ORGANIC PHOTOVOLATIC CELLS AND GRAPHENE …jm479qg5206... · organic photovolatic cells and...
Transcript of ORGANIC PHOTOVOLATIC CELLS AND GRAPHENE …jm479qg5206... · organic photovolatic cells and...
ORGANIC PHOTOVOLATIC CELLS AND GRAPHENE
TRANSPARENT CONDUCTORS
A DISSERTATION
SUBMITTED TO THE DEPARTMENT OF MATERIALS
SCIENCE AND ENGINEERING
AND THE COMMITTEE ON GRADUATE STUDIES
OF STANFORD UNIVERSITY
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
Junbo Wu
March 2011
http://creativecommons.org/licenses/by-nc/3.0/us/
This dissertation is online at: http://purl.stanford.edu/jm479qg5206
© 2011 by Wu Junbo. All Rights Reserved.
Re-distributed by Stanford University under license with the author.
This work is licensed under a Creative Commons Attribution-Noncommercial 3.0 United States License.
ii
I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.
Peter Peumans, Primary Adviser
I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.
Alberto Salleo, Co-Adviser
I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.
Michael McGehee
Approved for the Stanford University Committee on Graduate Studies.
Patricia J. Gumport, Vice Provost Graduate Education
This signature page was generated electronically upon submission of this dissertation in electronic format. An original signed hard copy of the signature page is on file inUniversity Archives.
iii
iv
Abstract
The Organic photovoltaic cell (OPV) is a promising technology because of its
potential for low-cost high-throughput roll-to-roll manufacturing. Significant
improvements have been achieved in power conversion efficiency (PCE) of OPV cells
during last two decades. While recent progress in raising the PCE has been
encouraging, the PCE of organic solar cells is still limited and needs to be improved to
meet the requirement for commercial applications. Further improvements in both
material properties and device architectures are necessary. Photocurrent generation in
an OPV cell is fundamentally different from the process that takes place in their
inorganic counterparts. A detailed understanding of the operation mechanisms of OPV
cells and optimization of the fundamental electronic properties of the system (or
material) are critical. In this work, I will first discuss major factors that limit the
efficiency of bilayer OPV cells, such as exciton binding energy, exciton diffusion
length, charge separation and open-circuit voltage. The exciton binding energy is one
of the key parameters that govern the operation of OPV cells, and determines the
required energy band offset between donor and acceptor, and thus the achievable
open-circuit voltage of the donor-acceptor combination. Exciton diffusion is a main
bottleneck limiting photocurrent of a bi-layer OPV cell, which depends on material
properties and film morphology. The energy loss between optical excitation and
extracted electrical power is mainly due to the energy band offset between donor and
v
acceptor in OPV cells. The PCE limit for single junction OPV cell can be estimated
based on the findings.
In the second part of this work, I will focus on transparent conductors, which are
essential components of thin-film optoelectronic devices. Sputtered Indium-Tin-Oxide
(ITO) is currently the most commonly used transparent electrode material, but it has a
number of shortcomings. There is a clear need for alternative transparent electrodes
whose optical and electrical performance is similar to that of ITO but without its
drawbacks. The next generation transparent conductor should also be lightweight,
flexible, cheap, environmental attractive, and compatible with large-scale
manufacturing methods. I will discuss the possibility of using graphene thin films as a
replacement for ITO. Theoretical estimates indicate that graphene thin films are
promising transparent electrodes for thin-film optoelectronic devices, with an
unmatched combination of sheet resistance and transparency. For the first time, we
demonstrated that solution-processed graphene thin films can serve as transparent
conductive anodes for both OPV cells and organic light-emitting diodes (OLEDs). The
graphene electrodes were deposited on quartz substrates by spin-coating of an aqueous
dispersion of functionalized graphene, followed by a vacuum anneal step to reduce the
sheet resistance. Small molecular weight organic materials and a metal cathode were
directly deposited on the graphene anodes, resulting in devices with a performance
comparable to control devices on ITO transparent anodes. Device modeling has been
explored to compare the performance between graphene-based device and ITO-based
control device. Transfer of graphene films to a foreign flexible substrate was also
vi
demonstrated which opens up new opportunities for low-cost flexible organic opto-
electronics.
vii
Acknowledgements
I am greatly honored to be supervised by my adviser, Prof. Peter Peumans. Peter has
deeply impressed me by his vast knowledge base and endless creative ideas. Peter is
always ready to provide constructive explanation and inspiring suggestion to my
project. In details, Peter taught me how to build up vacuum equipment, showed me
how to make organic solar cells, and guided me set up simulation model. During my
PhD life, Peter offered me various great research opportunities, and gave me virtually
absolute flexibility to work on them. His phenomenal intelligence and pleasant
personality made my PhD life at Stanford fruitful and memorable.
I am indebted to my collaborator Hector Becerril, who was postdoctoral fellow in Prof.
Zhenan Bao’s group and is now assistant professor in Brigham Young University. We
had very pleasant and productive collaboration work on graphene transparent
conductor. I wish the best to his career and his family.
I would like to give my special thanks to my committee members, Prof. Mike
McGehee, Prof. Alberto Salleo, Prof. Paul McIntyre, and Prof. Alexander Dunn. Prof.
Dunn was my defense chair, and I really appreciate his support upon my last minute
request. Prof. McIntyre was in my committee for both PhD qualification exam and
oral defense. I am deeply impressed and inspired by his concrete scientific knowledge,
rigorous research attitude and generous kindness to students. I would also like to thank
viii
Prof. Salleo for hooding me during the commencement as my co-adviser. I took three
classes from Prof. McGehee, and he introduced organic electronics into my life which
later becomes my PhD research focus. I also like to thank Prof. Bianxiao Cui, who
generously offered her help to my defense.
I am indebted to many other professors at Stanford, professors at Tsinghua University
and teachers in high school. I find myself lucky to have the opportunity to be trained
by many great teachers in my school life. Specially, I would like to thank my professor
Cewen Nan at Tsinghua University, who supervised my research for master degree
and helped me establish my professional life.
I am grateful to my fellow group members with whom I have had the pleasure to work
during my Stanford life. Mukul has always been ready to discuss all different kinds of
topics, and generously helped me with many questions and confusions. Shanbin and I
worked together to set up our organic small molecule purification system from scratch,
and we spent so much time to work together and chat about everything in our life.
Seung is a great thinker, and I am always benefited from our discussion. Moreover, I
am lucky to have the opportunity to continue our discussion when we joined the same
company for work. I am glad to share office space with Jung-Yong after we moved to
CISX 218, and our very frequent chatting was really cheerful. I am deeply inspired by
Albert’s enthusiasm about everything he is interested in. I always feel relaxed after
talking with Rostam about so many interesting topics. I admire Kevin’s persistence in
his technology and also like his cake very much. I am also honored to work with
Whitney Gaynor, Serena Faruque, Uraib Aboudi, Trudie Wang, Greyson Christoforo,
ix
Nicholas Sergeant, Chu-En Chang, Kyunglok Kim, Soo Jin Kim and Shrestha Basu
Mallick. I have been lucky to among this great group of people, and our discussion
contributes directly and indirectly to this work in one way or another.
I greatly thank many friends at Stanford who made my life very happy and exciting.
My special thanks should be given to Zhonghua Zhang, Wei Hu, Xiaoyan Li, Xirong
Jiang, Changhua He, Qihua Ran, Jun Wang, Shanbin Zhao, Dujuan Kang, Yongxing
Shen, Caifeng Zhu and Shuhong Liu, with whom I have spent countless time together
to discusse and practice poker, bridge, skiing, cocktail, and many more. I am also
indebted to my old friends from the past, with whom I grew up and from whom I
learned so much. Thank you all!
I would also like to thank Yuanyuan Liu for her support, understanding, and
encouragement since we met each other, and I really appreciate her help for improving
myself. I am deeply indebted to my parents for their unconditional and endless love
and support. They teach me how to be a good man, they encourage me to explore my
interest, they respect the decision I made, and they make me proud. I also like to thank
my sister for giving me confidence and support, and for taking care of my parents
when I am not around. Without their understanding and love, this work would not
have been nearly possible.
Finally, I would like to acknowledge various funding sources such as National Science
Foundation, Semiconductor Research Corporation, the Global Climate and Energy
Project at Stanford (GCEP), and the Center for Advanced Molecular Photovoltaics
x
(CAMP). I also like to thank administrative assistant Patricia Halloran-Krokel for
taking care of all the supporting work.
Thank you all!
Junbo Wu
March 2011 at Stanford
xi
Contents
Abstract .......................................................................................................................... iv
Acknowledgements ...................................................................................................... vii
List of Tables ............................................................................................................... xiv
List of Figures ............................................................................................................... xv
Chapter 1 Introduction ................................................................................................... 1
1.1 Organic electronic materials ................................................................................ 1
1.2 Organic opto-electronic devices .......................................................................... 3
1.3 Transparent conductors ........................................................................................ 7
1.4 Thesis overview ................................................................................................... 9
Bibliography ............................................................................................................. 11
Chapter 2 Organic Photovoltaic Cells ......................................................................... 15
2.1 Cost analysis of OPV system ............................................................................. 15
2.2 How OPV cells work ......................................................................................... 20
2.2.1 OPV cell operation ...................................................................................... 20
2.2.2 Fundamental problems in OPV operation ................................................... 24
2.3 Light absorption ................................................................................................. 25
2.3.1 Light trapping structure ............................................................................... 26
2.3.2 Bulk heterojunction ..................................................................................... 27
2.4 Exciton binding energy ...................................................................................... 29
2.4.1 Theoretical estimation of EB ....................................................................... 31
2.4.2 Measurement from EQE of OPV cell ......................................................... 32
2.4.3 Measurement from photoluminescence ...................................................... 36
2.4.4 Summary ..................................................................................................... 38
2.5 Exciton diffusion length ..................................................................................... 39
xii
2.5.1 Measure LD by EQE .................................................................................... 40
2.5.2 Measure LD by PL quenching ...................................................................... 42
2.6 Exciton dissociation ........................................................................................... 46
2.7 Charge separation .............................................................................................. 46
2.8 Charge collection ............................................................................................... 50
2.9 Voltage loss ........................................................................................................ 51
2.10 Efficiency estimation ....................................................................................... 56
2.11 Conclusion ........................................................................................................ 60
Bibliography ............................................................................................................. 61
Chapter 3 Organic Solar Cells with Solution-Processed Graphene Transparent Electrodes ..................................................................................................................... 67
3.1 Introduction ........................................................................................................ 67
3.2 Unique properties of gaphene ............................................................................ 68
3.3 Fabrication of large area graphene thin film ...................................................... 70
3.3.1 Fabrication method survey .......................................................................... 70
3.3.2 Solution-processed graphene film ................................................................ 72
3.4 Properties of Solution-processed graphene film ................................................ 74
3.4.1 Graphene film morphology ......................................................................... 74
3.4.2 Electrical and optical properties of reduced graphene film ........................ 76
3.5 OPV cells on graphene transparent electrode .................................................... 80
3.5.1 Device fabrication ....................................................................................... 80
3.5.2 OPV cell performance ................................................................................. 81
3.6 Conclusion ......................................................................................................... 85
Bibliography ............................................................................................................. 86
Chapter 4 Organic Light-Emitting Diodes on Solution-Processed Graphene Transparent Electrodes ................................................................................................. 90
4.1 Introduction ........................................................................................................ 91
4.2 Comparison of different transparent candidates ................................................ 92
4.3 Organic light-emitting diodes on graphene transparent electrode ..................... 97
4.3.1 OLEDs device fabrication ........................................................................... 97
xiii
4.3.2 Device measurement ................................................................................... 98
4.3.3 Device perofrmance .................................................................................... 99
4.3.4 Device simulation ...................................................................................... 101
4.4 Status of graphene film by CVD ....................................................................... 107
4.5 Conclusion ........................................................................................................ 109
Bibliography ........................................................................................................... 110
Chapter 5 Modeling of Interface Morphology at Organic/Metal Interface ............... 118
5.1 Introduction ...................................................................................................... 118
5.2 Simulation method ........................................................................................... 119
5.2.1 Basic idea .................................................................................................. 119
5.2.2 Simulation workspace ............................................................................... 119
5.2.3 Monte Carlo method ................................................................................. 122
5.3 Results and discussion ..................................................................................... 122
5.3.1 Substrate temperature / Deposition rate dependence ................................ 123
5.3.2 Diffusion coefficient dependence ............................................................. 126
5.3.3 Effect of drive-in potential ........................................................................ 127
5.3.4 Saturated metal atom within film .............................................................. 130
5.3.5 Cluster size on surface .............................................................................. 131
5.4 Conclusion ........................................................................................................ 132
Bibliography ........................................................................................................... 135
xiv
List of Tables
Table 2.1 Measured effective doping concentration………………………………….50
Table 2.2 Significant development work of OPV cells………………………………59
Table 4.1 Simulated fraction of emitted optical power into the various modes
supported by the OLED structures………………………………….……..104
Table 4.2 Sheet resistance and transmittance for different number of graphene
layer………………………………………………………………….…….108
xv
List of Figures
Fig. 1.1 Chemical structures of popular organic electronic materials. CuPc: copper
phthalocyanine; PTCBI: 3,4,9,10-perylene tetracarboxylic bisbenzimidazole;
BCP: bathocuproine; SubPc: boron subphthalocyanine chloride; C60:
fullerene; Alq3: tris(8-hydroxyquinoline) aluminum; Pentacene; NPD: N,N′-
diphenyl-N,N′-bis (1-naphthyl)-1,1′-biphenyl-4,4″-diamine; PCBM: [6,6]-
phenyl-C61-butyric acid methyl ester; ThCBM: thienyl analogue; MDMO-
PPV: poly[2-methoxy-5- (3,7-dimethyloctyloxy)-p-phenylene]vinylene;
P3HT: poly(3-hexylthiophene) ....................................................................... 2
Fig. 1.2 Examples of Organic Light-emitting Diodes and Organic Photovoltaic Cells . 5
Fig. 1.3 Schematic band diagram and I-V cure of Tang’ device .................................... 6
Fig. 1.4 Transparent conductor is the common component in OLEDs and OPV cells .. 7
Fig. 2.1 Dependence of levelized energy cost on module efficiency for organic
photovoltaic cells. The calculation assumes 1GWp installation in Los
Angeles with average insolation of 5.62kWh/m2/d. The electricity retail price
is plotted as black dashed line. The module cost of OPV cells is assumed be
to $48.8/m2 and $30/m2 respectively, for now and potentially in the neat
future. The circles indicate required module efficiency for OPV system with
different lifetime and module cost, in order to meet grid parity. LEC of
crystalline silicon system with module cost of $1.74/Wp and PCE of 14%,
and LEC of thin film system with module cost of $0.98/Wp and PCE of 9%
are plotted as plus (‘+’) and cross (‘x’) by assuming 25-year lifetime,
respectively.................................................................................................... 18
Fig. 2.2 Schematic of energy band diagram for an organic DA solar cell at flat band
condition. The five steps of photocurrent generation is indicated: (1) Light
xvi
Absorption; (2) Exicton Diffusion; (3) Exciton Dissociation; (4) Charge
Separation; and (5) Charge Collection. LA is typical optical absorption
length in OPV cells, and LED is typical exciton diffusion length near D/A
interface. ........................................................................................................ 21
Fig. 2.3 Light trapping structure: (a) V-shape light trapping structure;15 and (b)
resonant cavity-enhanced configuration by using high-reflectivity mirror16 27
Fig. 2.4 Schematic diagram of bi-layer planar heterojunction OPV cell (left) and bulk
heterojucntion OPV cell (right). Dashed yellow boxes indicate the volume
whitin exciton diffusion length from D/A interface. ..................................... 28
Fig. 2.5 Schematic of energy band diagram of donor-acceptor heterojunction. LUMOE
and HOMOE are the energy offsets of LUMO and HOMO between donor
and acceptor respectively. DBE , and ABE , the binding energy of exciton in
donor and acceptor materials respectively. ................................................... 30
Fig. 2.6 (a) Schematic of Franck-Condon principle. (b) Absorbance and
photoluminescence as a function of photon energy for PTCBI. Lowest
absorption peak and photoluminescence peak were indicated at 1.85eV and
1.58eV, respectively. ..................................................................................... 32
Fig. 2.7 Measured external quantum efficiency (EQE) (open square) as a function of
wavelength for the device of ITO/11nm SubPc/20nm PTCBI/13nm BCP/Ag.
Fitted EQE curves with different surface recombination velocity, S, for the
exciton from PTCBI. S=0 yields the best fit to the measured EQE. S was
assumed infinite for the exciton from SubPc. Exciton diffusion length is 50Å
for SubPc and 30 Å for PTCBI, respectively. Absorption coefficients of
SubPc and PTCBI were plotted for reference. .............................................. 34
Fig. 2.8 Normalized optical field intensity and normalized exciton concentration
within the device of ITO/11nm SubPc/20nm PTCBI/13nm BCP/Ag at
wavelength =585nm. Normalized exciton concentration was plotted by
xvii
assuming different surface recombination velocity, S, for the exciton from
PTCBI. .......................................................................................................... 35
Fig. 2.9 Photoluminescence as a function of wavelength for 10nm PTCBI film, 20nm
1:1 SubPc:PTCBI co-deposition film, and 20nm 1:1 CuPc:PTCBI co-
deposition film. Inset: schematic of exciton dissociation at donor-acceptor
interface. ........................................................................................................ 37
Fig. 2.10 Schematic of film structure for photoluminescence measurement ............... 38
Fig. 2.11 Measured EQE curve (open square) and theoretical fitting (solid line) for (a)
ITO/SubPc/PTCBI/BCP/Ag and (b) ITO/SubPc/C60/BCP/Ag. Independent
contributions from SubPc and C60 are plotted as dashed lines in (b)........... 41
Fig. 2.12 The PL intensity ratio PL1/PL2 of SubPc as function of thickness of SubPc
film (symbol). Fitted curves by using eq. (2.5) yields LD=15A for solid line
and LD=45A for dot-dashed line. Inset: schematic energy level diagram of
SubPc/C60 heterojunction. ............................................................................ 43
Fig. 2.13 Contour of short current density of SubPc/C60 cells as function of SubPc
thickness and C60 thickness. ......................................................................... 45
Fig. 2.14 (a) Effective doping concentration as a function of depletion layer thickness.
Inset: measured capacitance with applied voltage. (b) Schematic energy band
diagram when depletion happens within organic layers at low bias. (c)
Schematic energy band diagram after organic layers are totally depleted at
high bias. ....................................................................................................... 49
Fig. 2.15 Schematic energy band diagram of donor/acceptor heterojunction .............. 52
Fig. 2.16 Voltage loss for different solar cell material systems ................................... 53
Fig. 2.17 Open circuit voltage as function of interface gap for various small molecule
OPV cells. Dashed line indicates Voc = ΔEDA/e - 0.35. ............................... 54
Fig. 2.18 Assumption of efficiency simulation ............................................................ 57
xviii
Fig. 2.19 Simulated efficiency as a function of optical band gap. Best efficiency of
11.6% is predicted at band gap equals 1.65 eV. Significant development
work is shown on the plot for both small molecule and polymer OPV cells.
....................................................................................................................... 58
Fig. 3.1 Graphite crystal and monolayer graphene ....................................................... 69
Fig. 3.2 Graphene film and device made by scotch tape method ................................. 71
Fig. 3.3 Schematic of graphene-on-demand by cut-and-choose transfer-printing
(DCT)18 ......................................................................................................... 72
Fig. 3.4 Schematic of preparing large area graphene thin film by solution process .... 74
Fig. 3.5 AFM images for a reduced thin (<10nm) (a) and thick (>10nm) (b) graphene
film. The scan size is 3m in (a) and 8m in (b). ......................................... 75
Fig. 3.6 Transmittance at =550nm (triangles) and sheet resistance (circles) as a
function of the graphene film thickness for both reduction methods: vacuum
annealing at 1100oC (open symbols), and hydrazine treatment and argon
annealing at 400oC (filled symbols). ............................................................. 77
Fig. 3.7 Relationship between transmission and sheet resistance of reduced graphene
films ............................................................................................................... 78
Fig. 3.8 Transmittance as a function of wavelength for graphene films of various
thicknesses. The graphene films were reduced by the hydrazine treatment
and argon annealing at 400oC........................................................................ 79
Fig. 3.9 Molecule structure of organic materials and device structure of organic
photovoltaic cell ............................................................................................ 81
Fig. 3.10 Current density vs. voltage (J-V) curves for an OPV cell with layer structure
35nm CuPc/50nm C60/10nm BCP/100 nm Ag on graphene (circles) and ITO
(squares) in the dark (filled symbols) and under 85mW/cm2 AM1.5G
simulated solar illumination (open symbols). The sheet resistance and
xix
transmittance of the graphene film are ~100k/sq and 95%, respectively.
Inset: Schematic of the cells. ......................................................................... 83
Fig. 3.11 J-V curves for OPV cells on ITO (squares) and three different graphene
films (circles, triangles, and inverted triangles) under AM1.5G simulated
solar illumination. Using the data for the cell on ITO, the J-V curves of the
cells on graphene are reproduced by adding a resistive voltage drop to the
voltage axis (solid line, dashed line, and short dashed line). The series
resistances used to obtain these fits, R1,fit=110k, R2,fit=180k and
R3,fit=420k, are very close to the measured sheet resistances of the
graphene films used: R1=75k-150k, R2=200k and R3=200k-3M,
respectively.................................................................................................... 84
Fig. 4.1 Film thickness dependence of sheet resistance for different transparent
conductors. Two limiting values were calculated for graphene using =
2x105 cm2/Vs at Ni = 1012 cm-2 and = 104 cm2/Vs at Ni = 3x1012 cm-2. The
thicknesses of the Ag nanowire film and metal grating are mean thicknesses
obtained by spreading the mass of the metal over the entire substrate area. . 93
Fig. 4.2 Solar photon flux-weighted transmission vs. sheet resistance for different
transparent conductors. Two limiting lines of graphene film were calculated
by using two different refractive indexes n = 2.0+1.1i and n = 2.6+1.3i for
graphene, assuming = 2x105 cm2/Vs and Ni = 1012 cm-2. Two horizontal
lines represent the transmission through bare glass (dashed line, n = 1.463)
and PET (Polyethylene terephthalate) (dotted line, n = 1.575) by including
Fresnel reflections at both the air/(glass or PET) and (glass or PET)/air
interfaces. Solar transmission is calculated by integrating the product of the
spectrally resolved transmittance with the spectrally resolved AM1.5 photon
flux over the spectral range of 400-800nm. .................................................. 95
Fig. 4.3 Molecule structure of organic materials and device structure of organic light-
emitting diodes .............................................................................................. 98
xx
Fig. 4.4 Current density (filled symbols) and luminance (open symbols) vs. applied
forward bias for an OLED on graphene (squares) and ITO (circles), with
OLED device structure anode/PEDOT:PSS/NPD(50nm)/Alq3(50nm)/LiF/Al.
..................................................................................................................... 100
Fig. 4.5 External quantum efficiency (EQE) (filled symbols) and luminous power
efficiency (LPE) (open symbols) for an OLED on graphene film (squares)
and ITO glass (circles). The luminance, EQE, and LPE were adjusted to
reflect total front emission power according to the method described in the
Methods. ...................................................................................................... 101
Fig. 4.6 Schematic of light distribution of OLEDs. The light can be distributed into air
modes, substrate modes, wave-guide modes, surface plasma modes, anode
absorption and metal absorption. ................................................................ 102
Fig. 4.7 Measured light spectrum of OLED with the structure as shown in the plot . 103
Fig. 4.8 Simulated angularly-integrated spectral density of emitted power into
unbound (UB) air modes, substrate trapped (ST) modes, wave-guided (WG)
modes, and surface plasmon polariton (SPP) modes. ................................. 104
Fig. 4.9 Simulated spectrally-integrated emitted power per unit solid angle for OLEDs
on graphene film (blue dash-dotted line) and control ITO (red solid line)
glass. A Lambertian response (black dashed line) is plotted for reference. 105
Fig. 4.10 The fraction of power emitted as a function of normalized in-plane wave-
vector kxy for all wavelengths between 400nm and 800nm for an OLED on
ITO (a) and graphene (b), with device structure
anode/PEDOT:PSS/NPD(50nm)/Alq3(50nm)/LiF/Al as shown in Fig. 4.3.
The fraction of power coupled to air modes, substrate modes, wave-guided
modes and surface plasmon polariton modes is evaluated by integrating the
power distribution over the appropriate ranges of the in-plane wave-vector
kxy. .............................................................................................................. 106
xxi
Fig. 4.11 The relationship of solar transmission vs. sheet resistance for ITO and
graphene films. Texture-filled area for theoretically-estimated graphene, blue
crosses for solution-processed graphene, and red circle for CVD-growed
graphene. ..................................................................................................... 108
Fig. 5.1 Cubic array of organic molecules ................................................................. 120
Fig. 5.2 (a) Schematic of metal atom diffusion direction; (b) Probability distribution of
each event on 1D axis.................................................................................. 121
Fig. 5.3 Substrate temperature / deposition rate dependence. Region 1: above surface;
Region 2: organic film; Region 3: bottom electrode ................................... 124
Fig. 5.4 Comparison with experimental data .............................................................. 125
Fig. 5.5 Diffusion coefficient dependence. Keeping surface diffusion = 100, boundary
diffusion is increased from 20 to 50. ........................................................... 127
Fig. 5.6 Influence of drive-in potential ....................................................................... 128
Fig. 5.7 Track of diffusion. Diffusion path start from black dot and stop at green dot.
..................................................................................................................... 130
Fig. 5.8 Number and percent of atoms below surface with increasing deposition layer
..................................................................................................................... 131
Fig. 5.9 Deposition rate dependence of surface morphology ..................................... 132
1
Chapter 1 Introduction
This chapter is a brief introduction of organic electronic materials and organic opto-
electronic devices. The purpose is to provide some most relevant information and
explain the motivation of my work. I will start with some basics of organic electronic
materials, and then explain a brief development history of organic opto-electronic
devices, mainly organic photovoltaic (OPV) cells and organic light-emitting diodes
(OLEDs). The need for transparent conductors in organic opto-electronic devices is
then highlighted. Finally, the chapter is ended by the thesis overview.
1.1 Organic electronic materials
Most organic materials consist of carbon. Organic electronic materials are generally
categorized into two groups: small molecular weight organic materials and polymer.
Small molecular weight compounds consist of several to a few hundred atoms, while
polymers are characterized by a long molecular chain with repeated monomers. Fig.
1.1 shows chemical structures of popular organic electronic materials.
2
C60
PTCBI
SubPc
Pentacene
C60C60
PTCBI
SubPc
Pentacene
Fig. 1.1 Chemical structures of popular organic electronic materials.
CuPc: copper phthalocyanine; PTCBI: 3,4,9,10-perylene tetracarboxylic
bisbenzimidazole; BCP: bathocuproine; SubPc: boron subphthalocyanine
chloride; C60: fullerene; Alq3: tris(8-hydroxyquinoline) aluminum;
Pentacene; NPD: N,N′-diphenyl-N,N′-bis (1-naphthyl)-1,1′-biphenyl-4,4″-
diamine; PCBM: [6,6]-phenyl-C61-butyric acid methyl ester; ThCBM:
thienyl analogue; MDMO-PPV: poly[2-methoxy-5- (3,7-dimethyloctyloxy)-
p-phenylene]vinylene; P3HT: poly(3-hexylthiophene)
3
The common feature of organic electronic materials is the conjugation, that is,
alternation of single and double bonds. The structure represents the sp2 hybridization,
and allows delocalized the electrons among the ring, which forms the foundation of
electrical and optical properties of organic materials. The work in this thesis will
mainly focus on small molecular weight organic materials except for special
clarification.
One big advantage for organic electronic materials is that the electrical and optical
properties, such as conductivity and optical band gap, can be tailored by molecular
design, which will offer significant design freedom. In addition, organic molecule is
self-passivated, so the properties of the materials do not depend on lattice matching or
dangling bonds. Therefore, process cost could potentially be very low, such as spin
coating, spray, roll-to-roll process, and so on. Some potential concerns about organic
electronic materials are low conductivity, stability, high sensitivity and low purity.
The application development of organic electronic materials has been mainly focusing
on organic photovoltaic (OPV) cells, organic light-emitting diodes (OLEDs), organic
photo-detectors, organic laser, organic transistor, organic non-volatile memory, and so
on. OPV cells and transparent conductors for OPV cells and OLEDs will be the focus
of the thesis.
1.2 Organic opto-electronic devices
4
Fig. 1.2 shows some examples of OLED display, OLED lighting and OPV cells.
OLEDs and OPV cells are the most active areas in organic electronics.
Due to its superb image quality, low power consumption and ultra-thin device
structure, OLED has been developed for more than 20 years,1-8 and recently finds its
application in ultra-thin televisions and other display screens such as those on digital
cameras and mobile phones. On the other hand, OPV cell is still under development,
and very few products are available on the market for some niche application. Major
concerns of OPV cells are power conversion efficiency and reliability.
The first efficient bi-layer OPV cell was made by C.W. Tang.9 The cell is made on a
glass substrate coated with transparent conductor, indium tin oxide (ITO). As shown
in Fig. 1.3, copper phthalocynine (CuPc) and perylene tetracarboxylic derivative were
deposited on ITO/glass, and Ag was used as top electrode.
Light is absorbed in both organic layers and excitons formed upon light absorption
diffuse to the heterojunction interface and then dissociate. Holes and electrons travel
through CuPc and perylene tetracarboxylic derivative respectively, and get collected
by ITO and Ag respectively. The current-voltage curve of the device is shown in Fig.
1.3, where JSC is short circuit current density, VOC is the open circuit voltage, and PM is
maximum power point. Fill factor FF is defined via Eq. (1.1), and power conversion
efficiency is defined via Eq. (1.2),
OCSC
M
VJ
PFF
* (1.1)
5
Konarka OPV Cell
Samsung OLED Phone
UDC OLED Flexible Display
Sony OLED TV on Market GE OLED Lighting
Konarka OPV Cell
Samsung OLED PhoneSamsung OLED Phone
UDC OLED Flexible DisplayUDC OLED Flexible Display
Sony OLED TV on MarketSony OLED TV on Market GE OLED LightingGE OLED Lighting
Fig. 1.2 Examples of Organic Light-emitting Diodes and Organic
Photovoltaic Cells
6
ITO
/Gla
ss
Ag
CuPc PTCBI
ITO
/Gla
ss
Ag
ITO
/Gla
ss
Ag
CuPc PTCBIdonor acceptordonor acceptor
Fig. 1.3 Schematic band diagram and I-V cure of Tang’ device
IN
OCSC
IN
M
P
FFVJ
P
P ** (1.2)
where PIN is incident power. Tang introduced heterojunction in his device, which
helps dissociate exciton efficiently, and yield significant improvement of efficiency.
Significant progress has been made after Tang’s device, and achieved efficiency is
above 8% very recently.10 Main improvement comes from better OPV materials,11
7
bulk heterojunction donor/acceptor interface,12-14 light trapping technique,15, 16 and
multi-junction cells.10, 17
1.3 Transparent conductors
Organic optoelectronic devices have been attracting extensive interest because of their
potential to be deposited on flexible, light-weight substrates using low-cost fabrication
methods. An important aspect of optoelectronic thin-film devices is the transparent,
conductive electrode through which light couples in or out of the devices.
Fig. 1.4 Transparent conductor is the common component in OLEDs and
OPV cells
8
As shown in Fig. 1.4, transparent conductor (TCO) is the common component in both
OLEDs and OPV cells. Traditionally, indium tin oxide (ITO) is used in this type of
devices. However, indium is rare, expensive and difficult to recycle. Also metal oxides
such as ITO are brittle and therefore of limited use on flexible substrates.18 Moreover,
Indium is known to diffuse into the active layers of OLEDs, which leads to a
degradation of performance over time.19 There is a clear need for alternative
transparent electrodes whose optical and electrical performance is similar to that of
ITO but without its drawbacks. The next generation of optoelectronic devices requires
transparent conductive electrodes to be lightweight, flexible, cheap, environmental
attractive, and compatible with large-scale manufacturing methods.
Several alternative candidates have been demonstrated, including other transparent
conductive oxides,20, 21 very thin metal films,22 carbon nanotube (CNT) random
meshes,23, 24 metal nanowire random meshes25 and metal gratings.26 While substantial
progress has been made, many issues remain to be addressed, such as performance,
cost, lifetime, roughness, manufacturability, etc. In the second part of the thesis
Graphene, I will demonstrate solution-processed graphene is a promising candidate
due to its unique electrical and optical properties.
9
1.4 Thesis overview
The Organic photovoltaic cell (OPV) is a promising technology because of its
potential for low-cost high-throughput roll-to-roll manufacturing. To further improve
OPV cell performance, a detailed understanding of the operation mechanisms of OPV
cells and optimization of the fundamental electronic properties of the system (or
material) are critical. In this thesis, I will first investigate major factors that limit the
efficiency of bilayer OPV cells, and in the second part, I will focus on graphene
transparent conductors
In chapter 2, I start with cost analysis of OPV module, and then attempt to address
some of the important problems of the fundamental physical processes underlying the
operation of small molecular weight OPV cells, such as exciton binding energy,
exciton diffusion length, charge separation, voltage losses, and open circuit voltage. In
the end, achievable PCE of single junction OPV cells is estimated by using the
parameters concluded from the study.
In chapter 3, we demonstrate that solution-processed graphene thin films can serve as
transparent conductive anodes for organic photovoltaic cells. The graphene electrodes
were deposited on quartz substrates by spin-coating of an aqueous dispersion of
functionalized graphene, followed by a reduction process to reduce the sheet
resistance. Small molecular weight organic solar cells can be directly deposited on
such graphene anodes. The short-circuit current and fill factor of these devices on
graphene are lower than that of control device on indium tin oxide due to the higher
10
sheet resistance of the graphene films. We anticipate that further optimization of the
reduction conditions will improve the performance of these graphene anodes.
In chapter 4, our theoretical estimates indicate that graphene thin films can be used as
transparent electrodes for thin-film devices such as solar cells and organic light-
emitting diodes, with an unmatched combination of sheet resistance and transparency.
We demonstrate organic light-emitting diodes with solution-processed graphene thin
film transparent conductive anodes. The graphene electrodes were deposited on quartz
substrates by spin-coating of an aqueous dispersion of functionalized graphene,
followed by a vacuum anneal step to reduce the sheet resistance. Small molecular
weight organic materials and a metal cathode were directly deposited on the graphene
anodes, resulting in devices with a performance comparable to control devices on
indium-tin-oxide transparent anodes. The outcoupling efficiency of devices on
graphene and indium-tin-oxide is nearly identical, in agreement with model
predictions.
In chapter 5, a random walk model is proposed to simulate the interface morphology
when depositing metal atoms on the organic film. The paper mainly discusses the
influences of substrate temperature, deposition rate, diffusion coefficient and drive-in
potential. In comparison with experimental data, this simple model gives reasonable
interface morphology and metal distribution profile.
11
Bibliography
1. Tang, C. W.; VanSlyke, S. A. Organic electroluminescent diodes. Appl. Phys. Lett.
1987,51,913-915.
2. Baldo, M. A.; O'Brien, D. F.; You, Y.; Shoustikov, A.; Sibley, S.; Thompson, M.
E.; Forrest, S. R. Highly efficient phosphorescent emission from organic
electroluminescent devices. Nature 1998,395,151-154.
3. Sun, Y.; Giebink, N. C.; Kanno, H.; Ma, B.; Thompson, M. E.; Forrest, S. R.
Management of singlet and triplet excitons for efficient white organic light-emitting
devices. Nature 2006,440,908-912.
4. B. W. D'Andrade, R. J. Holmes,S.R.Forrest, Efficient Organic
Electrophosphorescent White-Light-Emitting Device with a Triple Doped Emissive
Layer. Adv Mater 2004,16,624-628.
5. Burroughes, J. H.; Bradley, D. D. C.; Brown, A. R.; Marks, R. N.; Mackay, K.;
Friend, R. H.; Burns, P. L.; Holmes, A. B. Light-emitting diodes based on conjugated
polymers. Nature 1990,347,539-541.
6. Friend, R. H.; Gymer, R. W.; Holmes, A. B.; Burroughes, J. H.; Marks, R. N.;
Taliani, C.; Bradley, D. D. C.; Santos, D. A. D.; Bredas, J. L.; Logdlund, M.; Salaneck,
W. R. Electroluminescence in conjugated polymers. Nature 1999,397,121-128.
12
7. He, G.; Pfeiffer, M.; Leo, K.; Hofmann, M.; Birnstock, J.; Pudzich, R.; Salbeck, J.
High-efficiency and low-voltage p-i-n electrophosphorescent organic light-emitting
diodes with double-emission layers. Appl. Phys. Lett. 2004,85,3911-3913.
8. Li, J.; Hu, L.; Wang, L.; Zhou, Y.; Gruner, G.; Marks, T. J. Organic Light-Emitting
Diodes Having Carbon Nanotube Anodes. Nano Letters 2006,6,2472-2477.
9. Tang, C. W. Two-layer organic photovoltaic cell. Appl. Phys. Lett. 1986,48,183-
185.
10. http://www.heliatek.com/ 2010.
11. Peumans, P.; Forrest, S. R. Very-high-efficiency double-heterostructure copper
phthalocyanine/C[sub 60] photovoltaic cells. Appl. Phys. Lett. 2001,79,126-128.
12. Peumans, P.; Uchida, S.; Forrest, S. R. Efficient bulk heterojunction photovoltaic
cells using small-molecular-weight organic thin films. Nature 2003,425,158-162.
13. Yang, F.; Shtein, M.; Forrest, S. R. Controlled growth of a molecular bulk
heterojunction photovoltaic cell. Nat. Mater. 2005,4,37-41.
14. Yang, F.; Sun, K.; Forrest, S. R. Efficient Solar Cells Using All-Organic
Nanocrystalline Networks. Adv Mater 2007,19,4166-4171.
15. Rim, S.; Zhao, S.; Scully, S. R.; McGehee, M. D.; Peumans, P. An effective light
trapping configuration for thin-film solar cells. Appl. Phys. Lett. 2007,91,243501-3.
13
16. Agrawal, M.; Peumans, P. Broadband optical absorption enhancement through
coherent light trapping in thin-film photovoltaic cells. Opt. Express 2008,16,5385-
5396.
17. Xue, J.; Uchida, S.; Rand, B. P.; Forrest, S. R. Asymmetric tandem organic
photovoltaic cells with hybrid planar-mixed molecular heterojunctions. Appl. Phys.
Lett. 2004,85,5757-5759.
18. Chen, Z.; Cotterell, B.; Wang, W.; Guenther, E.; Chua, S. A mechanical
assessment of flexible optoelectronic devices. Thin Solid Films 2001,394,201-205.
19. Lee, S. T.; Gao, Z. Q.; Hung, L. S. Metal Diffusion From Electrodes in Organic
Light-Emitting Diodes. Appl. Phys. Lett. 1999,75,1404.
20. Kim, H.; Gilmore, C. M.; Horwitz, J. S.; Pique, A.; Murata, H.; Kushto, G. P.;
Schlaf, R.; Kafafi, Z. H.; Chrisey, D. B. Transparent conducting aluminum-doped zinc
oxide thin films for organic light-emitting devices. Appl. Phys. Lett. 2000,76,259-261.
21. J. Cui, A. Wang, N. L. Edleman, J. Ni, P. Lee, N. R. Armstrong,T.J.Marks,
Indium Tin Oxide Alternatives - High Work Function Transparent Conducting Oxides
as Anodes for Organic Light-Emitting Diodes. Adv Mater 2001,13,1476-1480.
22. Pode, R. B.; Lee, C. J.; Moon, D. G.; Han, J. I. Transparent conducting metal
electrode for top emission organic light-emitting devices: Ca--Ag double layer. Appl.
Phys. Lett. 2004,84,4614-4616.
14
23. Hu L.; Gruner G.; Gong J.; Kim C.; Hornbostel B. Electrowetting Devices with
Transparent Single-Walled Carbon Nanotube Electrodes. Appl. Phys. Lett.
2007,90,093124/1.
24. Zhang, D.; Ryu, K.; Liu, X.; Polikarpov, E.; Ly, J.; Tompson, M. E.; Zhou, C.
Transparent, Conductive, and Flexible Carbon Nanotube Films and Their Application
in Organic Light-Emitting Diodes. Nano Letters 2006,6,1880-1886.
25. Lee, J. -Y.; Connor, S. T.; Cui, Y.; Peumans, P. Solution-Processed Metal
Nanowire Mesh Transparent Electrodes. Nano Lett. 2008,8,689-692.
26. M.-G. Kang, L. J. G., Nanoimprinted Semitransparent Metal Electrodes and Their
Application in Organic Light-Emitting Diodes. Adv Mater 2007,19,1391-1396.
15
Chapter 2 Organic Photovoltaic Cells
Organic photovoltaic (OPV) cells are a promising technology because of their
flexibility, light-weight, and potential low-cost high-throughput roll-to-roll production.
However, higher power conversion efficiency (PCE) needs to be realized in order for
OPV cells to be commercially attractive. In this chapter, I start with cost analysis of
OPV module, and then attempt to address some of the important problems of the
fundamental physical processes underlying the operation of small molecular weight
OPV cells, such as exciton binding energy, exciton diffusion length, charge separation,
voltage losses, and open circuit voltage. In the end, achievable PCE of single junction
OPV cells is estimated by using the parameters concluded from the study.
2.1 Cost analysis of OPV system
Photovoltaic (PV) cells are one of major renewable energy resources, and are expected
to supply a great portion of total power for all the human beings. While conventional
PV technologies, such as crystalline silicon solar cells, amorphous silicon solar cells,
CdTe thin film solar cells, CuInGaSe2 (CIGS) solar cells, and III-VI concentration
solar cells, have been making tremendous growth in production and installation in last
decade, the cost of electricity produced by PV technology is still significantly higher
16
than that of traditional electricity. In order to reach grid parity, substantial reduction in
PV system cost needs to be achieved. The cost of an installed PV system consists of
the cost of PV modules, and the other items so called balance of system (BOS), such
as module support frame, inverter, wiring, labor, and so on. A large fraction of BOS
cost of a PV system is proportion to installation area. Therefore, higher PCE of PV
module will yield lower BOS of the system per kilo Walt hour (kWh). All the current
PV manufactures are trying to further reduce PV module production cost, and improve
PCE at the same time.
OPV cells are one of the most emerging PV technologies, and are considered to
be third generation PV technology with the potential of significantly low cost. In fact,
the module cost of OPV cells could potentially be much lower than that of
conventional PV technologies, which is true for both solution-processed polymer OPV
cells,1 and vacuum-deposited small molecule OPV cells.2 Therefore, it is very
interesting to discuss the electricity cost produced by OPV system. As shown in Fig.
2.1, levelized energy cost (LEC) of a PV system is plotted as a function of module
efficiency for different system lifetime, assuming that the PV system is installed in
Los Angeles with 1 GW peak capacity (1GWp).
Levelized energy cost (LEC), C, is expressed as3
OMiPP
A CCfBOSI
BOSMC
0
0
1
(2.1)
where I0 (=1000W/m2) is the peak intensity of Sun, fi is the fraction of indirect cost
such as marketing and management overhead, and COM is the cost per kWh for
17
operation and maintenance (O&M). COM and fi are assumed to be 0.05¢/kWh and 10%
for the 1GWp system. BOSP is the BOS cost that scales linearly with total installed
Wp, and BOSA is the BOS cost that scales linearly with total installed module, which
are assumed to be $0.27/Wp and 40/m2 for a 1GWp system respectively.3. C0 is LEC
of a $1/Wp system without O&M and is obtained by
1)1(
)1(00
n
n
insDA r
rr
q
IC
(2.2)
where qins is insolation, ηDA is conversion efficiency from DC to AC, r is interest rate
and n is lifetime of the system. In this simulation, ηDA=0.8, r=7% and n=10 years are
assumed.
Based on the optimistic estimation for the cost of pure organic solar cells,4 the OPV
module cost M is assumed to be $48.8/m2, among which the material costs $23.4/m2
and ITO costs $7.9/m2 by itself. As seen in Fig. 2.1, levelized energy cost keeps
decreasing with module efficiency. To estimate the system requirement for matching
grid parity, the current retail price of electricity in Los Angeles, 11.5¢/kWh, is plotted
in Fig. 1.1 as dashed line.5 For a 5-year lifetime of OPV system, PCE of OPV module
needs to be 20% to meet grid parity. However, if the lifetime of OPV module could
reach 10 years or even 25 years, the requirement for OPV module efficiency is 9.5%
and 5%, respectively. Therefore, both PCE and reliability of OPV module are very
important in economical point of view.
18
0 5 10 15 20 250
5
10
15
20
25
30
X
+
7.5%5%
9.5% 20%
Optimistic estimation for the cost ofpure organic solar cell: $48.8/m2
* Material Cost: $23.4/m2
* ITO: $7.9/m2
electricity retail price: 11.5 cents/kwh
Los Angeles (ave insolation: 5.62 kwh/m2/d), 1GW installation
10y, $30
25y, $48.85y, $48.8
10y, $48.8
Leve
lized
Ene
rgy
Cos
t (c
ent
s/kw
h)
Module Efficiency (%)
Fig. 2.1 Dependence of levelized energy cost on module efficiency for
organic photovoltaic cells. The calculation assumes 1GWp installation in Los
Angeles with average insolation of 5.62kWh/m2/d. The electricity retail price
is plotted as black dashed line. The module cost of OPV cells is assumed be
to $48.8/m2 and $30/m2 respectively, for now and potentially in the neat
future. The circles indicate required module efficiency for OPV system with
different lifetime and module cost, in order to meet grid parity. LEC of
crystalline silicon system with module cost of $1.74/Wp and PCE of 14%,
and LEC of thin film system with module cost of $0.98/Wp and PCE of 9%
are plotted as plus (‘+’) and cross (‘x’) by assuming 25-year lifetime,
respectively.
19
The cost of $48.8/m2 for pure organic solar cells is estimated based on the current
status of materials and fabrication. Significant reduction of OPV module is anticipated
with increasing maturity of the technology and effort of exploring low-cost materials
and processes. If the cost of OPV module could be reduced to $30/m2, which is not
aggressive at all, the module efficiency needs only to be 7.5% in order to meet grid
parity by assuming system lifetime is 10 years. As a reference, the highest efficiency
of single-junction OPV cells has been improved to be over 8% very recently.6, 7
According to solarbuzz.com, lowest module price is $1.74/Wp for crystalline silicon
solar cells with module PCE of 14%, and $0.98/Wp for thin film solar cells with
module PCE of 9%, respectively. Also in Fig 2.1, levelized energy cost is indicated as
plus (‘+’) for the crystalline cells and cross (‘x’) for the thin film cells, by assuming
the system warrant is 25-year lifetime. We can see that both types of PV technologies
need improve further in order to meet grid parity even for Los Angeles, where
insolation is high and electricity is relatively expensive. For many other locations with
less insolation and electricity cost, more improvement needs to be realized in order for
PV system to match local grid parity.
Based on this analysis, three major parts of OPV cells need to be addresses, which are
high efficiency, low fabrication cost and long lifetime. Therefore, the rest of the
chapter will focus on understanding fundamental physics problem during OPV
operation, which is ultimately essential for improving PCE of OPV cells. In next two
20
chapters, I will discuss the possibility to use low-cost graphene film to replace ITO,
which is a big portion of material cost in OPV cells as mentioned previously.
2.2 How OPV cells work
Organic photovoltaic cells (OPV) have been attracting many interests because of their
flexibility, light weight, easy fabrication and potential low-cost production. Great
progress toward high efficiency has been made since the first donor/acceptor (D/A)
heterojunction cell was reported by C. W. Tang.8 Power conversion efficiency has
been increasing from 0.95% to more than 8% progressively,6, 9-11 however, it is still
less than current conventional PV technologies, and needs to be improved further.12 A
detailed understanding and optimization of fundamental electronic properties of the
system (or material) is indeed very critical. In this section, I will first describe the
fundamental operation mechanism of OPV cells, and then discuss the problems in
each physics step.
2.2.1 OPV cell operation
Photocurrent generation in OPV cells is fundamentally different from the counterpart
process in inorganic PV devices. The basic operation of a single junction OPV cell is
illustrated by using energy band diagram in Fig. 2.2. A basic single junction OPV cell
21
consists of two active organic layers, so called donor material and acceptor material,
both of which are responsible for generating photocurrent. Five steps are employed to
characterize this process as shown in the Fig. 2.2.
LA~LED
cathode
1 4 5
donor acceptor
anode cathode
3
2
LALA~LED~LED
cathode
11 4 5
donor acceptor
anode cathode
3
2
Fig. 2.2 Schematic of energy band diagram for an organic DA solar cell
at flat band condition. The five steps of photocurrent generation is indicated:
(1) Light Absorption; (2) Exicton Diffusion; (3) Exciton Dissociation; (4)
Charge Separation; and (5) Charge Collection. LA is typical optical
absorption length in OPV cells, and LED is typical exciton diffusion length
near D/A interface.
22
1. Light absorption. Upon the absorption of incident light, organic molecule will
transform itself from ground state to excited state, when electron-hole pair is
created. Right after its formation, the electron-hole pair will relax into a bound
state, which is called exciton. The binding energy of exciton prevents itself from
separation in absence of energy favorable condition, such as D/A (donor/acceptor)
interface and strong electrical field within the layer. Exciton usually sits on one
single molecule, and could either recombine or randomly diffuse to the nearby
molecule. This will lead to the next step, exciton diffusion.
2. Exciton diffusion. Exciton in organic layer could diffuse around via hopping
among adjacent molecules. The exciton diffusion process is random in terms of
direction. Presence of electrical filed will not help orient the diffusion direction,
because each exciton consists of one electron and one hole and appears neutral.
Average distance of exciton diffusion before its recombination is defined as
exciton diffusion length, LED. Exciton diffusion length is relatively short, about 5-
10nm for most of organic materials. If an exciton could meet D/A interface or
strong electrical field within its diffusion length, it will separate into electron and
hole.
3. Exciton dissociation. In a typical bi-layer OPV cell, electrical field within each
layer, donor or acceptor, is usually not strong enough to separate exciton.
Therefore, D/A interface is very essential to dissociate an exciton in OPV cells.
When an exciton in donor material diffuse to the D/A interface, the exciton will
transfer an electron to adjacent acceptor molecule, and leave the hole behind, as
shown in step 3 in Fig. 2.1. Similar process will happen when an exciton in
23
acceptor material diffuse to D/A interface. However, the energy band offset needs
to be larger than exciton binding energy in order to for exciton dissociation to
happen efficiently. Exciton dissociation will generate a electron polaron in the
lowest unoccupied molecular orbital (LUMO) of an acceptor molecule and a hole
polaron in the highest occupied molecular orbital (HOMO) of a donor molecule.
The electron and the hole generated via exciton dissociation is called geminate
electron-hole pair (GEHP), and they are still weakly bounded depending on the
initial separation distance.
4. Charge separation. It is still possible for GEHP to recombine due to Coulomb
attraction, although the diffusion process will help electron and hole further
separate because of D/A interface constrain. The direction and magnitude of
electrical field at D/A interface is very critical. If strong electrical field, e.g.,
106V/cm2, is present at D/A interface and also its direction helps electron and hole
diffuse away from the interface, charge separation will be very efficient.13, 14
Otherwise, the chance of GEHP recombination will be higher.
5. Charge collection. After electron and hole are completely separated from D/A
interface, they could travel through the organic layer and get collected by
corresponding electrodes. The process is usually very efficient for typical
operation condition. The electron and hole could still be recombined potentially
during its travel in organic layer, but the chance is generally very small.
Photocurrent generation in OPV cells can be described in five critical steps as
discussed above, and each step has a yield or quantum efficiency (QE) associated with
it. QE is generally defined as the ration of number out to number in. Internal quantum
24
efficiency of an OPV cell, ηIQE, is the ratio of the number of electrons that contribute
to photocurrent to the number of incident photons that are absorbed by active organic
layers. Similarly, external quantum efficiency of an OPV cell, ηEQE, is the ratio of the
number of electrons that contribute to photocurrent to the number of incident photons.
ηIQE measures the ability of OPV cell to convert absorbed photons into photocurrent
without considering absorption loss. The relationship is defined as follows.
CCCSCTEDAIQEAEQE (2.3)
Where ηA is optical absorption efficiency, ηED is the exciton diffusion efficiency, ηCT
is the exciton dissociation efficiency (also called charge transfer efficiency), ηCS is the
GEHP separation efficiency, and ηCC is the carrier collection efficiency.
The photocurrent under short-circuit conditions can be expressed,
dSqJ EQESC )( (2.4)
where S(λ) is the AM1.5 solar spectrum expressed in number of photons per unit area
per unit time per unit wavelength. Open circuit voltage (VOC) is determined together
by photocurrent and dark current of the cell.
2.2.2 Fundamental problems in OPV operation
As described in last section, photocurrent generation process in OPV cell is quite
different from conventional inorganic P-N junction. Till now, many fundamental
problems in OPV cell remains unclear. I will raise some of critical problems in this
sub-section, and try to address them in the following sections.
25
1. How to resolve trade-off between absorption and exciton diffusion?
2. How to determine the exciton binding energy?
3. What affects the exciton diffusion length?
4. Are the organic materials electrically doped?
5. What the voltage loss in an OPV cell?
6. What is the best possible single junction OPV cell efficiency?
2.3 Light absorption
Typical light absorption coefficient (α) of OPV materials is ~105/cm, with
absorption spectra width of ~200nm. To absorb incident light efficiently, organic
active layer needs to be thicker than optical absorption length, LA, which is calculated
from absorption coefficient, /1AL . Therefore, an ideal thickness of organic layer
is >100nm in order to absorb the majority of incident light. However, exciton created
by light absorption has to diffuse to D/A interface to dissociate. As I discussed in last
section, typical exciton diffusion length is only 5-10nm, which is much less than
optical absorption length. Therefore, the majority of exciton generated upon light
absorption will recombine within one organic layer for thick OPV cell, and thus will
not contribute to photocurrent. How to balance the conflict requirement of light
absorption and exciton diffusion?
26
2.3.1 Light trapping structure
One solution is to use light trapping technique on a very thin OPV device. Thin
device could increase the probability of exciton diffusing to D/A interface upon
absorption, and light trapping structure will allow un-absorbed light to bounce within
the device structure and gain multiple chances to be absorbed eventually.
Fig. 2.3a shows V-shape light trapping geometry.15 The active layer and
reflective metal electrode are deposited on a V-shaped transparent substrate coated
with a transparent electrode such as indium tin oxide (ITO). Incident optical rays
bounce off the solar cell structure multiple times as illustrated in the figure. The
scheme does not require modification of device structure itself, but fabricates the solar
cell on a V-shaped substrate with much larger scale than device thickness. A 52%
increase in power conversion efficiency was demonstrated experimentally. The
structure is applicable to a broad class of low-cost thin film solar cells.
Fig. 2.3b shows another light trapping scheme to enhance optical absorption of
thin OPV cell, by inserting a two-component aperiodic dielectric stack into the OPV
device structure.16 The stack is a generalized anti-reflection coating used in PV
technology. Specific designs for thin-film OPV cells increase the photocurrent under
AM1.5 illumination, averaged over all incident angles and polarizations, by up to 40%.
27
With low-cost techniques to manufacture such thin-film dielectric coatings, this
approach may be of practical use to enhance the efficiency of thin-film solar cells.
Some other light trapping techniques have been explored as well, such as metal
grating, buried nanoelectrodes, and scattered elements.15 However, the typical
required feature size is comparable to organic layer thickness, which makes it
challenging for low cost fabrication.
(b)(b)
Fig. 2.3 Light trapping structure: (a) V-shape light trapping structure;15
and (b) resonant cavity-enhanced configuration by using high-reflectivity
mirror16
2.3.2 Bulk heterojunction
In a bi-layer planar structure, only the exciton within the distance of exciton
diffusion length from D/A interface could diffuse to interface, dissociate, separate, and
thus contribute to photocurrent, as shown in the left of Fig. 2.4. On the other hand,
28
active organic layer needs to be thick enough to absorb the majority of incoming light.
One effective solution is to use bulk heterojunction structure as shown in the right of
Fig. 2.4. Bulk heterojunction structure is characterized by a 3-dimention interdigitate
D/A network, providing a large interface area for exciton to dissociate. As indicated
by dashed box, excitons are always generated within a diffusion length of a D/A
interface, which potentially leads to a higher cell efficiency than a planar
heterojunction. All the polymer OPV cells have bulk heterojunction structure by
nature. For small molecule OPV cell, special deposition techniques can be used to
produce bulk heterojunction structure, and significant improvements have been
achieved over their counterpart of planar structure.11, 13, 17, 17
ITO/Glass ITO/Glass
AgAg
donor
acceptor
ITO/Glass ITO/Glass
AgAg
donor
acceptor
Fig. 2.4 Schematic diagram of bi-layer planar heterojunction OPV cell
(left) and bulk heterojucntion OPV cell (right). Dashed yellow boxes indicate
the volume whitin exciton diffusion length from D/A interface.
29
Most commonly-used small molecule materials can only absorb photons with energy
larger than 1.7eV. However, about half of sunlight occurs with photon energy smaller
than 1.7eV. Investigation of low-bandgap OPV cells is very active in recent years,
which could potentially be employed as a sub-cell for multi-junction OPV cell to
improve overall power conversion efficiency. Tin(II) phthalocyanine (SnPc),18
Chloroaluminum phthalocyanine (ClAlPc),19 and lead phthalocyanine (PbPc)20 were
demonstrated very promising as low-bandgap donor material.
2.4 Exciton binding energy
The exciton binding energy (EB) is one of the key parameters that govern the
operation of OPV cells. Exciton is a bound electron-hole pair, and EB is the energy
needed to separate an exciton into free electron and hole. Reported EB ranges from 0.4
to 1.4 for organic materials,21, 22 which is much larger than thermal energy at room
temperature, so exciton dissociation probability is pretty low in a homogeneous
organic material unless a strong electrical field is applied.13 In a typical efficient OPV
cells, donor-acceptor (DA) heterojunction interface is required in order to separate
excitons generated upon light absorption.
As shown in Fig. 2.5, exciton in donor (acceptor) material will dissociate at D/A
interface efficiently if the energy offset of LUMO, LUMOE ( HOMOE ), is larger than
the binding energy DBE , ( ABE , ). Otherwise, exciton dissociation is much less efficient.
30
Therefore, by varying the energy offset between donor and acceptor, one could
measure exciton binding energy of a particular material roughly.
3,4,9,10-perylene tetracarboxylic bisbenzimidazole (PTCBI) is one of the most
important acceptor materials in small molecule OPV cells, and significant
improvement has been achieved for copper phthalocyanine (CuPc)/PTCBI
heterojunction structure.8, 11, 23 In this section, we use external quantum efficiency
(EQE) and photoluminescence to estimate the EB of PTCBI. The importance of energy
offset at DA interface of OPV cells is discussed.
EHOMO> EB, A ?
LUMO
HOMO
ELUMO> EB, D ?
Donor Exciton
LUMO
HOMO
ELUMO> EB, D ?
Donor Exciton Acceptor Exciton
Fig. 2.5 Schematic of energy band diagram of donor-acceptor
heterojunction. LUMOE and HOMOE are the energy offsets of LUMO and
HOMO between donor and acceptor respectively. DBE , and ABE , the binding
energy of exciton in donor and acceptor materials respectively.
31
2.4.1 Theoretical estimation of EB
Exciton is created upon photon absorption by organic molecule, and then
followed by molecule adaptation from both surrounding molecules and excited
molecule itself. Therefore, EB consists of two parts: electrostatic potential due to
charge attraction between electron and hole, and relaxation energy due to molecule
adaptation. The size of many popular small molecules, such as PTCBI, C60, CuPc,
and boron subphthalocyanine (SubPc), is on the order of 1 nm. For a Frenkel exciton,
where exciton exists on one single molecule, electron and hole are separated by 1nm
or so. Electrostatic potential of exciton is estimated to be 0.4±0.1eV by
using reE 02 4/ , given dielectric constant of organic materials is 0 = 3-4.
Lattice relaxation energy can be estimated from the energy shift between
absorption peak and photoluminescence peak, so-called Franck-Condon principle as
shown in Fig. 2.6a. Under illumination, organic molecule could be excited to different
energy levels of exciton state, which is corresponding to the multiple absorption peaks,
and then it would relax back to the lowest energy level of the exciton state by thermal
vibration. As next step, the molecule could further relax due to molecule adaptation to
exciton state, which creates Franck-Condon shift, the difference between the lowest
absorption peak and the highest luminescence peak. Fig. 2.6b shows the absorbance
and photoluminescence of PTCBI thin film deposited on glass subtract. The lowest
absorption peak of PTCBI is located at 1.85eV ( = 672nm), while the
photoluminescence peak is at 1.58eV ( = 785nm). The Franck-Condon shift, 1.85-
32
1.58 = 0.27eV, represents lattice relaxation energy. Therefore, EB of PTCBI is
estimated to be 0.7±0.1eV by including both electrostatic potential and molecule
relaxation energy.
Relaxation
Ground State
Exciton State
Ene
rgy
Nuclear Coordinates
AbsorptionLuminescence
1.5
1.8
2.1
2.4
2.7
1.58eV
1.85eV
Photon Energy (eV)
Absorbance/Photoluminescence (a.u.)
Absorbance Photoluminescence
Fig. 2.6 (a) Schematic of Franck-Condon principle. (b) Absorbance and
photoluminescence as a function of photon energy for PTCBI. Lowest
absorption peak and photoluminescence peak were indicated at 1.85eV and
1.58eV, respectively.
2.4.2 Measurement from EQE of OPV cell
Bi-layer small molecule OPV cells were made as a detector to probe EB of
PTCBI. CuPc and SubPc were used as donor material. ΔEHOMO of CuPc/PTCBI is
(a) (b)
33
0.9eV given HOMOCuPc = 5.2eV and HOMOPTCBI = 6.1eV,24 while ΔEHOMO of
SubPc/PTCBI is 0.5eV given HOMOSubPc = 5.6eV25 and HOMOPTCBI = 6.1eV. OPV
cells were fabricated on commercially obtained 130nm-thick ITO on glass (sheet
resistance <20/sq). All organic materials were obtained commercially and then
purified using thermal gradient sublimation. The organic thin films and metal cathode
were deposited at room temperature by thermal evaporation in high vacuum (~10-
7Torr). The layer structure of the devices is ITO/donor/PTCBI/bathocuproine
(BCP)/Ag. The Ag cathode was deposited through a shadow mask with circular
openings of 0.81 mm2. EQE was measured experimentally, and then fitted by transfer
matrix method developed for organic multilayer structure.26, 27
Fig. 2.7 shows measured EQE of SubPc/PTCBI cell, and fitting curves using
different surface recombination velocity from PTCBI, where surface recombination
velocity, S, indicates that how fast (or efficient) exciton would dissociate at the
interface and contribute to photocurrent. As seen in Fig. 2.7, the best fitting was
achieved when S = 0 is applied, which means the exciton from PTCBI would not
dissociate at SubPc/PTCBI interface. This is because the condition for exciton
dissociation is not satisfied since ΔEHOMO (0.5eV) of SubPc/PTCBI is smaller than EB
(0.7eV) of PTCBI. Fig. 2.8 shows the normalized optical field intensity and
normalized exciton concentration within the OPV structure of
ITO/SubPc/PTCBI/BCP/Ag under illumination at =585nm. Normalized exciton
concentration was plotted at different S. Exciton concentration at DA interface within
34
PTCBI layer goes to zero when S is infinite, while it is proportional to normalized
field within PTCBI layer when S = 0.
400 500 600 700 8000.00
0.05
0.10
0.15
0.20
0.25
0.30
0
1x105
2x105
3x105
4x105
5x105
Increasing interfacedissociation efficiencyof exciton from PTCBI
(c
m-1)
EQ
E
Wavelength (nm)
measured EQE S=0 S=1 S=inf
LD
SubPc = 50A
LD
PTCBI = 30A
PTCBI abs
SubPc abs
Fig. 2.7 Measured external quantum efficiency (EQE) (open square) as a
function of wavelength for the device of ITO/11nm SubPc/20nm
PTCBI/13nm BCP/Ag. Fitted EQE curves with different surface
recombination velocity, S, for the exciton from PTCBI. S=0 yields the best fit
to the measured EQE. S was assumed infinite for the exciton from SubPc.
Exciton diffusion length is 50Å for SubPc and 30 Å for PTCBI, respectively.
Absorption coefficients of SubPc and PTCBI were plotted for reference.
35
Our EQE measurement and fitting of ITO/CuPc/PTCBI/BCP/Ag (not shown)
indicates that S is infinite for CuPc/PTCBI interface, which is constant with others.26
This is naturally explained because ΔEHOMO (0.9eV) of CuPc/PTCBI is larger than EB
(0.7eV) of PTCBI. From the comparison between CuPc/PTCBI and SubPc/PTCBI,
energy offset of DA heterojunction of OPV cells is very important. On one hand,
energy offset can not be too small to dissociate exciton at DA interface, which
decreases photocurrent otherwise. On the other hand, smaller energy offset is expected
in order to have high VOC. Energy offset needs to be carefully optimized to maximize
PCE of OPV cells based on the knowledge of EB of organic molecules.
Sub
Pc
(11n
m)
ITO
Ag
BC
P (
13nm
)
PT
CB
I (20
nm)
Normalized Optical Field Intensity
Normalized Exciton Concentration optical field intensity S = inf S = 0 0 < S < inf
Fig. 2.8 Normalized optical field intensity and normalized exciton
concentration within the device of ITO/11nm SubPc/20nm PTCBI/13nm
BCP/Ag at wavelength =585nm. Normalized exciton concentration was
plotted by assuming different surface recombination velocity, S, for the
exciton from PTCBI.
36
2.4.3 Measurement from photoluminescence
Photoluminescence was employed to further verify the EB of PTCBI. Fig. 2.9
shows photoluminescence as a function of wavelength for 10nm PTCBI film, 20nm
1:1 SubPc:PTCBI co-deposition film, and 20nm 1:1 CuPc:PTCBI co-deposition film.
As shown in Fig. 2.10, the films were deposited on glass substrates as described above,
and glass/organic interface is generally assumed to be non-quenching interface. The
films were excited by a laser at wavelength = 550nm and then photoluminescence
was recorded. As expected, 10nm PTCBI film has a photoluminescence peak at =
785nm. In a co-deposition film, two materials are mixed in molecular level, and the
feature size is significantly less than exciton diffusion length. Therefore, each exciton
generated upon illumination would have a chance to reach DA heterojunction interface
before it recombines. Exciton would dissociate efficiently into electron in the acceptor
and hole in the donor, if the condition for exciton dissociation is met, as shown in the
inset of Fig. 2.9.
As seen from Fig. 2.9, there is no measurable photoluminescence signal for
CuPc:PTCBI co-deposition film, where ΔEHOMO (0.9eV) of CuPc/PTCBI is larger than
EB (0.7eV) of PTCBI. This indicates that all the excitons from PTCBI generated upon
excitation were dissociated, which is consistent with the conclusion from EQE fitting.
However, for SubPc:PTCBI co-deposition film, significant PTCBI photoluminescence
was observed, where ΔEHOMO (0.5eV) of SubPc/PTCBI is smaller than EB (0.7eV) of
37
PTCBI.. Therefore, most exciton from PTCBI could not dissociate at SubPc/PTCBI
interface. This is also consistent with our EQE measurement and fitting.
Fig. 2.9 Photoluminescence as a function of wavelength for 10nm
PTCBI film, 20nm 1:1 SubPc:PTCBI co-deposition film, and 20nm 1:1
CuPc:PTCBI co-deposition film. Inset: schematic of exciton dissociation at
donor-acceptor interface.
600 650 700 750 800 850
Ph
oto
lum
ine
sce
nce
Sig
na
l (a
.u.)
Wavelength (nm)
10nm PTCBI film 20nm SubPc:PTCBI
1:1 co-deposition film 20nm CuPc:PTCBI
1:1 co-deposition film
Δ
Δ
38
Excitation Laser
= 550nmLuminescence
?
Fig. 2.10 Schematic of film structure for photoluminescence
measurement
2.4.4 Summary
In this section, we have theoretically estimated that EB of PTCBI is 0.7±0.1eV by
including both electrostatic potential and molecule relaxation energy. Experimental
evidence from both EQE and photoluminescence of donor/PTCBI confirmed that EB
of PTCBI is larger than 0.5eV and smaller than 0.9eV. EB of 0.7eV for PTCBI is
similar to previously reported values measured by UPS/IPES technique, for other
organic small molecules, such as 0.6eV for CuPc, 0.6eV for PTCDA (perylene-
3,4,9,10-tetracarboxylic -3,4,9,10-dianhydride), and 1.0eV for -NPD (N,N’-
diphenyl-N,N’-bis(l-naphthyl)-1,1’ biphenyl-4,4’ diamine).21 Generally, polymer has
smaller exciton binding energy than small molecule due to its longer conjugated
chain.22 At DA interface in an OPV cell, the initial separation between the electron in
39
acceptor and the hole in donor after exciton dissociation is 4-5nm,12 and electrostatic
potential between them is about 0.1eV. Therefore, energy offset at DA interface could
potentially be 0.1eV smaller than EB without hurting exciton dissociation within OPV
cells significantly. Better understanding of exciton binding energy would facilitate
OPV design and thus help improve PCE of OPV cells significantly.
2.5 Exciton diffusion length
Exciton diffusion is a very critical step to generate photocurrent in an OPV cell.
Exponentially-measured exciton diffusion length (LD) is usually very short, 3-10nm to
be typical.26 Therefore, quantum efficiency of exciton diffusion is a major bottleneck
in planar bi-layer OPV cells. An improved understanding of exciton diffusion and the
development of organic materials and film structure with longer LD are anticipated.
Boron Subphthalocyanine Chloride (SubPc) is an interesting replacement for
CuPc because of its favorable energy level alignment.28, 29 SubPc also exhibits strong
absorption in the visible region and is suitable for use in tandem structures due to its
relatively narrow absorption spectrum. Optimized SubPc/C60 cell have efficiencies
that are nearly double that of comparable CuPc/C60 reference cells, mainly due to an
enhancement in VOC from 0.45V to ~1V. In this section, we measured LD of SubPc by
using EQE and photoluminescence (PL) quenching. It is found that LD of SubPc
depends on thin film structure strongly, varying from 1.5nm to 5nm.
40
2.5.1 Measure LD by EQE
EQE is the ratio of the number of charge carriers collected by of the solar cell to
the number of photons of a given energy shining on the cell. EQE reflects the response
of a solar cell to the various wavelengths in the spectrum of light shining on the cell.
EQE at certain wavelength is generally measured by shining a mono-wavelength light
on both target cell and reference diode simultaneously. EQE of target cell can then be
calculated based the ratio of electrical signal between target cell and reference diode
and known EQE of reference diode.
Fig. 2.11 shows measured EQE in open square for two different devices: (a)
ITO/SubPc/PTCBI/BCP/Ag and (b) ITO/SubPc/C60/BCP/Ag. As seen in Fig. 211b,
EQE curve is closely related to the absorption curves for both donor and acceptor
materials. Experimentally-measured EQE curve can be fitted by using transfer matrix
method and exciton diffusion equation,26, 27 where exciton diffusion lengths of both
donor and acceptor are free parameters. Since the absorption spectrums of donor and
acceptor materials are often not overlapped, exciton diffusion lengths of donor and
acceptor are easy to decouple in the fitting.
41
400 500 600 700 8000.00
0.05
0.10
0.15
0.20
0.25
SubPc/PTCBIL
ED = 50A
EQ
E
Wavelength (nm)
400 500 600 700 8000.0
0.1
0.2
0.3
0.4
SubPc/C60
EQ
E
(nm)
measured theory
SubPcC
60
LSubPc
D=45A
LC60
D=120A
Fig. 2.11 Measured EQE curve (open square) and theoretical fitting
(solid line) for (a) ITO/SubPc/PTCBI/BCP/Ag and (b)
ITO/SubPc/C60/BCP/Ag. Independent contributions from SubPc and C60 are
plotted as dashed lines in (b).
42
Exciton diffusion length is measured to be 45A~50A by fitting both EQE of
SubPc/PTCBI and SubPc/C60 solar cells. The value is typical and comparable to
many small molecule OPV materials, such as 30A for PTCBI and 10nm for CuPc.
2.5.2 Measure LD by PL quenching
We have measured the exciton diffusion length, LD, of SubPc by using acceptor
material C60 to provide exticon quenching at the SubPc/C60 interface, and three
schematic film structures are shown in the left of Fig. 2.12. Glass/organic interface is
assumed to be non-quenching interface. The energy level diagram is shown in the
inset of Fig. 2.12. The two different structures were used to measure LD of SubPc, (a)
ITO/C60/SubPc and (b) ITO/SubPc/C60. The thickness of the quenching material is
thin enough so that the optical interference effect can be ignored. The exciton
concentration profile upon excitation within SubPc is illustrated within the film in the
left of Fig. 2.12.
Photoluminescence (PL) is measured upon excitation for each film stack as a
function of SubPc thickness. Excitation wavelength is chosen to minimize the
absorption of C60 and maximize the absorption of SubPc. PL1 is the PL signal of
SubPc/Substrate without quenching surface, and PL2 is the PL signal of either
SubPc/C60/Substarte or C60/SubPc/Substrate with quenching surface. No
43
photoluminescence was detected from C60 films under the experimental conditions.
The ratio PL2/PL1 is plotted in the right of figure 2.12.
0 20 40 60 80 1000.0
0.2
0.4
0.6
0.8
1.0
glass/SubPc/C60
ITO/SubPc/C60
glass/C60/SubPc
SubPc C60
PL
Ra
tio
SubPc Thickness (A)
Experiment Data Model Fit, L
ED = 15A
Model Fit, LED
= 45A
Fig. 2.12 The PL intensity ratio PL1/PL2 of SubPc as function of
thickness of SubPc film (symbol). Fitted curves by using eq. (2.5) yields
LD=15A for solid line and LD=45A for dot-dashed line. Inset: schematic
energy level diagram of SubPc/C60 heterojunction.
Assume SubPc/C60 is the only sufficient quenching surface and the film is thin
enough so that uniform exciton generation throughout the thin film is applied. By
solving exciton diffusion equation, one can obtain
44
)]/2exp(1[
)]/2exp(1[11
1
2
D
DDED
Ldd
LdL
PL
PL
(2.5)
where, ED is the fraction of excitons reaching the quenching interface. d is the film
thickness, and LD is the exciton diffusion length of the film. Assuming negligible self
absorption, the PL signal is proportional to the average exciton density. Hence, by
measuring the PL signal both in the presence (PL2) and absence (PL1) of a quenching
interface as a function of the film thickness, one could obtain LD by fitting PL2/PL1 to
eq. (2.5).
The fitted curves in the right of Fig. 2.12 yields LD = 45A for film structure of
ITO/SubPc/C60, which is the same deposition sequence as used in solar cell.
Therefore, the LD measurement result by PL is consistent with the result from EQE
fitting. However, the fitted LD is 15A for structure of ITO/C60/SubPc, which is three
times smaller.
In comparison with growing on ITO glass, SubPc tends to form different
morphology and crystalline structure when growing on glass with organic layers (e.g.,
C60 or PTCBI), which causes the discrepancy in PL measurement for two structures.
Dependence of exciton diffusion length on film crystalline structure and molecule
stacking has been reported as well.30 X-ray diffraction studies suggest that structural
disorder plays an important role30 In general, it is believed that exciton transport is
affected by molecular packing, disorder, presence of impurities and film morphologies.
Also, optical interference effect needs to be considered when measuring exciton
diffusion length experimentally.31, 32
45
By using measured exciton diffusion length of SubPc, the contour of short current
density of SubPc/C60 cells was simulated as function of SubPc thickness and C60
thickness, as shown in Fig. 2.13. Also the short current density of the best SubPc/C60
OPV cell is plotted in the figure also.33 We can see that experimental result matches
simulation very well, which further confirm the exciton diffusion length is 45A-50A
for SubPc.
Fig. 2.13 Contour of short current density of SubPc/C60 cells as function
of SubPc thickness and C60 thickness.
46
2.6 Exciton dissociation
D/A interface is essential location for exciton to dissociate and contribute to
photocurrent because of strong exciton binding energy in OPV cells. As I have
discussed in the section of exciton binding energy, the energy offset (ΔEHOMO or
ΔELUMO) of D/A interface needs to be larger than exciton binding energy in order for
exciton to dissociate efficiently. Therefore, larger energy offset is preferred for exciton
dissociation. However, energy offset will increase the voltage loss, and thus decrease
the efficiency of the cell. An optimal energy offset should be similar to the exciton
binding energy of the OPV materials, which will guarantee efficiency exciton
dissociation and minimize the voltage loss at the same time.
2.7 Charge separation
Immediately after exciton dissociation at D/A interface in OPV cells, the initial
separation between the electron in acceptor and the hole in donor for a GEHP is 4-
5nm,13 and Coulomb electrostatic potential between them is about 0.1eV. The primary
driving force to further separate the GEHP is the electrical field at the interface.
According to transitional metal-insulator-metal (MIM) model of OPV cells, doping
concentration of organic layer is negligible, and therefore, a uniform electrical field
throughout the device is expected. An electrical field of less than 105V/cm is obtained
47
for a typical OPV cell.14 However, previous analysis indicates that such an electrical
field is not efficiency for GEHP separation, and majority of the GEHPs will
recombine at the interface.13
In fact, GEHP dissociation is very efficient for most efficient bi-layer OPV cells.
In these cases, both the donor and acceptor materials are usually unintentionally doped,
which will lead to a strong electrical field at D/A interface and assist GEHP
dissociation. Capacitance-voltage measurements were employed at a frequency of 10
kHz to determine the electrical doping concentrations in organic active layers. In the
inset of Fig. 2.14a, the C-V characteristics are shown for multiple devices with device
structure of ITO/PEDOT/SubPc/C60/BCP/Ag. The strong voltage dependence of the
capacitance indicates the presence of a depletion layer. The spatial profile of the
doping concentration was extracted as a function of depletion layer thickness by using
AoS
BdVCdAq
N/)/1(
222
(2.6)
where s is the dielectric constant, o is the permittivity of free space, A is the
area of the capacitor, and VA is applied voltage. Effective doping concentration NB, is
related to the p-type doping concentration in the donor material NA and the n-type
doping concentration in the acceptor material ND via Eq. (2.7). Therefore, NB is
determined by the layer with the lower doping concentration. Depletion layer
thickness, W, can be calculated from measured capacitance by using Eq. (2.8).
)( DA
DAB NN
NNN
(2.7)
48
CAW os / (2.8)
Fig. 2.14 shows effective doping concentration of SubPc/C60 as a function of
depletion layer thickness for the structure of ITO/PEDOT/SubPc/C60/BCP/Ag. The
device has a minimum depletion layer thickness of 37 nm, and effective doping
concentration is evaluated as 3.5*1017cm-3. The schematics of energy band diagram
are shown in Fig. 2.14b and Fig. 2.14c for partially-depleted scenario at low bias and
totally-depleted scenario at high bias, respectively. Measured doping concentration
can be used to explain the different current-voltage behavior for the same device
structure.14 It turns out that significant doping in organic materials is essential for the
efficient operation of OPV cells. Table 2.1 summarizes measured effective doping
concentration for different donor/acceptor organic material systems. As rule of thumb,
doping concentration is 1017cm-3-1018cm-3 for many efficient OPV materials.
49
33 36 39 42 45 48
1017
1018
1019
1020
1021
-4 -3 -2 -1 0 1
Total Depoletion
47nm
Applied Voltage (V)
Cap
acita
nce
Effective Doping Concentration: 3.5*1017cm-3
ITO/PEDOT/SubPc(11nm)/C60(35nm)/BCP(13nm)/Ag
Eff
ectiv
e D
opin
g C
once
ntra
tion
[cm
-3]
Depletion Layer Thickness [nm]
Fig. 2.14 (a) Effective doping concentration as a function of depletion
layer thickness. Inset: measured capacitance with applied voltage. (b)
Schematic energy band diagram when depletion happens within organic
layers at low bias. (c) Schematic energy band diagram after organic layers are
totally depleted at high bias.
-10 -5 0 5 10 15 20 25 30 35-8.5
-8
-7.5
-7
-6.5
-6
-5.5
-5
-4.5
-4
-3.5
position from junction(nm)
rela
tive
ener
gy le
vel(e
V)
LUMO
HOMO
EF
-10 -5 0 5 10 15 20 25 30 35-6.5
-6
-5.5
-5
-4.5
-4
-3.5
position from junction(nm)
rela
tive
ener
gy le
vel(e
V)
LUMO
HOMO
EF
50
Table 2.1 Measured effective doping concentration
Donor/Acceptor Effective doping concentration (cm-3)
CuPc/PTCBI 8*1017
CuPc/C60 2*1017
SubPc/PTCBI 8*1017
SubPc/C60 3.5*1017
Many of the well-studied and efficient small molecule OPV cells have
unintentional doping occurred during either material synthesis or device fabrication.
Doping mechanism is not clear for most materials. However, for certain materials, one
could indeed add or remove doping intentionally and as a result the electrical
performance can be modified.14, 34
2.8 Charge collection
After exciton dissociation and GEHP separation, charges travel through either
donor or acceptor layer, and then get collected to contribute photocurrent by
corresponding electrodes. In a small molecule planar cell, efficiency of charge
collection is very high, ~100%. However, charges could be trapped or recombined
during collection for bulk heterojunction OPV cell, and efficiency depends on the
51
morphologies of D/A interfaces. The morphologies should be optimized by overall
consideration of exciton dissociation, charge separation and charge collection.
2.9 Voltage loss
In previous sections, I have discussed the process and main aspect of photocurrent
generation. For OPV cell, another important issue is the open-circuit voltage of the
cell. The VOC of most small molecular weight OPV cells is <0.6V, so more than 1eV is
lost when compared with material optical gap which usually exceeds >1.6eV.
Improving VOC is a major issue in high efficiency OPV cells. Active research has been
conducting to understand the physical origin of open-circuit voltage in organic
heterojunction solar cells. Rather than due to the work function differences between
the metal electrodes, Voc is believed to originate from the difference between the
highest occupied molecular orbital (HOMO) of the donor material and lowest
unoccupied molecular orbital (LUMO) of the acceptor.
Recent studies in small molecule multilayer heterojunction devices and in polymeric
bulk heterojunction have confirmed that the magnitude of the open-circuit voltage
correlates with the energy difference between the HOMO of the donor and LUMO of
the acceptor, provided that good electrical contacts are formed between the electrodes
and the organic layers.35
Fig. 2.16 summaries band gap and Voc for the most popular inorganic solar cells and
organic solar cells. Voltage loss, which is defined as the difference between the optical
52
band gap of active material and open-circuit voltage (Voc), ranges from 0.4-0.8V for
typical inorganic solar cells. It is noticed that high quality inorganic solar cells give
smaller voltage loss, while disordered inorganic solar cells cause larger voltage loss,
i.e., 0.8eV for amorphous silicon solar cells. However, voltage loss in organic
heterojunction solar cells, which consists of energy band offset between donor and
acceptor ( LUMOE and HOMOE in Fig. 2.15) and energy differences between quasi
Fermi levels and energy band edges( DFDHOMO EE ,, and AFALUMO EE ,, in Fig. 2.15),
is generally larger than 0.8V.
ELUMO
EHOMO
Voc ΔEDA
ELUMO, A
EF, A
EHOMO, D
EF, D
ELUMO
EHOMO
Voc ΔEDA
ELUMO
EHOMO
Voc ΔEDA
ELUMO, A
EF, A
EHOMO, D
EF, D
Fig. 2.15 Schematic energy band diagram of donor/acceptor
heterojunction
53
As seen from Fig. 2.16, energy band offset, which is necessary for efficient exciton
dissociation, contributes the major part of the voltage loss. This is a big disadvantage
for OPV cells in comparison with inorganic solar cells. However, energy band offsets
in most current DA system are about or even larger than 1eV, which is more than
necessary according to our discussion in previous section. Voc and therefore power
conversion efficiency could gain significant enhancement if energy band offsets could
be decreased further down, e.g., 0.5eV, without harming exciton dissociation for most
systems.
Fig. 2.16 Voltage loss for different solar cell material systems
54
Several approaches have been proposed to explain the origin of Voc in organic solar
cells.35, 36 Nevertheless, a fast growing body of evidence is showing that VOC does not
depend on the work functions of the anode and the cathode, but stems from ΔEDA, the
difference between EHOMO,D and ELUMO,A.36-38 The establishment of an analytical
relationship between VOC and ΔEDA is complicated by several factors. The dark current,
photoconductivity, temperature, short circuits, carrier recombination, resistance related
to thickness and geometry all affect the experimental VOC values in a convoluted way.
Generally, linear relationship between VOC and ΔEDA holds. The higher ΔEDA is, the
higher the VOC is.
0.4 0.6 0.8 1.0 1.2 1.40.0
0.2
0.4
0.6
0.8
1.0
1.2
Op
en
Cir
cuit
Vo
ltag
e,
Vo
c (V
)
Interface Gap, EDA
(eV)
CuPc/PTCBI CuPc/TC-PTCDI CuPc/PTCDA CuPc/C60 SubPc/PTCBI SubPc/TC-PTCDI SubPc/PTCDA SubPc/C60 Pentacene/C60 NPD/C60 ClAlPc/C60
Fig. 2.17 Open circuit voltage as function of interface gap for various
small molecule OPV cells. Dashed line indicates Voc = ΔEDA/e - 0.35.
55
Fig. 2.17 shows the dependence of VOC on ΔEDA at RT under ~1sun illumination for
various DA heterojunction cells. The linear relationship can be explained by classic
semiconductor physics. As discussed earlier, heavy doping is present in most efficient
small molecule solar cells. Therefore, Fermi level can be calculated by classic
semiconductor physics formula for both donor and acceptor, and the built-in voltage,
Vbi, the difference between the Fermi Levels, will reflect VOC of the heterojunction.
))ln(())ln(()()( ,,CV
DAALUMOFAFDDHOMODA NnkTN
pkTEEEEEEqVoc
(2.9)
where, EF,D and EF,A are the Fermi levels of donor and acceptor, and NV and NC are the
carrier density of state of EHOMO,D and ELUMO,A, respectively. The size of small
molecules is about 1 nm cube, and each molecule can sustain one free carrier.
Therefore, NV = NC = 1021cm-3 is reasonable assumption. p and n are the free carrier
concentration of donor and acceptor, respectively, which is found to be around
1018cm-3 for many efficient small molecule bi-layer solar cells.13, 14
Using these values, the dependence of VOC on ΔEDA can be estimated by an empirical
relationship. As shown in the Fig. 2.17, the dash line stands for 35.0 DAEqVoc
with p = n = 1018cm-3. Most systems fall around the line, which indicates good match
between estimation and experimental value.
Having understood the origin of VOC, materials with high EG and DA pair with small
ΔELUMO or ΔEHOMO are desired for larger VOC. Although high ΔEDA=EG,D-ΔELUMO
contribute to high VOC, ΔELUMO need to be large enough to dissociate excitons as we
discussed in the earlier section. High EG is beneficial for high VOC while photons with
56
low energy could not be absorbed by high EG materials. Cells with high EG and/or
small ΔELUMO could give high VOC, but JSC could be low39 and thus EG and energy
offset ΔE are carefully optimized for high efficiency.
2.10 Efficiency estimation
As we explained in the previous section, EG is closely related to JSC and VOC, and thus
PCE in an OPV cell. Low EG materials can lead to high JSC due to absorption in broad
spectrum but VOC could be low. On the other hand, high EG materials could lead to low
JSC and high VOC. In this section, I will estimate the PCE as a function of optical band
gap, EG. This model assumes the same band gap EG for both donor and acceptor, and
perfect absorption above EG. The photocurrent under short-circuit conditions can be
expressed,
dSqJ EQESC )( (2.10)
where S(λ) is the AM1.5 solar spectrum expressed in number of photons per unit area
per unit time per unit wavelength. %85 IQEEQE is assumed based on perfect
absorption, and IQE is a cumulative parameter to reflect overall intrinsic process to
photocurrent generation after absorption.
Then VOC is defined by the empirical relationship Eq. (2.11) we concluded from
experimental data based on various organic donor/acceptor pairs.
35.0 DAEqVoc (2.11)
where DAE is the interface gap, and is evaluated by using Eq. (2.12).
57
LUMOGDA EEE (2.12)
Based on our estimation of exciton binding energy and reported value for OPV
materials, eVEE HOMOLUMO 5.0 is further assumed for efficient exciton
dissociation at D/A interface. The relationship is illustrated in Fig. 2.18.
PCE of a single junction OPV cell can then be obtained using
in
OCSCP P
FFVJ (2.13)
where Pin is incident power of light, which is 100mW/cm2 for AM1.5G, and
65.0FF is assumed.8
ELUMO = 0.5eV
EHOMO = 0.5eV
EDA
Voc =
EDA - 0.35eV
ELUMO = 0.5eV
EHOMO = 0.5eV
EDA
Voc =
EDA - 0.35eV
ELUMO = 0.5eV
EHOMO = 0.5eV
EDA
Voc =
EDA - 0.35eV
Fig. 2.18 Assumption of efficiency simulation
58
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.50.0
2.0
4.0
6.0
8.0
10.0
12.0
Eff
icie
ncy
(%
)
Band Gap, Eg (eV)
11.6% @Eg=1.65eV
XS1
XS2
XS3XS4
XS5
X P1
X P2
X P3
X P4
X
P6
P5
S6
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.50.0
2.0
4.0
6.0
8.0
10.0
12.0
Eff
icie
ncy
(%
)
Band Gap, Eg (eV)
11.6% @Eg=1.65eV
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.50.0
2.0
4.0
6.0
8.0
10.0
12.0
Eff
icie
ncy
(%
)
Band Gap, Eg (eV)
11.6% @Eg=1.65eV
XS1
XS2
XS3XS4
XS5
X P1
X P2
X P3
X P4
X
P6
P5
S6
Fig. 2.19 Simulated efficiency as a function of optical band gap. Best
efficiency of 11.6% is predicted at band gap equals 1.65 eV. Significant
development work is shown on the plot for both small molecule and polymer
OPV cells.
59
Table 2.2 Significant development work of OPV cells
Label Material and Cell Structure Year
S1 CuPc/PTCBI planer cell 1986
S2 CuPc/PTCBI planer w/ light trapping cell 2000
S3 CuPc/PTCBI bulk heterojunction cell 2004
S4 CuPc/C60 planer cell 2001
S5 CuPc/C60 bulk heterojunction cell 2007
S6 Small molecule tandem cell by Heliatek 2010
P1 PPV/PCBM polymer cell 1995
P2 MDMO-PPV/PCBM polymer cell 2001
P3 P3HT/PCBM polymer cell 2005
P4 PCDTBT/PCBM polymer cell 2009
P5 PBDTTT-CF/PCBM polymer cell 2009
P6 polymer cell by Solarmer 2009
As shown in Fig. 2.19, the optimized single junction OPV cell has a peak at ηP=11.6%
when EG=1.65eV, JSC=22.3mA/cm2 and VOC=0.80V. When EG is high, photocurrent
will decrease and limit the PCE. If EG is low, VOC will limit the PCE. Some significant
development work in both small molecule and polymer OPV cells has been labeled in
Fig. 2.19, and material/cell structure and year are summarized in Table 2.2.
In comparison with theoretical maximum efficiency of inorganic single junction cell,
the requirement of D/A interface energy offset in OPV cells limits the maximum
efficiency significantly. Peak efficiency happens at EG=1.65eV, which means the
60
photon with energy less 1.65eV could not be absorbed. About half of the sunlight
energy is wasted. Multi-junction OPV cell is a promising approach to address the issue.
2.11 Conclusion
The chapter discusses cost analysis of OPV module, operation mechanism of OPV cell,
and some fundamental physics problem underlying the operation. In the end,
achievable PCE of single junction OPV cells is estimated by using the parameters
concluded from the study.
61
Bibliography
1. Hadipour, A.; de Boer, B.; Wildeman, J.; Kooistra, F. B.; Hummelen, J. C.;
Turbiez, M. G. R.; Wienk, M. M.; Janssen, R. A. J.; Blom, P. W. M. Solution-
Processed Organic Tandem Solar Cells. Advanced Functional Materials
2006,16,1897-1903.
2. Forrest, S. R. Ultrathin Organic Films Grown by Organic Molecular Beam
Deposition and Related Techniques. Chem. Rev. 1997,97,1793-1896.
3. Zweibel, K. Terawatt Challenge for Thin-Film PV. 2005, NREL/TP-520-
38350, .
4. Kalowekamo, J.; Baker, E. Estimating the manufacturing cost of purely
organic solar cells. Solar Energy 2009,83,1224-1231.
5. Energy, I. A. Average Retail Price of Electricity to Ultimate Customers by
End-User Sector, by State. 2008).
6. http://www.heliatek.com/ 2010.
7. http://www.solarmer.com/ 2010.
8. Tang, C. W. Two-layer organic photovoltaic cell. Appl. Phys. Lett.
1986,48,183-185.
9. Xue, J.; Rand, B. P.; Uchida, S.; Forrest, S. R. A Hybrid Planar-Mixed
Molecular Heterojunction Photovoltaic Cell. Adv Mater 2005,17,66-71.
62
10. Peumans, P.; Forrest, S. R. Very-high-efficiency double-heterostructure
copper phthalocyanine/C[sub 60] photovoltaic cells. Appl. Phys. Lett. 2001,79,126-
128.
11. Yang, F.; Shtein, M.; Forrest, S. R. Controlled growth of a molecular bulk
heterojunction photovoltaic cell. Nat. Mater. 2005,4,37-41.
12. Noufi, R.; Zweibel, K. High-Efficiency CdTe and CIGS Thin-Film Solar
Cells: Highlights and Challenges. Photovoltaic Energy Conversion, Conference
Record of the 2006 IEEE 4th World Conference on 2006, 1, 317-320.
13. Peumans, P.; Forrest, S. R. Separation of geminate charge-pairs at donor-
acceptor interfaces in disordered solids. Chemical Physics Letters 2004,398,27.
14. A. Liu, S. Zhao, S.-B. Rim, J. Wu, M. Könemann, P. Erk,P.Peumans,
Control of Electric Field Strength and Orientation at the Donor-Acceptor Interface in
Organic Solar Cells. Adv Mater 2008,20,1065-1070.
15. Rim, S.; Zhao, S.; Scully, S. R.; McGehee, M. D.; Peumans, P. An effective
light trapping configuration for thin-film solar cells. Appl. Phys. Lett. 2007,91,243501-
3.
16. Agrawal, M.; Peumans, P. Broadband optical absorption enhancement
through coherent light trapping in thin-film photovoltaic cells. Opt. Express
2008,16,5385-5396.
63
17. Yang, F.; Sun, K.; Forrest, S. R. Efficient Solar Cells Using All-Organic
Nanocrystalline Networks. Adv Mater 2007,19,4166-4171.
18. Rand, B. P.; Xue, J.; Yang, F.; Forrest, S. R. Organic solar cells with
sensitivity extending into the near infrared. Appl. Phys. Lett. 2005,87,233508.
19. Bailey-Salzman, R. F.; Rand, B. P.; Forrest, S. R. Near-infrared sensitive
small molecule organic photovoltaic cells based on chloroaluminum phthalocyanine.
Appl. Phys. Lett. 2007,91,013508-3.
20. Dai, J.; Jiang, X.; Wang, H.; Yan, D. Organic photovoltaic cells with near
infrared absorption spectrum. Appl. Phys. Lett. 2007,91,253503.
21. Hill, I. G.; Kahn, A.; Soos, Z. G.; Pascal, J.,R.A. Charge-separation energy in
films of π-conjugated organic molecules. Chemical Physics Letters 2000,327,181-188.
22. Knupfer, M. Exciton binding energies in organic semiconductors. Appl. Phys.
A 2003,77,623-626.
23. Peumans, P.; Uchida, S.; Forrest, S. R. Efficient bulk heterojunction
photovoltaic cells using small-molecular-weight organic thin films. Nature
2003,425,158-162.
24. Antoine Kahn, Norbert Koch,Weiying Gao, Electronic structure and
electrical properties of interfaces between metals and pi-conjugated molecular films.
Journal of Polymer Science Part B: Polymer Physics 2003,41,2529-2548.
64
25. Mutolo, K. L.; Mayo, E. I.; Rand, B. P.; Forrest, S. R.; Thompson, M. E.
Enhanced Open-Circuit Voltage in Subphthalocyanine/C60 Organic Photovoltaic
Cells. J. Am. Chem. Soc. 2006,128,8108-8109.
26. Peumans, P.; Yakimov, A.; Forrest, S. R. Small molecular weight organic
thin-film photodetectors and solar cells. J. Appl. Phys. 2003,93,3693-3723.
27. Pettersson, L. A. A.; Roman, L. S.; Inganas, O. Modeling photocurrent
action spectra of photovoltaic devices based on organic thin films. J. Appl. Phys.
1999,86,487-496.
28. Mutolo, K. L.; Mayo, E. I.; Rand, B. P.; Forrest, S. R.; Thompson, M. E.
Enhanced Open-Circuit Voltage in Subphthalocyanine/C60 Organic Photovoltaic
Cells. J. Am. Chem. Soc. 2006,128,8108-8109.
29. Gommans, H. H. P.; Cheyns, D.; Aernouts, T.; Girotto, C.; Poortmans, J.;
Heremans, P. Electro-Optical Study of Subphthalocyanine in a Bilayer Organic Solar
Cell. 2007,17,2653-2658.
30. Rim, S.; Fink, R. F.; Schoneboom, J. C.; Erk, P.; Peumans, P. Effect of
molecular packing on the exciton diffusion length in organic solar cells. Appl. Phys.
Lett. 2007,91,173504-3.
31. Scully, S. R.; McGehee, M. D. Effects of optical interference and energy
transfer on exciton diffusion length measurements in organic semiconductors. J. Appl.
Phys. 2006,100,034907-5.
65
32. Rim, S.; Peumans, P. The effects of optical interference on exciton diffusion
length measurements using photocurrent spectroscopy. J. Appl. Phys.
2008,103,124515-5.
33. Gommans, H. H. P.; Cheyns, D.; Aernouts, T.; Girotto, C.; Poortmans, J.;
Heremans, P. Electro-Optical Study of Subphthalocyanine in a Bilayer Organic Solar
Cell. Advanced Functional Materials 2007,17,2653-2658.
34. Werner, A. G.; Li, F.; Harada, K.; Pfeiffer, M.; Fritz, T.; Leo, K. Pyronin B
as a donor for n-type doping of organic thin films. Appl. Phys. Lett. 2003,82,4495-
4497.
35. Potscavage, J.,W.J.; Yoo, S.; Kippelen, B. Origin of the open-circuit voltage
in multilayer heterojunction organic solar cells. Appl. Phys. Lett. 2008,93,193308.
36. Rand, B. P.; Burk, D. P.; Forrest, S. R. Offset energies at organic
semiconductor heterojunctions and their influence on the open-circuit voltage of thin-
film solar cells. Phys. Rev. B 2007,75,115327.
37. Cheyns, D.; Poortmans, J.; Heremans, P.; Deibel, C.; Verlaak, S.; Rand, B. P.;
Genoe, J. Analytical model for the open-circuit voltage and its associated resistance in
organic planar heterojunction solar cells. Phys. Rev. B 2008,77,165332.
38. Cravino, A. Origin of the open circuit voltage of donor-acceptor solar cells:
Do polaronic energy levels play a role? Appl. Phys. Lett. 2007,91,243502-3.
66
39. Uhrich, C.; Schueppel, R.; Petrich, A.; Pfeiffer, M.; Leo, K.; Brier, E.; Kilickiran,
P.; Baeuerle, P. Organic Thin-Film Photovoltaic Cells Based on Oligothiophenes with
Reduced Bandgap. Advanced Functional Materials 2007,17,2991-2999.
67
Chapter 3 Organic Solar Cells with Solution-Processed Graphene Transparent Electrodes
We demonstrate that solution-processed graphene thin films can serve as transparent
conductive anodes for organic photovoltaic cells. The graphene electrodes were
deposited on quartz substrates by spin-coating of an aqueous dispersion of
functionalized graphene, followed by a reduction process to reduce the sheet
resistance. Small molecular weight organic solar cells can be directly deposited on
such graphene anodes. The short-circuit current and fill factor of these devices on
graphene are lower than that of control device on indium tin oxide due to the higher
sheet resistance of the graphene films. We anticipate that further optimization of the
reduction conditions will improve the performance of these graphene anodes.
3.1 Introduction
Organic optoelectronic devices such as organic photovoltaic (OPV) cells,1, 2 organic
light-emitting diodes (OLEDs),3, 4 and organic photodetectors5, 6 have attracted interest
because they can be deposited on flexible, light-weight substrates using low-cost
68
fabrication methods. An important aspect of optoelectronic thin-film devices is the
transparent, conductive electrode through which light couples in or out of the devices.
Indium tin oxide (ITO) is widely used but may be too expensive7 for an application
such as solar cells. Moreover, metal oxides such as ITO are brittle and therefore of
limited use on flexible substrates.8 A substitute for ITO with a similar performance but
lower cost is clearly needed. Recent work on solution-processed random meshes of
carbon nanotubes9-11 and metal nanowires12, 13 has resulted in solution-processed,
conductive, transparent electrodes with transparencies and sheet resistances that
approach or exceed that of ITO. However, the electrodes realized using the above
approaches have a roughness that is comparable or larger than a typical device
thickness, leading to frequent shorts.
3.2 Unique properties of gaphene
As shown in Fig. 3.1, graphene is a monolayer of carbon atoms densely packed into a
benzene-ring structure, which is the basic building block of graphite crystal, carbon
nano-tubes, fullerenes, and so on. For example, carbon nano-tubes are usually thought
of as graphene sheets rolled up into nanometer-sized cylinders. The two-dimensional
conductor graphene is a promising transparent conductor because of its optical and
electrical properties.14-16 However, Planar graphene itself has been presumed not to
exist in the free states in the past, being unstable with respect to the formation of
curved structures such as fullerenes and nano-tubes. In principle, electrons in
69
individual graphene sheets delocalize over the complete sheet, providing ballistic
charge transport in an one-atom-thick material with very little optical absorption. In
practice, however, large-area graphene films produced via solution processing of
functionalized graphene contain multiple grain boundaries and incorporate lattice
defects and oxidative traps that increase the electrical resistance of the material. As a
result, the films must be made thicker than one atomic layer to obtain practical sheet
resistances. Recently, such graphene thin films were used as an anode in dye-
sensitized solar cells.17 In this work, we demonstrate that very thin films of graphene
can be used as anode in solid-state thin-film OPV cells with a performance that
approaches that of ITO while not exhibiting the roughness problems of carbon nano-
tube or metal nanowire meshes. The trade-off between transparency and sheet
resistance is also investigated.
Fig. 3.1 Graphite crystal and monolayer graphene
70
3.3 Fabrication of large area graphene thin film
3.3.1 Fabrication method survey
Given the unique properties of graphene film, researchers have been trying to develop
various methods to make thin-films of graphene, such as the scotch tape-method14,
transfer-printing18, epitaxial growth19, 20, and so on.
3.3.1.1 Scotch tape method
The scotch tape method is first developed by Andre Geim and Konstantin
Novoselov,14 who won the Nobel Prize in Physics 2010 because of their
groundbreaking experiments regarding the graphene. Basically, one could start with a
piece of high quality graphite flake, and then repeatedly peel the flake off by using
scotch tape. As a result, few-layer graphene or single-layer graphene is created, which
could be used to study its electronic and optical properties. Fig. 3.2 shows graphene
film prepared by scotch tape method and corresponding device by using the film.
71
Fig. 3.2 Graphene film and device made by scotch tape method
3.3.1.2 Transfer-printing
The method18, so called graphene-on-demand by cut-and-choose transfer-printing
(DCT), can be used to transfer a piece of high-quality graphene in a device active-area
on substrate, which is often a size of tenths of micrometers. As shown in Figure 3.3,
DCT uses pillars (protrusions) of a stamp to cut and exfoliate a high-quality graphene
piece (one piece per pillar) from graphite, and then inspects them. If an inspected
graphene is good, it will be printed on the device active-area of a wafer with high
placement precision. By using the method, working field-effect-transistors have been
achieved with good transport properties. Such an approach has the potential to build
graphene integrated circuits on large wafers.
72
Fig. 3.3 Schematic of graphene-on-demand by cut-and-choose transfer-
printing (DCT)18
It is also known that Graphene can grow epitaxially on single crystal silicon
carbide substrates both in theory and in experiment 19, 20. However, all these methods
are not suitable for making low-cost, large-area graphene thin film. Solution process
was used to make transparent conductive graphene film in this work.
3.3.2 Solution-processed graphene film
73
The graphene electrodes used in this work were deposited on quartz slides by
spin-coating dispersions of functionalized graphene in water, as shown in Fig. 3.4. The
functionalized graphene material was prepared through a modified Hummers method21,
22 as described elsewhere.23 Briefly, a graphite crystal was chemically oxidized by
treatment with various solutions of NaNO3, KMnO4, concentrated H2SO4 and 30 wt%
H2O2, washed with HCl and purified water, and ultrasonicated to exfoliate individual
graphene oxide sheets. This procedure produces a loose brown powder which can be
dispersed in water at loadings of up to ~15mg/mL. The spin-coating rate was
increased every 30s from an initial value of 500rpm to 800rpm to gradually spread the
water dispersion on the quartz, and finally to 1600rpm dry the film. Residual water
was removed by heating these films to 100C in a vacuum oven for several hours.
The resulting functionalized graphene films are not conductive and the material
must be reduced to obtain acceptable sheet resistances. The reduction process removes
oxidized functionalities from the film and partially restores electron delocalization,
increasing both the light absorption and electrical conductivity of the film. The
effectiveness of the reduction of functionalized graphene films by chemical and
thermal methods has been characterized17, 23 with chemical treatments being less
effective than thermal procedures. The dependence of optical and electrical properties
of reduced graphene film on the reduction process will be discussed elsewhere.23 In
this work, the graphene films were reduced either by vacuum annealing at 1100oC or
by a combination of a hydrazine treatment and Ar annealing at 400oC.
74
Fig. 3.4 Schematic of preparing large area graphene thin film by solution
process
3.4 Properties of Solution-processed graphene film
3.4.1 Graphene film morphology
Figure 3.5a shows a typical atomic force microscopy (AFM) image of a
graphene film after reduction using a hydrazine treatment and Ar annealing with a
thickness <10nm. The surface of the film presents some corrugations that possibly
75
arise from partial roll-up of the edges of individual graphene sheets. The root mean
surface roughness is Rq~3nm. Figure 3.5b shows the morphology of a thicker
graphene film (>10nm) exhibiting extended wrinkles leading to Rq~9 nm. In both
cases, the film surfaces are free from spikes which would lead to shorts in thin-film
optoelectronic devices. The thick, spin-coated conductive polymer buffer layer that is
used in devices with carbon nanotube or metal nanowire mesh transparent electrodes
is therefore not necessary to prevent shorts when using these graphene electrodes.
Since the film thicknesses exceed the thickness of a single layer of graphene, the films
used here consist mostly of multiple layers of graphene.
Fig. 3.5 AFM images for a reduced thin (<10nm) (a) and thick (>10nm)
(b) graphene film. The scan size is 3m in (a) and 8m in (b).
(a) (b)
76
3.4.2 Electrical and optical properties of reduced graphene film
Figure 3.6 shows the optical transmittance (triangles, measured at =550nm) and
sheet resistance (circles) vs. thickness for graphene films reduced by both vacuum
annealing (open symbols) and a combination of a hydrazine treatment and Ar
annealing (closed symbols). The film thickness was measured using atomic force
microscopy across edges and steps in the films. The sheet resistance was measured in
a 2-electrode configuration and includes the contact resistance between the graphene
film and vapor-deposited gold electrodes used to probe the films. Both transmittance
and sheet resistance decrease with increasing film thickness. Reduction by vacuum
annealing produces graphene films with slightly better transparency and conductivity
compared to the films reduced using the hydrazine treatment and Ar annealing. For the
film thicknesses <15nm, the optical transmittance is >80%, while the sheet resistance
varies from 5k/sq to 1M/sq.
77
0 5 10 15 20
50
60
70
80
90
100
103
104
105
106
107
Graphene filmQuartz substrate
1100oC vacuum annealingHydrazine + argon annealing
She
et R
esis
tanc
e (o
hm/s
q.)
Tra
nsm
ittan
ce (
%)
Graphene Film Thickness (nm)
Transmittance @ =550nm Sheet Resistance
Fig. 3.6 Transmittance at =550nm (triangles) and sheet resistance
(circles) as a function of the graphene film thickness for both reduction
methods: vacuum annealing at 1100oC (open symbols), and hydrazine
treatment and argon annealing at 400oC (filled symbols).
78
102 103 104 105 106 107
20
40
60
80
100
Tra
nsm
ittan
ce (
%)
Sheet Resistance (ohm/sq.)
Fig. 3.7 Relationship between transmission and sheet resistance of
reduced graphene films
High resolution XPS analysis indicates that the majority of oxygen content is
decreased during the reduction of functionalized graphene film. Reduction process
increases the film absorption while recovering the conductivity of the film. Therefore,
optimal combination of transmission and electrical conductivity needs to be explored
for transparent conductor applications. As shown in Fig. 3.7, high transmission film
has the issue of high sheet resistance, while low resistance film suffers the problem of
79
low transmission. To serve as a transparent electrode, the film is preferred to have
more than 80% of transmission.
The dependence of transmittance on wavelength is shown in Fig. 3.8 for three
graphene films with different thicknesses, reduced by the hydrazine treatment and
argon annealing. The optical transparency decreases monotonically for decreasing
wavelengths over the spectral range =400-1800nm.
400 800 1200 160060
70
80
90
100
Tra
nsm
ittan
ce (
%)
Wavelength (nm)
Film thickness 3.1nm 4.2nm 7.0nm
Fig. 3.8 Transmittance as a function of wavelength for graphene films of
various thicknesses. The graphene films were reduced by the hydrazine
treatment and argon annealing at 400oC.
80
The sheet resistance of the graphene films shown here is much higher than that of
ITO films for the same transparency. This may be because the functionalized graphene
is not completely reduced. Furthermore, since he average size of the sheets was
(364120)nm measured along the longest axis, charge carriers have to hop between
individual flakes. Finally, the films are not uniform and thinner regions of the film
may act as bottlenecks that limit the overall in-plane conductivity. Further work is
required to optimize the process conditions for improved optical and electrical
properties.
3.5 OPV cells on graphene transparent electrode
3.5.1 Device fabrication
Bilayer small molecule organic photovoltaic cells were fabricated on graphene
films on quartz and on commercially obtained 130nm-thick ITO on glass (sheet
resistance <20/sq) as transparent anode. All organic materials were obtained
commercially and then purified using thermal gradient sublimation.24
The organic thin films and metal cathode were deposited at room temperature by
thermal evaporation in high vacuum (~10-7Torr). The layer structure of the devices is
anode/copper phthalocyanine (CuPc)/fullerene (C60)/bathocuproine (BCP)/100nm Ag,
81
as shown in Fig. 3.9. The Ag cathode was deposited through a shadow mask with
circular openings of 0.81 mm2.
Fig. 3.9 Molecule structure of organic materials and device structure of
organic photovoltaic cell
3.5.2 OPV cell performance
The current density-voltage (J-V) curves were measured at room temperature
under 85mW/cm2 AM1.5G simulated solar illumination from a filtered 500W Xe arc
ITO or Graphene
CuPc (35nm)
C60 (50nm)
BCP (10nm)
Ag
Glass
C60
82
lamp. The illumination intensity was measured using a calibrated silicon photodiode.
No corrections were made for the spectral mismatch.
The thickness of graphene films used to fabricate OPV cells is between 4nm and
7nm, and the corresponding values of the transmittance and sheet resistance are 85%
to 95%, and 100k/sq to 500k/sq, respectively. The J-V curves of a typical cell with
layer structure 35nm CuPc/50nm C60/10nm BCP on graphene and ITO are shown in
Fig. 3.10, where the graphene film was reduced by vacuum annealing. The short
circuit current density, JSC, open circuit voltage, VOC, fill factor, FF, and power
conversion efficiency, , are 2.1mA/cm2, 0.48V, 0.34, and 0.4%, respectively, for the
cell on graphene, and 2.8mA/cm2, 0.47V, 0.54, and 0.84%, respectively, for the cell on
ITO. The efficiency of the ITO control device is well below the best published data25,
26 for the materials system used, because the device structure and deposition
conditions were not optimized. In spite of different work functions of graphene (4.4-
4.65eV)27, 28 and ITO (4.5-5.1eV),29 the difference in VOC is <100mV. This indicates
that VOC is determined by the difference between Fermi levels of the unintentionally
doped donor (CuPc) and acceptor (C60) materials.30
The lower efficiency of the cell on graphene is mainly due to a reduced JSC and
FF compared to the cell on ITO, both of which are due to the high sheet resistance of
the graphene film. Furthermore, a smaller shunt resistance is seen for devices on
graphene electrodes, leading to an increase in current in reverse bias, in the dark and
under illumination. Using experimental data for cells on ITO, J-V curves for the cells
83
on graphene can be reproduced by including an extra voltage drop due to a series
resistor to the voltage axis.
-0.4 -0.2 0.0 0.2 0.4 0.6-4
-2
0
2
4
6
ITO control device Graphene device
Ag electrodeBCPC60CuPcGraphene filmQuartz substrate
dark illuminated
Cur
rent
Den
sity
(m
A/c
m2 )
Voltage (V)
Fig. 3.10 Current density vs. voltage (J-V) curves for an OPV cell with
layer structure 35nm CuPc/50nm C60/10nm BCP/100 nm Ag on graphene
(circles) and ITO (squares) in the dark (filled symbols) and under 85mW/cm2
AM1.5G simulated solar illumination (open symbols). The sheet resistance
and transmittance of the graphene film are ~100k/sq and 95%, respectively.
Inset: Schematic of the cells.
84
-0.4 -0.2 0.0 0.2 0.4 0.6-4
-3
-2
-1
0
1
2
3
-0.4 -0.2 0.0 0.2 0.4 0.6
Series resistor() Shunt resistor()110k 450k180k 450k420k Infinite
measure model
Cur
rent
Den
sity
(m
A/c
m2 )
Voltage (V)
ITO control Graphene 1 Graphene 2 Graphene 3
Fig. 3.11 J-V curves for OPV cells on ITO (squares) and three different
graphene films (circles, triangles, and inverted triangles) under AM1.5G
simulated solar illumination. Using the data for the cell on ITO, the J-V
curves of the cells on graphene are reproduced by adding a resistive voltage
drop to the voltage axis (solid line, dashed line, and short dashed line). The
series resistances used to obtain these fits, R1,fit=110k, R2,fit=180k and
R3,fit=420k, are very close to the measured sheet resistances of the
graphene films used: R1=75k-150k, R2=200k and R3=200k-3M,
respectively.
85
As shown in Fig. 3.11, the modeled J-V curves match the experimental data of cells
on graphene. The series resistances used to obtain these fits, R1,fit=110k, R2,fit=180k
and R3,fit=420k, are very close to the measured sheet resistances of the graphene
films used: R1=75k-150k, R2=200k and R3=200k-3M, respectively.
3.6 Conclusion
In conclusion, we demonstrated that graphene thin films can be used as a
transparent conductive electrode for OPV cells. The graphene thin films were prepared
by solution processing of functionalized graphene followed by a reduction step to
improve the conductivity. For film thicknesses <15nm, the optical transmittance
is >80%, while the sheet resistance varies from 5k/sq to 1M/sq. Bilayer small
molecular weight OPV cells were demonstrated on 4nm to 7nm-thick graphene thin
films. Compared to control devices on ITO, it was shown that the lower JSC and VOC
are mainly caused by the high sheet resistance of the graphene thin films. Improved
graphene film treatments will be required to improve the sheet resistance without
compromising transmittance. In addition, processes need to be developed that are
compatible with low-cost flexible substrates.
86
Bibliography
1. Tang, C. W. Two-layer organic photovoltaic cell. Appl. Phys. Lett. 1986,48,183-
185.
2. Peumans, P.; Yakimov, A.; Forrest, S. R. Small molecular weight organic thin-film
photodetectors and solar cells. J. Appl. Phys. 2003,93,3693-3723.
3. Tang, C. W.; VanSlyke, S. A. Organic electroluminescent diodes. Appl. Phys. Lett.
1987,51,913-915.
4. Baldo, M. A.; O'Brien, D. F.; You, Y.; Shoustikov, A.; Sibley, S.; Thompson, M.
E.; Forrest, S. R. Highly efficient phosphorescent emission from organic
electroluminescent devices. Nature 1998,395,151-154.
5. Yu, G.; Cao, Y.; Wang, J.; McElvain, J.; Heeger, A. J. High sensitivity polymer
photosensors for image sensing applications. Synth. Met. 1999,102,904-907.
6. Peumans, P.; Bulovic, V.; Forrest, S. R. Efficient, high-bandwidth organic
multilayer photodetectors. Appl. Phys. Lett. 2000,76,3855-3857.
7. Forrest, S. R. The path to ubiquitous and low-cost organic electronic appliances on
plastic. Nature 2004,428,911-918.
8. Chen, Z.; Cotterell, B.; Wang, W.; Guenther, E.; Chua, S. A mechanical assessment
of flexible optoelectronic devices. Thin Solid Films 2001,394,201-205.
87
9. Wu, Z. C.; Chen, Z. H.; Du X.; Logan, J. M.; Sippel J.; Nikolou M.; Kamaras K.;
Reynolds, J. R.; Tanner, D. B.; Hebard, A. F.; Rinzler, A. G. Transparent, Conductive
Carbon Nanotube Films. Science 2004,305,1273.
10. Contreras, M. A.; Barnes T.; van de, L. J.; Rumbles G.; Coutts, T. J.; Weeks C.;
Glatkowski P.; Levitsky I.; Peltola J.; Britz, D. A. Replacement of Transparent
Conductive Oxides by Single-Wall Carbon Nanotubes in Cu(In,Ga)Se2-Based Solar
Cells. J. Phys. Chem. C 2007,111,14045.
11. Rowell, M. W.; Topinka, M. A.; McGehee, M. D.; Prall H.; Dennler G.; Sariciftci,
N. S.; Hu L.; Gruner G. Organic Solar Cells with Carbon Nanotube Network
Electrodes. Appl. Phys. Lett. 2006,88,233506/1.
12. M.-G. Kang, L. J. G., Nanoimprinted Semitransparent Metal Electrodes and Their
Application in Organic Light-Emitting Diodes. Adv Mater 2007,19,1391-1396.
13. Lee, J. -.; Connor, S. T.; Cui, Y.; Peumans, P. Solution-Processed Metal
Nanowire Mesh Transparent Electrodes. Nano Lett. 2008,8,689-692.
14. Novoselov, K. S.; Geim, A. K.; Morozov, S. V.; Jiang D.; Zhang Y.; Dubonos, S.
V.; Grigorieva, I. V.; Firsov, A. A. Electric Field Effect in Atomically Thin Carbon
Films. Science 2004,306,666.
15. Watcharotone S.; Dikin, D. A.; Stankovich S.; Piner R.; Jung I.; Dommett, G. H.
B.; Evmenenko G.; Wu S.; Chen S.; Liu C.; Nguyen, S. T.; Ruoff, R. S. Graphene-
Silica Composite Thin Films as Transparent Conductors. Nano Lett. 2007,7,1888.
16. Gomez-Navarro C.; Weitz, R. T.; Bittner, A. M.; Scolari M.; Mews A.; Burghard
M.; Kern K. Electronic Transport Properties of Individual Chemically Reduced
Graphene Oxide Sheets. Nano Lett. 2007,7,3499.
88
17. Wang, X.; Zhi, L.; Mullen, K. Transparent, Conductive Graphene Electrodes for
Dye-Sensitized Solar Cells. Nano Lett. 2008,8,323-327.
18. Liang, X.; Fu, Z.; Chou, S. Y. Graphene Transistors Fabricated via Transfer-
Printing In Device Active-Areas on Large Wafer. Nano Letters 2007,7,3840-3844.
19. Varchon, F.; Feng, R.; Hass, J.; Li, X.; Nguyen, B. N.; Naud, C.; Mallet, P.;
Veuillen, J. -.; Berger, C.; Conrad, E. H.; Magaud, L. Electronic Structure of Epitaxial
Graphene Layers on SiC: Effect of the Substrate. Phys. Rev. Lett. 2007,99,126805.
20. de Heer, W. A.; Berger, C.; Wu, X.; First, P. N.; Conrad, E. H.; Li, X.; Li, T.;
Sprinkle, M.; Hass, J.; Sadowski, M. L.; Potemski, M.; Martinez, G. Epitaxial
graphene. Solid State Commun. 2007,143,92-100.
21. Hummers Wm. S.; Jr Preparation of Graphitic Oxide. J. Am. Chem. Soc.
1958,80,1339.
22. Hirata M.; Gotou T.; Horiuchi S.; Fujiwara M.; Ohba M. Thin-Film Particles of
Graphite Oxide 1: High-Yield Synthesis and Flexibility of the Particles. Carbon
2004,42,2929.
23. Becerril, H. A.; Mao, J.; Liu, Z.; Stoltenberg, R. M.; Bao, Z.; Chen, Y. Evaluation
of Solution-Processed Reduced Graphene Oxide Films as Transparent Conductors.
ACS Nano 2008,2,463-470.
24. Forrest, S. R. Ultrathin Organic Films Grown by Organic Molecular Beam
Deposition and Related Techniques. Chem. Rev. 1997,97,1793-1896.
25. Peumans, P.; Forrest, S. R. Very-high-efficiency double-heterostructure copper
phthalocyanine/C[sub 60] photovoltaic cells. Appl. Phys. Lett. 2001,79,126-128.
89
26. Xue, J.; Uchida, S.; Rand, B. P.; Forrest, S. R. Asymmetric tandem organic
photovoltaic cells with hybrid planar-mixed molecular heterojunctions. Appl. Phys.
Lett. 2004,85,5757-5759.
27. Schroeder, P. G.; Nelson, M. W.; Parkinson, B. A.; Schlaf, R. Investigation of
band bending and charging phenomena in frontier orbital alignment measurements of
para-quaterphenyl thin films grown on highly oriented pyrolytic graphite and SnS2.
Surf. Sci. 2000,459,349-364.
28. Ago, H.; Kugler, T.; Cacialli, F.; Petritsch, K.; Friend, R. H.; Salaneck, W. R.;
Ono, Y.; Yamabe, T.; Tanaka, K. Workfunction of purified and oxidised carbon
nanotubes. Synth. Met. 1999,103,2494-2495.
29. Nuesch, F.; Rothberg, L. J.; Forsythe, E. W.; Toan Le, Q.; Gao, Y. A
photoelectron spectroscopy study on the indium tin oxide treatment by acids and
bases. Appl. Phys. Lett. 1999,74,880-882.
30. A. Liu, S. Zhao, S.-B. Rim, J. Wu, M. Könemann, P. Erk,P.Peumans, Control of
Electric Field Strength and Orientation at the Donor-Acceptor Interface in Organic
Solar Cells. Adv Mater 2008,20,1065-1070.
90
Chapter 4 Organic Light-Emitting Diodes on Solution-Processed Graphene Transparent Electrodes
Theoretical estimates indicate that graphene thin films can be used as transparent
electrodes for thin-film devices such as solar cells and organic light-emitting diodes,
with an unmatched combination of sheet resistance and transparency. We demonstrate
organic light-emitting diodes with solution-processed graphene thin film transparent
conductive anodes. The graphene electrodes were deposited on quartz substrates by
spin-coating of an aqueous dispersion of functionalized graphene, followed by a
vacuum anneal step to reduce the sheet resistance. Small molecular weight organic
materials and a metal cathode were directly deposited on the graphene anodes,
resulting in devices with a performance comparable to control devices on indium-tin-
oxide transparent anodes. The outcoupling efficiency of devices on graphene and
indium-tin-oxide is nearly identical, in agreement with model predictions.
91
4.1 Introduction
Organic light-emitting diodes (OLEDs) are a promising electronic display and solid-
state lighting technology because of their combination of high luminous efficiency and
compatibility with a wide variety of substrates. Extensive research efforts1-7 over the
last two decades have resulted in improvements in the luminous efficiency, lifetime,
and color gamut of both small molecular weight and polymer OLEDs since the
introduction of the bilayer organic electroluminescent heterostructure diode by C. W.
Tang.8
An important component of OLEDs is the transparent conductive electrode
through which light couples out of the devices. Indium-tin-oxide (ITO) is traditionally
used as transparent conductor, but has a number of disadvantages. First, ITO may be
too expensive9 for use in OLEDs for lighting because of the cost of Indium and the
low-throughput deposition process used. Second, metal oxides such as ITO (about
150nm thick to ensure electrical performance) are brittle and therefore of limited use
on flexible substrates.10 Third, Indium is known to diffuse into the active layers of
OLEDs, which leads to a degradation of performance over time.11 There is a clear
need for alternative transparent electrodes whose optical and electrical performance is
similar to that of ITO but without its drawbacks. Several alternative transparent
electrodes have been demonstrated, including other transparent conductive oxides,12, 13
very thin metal films,14 carbon nanotube (CNT) random meshes,7, 15 metal nanowire
random meshes16 and metal gratings.17 While substantial progress has been made,
92
many issues remain to be addressed, such as performance, cost, lifetime, roughness,
manufacturability, etc. In this work, we demonstrate OLEDs on transparent graphene
electrodes, with a device performance that is competitive with that of control devices
on ITO.
4.2 Comparison of different transparent candidates
Graphene has promise as a transparent conductor because of its unique electronic
structure.18-20 In principle, charge carriers in an individual graphene sheet delocalize
over the entire sheet, and can travel thousands of inter-atomic distances without
scattering. Since graphene is a zero-gap semiconductor with a very high Fermi
velocity vF = 106m/s, individual graphene sheets have very high in-plane
conductivities.21 Graphene can be chemically doped at doping levels of Ni = 1012 cm-2
while maintaining charge carrier mobilities of μ = 105cm2/Vs or higher.20, 22, 23 Based
on these values, the sheet resistance of graphene is:
NNNeR
ish
4.621
(4.1)
where N is the number of monolayers of graphene. The sheet resistance of a graphene
film is plotted as a function of film thickness (assuming that each graphene layer is
0.34nm thick) in Figure 4.1, and compared to experimental data for ITO thin films24
93
and other recently-developed alternative transparent electrodes such as CNT random
meshes,7, 25, 26 thin metal films,27, 28 metal gratings17 and Ag nanowire networks.29
1 10 10010-1
100
101
102
103
ITO_ref.24 CNT film 1_ref.25 CNT film 2_ref.7&26 Al thin film_ref.28 Metal grating_ref.17 Ag thin film_ref.27 Ag nanowire_ref.29
Graphene (calculation) Graphene (this work)
Ni = 3x1012cm-2
= 104cm2/Vs
Ni = 1012cm-2
= 2x105cm2/Vs
She
et R
esis
tanc
e (
/sq.
)
Thickness (nm)
Fig. 4.1 Film thickness dependence of sheet resistance for different
transparent conductors. Two limiting values were calculated for graphene
using = 2x105 cm2/Vs at Ni = 1012 cm-2 and = 104 cm2/Vs at Ni = 3x1012
cm-2. The thicknesses of the Ag nanowire film and metal grating are mean
thicknesses obtained by spreading the mass of the metal over the entire
substrate area.
A range of resistivities was calculated for graphene films by using the limiting values
for mobility and carrier density of = 2x105cm2/Vs and Ni = 1012cm-2, and =
94
104cm2/Vs and Ni = 3x1012cm-2.22, 30 To achieve a sheet resistance of 10Ω/, >100 nm
thick films of CNTs and ITO are needed. In contrast, <10 nm of graphene is in theory
sufficient to achieve a sheet resistance <10Ω/. Only metal thin films and gratings
approach the sheet resistance of graphene for a thickness between 10 and 20 nm.
Figure 4.2 summarizes the optical transmission averaged over the 400-800 nm
spectral range weighted by the AM1.5G solar spectrum, as a function of sheet
resistance for different candidate transparent conductors for the structure
air/glass/transparent conductor/air or air/polyethylene terephthalate (PET)/transparent
conductor/air.7, 17, 25, 26, 29, 31 The sheet resistance of graphene was calculated as
described above assuming = 2x105cm2/Vs and Ni = 1012cm-2. The transmission of
graphene was calculated using the transfer matrix method32 and spectrally averaged as
described above. The literature reports a range of values for the optical properties of
graphene, varying from n = 2.0+1.1i 33 to n = 2.6+1.3i 34, 35 with the latter value being
that of graphite. A range of transmissivities was calculated for graphene using these
extreme values. This analysis shows that graphene has the potential to achieve a
superior combination of sheet resistance and solar averaged optical transmission.
95
1 10 100 1000
40
50
60
70
80
90
100
ITO_ref.29 CNT film 1_ref.25 CNT film 2_ref.7&26 Ag grating_ref.29 (theory) Metal grating_ref.17 Ag thin film_ref.31 Ag nanowire_ref.29
Graphene (calculation) Graphene (this work)
PET
Glass
n = 2.6+1.3i
n = 2+1.1i
Ni = 1012cm-2
= 2x105cm2/Vs
Sol
ar T
rans
mis
sion
(%
)
Sheet Resistance (/sq.)
Fig. 4.2 Solar photon flux-weighted transmission vs. sheet resistance for
different transparent conductors. Two limiting lines of graphene film were
calculated by using two different refractive indexes n = 2.0+1.1i and n =
2.6+1.3i for graphene, assuming = 2x105 cm2/Vs and Ni = 1012 cm-2. Two
horizontal lines represent the transmission through bare glass (dashed line, n
= 1.463) and PET (Polyethylene terephthalate) (dotted line, n = 1.575) by
including Fresnel reflections at both the air/(glass or PET) and (glass or
PET)/air interfaces. Solar transmission is calculated by integrating the
product of the spectrally resolved transmittance with the spectrally resolved
AM1.5 photon flux over the spectral range of 400-800nm.
96
As mentioned in chapter 3, various methods to make thin-films of graphene have
been developed such as the scotch tape-method18 and epitaxial growth,36 but these
methods are not suitable for low-cost, large-area applications. The graphene electrodes
used in this work were deposited on quartz slides by spin-coating water-based
dispersions of functionalized graphene (see Methods). The functionalized graphene
material was prepared through a modified Hummers method as described elsewhere.37,
38 The resulting functionalized graphene films must be reduced, increasing both the
light absorption and electrical conductivity of the film. High-temperature vacuum
annealing was used here, which was shown to be more effective than chemical
reduction methods.38, 39 The sheet resistance and solar transmittance of the resulting
films are plotted in Fig. 4.1 and Fig. 4.2 (crosses). The thickness of the graphene films
used is ~7nm, and the corresponding sheet resistance and transmission are ~800Ω/
and 82% at 550nm, respectively. The sheet resistance is >2 orders of magnitude higher
than our theoretical estimates because large-area graphene films produced via solution
processing of functionalized graphene contain multiple grain boundaries and
incorporate lattice defects and oxidative traps that limit charge carrier transport.40, 41
Graphene thin films have been used as an anode in dye-sensitized solar cells42 and
organic solar cells.39,43, 44 Electroluminescent thin-film devices have different
requirements for the transparent electrode because both carrier injection and light
outcoupling must be considered. OLEDs on ITO benefit from optical interference
effects in the ITO layer to control the outcoupling efficiency and radiation pattern.45-47
Here, we demonstrated OLEDs on graphene thin films with electrical and optical
performance comparable to that of control devices on ITO.
97
4.3 Organic light-emitting diodes on graphene transparent electrode
4.3.1 OLEDs device fabrication
Small molecule organic light-emitting diodes were fabricated on graphene films
on quartz and on commercially obtained 130nm-thick ITO on glass (sheet resistance
<20/sq) as transparent anode, as shown in Fig. 4.3. The ITO substrates were cleaned
in acetone, and isopropyl alcohol followed by UV-ozone treatment for 15 minutes.
Poly(3,4-ethylenedioxythio- phene)poly(styrenesulfonate) (PEDOT:PSS) was spin-
coated on both graphene and ITO substrates followed by annealing in N2 glove-box
for 20 minutes. All organic materials were obtained commercially and then purified
using thermal gradient sublimation as least twice. The organic thin films and metal
cathode were deposited at room temperature by thermal evaporation in high vacuum
(~10-7 Torr). The deposition rate is 1-2 Å/s for organic materials, and 4-5 Å/s for metal
cathode. The layer structure of the devices is anode/Poly(3,4-ethylenedioxythiophene)
poly(styrenesulfonate) (PEDOT:PSS)/ N,N′-di-1-naphthyl-N,N′-diphenyl-1,1′-
biphenyl- 4,4′diamine (NPD)/ tris(8-hydroxyquinoline) aluminum (Alq3)/ lithium
fluoride (LiF)/Al. The Al cathode was deposited through a shadow mask with 1mm
diameter circular openings.
98
Fig. 4.3 Molecule structure of organic materials and device structure of
organic light-emitting diodes
4.3.2 Device measurement
Current-voltage characteristics were measured using an HP 4145B semiconductor
parameter analyzer. Luminance was measured by a flat calibrated silicon photodiode
that captures light emitted within a cone of 55, and then adjusted to include un-
captured emissive power. For a Lambertian emission source, the power emitted within
a cone is:
)2cos1(2
sin*cos2
0 0
'''
dd (4.2)
The fraction of optical power emitted by the OLED that is captured by our photodiode
is:
%67))180cos(1(
))110cos(1(
90
55
o
o
o
o
(4.3)
ITO (130nm)/G (7nm) PEDOT (40nm)
NPD (50nm) Alq3 (50nm) LiF (0.5nm)
Al
Glass
99
The luminous power reported is the value measured by the photodiode (corrected for
the spectral response of the photodiode), divided by the fraction captured (67%).
4.3.3 Device performance
Figure 4.4 shows the current density and luminance as a function of applied
forward bias for typical devices. OLEDs on graphene have a current drive and light
emission intensity comparable to control devices on ITO for current densities
<10mA/cm2. At current densities >10mA/cm2, the sheet resistance of graphene leads
to a voltage drop in the electrode. Luminance was measured by a flat calibrated silicon
photo detector that captures 67% of the emitted light (see Methods). The OLED turn-
on voltage (defined at 0.02cd/m2) at 4.5V and 3.8V, and reach a luminance of
300cd/m2 at 11.7V and 9.9V, for graphene and ITO anodes, respectively.
100
0 5 10 15 2010-7
10-5
10-3
10-1
101
103
10-1
101
103
105
Lum
inan
ce (
cd/m
2 )
Cur
rent
Den
sity
(m
A/c
m2 )
Applied Voltage (V)
Graphene ITO control
Fig. 4.4 Current density (filled symbols) and luminance (open symbols)
vs. applied forward bias for an OLED on graphene (squares) and ITO
(circles), with OLED device structure
anode/PEDOT:PSS/NPD(50nm)/Alq3(50nm)/LiF/Al.
In Figure 4.5, the external quantum efficiency (EQE) and luminous power
efficiency (LPE) of both graphene-based and ITO-based OLEDs are shown. The
graphene-based OLED performance nearly matches that of the ITO control device
despite the higher sheet resistance and different workfunction39 of the graphene anode.
The conductive PEDOT:PSS layer further removes the effects of small differences in
workfunction that may exist between ITO and graphene.
101
10-2 10-1 100 101 102 103 10410-2
10-1
100
0.0
0.3
0.6
0.9
1.2
1.5
Lum
inou
s P
ower
Effi
cien
cy (
lm/W
)
Ext
erna
l Qua
ntum
Effi
cien
cy (
%)
Current Density (mA/cm2)
Graphene ITO control
Fig. 4.5 External quantum efficiency (EQE) (filled symbols) and
luminous power efficiency (LPE) (open symbols) for an OLED on graphene
film (squares) and ITO glass (circles). The luminance, EQE, and LPE were
adjusted to reflect total front emission power according to the method
described in the Methods.
4.3.4 Device simulation
The nearly identical EQE is surprising because the fraction of optical power that
couples out of an OLED structure depends strongly on the thickness of the various
102
layers.48, 49 As shown in Fig. 4.6, light generated in organic layer within OLEDs can
be distributed into air modes, substrate modes, wave-guided modes, surface plasmon
polariton modes, anode absorption and metal absorption, where only air modes are the
light coupling out of the device and available to measure by photo-detector.
Fig. 4.6 Schematic of light distribution of OLEDs. The light can be
distributed into air modes, substrate modes, wave-guide modes, surface
plasma modes, anode absorption and metal absorption.
The angular and spectral distribution of light emitted from the organic active layer
was modeled as described elsewhere.50 Based on measured light spectrum of
NPD/Alq3 OLED in Fig. 4.7, Fig. 4.8 shows the simulated optical power emitted into
103
unbound (UB) air modes, substrate trapped (ST) modes, waveguided (WG) modes,
and surface plasmon polariton (SPP) modes at the metal cathode/organic interface.
The fraction of photons emitted into UB modes is the outcoupling efficiency. When
spectrally integrated, we find that 17% of the optical power couples out for this OLED
structure on both graphene and ITO, despite the differences in layer structure.47 The
detailed power distribution among different modes is summarized in Table 1.
200 300 400 500 600 700 800 900
0.0
0.2
0.4
0.6
0.8
1.0
Inte
nsity
(a.
u.)
Wavelength (nm)
501.24nm
ITO/NPD 80nm/Alq3 50nm/LiF 0.8nm/Al 50nm
Fig. 4.7 Measured light spectrum of OLED with the structure as shown
in the plot
104
Table 4.1 Simulated fraction of emitted optical power into the various modes
supported by the OLED structures
Air Modes
ST Modes
WG Modes
SPP Modes
Anode Absorption
Metal Absorption
Graphene 16.54% 22.86% 15.15% 40.54% 4.01% 0.90%
ITO 17.05% 23.51% 14.59% 40.77% 3.19% 0.88%
Air Modes
ST Modes
WG Modes
SPP Modes
Anode Absorption
Metal Absorption
Graphene 16.54% 22.86% 15.15% 40.54% 4.01% 0.90%
ITO 17.05% 23.51% 14.59% 40.77% 3.19% 0.88%
400 500 600 7000.0
1.0
2.0
3.0
4.0 Graphene ITO control
Air Modes(~17%)
Sub Modes(~23%)
W/G Modes(~15%)
SP Modes(~41%)
x10-15
Nor
mal
ized
Fra
ctio
nal P
ower
Wavelength (nm)
Fig. 4.8 Simulated angularly-integrated spectral density of emitted power
into unbound (UB) air modes, substrate trapped (ST) modes, wave-guided
(WG) modes, and surface plasmon polariton (SPP) modes.
105
Figure 4.9 shows the calculated spectrally-integrated power density of emitted
radiation per unit solid angle into UB optical modes at different azimuthal angles from
the surface normal. The emission pattern is nearly Lambertian for both graphene and
ITO-based OLEDs.
0.02
0.04
0.06
0
30
60
900.00
Lambertian
ITOGraphene
(b)
0.02
0.04
0.06
0
30
60
900.00
Lambertian
ITOGraphene
0.02
0.04
0.06
0
30
60
900.00
Lambertian
ITOGraphene
(b)
Fig. 4.9 Simulated spectrally-integrated emitted power per unit solid
angle for OLEDs on graphene film (blue dash-dotted line) and control ITO
(red solid line) glass. A Lambertian response (black dashed line) is plotted
for reference.
Figure 4.10 shows the simulated fraction of power emitted as a function of
wavelength and normalized in-plane wave-vector kxy for the OLED structure on
graphene and ITO (assuming a glass substrate). While the fraction of power emitted
into the various modes is nearly identical for devices on graphene and ITO (Fig. 4.8),
the power distribution patterns within each layer are quite different because of the
differences in layer structure. Specifically, the OLED structure on graphene does not
support a WG mode due to the thin overall dielectric structure.
106
(a) ITO control
(b) Graphene
Wav
elen
gth
(n
m)
Air modes
ST modes WG modes
SPP modes
Normalized wave-vector kxy
(a) ITO control(a) ITO control
(b) Graphene(b) Graphene
Wav
elen
gth
(n
m)
Air modes
ST modes WG modes
SPP modes
Normalized wave-vector kxy
Fig. 4.10 The fraction of power emitted as a function of normalized in-
plane wave-vector kxy for all wavelengths between 400nm and 800nm for an
OLED on ITO (a) and graphene (b), with device structure
anode/PEDOT:PSS/NPD(50nm)/Alq3(50nm)/LiF/Al as shown in Fig. 4.3.
The fraction of power coupled to air modes, substrate modes, wave-guided
modes and surface plasmon polariton modes is evaluated by integrating the
power distribution over the appropriate ranges of the in-plane wave-vector
kxy.
107
4.4 Status of graphene film by CVD
Based on the comparison of different transparent electrodes in section 4.2,
graphene thin film is one of the most promising candidates. Also, both organic
photovoltaic cells and organic light-emitting diodes have been demonstrated on
solution-processed graphene transparent electrodes with very comparable performance
as ITO control device. However, one of drawbacks of solution-processed graphene
film is high sheet resistance, as shown in Fig. 4.11 (blue crosses).
Very recently, Sukang Bae et. al. reported roll-to-roll production of very large area
graphene film (30in).51 Graphene is first grown on flexible copper substrate, and then
transferred to target substrate by using polymer support and etching of copper layer.
Graphene-based touch screen panel was demonstrated.
Transmittance of 97.4% and sheet resistance of 125 ohm/sq. was achieved for
monolayer graphene. Sheet resistance and transmittance for different number of
graphene layer was summarized in Table 4.2 and plotted in Fig. 4.11 (red circles). As
shown in the plot, graphene performance is approaching to theoretical estimation, very
comparable to ITO. This is very encouraging study for graphene transparent conductor.
108
Table 4.2 Sheet resistance and transmittance for different number of graphene layer
Number of Graphene Layer Sheet Resistance (Ω/sq.) Transmittance (%)
1 125 97.4
2 67 95.1
3 42.8 92.9
4 30 90
1 10 100 1000
40
50
60
70
80
90
100
ITO (a) RGO (this work) CVD R2R doping Graphene (b)
PET
Glass
n = 2.6+1.3i
n = 2+1.1i
So
lar
Tra
nsm
issi
on
(%
)
Sheet Resistance (/sq.)1 10 100 1000
40
50
60
70
80
90
100
ITO (a) RGO (this work) CVD R2R doping Graphene (b)
PET
Glass
n = 2.6+1.3i
n = 2+1.1i
So
lar
Tra
nsm
issi
on
(%
)
Sheet Resistance (/sq.)
Fig. 4.11 The relationship of solar transmission vs. sheet resistance for
ITO and graphene films. Texture-filled area for theoretically-estimated
graphene, blue crosses for solution-processed graphene, and red circle for
CVD-growed graphene.
109
4.5 Conclusion
In conclusion, we demonstrated that graphene thin films can be used as transparent
conductive anode in OLEDs. The graphene thin films were prepared by solution
processing of functionalized graphene followed by a reduction step to improve the
conductivity. The electrical and optical performance of a small molecule OLED on
graphene is similar to that of control devices on ITO despite marked differences in the
total thickness of the optical stack. The graphene film used in this work is ~7nm thick
and consists mostly of multiple layers of graphene. The film roughness is < 3nm,
which is adequate for use as an anode in OLEDs and organic solar cells. Both
transmittance and sheet resistance decrease with increasing film thickness. Thicker
graphene film would help reduce the sheet resistance but the increased optical
absorption due to the thicker film would decrease the photon outcoupling efficiency.
Further optimization of the graphene film quality and thickness may lead to improved
device performance. For graphene to become a viable alternative to ITO, further work
is needed to develop methods to deposit high-quality, thin layers of graphene on low-
cost plastic substrates in a cost-effective manner.52, 53
110
Bibliography
1. Baldo, M. A.; O'Brien, D. F.; You, Y.; Shoustikov, A.; Sibley, S.; Thompson, M. E.;
Forrest, S. R. Highly efficient phosphorescent emission from organic
electroluminescent devices. Nature 1998, 395, 151-154.
2. Sun, Y.; Giebink, N. C.; Kanno, H.; Ma, B.; Thompson, M. E.; Forrest, S. R.
Management of singlet and triplet excitons for efficient white organic light-emitting
devices. Nature 2006, 440, 908-912.
3. D'Andrade, B. W; Holmes, R. J.; Forrest, S. R. Efficient Organic
Electrophosphorescent White-Light-Emitting Device with a Triple Doped Emissive
Layer. Adv. Mater. 2004, 16, 624-628.
4. Burroughes, J. H.; Bradley, D. D. C.; Brown, A. R.; Marks, R. N.; Mackay, K.;
Friend, R. H.; Burns, P. L.; Holmes, A. B. Light-emitting diodes based on conjugated
polymers. Nature 1990, 347, 539-541.
5. Friend, R. H.; Gymer, R. W.; Holmes, A. B.; Burroughes, J. H.; Marks, R. N.;
Taliani, C.; Bradley, D. D. C.; Santos, D. A. D.; Bredas, J. L.; Logdlund, M.; Salaneck,
W. R. Electroluminescence in conjugated polymers. Nature 1999, 397, 121-128.
111
6. He, G.; Pfeiffer, M.; Leo, K.; Hofmann, M.; Birnstock, J.; Pudzich, R.; Salbeck, J.
High-efficiency and low-voltage p-i-n electrophosphorescent organic light-emitting
diodes with double-emission layers. Appl. Phys. Lett. 2004, 85, 3911-3913.
7. Li, J.; Hu, L.; Wang, L.; Zhou, Y.; Gruner, G.; Marks, T. J. Organic Light-Emitting
Diodes Having Carbon Nanotube Anodes. Nano Letters 2006, 6, 2472-2477.
8. Tang, C. W.; VanSlyke, S. A. Organic electroluminescent diodes. Appl. Phys. Lett.
1987, 51, 913-915.
9. Forrest, S. R. The path to ubiquitous and low-cost organic electronic appliances on
plastic. Nature 2004, 428, 911-918.
10. Chen, Z.; Cotterell, B.; Wang, W.; Guenther, E.; Chua, S. A mechanical
assessment of flexible optoelectronic devices. Thin Solid Films 2001, 394, 201-205.
11. Lee, S. T.; Gao, Z. Q.; Hung, L. S. Metal diffusion from electrodes in organic
light-emitting diodes. Appl. Phys. Lett. 1999, 75, 1404-1406.
12. Kim, H.; Gilmore, C. M.; Horwitz, J. S.; Pique, A.; Murata, H.; Kushto, G. P.;
Schlaf, R.; Kafafi, Z. H.; Chrisey, D. B. Transparent conducting aluminum-doped zinc
oxide thin films for organic light-emitting devices. Appl. Phys. Lett. 2000, 76, 259-261.
13. J. Cui, A. Wang, N. L. Edleman, J. Ni, P. Lee, N. R. Armstrong,T.J.Marks, Indium
Tin Oxide Alternatives - High Work Function Transparent Conducting Oxides as
Anodes for Organic Light-Emitting Diodes. Adv. Mater. 2001, 13, 1476-1480.
112
14. Pode, R. B.; Lee, C. J.; Moon, D. G.; Han, J. I. Transparent conducting metal
electrode for top emission organic light-emitting devices: Ca--Ag double layer. Appl.
Phys. Lett. 2004, 84, 4614-4616.
15. Zhang, D.; Ryu, K.; Liu, X.; Polikarpov, E.; Ly, J.; Tompson, M. E.; Zhou, C.
Transparent, Conductive, and Flexible Carbon Nanotube Films and Their Application
in Organic Light-Emitting Diodes. Nano Letters 2006, 6, 1880-1886.
16. Lee, J. -Y.; Peumans, P. Unpublished.
17. Kang, M.-G.; Guo, L. J. Nanoimprinted Semitransparent Metal Electrodes and
Their Application in Organic Light-Emitting Diodes. Adv. Mater. 2007, 19, 1391-1396.
18. Novoselov, K. S.; Geim, A. K.; Morozov, S. V.; Jiang, D.; Zhang, Y.; Dubonos, S.
V.; Grigorieva, I. V.; Firsov, A. A. Electric Field Effect in Atomically Thin Carbon
Films. Science 2004, 306, 666-669.
19. Zhang, Y.; Tan, Y.; Stormer, H. L.; Kim, P. Experimental observation of the
quantum Hall effect and Berry's phase in graphene. Nature 2005, 438, 201-204.
20. Geim, A. K.; Novoselov, K. S. The rise of graphene. Nat. Mater. 2007, 6, 183-191.
21. Fang, T.; Konar, A.; Xing, H.; Jena, D. Carrier statistics and quantum capacitance
of graphene sheets and ribbons. Appl. Phys. Lett. 2007, 91, 092109.
113
22. Schedin, F.; Geim, A. K.; Morozov, S. V.; Hill, E. W.; Blake, P.; Katsnelson, M. I.;
Novoselov, K. S. Detection of individual gas molecules adsorbed on graphene. Nat.
Mater. 2007, 6, 652-655.
23. Chen, J.; Jang, C.; Xiao, S.; Ishigami, M.; Fuhrer, M. S. Intrinsic and extrinsic
performance limits of graphene devices on SiO2. Nat. Nano 2008, 3, 206-209.
24. Kim, D.; Park, M.; Lee, H.; Lee, G. Thickness dependence of electrical properties
of ITO film deposited on a plastic substrate by RF magnetron sputtering. Appl. Surf.
Sci. 2006, 253, 409-411.
25. Kymakis, E.; Stratakis, E.; Koudoumas, E. Integration of carbon nanotubes as hole
transport electrode in polymer/fullerene bulk heterojunction solar cells. Thin Solid
Films 2007, 515, 8598-8600.
26. Gruner, G. Carbon nanotube films for transparent and plastic electronics. Journal
of Materials Chemistry 2006, 16, 3533-3539.
27. O'Connor, B.; Haughn, C.; An, K.; Pipe, K. P.; Shtein, M. Transparent and
conductive electrodes based on unpatterned, thin metal films. Appl. Phys. Lett. 2008,
93, 223304.
28. Camacho, J. M.; Oliva, A. I. Surface and grain boundary contributions in the
electrical resistivity of metallic nanofilms. Thin Solid Films 2006, 515, 1881-1885.
114
29. Lee, J. -Y.; Connor, S. T.; Cui, Y.; Peumans, P. Solution-Processed Metal
Nanowire Mesh Transparent Electrodes. Nano Lett. 2008, 8, 689-692.
30. Bolotin, K. I.; Sikes, K. J.; Jiang, Z.; Klima, M.; Fudenberg, G.; Hone, J.; Kim, P.;
Stormer, H. L. Ultrahigh electron mobility in suspended graphene. Solid State
Commun. 2008, 146, 351-355.
31. Charton, C.; Fahland, M. Optical properties of thin Ag films deposited by
magnetron sputtering. Surface and Coatings Technology 2003, 174-175, 181-186.
32. Heavens, O. S. Optical Properties of Thin Solid Films; Dover: New York, 1965;
pp 46-95.
33. Ni, Z. H.; Wang, H. M.; Kasim, J.; Fan, H. M.; Yu, T.; Wu, Y. H.; Feng, Y. P.;
Shen, Z. X. Graphene Thickness Determination Using Reflection and Contrast
Spectroscopy. Nano Letters 2007, 7, 2758-2763.
34. Blake, P.; Hill, E. W.; Neto, A. H. C.; Novoselov, K. S.; Jiang, D.; Yang, R.;
Booth, T. J.; Geim, A. K. Making graphene visible. Appl. Phys. Lett. 2007, 91, 063124.
35. Nair, R. R.; Blake, P.; Grigorenko, A. N.; Novoselov, K. S.; Booth, T. J.; Stauber,
T.; Peres, N. M. R.; Geim, A. K. Fine Structure Constant Defines Visual Transparency
of Graphene. Science 2008, 320, 1308.
36. Rollings, E.; Gweon, G. -H.; Zhou, S. Y.; Mun, B. S.; McChesney, J. L.; Hussain,
B. S.; Fedorov, A. V.; First, P. N.; de Heer, W. A.; Lanzara, A. Synthesis and
115
characterization of atomically thin graphite films on a silicon carbide substrate.
Journal of Physics and Chemistry of Solids 2006, 67, 2172-2177.
37. Hummers, W. S.; Offeman, R. E. Preparation of Graphitic Oxide. J. Am. Chem.
Soc. 1958, 80, 1339-1339.
38. Becerril, H. A.; Mao, J.; Liu, Z.; Stoltenberg, R. M.; Bao, Z.; Chen, Y. Evaluation
of Solution-Processed Reduced Graphene Oxide Films as Transparent Conductors.
ACS Nano 2008, 2(3), 463-470.
39. Wu, J.; Becerril, H. A.; Bao, Z.; Liu, Z.; Chen, Y.; Peumans, P. Organic solar cells
with solution-processed graphene transparent electrodes. Appl. Phys. Lett. 2008, 92,
263302.
40. Mkhoyan, K. A.; Contryman, A. W.; Silcox, J.; Stewart, D. A.; Eda, G.; Mattevi,
C.; Miller, S.; Chhowalla, M. Atomic and Electronic Structure of Graphene-Oxide.
Nano Letters 2009, 9, 1058-1063.
41. Mattevi, C.; Eda, G.; Agnoli, S.; Miller, S.; Mkhoyan, K. A.; Celik, O.;
Mastrogiovanni, D.; Granozzi, G.; Garfunkel, E.; and Chhowalla M. Evolution of
Electrical, Chemical, and Structural Properties of Transparent and Conducting
Chemically Derived Graphene Thin Films. Adv. Funct. Mater. 2009, 19, 1-7.
42. Wang, X.; Zhi, L.; Mullen, K. Transparent, Conductive Graphene Electrodes for
Dye-Sensitized Solar Cells. Nano Letters 2008, 8, 323-327.
116
43. Eda, G.; Lin, Y.; Miller, S.; Chen, C.; Su, W.; Chhowalla, M. Transparent and
conducting electrodes for organic electronics from reduced graphene oxide. Appl.
Phys. Lett. 2008, 92, 233305.
44. Tung, V. C.; Chen, L.; Allen, M. J.; Wassei, J. K.; Nelson, K.; Kaner, R. B.; Yang,
Y. Low-Temperature Solution Processing of Graphene−Carbon Nanotube Hybrid
Materials for High-Performance Transparent Conductors. Nano Letters 2009, 9, 1949-
1955.
45. Greenham, N. C.; Friend, R. H.; Bradley, D.D.C. Angular Dependence of the
Emission from a Conjugated Polymer Light-Emitting Diode: Implications for
efficiency calculations. Adv. Mater. 1994, 6, 491-494.
46. Madigan, C. F.; Chen, L.-M.; Sturm, J. C. Improvement of output coupling
efficiency of organic light-emitting diodes by backside substrate modification. Appl.
Phys. Lett. 2000, 76, 1650-1652.
47. Agrawal, M.; Sun, Y.; Forrest, S. R.; Peumans, P. Enhanced outcoupling from
organic light-emitting diodes using aperiodic dielectric mirrors. Appl. Phys. Lett. 2007,
90, 241112.
48. D'Andrade, B. W.; Brown, J. J. Organic light-emitting device luminaire for
illumination applications. Appl. Phys. Lett. 2006, 88, 192908.
117
49. Lu, M. -H.; Weaver, M. S.; Zhou, T. X.; Rothman, M.; Kwong, R. C.; Hack, M.;
Brown, J. J. High-efficiency top-emitting organic light-emitting devices. Appl. Phys.
Lett. 2002, 81, 3921-3923.
50. Agrawal, M.; Peumans, P. Design of non-periodic dielectric stacks for tailoring the
emission of organic lighting-emitting diodes. Opt. Express 2007, 15, 9715-9721.
51. Bae, S.; Kim, H.; Lee, Y.; et.al., Roll-to-roll prodection of 30-inch graphene films
for transparent electrodes. Nat. Nano. 2010, 5(8), 574-578.
52. Kim, K. S.; Zhao, Y.; Jang, H.; Lee, S. Y.; Kim, J. M.; Kim, K. S.; Ahn, J.; Kim,
P.; Choi, J.; Hong, B. H. Large-scale pattern growth of graphene films for stretchable
transparent electrodes. Nature 2009, 457, 706-710.
53. Reina, A.; Jia, X.; Ho, J.; Nezich, D.; Son, H.; Bulovic, V.; Dresselhaus, M. S.;
Kong, J. Large Area, Few-Layer Graphene Films on Arbitrary Substrates by Chemical
Vapor Deposition. Nano Letters 2009, 9, 30-35.
118
Chapter 5 Modeling of Interface Morphology at Organic/Metal Interface
5.1 Introduction
Due to their exciting electronic, optical, and optoelectronic properties, organic
semiconductors have experienced a tremendous increase in research activity over the
last few years. Some organic electronic devices have been demonstrated, such as
organic light-emitting diode display (OLED), organic solar cell, organic thin film
transistor, and so on 1, 2. And some new structures, such as organic bistable device
(OBD), have been reported recently3, 4. Most these devices operate with one or several
functional organic layers with desired physical properties and metal-organic contacts
to inject/extract charge carriers. In some of these device geometries, metal is needed to
deposit on the top of the organic film. The proper function of the contact is critical for
the performance of the device, and it depends on the interfacial morphology of the
metal-organic interface. Until now, many experimental data has been reported. So it is
quite interesting and useful to simulate the interface morphology at organic/metal
interfaces.
In this chapter, a simple model is proposed to simulate the interface morphology when
metal atoms are deposited on the surface of organic thin film. We mainly discuss how
119
the interface morphology depends on the conditions employed during deposition of the
metal layer, such as deposition rate, substrate temperature and drive-in potential, and
the properties of involved materials (e.g., diffusion coefficient).
5.2 Simulation method
5.2.1 Basic idea
Basically, when a metal atom is deposited on the organic surface, it has three
probabilities.5-8 First, metal atom has a chance to stick to organic molecule and does
not move any more, referring to stick probability. Second, metal atom can randomly
hop around on the surface, which results in diffusion coefficient on the surface, D//.
Third, metal atom can also hop down to penetrate the organic surface with diffusion
coefficient D, which happens along the boundary between organic molecules
generally. Therefore, random walk model (Monte Carlo) is employed to simulate the
diffusion processing of metal atoms. To start, let’s set up the workspace at first.
5.2.2 Simulation workspace
Assume that organic molecule is cubic and its size is 1nm*1nm*1nm. The molecules
are closely packed in simple cubic lattice, and dimension is 10*10*10, as shown in Fig.
120
5.1 (a). And there are boundaries between organic molecules, along which metal atom
can diffuse with diffusion coefficient D. To simplify, I assume the surface of organic
film is flat.
(a) (b)
Fig. 5.1 Cubic array of organic molecules
Metal atom is much smaller than organic molecule, and I assume the dimension of
metal atom is 1Å*1Å*1Å. Metal atom can randomly arrive at the surface with
negligible initial velocity. Upon its arrival on the surface, it can either stick there or
hop around. Like atom 1 in Fig. 5.1 (b), it can hop to four nearest neighborhood sites
on the surface with certain diffusion coefficient (diffusion_on _surface), and each
hopping step is 1Å. When metal atom appears on the boundary region, like atom 2,
besides hopping on the surface, it can also hop down to penetrate the surface with
another diffusion coefficient (diffusion_ along_boundary). Once medal atom
penetrates the surface, it can still diffuse along the boundary in this model. If a metal
atom meets other atom(s), like atom 3, they will form a cluster, and diffusion
coefficient is expected to decrease significantly so that even small cluster may be
considered as immobile. So I assume they are immobile in my model when two metal
121
atoms meet. By doing so, the disadvantage is that the shape of the cluster can not be
optimized to minimize the surface energy.
(a)
(b)
Fig. 5.2 (a) Schematic of metal atom diffusion direction; (b) Probability
distribution of each event on 1D axis
Stick probability
Down diffusion (+z)
Forward diffusion (-x)
Backward diffusion (+x)
Left diffusion (-y)
Right diffusion (+y)
Up diffusion (-z)
0 1
122
5.2.3 Monte Carlo method
Generally, each metal atom can hop to the nearest neighborhood site along six
possible directions, as shown in Fig. 5.2 (a). The probability to each direction depends
on the hopping rate. By using equation (1) to normalize as follows,
N
i i
jj
R
RP
1
(5.1)
I get have the probability of each event. As shown in Fig. 5.2 (b), these probabilities
are put on the axis between 0 and 1, and hopping direction can be determined by
matching a random number to the axis.
5.3 Results and discussion
This section will mainly discuss how the interface morphology depends on the
conditions employed during deposition. Except for specification, only the metal atoms
below the surface are shown for all morphology figures, and the metal atoms above
the surface are not. The influence of stick probability is not considered in this
simulation, which is assumed to be zero.
123
5.3.1 Substrate temperature / Deposition rate dependence
Substrate temperature and deposition rate are most important parameters controllable
during deposition. Higher temperature as well as lower deposition rate leads to larger
hopping rate, so atoms are expected to diffuse more before they meet other metal
atom(s) to form a cluster. And more atoms are expected to penetrate the surface under
higher substrate temperature and lower deposition rate.5
In this model, steps_of_walk represents metal diffusion steps on the surface during
unit time gap. And it will decrease by the ratio of diffusion coefficient on the surface
to that along the boundary while metal penetrates the surface and diffuses in the film.
Basically, high substrate temperature and/or low deposition rate can be achieved by
increasing steps_of_walk. As shown in Fig. 5.3 (a), the number and percent of atoms
penetrating the surface are calculated with increasing steps_of_walk, and the result is
the average over 100 simulations at each condition. As expected, the number of atoms
below the surface increases with increasing steps_of_walk when steps_of_walk is
relatively small. However, it does not change much when steps_of_walk is above
some number (10 in this simulation), and even decreases a little when steps_of_walk is
large.
124
100 101 102 103 104 105
100
200
300
400
0.003
0.004
0.005
0.006
0.007
0.008
0.009
0.010
0.011
num
ber
of a
tom
s be
low
the
sur
face
steps_of_walkincreasing substrate temp. OR decreasing deposition rate
pe
rce
nt
(a)
(b) steps_of_walk = 10 (c) steps_of_walk = 1000
Fig. 5.3 Substrate temperature / deposition rate dependence. Region 1:
above surface; Region 2: organic film; Region 3: bottom electrode
① ② ③ ① ② ③
125
(a) Tsub = -120oC, R = 23Å/min (b) steps_of_walk = 1
(c) Tsub = +70oC, R = 0.35Å/min (d) steps_of_walk = 10000
Fig. 5.4 Comparison with experimental data
The metal distribution profiles with different steps_of_walks are compared in Fig. 5.3
(b) and (c). The vertical axis is the number of metal atoms in log scale, and horizontal
axis is perpendicular to the film surface. The region 1, 2 and 3 are the region above
surface, within organic film and within bottom electrode, respectively. It is clear that
the metal atoms diffuse deeper in Fig. 5.3 (c), which is under higher substrate
126
temperature and lower deposition rate. Appearance of periodic peaks in region 2 is due
to boundary planes at the corresponding locations, which are parallel to the surface.
Fig. 5.4 shows the comparisons between my simulations and experimental data.5
Deposition conditions for Fig. 5.4 (a) is Tsub = -120oC and R = 23Å/min, while
deposition conditions for Fig. 5.4 (c) is Tsub = +70oC and R = 0.35Å/min. Obviously,
organic film is quite clear with little metal clusters for Fig. 5.4 (a). However, a lot of
metal clusters appear in the organic film for Fig. 5.4 (c), which is due to higher
substrate temperature and lower deposition rate. As shown in Fig. 5.4 (b) and (d), my
simulation indicates very similar morphology by changing steps_of_walk. How to
relate simulation parameter to real deposition conditions (e.g., substrate temperature
and deposition rate) need to consider further.
5.3.2 Diffusion coefficient dependence
Hopping rate or diffusion coefficient is strongly influenced by the interaction between
the metal and the organic molecules. Metal atoms that interact more strongly on the
organic film may exhibit less inter-diffusion into the organic film (i.e., boundary
diffusion).
By increasing the boundary diffusion, more metal atoms diffuse into organic film, and
they diffuse deeper as shown in Fig. 5.5. It is reasonable because larger boundary
diffusion gives metal atom more chance to diffusion down to penetrate the surface.
127
0 20 40 60 80 100 120 14010-1
100
101
102
103
104
bottom electrode
above surface organic film
num
ber
of m
eta
l ato
m
Z
boundary diffusion = 50 boundary diffusion = 20
Fig. 5.5 Diffusion coefficient dependence. Keeping surface diffusion =
100, boundary diffusion is increased from 20 to 50.
5.3.3 Effect of drive-in potential
Sometimes, potential drive-in force may influence the interface morphology, such as
electrical field, magnetic field, and so on. It can be intended or occasional. It is
interesting to simulate its influence.
128
(a) morphology without drive-in (b) morphology with drive-in
(c) metal distribution profile without
drive-in
(d) metal distribution profile with drive-in
Steps_of_walk = 100
Total deposition layer = 5
Atom below surface = 380.51 ± 28.71
Percent = 0.0076 ± 0.0006
Steps_of_walk = 100
Total deposition layer = 5
Atoms below surface = 2169.60 ± 97.31
Percent = 0.0434 ± 0.0019
Fig. 5.6 Influence of drive-in potential
129
Fig. 5.6 shows the influence of drive-in potential on metal clusters morphology below
the surface and metal atom profile. Compared to Fig. 5.6 (a) and (c), some kind of
drive-in force is employed in Fig. 5.6 (b) and (d) to double the downward diffusion
coefficient while keeping other parameters the same. Clearly, the drive-in potential
can make a big difference. As shown in Fig. 5.6 (b), much more metal atoms will
penetrate the surface, diffuse down, and eventually stick to the bottom electrode under
drive-in potential. The number of atoms below the surface increases by more than 5
times from Fig. 5.6 (a) to Fig. 5.6 (b). This phenomenon can be utilized or avoided
depending on applications.
By tracking the diffusion path of metal atom, one can come up some idea about how
the drive-in potential influences morphology. Fig. 5.7 shows all metal atoms
depositing on the surface. Upon arrival on the surface in Fig. 5.7 (a), the metal atom
will hop around and spend much time on the surface, and eventually stick to other
atom(s) to form cluster, as shown in red. However, in Fig. 5.7 (b) with drive-in
potential, the metal atom is quite easy to penetrate the surface, diffuse down and stick
to bottom electrode.
130
(a) without drive-in potential (b) with downward drive-in potential
Fig. 5.7 Track of diffusion. Diffusion path start from black dot and stop
at green dot.
5.3.4 Saturated metal atom within film
When more and more metal atoms are deposited, most atoms will stick to other atom(s)
to form a cluster upon their arrival on the surface and have no chance to diffuse further
into organic film. Therefore, atoms below the surface will almost keep constant when
deposited atoms are more than one nominal metal layer, as shown in Fig. 5.8.
131
0.01 0.1 1 100
50
100
150
200
250
300
350
400
450
0.0
0.1
0.2
0.3
0.4
0.5
0.6
perc
ent
num
ber
of a
tom
s be
low
sur
face
nominal deposition layer
Fig. 5.8 Number and percent of atoms below surface with increasing
deposition layer
5.3.5 Cluster size on surface
Now assuming that metal atoms only hop around on the surface and do not hop down
to penetrate the surface. When depositing the same number of metal atoms, high
deposition rate will result in many small clusters as shown in Fig. 5.9 (a), however,
low deposition rate will lead a few big clusters as shown in Fig. 5.9 (b). At low
deposition rate, metal atom has more time to hop around and finally stick to an existed
cluster. At high deposition rate, metal atom has a big chance to meet a newcomer to
132
form an additional cluster, which is account for larger cluster number and small cluster
size.
(a) high deposition rate (b) low deposition rate
Fig. 5.9 Deposition rate dependence of surface morphology
5.4 Conclusion
In this chapter, a random walk model is proposed to simulate the interface morphology
when depositing metal atoms on the organic film. The paper mainly discusses the
influences of substrate temperature, deposition rate, diffusion coefficient and drive-in
potential. In comparison with experimental data, this simple model gives reasonable
interface morphology and metal distribution profile. Here are some main conclusions.
133
1. Decreasing deposition rate or increasing substrate temperature will drive metal
atoms into the organic film deeper, but the number of metal atoms below the
surface will not monotonously increase with decreasing deposition rate;
2. Drive-in potential can make a big difference for interface morphology and
metal atoms distribution profile;
3. The larger the diffusion coefficient along the boundary is, the more metal
atoms are driven below the surface and the deeper;
4. After depositing one nominal metal layer, the atom number below the surface
is almost the constant and will not monotonously increase as before;
5. Considering metal atoms morphology on the surface, fast deposition rate will
give many small metal clusters, while low deposition rate will result in less big
metal clusters.
In the end, some ideas are proposed as below and further work can be done based on
this model.
1. How to relate the simulation parameters to the real properties of employed
materials, such as atom hopping rate, diffusion active energy, and so on.
2. More simulations can be done by optimizing the shape of metal cluster to
account for surface energy and changing the shape and array of organic
molecules to approach the reality.
3. In present model, substrate temperature and deposition rate are controlled by
one parameter. However, the interaction between metal and organic may
134
change with temperature, and higher temperature may enhance the diffusion
coefficient to penetrate the surface. So substrate temperature and deposition
rate should be varied separately to account for this.
4. The influences of surface roughness and stick probability need consider.
135
Bibliography
1. C. W. Tang, Appl. Phys. Lett. 1986, 48 (2), 183; C. W. Tang, and S. A. Vanslyke,
Appl. Phys. Lett. 1987, 51 (12), 913; C. W. Tang, S. A. Vanslyke, and Ch. Chen, J.
Appl. Phys. 1989, 65 (9), 3610.
2. S. R. Forrest, Chem. Rev. 1997, 97, 1793; P. Peumans, A. Yakimov, and S. R.
Forrest, J. Appl. Phys. 2003, 93 (7), 3693; A lot of recent papers can be found.
3. L. P. Ma, J. Liu, Y. Yang, Appl. Phys. Lett. 2002, 80 (16), 2997.
4. L. D. Bozano, B. W. Kean, V. R. Deline, J. R. Salem, and J. C. Scott, Appl. Phys.
Lett. 2004, 84 (4), 607.
5. A. C. Durr, et. al., J. Appl. Phys. 2003, 93 (9), 5201.
6. D. Q. Yang, and E. Sacher, J. Phys.: Condens. Matter, 2002, 14, 7097.
7. B. D. Silverman, Macromolecules, 1991, 24, 2467.
8. P. Jensen, Rev. Mod. Phys. 1999, 71, 1695.