ), as shown in Figure 6 9. The V oc of the devices without and with a ZnO layer closely follows a linear relationship with lnP 0 as derived from the Shockley diode characteristic s 19 T he FF varies within 0.4 0.5 across the illumination range and the device with a ZnO layer has slightly higher FF, particularly at high P 0 The ratio of J sc to P 0 c orresponding to the external quantum efficiency, decreases ~30% as the P 0 increases for both cells, suggesting the presence of strong bimolecular recombination 179 and the addition of a ZnO NP layer does not alleviate this loss. The presence of the recombination and trapping sites in the nanocrystal surface / interface together with the
134 Figure 6 9 (color) Effect of a ZnO NP layer on the performance of hybrid PV cells using a low gap polymer as don or materi al tested under variable illumination intensity. relatively low mobilities of PCPDTBT and CdSe nanoparticles are the possible reason for such strong bimolecular recombination. 61 Hence, the device without a ZnO layer leads to a highest p = 2.8% at P 0 = 10 mW/cm 2 ; while the device with a ZnO layer yields a maximum p = 3.7% at P 0 = 40 mW/cm 2 both are slightly higher than those at 1 sun condition. The EQE has also been measured as a function of wavelength, as shown in Figure 6 10a. The d infrared photons. T which is of benefit for solar c ells based on low gap polymers and with low electron mobility The J sc yielded by integrating the EQE spectra with the standard 1 sun AM 1.5G spectrum for the devices without and with a ZnO layer are 6.6 mA/cm 2 and 8.7 mA/cm 2 respectively, which are withi n the 5 10% errors of the experimental results. The i nternal quantum efficiency which is the ratio of the numbers of the collected electrons to the numbers of absorbed photons, has also been calculated by dividing the EQE by
135 the to tal light absorption eff iciency. The total light absorption efficiency of the active layer was corrected based on the published method. 180 It is interesting to find the IQE of the device with a ZnO layer is higher than that of the device without a ZnO layer at almost entire photoresponsive range, suggesting the device with a ZnO layer is more efficient in charge transport / collection The high est IQE for the device with a ZnO layer is ~80%, suggesting it is possible to achieve very high efficiency hybrid solar cells. Additionally, the high IQE also enables to enhance device efficiency by managing light absorption In fact, t he early demonstrat ion using close packed transparent polymer microlens arrays to enhance light harvesting witnessed 20 30% increase in p 181 Figure 6 10 (color) Effect of the ZnO NP layer on t he external and internal quantum efficiencies of PCPDTBT:CdSe hybrid PV cells 6 .3.2 Role of the ZnO NP Layer The roles of the ZnO NP layer in hybrid PV cells have alr eady been discussed in chapter 5 which indicates the combination of electronic, optical, and morphological effects. Here we do not intend to over emphasize these effect s but rather, illustrate the specific effects upon the low gap polymer. Similarly, the ZnO NP layer smoothens the PCPDTBT:CdSe layer from R rms of 5.9 nm to 3 .0 nm, blocks the hole diffusion to the
136 cathode, and eliminates the direct contact at the hybrid Al catho de interface to prevent exciton quenching. In particular, the ZnO layer can shift the higher optical intensity spectral range within the active layer to long wavelengt hs. The optical field profiles calculated based on transfer ma trix theory 14 as shown in Figure 6 11 do show that the optical intensity of the PCPDTBT:CdSe hybrid active layer shifts to long wavelength when a Zn O layer included. This optical effect has been reflected in the EQE, which is mu ch higher at enables to tune the absorption of the active layer by varying the ZnO layer thickness and consequently the device photocurrent. This argument has been well demonstrated based on optical simulation with an assumption of unity IQE. Figure 6 11c shows the J sc of the device is as a function of both the thickness of the active layer and the t hickness of the ZnO layer. T he maximum J sc as predicted by optical simulation appears at positi ons where the thickness of an active layer is 70 100 nm and the thickness of a ZnO layer is ~20 nm, which is consistent with our experimental results. 6 .3.3 Device S tability The unencapsulated PCPDTBT:CdSe/ZnO device stability has also been tracked in a daily / wee kly manner. As shown in Figure 6 12, the J sc falls more sharply at the initial testing days and then gradually reaches to a stable plateau. The p follows the similar trend as J sc because both V oc and FF are relatively stable within the testing period. A closer observation reveals a slight increase in V oc which is attributed to the decrease in dark current during the tracking period. T he slight decrease in FF is likely due to the decrease in charge carrier mobilities originated from the morphol ogical change of the hybrid active layer or just the electron mobility decrease due to the insulating passivation of CdSe surface by moisture /
137 Figure 6 11 (color) Optical intensity profiles of the PCPDTBT:CdSe hybrid PV cells without (a) and with (b) a ZnO layer, and the calculated short circuit current density of the PCPDTBT:CdSe hybrid PV cells as a function of active layer and ZnO layer thicknesses (c). both the J sc and p retain ~70% of their original values, whic h is similar to that of the P3HT device. 171 The environmental stability enhancement is again mainly due to the prevention of moisture / oxygen by the ZnO layer as discussed in chapter 5 Figure 6 12 (color) Evolution of the photovoltaic performance parameters of an unencapsulated PCPDTBT:CdSe hy brid PV cell with a ZnO NP layer upon exposure to the ambient condition.
138 6 .4 E nhancing Photovoltage Using a Deep HOMO Polymer The open circuit voltage, another important factor determining the efficiency of a PV cell, is also tunable through tailoring the energy levels of conjugated polymers in organic based solar cells. 126 Here we demonstrated the V oc enhancement in hybrid PV cells by using a polymer with deep HOMO. As shown in Figure 6 1, a new copolymer PB n DT has a HOM O level of 5.36 eV, lower than 5.1 eV of P3HT and 5.2 eV of PCPDTBT. 1 75 I ts optical gap is 2.0 eV, only slightly higher than P3HT ( 1.9 eV) (Figure 6 2). Also PB n DT shows the same order of magnitude in hole mobility as P3HT (10 3 cm 2 v 1 s 1 ) as measured by SCLC in a BHJ structure. 175 Hence, these properties suggest that PB n DT is a promising donor material for BHJ polymer based solar cell s with potential high V oc and efficiency. Figure 6 13a shows the J V characteristic of hybrid PV cells based on CdSe nano particles (~6.8 nm) and the medium gap polymer P3HT or PB n DT. The PB n DT cell shows J sc as high as 7.9 mA/cm 2 V oc = 0.82 V, FF = 0.50 and p = 3.2%, indicating higher photocurrent and photovoltage than those of the P3HT cell s To the best of our knowledge, this is the first hybrid PV cell showing efficiency over 3% using a medium gap polymer. :CdSe system is ~1 .1 1.3eV (setting 7 nm CdSe CB as 3.8 4.0 eV), around 0.2 0.3 eV lower than that of the PB n DT :CdSe system (1.3 1.6V), suggesting there is more significant loss in photovoltage in the PB n DT :CdSe system and further optimization is necessary to realize its fu ll potential. The PB n of P3HT cell i s < 45% at this region (Figure 6 13b). Besides, the PB n DT cell has higher nm that mainly corresponds to the polymer absorption, sugge sting that PB n DT is more efficient in hole transport in the BHJ
139 Figure 6 13. (color) J V characteristics under 1 sun AM 1.5G illumination (a) and EQE (b) of hybrid PV devices based on CdSe nanorods and medium gap polymers. structure. This is also conf irmed by the pioneer report, in which PB n DT:fullerene solar 175 Compared to the P3HT cell, the PB n DT cell shows 0.06 V increase in V oc mainly due to the larger energy difference between nanocrystal LUMO and polymer HOMO. The V oc of PB n DT based hybrid PV cells can be as high as 0.89 V if CdS e nanorods were used as accepto r material ( due to the raising in nanorod CB, Figure 6 14). Such a high V oc together with rational J sc make it very promising for polymer based tandem solar cells.
140 Figure 6 14. (color) J V characteristic of a hybrid PV de vice based on CdSe nanorods and PB n DT. The nanorods have width 3.2 nm and AR10. The a ctive layer was treated with ethanedithol before Al deposition (see Chapter 7) The PB n DT:CdSe NR device was processed from chlorobenzene :pyridine co solvent. 6 .5 Summary In this chapter we demonstrate both the photocurrent and photovoltage of hybrid photovoltaic cells can be improved by customizing conjugated polymers and optimizing processing conditions. The processing of PCPDTBT:CdSe hybrid has been comprehensively stud ied, which indicates the processing solvent and annealing temperature have a profound impact upon the phase separation and ultimately the device performance. In particular, the surface roughness and domain size of nanoparticle based hybrid films are drasti cally dependent on the processing solvent. Besides, appropriate thermal annealing has also been shown to be of benefit in enhancing device performance, probably due to more optimal phase segregation and further removal of exciton and charge carrier recombi nation/trap centers.
141 The photoresponsive wavelength of hybrid PV cells has been extended to ~850 nm using the low gap polymer PCPDTBT. The concept of including a ZnO layer between the active layer and the cathode has been further proved in the hybrid PV ce lls based on PCPDTBT, which leads to J sc ~ 9 mA/cm 2 and p 3.5% under 1 sun AM 1.5G illumination. Similarly, the ZnO layer smoothens the active layer, serves as optical spacer, eliminates the direct contact between the hybrid and the cathode, assists in extraction of photogenerated electron, and prevents hole d iffusion to the cathode. In addition, the tracking study in a 2 month period indicates the PCPDTBT cell with a ZnO layer is more resistive to the moisture and oxygen penetration, resulting in ~70% efficiency retained. Moreover, we also demonstrate the impr ovement in the open circuit voltage of hybrid PV cells using a conjugated polymer with deep HOMO (PB n DT FTAZ). The PB n DT:CdSe NP cell shows V oc > 0.8 V and p ~3.2%, which for the first time, the hybrid PV cells based on a medium gap polymer and spherical CdSe nanoparticles show p >3%. The V oc reaches to as high as 0. 8 9 V when CdSe nanorods are used as the acceptor materials in hybrid PV cells, suggesting t he promising candidate in polymer based multi junction solar cells.
142 CHAPTER 7 ENGINEERING POLYMER NANOCRYSTAL INTERFAC E BY CHEMICAL TREATMENT 7 .1 Introduction Though significant efforts h ave been made to tailor both conjugated polymers and colloidal nano crystals, the performance of organic inorganic hybr id photovoltaic cells still lag s behind their all organic count erpart The reason for t his disparity lies in the fundamental difference between two materials systems In particular, t he complex surface of colloidal nanocrystals creates a polymer nanocrystal interface that governs the electronic and morphological interactions and consequ ently affects the performance of organi c inorganic hybrid solar cells. 182 The primary concern to introduce inorganic nanocrystals into organic solar cells is to enhance the electron affinity and transport, 33 which, to a large extent, has been compromised by the fact that the charge carrier mobility of colloidal nanocrys tals is dropped by several orders of magnitude to a level of typical organic semiconductors. 61 Again, such low mobility is attributed to the abundant surface and interface of nanocrystals involving organic surfactants. Other than that, the pr esence of these organic surfactants can also serve as recombination centers / traps for photogenerated excitons and charge carriers, which further deteriorates the photocurrent generation in a solar cell involving colloidal nanocrystals. Therefore, these c hallenges undermine the potential advantages of organic inorganic hybrid photovoltaic cells. In this chapter, we stress the importance of engineering the polymer nanocrystal interface to bridge the performance gap between hybrid solar cells and all organic counterparts. The different types of ligands present upon nanocrystal surface s can be selectively removed by various chemical treatment s In particular, the removal of
143 negatively charged X type ligands significantly improves the charge transport propertie s as well as reduces the electronic trap density which together lead to significant enhancement in the per formance of hybrid solar cells. 7 .2 Synthesis and Processing of CdSe N anorods CdSe nanorods were synthesized according to the published procedures. 30 Briefly, 0.205 g of CdO (Alfa Aesar, 99%), 2.90 g of TOPO (Sigma Aldrich, 99%), and 0.851 g of TDPA (PCI synthesis, 97%) were mixed and heated to 320 o C under N 2 flow. The mixture was maintained at this temperature to get a clear solution, and then naturally cooled to room temperature. After 24 h aging, the mixture was re heated to 320 o C under N 2 flow. Selenium stock solution (62 mg Se, Sigma Aldrich, 99%; 0.30 mL tributylphosphine (TBP), Sigma Aldrich, 90%; 1.0 mL TOP, Sigma Aldrich, 90%; and 0.30 mL toluene) was injected, and the reaction was kept for 40 min at 300 310 o C for nanorods with low AR ~2 and length 14 nm, and at 250 260 o C for nanorods with high AR ~7 and length 32 nm (Figure 7 1) The reaction was terminated by naturally cooling to room temperature within 10 15 min As prepared CdSe nanorods were purified by dissolving in toluene and precipitation with methanol 6 times. After purification, nanorods we re ligand exchanged in pyridine. CdSe nanorods (~80 mg) were mixed with 15 mL pyridine, and then refluxed for 24 h under N 2 flow. The nanorods were recovered by precipitation with hexane and centrifugation. Polymer:CdSe nanorod hybrids were prepared by dis persing 30 mg CdSe nanorods into 1 mL chloroform followed by mixing w ith 3.0 mg of P3HT or PCPDTBT.
144 Figure 7 1. TEM images of CdSe nanorods with different aspect ratio. 7 .3 CdSe Nanorods and Polymer:CdSe Hybrid Film C haracterization 7 .3.1 TEM M easurement The size of the CdSe nanorods was characterized using a JEM 2010F transmission electron microscope (TEM) with a 200 keV electron beam energy. CdSe nanorods in toluene solution were dropped on to a carbon coated copper grid and natu rally dried before measur ement. The hybrid thin film samples were prepared as follows for TEM measurement. First, P3HT:CdSe hybrid film was spin coated upon ITO substrates pre coated with a layer of PEDOT:PSS. The active layer was treated by ethanedithiol ( EDT ) in acetonitrile wit h volume ratio of 1% if necessary and then annealed at 120 o C for 20 min. After that, the hybrid films were immersed into deionized water to dissolve the PEDOT:PSS layer and make the hybrid active layer floated upon the water surface. The floated active la yer was picked up by the carbon coated copper grid. The TEM measurement was performed in a JEOL 200CX with accelerating voltage of 200 KeV.
145 7 .3.2 FTIR M easurement CdSe nanorods in chloroform solution with a concentration of 5 mg/mL were spin coated on top of glass substrates, and then treated with EDT or without any more treatment. Then nanorod films were annealed at 120 C for 20 min and re dissolved into chloroform and uniformly deposited onto a sodium chloride substrate. T he FTIR spectra were collected using a Perkin FTIR spectrometer For quantitative calculation, we chose the absorption spectrum of the CH 3 group, si nce both the TOPO and PA molecules have methyl group s Here we only consider TDPA for PA molecules and PPA excluded for relatively low concentration. Since only TOPO molecule s can be removed during pyridine exchange and remaining PA molecule s removed after EDT treatment and based on Lambert where A is is extinction coefficient, b is path length, and c is concentration, we can quantitatively determine the percentage of organic surfactants removed after each treatment. Th e concentration for purified sample and pyridine exchanged sample is the same and twice of that of EDT treated sample. 7 .3.3 NMR M easurement For 31 P NMR measurement, saturated solutions of the CdSe nanorods in CDCl 3 were used in all NMR recordings. For th e sample exchanged by pyridine, we collected the ligands by removing the precipitated CdSe nanorods and evaporating the solvents. The collected white solids were re dissolved in CDCl 3 for NMR study. For the ligands exchanged by the EDT, pyridine exchanged nanorods in chloroform were added by a small amount of triethylamine and EDT, resulting in the precipitation of CdSe nanorods and concomitantly the removal of ligands present upon nanorod surface. The
146 chloroform solution was evaporated and organic ligands were collected and re dissolved in CdCl 3 for NMR study. Each recorded spectrum is a result of 2048 scans. The NMR experiments were done for solutions with and without added phosphoric acid standard vessel, so that the chemical shifts for 31 P are referenced to the phosphoric acid standard ( 31 P at 0 ppm). 31 P NMR spectra were recorded on a Varian VXR 300 spectrometer (300 MHz). 7 .3.4 XPS M easurement XPS data were collected using a PHI 6100 X ray Photoelectron Spectrometer using a magnesium anode, a step size of 0.1 eV and a pass energy 17.9 eV. The signal to noise ratio was improved by accumulating data over 90 scans. The samples were prepared by spin coating CdSe nanorods (30 mg/mL in chloroform) upon Si substrates. For the samples with thermal treatment, no more EDT treatment was performed. The XPS peaks were corrected for sample charging by referencing the main carbon peak to 284.6 eV, and the experimental data were fitted with Gaussian Lorentian profiles. The binding energy of XPS peak has been corrected by referencing C 1s to 284.6eV. 7 .3.5 AFM M easurement AFM measurement was performed on the devices with a Veeco Innova scanning probe microscope operating in tapping mode. The hybrid films were deposited upon ITO substrates pre coated with a layer of PEDOT:P SS. The active layer was treated by EDT if necessary and then annealed at 120 o C for 20 min before measurement. 7 .4 Effect of EDT Treatment on D evice P erformance One dimensional nanorods have been regarded as an advantage to provide direct charge transport pathway s in electronic devices, which has already been
147 demonstrated in hybrid PV cells by several groups 16 170 182 and in Chapter 4 as well. Thus, we synthesized CdSe nanorods using the high temperature injection method alkylphosphonic acids and TOPO as main capping ligands and TOPO as solvent The volume, length aspect ratio (AR) of CdSe nanorods can be tailored by controlling the reaction parameters Here CdSe nanorods with two different aspect ratios were used for fabri cating hybrid PV cells (Figure 7 1). The polymer:CdSe hybrid PV cells have a structure of IT O/ PEDOT:PSS/ polymer:CdSe/ Al. The hybrid active layer was deposited by spin coating the chloroform solution composed of pyridine exchanged CdSe nanorods and conjugated polymers (P3TH or PCPDTBT) and have a thickness ~ 80 100 nm, and then was treated by a cetonitrile solution containing 1% EDT for 1 min. The devices were annealed at 120 o C for 20 min at a N 2 filled glove box before Al deposition. All the devices were tested in the air under simulated AM 1.5 G solar illumination. A phenomenon of time depende nt performance was observed for all the devices, which had also been observed in our previous study on P3HT:CdSe hybrid PV devices based on either spherical nanopa rticles 61 or anisotropic nanorods ( Chapter 4 ) The device performance would rea c h a plateau af stable at the plateau for several hours ( Figure 7 2 ), where all the measurements have been conducted. Similarly, this aging effect has been attributed to the adsorption of water and oxygen molecules to CdSe nanorods surface through diffusion, which leads to the surfa ce passivation and oxid ation. The effect of EDT treatment in enhancing the hybrid PV device performance is primarily evident in the J V characteristics under simulated 1 sun illumination as
148 Figure 7 2. (color) Air exposure time dependent J V characteristics of PCPDTBT:CdSe NR hybrid PV cells under dark and 1 sun AM 1.5G illumination. shown in Figure 7 3 a and the corresponding PV parameters listed in Table 7 1. The device based on P3HT and CdSe nanorods with AR ~ 7 shows a significant increase in J sc from 5.9 mA/cm 2 to 7.4 mA/cm 2 after the EDT treatment T ogether with a slight enhancement in both V oc and FF, the p has been increased from 2 .2% to 2.9% (Table 7 1) The efficiency of hybri d PV device can be further enhanced by using low gap polymers to harvest abundant infrared photons As shown in Figure 7 3b and Table 7 1, a hybrid PV cell based on a low gap polymer PCPDTBT and CdSe nanorods shows a p = 3.3% under 1 sun illumination W he n the active layer treated with EDT, a substantial increase in J sc from 9.3 mA/cm 2 to 12.8 mA/cm 2 and a slight enhancement in
149 both V oc and FF have been observed, leading to p = 4. 8 % under 1 sun illumination and after spectral mismatch correction. To the b est of our knowledge, this is a record efficiency for bulk heterojunction hybrid PV cells based on conjugated polymers and colloidal nanocrystals. 115 The hybrid PV devices were further characterized under variable illumination intensities ( P 0 ). As shown in Figure 7 3 for both P3HT:CdSe and PCPDT BT:CdSe devices, the ratio of J sc to P 0 corresponding to the external quantum efficiency, decreases as P 0 increases from 0.07 sun to 1.3 s un for the devices both without and with EDT treatment. This behavior suggests the presence of strong bimolecular rec ombination within the active layer and leads to the loss in photogenerated charge carriers. 179 In particular, at high P 0 t his loss becomes more serious due to the high carrier concentration and low charge c arrier mobility in the active layer. Manipulating this loss is a must and still a challenge for bulk heterojunction organic based solar cells. The EDT treatment leads to ~10% and ~35% increase in J sc / P 0 for the P3HT and PCPDTBT devices, respectively, sugge sting partial suppression of the bimolecular recombination 179 Partly offset by the increase in V oc of P 0 183 a maximum p of 2.6% and 3.2% both appeared at 0.07 sun, have been observed for P3HT:CdSe devices without and with EDT treatment, respectively. T he maximum p of 3.7% (at P 0 = 0.07 sun without EDT treatment ) and 5.2% (at P 0 = 0.12 sun with EDT treatment ) have been observed for the PCPDTBT devices which is comparable to the all organic PV cells based on the same donor PCPDTBT and PC71BM with additives during active layer processing. 135
150 Figure 7 3. (color) Performance enhancement in polymer:nanocrystals hybri d photovoltaic cells upon the EDT treatment. J V characteristics and Illumination power dependence of p ratio of short circuit current density ( J sc ) to P 0 V oc and FF for P3HT:CdSe (a) and PCPDTBT:CdSe (b ) hybrid PV devices with and without EDT treat ment. The external quantum efficiency spectra have been recorded to further evaluate the device performance upon EDT treatment (Figure 7 4 ) P3HT:CdSe device shows higher EQE at wavelengths from 350 nm < < 60 0 nm with EDT treatment and a maximum EQE of 5 6 375 nm. Meanwhile, the device based on PCPDTBT shows a b roader photoresponse range at 300 nm < < 850 nm due to its lower optical gap (1.45 eV). The EDT treatment leads to higher EQE in the entire photoresponsive spectral range for nm suggesting better balanced charge transport in the active layer In particular, the fact that the EDT treated devices show higher EQE at the short wavelength ( < 5 0 0
151 Table 7 1. Summary of photovoltaic performance under 1 sun AM 1.5 G illumination for polymer:colloidal nanocrystal hybrid photovoltaic cells. Polymer CdSe NRs EDT J SC (mA/cm 2 ) V OC (V) FF P (%) P P (%) P3HT Low AR (a) no 4.9 0.73 0.39 1.4 0.1 yes 7.1 0.71 0.51 2.6 0.2 86 PCPDTBT Low AR (a) no 6.6 0.72 0.39 1.9 0.1 yes 9.9 0.72 0.46 3.3 0.2 74 P3HT High AR (b) no 5.9 0.72 0.52 2.2 0.2 yes 7.4 0.73 0.54 2.9 0.2 32 PCPDTBT High AR (b) no 9.3 0.72 0.49 3.3 0.2 yes 12.8 0.74 0.50 4.7 0.3 42 (a) Low AR = 2 (length: 14 nm, width: 6.4 nm); (b) High AR = 7 (length: 32 nm, width: 4.4 nm). nm ) region indicates electron transport has been significantly improved, since the high energy photons are more prone to be absorbed close to the anode side in a normal structure cell 177 The performance of hybrid PV devices is also dependent on CdSe na noro Figure 7 4 (color) Effect of the EDT treatment on EQE of hybrid polymer:CdSe NR PV devices.
152 length and AR 16 partly because nanorod s form the matrix for charge transport and their surface s act as exciton dissociation and charg e transfer sites. When nanorods with length 14 nm and AR ~2 used a s acceptor, as shown in Figure 7 5 and Table 7 1, P3HT:CdSe hybrid PV device shows p = 1.4%, ~50% lower than those devices based on nanorods with AR ~ 7. However, when the active layer treat ed with EDT, the device exhibits a substantial improvement in J sc from 4.9 mA/cm 2 to 7.1 mA/cm 2 and also in FF from 0.39 to 0.51, which leads to ~85% enhancement in p significantly higher than 32% enhancement for those devices with high AR nanorods. Simi larly, when low AR nanorods blended with PCPDTBT, the device without EDT treatment shows p = 1.9% together with low J sc and FF ; however, the device with EDT treatment witnesses a sharp increase in J sc and FF thus resulting in 73% improvement in p The r elative low performance for low AR nanorod devices is likely due to the poor charge transport and much richer surface chemistry that may result in more exciton and charge carrier recombination sites. The more significant enhancement in device performance w ith EDT treatment also suggests the effect of the nanorod surface properties. Figure 7 5 (color) J V characteristics (a) and EQE (b) of P3HT:CdSe NR PV devices based on shorter nanorods (AR: ~2).
153 The effect of chemical treatment on device performance has been further investigated by using different additives. Here we select two other additives to perform the same chemical treatment as EDT during device fabrication. One additive is 1, 8 octanedithiol (ODT) that has a similar chemical structure as EDT bu t with longer alkyl chain ; the other one is 1,8 diiodooctane (DIO) that has the same alkyl chain length as ODT but has anchoring group iodine. The J V characteristics of hybrid PV cells based on CdSe nanorods with AR ~5 and under various chemical treatme nt were sho wn in Figure 7 6a. The device with the EDT treatment shows increased J sc and FF under 1 sun illumination compared to the device without treatment, resulting in a relative 20% enhancement in p The device with the ODT treatment shows a significa nt decrease in J sc and FF compared to the untreated device. Though there is a slight increase in V oc the p has been reduced by 25%, suggesting poorer charge transport with long chain additive. The device performance has been further decreased when DIO u sed as treating agent, with less than 50% retained in p It is very likely that DIO just exists as foreign additive within the active layer or weakly bound to the nanorod surface, and it is not likely that the as bound ligands could be exchanged by DIO du e to its relatively weak affinity. Thus, rather than acting to passivate nanorod surface, DIO may act as recombination centers and/or insulating si tes to hinder charge transport. The EQE of the hybrid PV cells using different additives has also be en record ed as shown in Figure 7 b. Similar to the device based on nanorods with AR ~7, the device with EDT treatment shows an enhancement in EQE at the entire wavelength range compared to the untreated device. However, t he ODT and DIO devices show significant decre ase in EQE in the entire wavelength range, which is consistent with the decrease
154 Figure 7 6 (color) J V characteristics (a) and EQE (b) of P3HT:CdSe NR PV devices with various chemical treatment. in photocurrent reflected in the J V characteristics The uniform increase and decrease in EQE with chemical treatment further emphasizes the importance of the polymer nanocrystal interface, which is the junction for exciton dissociation and the path for charge carrier transport. 7 5 Effect of EDT T reatment on Nanorods and Hybrid Films 7 .5.1 UV Vis A bsorption To understand the nature of the EDT treatment on the hybrid materials and the effect on the device performance, a series of optical, electrical, chemical, and structural characterization on hybrid mater ials have been carried out. Figure 7 7 shows the UV Vis absorption spectra of P3HT:CdSe and PCPDTBT:CdSe hybrid films. The absorption of P3HT:CdSe and PCPDTBT:CdSe hybrid films is the superposition of the absorption of the individual polymer and CdSe. When treated by EDT, these films do not show apparent change in absorptio n, which suggests the EDT treatment does not result in changes in the bulk properties of the active layer, such as material gain / loss or
155 structural ordering Again, any potential change s are likely concerned with the interfaces between the mate rials. Figure 7 7 (color) UV Vis absorption spectra of polymer:CdSe hybrid films. 7 .5.2 Hybrid Film M orphology The morphologies of the PCPDTBT:CdSe hybrid film s wi thout and with EDT treatment we re also pro bed by tapping mode AFM. As shown in Figure 7 8, large domains with size ~ 200 600 nm are observed for the film without the EDT treatment (Figure 7 8 a). These domains do not correspond to either pure polymer or nanocrystal domains as suggested by the uniform phase image (Figure 7 8 b). The close up images shown in the inset reveal that each large domain consists of many smaller domains with size of ~ 30 50 nm and these small domains are compose d of both PCPDTBT and CdSe nanorods The hybrid l ayer shows no noticeable change in the surface topology and the root mean square surface roughness remains at R rms = 10 nm after the EDT treatment (Figure 7 8c) which is different to the appearance of extra pin holes and micro cracks in PbS quantum dot th in film after treatment by small molecules 184 185 and also different to polymer solar cells proces sed with additives that leads to drastic
156 Figure 7 8 (color) Surface m orphology of PCPDTBT:CdSe hybrid films Tapping mode atomic force microscopy (AFM) topographies and their corresponding phase images of PCPDTBT:CdSe NR hybrid films without EDT treatm ent ( a, b ) and with EDT treatment (c, d) (scale bar: 1 m). Inset: close up topographical image and the corresponding phase images (scale bar: 200 nm). change in phase separation 135 186 However, the phase image (Figure 7 8 d) does show an enhanced contrast for some regions among the small domains, suggesting a more rigid surface after the EDT treatment that may be due to the further removal of organic ligands from Cd Se surface (de tails in section 7 .6) To further investigate the possible morphological change, we employed TEM to image P3HT:CdSe hybrid films As shown in Figure 7 9 a c there are a number of white and black areas spreading across the entire image. The e nlarged image indicates that these white and black areas are the P3HT:CdSe hybrid but have different nanorod loading concentrations with dark region having higher nanorod loading (Figure 7 9 b c). CdSe nanorods are uniformly spreading over the hybrid film and not nanorod
157 Figure 7 9 TEM images of P3HT:CdSe hybrid films without (a) and with (b) the EDT treatment. aggregation has been observed. When E DT treated, as shown in Figure 7 9 d f, there is no significant difference in morphology compared to the un treated samples. Neither higher degree of nanorod segregation nor phase separation has been observed for the treated samples. In addition the open structure of hybrid film as clearly indicated i n the magnified images has the advantage of EDT penetration i nto the inside for more efficient removal of the capping ligands (Figure 7 9 c, & e) 7 .5. 3 Chemical Treatment on Surface Chemistry of CdSe Nanocrystals The EDT treatment does have a profound impact on the surface chemistry of CdSe nanorods as verified by FTIR, NMR, and XPS characterization. Organic surfactants TOPO and tetradecylphosphonic acid ( TDPA ) were involved during synthesis, which may bond to the nanocrystal surface to affect nanocrystal growth during synthesis and to stabilize nanocr ystals after s ynthesis. 27 28 Figure 7 10a shows
158 the FTIR transmittance spectra of CdSe nanorods purified after synthesis, ligan d exchanged in pyridine, and treated by EDT. The absorption peaks at 2921 cm 1 and 2847 cm 1 are due to the C H stretching vibration in CH 3 groups from either TOPO or TDPA. 187 The intensities of these absorption peaks are decreased after ligand exchange in pyridine and nearly vanish after EDT treatment, suggesting the alkyl chain ligands can be partly removed both by ligand exchange with pyridine and by EDT treatment. Also the presence of the abs orption peaks at 1098 cm 1 and 931 cm 1 that correspond to the stretching vibration of P=O and P O, respectively, suggests the abundance of TOPO and TDPA on the nanocrystal surface. These phosphor containing ligands are only partially removed after ligand exchange with pyridine, while they are more effectively removed by the EDT treatment. The cleavage processes of organic ligands upon CdSe surface have been further investigated by phosphor nuclear magnetic resonance ( 31 P NMR). The purified CdSe nanorods di spersed in deuterochloroform (CDCl 3 ) with concentration ~50 mg/mL does not show any distinctive signal in 31 P NMR. This behavior may be explained by the fact that the nanorod concentration (or the phosphor insufficient or the signal of the bounded ligands may be broadened due to the slow rotational correlation time of the nanorod s. However, as shown in Figure 7 11, after ligand exchange with pyridine, the exchanged ligands collected after precipitating nanorods do show a very sharp peak at 50.8 ppm, which can be assigned to the free TOPO. 56 188 191 The appearance of TOPO al so suggests it cannot be fully removed by common purification with toluene dissolution and methanol precipitation. Pyridine exchanged nanorods in CDCl 3 have been further treated by the EDT with a small
159 amount of triethylamine (TEA) as a Lewis base (acetoni trile is not very miscible in CDCl 3 ) The nanorods were aggregated and precipitated after this treatment that was isolated by centrifugation, and the CDCl 3 solution was used for NMR characterization. Th e EDT exchanged ligands exhibit three sharp peaks loca ted at 2 6.7 ppm, 21.5 ppm and 18.7 ppm ( Figure 7 11 ) The peaks at 26.7 ppm and 18.7 ppm can be assigned to the phosphors in TDPA and in P P (di n tetradecyl) dihydrogen pyrophosphonic acid (PPA, created by the condensation of two TDPA molecules at high temperature or as impurity in source materials ), respectively T he peak at 21.5 ppm may be due to the formation of either TDPA TEA salt or PPA TEA salt or an unknown alkyl phosphonic Figure 7 10. (color) FTIR spectra of CdSe nanorods upon various treatm ent s (a). Full spectra, (b) absorption spectra of CH 3 group.
160 acid specie s originating from the s ource materials. 56 188 191 Thus, the NMR study indicates that pyridine TOPO that typically bonds to the nanocrystals through weak van der Waals interaction; and the EDT treatment further remove the TDPA and PPA that usually bonds to the nanocrystal surface through ionic or covalent interaction. Figure 7 11. (color) 31 P NMR spectra of ligands exchanged by pyridine (black curve) and ligands exchanged by EDT (red curve) from nanorods. The selective removal of surface ligands by chemical treatment enables us to quantitatively estimate the population of each type of ligands. Here by integrating the absorption peaks of methyl group at 2921 cm 1 and 2847 cm 1 in FTIR spectra and based on Lambert Beer law, 187 6510% of organic ligands (primarily TOPO) are removed after ligand exchange with pyridine and 9010% of total ligands can be cleav ed after further EDT treatment. Though FTIR and NMR characterizations give a clear picture that surface ligands on the nanocrystals can be removed step by step by pyridine exchange and then by EDT treatment, the detailed cleavage processes and the chemical bonding nature are still uncle ar. Here XPS was employed to probe the chemical stat es of various elements.
161 Figure 7 12 a shows the C 1s XPS spectra for nanorods processed using various conditions. The intensity of the C 1s peak is reduced approximately by one half after ligand exchange i n pyridi ne. Additional EDT treatment without thermal annealing results in a slight increase in the C 1s intensity, which could be attributed to the absorption of EDT molecule to the CdSe surface. F urther thermal annealing of the EDT treated sample leads to a sharp reduction in the C 1s peak intensity, suggesting the removal of most organic ligands from the nanocrystal surface. The P 2p spectra appear to indicate a chemical shift of 0.5 to 0.8 eV between the purified, pyridine exchanged, and EDT treated nano crystal samples (Figure 7 12b) ; though the broad spectral distribution makes the result somewhat difficult to define. It is very likely that the chemical environments of these ligands have been changed after chemical treatment as the peak of fitted curve shifts to higher binding energy that may correspond to the free alkylphosphonic acid species (Figure 7 12b, blue curve) However, it is more conclusive for the EDT treated sample after thermal annealing, where the XPS spectrum clearly shows the disappearan ce of P in the nanorod film, suggesting the removal of phosphor containing ligands and being consistent with the FTIR result. Thus, the XPS results suggest that the EDT treatment process first results in the release of the alkyl phosphonic acids from the na nocrystal surface, which are then removed from the fil ms following thermal annealing. Other than removing X type ligands from nanocrystal surface with the EDT treatment, surface passivation is also envisioned for the strong affinity of mercapto group to th e nanocrystals. It is challenging to investigate using XPS due to overlapping of the binding energies of S 2s and 2p states with those of Se 3s and 3p states,
162 respectively. A s shown in Figure 7 12 c & d, the Se peaks at 160.0 eV (Se 3p 3/2 ), 165.5 eV (Se 3p 1/2 ), and 229.0 eV (Se 3s), which match well with the published XPS data for CdSe, 192 can be clearly identified. However, we can also identify peaks at 162.1 eV, 163.8 eV, and 227.1 eV, which correspond to S 2p 3/2 2p 1/2 and 2s states, respectively. Moreover, the binding energies of the S 2s and S 2p 3/2 states are 0.5 to 1 eV higher than the reported values for CdS but are ~1 eV lower than those for organic thiols. The se peaks are matched with those of metal thiolate. 193 The surface S to Se ratio is estimated to be 3010% based on the individual peak intensities. Thus, the EDT treatment results in the passivation of CdSe s urface with a chemical state of Cd thiolate, rather than either CdS or EDT molecule. Figure 7 12. (color) XPS high resolution spectra of CdSe nanorods upon various treatment. (a). C 1s, (b). P 2p, (c). Se 3p and S 2p, and (d) Se 3s and S 2s.
163 7 .5.4 Charge Transport of Hybrid F ilms The chemical characterization gives a clear picture of the removal of surface ligands, which could result in the change in charge transport properties of hybrid films Here we fabricated electron only and hole only devices to det ermine the charge transport behaviors of PCPDTBT:CdSe hybrid films without and with the EDT treatment. The electron only device has a structure of Al/ active layer/ Al and the hole only device has a structure of ITO/ active layer/ Au. The J V characteris tics of the se devices are shown in Figure 7 13. According to (eq.3 7) and inserting the parameters the same as used in Chapter 4 it is shown that the electron mobility has been increased from e = 310 6 cm 2 /Vs to 610 5 cm 2 /Vs after the EDT treatment. T he hole mobility remains at h = 110 5 cm 2 /Vs after the EDT treatment, suggesting the EDT treatment mostly has an impact upon the nanocrystals. 7 .6 Nature of the EDT Treatment on Device Performance The various chemical characterizations clea rly identify the existence of charged X type ligands bounded to the CdSe surface. These X type ligands have been regarded to form surface defect states that could quench exciton s and trap charge carriers, which in turn result in the loss of photocurrent an d photovoltage in a hybrid PV cell. Removal of these ligands from the nanocrystal surface and simultaneous passivation by formation of Cd thiolate upon the EDT treatment lead to the removal of these surface defect states, which is certainly of benefit in r educing the exciton and charge recombination loss at the polymer nanocrystal interface. This benefit has been well reflected in the huge enhancement in J sc and also slight increase in V oc of the hybrid PV cells with EDT treatment. Furthermore, the increa se in electron mobility of hybrid film with the EDT treatment also confirm s the reduction in charge recombination, which
164 Figure 7 1 3 (color) J V characteristics of electron only devices with a structure of Al/PCPDTB T:CdSe/Al using CdSe nanorods (AR: ~7) (a) and hole only devices with a structure of ITO/PCPDTBT:CdSe/Au (b) The dashed blue results in more favored collection of photogenerated charges as reflected by the increase in J sc and FF of the EDT trea ted PV devices. The higher level of performance enhancement for the devices with low AR nanorods can be also explained by the improvement in charge transport. Due to their short lengths, charge carriers along nanorods are more prone to be trapped and reco mbined for the inefficient transport that leads to relatively lower overall device performance. 16 R emoval of these charge recombination centers upon the EDT
165 treatment reduces the trapping of the photogenerated charges, thus resulting in m ore significant impact on the device performance. The effect of the additives on the device performance can also be attributed to the charge transport. Similar to the EDT treatment, though the ODT treatment can also remove the X type ligands, its long alky l chain hinders the electronic interaction and decreases charge transport, thus resulting in low J sc and FF. The X type ligands are not able to be removed by the DIO treatment. Together with its long alkyl chain, the DIO treatment could more seriously dete riorate the charge transport, as reflected in the device performance of the hybrid PV cells. 7 .7 Surface Chemistry of CdSe N anorods S urface chemistry study enables us to give a clearer picture in a CdSe nanorod As shown in Figure 7 14, the nanocrystals ge nerally have a composition of a CdSe core, a monolayer of Cd 2+ shell, and a monolayer of mixed L type (TOPO, TOP, and TOPSe) and X type ligands (alkyl phosphonic acids). The L type ligand s bond to the CdSe surface s through weak van der Waals interaction, w hile the X type ligand s bond to Cd 2+ through much stronger Coulombic interactions. In particular, an X type ligand with a single phosphonic acid group such as TDPA preferentially bonds to Cd cations as a monodentate hydrogen phosphonate, and PPA with two p hosphonic acid groups tends to bond Cd cations as bidentate hydrogen phosphonates. The neutral molecule TOPO preferentially binds to the Cd sites through oxygen rather than Se based on the ab initio calculation. 55 The preferential facets for TOPO binding are (11 0) and (01 0) facets with binding energies of 1.23 eV and 1.37 eV, respectively. These two facets are also the preferential binding facets of PA molecules. The ab initio calculation also
166 Figure 7 14. (color) Schematic illustration of surface chemistry of colloidally synth esized CdSe nanorods. indicates that binding of PA molecule to CdSe is much stronger than that of TOPO and this bindi ng is also through the oxygen atom s The ligand exchange interaction in pyridine, as revealed by 31 P NMR, only removes the L type ligands. Based on the chemical characterization results and the study on the surface chemistry of colloidal nanocrystals by several other groups 56 188 191 we propose the following cleavage mechanism of X type ligands (phosphonic acid species) from CdSe surface upon the EDT treatment. As shown in Figure 7 15 a free EDT molecule adsorb s to the CdSe surface bonds weakl y with a Cd cation, and weakens the S H bond in EDT. In a concerted process with the assistance of a Lewis base acetonitrile (also used as the solvent), the adsorbed EDT molecule deprotonates into EDT anions. Third, th e strong nucleophilic EDT anion attack s X type ligand bounded Cd cation (Cd X), following by replacing X type ligands and simultaneously the formation of Cd thiolate. The overall reaction is that alkyl phosphonic acid specie s are exchanged by EDT and Cd 2+ reacts with EDT to form Cd thiolate ( Fi gure 7 15 ) This reaction is a typical nucleophilic substitution reaction.
167 Figure 7 15. (color) Prop osed cleavage mechanism of alkylphosphonic acid molecules from CdSe nanocrystals upon the EDT treatment. 7 .8 Summary Interface engineering has been demons trated to be of critical importance in enhancing the performance of organic inorganic hybrid photovoltaic cells. The power conversion efficiency of the hybrid PV cells based on CdSe nanorods and low gap polymer PCPDTBT under AM 1.5 G illumination reaches t o a record level of 5% upon EDT treatment, which is comparable to the state of the art organic counterparts using fullerene derivative and PCPDTBT as the active layer. In general, depending on the conjugated polymers and nanorod size, ~30 90% enhancement in p has been observed for polymer:nano crystal hybrid PV cells with EDT treatment.
168 A series of optical, electrical, chemical, and structural characterization have been performed to understand the nature of the EDT treatment on device performance and surface chemistry of CdSe nanocrystals. The results indicate that the EDT treatment results in the removal of the charged X type ligands and simultaneous passivation of surface defects by the formation of Cd thiolate, which leads to reduced recombination and charg e trapping sites upon the nanocrystal surface and i mproved electron transport. In addition, the surface chemistry of CdSe nanocrystals after each step of chemical treatment has been depicted; and the reaction mechanism of X type ligand bounded Cd Se nanocry stals with EDT has also been proposed as a nucleophilic substitution reaction
169 CHAPTER 8 GRAFTING CONJUGATED OLIGOMERS TO COLLOID AL NANOCRYSTALS 8 .1 Introduction Chapter 7 strengthens the importance of interface engineering by chemical treatment in achiev ing high efficiency organic inorganic hybrid photovoltaic cells. The development in synthetic chemistry together with the unsaturated surface nature of colloidal nanocrystals make s it possible to directly interface organic donor with inorganic acceptor. 194 196 As shown in Figure 8 1, s uch an organic in organic bond could be a relatively short diffusion length in organic based system. The molecular interaction transport; and the photogenerated electron can be exclusively extracted through the n anocrystal phase. D irect grafting of organic molecule to colloidal nanocrystals may carry additional advantages compared to the typical ligand exchange method for removing the alkyl chain ligands on the nanocrystal surface originated during synthesis Firs t, the traditional ligand exchange using small molecules such as pyridine results in the remaining of small molecules upon nanocrystal surface, while the direct grafting only leaves the donor materials. Second, the ligand exchanged colloidal nanocrystals a re typically less dispersible in organic solvents, 61 which leads to unfavorable phase separation when blended with conjugated polymers; while the organic grafted nanocrystal is a supermolecule that has similar processing robustness as individ ual organic molecule s Third, the absorption of either organic or inorganic occurs close to the organic inorganic interface, which significantly reduces the exciton diffusion path
170 Figure 8 1. (color) Schematic drawing of colloidal nanocrystals grafted b y conjugated oligomers through strong chemical interaction. The grafting enables direct charge transfer at the organic inorganic interface immediately after exciton generation. The molecular interaction among conjugated oligomers also enables the packing o f oligomers to form a hole transporting channel. (Figure 8 1) In addition, the anchoring group at the organic inorganic interface can be tailored to facilitate the charge transfer and transport and reduce recombination traps. In this chapter, we attempt t o engineer the organic inorganic interface by directly grafting photoactive conjugated oligomers to colloidal nanocrystals. The grafting of oligomers to nanocrystals has been studied by various characterization techniques. Hybrid PV cells based on these ol igomer grafted nanocrystal supermolecules have also been fabricated and evaluated. 8 .2 Design and Properties of Functional O ligomers Though small molecules or oligomers have the advantages of defined chemical and conformational structures and reproducibili ty they are generally used in vacuum deposited organic photovoltaic devices, partly because these small molecules tend to crystallize and/or aggregate, rather than form bi continuous network with acceptor materials when processed into thin film s using wet chemistry This situation changes as wet chemistry processing approach improves. Recently, organic solar cells based on
171 Figure 8 2 (color) Chemical structures of phosphonic acid functionalized oligomers small molecule and PC70BM with p = 6.7% have been reported using a small percent of solvent additive during the film formation. 117 Here we design three monofunctionalized oligomers with defined chemical structure for interfacing with nanocrysta ls for hybrid PV cells (Figure 8 2 ). 197 The oligomers are asymmetric conjugated molecules with a phosphonic acid anchoring group at one end, which has the strongest affinity with CdSe nanocrystals in both the theoretical and experimental perspectives (Chapter 7) The energy gaps and HOMO and LUMO levels of the oligomers can be tailored by varying the conjugation degree and nature of the aromatic rings. The OPE A based on phenylene ethynylene typically has large energy gap, with similar properties to the PPV derivatives. The T6 A based on thiophene mimics the well known P3HT with good crystallinity high hole mobility and lower gap than OPE A. The T4BTD A introduces th e donor accept concept for designing low gap organic molecules: two electron rich thiophene groups and one electron deficient BTD group are linked together through cross coupling reaction.
172 Figure 8 3 (color) UV Vis absorption spectra of functional oligo mers and 6 nm CdSe particles. The UV visible absorption spectra of the oligomers in solution were shown in Figure 8 3 The oligomers exhibit wide absorption wavelength bands, the peaks of which locate at 323 nm and 379 nm for OPE, 428 nm for T6, and 360 nm and 510 nm for BTD, respectively. The optical gaps of these oligomers, determined based on the onset of the absorption, are relatively large for OPE (2.8 eV) and T6 (2.4 eV); while the gap has been reduced to 2.0 eV for T4BTD due to the energy mixing of t he BTD acceptor unit with thiophene donor unit. E nergy levels of the oligomers determined by cyclic voltammetry ( CV ) and UV Vis absorption were shown in Figure 8 4 In general, the oligomers have suitable energy levels to form a type II st aggered structure with nanocrystals. Both OPE and T6 show very shallow LUMO levels that could prevent electron leakage from acceptor; while the OPE HOMO is very deep, close to the nanocrystal CB, which may result in significant charge recombination due to hole leakage from OPE to nanocrystals. The HOMO and LUMO levels of T4BTD move downward significantly while they are still aligned with nanocrystals to create enough energy offset (> 0.3 eV) for charge transfer.
173 Figure 8 4 (color) Schematic energy level diagram of conjugated oligomers and CdSe nanocrystals. 8 .3 Oligomer grafted N anocrystal H ybrids The oligomer grafted nanocrystal hybrids were prepared by mixing oligomers and nanocrystals in solution for 2 h and then purified to remove the un grafted olig omers through several cycles of dispersion in good solvent (chloroform) and precipitation in poor solvent (methanol). 197 The resulting oligomer:CdSe hybrids ha ve characteristic colors of the oligomers in solution. This grafting was firstly confirmed by thermogravimetric anal ysis (TGA), as shown in Figure 8 5 The native nanocrystals contain ~18% organic ligands corresponding to the OA and TOPO; while the graftin g leads to ~4% 8% increase in organic content (22%, 26%, and 24% for OPE, T6, and T4BTD, respectively). A higher organic content is likely due to higher molecular weights of oligomers (OPE A: 815 g/mol; T6 A: 827 g/mol; T4BTD A: 797 g/mol; OA: 282 g/mol; TOPO: 415 g/mol). However, the degree of ligand exchange and the unknown distribution of the native ligands upon nanocrystal surface make it difficult to quantitatively estimate the number of oligomers based on TGA results.
174 Figure 8 5 (color) TGA of as synthesized and oligomer grafted CdSe nanoparticles. Fluorescence quenching experiments were also carried out to monitor the grafting. 197 198 Fluorescence of oligomers with a phosphonic acid group can be rapidly quenched with addition of CdSe nanoc rystals and vice versa (Figure 8 6 ); while the FL quenching of oligomers with an ester group is less efficient b y adding nanocrystals or vice versa (Figure 8 6 ). It is also suggested, due to the dual FL quenching of OPE A:CdSe and T6 A:CdSe, the quenching occurs via a charge transfer process rather than direct energy transfer. Through FL quenching experiments, it ca n be estimated that each nanocrystal has been covered by ~ 50 conjugated oligomers. Transmission electron microscop y is also employed to study the oligomer CdS e grafting. As shown in Figure 8 7 the native nanocrystals are monodisperse in size and uniforml y distributed; while the oligomer grafted nanocrystals tend to aggregate, which may be due to the strong intermolecular interaction of oligomers among neighboring nanocrystals.
175 Figure 8 6 (color) Evolution (a) of the fluorescence of OPE E and OPE A upon addition of CdSe NPs into the solution, and evolution (b) of the peak fluorescence intensities for the ester (squares) and acid (circles) forms of OPE (black line), T6 (blue line) and T4BTD (red line) upon incremental addition of CdSe NCs. 8 .4 Oligomer gr afted nanocrystal H ybrid Photovoltaic C ells To test the validity of the oligomer grafting nanocrystal hybrids for photovoltaic application, we then fabricated oligomer:CdSe hybrid PV cells. The hybrids were spin coated from CF solution with a thickness ~60 80 nm and then a ZnO NP layer was deposited to smoothen the active layer, prevent hole leakage, facilitate electron extraction, and manage light absorpti on as suggested in Chapter 5 and Chapter 6 The active layer and the ZnO layer were co annealed at 150 o C for 30 min at a glove box before Al deposition. The final device structure was show n in the inset of Figure 8 8 a. The photovoltaic effect has been observed for all oligomer:CdSe cells, and the J V characteristics of the oligomer: CdSe cells are shown in Figure 8 8 a. The OPE:CdSe cell show a J sc of 0.3 mA/cm 2 V oc of 0.41 V, and p of 0.03%. The low J sc is likely and partly due to the narrow absorption wavelength range of OPE and relatively low absorption coefficient of CdSe nanocrystals. The J sc has be en increased when a
176 Figure 8 7 TEM images of as synthesized and oligomers grafted CdSe nanocrystals. lower gap oligomer T6 was used, while the T6:CdSe cell shows very small V oc (0.29 V), which may result from unfavorable large phase separation of the h ybrid. The low gap oligomer T4BTD leads to the further increasing in J sc and a high V oc (0.68 V) has been observed in this cell, suggesting promising photovoltaic application of the T4BTD:CdSe hybrid system if the processing conditions were optimized. The EQE of these oligomer:CdSe cells were also collected and shown in Figure 8 8 b. The OPE:CdSe cell shows two EQE peaks of 10% at and a small shoul The EQE of the T6:CdSe cell is much broader and higher than the OPE cell; and the T4BTD:CdSe cell exhibits the broadest and highest EQE. The two EQE bands centered at 370 nm and 505 nm correspond to the T4BTD absorption. Again, a small shoulder at
177 Figure 8 8 (color) J V characteristics (a) under 1 sun AM 1.5G illumination and EQE (b) of hybrid PV cells ba sed on oligomers grafted nanocrystals. both oligomers and CdSe nanocrystals contribute to the photocurrent generation in all oligomer grafted CdSe hybrid solar cells. T he efficiencies of these oligomer:CdSe cells are much lower than their polymer counterparts (Chapter 5) The absorption coverage could be a significant issue for the lower efficiencies, particularly for OPE and T6 based cells. However, the T4BTD cell has similar absorption coverage to the P3HT and PB n DT cells. In fact, the oligomer:CdSe hybrid morphology may be the m ain reason for the such low performance As shown in Figure 8 9 the T4BTD:CdSe hybrid film shows several large aggregation domains (> 300 nm in parallel length, and vertically > 150 nm). This
178 Figure 8 9 (color) Tapping mode AFM topological image of T4BTD A:CdSe nanoparticle hybrid. significant aggregation is detrimental for exciton dissociation and charge carrier extraction. Hence, processing is still an issue for high efficiency hybrid PV cells using small molecules as donor materials. 8.5 Summary In this chapter we demonstrate the direct interfacing of CdSe nanocrystals by conjugated oligomers. Three conjugated oligomers bearing a phosphonic acid group were designed with differe nt energy gaps and energy levels. The TGA, PL quenching, and TEM study clearly indicate that the oligomers are intensively grafted to the CdSe nanocrystal through a strong chemical interaction (likely covalent). T hough t he oligomer:CdSe hybrid PV cells exh ibit photovoltaic effect as reflected in J V measurement, the efficiencies of these cells are relatively low at this moment. The EQE shows that both oligomers and CdSe nanocrystals contribute to the photogeneration. High performance PV cells based on ol igomer grafted CdSe hybrids is anticipated if the
179 processing could be improved and the anchoring group could be more systematically investigated.
180 CHAPTER 9 SOLUTION PROCESSED METAL OXID E ANODE INTERLAYER 9 .1 Introduction The relative low work function an d rough surface of indium tin oxide (ITO) substrate typically require an exotic interlayer that facilitates the charge injection/extraction in organic based photovoltaic cells. Due to its high conductivity, high transparency, low temperature aqueous soluti on processing, and commercial availability, PEDOT:PSS has witnessed great success as anode interlayer during the progress of organic based solar cells. However, an increasing number of studies reveal the hygroscopic and acidic nature of the PEDOT:PSS is on e of the main cause s of device degradation. 173 Plus, the relatively low work function (~5.0 eV) makes PEDOT:PSS not the ideal choice for organic solar cells using organic semiconductors with deep HOMO level. Theoretically, an ideal anode interlayer material may include the following characteristics: i ).proper energy level alignme nt with donor materials for charge injection/extraction; ii ). high transparency and conductivity to reduce photocurrent loss; iii ). comparability with roll to roll fabrication technique; iv ). structural an d chemical robustness to avoid inhomogeneous chemical composition and morphology; and v ). chemically stable and environmentally friendly. Transition metal oxides (TMOs), including molybdenum oxide (MoO 3 ), nickel oxide (NiO), vanadium oxide (V 2 O 5 ), and tungsten oxide (WO 3 ), have then arisen as promisin g anode interlayer candidates. 136 137 199 203 These TMOs are transparent in visible to near infrared spectrum region, electrically conductive due to chemical non stoichiometry, and chemically and structurally stable in ambient condition. How ever,
181 these TMOs are typically deposited upon ITO substrates at high temperature (> 400 o C) and at high vacuum 199 202 which complicates the device fabricati on and incompatible with roll to roll processing. Recently, MoO x thin film prepared by solution processing at relatively low temperature (< 200 o C) has been reported by several groups. 204 208 These MoO x films are either prepared in the form of nanoparticle precursor or converted from molybdenum oxide precursors. NiO precursor nanoparticles can be also prepar ed from solution chemistry, but they require relatively high temperature (> 4 00 o C) to convert these precursor nanoparticles into NiO thin films. 209 210 Tungsten o xide that is less harmful to the environment than other TMOs has received less attention using solution processing, 204 though it has been extensively studied i n organic solar cells using vacuum deposition. 201 In this chapter, we demonstrate the low temperature solution processing of tungsten oxide interlayer for organic based photovoltaic cells. The sol ution processed tungsten oxide films are very smooth, highly transparent, and morphologically amorphous. The organic PV devices using solution processed tungsten oxide as anode interlayer show comparable efficiency to thos e based on archetypal PEDOT:PSS and other anode interlayer. 9 .2 Synthesis and C haracterization of Tungsten Oxide Precursor and Thin F ilm 9 .2.1 Synthesis of Tungsten Oxide P recursor Tungsten oxide precursor solution was prepared by dissolving a saturated am ount (~22 mg/mL at room temperature) of ammonium tungsten oxide hydrate ((NH 4 ) 10 H 2 W 12 O 42 4H 2 O, here denoted as source material, Alfa Aesar) in deionized water and stirred for 1 h at 80 90 o C at air. A clear aqueous solution was obtained after cooling to room temperature, and not any other treatment is needed. The saturated
182 precursor solutio n can be directly diluted by deionized water at room temperature to obtain precursor solution with various concentrations. 3 nm, 6 nm, and 10 nm thin films were prepared by spin coating tungsten oxide precursor solutions wi th concentrations of 6 mg/mL, 11 mg/mL, and 22 mg/mL, respectively, at 5000 rpm for 1 min. The precipitated crystals of ammonium tungsten oxide hydrate (here denoted as precipitated crystal) were collected from an aqueous solution containing over saturated ammonium tungsten oxide at room temperature. Before thermogravimetric analysis (TGA) measurement, the precipitated crystals were naturally dried at room temperature for a few days. This drying is supposed to lose some crystalline water molecules incorporated during dissolution at elevat ed temperature. 9 .2.2 Characterization of Tungsten Oxide Thin F ilm The tungsten oxide solution was spin coated upon soda lime glass for transmittance and X ray diffraction (XRD) measurement, upon ITO substrates for atomic force microscopy (AFM) measurement and upon Si substrates for X ray photoelectron spectroscopy (XPS), ultraviolet photoelectron spectroscopy (UPS), and thickness measurements. All thin films were thermally annealed at 150 o C for 30 min. The surface morphologies of thin films and ITO subst rate were carried out in a Veeco Nanoscope scanning probe microscopy operating in a tapping mode. XRD measurement was ray diffractometer using Cu K radiation and operating at 40 kV and 45 mA. Thermogravimetric analysis was carried out in a TGA Q5000 at a heating rate of 5 o C at the air. The thickness of tungsten oxide thin film was determined by an ellipsometer based on a Cauchy model.
183 Figure 9 1. (color) TGA of ammonium tungsten oxide bulk materials. 9 .3 Solution processed Tungsten Oxide Thin F ilm Tungsten oxide precursor solution was prepared by dissolving ammonium tungsten oxide in deionized water at 80 o C The resulting solution is transparent, colorless, and stable in ambient condition, which has not shown any change in color and transparency after storage for over half a year. This dissolution process is not intended to thermally decompose the ammonium tungsten oxide into tungsten oxide but r ather, to disperse it in aqueous solution as tungsten oxide precursor TGA result of precipitated crystal from the precursor solution indicates the precursor in aqueous solution retains the chemical stoichiometry (Figure 9 1). Also the pH value of the aqu eous solution before and after dissolution of ammonium tungsten oxide precursor remains neutral, indicating there is no or limited ionization occurring. TGA thermogram of ammonium tungsten oxide hydrate crystal precipitated from its precursor solution sho ws that it starts to decompose at a very low temperature and dehydration occurs at ~50 o C (Figure 9 1), which is due to the excess incorporation of
184 water molecules to starting source material. 211 T he TGA thermogram of ammonium tungste n oxide hydrate source material shows a delayed behavior, where complete dehydration of absorbed water molecules occurs at ~160 o C and with a ~ 40 o C stage. After initial dehydration, the precipitated crystal shows a gradually thermal decomposition behavior at 50 300 o C where the excess water molecule and ammonium are removed, while the source material exhibits a sharp decomposition behavior at this temperature regime. This suggests the incorporation of water molecule restructures the ammonium tungsten oxi de framework. At temperature > 400 o C, further dehydration occurs for both samples to yield a stable and stoichiometric tungsten oxide (WO 3 ). The early decomposition behavior of precipitated crystal also suggests the possibility of decomposing the precurso r thin film into tungsten oxide at an even lower temperature due to dimension reducing X ray photoelectron spectroscopy was used to study the chemical states of tungsten oxide based thin films thermally treated at various temperatures. As shown in Figure 9 2, the O 1s spectra of the thin films were gradually shifted to lower binding energies as annealing temperature increased. The film annealed at 100 o C shows two O 1s peaks locating both at 532.0 eV that corresponds to ammonium tungsten oxide as indicated in the non annealed sample and at 530.5 eV that corresponds to tungsten oxide as demonstrated in the samples annealed at 150 o C and 200 o C. The W 4f spectra of these films also show a slight shift in binding energy as annealing temperature increases, and the binding energies at 37.7 eV and 35.6 eV corresponds to the characteristic W 6+ 4f 5/2 and 7/2 of tungsten oxide. Other than W 6+ there is no W state revealed in the XPS spectra for all the tested samples, which is also proved by the
185 Figure 9 2. (colo r) XPS of tungsten oxide precursors annealed at different temperature. fact that film color did not show any visible variation during annealing. 211 In particular, the intensity of W 4f spectra significantly increases for the annealed s amples, suggesting the shrink of precursor thin film and making W more sensitive for X ray electron penetration. Integration of the intensity of O 1s and W 4f spectra reveals the o C has an O/W of (2.9 3.1)/1. The slight non stoichiometric composition may be due to the incomplete decomposition of water molecule at relatively o C) and to oxygen deficiency under anneali ng at the inert condition. We further probed the morphological structure of these thin films prepared with different annealing temperature by X ray diffraction. The annealed thin film shows
186 almost the same XRD pattern as the non annealed precursor thin fil m and not any identical sharp peak appears (Figure 9 3), suggesting the amorphous nature of these tungsten oxide thin films. Being amorphous could be of benefit for organic electronics, since the device performance will not be limited by the spatial/struct ural non uniformity. 201 Figure 9 3. (color) XRD patterns of tungsten oxide precursors annealed at different temperature. Surface morphology of tungsten oxide thin films deposited upon ITO substrates was probed by tapping mode AFM. Figure 9 4 a & 9 4 b show the AFM images of 3 nm and 10 nm WO 3 films deposited upon ITO substrates, respectively. The 3 nm film shows some grains, while these grains have size much smaller than those of ITO (Figure 9 4c). Such a thin fil m may not fully cover the ITO surface and more likely fill its pin holes, which thus leads to significant reduction in the root mean square roughness (R rms ) from 1.7 nm for ITO to 0.82 nm for ITO with a 3 nm WO 3 film. The 10 nm WO 3 film is very homogeneous and does not show any grains, pin holes, or microstructures, which
187 Figure 9 4. (color) Tapping mode AFM topological images of (a) ITO with a 3 nm WO 3 (b) ITO with a 10 nm WO 3 (c) bare ITO substrate, and (d) ITO with a 20 nm PEDOT:PSS layer. The root mean square roughness (R rms ) of sWO 3 (3 nm), sWO 3 (10 nm), ITO, and PEDOT:PSS are 0.82 nm, 0.34 nm, 1.7 nm, and 1.0 nm, respectively. results in a very smooth surface with R rms = 0.34 nm, even smooth er than those vacuum deposited WO 3 films. Such a smooth s urface, as suggested by the XRD result, is due to the amorphous nature of WO 3 thin film. As a comparison, PEDOT:PSS, the typical anode interlayer for organic solar cells, has an R rms of 1.0 nm deposited upon ITO substrate (Figure 9 4d). Transparency is ano ther important factor of anode interlayer in organic electronics. Figure 9 5 shows the transmittance of different thin films deposited upon soda lime glass. The transmittance of the tungsten oxide thin film follows exactly with the sola lime glass at the e n
188 Figure 9 5 (color) T ransmittance of various interlayers deposited upon glass substrates. wavelength regi on for both films. Compared to 20 nm PEDOT:PSS deposited upon glass, 10 nm WO 3 film shows 3 suggesting the additional advantage of WO 3 as anode interlayer in organic solar cells utilizing low gap small m olecules / polymers as light harvesting materials. 9 .4 PV D evices with Tungsten Oxide Anode I nterlayer 9 .4.1 Solution processed P olymer S olar C ells The chemical, morphological, and optical properties of the solution processed tungsten oxide thin films sugg est much promising as anode interlayer and/or charge recombination layer in organic solar cells. As a proof, we first demonstrated polymer solar cells using archetypal P3HT:PCBM as the active layer. In order to maximize device performance, the PEDOT:PSS an d WO 3 thin films were all annealed at 150 o C before active layer deposition. The J V characteristics of the P3HT:PCBM devices were shown in Figure 9 6 and their photovoltaic performance parameters were summarized in Table 9 1.
189 The device using bare ITO s hows high injection barrier and series resistance (17 2 ), resulting in low photocurrent and voltage. However, i ntroducing a n ultrathin 3 nm WO 3 film as anode interlayer into P3HT:PCBM device drastically decreases series resistance to 5.6 2 and the device shows short circuit current J sc = 8.0 mA/cm 2 open circuit voltage V oc = 0.621 V, fill factor FF = 0.56 and ultimate p = 2.8%, all of which are analogous to the device using PEDOT:PSS as an anode interlayer (( J sc = 8.3 mA/cm 2 V oc = 0.633 V, FF = 0.56, and p = 2.9%). Slight ly increasing film thick ness to 6 nm and 10 nm, the WO 3 devices show exact ly the same J sc as the PEDOT:PSS device but slightly higher V oc due to lower dark current (Figure 9 6a ). In particular, FF has been significantly increased from 0.55 to 0.61 probably due to more balanced ch arge injection and collection, resulting in enhancement in p from 2.9% of the PEDOT:PSS device and 2.8 % of the sMoO x device to 3.2% of the 10 nm WO 3 device (Table 9 1) This efficiency is also comparable to the state of the art P3HT:PCBM devices using either vacuum deposited or solution processed TMOs anode inte rlayers reported by other groups. 137 Moreover, as shown in Figure 9 6a the dark current of the WO 3 devices has been drastically decreased compared to the devices using either bare ITO or PEDOT:PSS, s uggesting effecti ve blocking in leakage current. The EQE as a function of wavelength has also been recorded for the P3HT:PCBM devices using various anode interlayers. As shown in Figure 9 6c the device s with an anode interlayer show much higher EQE at all the photoresponsive regime. The WO 3 devices exhibit slightly higher EQE than the PEDOT:PSS device at 3 films.
190 Figure 9 6. (color) J V characteristics of P3HT:PCBM solar cells using different anode interfacial layer under dark (a) and 1 sun AM 1.5 G illumination (b), and their corresponding EQE (c). The device performance has also been found to be dependent on the annealing temperature of the tungsten oxide film. As shown in Fi gure 9 7, the device using WO x films without annealing exhibits an S shape of J V characteristic and results in low J sc 2 ) rising from
191 Table 9 1. Summary of photovoltaic performance under 1 sun AM 1.5 G illumination for P3HT:CdSe hybrid photovoltaic cells. Anode interlayer J sc (mA/cm 2 ) V oc (V) FF p (%) R s ( cm) Bare ITO 6.9 0.492 0.32 1.1 17 PEDOT:PSS (20 nm) 8.3 0.633 0.56 2.9 3.9 sWO 3 (3 nm) 8.0 0.621 0.56 2.8 5.6 sWO 3 (6 nm) 8.3 0.639 0.59 3.1 4. 3 sWO 3 (10 nm) 8.3 0.641 0.61 3.2 4. 2 sMo O x (10 nm) 8. 1 0.6 12 0. 56 2.8 3.0 sMoO x was prepare d based on reference 206 at ambient condition. insulating nature of the ammonium tungsten oxide precursor interlayer prepared at room temperature. When the annealing temperature increased to 50 o resistance was reduced by 75% to 8.0 2 due to the partial decomposition of the tungsten oxide precursor, correspondingly resulting in better J sc (6.3 mA/cm 2 ) and FF (0.49). The device performance has been further increased to the best values with J sc = 9.0 mA/cm 2 V oc = 0.639 V, FF=0. 58, and p = 3.3% when the tungsten oxide film annealed at 150 o C, at which the series resistance has been reduced to ~ 4 2 F urther increasing annealing temperature to 200 o C results in an S shape in J V characteristic due to the increased series r 2 ) and consequently the device performance h as been significantly reduced. This increased series resistance probably contributes from the oxygen deficiency in ITO or tungsten oxides during annealing. In brief, the chemical composition of the tungsten oxide film is highly dependent on the annealing temperature, which in turn reflects in the device series resistance and efficiency
192 Figure 9 7. (color) Effect of annealing temperature of WO 3 interlayers on the performance of P3HT:PCBM solar cells. 9 .4.2 Vacuum deposited S mall molecule Solar C ells The vacuum growth of organic small molecule thin film has been regarded to be orientation, and domain size, which i n turn affects the performance of the PV cells. 212 As discussed, tungsten oxide film deposited upon ITO substrate is very smooth and transparent, which may be of benefit for the growth of organic molecules and then their respective electronic devices. We first demonstrate the robustness of the tungsten oxide film in bilayer CuP c/C60 cell. As shown in Figure 9 8 a the device deposited upon pure ITO shows J sc = 3.8 mA/cm 2 V oc = 0.48 V, FF = 0.60, and p = 1.1%; while the devices with a WO 3 layer show slightly increased J sc = 4.1 mA/cm 2 and V oc = 0.51 V but similar FF = 0.60, resulting in ~18% enhancement in p (1.3%). The enhancement in V oc attributes to the suppression of dark current when a WO 3 layer included, and the increase in J sc is likely due to better hole extraction. As shown in Figu re 9 8b, the EQE further confirms the
193 increase in J sc The device with a WO 3 that mainly contributes from CuPc absorption, suggesting photogenerated hole can be more efficiently extracted when dissociat ed at th e CuPc C60 interface. Figure 9 8. (color) Effect of a WO 3 interlayer on the bilayer CuPc/C60 solar cells. (a, b) J V characteristics under 1 sun AM 1.5G illumination and dark condition; (c) the corresponding EQE. The effect of an anode interlayer may be more prominent in device using deep HOMO molecules. In order to prove the universality of the solution processed tungsten oxide thin film, we fabricate d mixed planar SubPc:C60 (1:4) /C60 cells. As shown in Figure 9 9, the devices with a WO 3 layer show an increased V oc ( ~0.08 V), while J sc and FF remain the same as the device directly deposited upon the ITO substrate. This
194 enhancement in V oc is again due to more than one order of magnitude decrease of dark current (Figure 9 9a). Figure 9 9. (color) Effect of a WO 3 interlayer on the mixed planar SubPc:C60/C60 solar cells. (a, b) J V characteristics under 1 sun AM 1.5G illumination and dark condition. 9 5 Summary In this chapter we d emonstrate the low temperature solution processing of tungs ten oxide thin films and their application in organic photovoltaic cells. The XPS and TGA studies indicate that the solution processed tungsten oxide precursor has been converted to tungsten oxide after annealing at a mild temperature (150 o C). The
195 resulti ng tungsten oxide film is very smooth, transparent, and amorphous, which are the very characteristics as anode interfacial layer for organic based solar cells. The tungsten oxide film as anode interlayer has been firstly demonstrated in polymer solar cells The P3HT:PCBM cell with a WO3 layer shows comparable or even better device performance compared to the cell using PEDOT:PSS interlayer. In particular, the FF of the cell with WO 3 layer has been significantly increased from 0.56 to 0.61, suggesting that t he WO 3 layer may facilitate charge extraction. The polymer solar cells also exhibit annealing temperature dependent performance of the WO 3 layer, which is due to and an indicator of the conversion efficacy of tungsten oxide precursor to tungsten oxide. The tungsten oxide interlayer also shows robustness in organic based solar cells, which is also applicable in vacuum deposited small molecule solar cells. Using a 3 nm 10 nm WO 3 interlayer, both the bilayer CuPc/C60 cell and the mixed planar S ubPc:C60/C60 c ell show slightly increased J sc and V oc resulting in ~ 10 20% enhancement in p
196 CHAPTER 10 CONCLUSIONS AND FUTU RE WORK 10 .1 Conclusions Though organic inorganic hybrid photovoltaic cells containing colloidal nanocrystals have shown significant advance, the pace is in a slower manner compared to that of all organic counterparts, partly because it involves much more complicate electronic, chemical, and morph ological interfaces. This dissertation devotes to better understanding the fundamental chemical and physical properties of organic inorganic hybrid materials, and their relation ship with developing high performance organic inorganic hybrid photovoltaic cells. In particular, by tailoring both conjugated polymers and colloidal nanocrystals, engineering polymer nanocrystal interface, and optimizing device architecture, the power con version efficiency of bulk heterojunction organic inorganic hybrid PV cells has been increased from 1 3% in literature to 3 5% present in this dissertation (Figure 10 1). 115 10 .1.1 Organic inorganic Hybrid M aterials Colloidal inorganic nanocrystals can be designed and tailored in the perspective of structure and properties such as the size, shape, composition, and surface chemist ry. First, as shown in Chapter 4 the hybrid PV cells show performance dependence on nanoparticle size, which attributes to increased electron mobility and decreased surf ace defect density as nanoparticle size increases. Second, hybrid PV cells also show an aging effect as exposed to the air, attributing to the interaction between CdSe surface with moisture / oxygen and also suggesting the complex surface property of collo idal nanocrystals. Moreover, compared to spherical nanocrystals, the elongated counter p arts show enhanced performance in hybrid PV cells, which is due to the
197 Figure 10 1 (color) The plot of power conversion efficiency ( p ) vs year in bulk heterojunction organic inorganic hybrid photovoltaic cells. This plot summarizes the major advance, and the champion efficiencies of hybrid PV cells in other groups based on each kind of colloidal nanocrystals have also been shown. The reference 17, 18, 115, and146 The black and red curves are the eye guiders of the The dashed red line indicates the starting date o f this work. increased charge transport property in hybrid matrix. In addition, composition varied cadmium chalcogenides nanorods were also designed and synthesized. The initial result shows hybrid PV cells based on CdS nanorods and P3HT are promising for further investigation. T ailoring both energy gap and energy level s of colloidal nanocrystals, to a large extent, is constrained to their intrinsic properties of bulk materials; while conjugated polymers are more feasible and not limited to starting organic unit or group. The archetypal P3HT has been mostly demonstrated as donor materials in hybrid PV cells,
198 while its relatively large energy gap (1.9 eV) limits the photon harvesting that leads to relatively low short circuit current ( J sc ~5 6 mA/cm 2 ) and n arrow spectral response (~ 650 nm). Besides, its energy levels are not optimally aligned with those of colloidal nanocrystal to maximize open circuit voltage. In chapter 6 a low gap polymer PCPDTBT that extends spectral response to ~850 nm was utilized as donor materials for hybrid PV cells, resulting in enhanced J sc due to the photocurrent contribution from near infrared region. The optimized device using PCPDTBT and CdSe nanoparticles shows J sc ~ 9 mA/cm 2 and p 3.5%. Moreover, the open circuit voltage of hybrid PV cells can also be manipulated by choosing conjugated polymer with optimal energy level alignment with nanocrystal acceptor. This was demonstrated using a polymer PB n DT FTAZ with deep HOMO level (~5.3 6 eV), resulting in hybrid PV cells with open circuit voltage as high as 0.9V and p > 3% ( Chap ter 6) The processing of organic inorganic hybrid materials also has a vast impact on the device perfor mance. As discussed in chapter 6 the domain size and sur face roughness of PCPDTBT:CdSe hybrid films have shown to be highly dependent upo n processing solvent s, which in turn affects the hybrid PV cell performance. Thermal annealing that may further lead to optimal phase segregation and removal of exciton and ch arge carrier recom bination/trap centers present on CdSe nanocrystals has shown to be an effect ive manner to enhance the device performance as demonstrated in both P3HT:CdSe and PCPDTBT:CdSe cells. 10 .1.2 Organic inorganic I nterface Surface property of coll oidal nanocrystals has been found to be of particular importance in bulk heterojunction organic inorganic hybrid PV cel ls, as demonstrated in Chapter 7 and Chapter 8 A significant improvement in performance of hybrid PV cells
199 using CdSe nanorods and conju gated polymers has been observed by engineering the polymer CdSe interface with the EDT treatment. The devices with the EDT treatment show ~30 90% enhancement in p depending on the aspect ratio of CdSe nanorods, resulting in maximum p ~3% for P3HT:CdSe cells and record p ~5% for PCPDTBT:CdSe cells under AM 1.5G illumination, respectively. Surface chemistry of CdSe nanorods with various chemical treatment s was pr obed by FTIR, NMR, and XPS characterization s These characterizations unanimously indicate that t he ligand exchange in pyridine selectively removes neutral L type ligands such as TOPO yet the EDT treatment removes charged X type ligands such as alkylphosph onic acids and concomitantly passivates the nanocrystal surface by formation of Cd thiolate. The removal of potential exciton and charge carrier recombination sites together with increased electron mobility of hybrid thin film upon X type ligand cleavage a re the very nature of performance enhancement in hybrid PV cells. The chemical characterization also further proves the chemical composition of colloidal nanocrystals: a CdSe core with a monolayer of Cd 2+ and a monolayer of mixed L type and X type ligands The neutral L type ligands bond to the CdSe surface through dative interaction, and the charge X type ligands bond to the Cd 2+ through ionic interaction. Finally, the reaction mechanism between Cd X ligand and EDT has also been regarded as a nucleophilic reaction The unsaturated surface of colloidal nanocrystals enables to directly interface with organic donor molecules to generate organic inorganic hybrid supermolecules. Oligomers bearing a chemically reactive phosphonic acid group were designed and
200 syn thesized. Various chemical, optical, and morphological characterizations demonstrated that the oligomers are bounded to the CdSe surface through strong chemical interaction, resulting in stable oligomer:CdSe hybrid supermolecules and substantial energy tra nsfer between oligomer and CdSe nanocrystal under optical stimulus. The hybrid PV cells based on these oligomer:CdSe nanocrystal hybrid supermolecules showed significant photovoltaic effect and also photocurrent contribution from both oligomers and CdSe na nocrystals. 10 .1.3 Semiconductor Metal Interface Semiconductor metal interface is another aspect that can be optimized in an organic inorganic hybrid PV cell. Introducing an interlayer to the anode or cathode is the typical strategy to engineer the semic onductor metal interface ( device structure ) to facilitate charge injection/extraction and ma nage light absorption distribution. Hybrid PV cells using CdSe nanoparticles with a ZnO NP layer show ~30 90% enhancement in p depending on the nanoparticle size and conjugated polymers which leads to maximum p 2.4% for P3HT:CdSe cells and 3.5% for PCPDTBT:CdSe cells under simulated 1 sun AM 1.5G illumination as discussed in Chapter 5 and Chapter 6 The morphological, electronic, and optical studies indicate the ZnO NP layer could optimally re distribute the optical field within the active layer, smooth en the active layer to provide better contact with the Al cathode, block the hole transport to the cathode, facilitate the electron extraction, and also prevent the exciton recombination at the active layer Al interface. In addition, the ZnO layer also physically isolates or retards the permeation of moisture / oxygen from cathode side, which leads to dramatic improvement in device environmental stability with ~70% p retained after storage at the am bient condition for > 2 months.
201 On the other hand, anode modification has also received much attention for organic based photovoltaic cells. Though PEDOT:PSS has shown much success in solution processed polymer solar cell s, the intrinsic acidity and hygroscopicity make it detrimental t o device stability. In chapter 9 tungsten oxide, one of the less harmful transition metal oxides, has been prepared by converting its precursor by annealing at mild temperature. Various char acterization s indicate the solution processed tungsten oxide thin film is smooth, transparent, and structurally amorphous. The tungsten oxide thin film has been demonstrated as anode interlayer in both solution processed P3HT:PCBM cell s and vacuum evaporat ed small molecule cells. The P3HT:PCBM cells show similar J sc and V oc but increased FF compared to the device using PEDTO:PSS, suggesting the tungsten oxide interlayer facilities charge extraction. In addition, the bilayer CuPc/C60 and mixed planar SubPc:C 60/C60 devices show slightly increased J sc and V oc compared to the devices deposited upon bare ITO substrates. 10 .2 Future Work Although bulk heterojunction organic inorganic hybrid photovoltaic cells have witnessed significant progress in the past severa l years, further research effort is still needed to move this technology into commercial interest. Hence, the theme for future work is again to improve the power conversion efficiency of hybrid PV cells to >10% and device environmental stability > 5 10 yea rs. Moreover, the study in organic inorganic hybrid materials can also be extended to the optoelectronic devices such as light emitting diodes, photodetector, etc.
202 10 .2.1 Organic inorganic Hybrid Materials 10 .2.1.1 Colloida l n anocrystals Though ZnO, CdS, PbS, etc. have shown some success, so far CdSe is still the mostly investigated nanocrystals as acceptor in hybrid PV cells, partly because of its maturity in synthesis, better energy level alignment with conjugated polymers, and suitable energy gaps. None theless, the most efficient hybrid PV cells based on CdSe nanocrystals only show J sc ~12 mA/cm 2 V oc ~0.7 0.8V, and FF ~0.50 0.55 There is still much room to improve the photovoltaic performance by tailoring colloidal nanocrystals. The ideal colloidal nanocrystals for hybrid PV application may include the following chemical and physical properties: high electron mobility (~10 2 10 3 cm 2 /Vs), energy gap ~1.2 1.5 eV, LUMO level ~3.7 4.0 eV, few surface defects, and flexible surface chemistry. Colloidal n anocrystals with anisotropic structure have been shown higher electron transport behavior, which have been frequently demonstrated in literature and in this work as well. Aligning elongated nanocrystals vertically upon the substrates could be another strat egy to improve the electron mobility of colloidal nanocrystals within a polymer:nanocrystal matrix. Besides, exploring a new approach to passivate nanocrystal surface is also of critical importance in improving electron transpor t, as exemplified in chapter 7 Though the natural semiconductors are limited, synthetic chemistry enable to develop nanocrystals non existed in nature. Alloying two or more semiconductors to produce ternary or quaternary nanocrystals is a choice to optimize energy level and energy g ap for PV application based on donor acceptor heterojunction. Band gap and composition engineering on a nanocrystal (BCEN) as proposed by Peng, is the very
203 concept for the future development of colloidal nanocrystal for PV application and beyond. Engineeri ng surface property of colloidal nanocrystals, as demonstrated in chapter 7 is critical in hybrid PV application. The early study in nanocrystal surface tailoring including this work is still in its fancy and requires much more study. Other than organic c apping molecules, inorganic ligands as demonstrated in field effect transistor or organic and inorganic hybrid ligands may also be promising to passivate the nanocrystal surface for hybrid PV cells. The physical mechanism in enhancing the photovoltaic perf ormance by surface pas sivation also needs more study. 10 .2.1.2 O rganic s emiconductors Thus far, the organic semiconductors in hybrid PV cells are mostly directly borrowed from study in all organic solar cells such as P3HT, PCPDTBT, PBnDT, etc. There are ve ry few organic molecule s (the functional oligomers in C hapter 7) specially tailored hybrid PV cells. T he concept for designing organic semiconductors for hybrid PV cells is more or less similar to that for all organic PV cells. The energy gap, energy level hole mobility, and solvent affinity ( processability ) are the parameters featured in designing organic semiconductor for BHJ PV application. Other than that, organic semiconductors with a reactive group may be advantage for blending with colloidal nanocry stals. This reactive group could not be a charge or exciton sink, but must bind to colloidal nanocrystals through strong chemical interaction. Phosphonic acid (PA) group may be not an ideal choice, since the bonding between PA and colloidal nanocrystal is through oxygen and the low energy orbital level of oxygen may sink the passing electron during charge transfer or transport that lead to loss in photocurrent. M ercapto group could be a better choice, since S has much lower electronegativity and higher
204 orbi tal level and SH can lose an electron to bind with colloidal nanocrystals (Chapter 7) 10 .2.1.3 O rganic inorganic h ybrids The typical approach to prepare organic inorganic hybrids is by directly blending organic semiconductors with colloidal nanocrystals in organic solvent mixtures. The solubility of organic semicond uctors is typically given by alkyl side chain; while the dispersity of colloidal nanocrystals in org anic solvents is provided by surface ligands. These surface ligands are typically insulating for charge transport and also able to serve as recombination / trapping centers for photogenerated excitons and charge carriers. An alternative approach to prepare organic inorganic hybrids is by in situ synthesis. For example, P3HT:CdS hybrids prepared b y in situ wet chemistry method as the active layer leads to p ~3%; 156 low gap polymer (PSiF):CuInS hybrid prepared by decomposing the CuInS precursors within the polymer matrix as active layer results in p 2.8%, 161 several times higher than those cells prepared by blending CuInS nanoparticles with conjugated poly mers. Nonetheless, the challenge for this approach is to select appropriate precursors and solvents for colloidal nanocrystal growth at relatively low temperature. The solvents should not only have affinity with precursors, but also the organic semiconduc tors. So far, though there are very limited nanocrystals that can be grown within a polymer matrix at low temperature, it is still a promising approach to integrate hybrid materials preparation and device application. 10 .2.2 Multi junction Organic inorgani c Hybrid PV C ells Stacking cells in series is another approach to improve the efficiency of hybrid solar cells. The photogeneration can be maximized by readily designing the optical distribution in each subcell and concurrently the photovoltage of each sub cell can be
205 added. The tandem all organic solar cells have shown some success recently, while there is still no literature report of solution processed tandem hybrid solar cells, partly because the device fabrication is much more complicate than single jun ction cell. Figure 10 2 shows a tandem cell composed of a front solution processed hybrid subcell, a back vacuum deposited small molecule subcell, and ZnO/PEDOT:PSS/MoOx charge recombination zone (CRZ). The broad spectral response of PCPDTBT:CdSe cell coul d harvest near infrared solar photon. Though the SubPc/C60 cell can only absorb the visible photon, it can produce large open circuit voltage. The preliminary result shows J sc of the tandem cell close to the J sc of the current limiting SubPc/C60 cell; and the V oc of the tandem cell is close to the addition of those of the two subcells, suggesting the efficient charge recombination at CRZ. High efficiency tandem hybrid Figure 10 2. (color) J V characteristics of a tandem cell containing a PCPDTBT:CdSe h ybrid cell and a small molecule cell using ZnO NPs as inter connecting layer and the corresponding subcells.
206 solar cells are anticipated if the front and back subcells were optimized to balance photon harvesting. In addition, by replacing SubPc/C60 with PB nDT:CdSe, all solution processed tandem hybrid solar cells with both large V oc and high efficiency are envisioned.
207 APPENDIX LIST OF PUBLICATIONS AND CONFERENCE PRESENTATIONS Publications 1. Renjia Zhou John P. Mudrick, Weiran Cao, temperature solution processed tungsten oxide interlayer for efficient organic 2. Renjia Zhou Romain Stalder, Dongping Xie, Ying Zheng, Weiran Cao, Yixing Yang, Marc Plaisant, Kirk S. Schanze, Paul H. Holloway, John R Reynolds, Jiangeng processed polymer:colloidal nanocrystal ACS Nano 2012. 3. Romain Stadler, Dongping Xie, Renjia Zhou Jiangeng Xue, Kirk S. S chanze, gap conjugated oligomers grafted to CdSe Chemistry of Materials 2012 24 3143 4. Renjia Zhou ChemPhysChem, 2 012, 13 2471. 5. Renjia Zhou Ying Zheng, Lei Qian, Yixing Yang, Paul Holloway, Jiangeng processed, nanostructured hybrid organic inorganic solar cells with Nanoscale 2012, 4 3507. 6. Jason D. Myers, Weiran Ca o, Vincent Cassidy, Sang Hyun Eom, Renjia Zhou enhanced efficiency in organic Energy & Environmental Science, 2012, 5 6900.
208 7. Jihua Yang, Lei Qian, Re njia Zhou Aiwei Tang, Paul Holloway, and Jiangeng stable hybrid inorganic/organic solar cells with a ZnO Journal of Applied Physics 2012, 111 044323. 8. O. M. Ntwaeaborwa, Renjia Zhou Lei Qian Shrey as S. Pitale, J. Xue, H. C. fabrication annealing effects on the performance of Physica B 2012, 407 1631. 9. Lei Qian, Jihua Yang, Renjia Zhou Aiwei Tang, Ying Zheng, Te ngkuan CdSe Solar Cells Journal of Materials Chemistry, 2011, 21 3814. 10. Jihua Yang, Aiwei Tang, Renjia Zhou J size and device aging on performance of hybrid poly(3 hexylthiophene):CdSe Solar Energy Materials and Solar Cells, 2011, 95 476. Conference Presentations 1. (Invited) Jiangeng Xue*, Renjia Zho u talk, Tampa, Oct. 2012. 2. (Invited) Jiangeng Xue*, Renjia Zhou inorganic Hybrid Materials for Photov IUMRS International Conference on Electronic Materials ( IUMRS ICEM 2012 ) Invited talk, Pacifico Yokohama, Yokohama, Japan, Sept. 2012. 3. Renjia Zhou Ying Zheng, Lei Qian, Paul H. Holloway, and Jiangeng Xue, efficiency, Solut ion processed Hybrid Organic
209 Annual Joint Symposium of the FLAVS and FSM Invited talk for young leaders section, Orlando, Mar. 2012. 4. (Invited) Renjia Zhou Ying Zheng, Lei Qian, Paul H. Holloway, and Jiangeng h efficiency, Solution processed Hybrid Polymer:Colloidal Nanocrystal Global Organic Photovoltaics Invited talk, Hangzhou, Oct. 2011. 5. Renjia Zhou ,* Ying Zheng, Dongping Xie, Weiran Cao, Yixing Yang, Romain Stalder, Marc Plaisant, Kirk S. Schanze, Paul H. Holloway, John R. Reynolds, and efficiency solution processed hybrid polymer:colloidal nanocrystals SPIE Optics & Photonics Oral, San Diego, Aug. 2011. 6. Renjia Zhou ,* Ying Zheng, Dongping Xie, Weiran Cao, Yixing Yang, Romain Stalder, Marc Plaisant, Kirk S. Schanze, Paul H. Holloway, John R. Reynolds, and inorganic solar cells based on blends of Annual Joint Symposium of the FLAVS and FSM Poster, Orlando, Mar. 2011. 7 Renjia Zhou Lei Qian, Ying Zheng, Paul H. Holloway, and Jiangeng Xue stable Hybrid Organic SPIE Optics & Photonics Oral presentation, San Diego, Aug. 2010. 8 Renjia Zhou ,* Qian Lei, Ying Zheng, Paul Holloway, Jiangeng Xue, Efficient and air stable hybrid organic inorganic photovoltaic cells. The 9 th International Electron Systems (F 9) Poster, Atlanta, May 2010. 9. (Invited) Jiangeng Xue,* Renjia Zhou Lei Qian, Ying Zheng, Jihua Yang, Aiwei Tang, Paul H. Holloway, Hybrid Photovoltaic Cells based on Conjugated
210 Polymers and CdSe Nanoparticles, The International Conference on Nanophotonics Invited talk, Epocal Tsukuba, Japan, May 2010. 10. Renjia Zhou ,* Lei Qian, Ying Zheng, Paul Holloway, Jiangeng Xue, Efficient and air stable hybrid organic inorganic solar cells based o n a low gap polymer and CdSe nanoparticles. Annual Joint Symposium of the FLAVS and FSM Poster, Orlando, Mar. 2010. 11. Jihua Yang, Lei Qian, Renjia Zhou ,* Aiwei Tang, Paul Holloway, Jiangeng Xue, Improving the efficiency and air stability of hybrid P3HT /CdSe solar cells with a ZnO buffer layer. MRS Fall Meeting Oral, Boston, Dec. 2009. 12. Jihua Yang, Aiwei Tang, Renjia Zhou ,* Jiangeng Xue, Hybrid P3HT/CdSe photovoltaic cells: effects of nanocrystal size and device aging. MRS Fall Meeting Poster, Bosto n, Dec. 2009.
211 LIST OF REFERENCES 1. Chu, S.; Majumdar, A. Nature 2012 488 294 303. 2. Luque, A.; Hegedus, S. Handbook of Photo voltaic Science and Engineering John Wiley & Sons Ltd 2002 3. Platzer, M. D. U.S. Solar Photovoltaic Ma nufacturing: Industry Trends, Global Competition, Federal Support Congretional Research Service, 2012 4. Green, M. A.; Emery, K.; Hishikawa, Y.; Warta, W.; Dunlop, E. D. Prog Photovoltaics 2012 20 12 20. 5. Mitzi, D. B.; Yuan, M.; Liu, W.; Kellock, A. J.; Chey, S. J.; Deline, V.; Schrott, A. G. Adv Mater 2008 20 3657 3662 6. Bag, S.; Gunawan, O.; Gokmen, T.; Zhu, Y.; Todorov, T. K.; Mitzi, D. B. Energ Environ Sci 2012 5 7060 7065. 7. Guo, Q. ; Hillhouse, H. W.; Agrawal, R. J Am Chem Soc 20 09 131 11672 11673 8. Oregan, B.; Gratzel, M. Nature 1991 353 737 740. 9. Hagfeldt, A.; Gratzel, M. Chem Rev 1995 95 49 68. 10. Bach, U.; Lupo, D.; Comte, P.; Moser, J. E.; Weissortel, F.; Salbeck, J.; Spreitzer, H.; Gratzel, M. Nature 1998 395 583 585. 11. Chung, I.; Lee, B.; He, J. Q.; Chang, R. P. H.; Kanatzidis, M. G. Nature 2012 485 486 489 12. Lee, M. M.; Teuscher, J.; Miyasaka, T.; Murakami, T. N.; Snaith, H. J. Science 2012 338 643 647. 13. Tang, C. W. Appl Phys Lett 1986 48 1 83 185. 14. Peumans, P.; Yakimov, A.; Forrest, S. R. J Appl Phys 2003 93, 3693 3723. 15. Forrest, S. R. Nature 2004 428 911 918. 16. Huynh, W. U.; Dittmer, J. J.; Alivisatos, A. P. Science 2002 295 2425 2427. 17. Greenham, N. C.; Peng, X. ; Alivisat os, A. P. Phys Rev B 1996 54 17628 17637. 18. Zhou, R.; Xue, J. ChemPhysC hem 2012 13 2471 2480.
212 19. Sze, S. M.; Ng, K. Physics of semiconductor devices, 3rd edition John Wiley and Sons, Inc 2007. 20. Rossetti, R.; Ellison, J. L.; Gibson, J. M.; Bru s, L. E. J Chem Phys 1984 80 4464 4469. 21. Alivisatos, A. P. Science 1996 271 933 937. 22. Alivisatos, A. P. J. Phys. Chem. 1996 100 13226 13239. 23. Xia, Y.; Yang, P. Adv Mater 2003 15 351 355. 24. Brus, L. E. J Chem Phys 1984 80 4403 44 09. 25. Yu, W. W.; Qu, L.; Guo, W.; Peng, X. Chem Mater 2003 15 2854 2860. 26. Smith, A. M.; Nie, S. Acc. Chem Res 2010 43 190 200. 27. Yin, Y.; Alivisatos, A. P. Nature 2005 437 664 670. 28. Peng, X.; Manna, L.; Yang, W.; Wickham, J.; Scher, E.; Kadavanich, A.; Alivisatos, A. P. Nature 2000 404 59 61. 29. Ji, X.; Copenhaver, D.; Sichmeller, C.; Peng, X. J Am Chem Soc 2008 130 5726 5735. 30. Peng, Z.; Peng, X. J Am Che m Soc 2002 124 3343 3353. 31. Murray, C. B.; Kagan, C. R.; Bawendi, M. G. Annu Rev Mater Sci 2000 30 545 610. 32. Colvin, V. L.; Schlamp, M. C.; Alivisatos, A. P. Nature 1994 370 354 357. 33. Huynh, W. U.; Dittmer, J. J.; Alivisatos, A. P. Science 2002 295 2425 2427. 34. Bruchez, M.; Moronne, M.; Gin, P.; Weiss, S.; Alivisatos, A. P. Science 1998 281 2013 2016. 35. Chan, W. C. W.; Nie, S. Science 1998 281 2016 2018. 36. Alivisatos, A. P. J. Phys. Chem. 1996 100 13226 13239. 37. McBride, J.; Treadway, J.; Feldman, L. C.; Pennycook, S. J.; Rosenthal, S. J. Na no Lett 2006 6 1496 1501. 38. Gaponenko, S. V. Optical Properties of Semiconductor Nanocrystals Cambridge University Press 1998. 39. Wei, S.; Zunger, A. Appl Phys Lett 1998 72 2011 2013.
213 40. Moreels, I.; Lambert, K.; Smeets, D.; De Muynck, D.; No llet, T.; Martins, J. C.; Vanhaecke, F.; Vantomme, A.; Delerue, C.; Allan, G.; Hens, Z. ACS Nano 2009 3 3023 3030. 41. Peng, X. Nano Res 2009 2 425 447. 42. Peng, Z.; Peng, X. J Am Chem Soc 2001 123 183 184. 43. Peng, Z.; Peng, X. J Am. Chem S oc 2001 123 1389 1395. 44. Talapin, D. V.; Lee, J.; Kovalenko, M. V.; Shevchenko, E. V. Chem Rev 2010 110 389 458. 45. Murray, C. B.; Norris, D. J.; Bawendi, M. G. J Am Chem Soc 1993 115 8706 8715. 46. Peng, X. Acc. Chem Res 2010 43 1387 1 395. 47. Evans, C. M.; Evans, M. E.; Krauss, T. D. J Am Chem Soc 2010 132 10973 10975. 48. Ruberu, T. P. A.; Albright, H. R.; Callis, B.; Ward, B.; Cisneros, J.; Fan, H. J.; Vela, J. A CS Nano 2012, 6 5348 5359. 49. Yu, W. W.; Peng, X. Angew Chem I nt Ed. 2007 46 2559 2559. 50. Evans, C. M.; Guo, L.; Peterson, J. J.; Maccagnano Zacher, S.; Krauss, T. D. Nano Lett 2008 8 2896 2899. 51. Luther, J. M.; Law, M.; Beard, M. C.; Song, Q.; Reese, M. O.; Ellingson, R. J.; Nozik, A. J. Nano Lett 2008 8 3488 3492. 52. Ma, W.; Luther, J. M.; Zheng, H.; Wu, Y.; Alivisatos, A. P. Nano Lett 2009 9 1699 1703. 53. Manna, L.; Scher, E. C.; Alivisatos, A. P. J Am Chem Soc 2000 122 12700 12706. 54. Li, Z.; Peng, X. J Am Chem Soc 2011 133 6578 6586 55. Puzder, A.; Williamson, A. J.; Zaitseva, N.; Galli, G.; Manna, L.; Alivisatos, A. P. Nano Lett 2004 4 2361 2365. 56. Owen, J. S.; Park, J.; Trudeau, P. E.; Alivisatos, A. P. J Am Chem Soc 2008 130 12279 12281. 57. Gomes, R.; Hassinen, A.; Sz czygiel, A.; Zhao, Q.; Vantomme, A.; Martins, J. C.; Hens, Z. J Phys Chem Lett 2011 2 145 152.
214 58. Cros Gagneux, A.; Delpech, F.; Nayral, C.; Cornejo, A.; Coppel, Y.; Chaudret, B. J Am Chem Soc 2010 132 18147 18157. 59. Huynh, W. U.; Dittmer, J J.; Libby, W. C.; Whiting, G. L.; Alivisatos, A. P. Adv Funct Mater 2003 13 73 79. 60. Barkhouse, D. A. R.; Pattantyus Abraham, A. G.; Levina, L.; Sargent, E. H. A CS Nano 2008, 2 2356 2362. 61. Yang, J.; Tang, A.; Zhou, R.; Xue, J. Sol Energ Mat e r. Sol C ells 2011 95 476 482. 62. Ekimov, A. I.; Hache, F.; Schanneklein, M. C.; Ricard, D.; Flytzanis, C.; Kudryavtsev, I. A.; Yazeva, T. V.; Rodina, A. V.; Efros, A. L. J Opt Soc Am er. B 1994 11 524 524. 63. Norris, D. J.; Bawendi, M. G. Phys Re v B 1996 53 16338 16346. 64. Sargent, E. H. Nature Photon 2009 3 325 331. 65. Nozik, A. J. Physica E 2002 14 115 120. 66. Sukhovatkin, V.; Hinds, S.; Brzozowski, L.; Sargent, E. H. Science 2009 324 1542 1544. 67. Nozik, A. J.; Beard, M. C.; Luthe r, J. M.; Law, M.; Ellingson, R. J.; Johnson, J. C. Chem Rev 2010 110 6873 6890. 68. Semonin, O. E.; Luther, J. M.; Choi, S.; Chen, H. Y.; Gao, J. ; Nozik, A. J.; Beard, M. C. Science 2011 334 1530 1533. 69. Li, J.; Wang, Y.; Guo, W.; Keay, J. C.; Mis hima, T. D.; Johnson, M. B.; Peng, X. J Am Chem Soc 2003 125 12567 12575. 70. Zhang, J.; Tang, Y.; Lee, K.; Ouyang, M. Science 2010 327 1634 1638. 71. Kim, S.; Fisher, B.; Eisler, H. J.; Bawendi, M. J Am Chem Soc 2003 125 11466 11467. 72. Smi th, A. M.; Mohs, A. M.; Nie, S. Nat ure Nanotechnol 2009 4 56 63. 73. Greenham, N. C.; Peng, X.; Alivisatos, A. P. Phys Rev B 1996 54 17628 17637. 74. Barkhouse, D. A. R.; Pattantyus Abraham, A. G.; Levina, L.; Sargent, E. H. ACS Nano 2008 2 2356 2 362. 75. Tang, J.; Sargent, E. H. Adv Mater 2011 23 12 29.
215 76. Talapin, D. V.; Murray, C. B. Science 2005 310 86 89. 77. Kovalenko, M. V.; Scheele, M.; Talapin, D. V. Science 2009 324 1417 1420. 78. Lee, J.; Kovalenko, M. V.; Huang, J.; Chung, D.; Talapin, D. V. Nat ure Nanotechnol 2011 6 348 352. 79. Mott, N. F. Adv Phys 1967 16 49 144 80. Mott, N. F. Philos Mag 1969 19 835 852 81. Efros, A. L.; Shklovskii, B. I. J Phys C Solid State 1975 8 L49 L51. 82. Efros, A. L.; Shklovskii, B. I. Physica Status Solidi B Basic Res 1976 76 475 485. 83. McDonald, S. A.; Konstantatos, G.; Zhang, S.; Cyr, P. W.; Klem, E. J. D.; Levina, L.; Sargent, E. H. Nat ure Mater 2005 4 138 142 84. Sargent, E. H. Nat ure P h on ton. 2012 6 133 135. 85. Ma, W.; Swisher, S. L.; Ewers, T.; Engel, J.; Ferry, V. E.; Atwater, H. A.; Alivisatos, A. P. ACS Nano 2011 5 8140 8147. 86. Pattantyus Abraham, A. G.; Kramer, I. J.; Barkhouse, A. R.; Wang, X. H.; Konstantatos, G.; Debnath, R.; Levina, L.; Raabe, I.; Nazeer uddin, M. K.; Gratzel, M.; Sargent, E. H. ACS Nano 2010 4 3374 3380. 87. Tang, J.; Kemp, K. W.; Hoogland, S.; Jeong, K. S.; Liu, H.; Levi na, L.; Furukawa, M.; Wang, X. ; Debnath, R.; Cha, D. K.; Chou, K. W.; Fischer, A.; Amassian, A.; Asbury, J. B.; Sarge nt, E. H. Nat ure Mater 2011 10 765 771. 88. Pope, M.; Swenberg, C. E. Electronic processes in organic crystals and polymers Oxford University Press 1999. 89. Tang, C. W.; Vanslyke, S. A. Appl Phys Lett 1987 51 913 915. 90. Burroughes, J. H.; Brad ley, D. D. C.; Brown, A. R.; Marks, R. N.; Mackay, K.; Friend, R. H.; Burns, P. L.; Holmes, A. B. Nature 1990 347 539 541. 91. Dimitrakopoulos, C. D.; Malenfant, P. R. L. Adv Mater 2002 14 99 117 92. Sun, S. S. ; Sariciftci, N. S. Organic Photovoltai cs: Mechanisms, Materials, and Devices CRC press 2005. 93. Lennard Jones, J. E. Cambridge Philos. Soc. 1931 169 469. 94. Barth, S.; Bassler, H. Phys Rev Lett 1997 79 4445 4448.
216 95. Knupfer, M.; Peisert, H.; Schwieger, T. Phys Rev B 2002, 65 033 204. 96. Murphy, C. B.; Zhang, Y.; Troxler, T.; Ferry, V.; Martin, J. J.; Jones, W. E. J Phys Chem B 2004 108 1537 1543. 97. Bergemann, K. J.; Forrest, S. R. Appl Phys Lett 2011 99 243303. 98. Murphy, A. R.; Frechet, J. M. J. Chem Rev 2007 107 1066 1096. 99. Coropceanu, V.; Cornil, J.; da Silva, D. A.; Olivier, Y.; Silbey, R.; Bredas, J. L. Chem Rev 2007 107 2165 2165. 100. Shirota, Y.; Kageyama, H. Chem Rev 2007 107 953 1010. 101. Karl, N.; Marktanner, J. Mol Cryst Liq Cryst 2001 355 149 173. 102. Wright, J. D. Molecular Crystals, 2nd ed. Cambridge University Press 1999. 103. Rand, B. P.; Xue, J.; Uchida, S.; Forrest, S. R. J Appl Phys 2005 98 124902 104. Zhu, X.; Yang, Q.; Muntwiler, M. Acc. Chem Res 2009 42 1779 17 87. 105. Heremans, P.; Cheyns, D.; Rand, B. P. Acc. Chem Res 2009 42 1740 1747. 106. Xue, J. Polym Rev 2010 50 411 419. 107. Yu, G.; Gao, J.; Hummelen, J. C.; Wudl, F.; Heeger, A. J. Science 1995 270 1789 1791. 108. Peet, J.; Heeger, A. J.; Bazan G. C. Acc. Chem Res 2009 42 1700 1708. 109. Heeger, A. J. Chem Soc Rev 2010 39 2354 2371. 110. Moon, J. S.; Takacs, C. J.; Sun, Y. ; Heeger, A. J. Nano Lett 2011 11 1036 1039. 111. Clarke, T. M.; Durrant, J. R. Chem Rev 2010 110 6736 6767. 112. Vandewal, K.; Tvingstedt, K.; Gadisa, A.; Inganas, O.; Manca, J. V. Nature Mater 2009 8 904 909. 113. Veldman, D.; Meskers, S. C. J.; Janssen, R. A. J. Adv Funct Mater 2009 19 1939 1948. 114. Vandewal, K.; Tvingstedt, K.; Gadisa, A.; Inganas, O.; Manca, J. V. Phys Rev B 2010 81 125204 115. Wright, M.; Uddin, A. Sol Energ Mat er. Sol C ells 2012 107 87 111.
217 116. Sariciftci, N. S.; Smilowitz, L.; Heeger, A. J.; Wudl, F. Science 1992 258 1474 1476. 117. Sun, Y.; Welch, G. C.; Leong, W. L.; Takacs, C. J.; Bazan, G. C.; Heeger, A. J. Nat ure Mater 2012 11 44 48. 118. Yoshino, K.; Morita, S.; Kawai, T.; Araki, H.; Yin, X. H.; Zakhidov, A. A. Synth. Met als 1993 56 2991 2996. 119. Sariciftci, N. S.; Braun, D.; Zhang, C.; Srdanov, V. I.; H eeger, A. J.; Stucky, G.; Wudl, F. Appl Phys Lett 1993 62 585 587. 120. Shaheen, S. E.; Brabec, C. J.; Sariciftci, N. S.; Padinger, F.; Fromherz, T.; Hummelen, J. C. Appl Phys Lett 2001 78 841 843. 121. Li, G.; Shrotriya, V.; Huang, J. ; Yao, Y.; Moriarty, T.; Emery, K.; Yang, Y. Nat ure Mater 2005 4 864 868. 122. Kim, Y.; Cook, S.; Tuladhar, S. M.; Choulis, S. A.; Nelson, J.; Durrant, J. R.; Bradley, D. D. C.; Giles, M.; Mcculloch, I.; Ha, C. S.; Ree, M. Nat ure Mater 2006 5 197 203. 123. Dang M. T.; Hirsch, L.; Wantz, G. Adv Mater 2011 23 3597 3602. 124. van Mullekom, H. A. M.; Vekemans, J. A. J. M.; Havinga, E. E.; Meijer, E. W. Mat er. Sci Eng R 2001 32 1 40. 125. Muhlbacher, D.; Scharber, M.; Morana, M.; Zhu, Z.; Waller, D.; Gaudia na, R.; Brabec, C. Adv Mater 2006 18 2884 2889 126. Chen, H.; Hou, J.; Zhang, S.; Liang, Y.; Yang, G.; Yang, Y.; Yu, L.; Wu, Y.; Li, G. Nat ure Photon 2009 3 649 653. 127. Liang, Y.; Feng, D. ; Wu, Y.; Tsai, S T.; Li, G.; Ray, C.; Yu, L. J Am Chem Soc 2009 131 7792 7799. 128. Park, S. H.; Roy, A.; Beaupre, S.; Cho, S.; Coates, N.; Moon, J. S.; Moses, D.; Leclerc, M.; Lee, K.; Heeger, A. J. Nat ure Photon 2009 3 297 303 129. Liang, Y.; Yu, L. Acc. Chem Res 2010 43 1227 1236. 130. Amb, C. M .; Chen, S.; Graham, K. R.; Subbiah, J.; Small, C. E.; So, F.; Reynolds, J. R. J Am Chem Soc 2011 133 10062 10065. 131. Liang, Y.; Wu, Y.; Feng, D.; Tsai, S.; Son, H.; Li, G.; Yu, L. J Am Chem Soc 2009 131 56 57 132. Wakim, S.; Beaupre, S.; Bl ouin, N.; Aich, B. R.; Rodman, S.; Gaudiana, R.; Tao, Y.; Leclerc, M. J Mater Chem 2009 19 5351 5358.
218 133. Orimo, A.; Masuda, K.; Honda, S.; Benten, H.; Ito, S.; Ohkita, H.; Tsuji, H. Appl Phys Lett 2010 96 043305 134. Ma, W. ; Gopinathan, A.; He eger, A. J. Adv Mater 2007 19 3656 3659 135. Peet, J.; Kim, J. Y.; Coates, N. E. ; Ma, W. ; Moses, D.; Heeger, A. J.; Bazan, G. C. Nat ure Mater 2007 6 497 500. 136. Chen, L.; Xu, Z.; Hong, Z.; Yang, Y. J Mater Chem 2010 20 2575 2598. 137. Meyer, J.; Hamwi, S.; Kroger, M.; Kowalsky, W.; Riedl, T.; Kahn, A. Adv Mater 2012 24 5408 5427. 138. Brabec, C. J.; Shaheen, S. E.; Winder, C.; Sariciftci, N. S.; Denk, P. Appl Phys Lett 2002 80 1288 1290. 139. Gilot, J.; Barbu, I.; Wienk, M. M.; Janss en, R. A. J. Appl Phys Lett 2007 91 113520 140. Xue, J.; Uchida, S.; Rand, B. P.; Forrest, S. R. Appl Phys Lett 2004 85 5757 5759. 141. Kim, J. Y.; Lee, K.; Coates, N. E.; Moses, D.; Nguyen, T. Q.; Dante, M.; Heeger, A. J. Science 2007 317 222 225. 142. Peng, Z.; Peng, X. J Am Chem Soc 2001 123 183 184. 143. Yu, W. W.; Peng, X. Angew Chem Int Ed. 2002 41 2368 2371. 144. Robinson, R. D.; Sadtler, B.; Demchenko, D. O.; Erdonmez, C. K.; Wang, L. W.; Alivisatos, A. P. Science 2007 317 355 358. 145. Peng, Z.; Peng, X. J Am Chem Soc 2001 123 1389 1395. 146. Milliron, D. J.; Gur, I.; Alivisatos, A. P. MRS Bull 2005 30 41 44. 147. Manna, L.; Milliron, D. J.; Meisel, A.; Scher, E. C.; Alivisatos, A. P. Nat ure Mater 2003 2 382 385 148. Milliron, D. J.; Hughes, S. M.; Cui, Y.; Manna, L.; Li, J. B.; Wang, L. W.; Alivisatos, A. P. Nature 2004 430 190 195. 149. Ruberu, T. P. A.; Vela, J. ACS Nano 2011 5 5775 5784. 150. Cordero, S. R.; Carson, P. J.; Estabrook, R. A.; Strouse, G. F .; Buratto, S. K. J Phys Chem B 2000 104 12137 12142. 151. Koberling, F.; Mews, A.; Basche, T. Adv Mater 2001 13 672 676.
219 152. van Sark, W. G. J. H. M.; Frederix, P. L. T. M.; Bol, A. A.; Gerritsen, H. C.; Meijerink, A. Chem PhysC hem 2002 3 871 8 79. 153. Muller, J.; Lupton, J. M.; Rogach, A. L.; Feldmann, J.; Talapin, D. V.; Weller, H. Appl Phys Lett 2004 85 381 383. 154. Dembski, S.; Graf, C.; Kruger, T.; Gbureck, U.; Ewald, A.; Bock, A.; Ruhl, E. Small 2008 4 1516 1526. 155. Pechstedt, K. ; Whittle, T.; Baumberg, J.; Melvin, T. J Phys Chem C 2010 114 12069 12077. 156. Lia o, H.; Chen, S.; Liu, D. Macromolecules 2009 42 6558 6563. 157. Masala, S.; Del Gobbo, S.; Borriello, C.; Bizzarro, V.; La Ferrara, V.; Re, M.; Pesce, E.; Minarini, C.; De Crescenzi, M.; Di Luccio, T. J Nanopart Res 2011 13 6537 6544. 158. Ren, S.; Chang, L.; Lim, S. K.; Zhao, J.; Smith, M.; Zhao, N.; Bulovic, V.; Bawendi, M.; Gradecak, S. Nano Lett 2011 11 3998 4002. 159. Oosterhout, S. D.; Wienk, M. M.; van Bavel, S. S.; Thiedmann, R.; Koster, L. J. A.; Gilot, J.; Loos, J.; Schmidt, V.; Janssen, R. A. J. Nat ure Mater 2009 8 818 824. 160. Liu, C.; Holman, Z. C.; Kortshagen, U. R. Nano Lett 2009 9 449 452. 161. Rath, T.; Edler, M.; Haas, W.; Fischereder, A.; Moscher, S.; Schenk, A.; Trattnig, R.; Sezen, M.; Mauthner, G.; Pein, A.; Meischler, D.; Bartl, K.; Saf, R.; Bansal, N.; Haque, S. A.; Hofer, F.; List, E. J. W.; Trimmel, G. Adv Energ. Mater 2011 1 1046 1050. 162. Seo, J.; Cho, M. J.; Lee, D.; Cart wright, A. N.; Prasad, P. N. Adv Mater 2011 23 39 84 3988 163. Jorgensen, M.; Norrman, K.; Krebs, F. C. Sol Energ Mat er. Sol C ells 2008 92 686 714. 164. Lee, J. K.; Coates, N. E.; Cho, S.; Cho, N. S.; Moses, D.; Bazan, G. C.; Lee, K.; Heeger, A. J Appl Phys Lett 2008 92 243308 165. Yip, H. L.; Hau, S. K.; Baek, N. S.; Ma, H.; Jen, A. K. Y. Adv Mater 2008 20 2376 2382 166. Ntwaeaborwa, O. M.; Zhou, R. ; Qian, L.; Pitale, S. S.; Xue, J.; Swart, H. C.; Holloway, P. H. Physica B 2012 407 1 631 1633.
220 167. Qian, L.; Yang, J.; Zhou, R.; Tang, A.; Zheng, Y.; Tseng, T.; Bera, D.; Xue, J.; Holloway, P. H. J Mater Chem 2011 21 3814 3817. 168. Qian, L.; Zheng, Y.; Choudhury, K. R.; Bera, D.; So, F.; Xue, J. G.; Holloway, P. H. Nano Today 2010 5 384 389. 169. Sun, B.; Marx, E.; Greenham, N. C. Nano Lett 2003 3 961 963. 170. Sun, B.; Greenham, N. C. Phys Chem Chem Phys 2006 8 3557 3560. 171. Yang, J.; Qian, L.; Zhou, R.; Zheng, Y.; Tang, A.; Holloway, P. H.; Xue, J. J Appl Phys 2012 111 044323 172. Myers, J. D.; Tseng, T. K.; Xue, J. Org Electron 2009 10 1182 1186. 173. Norrman, K.; Madsen, M. V.; Gevorgyan, S. A.; Krebs, F. C. J Am Chem Soc 2010 132 16883 16892. 174. Scharber, M. C.; Wuhlbacher, D.; Koppe, M.; Denk, P.; Waldauf, C.; Heeger, A. J.; Brabec, C. L. Adv Mater 2006 18 789 794 175. Price, S. C.; Stuart, A. C.; Yang, L.; Zhou, H. ; You, W. J Am Chem Soc 2011 133 4625 4631. 176. Sun, B.; Snaith, H. J.; Dhoot, A. S.; Westenhoff, S.; Greenham, N. C. J App l Phys 2005 97 014914. 177. Zhou, R.; Zheng, Y.; Qian, L.; Yang, Y.; Holloway, P. H.; Xue, J. Nanoscale 2012 4 3507 3514. 178. Dayal, S.; Kopidakis, N.; Olson, D. C.; Ginley, D. S.; Rumbles, G. Nano Lett 2010 10 239 242. 179. Pivrikas, A.; Juska, G.; Mozer, A. J.; Scharber, M.; Arlauskas, K.; Sariciftci, N. S.; Stubb, H.; Osterbacka, R. Phys Rev Lett 2005 94 176806 180. Burkhard, G. F.; Hoke, E. T.; McGehee, M. D. Adv Mater 2010 22 3293 3297 181. Myers, J. D.; Cao, W.; Cassidy, V.; Eom, S. H.; Zhou, R.; Yang, L.; You, W.; Xue, J. Energ Environ Sci 2012 5 6900 6904. 182. Wu, Y.; Zhang, G. Nano Lett 2010 10 1628 1631. 183. Xue, J. ; Rand, B. P.; Uchida, S.; Forrest, S. R. Adv Mater 2005 17 66 71 184. Koleilat, G. I.; Levina, L.; Shukla, H.; Myrskog, S. H.; Hinds, S.; Pattantyus Abraham, A. G.; Sargent, E. H. ACS Nano 2008 2 833 840. 185. Sarasqueta, G.; Choudhury, K. R.; So, F. Chem Mater 2010 22 3496 3501.
221 186. Lee, J. K.; Ma, W. L.; Brabec, C. J.; Yuen, J.; Moon, J. S.; K im, J. Y.; Lee, K.; Bazan, G. C.; Heeger, A. J. J Am Chem Soc 2008 130 3619 3623. 187. von Holt, B.; Kudera, S.; Weiss, A.; Schrader, T. E.; Manna, L.; Parak, W. J.; Braun, M. J Mater Chem 2008 18 2728 2732. 188. Kopping, J. T.; Patten, T. E. J Am Chem Soc 2008 130 5689 5698. 189. Wang, F.; Tang, R.; Kao, J. L. F.; Dingman, S. D.; Buhro, W. E. J Am Chem Soc 2009 131 4983 4994. 190. Morris Cohen, A. J.; Donakowski, M. D.; Knowles, K. E.; Weiss, E. A. J Phys Chem C 2010 114 897 906 191. Gomes, R.; Hassinen, A.; Szczygiel, A.; Zhao, Q. A.; Vantomme, A.; Martins, J. C.; Hens, Z. J Phys Chem Lett 2011 2 145 152. 192. Hao, E.; Sun, H. ; Zhou, Z.; L iu, J. ; Yang, B.; Shen, J. Chem Mater 1999 11 3096 3102. 193. Luther, J. M.; Law M.; Song, Q.; Perkins, C. L.; Beard, M. C.; Nozik, A. J. ACS Nano 2008 2 271 280. 194. Milliron, D. J.; Alivisatos, A. P.; Pitois, C.; Edder, C.; Frechet, J. M. J. Adv Mater 2003 15 58 61 195. Liu, J.; Tanaka, T.; Sivula, K.; Alivisatos, A. P.; Frechet, J. M. J. J Am Chem Soc 2004 126 6550 6551. 196. Briseno, A. L.; Holcombe, T. W.; Boukai, A. I.; Garnett, E. C.; Shelton, S. W. ; Frechet, J. J. M.; Yang, P. Nano Lett 2010 10 334 340 197. Stalder, R.; Xie, D.; Zhou, R.; Xue, J.; Reynolds, J. R.; Schanze, K. S. Chem Mater 2012 24 3143 3152. 198. Park, Y.; Advincula, R. C. Chem Mater 2011 23 4273 4294. 199. Shrotriya, V.; Li, G.; Yao, Y.; Chu, C. W.; Yang, Y. Appl Phys Lett 2006 88 073508 200. Irwin, M. D.; Buchholz, B.; Hains, A. W.; Chang, R. P. H.; Marks, T. J. P roc. Natl Acad Sci U S A 2008 105 2783 2787. 201. Han, S.; Shin, W. S.; Seo, M.; Gupta, D.; Moon, S. J.; Yoo, S. Org Electron 2009 10 791 797. 202. Kim, D. Y.; Subbiah, J.; Sarasqueta, G.; So, F.; Ding, H. J.; Irfan; Gao, Y. L. Appl Phys Lett 2009 95 093304
222 203. Hancox, I.; Rochford, L. A.; Clare, D.; Sullivan, P.; Jones, T. S. Appl Phys Lett 2011 99 013304 204. Girotto, C.; Voroshazi, E.; Cheyns, D.; Heremans, P.; Rand, B. P. ACS Appl Mater Inter face s 2011 3 3244 3247. 205. Choi, H.; Kim, B.; Ko, M. J.; Lee, D. K.; Kim, H.; Kim, S. H.; Kim, K. Org Electron 2012 13 959 968. 206. Jasieniak, J. J.; Seifter, J.; Jo, J.; Mates, T.; Heeg er, A. J. Adv Funct Mater 2012 22 2594 2605. 207. Murase, S.; Yang, Y. Adv Mater 2012 24 2459 2462. 208. Zilberberg, K.; Gharbi, H.; Behrendt, A.; Trost, S.; Riedl, T. ACS Appl Mater Inter face s 2012 4 1164 1168. 209. Jung, J.; Kim, D. L.; Oh, S. H.; Kim, H. J. Sol Energ Mat er. Sol C ells 2012 102 103 108. 210. Jung, J.; Oh, S. H.; Yoon, D. H.; Kim, H. J. J Nanosci Nanotechno 2012 12 1165 1169. 211. Lassner, E.; Schubert, W. D. T ungsten: properties, chemistry, technology of the element, alloys, and chemical compounds Kluwer Academic / Plenum Publishers 1999. 212. Zhao, W.; Mudrick, J. P.; Zheng, Y.; Hammond, W. T.; Yang, Y.; Xue, J. Org Electron 2012 13 129 135.
223 BIOGRAPHICAL SKETCH Renjia Zhou was born in Sanming, Fujian Province China. He studied in Polymer Science and E ngineering and obtained the Bachelor of Science from Zhejiang University at 2006. He continued the study at Zhejiang University and majored in Chemistry for a Master of Science from 2006 to 2008 and worked on ch emistry of colloidal nanocrystals and carbon nanotubes with Prof. Hongzheng Chen and Prof. Mang Wang Then he moved to University of Florida to pursue a Doctor of Philosophy at Department of Materials Science and Engineering since 2008 where he has worked with Prof. Jiangeng Xue and focused on development of colloidal nanocrystals for optoelectronic devices, particularly organic inorganic hybrid photovoltaic cells He has published 20 peer reviewed papers in the field of nanostructured materials and optoel ectronic devices He will head to the Molecular Foundry at Lawrence Berkeley National Laboratory to start a postdoctoral position