University of Southern California Dissertations and Theses

Molecular and morphological effects on the operational parameters of organic solar cells

(USC Thesis Other)

Molecular and morphological effects on the operational parameters of organic solar cells

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content i

Molecular and Morphological Effects on the Operational
Parameters of Organic Solar Cells

By

Patrick Erwin

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(CHEMISTRY)

August 2015

ii

Dedication

To my family who readied me for this challenge,

To my love Jena who helped me finish

iii

Acknowledgments

I first want to thank Prof. Mark Thompson who gave me this opportunity to grow and
enrich myself. He taught me the fascinating science of organic photovoltaics and allowed me the
chance to study new and interesting systems. I became the scientist I am because of him. I want
to thank my committee members Prof. Richard Brutchey and Prof. Andrea Armani for their
participation in this process. I also want to thank Prof. Surya Prakash and Prof. Barry Thompson
for their time on my qualification committee.
I owe my inclination for science to my father who helped instill in me not just a capacity
for learning, which although helpful is insufficient to drive discovery, but the ability for
independent critical thought. This profoundly powerful skill which is so integral to science yet is
so infrequently taught may be the single most important factor for my success.
My first opportunity in scientific research was given to me by my first professor Dr.
Patrick Farmer who taught me much about how to deal with the successes and inevitable failures
of academic research and how to tell the difference. My graduate research career would not have
been what it was if it were not for the mentorship I received from Dr. Cody Schlenker and Dr. M.
Dolores Perez who taught me everything from the finer points of exciton migration to the proper
way to load a column. I will be forever indebted for their guidance and in fact got my first paper
by executing the “future research” section of Dolores’s last group meeting. Indeed much of this
research would not be possible without my collaborators. For this I want to thank Dr. Kenneth
Graham, Prof. Mike McGehee, Dr. Sarah Conron, Dr. Kathryn Allen, John Chen, Jessica
Golden, and Dr. Mike Dimitriou. I am also thankful for the friendships I made which have
helped me persevere through the toughest moments. The conversations I have had with John,
iv

Sarah and Denise on topics ranging from scientific to inane kept me lighthearted when the
research did its best to break my spirit. I would thank all of my co-workers as well who have
provided invaluable discussions on a myriad of scientific topics throughout my career.
Finally, I want to thank my love Jena for all of her support. The assistance you provided
both emotional and logistical has been beyond what could be expected and none of my success
would be possible without you. I love you very much.

v

Table of Contents

Dedication ……………………………………………………………………………………… ii

Acknowledgments …………………………………………………………………………….. iii

List of Tables ………………………………………………………………………………… vii

List of Figures ……………………………………………………………………………..……viii

Abstract ………………………………………………………………………………………… xii

Chapter 1 Introduction to Organic Photovoltaics ……………………………………………….. 1
1.1 Justification for Solar Research …………………………………………………………. 1
1.2 Photovoltaics …………………………………………………………………………….. 4
1.2.1 Silicon Solar Cells ……………………………………………………………….. 5
1.2.2 Organic Photovoltaics …………………………………………………………… 7
1.3 OPV Processes …………………………………………………………………………. 10
1.3.1 Adsorption ……………………………………………………………………… 12
1.3.2 Exciton Diffusion ………………………………………………………………. 15
1.3.3 Charge Transfer ………………………………………………………………... 19
1.3.4 Charge Separation ……………………………………………………………… 22
1.3.5 Charge Transport ………………………………………………………………. 27
1.3.6 Charge Collection ……………………………………………………………… 28
1.4 Characterization of OPV Devices ……………………………………………………… 30
1.4.1 Mismatch Factor Corrections and Proper Testing Conditions …………………. 32
1.4.2 Electrical Parameters and the Generalized Shockley Equation ………...……… 34
1.5 Overview of work ……………………………………………………………………… 36
1.6 Chapter 1 References …………………………………………………………………... 37
Chapter 2 Steric Bulk and Open Circuit Voltage ………………………………………………. 41
2.1 Introduction …………………………………………………………………………….. 41
2.2 The Physical Origin of V
OC
……………………………………………………….…… 42
2.2.1 Shockley Diode Equation ……………………………………………………… 43
2.2.2 Relation Between Steric Bulk and V
OC
………………………………...……… 46
2.3 Experimental …………………………………………………………………………… 47
2.4 Results and Discussion ………………………………………………………………… 48
2.4.1 Material Characterization ………………………………………………………. 48
2.4.2 Device Characterization ………………………………………………………... 53
2.4.3 Interfacial Nature of V
OC
……………………………………………...……….. 56
2.5 Conclusions …………………………………………………………………………….. 58
vi

2.6 Chapter 2 References …………………………………………………………………... 60
Chapter 3 Relation Between E
CT
and V
OC
……………………………………………………... 63
3.1 Introduction …………………………………………………………………………….. 63
3.2 Experimental …………………………………………………………………………… 64
3.3 Re-Evaluation of E
a
……………………………………………………………………. 64
3.4 Results and Discussion ………………………………………………………………… 66
3.4.1 Relation Between E
CT
and V
OC
………………………………………………… 66
3.4.2 E
CT
Measurements of PDI ……………………………………………………… 69
3.4.3 Exposure Study of Tetracene/C
60
……………………………………………… 74
3.5 Conclusions …………………………………………………………………………….. 77
3.6 Chapter 3 References …………………………………………………………………... 78
Chapter 4 Morphology of the D/A Interface in PHJs ………………………………………….. 80
4.1 Introduction …………………………………………………………………………….. 80
4.2 Experimental …………………………………………………………………………… 81
4.3 Neutron Reflectometry …………………………………………………………………. 82
4.4 Results and Discussion ………………………………………………………………… 86
4.4.1 Interfacial Depth in PHJs ………………………………………………………. 86
4.4.2 Surface Roughness and Interfacial Depth ……………………………………… 90
4.4.3 Structure of bDIP/C
60
Film Stack ……………………………………………… 92
4.4.4 PMHJ bDIP/C60 Devices ………………...…………………………………... 104
4.5 Conclusions …………………………………………………………………………… 105
4.6 Chapter 4 References …………………………………………………………………. 107
Chapter 5 Multi-Chromophoric Arrays in OPVs ……………………………………………... 111
5.1 Introduction …………………………………………………………………………… 111
5.2 Multi-Chromophoric Arrays ………………………………………………………….. 112
5.3 Experimental ………………………………………………………………………….. 115
5.4 Results and Discussion ……………………………………………………………….. 118
5.4.1 BDP-Por Devices ……………………………………………………………... 118
5.4.2 Blended Chromophore Systems ………………………………………………. 120
5.5 Conclusions …………………………………………………………………………… 127
5.6 Chapter 5 References …………………………………………………………………. 129

vii

List of Tables

Table 2.1: Comparison of averaged device parameters for ITO/CuPc/PDI/BCP//Al
devices showing the clear correlation between the dark current and open
circuit voltage …...………………………………………………………… 56

Table 2.2: The parameters for the ITO/CuPc/dmPh-PDI (X)/H-PDI/BCP//Al device.
Again there is a direct correlation here between J
S
and V
OC
.……………... 58

Table 3.1: Parameters from ITO/Ancene/C60/BCP//Al devices ………...……………. 68

Table 3.2: Comparison of device parameters from I-V curves of the structure
ITO/ZnPc/PDI/BCP//Al devices ………………………………………...... 72

Table 3.3: Comparison of the parameters from the exposure study of
ITO/Tetracene/C
60
/BCP//Al devices show similarity between the air and
oxygen exposed devices …………………………………...……………… 75

Table 4.1: The interfacial depths, as measured by neutron reflectometry, between various
donors and C
60
…………………………………………………………….. 90

Table 4.2: Comparison of AFM measured maximum height and neutron reflectometry
measured interfacial depth ………………………………………………… 92

Table 4.3: Summary of PV parameters for PMHJ ITO/bDIP/C
60
/BCP//Al devices and the
control PHJ structure …………………..………………………………… 105

viii

List of Figures

Figure 1.1 Distribution of global solar insolation in the winter (top) and spring (bottom).
(Provided by NASA athttp://earthobservatory.nasa.gov/) …………………. 3

Figure 1.2 The 2014 NREL graph on the improving efficiencies of various types of PVs.
(Provided by NREL at www.nrel.gov) ……………………………………... 7

Figure 1.3 Common materials used small molecule OPVs are listed here. The materials
are divided into groups by function in the devices of donor, acceptor, and
buffer materials ……………………………………………………………... 9

Figure 1.4 (a) The electronic structure of modern OPVs. The device architecture of
common OPVs is displayed showing both (b) planar (PHJ) and (c) bulk
heterojunction (BHJ) style devices with photo active regions sandwiched
between conducting electrodes ………………………………………….… 10

Figure 1.5 The diagram shows of the 5 steps that an OPV undergoes during the
generation of photocurrent ………………………………………………... 11

Figure 1.6 (a) The Jablonski diagram of common excited states and relaxation pathways
involved in the evolution of energy in OPVs. (b) The generic electron
configurations of the states involved ……………………………………… 13

Figure 1.7 (a) The incident solar flux in photon flux per wavelength. (b) The
absorbtivities of silicon (black trace) and PDI/C
60
film (red trace) per
centimeter, showing the magnitude difference. (c) The traces the absorbed
photons from the solar flux for a 300 nm film of each material and shows the
inherent advantage of organic materials …………………………………... 14

Figure 1.8 The difference between Forster transfer (FRET) and Dexter transfer (DET) is
diagramed …………………………………………………………………. 18

Figure 1.9 A generalized charge transfer diagram as envisaged by Marcus’ theory where
each state is shown as a harmonic oscillator ……………………………… 21

Figure 1.10 A potential energy diagram that visualizes Onsager Theory for
auto-ionization. The absorption event frees an electron to the Onsager
radius, a, forming a CT state ……………………………………………… 24

Figure 1.11 shows a general energy level diagram that summarizes the CT and CS
processes. The exciton in the blue levels gets generated and forms a charge
transfers (k
CT
). Taken from reference 35 ………………………………….. 26

ix

Figure 1.12 A typical I-V curve for a CuPc/C
60
device with several of the most important
parameters diagramed ……………………………………………………... 30

Figure 1.13 The equivalent circuit diagram of a solar cell is diagramed ……………… 34

Figure 2.1 The structure of three PDI derivatives are shown (left) with the dihedral angle
of the R group of each derivative listed (middle). The conformation of the
molecule is calculated (right) …………………………………………….. 49

Figure 2.2 (a) The solution spectra, of the three PDIs are shown. (b) The thin film
adsorption spectra are plotted and shows the effect steric bulk has on
molecular packing ………………………………………………………… 50

Figure 2.3 The excitation (dashed line) and emission (solid line) spectra, measured at
77K, of the three PDI derivatives in this study are presented …………….. 52

Figure 2.4 (a) Current-voltage curves of the ITO/CuPc/PDI/BCP//Al devices are shown
on both semi-log and linear plots. (b) EQE spectrums of the same PDI
devices …………………………………………………………………….. 54

Figure 2.5 Devices with the structure ITO/CuPc/dmPh-PDI(X)/H-PDI/BCP//Al with
varying thicknesses of dmPh-PDI are shown. By varying the interlayer
thickness the V
OC
can be tuned ……………………………………..…….. 57

Figure 3.1 (a) the I-V curves of rubrene/C
60
and tetracene/C
60
are shown here plotted on
a semi-log plot. (b) The EQE curves of the same devices are analyzed using
equation 3.3 ……………………………………………………………….. 67

Figure 3.2 In many systems the correlation between E
CT
and V
OC
is extremely strong.
Here is a series of experiments showing these results. Most devices represent
polymer BHJs. Figure taken from reference 8 …………………………… 69

Figure 3.3 E
CT
measurements of ZnPc/PDI devices are shown here. The CT transition is
modeled as a simple harmonic oscillator and the extracted parameters are
overlaid ……………………………………………………………………. 70

Figure 3.4 The I-V curves for the devices of structure ITO/ZnPc/PDI/BCP//Al on a
semi-log plot ………….…………………………………………....…….... 71

Figure 3.5 I-V curves for an exposure study on ITO/Tetracene/C
60
/BCP//Al devices.
From this data it can be seen that it is the oxygen in air that is causing the
quick degradation of the V
OC
…………………………………………...… 74

Figure 4.1 The physics behind the neutron reflectivity experiment is diagramed, showing
how waves reflecting off of different interfaces will interfere as a function of
incident angle ……………………………………………………………... 83
x

Figure 4.2 The two extremes in morphology that can give rise to interfacial depth are
depicted …………………………………………………………………… 85

Figure 4.3 The measured NR spectrum (above) for a Si/SiO
2
/CuPC/C
60
film is shown
(black squares). ). A model stack (below) was constructed that is close to
400 Å of C
60
and 240 Å of CuPc as measured by the QCM …………...…. 87

Figure 4.4 The measured NR spectrum (above) for a Si/SiO
2
/C
60
/CuPc film is shown
(black squares). A model stack (below) was constructed that is close to 40
nm of C
60
and 30 nm of CuPc, as measured by the QCM, with an interfacial
region of 5 nm ………………………………………………………….…. 88

Figure 4.5 A 25 μm
2
AFM picture of a C60 film deposited on silicon (left). The analysis
statistics of this picture were calculated and are shown (right) ………….... 91

Figure 4.6 The reflectivity spectrum (bottom) and SLD profile (top) of the Si/bDIP/C
60

stack, showing a region of extended mixing …………………………….... 94

Figure 4.7 The reflectivity spectrum (bottom) and SLD profile (top) of a Si/C
60
/bDIP
stack, which forms a very discrete bilayer ……………………………...… 95

Figure 4.8 The reflectivity spectrum (bottom) and SLD profile (top) of Si/ bDIP/C
60

gradient blend film is shown which conforms closely to the structure
measured by the QCM …………………………………………………….. 98

Figure 4.9 The reflectivity spectrum (bottom) and SLD profile (top) of Si/C
60
:bDIP flat
blend film ……………………………………………………………….… 99

Figure 4.10 Comparing the QCM measured blend ratio for the Si/C60:bDIP film is
shown with the SLD profile shows that the two correlate closely ………. 100

Figure 4.11 The reflectivity spectrum (bottom) and SLD profile (top) of a
Si/bDIP/gradient/C
60
film stack ………………………………………….. 101

Figure 4.12 The reflectivity spectrum (bottom) and SLD profile (top) of a
Si/bDIP/Blend/C
60
film stack ………………………………………….… 102

Figure 4.13 (top) Representative I-V curves under light (solid lines) and dark (dashed
lines) conditions. (bottom) EQEs of the different blend ratio PMHJ devices
show increasing efficiency in the bDIP region ………………………..… 104

Figure 5.1The structures of the porphyrin PtTPBP and the BDP-Por multichromophoric
array are shown ………………………………………………………..… 113

xi

Figure 5.2 (a) The absorption spectra of neat PtTPBP and C
60
and the EQE of a
PtTPBP/C60 device are overlaid. (b) The absorbance spectra of PtTPBP,
BDP1, and the arithmetic sum of these two plotted against the absorption
spectrum of BDP-Por ……………………………………………………. 114

Figure 5.3 (a) J-V curves for BDP-Por (X nm)/C
60
(40 nm) devices are shown against the
solution processed PtTPBP (15 nm)/C
60
(40 nm) device. (b)The EQE plot
shows enhanced spectral response past 550 nm in the same region as the
BDP adsorption ………………………………………………………..… 120

Figure 5.4 (a) The structures of BDP1 and BDP2 are shwon. (b) The thin film
absorptivities of three different ratios of Por+BDP1 1:2 (red), 1:4 (blue), and
1:6 (green) before (squares) and after (circles) being exposed to high vacuum
conditions for 2 hrs ………………………………………………………. 122

Figure 5.5 (a) J-V curves of devices made with Por+BDP2 showing the 1:2 ratio with
highest efficiency. (b) The EQE spectra of the same devices shows a
shoulder increasing BDP2 concentration ………………………………... 123

Figure 5.6 (a) The J-V curves of an ITO/Por+BDP2/C
60
/BCP//Al device after varying
periods of illumination shows how the V
OC
starts around 0.64 V and
eventually stabilizes at 0.36 V. (b) The J-V curve of a device with a neat
BDP2 donor layer, i.e. ITO/BDP2/C
60
/BCP//Al, giving a V
OC
of 0.34 V.
(Inset) The EQE of the neat BDP2 based OPV ……………………….…. 125

xii

Abstract
The field of organic photovoltaics has received much attention in recent years and has
made great strides toward market relevancy. The need for new, clean and environmentally
friendly sources of energy has been made clear with the latest reports on global climate change.
Solar-to-electric energy conversion is the ideal candidate to produce the large scale power
needed with little impact to the planet, promising terawatt energy scale production with zero
emissions. However, the development of solar energy is hamstrung with high production costs
leading to prohibitively high capital investment necessary for implementation. In this, organic
photovoltaics has an opportunity to make an impact in the market by providing a technology
with inherently lower materials costs due to the promise of carbon-based dyes that can be
manufactured with roll-to-roll technologies. Before this objective can be achieved however,
conversion efficiencies need to be taken to economical levels. Key in the optimization of these
devices will be the understanding of the links between materials properties and device
parameters.
It is the objective of this thesis to elucidate some relationships between materials
properties of organic chromophores and the device parameters in the resulting solar cells. Much
insight has been taken from the inorganic silicon solar cell industry in predicting how organic
photovoltaics work, but there are many fundamental differences due to the molecular nature of
the materials and these divergences are explored in this work. Different series of organic
materials were developed with systematically altered chemical structures allowing conclusions to
be drawn about, for example, how the resulting open circuit voltage of the device is impacted by
the changes under study. Film morphologies created by the packing of these molecules and the
xiii

impact in the resulting device will also be presented here. Finally, the efficacy of
supramolecular chromophore films will be contrasted with blended films of the same
chromophore units. It is the hope that through this study of structure property relationships that
the next generation in organic photovoltaic materials.

1

Chapter 1. Introduction to Organic Photovoltaics
1.1 Justification for Solar Research
Most energy used by humans originated at one point from our sun.
1
In their beginning,
humans harvested the suns energy the way all life does, harvesting wild plants and eating other
animals. That energy, captured and stored by plants, comes from the light of the sun converted
to chemical energy through the process of photosynthesis. The first great advancement by man
in energy technology was the domestication of fire. This achievement gave humans access to
large quantities of the suns energy, recently converted and stored in wood, which let them unlock
even more energy from the food that was cooked by it. Fire, used as heating, allowed humans to
spread and become the preeminent species on the planet. Through the course of history
mankind improved these energy sources through domesticating animals, breeding high efficiency
crops and concentrating the energy in wood into charcoal. The sources of the energy however
remained the same. The primary mechanical energy of the times came from the manual labor of
humans and animals which again was provided to them through plants and photosynthesis. Even
wax, primarily used for lighting, is energy collected by bees from plants that harvested solar
energy. During the industrial revolution the use of hydraulic and wind power first came into
widespread use. Though these sources of mechanical energy no longer relied on plants for the
energy they are an even more directly reliant on solar energy as both the rainwater that drives the
waterwheel and the wind that drives the windmill are powered by heat energy garnered from the
sun. Eventually wood gave way to coal for heat and to power the newly invented steam engines
heralding in the modern age of energy use. This started the change from using recently living
biomass to fossilized biomass as our main energy supply. Now great the majority of energy used
by modern civilization comes in the form of fossil fuels like oil, coal, and most recently natural
2

gas. Still, the energy stored in these fossil fuels is disguised solar energy as it was harvested by
ancient plants over the course of hundreds of millions of years and slowly converted into its
current form deep in the earth. This newfound store of concentrated energy only accelerated our
thirst for more and we rapidly developed new ways to this energy. The cycle is self sustaining
and drove energy technology even faster, feeding off of the huge supply of energy of fossil fuels.
Despite photosynthesis being a relatively inefficient process with most plants garnering 0.2-2%
effective energy conversation,
2
several billion years of life has allowed for a buildup of a great
repository of energy in these fossil fuel deposits. However vast they may be though, these
deposits are a finite resource that humans only began to exploit about 200 years ago and, using
the most generous estimates, will be used up in the next few hundred.
3
With a growing thirst for
energy
4, 5
and a dwindling supply of currently used fuel, there is a great need to achieve the next
breakthrough in energy technology and learn how to effectively harvest solar energy directly to
ensure the availability of energy for the future.
Around 90,000 TW of electromagnetic energy is estimated to radiate down on the earth
averaged over time and area. This translates to more energy radiating down on earth in on and a
half hours (480 EJ) than the total energy usage of the world in 2001 (430 EJ). With the demand
for power around 15 TW, less than 0.2% of the world’s surface needs to be covered by 10%
efficient solar conversion systems for the power generated to be about equal to the total power
demand for the planet. Even considering the most power demanding country the United States,
the area needed is still of about 1.9% of the size of the country which is on par with the area of
the public roads. Conversely, if the roof top of a single family home with 100 m
2
in usable roof
space was covered in 15% solar cells and an average solar resource of 4.5 kWh/m
2
/day the

3

power generated, 67.5 kWh/day would be well above the average electricity consumption in the
US.
6
Though this is a project of enormous scale, both the technology and production capacity to

Figure 1.1 Distribution of global solar insolation in the winter (top) and spring (bottom).
(Provided by NASA at http://earthobservatory.nasa.gov/) It can be seen that solar energy is
both vast and relatively evenly distributed over the populated globe. This is further proved by
the fact that more than half of the worlds solar energy production is done in Germany, a
country both small and at a northern clime.
4

do it already exist today; all that remains to do is to make the economics of solar energy make
fiscal sense. Especially when viewed in the long term, solar energy conversation one of the only
real options for large scale sustainable energy.
There are a several effective techniques currently used to harvest solar energy. Biomass
farms, which employ some organism to convert sunlight into chemical energy, are a method that
is being developed. Advantages to this strategy include relatively cheap upfront costs to
producing the fuel and the natural consequence of removing carbon dioxide from the air. Major
disadvantages include the need to develop new infrastructure to harness the new fuel and low
overall efficiencies that top out at a maximum of 10%. Solar thermal systems are another
strategy for harnessing the suns energy directly. These systems roughly fall into two categories,
low-temperature ones that use the collected heat directly for local heating cooling and
ventilation and high temperature systems that produce electricity. But the perhaps the most
elegant and efficient solution to harnessing the power of the sun is to make use of the
photoelectric effect which can transform solar irradiance directly into electricity.

1.2 Photovoltaics
The photovoltaic (PV) effect is the physical process that powers to solar cells. Known since
the mid 19
th
century,
7
it is defined as the onset of an electrical potential across a material under
illumination. A material that demonstrates the PV effect which is capped by two electrodes will
generate a current and is the definition of a solar photovoltaic. The materials that demonstrate
this effect can vary widely in composition and can be liquid and solids as well as intricate
systems of materials.
8
The manner in which the PV effect arises depends entirely on the
materials used but in all cases it is dependent on some kind of asymmetry in the device which
5

drives positive carriers in one direction and negative carriers the other way.
9, 10
This is what
differentiates the materials that demonstrate the PV effect from those that show
photoconductivity where carriers are generated but there is no preferential polarity in the
device.
11
Devices that produce current as a result of incident radiation are known as PV cells.
These devices are also differentiated from cells that use sunlight to produce chemical bonds as is
done in photosynthesis
12
and solar fuels.
13
As mentioned, the manner in which the voltage is
generated is dependent on the material or materials used to construct the photoactive region and
so it is the different types of materials that can be used that define the different classes of PV
cells. In devices made from a single photoactive material the asymmetry is introduced by the
selective doping of impurities, to create a junction. In one side of the junction electron accepting
dopants will be introduced to create a region of positively charged carriers, or p doped, and in
another electron donating dopants are introduce creating a negatively charged region or n
doping; this is referred to as a pn junction. Alternatively a junction can be made by layering two
different materials which possess different energy levels. The material with a higher Fermi level,
or energy level of the highest lying electrons, will donate electrons to the material with the lower
Fermi level until these levels equilibrate. This creates a junction with positively and negatively
charged sides that acts as an asymmetric element to drive the different charges to their respective
contacts.
14

1.2.1 Silicon Solar Cells
The current leader in the photovoltaic industry is the crystalline silicon solar cell (c-si) which
relies on doping to create the pn junction. This type of device has been studied for over 60 years
first being developed at bell labs in the early ‘50s.
15
The physics of this device structure is now
6

mostly understood and c-si solar cells of up to 25% can be made. This is very near the 29%
maximum efficiency predicted for c-si solar cells by the Shockley-Queisser limit, given silicon’s
band gap of 1.1 eV.
16
Silicon PV research, however, was not initially driven by the need for low
cost power but was instead given most of its funding to produce light weight power for extra
terrestrial vehicles; an area in which the cost was relatively immaterial. Consequently it is not
necessarily the ideal material for producing large scale cheap power capable of competing with
fossil fuels. One of the problems keeping silicon back is the fact that silicon’s conduction band
and its valance band are not aligned in k space, making it an indirect absorber. This means that
an electron excited in the conduction band does not have the same momentum as one into the
valance band. During an excitation, momentum needs to be conserved and so the electrons
cannot be directly excited, instead relying on the coupling with crystal lattice phonons to take
place. This results in silicon having weak oscillator strength at energies nearing its band gap. In
the wavelengths relevant in photovoltaics, this translates to absorption coefficients of 1E
3
to 1E
4

cm
-1
. Consequently, silicon devices need to have a photoactive region of around100 μm thick to
absorb a significant portion, >90%, of solar radiation. This then requires that one of the minority
carriers needs to travel that distance from the pn junction to the back contact without being
trapped and recombining. For a minority carrier to have a lifetime long enough to make this
journey an extremely high purity material, free of any trapping impurities, is needed.
17
These
very high purity requirements are the reason why single crystal silicon has such a high
production cost, even after decades of perfecting production procedures, which makes c-si
devices cost prohibitively high. Yet despite these drawbacks this type of device dominates the
market share
18
for PVs leaving a great opportunity to develop new technologies.
7

1.2.2 Organic Photovoltaics
Economic viability is the deciding factor to making a PV technology available for
widespread use as a carbon neutral energy source.
19
With projected practical power conversion
efficiencies for organic photovoltaics (OPVs) reaching 10-15%
20
and promise of cheaper, easier
to process materials that would allow photovoltaics to be a more competitive energy source,
much research has been done the last decade.
21, 22
The promise and history of the development
of OPVs can be seen in brief in figure 1.2. As shown in this graph the most efficient PV cells are
III-V GaAs and mulitjunction devices (purple traces). These types of solar cells though are high

Figure 1.2 The 2014 NREL graph on the improving efficiencies of various types of PVs.
(Provided by NREL at www.nrel.gov) Traditional a-Si devices (blue trace) can be seen to
plateau in recent decades having nearly reached its theoretical maximum. In the lower right it
can be seen the rapidly improving efficiencies of emerging PVs (red traces) which is
composed mostly of OPV style devices. Though multi-junction PV cell vastly outperform the
other technologies, inherently high production costs mean they will remain only suitable for
space applications.
8

cost with applications restricted to aerospace where weight, not cost is the primary consideration.
Crystalline silicon devices (blue traces), the longest studied type, have plateau-ed in efficiency
being near their maximum, and yet still remain too expensive to directly compete with fossil fuel
technologies without government incentives. Inorganic (green) and organic (red) thin film
technologies alone still have rising record efficiencies among the PV technologies. However
CIGS, a-si and CdTe thin film devices depend on many rare earth elements deposited in vacuum
chambers and both requirements are major challenges on the production scale needed.
23
OPV
technologies, on the other hand, promise the cheapest materials being reliant only on earth
abundant organic elements and can be processed by inkjet and roll to roll techniques and
therefore may be cost competitive with the conventional technologies in the near future.
In this class of PV cell, organic, carbon based, molecular materials are used. Among their
advantages, they are inherently stronger absorbers due to the fact that the energy transitions in
the visible wavelengths are usually allowed processes which can have absorption coefficients
around 1,000 times larger than that of silicon. This allows the devices to contain much thinner
layers of photoactive material, reducing production costs related both to the scale of production
and the purity requirement of the finished product (see figure 1.6). In addition, because of the
amorphous or polycrystalline nature of the films made from these materials they can be made
into flexible light weight devices that could potentially open up new markets for portable and
easy to install PV systems, which cannot be filled by current technologies. Many of the materials
used being studied for use in OPVs are organic dyes that have been know in the dye industry for
decades and most are base their strong absorptions on π to π* transitions of extended aromatic
rings. Figure 1.3 demonstrates some of the materials under study, focusing on the molecular
9

dyes used in this dissertation. In addition many polymeric species such as P3HT are also getting
much attention in development as new, cheap, OPV materials.
OPVs devices are structurally different than c-si cells and take one of two architectures
either being planar or bulk heterojuctions as shown in figure 1.4 (b). In either case the active
layer is composed of two materials; a donor and an acceptor. In addition this region may be
sandwiched between one or more buffer layers (not shown) and capped with a transparent anode
and a metal cathode. The energy structure of relevant states is diagramed in figure 1.4, where the
highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO)
of the donor are higher in energy than the respective orbitals on the acceptor. The energy gap

Figure 1.3 Common materials used small molecule OPVs are listed here. The materials are
divided into groups by function in the devices of donor, acceptor, and buffer materials. The
function of each type will be discussed in the following sections.
10

between the HOMO of the donor and LUMO of the acceptor is what is known as the interface
gap ( ΔE
DA
).

1.3 OPV Processes
The process by which OPVs generate current is fundamentally different than the manner in
which inorganic technologies work. It is useful in the reading of this thesis to have a basic
knowledge in the photo-electrochemical processes involved in the generation of current in an
OPV. The process can be generally distilled into six discrete photochemical or electrochemical
steps that transform a photon impinging on the active layer to an electron emerging from the
device electrodes and is depicted in figure 1.4.

Figure 1.4 (a) The electronic structure of modern OPVs are displayed in a rough cartoon. The
energy levels of the acceptor and donor are shown relative to each other which are required for
properly functioning OPV. The device architecture of common OPVs is displayed showing
both (b) lamellar or planar and (c) bulk heterojunction (BHJ) style devices with photo active
regions sandwiched between a transparent anode and a reflective metal cathode.
11

Figure 1.5 The diagram shows of the 5 steps that an OPV undergoes during the
generation of photocurrent. Each process will be discussed in detail in the following
sections.
12

1.3.1 Absorption
1.3.1.1 Excitonic materials
In semiconductors, a photon with energy larger than the band gap causes an absorption event
where an electron is promoted from the valence band to the conduction band. The resulting
excited electron can “see” the coulombic potential of the positively charged hole it just vacated,
but due to the high dielectric constants and the covalent nature of their bonding of these materials
the two charges are heavily screened from each other. The results is that the radius of the
excitation is spread out over the distance of several lattice sites creating what is referred to as a
Wannier exciton. Because of this, the binding energy of the exciton is relatively low and on the
order of a few kT at room temperature. This means that the exciton will quickly dissociate and
generate free charge carriers.
In OPV materials strong covalent bonding only exists intra-molecularly with much weaker
vand er Waal’s interactions between molecules. As a result, the excitation is very localized,
usually to a single molecule and only extending to a couple of molecules at most. Because of the
low dielectric constant in organic molecules, there is no electronic screening between the
electron and the hole. They therefore become coulombically bound and form a quasi particle
called a Frenkle exciton. As these are strong interactions that range from 0.3 to 1.0 eV lower in
energy than the free charge carriers. There is not enough energy available at room temperature
to supply this much energy and separate these two particles, this means that the exciton cannot
dissociate into free charge. As a consequence, the photoactive layer of OPV devices need to be
composed of two different materials, referred to as a donor and acceptor material. It is at this
interface that the excitons will be separated into free charges. This will be discussed in length at
in a later section
13

1.3.1.2 Nature of transitions in organic materials
In the organic molecular materials studied here, one of the primary modes of absorption
is done by an exciting an electron from the S
0
to S
1
state, though other processes will take part
including charge transfer (CT) excitations and direct S
0
to T
1
transitions. S
0
to S
1
transitions in
molecular solids typically are thought of to be transitions between the HOMO to the LUMO,
which, in the case of the types of dyes used in OPVs that have large pi bonding systems, means a
π to π*. By design these types of transitions have large amounts of wave function overlap
between the HOMO and LUMO levels which results in having transitions with very high
oscillator strengths. This very strong adsorptivity allows the majority of the light to be absorbed
with only a few hundred nanometers of material allowing for the thin flexible devices that give
OPVs their promise. However, due to these materials being molecular in nature, their
transitions, relying on these discrete orbital to orbital transitions, tend to be narrow and often

Figure 1.6 (a) The Jablonski diagram here depicts some of the common excited states and relaxation
pathways involved in the evolution of energy on organic molecules employed in OPVs. (b) The generic
electron configurations of the states involved are shown.
14

preserve their vibronic features. Consequently, the transitions do not cover the entire solar
spectrum. In contrast, semiconductor materials absorb, to some degree, all electromagnetic
radiation above the band gap due to the dense nature of the states in the electronic bands. This
is why designing the right material or system of materials to use in the photoactive region of
OPVs is much more delicate and intricate.
1
2
3
4
5
1
2
3
4
5
6
400 500 600 700 800
0
1
2
3
4
5
6

Absorbed Photons
(counts/cm
2
s)
Wavelength nm
c)

 X 10
14
(cm
-1
)
b)

x10
14
(counts/cm
2
s)
a)

Figure 1.7 (a) The incident solar flux in photon flux per wavelength. (b) The
absorbtivities of silicon (black trace) and PDI/C
60
film (red trace) per centimeter,
showing the magnitude difference. (c) The traces the absorbed photons from the solar
flux for a 300 nm film of each material and shows the inherent advantage of organic
materials.
15

In figure 1.7 (a) the spectrum of the sun is plotted as describe by ASTM G173-03 AM
1.5G spectral irradiance in counts as a function of wavelength. In section (b) of the same figure
the adsorptivity of c-Si is plotted against the active region of the standard OPV device C
60
/CuPc.
This plot allows one to graphically see how much more OPV materials can absorb over the
standard PV material c-Si. In section (c) the number of absorbed photons per centimeter is
displayed. From this plot it is easy to visualize the inherent advantage of using organic absorbers
over traditional silicon devices.

1.3.2 Exciton Diffusion
1.3.2.1 Diffusion Length
The next step in the process to generating current is referred to as exciton diffusion. In
this step the newly formed exciton, generated at the point of absorption, needs to move toward
the donor acceptor (D/A) interface. This is because, as discussed earlier, the exciton cannot
dissociate in the bulk material due to the high binding energies inherent in molecular absorbers.
As excitons are not charged, the only force that drives them toward the D/A interface in the
concentration gradient that is set up by their generation in the bulk and quenching at the
interface. The distance that excitons can travel in a device before the exciton population decays
to 1/e, or roughly one third remaining, is the exciton diffusion length (L
D
) and is defined by
equation 1.1.
24

1.1
Here τ
0
is the lifetime of the excited state and D is the diffusivity. In an ideal device the number
of excitons reaching the interface over the number of absorbed photons will be near unity. This
can be difficult achieve because of the short exciton diffusion lengths that exist in molecular
16

organics. A situation can therefore arise that when a film is made thick enough to collect >90%
of the incoming photons the majority of the resulting excitons will be stranded inside the layer to
far from the D/A interface to reach it in their lifetimes. In an attempt to solve this problem much
research has gone into developing bulk-heterojunctions which are designed to have intimately
intermix the donor and acceptor materials in domains on the order of the L
D
. However if a
planar structure is to be made to work, the way to make efficient devices is to design materials
that have sufficiently long L
D
’s so that the diffusion length is on the order of the thickness
needed to absorb the majority of the incident photons. To do this one must understand the nature
of the exciton transport, or energy transport mechanism, which takes place in the material system
under study. One way to do this is to reduce the decay rate of the exciton, thereby increasing its
lifetime. A strategy used in c-Si devices is to use extremely high purity material which reduces
the non-radiative decay rate. Another way to extend the lifetime is to use a triplet exciton which,
due its spin forbidden nature, has long radiative lifetimes. Increasing the diffusivity is the other
way to enhance the L
D
. Doing this is a material specific process which depends on the type of
exciton transfer that dominates.

1.3.2.2 Energy transfer mechanisms
The two primary methods of energy transfer by which excitons migrate are dipole-dipole
interactions,
25
and electron coupling transfer.
26
Dipole-dipole interaction, also termed Forster
resonance energy transfer (FRET), allows for energy transfer by the coupling of the dipolar
electric fields between and excited donor (d*) and an acceptor (a). This mechanism was
described by Forster
27
using a classical model for the electronic interaction between two
transition electric dipole moments using equation 1.2.
17

1.2

1.3
In this equation it can be seen that the important quantities that lead to fast Forster energy
transfer (K
ET
) are μ
d
and μ
a
, the size of the two interacting transition dipole moments, and the
distance between them (R
da
). The dipole moments here are for the transitions D* → D and A →
A* respectively. Because of this mechanism, it is a “through space” transfer than doesn’t need
direct contact and can proceed over longer distances. Equation 1.3 expresses this rate explicitly
with measurable quantities where n is the index of refraction τ is the radiative lifetime of the
excited state donor, r is the intermolecular distance, j is the spectral overlap integral, and K is a
dipole-dipole orientation factor. The efficiency ( η) of this FRET is described in equation 1.4
where K
nr
and K
r
are the radiative and non-radiative decay rates.
28

1.4
This mechanism is the primary way in which singlet excitons migrate because of their fast
radiative rates which give rise to fast K
FRET
. These fast radiative rates are due to the fact that the
process is spin allowed. Also importantly singlet state transitions present in the π extended dyes
studied here have large oscillator strengths, leading to a large donor acceptor overlap integrals
(j).
The second method for exciton transport proceeds by electron exchange also referred to
as Dexter energy transfer (DET). This exchange is directly related to the overlap of the excited
18

state of the donor and the ground state of the acceptor. The rate of energy transfer is expected to
fall of exponentially with distance as the wave function overlaps decrease.

) 1.5
Where K is related to the specific orbital overlap interactions, J is the normalized spectral
overlap, R
DA
is the separation between D* and A, and

is the separation when the molecules
are in Van de Waals contact. This mechanism is much more difficult to relate to experimentally
observable parameters and so is still difficult quantify. Though both singlets and triplet excitons
can transfer by this mechanism, triplets primarily use this method for diffusion because it does
not rely on radiative rates, like FRET, which is very slow in triplet states. As this mechanism
only can transfer excitons to neighboring molecules it is seen as an inherently slower transport
path. However, because of the increased lifetimes of triplet states, which can easily last for
milliseconds, the diffusion lengths of triplets can be very long.

Figure 1.8 The difference between Forster transfer (FRET) and Dexter transfer (DET) is
diagramed. In FRET the electron only changes energy not the molecule and the energy is
transferred through a virtual photon to another molecule in a dipole-dipole interaction. In
DET the excited electron gets transferred to another molecule with its extra energy as a
relaxed electron is simultaneously transferred back to the first chromophore.
19

Figure 1.8 demonstrates the difference between FRET and DET. FRET can be seen as
analogous with “trivial” energy transfer where the energy donor emits a photon, either through
phosphorescence or fluorescence, and an energy acceptor absorbs it. Though in FRET no photon
is ever emitted and the transfer is immediate, the process involves all of the same parameters and
the energy is transferred from one electron to another while the electrons themselves stay on
their respective molecules. In this example it can be easily seen that phosphorescence would be
an inefficient mechanism for transfer due to the very long radiative rates. In DET the electrons
retain their energy levels and are transferred from the energy donor to acceptor.

1.3.3 Charge Transfer
Following the migration of an exciton to the D/A interface efficient charge transfer
quenching of the excited state is required for photocurrent generation. In this step an excited
electron will be transferred from a donor molecule at the interface to an adjacent acceptor
molecule’s LUMO, leaving a hole behind in the donors HOMO. This process creates a
hole-electron pair across the interface that, because of the small amount of screening in low
dielectric constant organic solids, is coulombically bound. This coulombically bound
hole-electron pair is referred to as a charge transfer (CT) state.

1.3.3.1 Thermodynamic requirements for CT
In principle the thermodynamic requirements of this process are straightforward; the
energy of the CT state, as generally defined as the energy gap between the HOMO of the donor
and the LUMO of the acceptor, must be lower in energy than the exciton provided. The
thermodynamic driving force can be quantified by ΔH= -E
i
+E
a
-E
0
where E
i
and E
a
are the
20

electron ionization potential of the donor and electron affinity of the acceptor, respectively, and
E
0
is the exciton energy. Defining the energy of these levels is still a matter of ongoing research
and debate
29, 30
but the most direct measurements are inverse photoelectron spectroscopy (IPES)
for LUMOs and ultraviolet photoelectron spectroscopy (UPS) for HOMOs. However, these
measurements require expensive and difficult to operate machinery that is still plagued by
experimental error due to sample degradation with highly variable results. More conventionally,
these levels are approximated by measuring the solution oxidation and reduction potentials
against an internal standard and using the right correlations
31, 32
to convert them to orbital
energies. While these measurements are reproducible and easy to perform, the measurement is
not a direct measurement of the energy levels. While only a slight enthalpic driving force is
necessary for CT state formation, a larger driving force is required to achieve a fast enough rate
efficient photocurrent generation.

1.3.3.2 Kinetics of CT generation
Marcus’ theory of semi-classical non-adiabatic electron transfer is a model used to
characterize many electron transfer reactions and has seen much use in describing the formation
of CT states in OPVs.
33
This theory considers the reactant and product as a single potential
energy surface modeled as a harmonic oscillator with the horizontal axis being a reaction
coordinate that includes all nuclear motion of the system. The CT state has its own potential
energy surface and the theory dictates that the electron transfer can only happen at the
intersection of the two curves to properly satisfy both the conservation of energy and the
21

Frank-Condon principle which states that the electron transfer happens on a time scale much too
fast for any movement of heavy atoms (See figure 1.9). This point of intersection then represents
the conformation the system needs to achieve, through thermal motion, for the iso-energetic

Figure 1.9 A generalized charge transfer diagram as envisaged by Marcus’ theory where each
state is shown as a harmonic oscillator. Here Q is an aggregated reaction coordinate that
encompasses all relevant parameters (mostly molecular position, motion and deformation) and E
is state energy. An absorption event will take the system from D/A curve to D*/A curve on a
vertical path. D*/A will then need to gain the activation energy

to move to the crossing
point of D*/A and D
+
/A
-
.

is the driving force of the CT process and λ is the energy
difference between the exciton and the CT state at the CT states ground state conformation.
22

formation of the CT state and so describes the activation energy of the reaction. The rate
constant for such an electron transfer k
ET
, can then be described in terms of a Fermi’s Golden
rule type analysis (EQ 1.6)

1.6
In this equation V refers to electronic coupling between the reactant and the product
states and depends on the wave function overlap of the donor and acceptor. The exponential
term in Eq. 1.6 is the Frank-Condon factor and predicts that as

increases becomes more
negative, more driving force, the rate will increase. When

the electron transfer has
no activation barrier and as the magnitude of

continues to increase the rate begins to
decrease and is described as the Marcus inverted region. While this theory has seen much use
in describing OPVs it has several deficiencies. One shortcoming is that it does not consider
aggregation effects in the solid state, only treating the highest lying orbitals as discrete
surfaces. In addition, Marcus inverted regions have not been observed in these systems,
though this may be due to the energetic requirements of the system restricting it to the
classical region.

1.3.4 Charge Separation
Once formed, the CT state can undergo one of two decay pathways. Unproductively, the
hole and electron can recombine into some neutral species; either returning to the excited state it
came from, or to the ground state, dissipating the energy as thermal motion, in a process referred
to as geminate recombination. Alternatively, the hole and electron can separate, breaking the
coulombic binding forces, and becoming free charges, also referred to as a charge separated CS
23

state. Assuming that CT state formation is near unity, which is true for appropriately chosen
material pairs, the relative rates of these two processes determines the overall efficiency of
charge generation. The first qualitative description of CT state behavior was published by
Onsager in a theory now named after him.
34
He proposed a model that calculates the probability
that a coulombically bound CT state in a weak electrolyte which undergoes Brownian motion
will charge separate, generating free charges, or recombine. The model supposes that absorption
of a photon results in a localized hole and a “hot” electron, with excess thermal energy, which
will go through rapid random motion thermalizing at some distance from the hole. This is
diagramed in figure 1.10 where a electron is excited and moved away from the hole by some
distance a, the thermalization length.
35

The competition between full ionization and recombination is then dependant on the
coulombic potential felt by this CT state. Onsager proposed that this was defined by a coulomb
capture radius, or Onsager radius, r
c
, which is the distance at which the coulomb attraction
energy equals the thermal energy k
B
T, described in equation 1.7.

1.7
Where the e is the elementary charge,

is the dielectric constant of the medium,

is the
permittivity of vacuum,

is the Boltzmann’s constant, and T is temperature. Now if the
thermalization length a is greater than the Onsager radius the carriers are considered to be fully
dissociated. If it is shorted then the CT state will dissociate into free charges with an escape
probability P(E) or recombine with the probability of 1-P(E) that is dependent on the electric
24

field felt by the CT state. Equation then shows how materials with large dielectric constants
(

like silicon will directly generate free charges from an absorption event, while
organics with lower dielectric constants (

will have to form CT states will some chance
of generating free carriers.

1.8
Equation 1.8 shows the escape probability of a CT state in a homogenous material with an
applied electric field of E.
36
This relationship holds for small applied fields having a linear
relationship which is independent of a.

Figure 1.10 A potential energy diagram that visualizes Onsager Theory for auto-ionization.
The absorption event frees an electron to the Onsager radius, a, forming a CT state. This
electron then needs to gain energy necessary to get beyond KBT to dissociate into free
charges before it recombines. Adapted from reference 35.
25

For typical organics used in OPVs P(E) is close to zero in a neat material again
displaying the need for a D/A interface for exciton dissociation. Onsager theory has successfully
been adapted to multiple systems and many people have attempted to do so for organic D/A
interfaces but the area is still the source of much debate. The importance of D/A interface was
highlighted in a dipolar layer model suggested by Arkhipov et al.
37
This model proposed that
formation of a CT state at the interface induced the formation of interfacial partial dipoles which
raises effective dielectric constant of the medium and allowing for escape similar to inorganic
semiconductors. This model required that the polymer chains were aligned parallel to the
interface with a ordered array of acceptor molecules, implying that both molecular orientation
and relative energetic are required for efficient charge separation. Peumans and Forrest
38

developed a model to investigate D/A interface of organic systems by using a kinetic Monte
Carlo system to Onsager theory. They said that the Arkhipov model had too many restrictions,
suggesting instead that the presence of interfacial CT state dissociated because of the
confinement of the available volume for geminate recombination to the interface. Their model
showed that if electron mobility exceeded hole mobility by more than a factor of 100, the escape
probability increased due to the larger space it could sample prior in the lifetime of the CT state.
This result correlates to the findings of another study where it was found that the presence of
traps, immobilizing one of the charges of the CT pair enhanced dissociation probability.
39

Further studies have shown that the electric field at the D/A interface, generated by presence of
permanent dipoles formed at the interface in a process analogous to band bending in inorganic
semiconductors, is centrally important in charge separation.
40
This induced electric field is not
well characterized by current theory due in part to a lack of knowledge about the structural
nature of these interfaces. Further controversy has also been thrown into the discourse with
26

finding which indicate that in rr P3HT:PCBM solar cells exciton quenching results in the direct
formation of free charges with no CT state formation.
41
This may imply that the mechanism by
which free charges are form is material system dependent.

1.3.4.1 Overall CS mechanism
Regardless of the exact mechanism the overall process follows the scheme shown in
figure 1.11. Here S
1
an exciton created by and absorption event will charge Transfer to a hot
exciton (k
CT
). This may directly lead to free charges (k
CS*
) or thermalize to a ground CT state

Figure 1.11 shows a general energy level diagram that summarizes the CT and CS processes.
The exciton in the blue levels gets generated and forms a charge transfers (k
CT
). This may
directly lead to free charges (k
CS*
) or generate a hot CT state which thermalizes (

). This
state can either decay to the ground state (k
triplet
, k
GR
) or charge separate (k
CS
). One caveat to
this figure is that the triplet state T1 may be incompetent for CT (as shown here) or high
enough in energy to also CT. Taken from reference 35.
27

(

). This CT state can either decay by transferring to a T
1
exciton (k
triplet
), or vibrationaly
return to the ground state (k
GR
), or charge separate (k
CS
) into free charges. One caveat to this
figure is that the triplet state T
1
may be incompetent for CT (as shown here) or high enough in
energy to also CT. The free carriers can then return to form a CT state or leave the interface and
thermalize to the energy of the bulk charge carriers. The figure also displays how these
processes will all be exothermic in nature as

and

must be negative. This diagram also
brings up one last point that charge separation is entropically driven as there are more free carrier
states than there are bound states.

1.3.5 Charge Transport
The next step required for the generation of photocurrent is to transport the newly formed
free charges from the D/A interface where they were created to the contacts of the device. These
carriers will be at an energy level equal to the quasi Fermi level of the material and can be
approximated by the E
a
for the donor and E
i
for the acceptor. Just as the nature of molecular
organic materials restricts the exciton in size, so too does it localize the electrons and holes to
single molecules. The electron and hole can then be thought of as radical anions and cations,
resulting from the reduction or oxidation of a single molecule in the solid that then polarizes the
surrounding medium resulting in a quasi-particle referred to as a polaron. As there is no band
structure in the organic solids as there is in inorganic semiconductors, the polaron will then
transfer from one molecule to another in a hopping mechanism that can be described by Marcus`
theory, as described above.
42, 43
This case differs only in that the two electron donor and
acceptor molecules are the same and so the two state curves have the same energy. The state
energy curve of D
+
/D will then cross D/D
+
with some activation energy but only the internal
28

built in bias pushing the electron toward the contacts will provide the driving force (

). The
rate of conductivity in these systems is then directly affected by the reorganization energies the
polaron has on the molecule and the motion of the nearby polarized molecules involved in these
polaron transfer reactions.
Additionally, the rate of polaron hopping can be decreased by the presence of trap states.
These traps would be shown on a Marcus’ diagram lowering the D
+
/D state curve in energy.
These trap states can be a consequence of impurities in the conducting material or result from the
in-homogeneity of the amorphous material. These shallow trapping states will be present in pure
phases or on the edges of small crystalline domains, slowing the hopping rate dependent on the
magnitude of trapping relative to kT, sometimes making an otherwise promising material to be
unusable. As a consequence the mobility will be dependent on the morphology of the films,
which can range from 10
-6
to 10
-3
cm
2
V
-1
s
-1
in these amorphous films. For this reason the use of
crystalline materials for OPVs is under investigation to improve performance.
44

1.3.6 Charge Collection
Once the polarons have reached the electrode, charge needs to be transferred to the contact
and collected by the external circuit to produce photocurrent. The process of removing the
polaron from the device is called charge injection and is mediated by the organic inorganic
interface properties. For efficient charge extraction to occur the contact needs to be Ohmic,
meaning that no depletion region can form at that interface. This is achieved by correctly
selecting the electrode material so that no charge transfer happens at steady state to prevent any
diode from forming in a Schottky contact, which would drive charges away from the interface.
These materials need to be chosen such that the Fermi level of the material matches that of the
29

organic it is in contact with. In the OPVs studied here in this paper, the contacts are a
transparent anode composed of indium tin oxide (ITO) and a low work function metal such as
aluminum (al) or silver (ag) for the cathode. Even when the system meets the thermodynamic
requirements for charge collection there is evidence to suggest that electron transfer kinetics can
play an important role in determining the charge extraction efficiency.
45

One of the biggest challenges still remaining in this step is not the collection itself but
preventing the interface from quenching excitons. To this end, non-photoactive layers, called
buffer layers, are inserted before the electrode materials which have the correct transport
energies but high enough excitation energies that they prevent excitons from reaching these
quenching interfaces. In many OPVs and the ones discussed here, BCP is used as the buffer
layer between the acceptor and the metal contact. This serves well in its roll to stop exciton
quenching, but may not be the best candidate due to a relatively high resistance. As a second
purpose this buffer layer has been shown to protect the photoactive material from any deleterious
effect that could occur from the deposition of hot metal atoms to the surface. Several materials,
including MoO
3
and Poly(3,4-ethylenedioxythiophene) Polystyrene sulfonate (PEDOT:PSS) are
being used as anode side buffer layer. These and more are still an area of active research.
30

1.4 Characterization of OPV devices
In the dark, solar cells act as diodes which preferentially allow current to flow in one
direction through asymmetric conductivity. A diode has four distinct behaviors at different
applied voltages. At extremely negative potentials, beyond peak inverse voltage (PIV), the
device breaks down and allows free flow of electrons, usually resulting in permanent damage of
the device. At potentials slightly more positive than the PIV, called reverse bias, the flow of
current is resisted allowing only a very small reverse saturation current to flow which is
generated in most OPVs by thermal excitations at the D/A interface.
46
As the potential is raised,

Figure 1.12 A typical I-V curve for a CuPc/C
60
device with several of the most important
parameters diagramed. J
SC
is the Y intercept and is the current density at zero applied
voltage. V
OC
is the X intercept and is the Voltage needed to stop all current flow. The red
box of area P
max
is the power generated when operated at the maximum power point (MPP).
FF is the fill factor and represents the “square-ness” of the curve and is calculated by the
formula

31

in this region, the current stays constant or changes little until the potential reaches the knee
voltage, or rectification potential, where at this point the current begins to increase exponentially.
Finally, at very high voltages the current asymptotically approaches a linear increase
approaching the bulk resistance of the materials.
Once light is cast on the device an electric potential is developed across the contacts and a
photocurrent begins to flow. The behavior of the cell remains the generally the same with the
addition of a constant offset current called the photocurrent. In ideally operating PVs this
photocurrent is independent of applied voltage. The power produce by a solar cell at any applied
voltage can be calculated by determining the product of the current and voltage on that point on
the curve. When the voltage applied to the cell is positive but below the rectification voltage,
this photocurrent flows against the applied voltage and power is produced.
A typical current voltage (J-V) curve is shown in figure 1.12. The device response in the
absence of incident illumination, or dark curve (blue), shows the diodic response with a knee
voltage at around 0.4 V. The device response under illumination, or light curve (red), is plotted
next to it with a nearly constant offset due to the induced photocurrent. Some non-ideal behavior
in the device can be seen as the light curve is not flat in reverse bias, implying some voltage
dependence on the photocurrent. From this plot three important values are reported that sum up
the performance of the device adequately: short circuit current, open circuit voltage, and fill
factor. Short circuit current (J
SC
) is the taken from the Y intercept on the light curve and
represents the current that flows through the cell under illumination while zero voltage is applied
to the solar cell. The open circuit voltage (V
OC
) is the X intercept of the red curve and represents
the potential that needs to be applied to produce zero current in the solar cell. Finally fill factor
(FF) is the parameter that represents the “square-ness” of the curve. This is important because,
32

with a given Jsc and Voc, the power produce from a cell is maximized by having a square curve.
It is calculated by first determining the maximum power point (MPP) which is the point on the
light curve that has the largest negative product of current and voltage. This represents the
operating voltage at which the PV produces the most power. Dividing the MPP by the product
of J
SC
and V
OC
will produce the FF, a number between 0 and 1 with 1 being the “squarest” and
best curve possible. With these parameters it is possible to calculate the power output (P
max
) and
efficiency ( η) as shown in equations 1.9 and 1.10

1.9

1.10
Here P
in
is the incident power from the sunlight impinging on the PV. This incident power is
taken from the ASTM G-173-03 AM 1.5 solar spectrum, which is agreed upon as the test
conditions for standardization of solar cells by the National Renewable Energy Laboratory
(NREL) and American Society for Testing and Materials (ASTM). This spectrum corresponds
to the spectrum one would measure in a clear sky when the sun is passing through 1.5 air mass,
or at the angel of 48.2°.
47

1.4.1 Mismatch Factor Corrections and Proper Testing Conditions
As stated above the performance of PV cells is reported against international standards,
which is most commonly ASTM G-173-03 AM 1.5 solar spectrum in the United States. It is,
however, nearly impossible to test under this spectrum as there is not lamp that perfectly emits
this spectrum and in nature only occurs a few minutes a day even in perfect weather. Therefore,
33

there were methods developed to predict what the performance would be under these conditions,
by testing under different illumination conditions and calibrating with a known standard.
48

The data needs to be corrected for spectral differences and this is done by first calculating the
spectral mismatch factor (M) with equation 1.11.

1.11
Where S
R
is the reference cell responsivity, S
T
is the responsivity of the cell being tested,
and E
S
is the spectrum of the lamp being used to test the cell. The reference cell is selected so
that the mismatch factor’s deviation from unity is minimized. The integrals here require only the
normalized spectrums. E
Ref
is the standard ASTM G-173-03 AM 1.5 AM1.5 spectrum, the
reference cells are sent to NREL where S
R
is measured, and E
S
is measured with a calibrated
photo detector. S
T
is the only unknown and it is measured by calculating the quantum efficiency
as a function of wavelength and relating it with the equation 1.12 where q, h, and c are the
fundamental charge, planks constant and the speed of light respectively.

1.12
Once M is calculated it can be used to correct the measured currents in the test cell under
the source spectrum (I
T,S
) to what the current would be under the reference spectrum (I
T,R
) by
using equation 1.13. Here I
R,R
is the reference cell under the reference spectrum and I
R,S
is the
reference cell under the source spectrum.

1.13

1.14
34

By setting the lamp intensity so that the reference cell produces the same J
SC
that it did
when it was calibrated under E
Ref
. As we have set the source spectrums intensity so that I
R,S
is
equal to I
R,R
the equation simplifies further to equation 1.14. This equation shows how, now
knowing the mismatch factor, the current can be corrected to the currents under the standard
conditions.

1.4.2 Electrical Parameters and the Generalized Shockley Equation

To analyze the diode an equivalent model circuit must be built (figure 1.13) which
simplifies the device by reducing it to a series of linear, passive elements but which retains all of
the electrical characteristics of the circuit. In this figure the circuit is distilled down to four basic
elements in a PV. In this diagram V is the applied potential across the device, J is the total

Figure 1.13 The equivalent circuit diagram of a solar cell is diagramed. J
ph
is the
photocurrent generated by the photoactive region under illumination. The second element
in parallel with that represents the diodic nature of the device that allows the flow of current
opposite the photocurrent beyond a specific applied voltage. R
P
is the shunt resistance of
the devices that allows current to leak past the diodic element at voltages lower than the
knee potential. R
S
is the resistance in series with the diode that represents the bulk
resistance of the device and dominates the current at high forward biases.
35

current density that flows through the cell, J
ph
is the photo induced current driven against the
applied voltage, R
p
is the parallel resistance and R
s
is the series resistance. To fully characterize
the diodic properties of a solar cell it is necessary to employ a equation like the generalized
Shockley equation (eq. 1.11).
49
This equation was developed to describe the electrical
performance of a p-n junction of a classical inorganic PV and gives the current that flows
through the cell as a function of applied voltage.

1.11

1.12
The total current that flows through the device can then be described in the generalized
Shockley equation 1.11. In this equation k is Boltzmann’s constant, q is the fundamental charge,
and T is temperature. J
S
is the diode’s saturation current density which describes how much
current flows in reverse bias in the dark, n is the diode ideality factor, which describes how the
saturation current is generated and is equal to two (2) when carriers are thermally promoted. In a
derivation that will be explained in chapter 2, equation 1.11 can be simplified to equation 1.12
when considering an OPV in the dark at small applied currents. Equation 1.12 can then be used
by fitting it to the dark curve by adjusting the parameters n, J
S
and R
S
. If the ideality factor, n, is
close to the assumed value of 2, then the model can be considered valid and the J
s
and R
S
can be
considered good approximations of the dark saturation current and series resistance, respectively.
These parameters can then be used to make assumptions about the nature of the performance of
the devices and its relationship to the materials properties of the molecules used.

36

1.4.3 Overview of Work
Despite much progress over the last decade in advancing the science of organic
photovoltaics, basic research is still required due to the many challenges that remain. Though
efficiencies are improving for the record devices, a systematic understanding of the structure
property relationships for OPV materials still remains elusive. This work will focus on
attempting to relate molecular materials properties and film morphologies with device operating
parameters. In chapter 2 and 3 we will investigate the materials properties that impact the open
circuit voltage in OPVs and how to predict it in an a priori way. In Chapter 4 the morphological
nature of the D/A region in laminar bilayer OPVs is investigated with neutron reflectivity. The
long used model of discrete layers will be challenged and the modes and reasons behind the true
structure of these devices will be investigated. Finally, in Chapter 5 the multi-chromophore
array BODIPY-PtTBP will be used in devices and compared to a blended system of the same
components to determine the advantages, or lack thereof, of the covalent linkage in these
supra-molecules will be discussed.

37

1.5 Chapter 1 References

1. Crosby, A. W., Children of the Sun: A history of Humanity's Unappeasable Appetite for
energy. W. W. Norton & Company Ltd.: New York, 2006.
2. Stryer, L., Biochemistry. W.H.Freeman & Co Ltd: Toledo, OH, 1981.
3. Annual Energy Outlook 2014. In Administration, U. S. E. I., Ed. 2014.
4. Hoffert, M. I.; Caldeira, K.; Jain, A. K.; Haites, E. F.; Harvey, L. D. D.; Potter, S. D.;
Schlesinger, M. E.; Schneider, S. H.; Watts, R. G.; Wigley, T. M. L.; Wuebbles, D. J., Energy
implications of future stabilization of atmospheric CO2 content. Nature 1998, 395, (6705), 881-
884.
5. Hoffert, M. I.; Caldeira, K.; Benford, G.; Criswell, D. R.; Green, C.; Herzog, H.; Jain, A.
K.; Kheshgi, H. S.; Lackner, K. S.; Lewis, J. S.; Lightfoot, H. D.; Manheimer, W.; Mankins, J.
C.; Mauel, M. E.; Perkins, L. J.; Schlesinger, M. E.; Volk, T.; Wigley, T. M. L., Advanced
Technology Paths to Global Climate Stability: Energy for a Greenhouse Planet. Science 2002,
298, (5595), 981-987.
6. In Office of Energy Statistics: Washington, DC, 2014.
7. Becuerel, E., "Mémoire sur les effets électriques produits sous l'influence des rayons
solaires". Comptes Rendus 1839, 9, 561–567.
8. Copeland, A. W.; Black, O. D.; Garrett, A. B., The Photovoltaic Effect. Chemical
Reviews 1942, 31, (1), 177-226.
9. O'Regan, B.; Gratzel, M., A low-cost, high-efficiency solar cell based on dye-sensitized
colloidal TiO2 films. Nature 1991, 353, (6346), 737-740.
10. Tang, C. W., Two ‐layer organic photovoltaic cell. Applied Physics Letters 1986, 48, (2),
183-185.
11. Petritz, R. L., Theory of Photoconductivity in Semiconductor Films. Physical Review
1956, 104, (6), 1508-1516.
12. Wasielewski, M. R., Photoinduced electron transfer in supramolecular systems for
artificial photosynthesis. Chemical Reviews 1992, 92, (3), 435-461.
13. Gray, H. B., Powering the planet with solar fuel. Nat Chem 2009, 1, (1), 7-7.
14. Neamen, D. A., Semiconductor Physics and Devices, 3rd ed. McGraw-Hill: New York,
2003.
38

15. Chapin, D. M.; Fuller, C. S.; Pearson, G. L., A New Silicon p ‐n Junction Photocell for
Converting Solar Radiation into Electrical Power. Journal of Applied Physics 1954, 25, (5), 676-
677.
16. Queisser, H. J., Detailed balance limit for solar cell efficiency. Materials Science and
Engineering: B 2009, 159–160, (0), 322-328.
17. Green, M. A., Silicon solar cells: evolution, high-efficiency design and efficiency
enhancements. Semiconductor Science and Technology 1993, 8, (1), 1.
18. Goetzberger, A.; Hebling, C.; Schock, H.-W., Photovoltaic materials, history, status and
outlook. Materials Science and Engineering: R: Reports 2003, 40, (1), 1-46.
19. Lewis, N. S.; Nocera, D. G., Powering the planet: Chemical challenges in solar energy
utilization. Proceedings of the National Academy of Sciences 2006, 103, (43), 15729-15735.
20. 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. Physical
Review B 2007, 75, (11), 115327.
21. Spanggaard, H.; Krebs, F. C., A brief history of the development of organic and
polymeric photovoltaics. Solar Energy Materials and Solar Cells 2004, 83, (2–3), 125-146.
22. Krebs, F. C.; Gevorgyan, S. A.; Gholamkhass, B.; Holdcroft, S.; Schlenker, C.;
Thompson, M. E.; Thompson, B. C.; Olson, D.; Ginley, D. S.; Shaheen, S. E.; Alshareef, H. N.;
Murphy, J. W.; Youngblood, W. J.; Heston, N. C.; Reynolds, J. R.; Jia, S.; Laird, D.; Tuladhar,
S. M.; Dane, J. G. A.; Atienzar, P.; Nelson, J.; Kroon, J. M.; Wienk, M. M.; Janssen, R. A. J.;
Tvingstedt, K.; Zhang, F.; Andersson, M.; Inganäs, O.; Lira-Cantu, M.; de Bettignies, R.;
Guillerez, S.; Aernouts, T.; Cheyns, D.; Lutsen, L.; Zimmermann, B.; Würfel, U.; Niggemann,
M.; Schleiermacher, H.-F.; Liska, P.; Grätzel, M.; Lianos, P.; Katz, E. A.; Lohwasser, W.;
Jannon, B., A round robin study of flexible large-area roll-to-roll processed polymer solar cell
modules. Solar Energy Materials and Solar Cells 2009, 93, (11), 1968-1977.
23. Schlenker, C. W.; Thompson, M. E., The molecular nature of photovoltage losses in
organic solar cells. Chemical Communications 2011, 47, (13), 3702-3716.
24. Schlenker, C.; Thompson, M., Current Challenges in Organic Photovoltaic Solar Energy
Conversion. In Unimolecular and Supramolecular Electronics I, Metzger, R. M., Ed. Springer
Berlin Heidelberg: 2012; Vol. 312, pp 175-212.
25. Turro, J. N., Principles of Molecular Photochemistry: An introduction. University
Science Book: Sausalito, Californi, 2009.
26. Dexter, D. L., A Theory of Sensitized Luminescence in Solids. The Journal of Chemical
Physics 1953, 21, (5), 836-850.
27. Forster, T., Inter. Ann. Physik 1948, 2, 55-75.
39

28. Michaelis, J., Quantitative Distance and Position Measurement using Single-Molecule
FRET. In Single Particle Tracking and Single Molecule Energy Transfer, Wiley-VCH Verlag
GmbH & Co. KGaA: 2009; pp 191-214.
29. Tsutsumi, K.; Yoshida, H.; Sato, N., Unoccupied electronic states in a hexatriacontane
thin film studied by inverse photoemission spectroscopy. Chemical Physics Letters 2002, 361,
(5–6), 367-373.
30. Kanai, K.; Akaike, K.; Koyasu, K.; Sakai, K.; Nishi, T.; Kamizuru, Y.; Nishi, T.; Ouchi,
Y.; Seki, K., Determination of electron affinity of electron accepting molecules. Applied Physics
A 2009, 95, (1), 309-313.
31. D’Andrade, B. W.; Datta, S.; Forrest, S. R.; Djurovich, P.; Polikarpov, E.; Thompson, M.
E., Relationship between the ionization and oxidation potentials of molecular organic
semiconductors. Organic Electronics 2005, 6, (1), 11-20.
32. Djurovich, P. I.; Mayo, E. I.; Forrest, S. R.; Thompson, M. E., Measurement of the lowest
unoccupied molecular orbital energies of molecular organic semiconductors. Organic
Electronics 2009, 10, (3), 515-520.
33. Marcus, R. A., On the Theory of Oxidation ‐Reduction Reactions Involving Electron
Transfer. I. The Journal of Chemical Physics 1956, 24, (5), 966-978.
34. Onsager, L., Initial Recombination of Ions. Physical Review 1938, 54, (8), 554-557.
35. Clarke, T. M.; Durrant, J. R., Charge Photogeneration in Organic Solar Cells. Chemical
Reviews 2010, 110, (11), 6736-6767.
36. Tachiya, M., Breakdown of the Onsager theory of geminate ion recombination. The
Journal of Chemical Physics 1988, 89, (11), 6929-6935.
37. Arkhipov, V. I.; Heremans, P.; Bässler, H., Why is exciton dissociation so efficient at the
interface between a conjugated polymer and an electron acceptor? Applied Physics Letters 2003,
82, (25), 4605-4607.
38. Peumans, P.; Forrest, S. R., Separation of geminate charge-pairs at donor–acceptor
interfaces in disordered solids. Chemical Physics Letters 2004, 398, (1–3), 27-31.
39. Offermans, T.; Meskers, S. C. J.; Janssen, R. A. J., Charge recombination in a poly(para-
phenylene vinylene)-fullerene derivative composite film studied by transient, nonresonant, hole-
burning spectroscopy. The Journal of Chemical Physics 2003, 119, (20), 10924-10929.
40. Koehler, M.; Santos, M. C.; da Luz, M. G. E., Positional disorder enhancement of exciton
dissociation at donor∕acceptor interface. Journal of Applied Physics 2006, 99, (5), -.
41. Howard, I. A.; Mauer, R.; Meister, M.; Laquai, F., Effect of Morphology on Ultrafast
Free Carrier Generation in Polythiophene:Fullerene Organic Solar Cells. Journal of the
American Chemical Society 2010, 132, (42), 14866-14876.
40

42. Coropceanu, V.; Cornil, J.; da Silva Filho, D. A.; Olivier, Y.; Silbey, R.; Brédas, J.-L.,
Charge Transport in Organic Semiconductors. Chemical Reviews 2007, 107, (4), 926-952.
43. Nelson, J.; Kwiatkowski, J. J.; Kirkpatrick, J.; Frost, J. M., Modeling Charge Transport in
Organic Photovoltaic Materials. Accounts of Chemical Research 2009, 42, (11), 1768-1778.
44. Ma, W.; Yang, C.; Gong, X.; Lee, K.; Heeger, A. J., Thermally Stable, Efficient Polymer
Solar Cells with Nanoscale Control of the Interpenetrating Network Morphology. Advanced
Functional Materials 2005, 15, (10), 1617-1622.
45. Hirose, Y.; Kahn, A.; Aristov, V.; Soukiassian, P.; Bulovic, V.; Forrest, S. R., Chemistry
and electronic properties of metal-organic semiconductor interfaces: Al, Ti, In, Sn, Ag, and Au
on PTCDA. Physical Review B 1996, 54, (19), 13748-13758.
46. Fahrenbruch, A. L.; Aranovich, J., Heterojunction phenomena and interfacial defects in
photovoltaic converters. In Solar Energy Conversion, Seraphin, B., Ed. Springer Berlin
Heidelberg: 1979; Vol. 31, pp 257-326.
47. NREL ASTM G173-03 AM 1.5 spectrum http://rredc.nrel.gov/solar/spectra/am1.5/
48. Shrotriya, V.; Li, G.; Yao, Y.; Moriarty, T.; Emery, K.; Yang, Y., Accurate Measurement
and Characterization of Organic Solar Cells. Advanced Functional Materials 2006, 16, (15),
2016-2023.
49. Kippelen, B.; Bredas, J.-L., Organic photovoltaics. Energy & Environmental Science
2009, 2, (3), 251-261.

41

Chapter 2. Steric bulk and Open Circuit Voltage

2.1 Introduction
As stated in section 1.4, one of the device parameters that is used to quantify the
efficiency of an OPV is the open circuit voltage (V
OC
) which is the voltage generated when the
circuit is opened in an OPV. Optimizing this parameter is an active field of research with much
potential with voltages of over 1.0 V now achieved in efficient devices.
1, 2
However, the V
OC
of
many OPVs are still as low as half the voltage of the incident photons absorbed at the lowest
energy transition leaving much room for improvement in many systems.
3
In many systems it is
not clear which materials properties contribute to this parameter. Understanding the origin of this
parameter and what materials properties that lead to it, is crucial in optimizing V
OC
while
maintaining the other performance parameters to achieve high efficiency OPVs. In addition, if
the origin of a material system’s V
OC
can be derived from the properties of the materials
themselves, large databases of materials could be screened without needing to construct every
donor acceptor pair, allowing for a more intelligent design of new OPV systems with less
dependence on serendipity.
In this chapter we will discuss an investigation into how materials properties contribute to
the V
OC
in OPV’s. The dependence of V
OC
on the reverse saturation dark current (Js) will be
describe though equations derived from a detailed balance of charges in OPVs and the properties
of the molecular system that give rise to the J
S
will be discussed and insight will be given to what
can be done in future systems to minimize this voltage loss. A system of perylene diimide
materials that has the steric bulk systematically modified will be analyzed and the interfacial
nature of this parameter will be discussed.

42

2.2 The Physical Origin of Voc
There are a variety of external forces that impact the final photo-voltage of an OPV
independent of the properties of the organic photo-active materials that make up the active
region. It is important to understand these forces so that standard conditions can be set up that
eliminate them when studying the active region. The first of these forces is the difference in the
work functions of the electrode materials that sandwich the photoactive region that extract the
charges from the device. This value, referred to as the built in potential, determines the
maximum voltage that can be achieved in the OPV.
4, 5
It has been shown clearly that when the
built-in potential is small it correlates linearly with the V
OC
of the resulting OPV. However,
once this potential is made large enough the V
OC
plateaus at a potential determined by the
photo-active material system and further increases in the built-in potential no longer raise the
V
OC
. In this investigation the built-in potential, set up between the ITO and aluminum
electrodes, is sufficiently large as to not be impacting the V
OC
of the OPVs. Light intensity has
been proven through fundamental principles to haves a logarithmic relationship with the V
OC
in
OPVs as it does in semiconductor solar cells and is standardized at one sun.
6
Finally,
temperature has been shown to have an inverse relationship with V
OC
with the origin of this
relationship coming from higher temperatures promoting thermally excited carriers at the D/A
interface at a faster rate. These thermally generated carriers are in opposition to photocurrent
under forward bias, working to reduce the V
OC
of the device.
7
These devices are tested under
STP conditions to eliminate these contributions.
The first material property that was proposed to be related to the V
OC
in OPV devices is
the energy difference between the highest occupied molecular orbital (HOMO) of the donor and
the lowest unoccupied molecular orbital (LUMO) of the acceptor, which is referred to as the
43

interfacial energy gap (ΔE
DA
).
8-10
This parameter is a natural extension from the band gap in
inorganic semiconductors which is the determining factor of V
OC
in those systems. Mutolo et.
al. showed this correlation by using a new donor molecule Subthalocyanine (SubPC) paired with
C
60
in a device and compared to the standard OPV based on the CuPc/C
60
heterojuction.
11
It was
measured that the ΔE
DA
in the SubPC/C
60
device is 0.4 eV larger than in the CuPc/C60 system.
In the resulting SubPC/C
60
devices the V
OC
is correspondingly increased by 0.5 V over the
CuPc/C
60
device. However, as seen in this study, V
OC
and ΔE
DA
do not correlate linearly and are
always offset from each other with ΔE
DA
being higher than the V
OC
. This discrepancy between
ΔE
DA
and V
OC
is referred to as the intrinsic “voltage loss” of the device.
3, 12
This voltage loss is
not consistent across different material pairs and is not predicable, making ΔE
DA
a poor indicator
of V
OC
in many systems. It is therefore important to discover the nature of this correlation so
that molecular systems with low voltage loss can be designed in an intelligent manner.

2.2.1 Shockley Diode Equation
The current-voltage (J-V) characteristics in OPVs often behave similarly to their
inorganic counterparts and much insight has been taken from that field in understanding OPVs.
Consequently, OPV performance can be modeled by the generalized Shockley equation
13, 14

(eq. 2.1). The Shockley diode equation was developed for p-n junctions with well defined band
structure in which excitons result in delocalized free charge carriers, OPVs, however, operate
with tightly bound, localized excitons which undergo hopping transport as described in section
1.3. Though there are discrepancies in the properties of organic and inorganic materials, the
Shockley diode equation has been shown to accurately parallel the operation of OPVs.
15
The
equation for current (J) according to the Shockley diode equation is as follows.
16

44

2.1
R
P
and R
S
are the parallel and series resistances, respectively. J
S
is the saturation dark
current, q is the fundamental charge, n is the diode ideality factor and J
Ph
is the photocurrent.
This equation can be simplified by assuming that the parallel resistance is orders of magnitude
larger than the series resistance (R
P
>> R
S
), causing the resistance ratio to reduce to one and that
the photocurrent is voltage independent and is therefore equal to the short circuit current
(J
Ph
=J
SC
) at all voltages. This results in Equation 2.1 reducing to:

2.2
The first term in this equation represents the thermally generated current described by the diode
ideality factor which, if around 2 as is the case in the OPVs studied here, is dominated by
recombination. The second term is the opposed light generated carriers traveling against the
internal bias of the diode.
17, 18
This equation can then be solved for voltage under open circuit
conditions (J=0, V= V
OC
). This means that we are finding the voltage at which J
S
and J
SC
are
equal and opposing. The resulting equation is:
19

2.3
It is safe to assume that the photocurrent in reverse bias is orders of magnitude higher
than the dark saturation current (J
SC
>>J
S
) and therefore J
SC
/J
S
>>1. This assumption lets us
simplify equation 2.3:

2.4
45

Here we can see that V
OC
has a direct correlation with the log of the ratio between
photocurrent and dark current. This equation suggests a clear mechanism to increase V
OC
.
Given that photocurrent will always be in the tens of mA/cm
2
when illuminated in 1 sun
conditions, it is the dark current that will have the determining impact on the open circuit voltage
as it can vary over several orders of magnitude between different materials systems.
19, 20
The
origin of the J
S
can be investigated further to give more insight into modifying this parameter.
As this current is generally due to thermal promotion of charges at the interface when the ideality
factor is close to 2, as it is in OPV’s, the saturation current can be described according to.
21, 22

2.5
In Eq. 2.5 E
a
is the activation energy required to promote a charge carrier at the interface.
Conceptually, this requires an electron to be taken from the HOMO of the donor and excited into
the LUMO of the acceptor material and so will be assumed here as equal to ΔE
DA
. The pre-
exponential factor J
S0
, which will be referred to here as the D/A coupling factor, depends on a
variety of materials properties including the bulk charge conductivities of the neat materials, the
reorganization energy of the carrier promotion, the density of states at the interface, and the
electronic coupling of the D/A pair. Substitution of equation 2.5 into 2.3 gets us:

2.6
Equation 2.6 now shows two different terms that determine V
OC
. The second term shows
the linear relationship with the thermalized carrier activation energy that has been related to
ΔE
DA
, which has been shown in literature.
11
The first term is a logarithmic dependence on the
ratio J
SC
/J
S0
which has been termed the “voltage loss,” and is dependent on the ratio of the
forward and reverse currents. This equation also reiterates some of the other V
OC
dependencies
46

discussed earlier, including an inverse linear relationship with temperature and the logarithmic
relationship with photocurrent and by extension illumination intensity.
To summarize the implications of these equations, the open circuit voltage results from
the competition of rates between the photocurrent and the reverse saturation current. Equation
2.2 suggests two different ways to decrease the magnitude of J
S
. The first is to increase the
energy gap that is needed to be overcome for the generation of dark carriers and the other is to
reduce coupling (J
SO
) at the interface to reduce the probability that the carriers are generated
even when sufficient energy is provided.

2.2.2 Relation Between Steric Bulk and V
OC

It has been suggested in literature that J
S
can be reduced by adding steric bulk to a donor
molecule when they are paired with C
60
in an OPV.
3
In this investigation, similar molecules
were made with varying amounts of steric bulk around their π systems. This I generally
achieved by adding substituents to the chromophores that do not contain valence orbitals,
therefore forcing valence orbitals on neighboring molecules to be further removed. When they
were paired with C
60
in an OPV, the resulting devices tended to have lower J
S
’s and higher V
OC
’s
than would be expected when compared their ΔE
DA
’s. In one particular case, tetracene was
compared to rubrene, which have very similar HOMO energies of 5.1 eV and 5.3 eV
respectively, but the latter has four additional phenyl rings attached to the central benzene rings.
Despite the similarities in HOMO energies, and therefore ΔE
DA
’s, the V
OC
’s of these devices are
very different at 0.55 V and 0.92 V, respectively. When analyzing these devices using Eq. 2.6
we see that the difference in performance comes from the wildly different J
S0
values which are
three orders of magnitude different. A similar result comes from comparing the performances of
47

CuPc and PtTPBP. In this case the HOMO values indicate that PtTPBP should have a lower V
OC

but after making the devices produces an improved V
OC
of 0.65 V over 0.44 V for CuPc. Again
this improvement is due to the dramatically decreased J
S0
value, in this case by five orders of
magnitude.
Both of these sets of devices support the claim that increased steric bulk can decrease the
J
S0
value and resulting in a higher V
OC
in the resulting device. It is, however, an imperfect study
as the there are many significant differences between these molecule pairs. In the process of
adding the steric bulk to the donor, differences in electronic structure and energy levels were also
introduced. It also remains to be proven that this effect can take place in pairs of materials that
do not include C
60
. Therefore, in this chapter, a series of acceptor side materials are developed
and studied in OPV devices which limit the changes as much as possible to steric bulk. It is also
reinforced that this is an interfacial parameter as the material present at the D/A interface is that
which dictates the V
OC
of the overall device.

2.3 Experimental
Photovoltaic devices of the composition ITO/CuPc (200A)/ PDI (200A)/
BCP(100A)//Al (1000A) were made with three different PDI compounds as the acceptor
material. These materials were made by refluxing 3,4,9,10-perylenetetracorboxylic dianhydride
(PTCDA) with the appropriate amine. The solution was then poured into an excess of water and
the precipitate was collected by filtration. The products were then purified by vacuum gradient
sublimation. Devices were grown by vacuum thermal evaporation (VTE) on indium tin oxide-
coated glass substrates that had been solvent cleaned and baked with UV ozone for 10 minutes.
The organic materials, PDI, copper phthalocyanine (CuPc) (Aldrich), C60 (MTR unlimited), and
48

2,9-dimethyl-4,7-diphenyl-1,10-phenanthroline (bathocuproine, BCP) (Aldrich) were purified by
thermal gradient sublimation in vacuum prior to use. Aluminum (99.999% pure, Alfa Aesar)
was used as received and evaporated through a shadow mask to for 1 mm diameter cathodes.
The current-voltage (J-V) curves were measured in the dark and under simulated 1 sun, AM1.5
illumination conditions. These were analyzed by established procedures,
23
obtaining the
parameters J
SC
, V
OC
, R
SA
, FF, n, J
S,
and η. The interface gap, ΔE
DA
, was calculated using the
CuPc HOMO value (5.2 eV)
24
and the LUMO of the PDI acceptors used.
25, 26
These were
measured by ultraviolet photoelectron spectroscopy (UPS) and inverse photoelectron
spectroscopy (IPES) respectively. The LUMO for dmPh-PDI was assumed to be equal to that of
Ph-PDI.

2.4 Results and Discussion
2.4.1 Material Characterization
A series of PDI acceptors was generated by varying the alkyl substituent at the nitrogen
position. The core structure of this class of compounds can be seen on the left of figure 2.1. The
objective of this series was to modify the amount of steric crowding around the perylene π-
system without significantly affecting the electronic energy levels or absorption. This was
achievable in this system do to the fact that the valence orbitals of PDI contain a nodal plane that
bisects the molecule lengthwise putting very little density on the nitrogen. The affect of the
addition of steric bulk to the PDI moiety can be seen when visualized by DFT modeling
software. The pictures on the right side of Figure 2.1 are end on views the three PDI materials,
49

based on literature x-ray crystal structures
25, 27
and the DFT calculated LUMO orbitals. The
figures depict the LUMO orbitals, shown in blue and red, interacting with a planar donor such as
CuPc. When R group attached to the nitrogen is hydrogen (H-PDI), the π-system is completely
exposed for intermolecular interactions with a planar donor allowing for greatest interaction.
However, when the R group is a phenyl group (Ph-PDI), the phenyl rings are rotated out of plane
with the main perylene core at a dihedral angle of 60 ⁰ degrees. This partially obstructs access to
the π-system and would weaken the intermolecular interactions with a planar donor. Finally,
when the R group becomes 2,6 dimethylphenyl (dmPh-PDI), the phenyl rings are then rotated
89⁰ degrees out of plane which should severely hinder the π-electrons of the perylene core from
participating in strong intermolecular interactions with its nearest neighbors.

Figure 2.1 The structure of three PDI derivatives are shown (left) with the dihedral angle of
the R group of each derivative listed (middle). The conformation of the molecule is
calculated (right) and shown end on with its impact on packing is shown in cartoon form.
50

400 500 600 700
a)
dmPh-PDI
PDI
Ph-PDI

Wavelength (nm)
Absorbance (a.u.)
Solution Adsorption Spectrum

0
1x10
5
2x10
5
3x10
5
4x10
5
5x10
5
400 500 600 700 800
Absorption Coefficient (cm
-1
)
PTCDI
Ph-PTCDI
dmPhPTCDI

Wavelength (nm)
Thin Film Adsorption Spectra
b)

Figure 2.2 (a) The solution spectra, of the three PDIs are shown offset from one another for
clarity. The substitution has been shown to make no changes to the electronic structure of the
molecule as the spectra are identical. (b) The absorption spectra of 20 nm thin films show
that this additional steric bulk has broken up the packing as predicted making the bulkiest
derivative have a spectrum close to that of solution
51

The electronic properties of this series of the PDI derivatives have been characterized by
UV-Vis absorption spectra in fluid solution and thin film, excitation-emission spectroscopy,
IPES measurements, and electrochemistry. Since both nitrogens reside within the nodal plane of
the π orbitals, substitutions of the R group has little impact on the valence orbitals including the
HOMO and LUMO. This is evident in the solution absorption spectra (Figure 2.2 Top ) as the
three compounds used have nearly identical spectra, with absorption peaks corresponding to the
0-0 through 0-3 vibronic levels in the S
0
-S
1
transition.
28
However, the alkyl substitutions have
achieved the goal of significantly breaking up intermolecular interactions in the thin film, a large
effect is seen in the solid state absorption spectra. . In the thin film absorption spectrum of
dmPh-PDI, when the 2,6 dimethylphenyl group is the R group, there is only slight broadening of
the spectrum, producing a curve very similar to that of the solution measurement, indicating that
there are no strong π-π interactions between nearest neighbors in the solid state due to the added
steric bulk. The thin film absorption spectrum of Ph-PDI shows more broadening of its features
with a slight shoulder at growing in at 575nm. In addition there is a re-ordering of the vibronic
level transition intensities, with the higher vibronic transitions being enhanced over the 0-0
transition, which can be attributed to some π-π interactions in the solid state. In the thin film
measurement of H-PDI, the spectrum shows intense broadening with a complete loss of the
vibrational structure. In addition, an intense band near 575 nm has appeared which is indicative
of an aggregate absorption event.
28
These types of bands are attributed to a charge transfer (CT)
excitation between two adjacent PDI molecules. These features, especially the presence of the
CT state, are indicative of strong intermolecular π-π stacking interactions between adjacent H-
52

PDI molecules and prove that close association with the PDI π orbitals are possible when no
steric bulk is added.
The excitation emission spectra were taken and the excitation spectra paralleled the thin
film absorption spectra closely. The emission spectra show a blue shift in energy in the solid
states as the R group becomes larger, again indicating that the addition of steric bulk decreases

400 500 600 700 800
0.0
0.2
0.4
0.6
0.8
1.0
Ph-PDI
PDI
dmPh-PDI

Nomalized count
Wavelength (nm)

Figure 2.3 The excitation (dashed line) and emission (solid line) spectra, measured at 77K, of
the three PDI derivatives in this study are presented. The relation between the excitation and
the thin film absorption spectra can be seen. The emission is red shifted and broadened in the
chromophores with less steric bulk. The emission spectra are broad and featureless for all
derivatives and red shifts as the steric bulk is removed.
53

π-π interactions in the thin film. The IPES data available in literature substantiates the idea that
alkyl substitution at the nitrogen position does not significantly change the electronic structure of
the resulting PDI molecule even in the thin film as there is only 0.09 eV difference in LUMO
levels in the thin films of H-PDI and Ph-PDI (-4.04 eV Vs. -3.95 eV).
25, 26
The dmPh-PDI
derivative has not been studied by IPES but titan DFT calculations suggest that the difference
between it and Ph-PDI (-3.61 eV vs. -3.58eV) should be smaller than those between Ph-PDI and
H-PDI (-3.61 eV vs. -3.70eV).

2.4.2 Device Characterization
Devices were made by pairing the three PDI materials with CuPc and the results are
shown in Figure 2.4. Section A) of that figure shows the dark J –V curves of the three PDI based
devices on a semi-logarithmic scale. In this plot the dark curves show a marked decrease of three
orders of magnitude (from 10
4
to 10
7
) in dark saturation current in the reverse bias as steric bulk
is added to the PDI moiety with H-PDI having the highest dark saturation current and dmPh-PDI
having the lowest. The light and dark curves of the same devices are shown on a linear scale in
figure 2.4 B). In this plot the effect of the changes in dark current can be seen, as the devices
more sertic bulk, and lower J
S
, produce higher V
OC
’s. The device with H-PDI as the acceptor
produces a V
OC
of 0.28 V. This is increased to 0.35 V for the Ph-PDI device, and then is further
pushed to 0.50 V for dmPh-PDI, the molecule with the most steric bulk. The parameters
extracted from these plots are summarized in Table 2.2. From the data in table 2.1 a strong
logarithmic correlation can be seen between J
s
, which decreases by orders of magnitude, and the
Voc of the devices, which increase linearly, as is required by equation 2.4.
54

-0.8 -0.4 0.0 0.4 0.8
-8
-4
0
4
8
10
-8
10
-6
10
-4
10
-2
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
a)
(A/cm
2
) (mA/cm
2
)
Voltage (V)

Current density
dmPh-PDI
Ph-PDI
H-PDI

Power Density (mW/cm
2
)
400 500 600 700 800
0
5
10
15
20

b)
H-PDI
Ph-PDI
dmPh-PDI
EQE (%)
Wavelength (nm)

Figure 2.4 (a) Current-voltage curves of the ITO/CuPc/PDI/BCP//Al devices are shown on
both semi-log and linear plots. Comparison of the curves shows that bulkier PDI materials
produce substantially higher V
OC
’s. This increase in V
OC
corresponds to a decrease in the
dark current as is dictated in equation 2.4. (b) EQE spectrums of the same devices
55

By design the LUMO’s of these three materials is virtually unchanged and therefore so is ΔE
DA
.
This means that, considering equation 2.5, that the boost in V
OC
is due to a decrease in J
S0
. It is
therefore postulated that as steric bulk is added to the PDI moiety the geometry of the D/A
interface is changed, restricting CuPc access to the π electron orbitals, resulting in a decrease of
the coupling between the donor and acceptor resulting in the marked decrease in J
SO
. As part of
this analysis the open circuit voltages were calculated according to eq. 2.5, and listed in table 2.1.
This calculation uses the measured J
S
, J
SC
and calculated ΔE
DA
parameter and projects the
excepted V
OC
. These calculated values matched the experimental values well implying that the
formulas and assumptions used were appropriate for the devices under study. The fit is not as
good for the dmPh-PDI device due to the fact that the device has a series resistance that is larger
by two orders of magnitude than the others, which begins to violate one of the assumptions used
to derive equation 2.4, where it is assumed that the series resistance is very small.
This series of acceptors also demonstrates that while the addition of steric bulk around the π-
system decreases unfavorable donor-acceptor interactions that lead to high J
SO
’s, it can also
reduce favorable self exchange coupling interactions between molecules, which results in higher
series resistances (R
SA
’s). These higher resistances frequently lead to lower fill factors (FF) and
less square rectification, which can be seen in the dmPh-PDI device (Fig. 2.4B). Additionally,
J
SC
of this PDI device series is affected by the addition of bulky R groups in other ways. First,
the addition of bulky R groups increases the size of each molecule which results in having fewer
chromophores per unit volume. Because overall oscillator strength of each PDI is very similar,
56

this results in a reduction in the optical densities seen in Fig. 2.2.Bottom. Secondly, the
addition of bulkier R groups significantly affects the wavelength dependent external quantum
efficiencies (EQEs) which can be seen in Fig. 2A. These EQEs strongly influenced by the film
absorptions, as well as other factors including excition diffusion length, charge trapping, and
charge generation/recombination rates. The overlap of these EQE curves with the solar spectrum
generates the J
SC
’s of the three devices. The addition of steric bulk obstructing access to the π-
system leads to both favorable and unfavorable materials properties, resulting in a tradeoff that
needs to be balanced to maximize power conversion efficiencies. This can be seen here as the
performance is optimized for the Ph-PDI device, having the highest conversion efficiencies.

2.4.3 Interfacial Nature of V
OC

It has been concluded here that the V
OC
is determined not by the bulk properties as much
as those at the D/A interface. To experimentally reinforce this point a series of devices were
made that inserted the bulky PDI molecule at the D/A interface of an H-PDI/CuPc device. The
resulting J-V curves of these devices is shown in figure 2.5 where the structure of these devices
are a variable amount of dmPh-PDI sandwiched between 200 Å of H-PDI and 200 Å of CuPc.

Table 2.1 Averaged device parameters for ITO/CuPc/PDI/BCP//Al devices showing the
clear correlation between the dark current (J
S
) and open circuit voltage (V
OC
).
57

In figure 2.5 four curves are shown in both semi-log and linear plots of the devices with 0 to 100
Å of dmPh-PDI inserted at the D/A interface. The data is compiled in table 2.2. Here it can be
seen that as the sterically hindered molecule dmPh-PDI layer is increased in depth the dark
saturation current is decreased with a corresponding increase in V
OC
. When the interlayer
becomes 100 A the V
OC
is equal to that of the neat dmPh-PDI. It is likely that the blended V
OC
’s
produced by the thinner interlayers are due to the fact that the layer is not contiguous presenting
mixed D/A interface with a weight averaged V
OC
. This study proves first that V
OC
is a
predominantly an interfacial property that arises from the interaction of the donor and acceptors
-0.8 -0.4 0.0 0.4 0.8
-8
-4
0
4
8
10
-8
10
-6
10
-4
10
-2
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
0 Å dmPh-PDI
25 Å dmPh-PDI
50 Å dmPh-PDI
100 Å dmPh-PDI
(A/cm
2
)
(mA/cm
2
)
Voltage (V)

Current density
Power Density (mW/cm
2
)

Figure 2.5 Devices with the structure ITO/CuPc/dmPh-PDI(X)/H-PDI/BCP//Al with varying
thicknesses of dmPh-PDI are shown. The device with no interlayer (Black) has the lowest V
OC

which increases as the interlayer grows. The device with 100 Å of dmPh-PDI has the highest
V
OC
which is similar to that of the bulk dmPh-PDI device.
58

at a molecular level. In addition it serves as a test case that shows that multiple acceptor
materials can be used each tailored to their different functions. Here we have used dmPh-PDI
for its interfacial property producing a high V
OC
, while using H-PDI for its bulk properties of
higher conductivity yielding a device with low resistance and high V
OC
. However, in this test
case not all the materials properties are optimized as we have introduced a voltage dependence
on the photo-current, as can be seen from the sloped nature of the curve prior to the rectification
potential. This is most likely due to trapping of the charge carriers at the interface due to
misaligned energy levels. In addition excitons will be energetically encouraged to transfer from
the dmPh-PDI to H-PDI away from the interface, further decreasing the photo-current.

2.5 Conclusions
In this chapter a series of PDI acceptors where prepared which primarily vary only in the
amount of sertic bulk around their electronic core. The series was especially tailored to do this
without changing the electronics of the valence orbitals, so that the effect of steric bulk can be

Table 2.2 The parameters for the ITO/CuPc/dmPh-PDI (X)/H-PDI/BCP//Al device.
Again there is a direct correlation here between J
S
and V
OC
. It can be seen here that the
devices can be optimized using multiple materials in the acceptor layer, reaching a
maximum efficiency at 50 Å of dmPh-PDI at the interface.
59

studied by itself. Using this acceptor series to make OPVs with CuPc we were able to prove that
the amount of sertic bulk can have a dramatic impact on the resulting V
OC
in the device, by
nearly doubling the voltage produced without changing other factors. This modifies the
previously theory that the primary factor determining V
OC
was simply the energy gap between
donor and acceptor. However, it was shown at the same time that the same steric bulk that
benefits the open circuit voltage can be harmful to other device parameters, primarily by
introducing additional series resistance in the device.
It was further proved that the effect that steric bulk has on the devices V
OC
is the result of
it breaking up intermolecular interactions at the donor acceptor interface of the device. It was
proved that only a small amount of sertically hindered acceptor material was needed to be added
to the D/A interface to have major effects on the resulting voltage. Through this work it is
suggested that further improvement in device performance can be achieved by tailoring different
acceptor materials for their interfacial and bulk properties and used in multilayer stacks. In this
way one could avoid some of the negative properties that come along with using sertically
hindered materials in OPVs.

60

2.6 Chapter 2 References

1. Chen, L.; Huang, L.; Yang, D.; Ma, S.; Zhou, X.; Zhang, J.; Tu, G.; Li, C., A non-
fullerene acceptor with all "A" units realizing high open-circuit voltage solution-processed
organic photovoltaics. Journal of Materials Chemistry A 2014, 2, (8), 2657-2662.
2. Peng, Y.; Zhang, L.; Andrew, T. L., High open-circuit voltage, high fill factor single-
junction organic solar cells. Applied Physics Letters 2014, 105, (8), 083304.
3. Perez, M. D.; Borek, C.; Forrest, S. R.; Thompson, M. E., Molecular and Morphological
Influences on the Open Circuit Voltages of Organic Photovoltaic Devices. Journal of the
American Chemical Society 2009, 131, (26), 9281-9286.
4. Mihailetchi, V. D.; Blom, P. W. M.; Hummelen, J. C.; Rispens, M. T., Cathode
dependence of the open-circuit voltage of polymer:fullerene bulk heterojunction solar cells.
Journal of Applied Physics 2003, 94, (10), 6849-6854.
5. Lo, M. F.; Ng, T. W.; Liu, T. Z.; Roy, V. A. L.; Lai, S. L.; Fung, M. K.; Lee, C. S.; Lee,
S. T., Limits of open circuit voltage in organic photovoltaic devices. Applied Physics Letters
2010, 96, (11), -.
6. Koster, L. J. A.; Mihailetchi, V. D.; Ramaker, R.; Blom, P. W. M., Light intensity
dependence of open-ciruit voltage of plymer:fullerene solar cells. Applied Physics Letters 2005,
94, (10), 123509.
7. 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. Physical
Review B 2007, 75, (11), 115327.
8. Gadisa, A. M.; Svensson, M.; Andersson, M. R.; Inganas, O., Correlation between
oxidation potential and open-circuit voltage of composite solar cells based on blends of
polythiophenes/ fullerene derivative. Applied physics Letters 2004, 84, (9), 1609-1611.
9. Kooistra, F. B.; Knol, J.; Kastenberg, F.; Popescu, L. M.; Verhees, W. J. H.; Kroon, J.
M.; Hummelen, J. C., Increasing the Oen Circuit Volatage of Bulk-Heterojunction Solar Cells by
Raising the LUMO Level of the Acceptor. Organic Letters 2007, 9, (4), 551-54.
10. Mutolo, K. l.; Mayo, E. I.; Rand, B. P.; Forrest, S. R.; Thompson, M. E., Enhanced open-
circuit voltage in subphthalocyanine/C-60 organic photovoltaic cells. Journal of American
Chemical Society 2006, 128, (25), 8108-8109.
61

11. 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. Journal of the
American Chemical Society 2006, 128, (25), 8108-8109.
12. Erwin, P.; Thompson, M. E., Elucidating the interplay between dark current coupling and
open circuit voltage in organic photovoltaics. Applied Physics Letters 2011, 98, (22), 223305.
13. Shockley, W., Bell System Tech. J. 1949, 28, 435.
14. Shockley, W., Electrons and holes in semiconductors with applications to trasistor
electronics. D. Van Nostrand, New York 1950.
15. Giebink, N. C.; Wiederrecht, G. P.; Wasielewski, M. R.; Forrest, S. R., Ideal Diode
equation for organic heterojunctions. Iderivation and application. Physical Review B 2010, 82,
155305.
16. Sze, S. M., Physics of Semiconductor devices. In 2nd ed.; Wiley-Interscience: New York,
NY, 1981; p 878.
17. Bube, H. R.; Fahrenbruch, A. L., Advances in Electronics and electron Physics.
Academic: New York 1981, 163.
18. Fahrenbruch, A. L.; Aranovich, J., Solar Energy Conversion - Solid-State Pysics Aspects.
Topics in Applied Hysics 1979, 31, 257.
19. Li, N.; Lassiter, B. E.; Lunt, R. R.; Wei, G.; Forrest, S. R., Open circuit voltage
enhancement due to reduced dark current in small molecule photovoltaic cells. Applied Physics
Letters 2009, 94, (2), -.
20. Potscavage, W. J.; Yoo, S.; Kippelen, B., Origin of the open-circuit voltage in multilayer
heterojunction organic solar cells. Applied Physics Letters 2008, 93, (19), -.
21. Nelson, J., The physics of Solar Cells. Imerial College Press: London, 2003.
22. Wurfel, P., Physics of Solar Cells: From Principles to New Concepts. Wiley-VCH:
Weinhem, 2005; Vol. 186.
23. Shrotriya, V.; Li, G.; Yao, Y.; Moriarty, T.; Emery, K.; Yang, Y., Accurate Measuremetn
and Characterization of Orgnic Solar Cells. Advanced Functional Materials 2006, 16, (15), 2016.
24. D'Andrade, B. W.; Datta, S.; Forrest, S. R.; Djurovich, P.; Polikarpov, E.; Thompson, M.
E., Relationship between the ionization and oxidation potentials of molecular organic
semiconductors. Organic Electronics 2005, 6, 11.
25. Hadicke, E.; Graser, F., Structures of eleven perylene-3,4:9,10-bis(dicarboximide)
pigments. Acta Cryst. 1986, C42, 189.
62

26. Zahn, D. E. T.; Gavrila, G. N.; Gorgoi, M., The transport gap of organic semiconductors
studied using the combination of direct and inverse photoemission. Chemical Physics 2006, 325,
99.
27. Sato, K.; Mizuguchi, J., N,N'-Diphenylperylene-3,4:9,10-bis(dicarboximide). Acta
Crystallographica Section E 2006, 62, (11), o5008-o5009.
28. Hoffmann, M.; Schmidt, K.; Fritz, T.; Hasche, T.; Agranovich, V. M.; Leo, K., Chemical
Physics 2000, 258, 73.

63

Chapter 3. Relation Between E
CT
and V
OC

3.1 Introduction
In chapter 2 the origin of open circuit voltage (V
OC
) was discussed and an equation 3.1
was derived from the Scholkey’s diode equation.
1-4

3.1
In this equation J
SC
is the short circuit current, J
S0
is the coupling factor and E
a
is the
energy barrier for thermal promotion on and electron across the interface. Additionally, n is the
diode ideality factor, k is the Boltzmann constant, T is temperature and q is the elementary
charge. Here V
OC
is related to two terms; the first is a kinetic term dependent on the ratio of
short circuit current to dark current coupling and the second is a thermodynamic term based in
the activation energy of thermally promoted charge carriers (E
a
).
5-7
In chapter 2 the E
a
was
assumed to be related to the difference between the highest occupied molecular orbital (HOMO)
of the donor and the lowest unoccupied molecular orbital (LUMO) of the acceptor or ΔE
DA
.
However, a lack of direct correlation between V
OC
and ΔE
DA
, though it can be explained through
changes in the kinetic term, suggests that there could be a better measure of E
a
.
8
As E
a
is a bulk
property it may not be the best representation of Ea in a system that we have shown is dominated
by localized, interfacial and molecular properties.
9

In this chapter a new insitu measurement for the estimation of Ea will be discussed and
extracted performed on a series of small molecule OPVs. The relation between this new measure
for Ea and V
OC
will be explored. Finally, the shortcomings of this measurement and some
exceptions to its good correlation with V
OC
will be discussed.
64

3.2 Experimental
Materials: Tetracene (Aldrich), rubrene (Aldrich), MoO
3
(Alfa Aesar), copper
phthalocyanine (CuPc) (Aldrich), C60 (MTR unlimited), and 2,9-dimethyl-4,7-diphenyl-1,10-
phenanthroline (bathocuproine, BCP) (Aldrich) were purchased and purified by thermal gradient
sublimation in a three zone oven prior to use. PDI materials were synthesized according to
literature procedures and also purified by sublimation.

Devices: OPV devices were grown by vacuum thermal evaporation on patterned indium
tin oxide glass substrates that had been solvent cleaned and baked with UV ozone for 10
minutes. Materials were deposited at 2 Å/s for all materials except tetracene which was
deposited at 15-25 Å/s. This rate is needed to prevent excess crystallization of the film.
Aluminum (99.999% pure, Alfa Aesar) was used as received and evaporated through a shadow
mask to for 1 mm cathodes stripes, resulting in with devices areas of 1 mm
2
. Current-voltage (J-
V) curves were measured in the dark and under simulated 1 sun, AM1.5 illumination conditions.
These were mismatch corrected and analyzed by established procedures,
10
obtaining the
parameters J
SC
, V
OC
, FF, n, J
S,
and η. E
CT
EQE measurements where performed in the McGehee
laboratory at Stanford as detailed in Ko et al.
11

3.3 Re-Evaluation of E
a
As stated in chapter 2, E
a
is most directly related to the energy barrier for the thermal
promotion of charge carriers at the interface. This electron transfer involves exciting a D/A pair
from the S
0
to S
1
or T
1
state, which principally would involve promoting an electron from the
donor’s HOMO to the acceptor’s LUMO orbital.
12
The resulting charge transfer state (CT state)
65

is composed of a positively charged donor and negatively charged acceptor molecular pair. This
is the reason that the difference in the energies of these molecular orbitals was used previously to
represent Ea. however this ignores any interaction the two molecules have at the D/A interface
that might impact the energetic of the system.
13

This CT state is theoretically the same that is generated after thermaliztation of a CT state
generated from exciton charge transfer which is described in section 1.3 and it is possible to
optically excite this CT state directly. However, the CT state absorptions have low extition
coefficients and a low density of states due to the fact that they only occur where donor and
acceptor meet. To get around this problem a technique was developed to probe this state by
doing a very sensitive spectral response measurement of the external quantum efficiency (EQE)
of the devices.
14-16
The CT transition can then be modeled as a simple harmonic oscillator and
the energy of the CT state (E
CT
) can be extracted. This is done by illuminating a device with a
known quantity of light and measuring the number of electrons that exit the device. With this
technique it is possible to get acceptable signal to noise ratios down to E
-7
percent as can be seen
in figure 3.1. In the resulting EQE spectrums such as those shown in figure 3.1 (b) the lowest
energy transition is assumed to be the CT absorption and can be analyzed to extract E
CT
.
17
The
energy dependent cross section for the CT absorbance, σ(E), is given in equation 3.2.

3.2

3.3
Here, V is the electronic coupling matrix element, λ is the reorganization energy, E
CT
is
the energy of the CT state. σ(E) can is proportional to EQE in the CT region by multiplying by
66

the number of CT states (N
CT
) and the internal quantum efficiency (IQE). By adding these
factors in we get equation 3.3, where the V
2
x N
CT
x IQE = f , relating EQE to the CT transition.
E
CT
can then be used as a new measure of E
a
in equation 3.1 which takes into account molecular
interactions between the D/A pair that are not included in the simpler parameter of ΔE
DA
.

3.4 Results and Discussion
3.4.1 Relation between E
CT
and V
OC

Rubrene and tetracene have been studied in OPVs to determine the origin of V
OC
in such
devices.
8
In this study it was seen that the V
OC
was not well correlated with the ΔE
DA
of the
device with an enhancement in the rubrene device. Similarly to the PDI series studied in chapter
2, it was determined that the enhancement was due to the added steric bulk on the rubrene
molecule which reduces the voltage loss due to the kinetic term in equation 3.1. Here this
system was re-evaluated using the E
CT
measurement. Devices were made of the structure
ITO/MoO
X
(5)/Donor/C
60
(40)/BCP(10)/Al(100) and analyzed by J-V and EQE measurement and
the results are shown in figure 3.1. In part (a) the light (solid) and dark (dotted) J-V curves on a
semi-log plot are presented. The data from these curves are summarized in table 3.1. In the plot
it can be seen that the rubrene device has a higher voltage than tetracene (0.94 V vs. 0.69 V)
which is accompanied by a lower reverse current. In part (b) of figure 3.1 the EQE curves of the
same two devices plotted on a semi-log plot. The CT absorption band is the lowest energy
transition and can be seen here as a clear shoulder in the plot. This shoulder is modeled as a
simple harmonic oscillator and analyzed using equation 3.3.
18, 19

67

Figure 3.1 (a) the I-V curves of rubrene/C
60
and tetracene/C
60
are shown here plotted on a semi-log
plot. (b) The EQE curves of the same devices are analyzed using equation 3.3 with the parameters
for the CT absorption overlaid.
68

The parameters for each curve are shown on the plot and shows that rubrene has a transition
0.25 V higher than tetracene (1.23 vs. 1.48). The difference in the V
OC
s of the same devices is
also 0.25 V, as can be seen in table A, and established an exactly linear relationship between V
OC

and E
CT
is established. The one to one correlation between these two parameters in this set of
devices mandates that there is no change in the value of the kinetic term for these two devices. f
, which can be related to the coupling of the D/A pair, is higher for the rubrene device than for
tetracene (2.8E
-4
vs. 9.4E
-6
). Though this parameter is conflated with other factors, this is
implies that the coupling of the tetracene/C
60
is lower than that in the rubrene/C
60
D/A pair,
opposite what has been previously purported.
8

This direct one to one correlation between E
CT
and V
OC
continues a trend measured in
polymer based bulk heterojunction (BHJ) OPVs.
17
This linear relationship is summarized in
figure 3.2 where qV
OC
is related to E
CT
with a slope of 1 and a y-intercept of -0.06 eV. All
points plotted here have a residual of ± 0.07 eV. This means that in these systems most to all of
the change in the V
OC
for can be explained due to changes in the energy of the D/A CT state with
no need to invoke coupling. The small molecules studied here follow this slope exactly though
they are high off of the trend line with residuals of 0.06 eV. These studies imply that there is a
flat 0.6 V “voltage loss” from E
CT
that cannot be circumvented regardless of the steric bulk
Device J
SC
V
OC
FF η
Tetracene/C
60
3.59 0.758 0.65 1.76
Rubrene/C
60
3.54 0.733 0.63 1.62

Table 3.1 Parameters from ITO/Ancene/C60/BCP//Al devices.
69

added. This is in direct opposition to the conclusions drawn in chapter 2 when studying a PDI
based chromophore system.
9

3.4.2 E
CT
Measurement PDI
To shed more light on the magnitude of impact E
CT
and J
SO
have on V
OC
, the same family
of PDI molecules in chapter one where studied with the E
CT
measurement. Devices of the
structure ITO/ZnPc/PDI/BCP//Al were mad and the results of two derivatives are shown in
figure 3.4 and 3.3. These devices use Zinc phthalocyanine as the donor layer and not copper
phthalocyanine, as was done in the study in chapter 2, due to a limitation with the E
CT

measurement. The CT state absorption is the lowest energy photocurrent generating state by

Figure 3.2 In many systems the correlation between E
CT
and V
OC
is extremely strong. Here is
a series of experiments showing these results. Most devices represent polymer BHJs. Figure
taken from reference 8.
70

Figure 3.3 E
CT
measurements of ZnPc/PDI devices are shown here. The CT transition is
modeled as a simple harmonic oscillator and the extracted parameters are overlaid.
71

-1.0 -0.5 0.0 0.5 1.0
10
-6
10
-5
10
-4
10
-3
10
-2
10
-1
10
0
10
1
10
2
10
3
ZnPc/H-PDI Bilayer
ZnPc/H-PDI Gradient Blend

J (mA/cm
2
)
Applied Bias (V)
a)

-1.0 -0.5 0.0 0.5 1.0
10
-6
10
-5
10
-4
10
-3
10
-2
10
-1
10
0
10
1
10
2
10
3
ZnPc/C
12
-PDI Bilayer
ZnPc/C
12
-PDI Gradient Blend
b)

J (mA/cm
2
)
Applied Bias (V)

Figure 3.4 The I-V curves for the devices measured in figure 3.3 are plotted here in semi-log
form so that the V
OC
and reverse current can be clearly seen. The relevant parameters are
summarized in table J.
72

definition, but to be resolved in the EQE spectrum E
CT
must be distinctively separated from the
next higher energy transition. Due to this requirement, the E
CT
of CuPc/PDI based devices could
not be measured due to the presence of low lying triplet absorptions of CuPc that obscure the CT
transition. For this study both bilayer devices and gradient blended devices were made and
studied. Gradient blend device refers to the technique of fabricating a layer of organic material
that is 100% component A and 0% component B at the start of the layer that linearly changes in
composition as a function of depth until it is 0% component A and 100%component B. By
making devices with this structure of D/A layers we are able to increase the density of CT states
in the device while maintaining some functionality, increasing the intensity of the CT transition
in the EQE spectrum. ZnPc/PDI BHJ devices with a flat blending ratio, as in traditional BJH
devices, do increase the density of CT states further but produces devices with open circuit like
Device J
SC
(mA/cm
2
)
V
OC
(V)
FF η
(%)
Gradient
ZnPc/H-PDI
3.44 0.35 0.56 0.68
Bilayer
ZnPc/H-PDI
1.54 0.50 0.45 0.35
Gradient
ZnPc/C
12
-PDI
1.20 0.59 0.54 0.39
Bilayer
ZnPc/C
12
-PDI
0.16 0.68 0.21 0.023

Table 3.2 Comparison of the parameters from I-V curves of the ITO/ZnPc/PDI/BCP//Al
devices.
73

curves with negligible currents making the EQE signal less intense. The EQE for both the
bilayer and gradient blend H-PDI/ZnPC are shown in figure 3.3 (a) and the effect of boosting the
CT signal can be seen. A clear shoulder appears in the blended device with an E
CT
of 1.28 eV,
while the bilayer device produces no clear CT signal and the E
CT
cannot be calculated. In figure
3.3 (b) the EQE the gradient blend C
12
-PDI/ZnPc device is plotted and a very slight shoulder
seen in the curve is fitted with a harmonic and an E
CT
of 1.32 eV is calculated. Similarly to the
H-PDI case the bilayer device does not show any measurable CT absorbance and an E
CT
cannot
be extracted. In figure 3.4 the I-V curves with the H-PDI (a) and C12-PCI (b) are plotted on a
semi log plot to highlight the reverse current and V
OC
. Here it can be seen that blending the
layers both reduces reverse bias reverse current and increases the V
OC
though at the expense of
photocurrent. Comparing the V
OC
’s from figure 3.4 with E
CT
’s calculated from figure 3.3, using
the correlation highlighted in figure 3.2, shows that gradient H-PDI/ZnPc device, having a V
OC

of 0.50 V and an E
CT
of 1.28 eV, has a residual of -0.18 eV, putting it off of the trend line by
three times the average. The gradient C
12
-PDI/ZnPc device has a V
OC
of 0.74 and an E
CT
of
1.32 eV giving it the more reasonable residual of 0.02 eV.
Here we have shown using a pair of PDI devices a few limitations to the E
CT

measurement. First is that most PDI devices could not be analyzed with the technique due to a
lack of CT signal in the EQE spectrum. Even in the derivatives where CT signals can be
measured they only appear in blended device structures. Finally, though the C
12
-PDI derivative
closely follows the trend seen previously, the H-PDI/ZnPc device deviates drastically. In this
device we measure a V
OC
substantially lower that would be predicted from the trend line. It
could be that the shoulder analyzed in the EQE is not due to the lowest energy CT state.
However, if this is an accurate measurement of the E
CT
of the blended H-PDI ZnPc device it
74

presents evidence that steric bulk, or the lack thereof in this case, can have substantial impact on
the V
OC
. This is the first evidence of that impact from E
CT
measurements which may be because
H-PDI is the most closely associating molecule studied.[REF] More work may be needed to
fully understand this outlier.

3.4.3 Exposure Study of Tetracene/C
60

After testing the tetracene/C
60
based devices reported here it was observed that the V
OC

they produced was substantially higher than those reported in literature (0.50 V vs. 0.76 V).
8

-0.8 -0.4 0.0 0.4 0.8
-4
0
4
8
10
-8
10
-6
10
-4
10
-2
N
2
H
2
O
O
2
Air
(A/cm
2
) (mA/cm
2
)
Voltage (V)

Current density

Figure 3.5 I-V curves for tetracene/C
60
devices exposed to four different conditions are presented
here. From this data it can be seen that it is the oxygen in air that is causing the quick degradation of
the V
OC

75

The primary difference in the devices reported here from those in literature is that these are not
exposed to atmospheric air between the deposition of the organics and the metal layer. This was
required previously so that shadow masks could be applied for the fabrication of the cathode.
With the system used to prepare the devices reported here a glove box is attached to the chamber
so that these manipulations take place in an inert nitrogen atmosphere. To determine what was
causing this discrepancy in the devices an exposure study was carried out and the results are
shown in figure 3.5 and tabulated in Table 3.2. In this study tetracene/C
60
devices were prepared
and exposed to different gasses between the organic and metal depositions. For the negative
control the device was exposed to the nitrogen gas in the glove box and the positive control
device was removed from the glove box and exposed to room air for 30 seconds. In addition,
one set of devices was exposed to 33% oxygen in nitrogen for 30 seconds

Device J
SC
V
OC
FF η
N
2
3.59 0.758 0.65 1.76
H
2
O 3.54 0.733 0.63 1.62
O
2
2.76 0.614 0.55 0.925
Air 2.92 0.622 0.52 0.945

Table 3.3 Comparison of the parameters from the exposure study of the device
ITO/Tetracene/C60/BCP//Al show similarity between the air and oxygen exposed devices.
76

study it is clear that exposure to oxygen is causing the reduction in V
OC
implying that some
oxidation process has occurred in the device. and the another was exposed to nitrogen bubbled
through a water trap for the same period. All devices were then pumped down in the
antechamber and returned to the glove box. From this
It is well known that tetracene can undergo a reversible oxidation of one of its internal
rings forming a peroxide bridge under atmospheric conditions.
20
It was previously thought that
due to the relatively slow rate that this reaction was not significant in the preparation of these
devices. However, it appears that within 30 seconds of exposure enough of the oxidized species
can form to substantially influence the V
OC
of these devices. This is most likely due to the fact
that, as studied in chapter two, material at the D/A interface is what is most important for
determining V
OC
. It is seen here that after only a short exposure duration, enough oxidization
occurs at this interface to substantially reduce the V
OC
.
This defect forming in tetracene devices shows another shortcoming of the E
CT

measurement. Studying tetracene/C
60
devices that have been exposed to air during fabrication
were measured for E
CT
and it was found that they produce EQE curves indistinguishable from
devices that had been prepared under inert conditions. As this system shows no change in the
E
CT
of devices with substantially different the trend seen in the study down in section 3.4.1 is no
longer held. It is presumed here that a lower lying CT state is now forming between the oxidized
tetracene and C
60
. It can therefore be assumed that though this new CT state can substantially
affect the V
OC
of the device the density of states in the device is far too low to be visible to the
EQE measurement and cannot be probed. This make it very important to make sure that the CT
state measured is truly the lowest lying CT state in the device under study.

77

3.5 Conclusions
Here the relatively new measurement of the lowest lying CT state of OPV’s has been
described and used to analyze a pair of acene/C
60
devices to relate the E
CT
and V
OC
. This study
expands on a trend well measured in polymer BJH devices that shows a strong one to one
correlation between these two parameters. However, expanding this measurement to other
bilayer and planar of Zn/PDI devices reveals some of the limitations to using this measurement .
Frequently, true bilayer devices have too little CT absorption to be analyzed through this
mechanism due to the face that these CT states only exist at the interface of the layers and do not
have the intensity to be observed. Constructing gradient blended PDI devices can alleviate this
problem but do not work for all derivatives. For the bulkier C
12
-PDI the one to one trend shown
in figure 3.2 is held. However, the ZnPc/H-PDI device drastically deviates from this trend
indicating that it may not be valid for very closely associating D/A pairs. Finally, we have
shown that low density defects can be created at the D/A interface that substantially change the
V
OC
but are not detectible by E
CT
measurement. This makes if very important to positively
identify the state being measured to insure that it is the lowest lying CT state in the device.

78

3.6 Chapter 3 References

1. Shockley, W., Bell System Tech. J. 1949, 28, 435.
2. Shockley, W., Electrons and holes in semiconductors with applications to trasistor
electronics. D. Van Nostrand, New York 1950.
3. Sze, S. M., Physics of Semiconductor devices. In 2nd ed.; Wiley-Interscience: New York,
NY, 1981; p 878.
4. Giebink, N. C.; Wiederrecht, G. P.; Wasielewski, M. R.; Forrest, S. R., Ideal Diode
equation for organic heterojunctions. Iderivation and application. Physical Review B 2010, 82,
155305.
5. 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. Physical
Review B 2007, 75, (11), 115327.
6. Bube, H. R.; Fahrenbruch, A. L., Advances in Electronics and electron Physics.
Academic: New York 1981, 163.
7. Fahrenbruch, A. L.; Aranovich, J., Solar Energy Conversion - Solid-State Pysics Aspects.
Topics in Applied Hysics 1979, 31, 257.
8. Perez, M. D.; Borek, C.; Forrest, S. R.; Thompson, M. E., Molecular and Morphological
Influences on the Open Circuit Voltages of Organic Photovoltaic Devices. Journal of the
American Chemical Society 2009, 131, (26), 9281-9286.
9. Erwin, P.; Thompson, M. E., Elucidating the interplay between dark current coupling and
open circuit voltage in organic photovoltaics. Applied Physics Letters 2011, 98, (22), 223305.
10. Shrotriya, V.; Li, G.; Yao, Y.; Moriarty, T.; Emery, K.; Yang, Y., Accurate Measurement
and Characterization of Organic Solar Cells. Advanced Functional Materials 2006, 16, (15),
2016-2023.
11. Ko, S.; Hoke, E. T.; Pandey, L.; Hong, S.; Mondal, R.; Risko, C.; Yi, Y.; Noriega, R.;
McGehee, M. D.; Brédas, J.-L.; Salleo, A.; Bao, Z., Controlled Conjugated Backbone Twisting
for an Increased Open-Circuit Voltage while Having a High Short-Circuit Current in
Poly(hexylthiophene) Derivatives. Journal of the American Chemical Society 2012, 134, (11),
5222-5232.
12. Sarangerel, K.; Ganzorig, C.; Fujihira, M.; Sakomura, M.; Ueda, K., Influence of the
Work Function of Chemically Modified Indium–Tin–Oxide Electrodes on the
Open-circuit Voltage of Heterojunction Photovoltaic Cells. Chemistry Letters 2008, 37, (7), 778-
779.
79

13. Mutolo, K. l.; Mayo, E. I.; Rand, B. P.; Forrest, S. R.; Thompson, M. E., Enhanced open-
circuit voltage in subphthalocyanine/C-60 organic photovoltaic cells. Journal of American
Chemical Society 2006, 128, (25), 8108-8109.
14. Vandewal, K.; Tvingstedt, K.; Gadisa, A.; Inganas, O.; Manca, J. V., On the origin of the
open-circuit voltage of polymer-fullerene solar cells. Nat Mater 2009, 8, (11), 904-909.
15. Vandewal, K.; Widmer, J.; Heumueller, T.; Brabec, C. J.; McGehee, M. D.; Leo, K.;
Riede, M.; Salleo, A., Increased Open-Circuit Voltage of Organic Solar Cells by Reduced
Donor-Acceptor Interface Area. Advanced Materials 2014, 26, (23), 3839-3843.
16. Vandewal, K.; Oosterbaan, W. D.; Bertho, S.; Vrindts, V.; Gadisa, A.; Lutsen, L.;
Vanderzande, D.; Manca, J. V., Varying polymer crystallinity in nanofiber poly(3-
alkylthiophene): PCBM solar cells: Influence on charge-transfer state energy and open-circuit
voltage. Applied Physics Letters 2009, 95, (12), 123303.
17. Graham, K. R.; Erwin, P.; Nordlund, D.; Vandewal, K.; Li, R.; Ngongang Ndjawa, G. O.;
Hoke, E. T.; Salleo, A.; Thompson, M. E.; McGehee, M. D.; Amassian, A., Re-evaluating the
Role of Sterics and Electronic Coupling in Determining the Open-Circuit Voltage of Organic
Solar Cells. Advanced Materials 2013, 25, (42), 6076-6082.
18. Vandewal, K.; Tvingstedt, K.; Manca, J. V.; Ingana; x; s, O., Charge-Transfer States and
Upper Limit of the Open-Circuit Voltage in Polymer:Fullerene Organic Solar Cells. Selected
Topics in Quantum Electronics, IEEE Journal of 2010, 16, (6), 1676-1684.
19. Lee, J.; Vandewal, K.; Yost, S. R.; Bahlke, M. E.; Goris, L.; Baldo, M. A.; Manca, J. V.;
Voorhis, T. V., Charge Transfer State Versus Hot Exciton Dissociation in Polymer−Fullerene
Blended Solar Cells. Journal of the American Chemical Society 2010, 132, (34), 11878-11880.
20. Venediktov, E. A.; Tulikova, E. Y., Kinetics of tetracene oxidation with singlet molecular
oxygen: Dependence on the physicochemical properties of the solvent. Kinetics and Catalysis
2015, 56, (1), 49-55.

80

Chapter 4. Morphology of the D/A Interface in PHJs

4.1 Introduction
As discussed in section 1.3, two of the key steps in photocurrent generation are charge
transfer (CT) and charge separation (CS). In these processes excitons, bound hole-electron pairs,
are converted into free charge carriers in the bulk.
1-3
In OPVs, which are based on a
heterojunction structure, CT takes place at the interfacial region of the device where the donor
and acceptor meet. This area is referred to as the donor-acceptor (D/A) interface and the
structure and morphology of this region strongly impacts the kinetics of the CT and CS
processes.
4
The critical role of this interfacial region in current generation has been previously
identified and several D/A interface structures have been employed in OPVs. The first and
simplest structure made is now termed a planar heterojunction (PHJ). These devices are based
on neat, discrete, and sequential donor and acceptor layers in which the D/A interface is the two
dimensional area between the layers. This structure generates devices with low bulk resistances
but suffers from low currents.
5, 6
In contrast to this is the bulk heterojunction (BHJ) which is
composed of one mixed layer composed of both donor and acceptor materials with the D/A
interface existing throughout the bulk film. These BHJ devices have shown that an increase in
the depth of the D/A interface is correlated to an increase in photocurrent, and is the primary
benefit seen in the BHJs.
7-11
It is reasoned that the increased D/A region allows for more photon
absorption events to be within the exciton diffusion length of a CT site as well as providing
more sites for CT events to occur.
12, 13
In addition, these mixed regions provide transient
delocalization effects that allow for excitations to be funneled away from trap states.
14
This has
lead to an increasing amount of research into merging the advantages of BHJs and PHJs.
15-19
In
81

these device structures termed planar-mixed heterojunctions (PMHJs) a mixed region is
intentionally created between two neat phases.
20
This has been shown as an effective way to
boost photocurrent versus a strict PHJ while maintaining high motilities. However, despite our
growing knowledge about the effects of D/A interfacial region,
21
there remains little direct
measurement on the structure of D/A interface.
22
Typically, when discussing PHJs, the layers
are depicted with discrete, abrupt interfaces between two neat films. It is obvious that this is an
oversimplification and there must be some roughness to this interface but to what extent and its
effects have not been shown. This work will use the technique of neutron reflectometry to
directly probe the structure of these types of interfaces and using that knowledge determines the
parameters that dictate the resulting structure of this region in OPVs.

4.2 Experimental
The materials DBP, DIP, were obtained from the Brütting group. bDIP and BDP-Por
were synthesized by Mr. Chen and Ms. Golden respectively according to literatures
procedures.
23, 24
DBP, bDIP, DIP, CuPc 99% (Aldrich) and C
60
99.5% (MTR Limited), were
purified by vacuum thermal gradient sublimation in a three zone oven. DPSQ due to stability
concerns was synthesized according to literature and purified by column chromatography.
25

BDP-Por and DPSQ were spin-coated into films at a rate of 4000 RPMs from chloroform
solutions with a concentration 1 mg/mL. All other materials were processed into films by
vacuum thermal evaporation at a rate of 2 Å/s. Neutron reflectometry measurements were
performed at the National Institute of Standards and Technology and taken on their NG7
horizontal neutron reflectometer (NR). The NR samples were prepared and stored in a nitrogen
atmosphere until just before NR measurement and then tested in air. Programs from the
82

reflpacksuite were used for elements of the data reduction and analysis.
26
Films made for AFM
analysis of surface roughness were fabricated on UV ozone cleaned silicon wafers and
performed on an Agilent 5420 AFM and analyzed with Peaks software. Roughness was
quantified by the 99% of the maximum height of the film (S
Z
).

4.3 Neutron Reflectometry
Reflectometry is the most direct technique for determining the structure of multilayer thin
film stacks. It is particularly unique in its ability to investigate the nature of buried interfaces,
and has become an established way to probe the nanometer scale structures of organic thin films
over the last decade.
27-29
Neutron reflectivity (NR) measures the scattering length density (SLD)
of a film stack as a function of depth. The SLD is the neutron scattering parameter related
directly to refractive index in zero order dynamical theory of diffraction and it represents the
strength of interaction between a given nucleus and the neutron beam.
30
By correlating materials
with their theoretical SLD parameters, the film composition as a function of depth in a film can
be calculated. In contrast to optical and X-ray reflectometry which measure the refractive index
and are therefore sensitive to the electrons in the film, NR studies the interaction with the nuclei
and consequently is more sensitive to and gives more contrast between the lighter elements such
as hydrogen, nitrogen, and carbon.
31
This gives NR the ability to elucidate the fine structure of
film stacks made with different organic materials, particularly the ubiquitously used acceptor
C
60
, and can shed light on these interfacial regions of organic films.
NR operates by impinging neutrons onto a film stack at glancing angles near the critical
angle and measures the resulting reflection. Figure 4.1 shows how neutrons that reflect off of the
83

first interface will have traveled a shorter distance than those that reflect off of a buried interface.
Because of this difference in path length the waves will either constructively or destructively
interfere as a function of the incident angle. The signal will therefore oscillate in intensity when
the angle of the neutrons is changed as the signal undergoes constructive and destructive
interference. By analyzing the resulting spectra the structure of the film stack under study can be
inferred. It is difficult, however, to directly invert the reflectometry spectrum to an SLD profile

Figure 4.1 The general concept behind neutron reflectivity is described here. The incident
beam strikes the first interface at point A. The beam is split into refracted and reflected parts
in a proportion determined by the change in SLD at the interface. The refracted beam will
then be reflected when it strikes the second interface and be refracted again as it exits the
film, being now parallel to the originally reflected component. However the phase of the
second beam has been changed based on the additional path traveled and will interfere with
the reflected wave. The nature of this interference will change based on the incident angle
of the beam. The result is “fringes” of constructive and destructive interference with a width
of 2π.
84

directly, due to the fact that there is not one unique solution to each reflectometry spectrum.
32

Instead, a model SLD profile is created and a theoretical reflectometry spectrum is calculated.
The calculated spectrum is then fitted to the measured spectrum in a least squares fashion by
optimizing the film parameters of thickness, interfacial roughness, and SLD. By grounding these
parameters to expected values, a physically meaningful model can be converged upon.
When building the model stack for generating simulated spectra, the bulk SLD values of
silicon and silicon dioxide were used and not allowed to vary. The depth and roughness of the
SiO
2
layer was allowed to vary as these can change with time after etching. The thicknesses of
the organic layers were measured during the preparation process with a quartz crystal monitor
(QCM) and used as a starting point for the fit. Estimate SLDs of organic materials were
calculated using equation 4.1.

4.1
In equation 1, N
a
is Avogadro’s number, ρ is the density, MM is the relative molecular mass
and b
ci
is the bound coherent scattering length. These values were used as a starting point for the
model and allowed to vary slightly during the fitting process. Fits were evaluated by minimizing
the χ
2
value in a least squares analysis to the data. This was balanced against the number of
layers used in the model stack to limit the number of fitting parameters. The depth of the
interfacial regions is defined in the fitting software as the interfacial roughness of the lower
layer.
85

When discussing the interfacial depth of these film stacks it is important to note that there
is a spectrum of morphologies a film can take which will still lead to the same interfacial depth
measurement by NR. At one extreme of the spectrum figure 4.2 (top) a blended region is formed
between the two neat layers that has a flat and planar shape. When a film of this structure is
measured by NR it will measure the interfacial depth of the blended region that is of mixed
composition. At the other extreme the film can take on the morphology of an interdigitated but
abrupt interface between the two neat layers figure 4.2 (bottom). In this film type NR will report
the maximum height of these features. The NR spectrum reports data as an average over the
sampled area and so will not be able to differentiate these two cases from each other.

Figure 4.2 The two extremes in morphology that can give rise to interfacial depth are
depicted. At one extreme the interface is flat but contains a mixed region of some gradient
(top). On the other extreme there is no blending gradient but the interface contains substantial
roughness giving rise to a mixed depth when averaged over the substrate (Bottom). These
two distinct morphologies of the interface are indistinguishable by NR.
86

4.4 Results and Discussion
4.4.1 Interfacial Depth in PHJs
The first film stack studied is of structure Si/SiO
2
/CuPc/C
60
that corresponds to the standard
OPV. The NR spectrum of the stack is shown in figure 4.3 where the reflectometry spectrum
(Top, Black squares) is plotted as reflectance versus Q.

4.2
Where θ is the incident angle of the radiation and is its wavelength. A model film stack
was built and optimized (bottom) and shows the SLD as a function of depth from the top
interface. The expected reflectance of this model is calculated and plotted with the measured
reflectance (Top, red line) to show the goodness of fit which is calculated as a χ
2
residual of 1.25.
The resulting model stack shown in the SLD profile has the structure of Si/CuPc(18nm)/C
60

(48nm) with an interfacial region of 8.1nm. This structure is consistent with the intended
structure of Si/CuPc(18 nm)/C
60
(40 nm) where the C
60
region has expanded because of a lower
density. This lower density is due to being deposited onto a more amorphous surface, in this case
CuPc, than when it was calibrated, which was done on a silicon substrate. As hypothesized there
is a mixed region of significant depth of 8.1 nm formed between the two layers. This natively
formed intermixed region is likely to have a substantial impact on the resulting device. To
understand the nature of this layer, its impact on device performance, and how it forms the same
materials were deposited in the reverse order of Si/SiO
2
/C
60
/CuPc shown in figure 4.4. The
fitted SLD profile (figure 4.4 bottom) gives a structure of
87

0.00 0.05 0.10 0.15
1E-7
1E-6
1E-5
1E-4
1E-3
0.01
0.1
1

Reflectance
Q (Å
-1
)

Figure 4.3 The measured NR spectrum (above) for a Si/SiO
2
/CuPC/C
60
film is shown
(black squares). A model stack (below) was constructed that is close to 400 Å of C
60
and
240 Å of CuPc as measured by the QCM. The calculated spectrum is mapped onto the data
(above, red line) and is shown to match the data well. The goodness of fit is represented by
a χ
2
value 1.99.
88

0.00 0.05 0.10 0.15
1E-7
1E-6
1E-5
1E-4
1E-3
0.01
0.1
1

Reflectance
Q (Å
-1
)
Figure 4.4 The measured NR spectrum (above) for a Si/SiO
2
/C
60
/CuPc film is shown
(black squares). A model stack (below) was constructed that is close to 40 nm of C
60
and
30 nm of CuPc, as measured by the QCM, with an interfacial region of 5 nm. The
calculated spectrum is mapped onto the data (above, red line) and is shown to match the
data well. The goodness of fit is represented by a χ
2
value 1.53. It can be
89

Si/SiO
2
/C
60
(40 nm)/CuPc(30 nm) which is within error of what was reported by the QCM and
indicates that there is an intermixed region of 4.0 nm. This intermixed region is less than half as
thick as in the standard structure and much more closely resembles the discrete interface
generally assumed. It is also notable that the C
60
layer now matches well with the QCM
measurement indicating that it is indeed denser when deposited on a silicon substrate instead of
an amorphous organic.
A large number of devices using these materials have been published in a variety of
structures and so some conclusions on the impact of this intermixed region can be correlated to
device performance.
33, 34
It has been shown previously that the short circuit current can be
enhanced substantially for some materials when making devices as a PMHJ, where there is an
intentional introduction of an intermixed region.
35
In particular, a CuPc/C
60
PMHJ OPV has
been reported with J
SC
enhancements of up to 50%. It is therefore reasonable to assume that
reducing a mixed region in the inverted structure would negatively impact the JSC of the device.
In fact this result has already been borne out by the inverted C
60
/CuPc devices reported by
Thompson et. al.
36
in which the J
SC
has been significantly decreased by almost 25% from the
standard structure with the same thicknesses. As mentioned, this correlation between intermixed
region and J
SC
has been discussed extensively in PMHJ structured devices but what this suggests
is that these devices are just extending mixed regions natively created in PHJs structures and that
this needs to be considered when analyzing PHJ devices.
23

To determine the nature and effects of these spontaneously generated mixed regions, a
series of different material pairs used in OPVs was investigated. The results of these
experiments can be seen in table 4.1. It can be observed from this short study of materials that
the depth of the interfacial region can vary widely and is order dependent in all cases studied.
90

What also stands out is that when C
60
is deposited first the depth of the resulting interfacial
region seems to be independent of the material deposited on top.

4.4.2 Surface Roughness and Interfacial Depth
A property that may contribute to the final interfacial depth of a stack is the surface
roughness after deposition of the first layer in that stack. A surface with large hills and valleys in
the first layer that are filled in during the deposition of the next layer would be seen in the SLD
depth profile as a mixed layer when averaged over the entire film. By this mechanism no true
blending is needed to generate the interfacial depth measured by NR. To characterize the surface
roughness a neat film of each material studied was fabricated on a silicon wafer and the surface
structure was measured by atomic force microscopy (AFM). When investigating the AFM data
it is assumed here that the 95% percentile peak to trough (S
Z
) is the best roughness parameter for
comparison to interfacial depths of the resultant stacks as this would be similar to the depth of
the film that would have mixed composition given no mixing.
Si/Donor/C
60
Si/C
60
/Donor
DPSQ* 30 Å -
bDIP 140 Å 43 Å
CuPc 81 Å 47 Å
DBP 19 Å 47 Å
DIP 49 Å -

Table 4.1 Listed here are the depths of interfacial regions between studied donors and C
60

for standard and inverted structures. It can be seen here that the region ranges from 19 to
140 Å in depth when the donor is deposited first but has only one value when C
60
is the
first material deposited. All materials were deposited by VTE at 2 Å/s except DPSQ
which was spin coated at 3000 RPM from CHCl
3
.
91

Shown in figure 4.5 is an AFM image of a neat C
60
film with the full roughness analysis
parameters and shows the S
Z
of the C
60
film is 4.8 nm. The S
Z
roughness parameters for all the
materials studied appear in table 4.2. As noted before (table 4.1), the stacks with the C
60
as the
lower layer consistently had 4-5 nm thick interface region regardless of the material deposited on
top. This correlates well with the with the 4.8 nm roughness of the C
60
film and implies that the
interfacial depth in a film stack was defined by the roughness of the C
60
film prior to the second
layer’s deposition. This indicates that there is little true mixing at the interface and that the

interfacial depth of the NR more like that depicted in figure 4.2 (bottom). When measuring DBP
and DPSQ the AFM determined roughness correlates well with the interfacial depth measured by
NR, with AFM Vs. NR measurements of 12 nm Vs. 19 nm and 40 nm Vs. 30 nm. Similarly to

Figure 4.5 A 25 μm
2
AFM picture of a C60 film deposited on silicon (left). The analysis
statistics of this picture were calculated and are shown (right). Of particular interest here is
the Sz statistic which measures 95% maximum height displacement. This is the
measurement that will be the closest to the interfacial roughness of the film stack, assuming
the interface does not change during the deposition of the second layer.
92

the C
60
case, this implies that the surface roughness determines the interfacial region and that
little mixing takes place. However, for the materials CuPc and bDIP, there is no clear correlation
between the P
Z
roughness and the NR calculated interfacial depth that in the bilayer. In these
cases the surface roughness of the bottom film is significantly smaller than the NR calculated
interfacial depth. It is therefore reasonable to assume that in the case of these material systems
additional mixing takes place during the deposition of the second layer leading to a structure
closer to that depicted in figure 4.2 (top). These results seem to imply that not only is the
interfacial depth material dependant but the limiting factor that determines the interfacial region
changes with material system

4.4.3 Structure of bDIP/C
60
Film Stack
As can be seen in table 4.1 the interface of the Si/bDIP/C
60
film stack is categorically
different than the other organic stacks studied here. Not only is the interfacial depth of this film
nearly twice that of the next largest interlayer (14 nm vs. 8 nm), but the structure of the layer is
Material S
Z
(nm)
Interfacial depth (nm)
DPSQ* 4.0
3.0
bDIP 3.7
14.0
CuPc 6.1
8.1
DBP 1.2
1.9
C60 4.8
4.5*
DIP 4.1
4.9

Table 4.2 The S
Z
parameters of all materials studied here are presented along with the
interfacial depth of the bilayers of structure (material/C
60
). *The C
60
interfacial depth is an
average measurement of all films of the structure (C
60
/material).
93

unique among those in this study. In all of the other organic stacks studied in section 4.4.1 the
interfacial region takes on a structure where there is a composition gradient that follows a
Gaussian distribution between the two neat films of donor and acceptor (figure 4.3 and 4.4). In
stark contrast to that is the case of the Si/bDIP/C
60
stack shown in figure 4.6. In this film stack
30 nm of bDIP is capped with 30 nm of C
60
as measured by QCM. The reflectance spectrum is
plotted in figure 4.6 and was fitted to the SLD profile with a χ
2
value of 1.75. Here it can be seen
that the normal Gaussian composition gradient pauses at a fixed blending ratio for several
nanometers (7-8 nm), forming an extended region of constant blend ratio between the two neat
layers. This unique structure to the D/A interface indicates that more is happening than the filling
of pre-existing hills and valleys as is indicated for many of the materials in section 4.4.2. To
investigate the origin of this unique structure, the reciprocal film stack (Si/C
60
/bDIP) was
fabricated, where again the intent was to make discrete layers of 30 nm but in the reverse order.
The result of the NR on the film is shown in figure 4.7 and is fitted to an SLD profile with a χ
2

value of 7.57. The SLD profile shows a Gaussian blending gradient between the neat layers and
shows the shallow (~4 nm) D/A interface region which was found in all the film stacks where
C
60
is the first layer. This profile is very close to the structure intended in the deposition scheme
where there are two neat layers with a fairly discrete interface. The film stack is noticeably
absent of any of the unique features shown in figure 4.6 having no extended zone of constant
94

-100 0 100 200 300 400 500 600 700
0.0
1.0x10
-6
2.0x10
-6
3.0x10
-6
4.0x10
-6
5.0x10
-6
6.0x10
-6
Mixed

SLD (Å
-2
)
Depth (Å)
bDIP
C
60
0.00 0.05 0.10 0.15
1E-7
1E-6
1E-5
1E-4
1E-3
0.01
0.1
1


= 1.75

Reflectance
Q (Å
-1
)
Figure 4.6 The reflectivity spectrum (bottom) and SLD profile (top) of the Si/bDIP/C
60
stack are
shown here. The unique interfacial region in this film stack can be seen with an extended region
of constant blend ratio forming in situ.
95

0 200 400 600 800 1000
0.0
1.0x10
-6
2.0x10
-6
3.0x10
-6
4.0x10
-6
5.0x10
-6
6.0x10
-6

SLD (Å
-2
)
Depth (Å)

bDIP
C
60
0.00 0.05 0.10 0.15
1E-7
1E-6
1E-5
1E-4
1E-3
0.01
0.1
1


= 7.57

Reflectance
Q (Å
-1
)
Figure 4.7 The reflectivity spectrum (bottom) and SLD profile (top) of Si/C
60
/bDIP stack. It can
be seen that when the stack is constructed in reverse order the interfacial region takes on a more
traditional structure.
96

blend ratio. These findings lead to the conclusion that both the order and materials of a bilayer
stack are important in determining the final structure of an organic film stack.
As seen in figure 4.6 not all film stacks form the structure intended during deposition. To
help elucidate the structures produced from different deposition schemes, as well as the accuracy
and fidelity of the SLD profile modeling software, a series of bDIP/C
60
film stacks were
fabricated focusing on introducing blended layers in the device. A gradient blend film stack was
fabricated using as rate ramp where at the start of the deposition the rater ratio of bDIP:C
60
was
2:0 Å/s. The ratio is linearly changed so that after two hundred seconds the rate ration is 1:1 and
after 400 seconds is 0:2. The result is a film stack of with a linear composition gradient from
pure bDIP at the silicon surface to pure C
60
at the top surface. The result of the NR experiment
on this film stack is shown in figure 4.8 and is fitted with a strong χ
2
value of 3.81. The SLD
profile generated shows a film composition very close to that which was expected from the QCM
measurements with the highest SLD at the top interface, close to that of pure C
60
, which linearly
decreases as bDIP ration increases. The SLD reaches that of pure bDIP at the bottom of film
stack right before it jumps up at the SiO
2
layer on the wafer at a depth of ~950 Å. This film
shows two things, first that the resulting SLD of a blend is equal to the weight average of the
SLDs of the components, and second that when depositing blends of bDIP/C
60
there is little
rearrangement in the film. The film also shows that the volume of the blended layers is greater
than of the isolated neat materials, leading to film stacks that are higher than would be
anticipated. In the case of this film stack 400 Å of each material was deposited when measured
individually but leading to a film stack almost 1000 Å high.
This gradient blend film stack is contrasted with a flat blended stack where the volume
ratio of 1:1 (bDIP:C
60
) is maintained across the entire film. The NR of the film stack is shown in
97

figure 4.9 and reflectance spectrum is fitted with a χ
2
value of 4.93. This plot shows a
distinctively different pattern from the gradient blend film stack having a flat SLD across the
entire film which is again generally what would be expected deposition scheme. The profile
varies from the flat SLD with some deviations early in the deposition at the film depth of 900-
1150 Å. To investigate the origins of these oscillations in SLD the exact rate ratios that occurred
during deposition where extracted from the deposition software. Figure 4.Z shows the blend
ratio of C
60
to bDIP as a function of depth in the film as measured by the QCM during deposition
of the film. Values above one indicate that the film is rich in C
60
and values lower than one
indicates where the film is rich in bDIP. As state before the QCMs measure thinner films than
occur in blended film so for comparison proposes this graph has been density corrected to map
onto the SLD profile. It can be seen that the blend ratio measured by the QCM in figure 4.10
oscillates from C
60
poor to C
60
rich early in the deposition before stabilizing at the desired 1:1
ratio later in the process. Comparing this blend ratio graph to the SLD profile in figure 4.9 it can
be seen that the profile has accurately measured the oscillations in the blend ratio that occurred
early in the deposition, showing a reduction of SLD as the C
60
concentration decreases and
increase in SLD with increasing C
60
concentration, finally leveling out as the blend ratio does.
The study of this film highlights the level of fidelity achieved from the NR and the SLD fitting
software, giving a high level of confidence when applying the SLD profiles of other stacks.
98

0 100 200 300 400 500 600 700 800 900 1000
0.0
1.0x10
-6
2.0x10
-6
3.0x10
-6
4.0x10
-6
5.0x10
-6
6.0x10
-6

SLD (Å
-2
)
Depth (Å)
bDIP C
60
0.00 0.05 0.10 0.15
1E-7
1E-6
1E-5
1E-4
1E-3
0.01
0.1
1


= 3.81

Reflectance
Q (Å
-1
)
Figure 4.8 The reflectivity spectrum (bottom) and SLD profile (top) of Si/C
60
:bDIP gradient
blend film is shown here. At the surface, or 0 Å depth, the film is entirely composed of C
60
. As
the depth increases the molecular ratio between the two components linearly shift toward bDIP
until at the silicon substrate the composition is entirely bDIP.
99

0 100 200 300 400 500 600 700 800 900 1000 1100 1200
0.0
1.0x10
-6
2.0x10
-6
3.0x10
-6
4.0x10
-6
5.0x10
-6
6.0x10
-6

SLD (Å
-2
)
Depth (Å)
bDIP:C
60
0.00 0.05 0.10 0.15
1E-7
1E-6
1E-5
1E-4
1E-3
0.01
0.1
1


= 4.93

Reflectance
Q (Å
-1
)
Figure 4.9 The reflectivity spectrum (bottom) and SLD profile (top) of Si/C
60
:bDIP flat blend
film is shown here. The profile shows an oscillating SLD at the start of the layer which dampens
out to a constant value close to the average SLD of the components.
100

The final film stacks studied were of the PMHJ structure where a neat layer of bDIP and
C
60
were separated by a mixed layer of either a gradient blend or a 1:1 flat blend. The NR data
for these two film stacks are shown in Figures 4.11 and 4.12 respectively. The NR of the two
films were fitted with SLD profiles with the good fitting parameters of 1.35 and 1.75,
respectively, and show structures that are very close to that measured by the QCMs. From these
SLD profiles it is clear that these film stacks have structures very similar to what would be
assumed from the QCM data. This leaves the simple bDIP/C
60
film stack as the only structure
that undergoes a morphological change during or after construction of the film stack.
0 200 400 600 800 1000 1200
0.0
0.5
1.0
1.5
2.0

C
60
: bDIP Ratio
Film Depth

Figure 4.10 The blend ratio for the film stack shown in figure 4.9 is shown. By comparing
the blend ratio to the SLD profile it can be seen that the two correlate closely with the
simulation able to measure the oscillations created early in the deposition.
101

-100 0 100 200 300 400 500 600 700 800 900 1000
0.0
1.0x10
-6
2.0x10
-6
3.0x10
-6
4.0x10
-6
5.0x10
-6
6.0x10
-6

SLD (Å
-2
)
Depth (Å)
bDIP
C
60
0.00 0.05 0.10 0.15
1E-7
1E-6
1E-5
1E-4
1E-3
0.01
0.1
1


= 1.35

Reflectance
Q (Å
-1
)
Figure 4.11 The reflectivity spectrum (bottom) and SLD profile (top) of a Si/bDIP/gradient/C
60

film stack is shown here. The SLD trace conforms to the QCM measured blend ratios
102

-100 0 100 200 300 400 500 600 700 800 900
0.0
1.0x10
-6
2.0x10
-6
3.0x10
-6
4.0x10
-6
5.0x10
-6
6.0x10
-6

SLD (Å
-2
)
Depth (Å)
bDIP C
60
0.00 0.05 0.10 0.15
1E-7
1E-6
1E-5
1E-4
1E-3
0.01
0.1
1


= 1.75

Reflectance
Q (Å
-1
)
Figure 4.12 The reflectivity spectrum (bottom) and SLD profile (top) of a Si/bDIP/Blend/C
60

film stack is shown here. The SLD trace conforms to the QCM measured blend ratios showing
interfacial regions between the C
60
/blend and blend/bDIP of 20 Å and 8 Å respectively.

103

4.4.4 PMHJ bDIP/C
60
Devices
In order to determine the effect of the unique film structures present in the bDIP/ C
60
film
stack on the performance of the OPV device based on it, a series of PMHJ devices where
synthesized. These devices were prepared with a mixed bDIP: C
60
layer with a variety of
blending ratios, between neat bDIP and C
60
layers. Three PMHJ devices were made of structure
ITO/bDIP 35 nm/ X:1 C
60
:bDIP 10 nm/ C
60
35 nm/BCP/Al where the blending ratios used were
1:1, 2:1, and 3:1 C
60
:bDIP. These devices have an active region similar to that of the film stack
depicted in figure 4.12 but with varying blending ratios in the mixed region. The mixed region
was deposited at a total rate of 0.4 Å/s so that the blending ratio could be more easily controlled.
All other layers were deposited at the standard 2 Å/s. The performances of these devices are
summarized in figure 4.13 and table 4.3. It can be seen here that the performance of the PMHJ
devices improved with increasing C
60
content of the blended film with the 1:3 bDIP: C
60
mixed
region giving the best device with an efficiency of 3.76%. It can be inferred from the poor
rectification of the 1:1 and 1:2 devices (figure 4.13 top) that the mixed layer in these devices
must have a reduced carrier mobility leading to the reduced performance. The performance of
the natural bilayer device is plotted on the same graph and it can be seen that it most closely
matches the performance of the 3:1 bend ratio indicating that is the ratio achieved in the natural
mixing taking place in the bilayer film stack. This correlates well with the SLD profile fitted to
the Si/bDIP/ C
60
film stack (figure 4.6). SLD of the mixed region in that stack is around
4.1E
-5
Å
-2
which translates to a mix ratio of around 1:3 given that the SLD of a mixed film is the
weight average of its components. The EQE plot in figure 4.13 bottom shows the photo response
104

-1.0 -0.5 0.0 0.5 1.0
-10
-5
0
5
10
1:1 bDIP:C
60
1:2 bDIP:C
60
1:3 bDIP:C
60
Bilayer
(mA/cm
2
)
Voltage (V)

Current density

400 500 600 700 800
0
5
10
15
20
25
30
35
40
45
50
55
1:1 bDIP:C
60
1:2 bDIP:C
60
1:3 bDIP:C
60
Bilayer
Quantum Efficiency (%)
Wavelength (nm)

Figure 4.13 (top) Representative I-V curves under light (solid lines) and dark (dashed lines)
conditions. (bottom) EQEs of the different blend ratio PMHJ devices show increasing efficiency
in the bDIP region.
105

of the PMHJ devices. In it we can see an increase in response in the bDIP region of 550 nm to
780 nm as the C
60
concentration increases. This is somewhat counterintuitive as the total
quantity of bDIP in the device is decreasing as the bDIP response increases. It therefore must be
due to an increase in the CT efficiency in the C
60
rich mixed layer. In addition, the presence of a
red shoulder around 750 nm in the bilayer and 1:3 devices, which is absent in the other devices,
reinforces the similarity between the 1:3 PMHJ and bilayer devices.
In several studies the PMHJ devices have been shown to be the optimal architecture,
outperforming both the BHJ and PHJ.
20, 37-39
This enhancement for PHJs has been determined to
be because of the increased interfacial area between donor and acceptor leading to increased
photocurrent.
4, 40, 41
We have shown here that the device base on a bDIP/ C
60
active region
adopts this type of structure natively when a sequential deposition of the components is
attempted. It is therefore concluded that this unintended formation of a mixed region is
responsible for the surprisingly high performance reported for this device.

4.5 Conclusions
It has been shown here that bilayer film stack can vary significantly in their final
structure depending on the material system being used. In case where C
60
is the first layer in the
Ratio (bDIP:C
60
) J
SC
(mA/cm
2
)

V
OC
(V) FF (%)
1:1 7.38 0.79 0.40 2.33
1:2 7.04 0.82 0.43 2.49
1:3 8.42 0.82 0.55 3.76
Bilayer 8.70 0.79 0.55 3.77

Table 4.3 Summary of PV parameters for ratio-dependent PMHJ devices and the control
bilayer of structure: ITO/bDIP/C
60
/BCP/Al.
106

stack, the interfacial depth is defined by the roughness of the C
60
regardless of the composition
of the following layer. This indicates that there the interface is abrupt with no mixed layer but
with inter-digitations as shown in figure 4.2 bottom. In the other extreme, as is the case in the
bDIP/ C
60
stack, the final structure can have little correlation with the deposition scheme having
the ability to form extended interfacial layers of mixed composition. Most of the materials
studied here fall somewhere between these two extremes. Through analysis of these systems it is
postulated that to form one of these mixed region two criteria need to be present. First that the
sublimation temperature of the second material has to be in excess of the T
g
of the first layer, and
second that the forming of a mixture should have a negative free energy change due to
dissolution of the material with each other. Overall this study shows that the structure of bilayer
film stacks can vary widely depending on technique and composition and cannot necessarily be
assumed to be the discrete bilayer as is usually depicted in the cartoon models. Furthermore, it is
shown here that this deviation from the ideal structure of the OPV can have significant
implication on the performance of the device and need to be taken into account when drawing
conclusions.

107

4.6 Chapter 4 References

1. Kanai, K.; Akaike, K.; Koyasu, K.; Sakai, K.; Nishi, T.; Kamizuru, Y.; Nishi, T.; Ouchi,
Y.; Seki, K., Determination of electron affinity of electron accepting molecules. Applied Physics
A 2009, 95, (1), 309-313.
2. Marcus, R. A., On the Theory of Oxidation‐Reduction Reactions Involving Electron
Transfer. I. The Journal of Chemical Physics 1956, 24, (5), 966-978.
3. Onsager, L., Initial Recombination of Ions. Physical Review 1938, 54, (8), 554-557.
4. Khlyabich, P. P.; Burkhart, B.; Rudenko, A. E.; Thompson, B. C., Optimization and
simplification of polymer–fullerene solar cells through polymer and active layer design. Polymer
2013, 54, (20), 5267-5298.
5. Menke, S. M.; Holmes, R. J., Energy-Cascade Organic Photovoltaic Devices
Incorporating a Host–Guest Architecture. ACS Applied Materials & Interfaces 2015, 7, (4),
2912-2918.
6. Lin, C.-F.; Zhang, M.; Liu, S.-W.; Chiu, T.-L.; Lee, J.-H., High Photoelectric Conversion
Efficiency of Metal Phthalocyanine/Fullerene Heterojunction Photovoltaic Device. International
Journal of Molecular Sciences 2011, 12, (1), 476-505.
7. Treat, N. D.; Chabinyc, M. L., Phase Separation in Bulk Heterojunctions of
Semiconducting Polymers and Fullerenes for Photovoltaics. Annual Review of Physical
Chemistry 2014, 65, (1), 59-81.
8. Savoie, B. M.; Jackson, N. E.; Chen, L. X.; Marks, T. J.; Ratner, M. A., Mesoscopic
Features of Charge Generation in Organic Semiconductors. Accounts of Chemical Research
2014, 47, (11), 3385-3394.
9. Roncali, J., Molecular Bulk Heterojunctions: An Emerging Approach to Organic Solar
Cells. Accounts of Chemical Research 2009, 42, (11), 1719-1730.
10. Grancini, G.; Polli, D.; Fazzi, D.; Cabanillas-Gonzalez, J.; Cerullo, G.; Lanzani, G.,
Transient Absorption Imaging of P3HT:PCBM Photovoltaic Blend: Evidence For Interfacial
Charge Transfer State. The Journal of Physical Chemistry Letters 2011, 2, (9), 1099-1105.
11. Ryuzaki, S.; Onoe, J., Basic aspects for improving the energy conversion efficiency of
hetero-junction organic photovoltaic cells. 2013 2013.
12. Jackson, N. E.; Savoie, B. M.; Marks, T. J.; Chen, L. X.; Ratner, M. A., The Next
Breakthrough for Organic Photovoltaics? The Journal of Physical Chemistry Letters 2015, 6, (1),
77-84.
108

13. Collins, B. A.; Bokel, F. A.; DeLongchamp, D. M., Organic Photovoltaic Morphology. In
Organic Photovoltaics, Wiley-VCH Verlag GmbH & Co. KGaA: 2014; pp 377-420.
14. Sweetnam, S.; Graham, K. R.; Ngongang Ndjawa, G. O.; Heumüller, T.; Bartelt, J. A.;
Burke, T. M.; Li, W.; You, W.; Amassian, A.; McGehee, M. D., Characterization of the Polymer
Energy Landscape in Polymer:Fullerene Bulk Heterojunctions with Pure and Mixed Phases.
Journal of the American Chemical Society 2014, 136, (40), 14078-14088.
15. Kekuda, D.; Huang, J.-H.; Ho, K.-C.; Chu, C.-W., Modulation of Donor−Ac ceptor
Interface through Thermal Treatment for Efficient Bilayer Organic Solar Cells. The Journal of
Physical Chemistry C 2010, 114, (6), 2764-2768.
16. Peumans, P.; Uchida, S.; Forrest, S. R., Nature 2003, 425, 158.
17. Xue, J.; Uchida, S.; Rand, B. P.; Forrest, S. R., Appl. Phys. Lett. 2004, 84, 3013.
18. Law, M.; Greene, L. E.; Johnson, J. C.; Saykally, R.; Yang, P., Nat. Mater. 2005, 4, 455.
19. McGehee, M. D., Mater. Res. Soc. Bull. 2009, 34, 95.
20. Xue, J.; Rand, B. P.; Uchida, S.; Forrest, S. R., A Hybrid Planar–Mixed Molecular
Heterojunction Photovoltaic Cell. Advanced Materials 2005, 17, (1), 66-71.
21. Kendrick, M. J.; Neunzert, A.; Payne, M. M.; Purushothaman, B.; Rose, B. D.; Anthony,
J. E.; Haley, M. M.; Ostroverkhova, O., Formation of the Donor–Acceptor Charge-Transfer
Exciton and Its Contribution to Charge Photogeneration and Recombination in Small-Molecule
Bulk Heterojunctions. The Journal of Physical Chemistry C 2012, 116, (34), 18108-18116.
22. Pandey, R.; Holmes, R. J., Organic Photovoltaic Cells Based on Continuously Graded
Donor–Acceptor Heterojunctions. Selected Topics in Quantum Electronics, IEEE
Journal of 2010, 16, (6), 1537-1543.
23. Chen, J. J.; Conron, S. M.; Erwin, P.; Dimitriou, M.; McAlahney, K.; Thompson, M. E.,
High-Efficiency BODIPY-Based Organic Photovoltaics. ACS Applied Materials & Interfaces
2014.
24. Whited, M. T.; Djurovich, P. I.; Roberts, S. T.; Durrell, A. C.; Schlenker, C. W.;
Bradforth, S. E.; Thompson, M. E., Singlet and Triplet Excitation Management in a
Bichromophoric Near-Infrared-Phosphorescent BODIPY-Benzoporphyrin Platinum Complex.
Journal of the American Chemical Society 2010, 133, (1), 88-96.
25. Wang, S.; Mayo, E. I.; Perez, M. D.; Griffe, L.; Wei, G.; Djurovich, P. I.; Forrest, S. R.;
Thompson, M. E., High efficiency organic photovoltaic cells based on a vapor deposited
squaraine donor. Applied Physics Letters 2009, 94, (23), 233304-3.
26. Kienzie, P. A., O'Donovan, K.V., Anker, J. F., Berk, N. F., Majkrzak, C. F.
109

27. Majkrzak, C. F.; Carpenter, E.; Heinrich, F.; Berk, N. F., When beauty is only skin deep;
optimizing the sensitivity of specular neutron reflectivity for probing structure beneath the
surface of thin filmsa). Journal of Applied Physics 2011, 110, (10), -.
28. Thomas, R. K.; Penfold, J., Neutron and X-ray reflectometry of interfacial systems in
colloid and polymer chemistry. Current Opinion in Colloid & Interface Science 1996, 1, (1), 23-
33.
29. Penfold, J.; Thomas, R. K., Neutron reflectivity and small angle neutron scattering: An
introduction and perspective on recent progress. Current Opinion in Colloid & Interface Science
2014, 19, (3), 198-206.
30. Jacrot, B., The study of biological structures by neutron scattering from solution. Reports
on Progress in Physics 1976, 39, (10), 911.
31. Kirby, B. J.; Kienzle, P. A.; Maranville, B. B.; Berk, N. F.; Krycka, J.; Heinrich, F.;
Majkrzak, C. F., Phase-sensitive specular neutron reflectometry for imaging the nanometer scale
composition depth profile of thin-film materials. Current Opinion in Colloid & Interface Science
2012, 17, (1), 44-53.
32. Lovell, M. R.; Richardson, R. M., Analysis methods in neutron and X-ray reflectometry.
Current Opinion in Colloid & Interface Science 1999, 4, (3), 197-204.
33. Uchida, S.; Xue, J.; Rand, B. P.; Forrest, S. R., Organic small molecule solar cells with a
homogeneously mixed copper phthalocyanine: C
60
active layer. Applied Physics Letters 2004,
84, (21), 4218-4220.
34. Xue, J.; Uchida, S.; Rand, B. P.; Forrest, S. R., Asymmetric tandem organic photovoltaic
cells with hybrid planar-mixed molecular heterojunctions. Applied Physics Letters 2004, 85,
(23), 5757-5759.
35. Xue, J.; Uchida, S.; Rand, B. P.; Forrest, S. R., 4.2% efficient organic photovoltaic cells
with low series resistances. Applied Physics Letters 2004, 84, (16), 3013-3015.
36. Trinh, C.; Bakke, J. R.; Brennan, T. P.; Bent, S. F.; Navarro, F.; Bartynski, A.;
Thompson, M. E., Power losses in bilayer inverted small molecule organic solar cells. Applied
Physics Letters 2012, 101, (23), -.
37. Ojala, A.; Burckstummer, H.; Hwang, J.; Graf, K.; von Vacano, B.; Meerholz, K.; Erk,
P.; Wurthner, F., Planar, bulk and hybrid merocyanine/ C
60
heterojunction devices: a case study
on thin film morphology and photovoltaic performance. Journal of Materials Chemistry 2012,
22, (10), 4473-4482.
38. Peumans, P.; Uchida, S.; Forrest, S. R., Efficient bulk heterojunction photovoltaic cells
using small-molecular-weight organic thin films. Nature 2003, 425, (6954), 158-162.
110

39. Hiramoto, M.; Suemori, K.; Yokoyama, M., Photovoltaic Properties of
Ultramicrostructure-Controlled Organic Co-Deposited Films. Japanese Journal of Applied
Physics 2002, 41, (4S), 2763.
40. Walker, B.; Tamayo, A. B.; Dang, X.-D.; Zalar, P.; Seo, J. H.; Garcia, A.; Tantiwiwat,
M.; Nguyen, T.-Q., Nanoscale Phase Separation and High Photovoltaic Efficiency in Solution-
Processed, Small-Molecule Bulk Heterojunction Solar Cells. Advanced Functional Materials
2009, 19, (19), 3063-3069.
41. Schünemann, C.; Wynands, D.; Wilde, L.; Hein, M. P.; Pfützner, S.; Elschner, C.;
Eichhorn, K.-J.; Leo, K.; Riede, M., Phase separation analysis of bulk heterojunctions in small-
molecule organic solar cells using zinc-phthalocyanine and C
60
. Physical Review B 2012, 85,
(24), 245314.

111

Chapter 5. Multi-Chromophoric Arrays in OPVs

5.1 Introduction
As described in section 1.3.1.2, organic dye materials have a natural advantage over most
inorganic semiconductors as their absorbtivities are usually orders of magnitude greater, being as
high as 10
5
- 10
6
cm
-1
. This is because, in most organic dyes including the ones used here, the
primary absorption in the visible region is due to π to π* or n to π* transitions in the aromatic
systems.
1
While silicon, owes its absorption to classically forbidden transitions, these π to π*
excitations are fully allowed and as consequence are more intense.
2
However, a significant
limitation to this method of photon absorption is that these state to state transitions tend to be
rather narrow compared to the inorganic systems which absorb continuously above the band gap.
When organic dyes, which rely on these discrete state to state transitions with limited spectral
coverage, are made into films, they typically will have near unity absorption in some parts of the
solar spectrum while being effectively transparent in others. It is therefore difficult to have a
single chromophore cover the entire spectrum of harvestable light, making it necessary to find
innovative ways to broaden the absorption spectra of OPVs to further increase efficiencies. In
this report we investigate the use of two intense chromophores, a porphyrin and a boron-dipyrrin,
covalently coupled in a multi-chromophoric array to achieve broadband absorption. In the
present report we discuss the use of the BDP-Por array as a donor layer in OPVs and contrast the
properties of those devices with analogous OPVs prepared with donor layers consisting of a
physical mixture of PtTPBP and BDP.

112

5.2 Multi-Chromophoric Arrays
Porphyrins and closely related materials are ubiquitous in nature for light absorption and
have been incorporated into modern small molecule OPVs.
3-5
These compounds typically have
two intense absorption peaks in the visible part of the spectrum. One of these is a strong but
narrow absorption in the blue part of the spectrum called the Soret band and other absorption is
in the red and called the Q band (figure 5.2(b)). Porphyrin absorption peaks are strongly
absorbing in these parts of the solar spectrum, but do not absorb a significant part of the visible
spectrum in the gap between these two transitions. Devices using PtTPBP have been made and a
characteristic J-V and EQE of these devices are shown in figure 5.2(a).
6
These devices perform

PtTPBP: R = Ph; BDP-Por: R=BDP
BDP
Figure 5.1The structures of the porphyrin PtTPBP and the BDP-Por
multi-chromophoric array are shown.
113

400 500 600 700
0
5
10
15
20
25
30
35
40
0.0
0.2
0.4
0.6
0.8
1.0
EQE
Quantum Efficiency
Wavelength (nm)
-0.4 0.0 0.4 0.8
-3
-2
-1
0
1
2
3

(mA/cm
2
)
Voltage (V)
Current density
a
a)
Absorption (AU)
PtTPBP Abs.
C
60
Abs.
300 400 500 600 700
0.0
0.2
0.4
0.6
0.8
1.0
b)
BDP
PtTPBP
BDP-Por
Sum BDP
+ PtTPBP

Absorbance (a.u.)
Wavelength (nm)

Figure 5.2 (a) The absorption spectra of neat PtTPBP and C
60
films are plotted against the EQE
of the device made from these materials. (b) The absorbance spectra of PtTPBP, BDP1, and the
arithmetic sum of these two plotted against the absorption spectrum of BDP-Por. The composite
spectra overlays nicely with the spectrum of the BDP-Por, showing that there is little ground state
coupling in the multi-chromophoric array.
114

well, but the gap in the porphyrin’s absorption spectrum can be seen to negatively impact the J
SC
.
The spectral responsivity of this device, shown in figure 4.2 (a), illustrates this problem, showing
lower solar conversion efficiencies between 500 and 600 nm, figure 5.2 (a). Were it not for the
strong C
60
absorbance between 400-500 nm, the spectral response would be even worse than
shown here.
BODIPYs are promising dyes that have recently been incorporated into OPV’s with
much success due to their desirable electronic properties and a high molar adsorptivity.
7-9
The
standard BODIPY absorption falls directly between Soret and Q-bands of the PtTPBP
(figure 5.2 (b)). The complementary absorption bands of BDP and PtTPBP provided an
opportunity to make a multi-chromophoric array with a broad absorption which may allow for an
increase in photo-collection and therefore photocurrent and power conversion efficiency relative
to the PtTPBP based OPV. With this objective these moieties were covalently bound together
making BODIPY-PtTPBP (BDP-Por) shown in figure 5.1.
10
The absorption spectrum of the
array is nearly identical to the arithmetic sum of the spectra of the separate PtTPBP and BDP
chromophores, indicating that there is poor ground state coupling between the chromophores.
This new molecule has a broadened absorption spectrum, with a AM1.5G photon capture
percentage in solution 60% greater than that of PtTPBP alone.
10
The absorption spectra of the
BDP-Por film changes little from the solution, with the film spectrum showing only a small red-
shift of both the BODIPY absorption (23 nm) and the Q-band of the porphyrin (13 nm).
Additionally, the film spectral features are slightly broadened relative to those in the solution
spectrum, which is attributed to increased -orbital overlap in the film. This is similar to the film
spectrum of PtTPBP as it is also slightly red shifted and broadened slightly relative to the
solution spectrum.
115

In many multi-chromophoric arrays reported elsewhere, energy collected by antennae
chromophores and funneled to a central core, in a uni-directional manner.
11-16
There the exciton
is isolated from its nearest neighbors by the antennae shell and consequently the exciton cannot
be effectively transported. We have investigated the dynamics of the intramolecular energy
transfer processes between the porphyrin and BODIPY moieties in BDP-Por and found this not
to be the case with this system.
10
Direct excitation of the BDP chromophore to its singlet excited
state is followed by rapid singlet energy transfer (ST) to the central porphyrin, with a rate
constant k
ST
of 7.8 x 10
11
s
-1
. The platinum porphyrin then rapidly intersystem crosses to its
triplet excited state, with the k
ISC
of 2.5 x 10
12
s
-1
. The triplet excitation then equilibrates
between the porphyrin triplet (1.62 eV) and the nearly degenerate BODIPY triplet (1.64 eV) with
a rate constant k
TT
of 1.0 x 10
10
s
-1
and a K
eq
of 0.61. Emission from BDP-Por comes exclusively
from the porphyrin triplet with a measured rate of 3.9 x 10
3
s
-1
and a quantum efficiency of 0.17
in toluene solution. This rapid equilibration of the energy from the BODIPY antennae to the
porphyrin core is an important difference from much of the literature in multi-chromophoric
arrays, since the exciton is distributed over the entire array and not trapped at the core.

5.3 Experimental
10,10'-(1,3-phenylene)bis(5,5-difluoro-5H-dipyrrolo[1,2-c:2',1'-f][1,3,2]diazaborinin-4-ium-
5-uide) (BDP2). Isophthaloyl dichloride (1 g, 4.95 mmol) was dissolved in dry dichloromethane
(80 ml) under N
2
. Four equivalents of 2,4-Dimethyl-3-ethylpyrrole (2.43 g, 19.7 mmol) was
added and the flask was fitted with a condenser and refluxed for 3 to 4 hrs. N,N-
Diisopropylethylamine (6.87 ml, 39.4 mmol) was added at reflux. After 15 minutes, the mixture
was cooled to room temperature and boron trifluoride etherate (5.59g, 39.4 mmol) was added in
116

one portion. After one hour, the reaction was quenched with saturated Na
2
S
2
O
3
(50 mL), washed
with saturated NaHCO
3
(2  50 mL) and water (2  50 mL). The organic layer was removed,
dried over MgSO
4
, filtered and concentrated. The product was purified by re-crystallizing from
DCM and MeOH to give 1.73 g (yield = 68%). Final purification was accomplished by and
sublimation in a three zone oven with temperatures of 300°C, 275°C and 250°C.
1
H NMR
(CDCl
3
): δ 7.61 (t, 1H), 7.54 (d, 1H), 7.40 (s, 1H), 2.53 (s, 12H), 2.31 (q, 8H), 1.51 (s, 12H),
1.00 (t, 12H).
13
C NMR (CDCl
3
): δ 154.74, 139.65, 138.14, 136.36, 133.76, 131.97, 131.28,
128.90, 53.89, 17.63, 14.93, 13.10. Excitation λ max (CH 2Cl 2) 529 nm Emision λ max (CH 2Cl 2)
553 nm. MALDI m/z for C 40H 48B 2F 2N 4 Calculated 682.4 Found 682.7. CHN analysis: 69.88%
C 7.28% H 8.32% N, theoretical: 70.4% C 7.09% H 8.21% N

BODIPY ethylene bridged Hexenoporphyrin The starting ethylene bridged pyrrole, 4,7-
dihydro-4,7-ethano-2H-isoindole, was synthesized according to literature procedure
17
. The
ethylene bridged Hexenoporphyrin was made by following literature procedures,
18
yielding a
mixture of four diastereomers after re-crystallization with dichloromethane/diethyl ether. The
diastereomers were not separated and the mixture was used in the next step.
1
H-NMR (400 MHz,
CDCl
3
), all four diastereomers, δ 8.83–8.70 (m, 8H), 8.21 (d, 8H, J = 8.00 Hz), 7.11–7.02 (m,
8H), 6.92–6.80 (m, 8H), 6.48 (d, 8H, J = 8 Hz), 3.74–3.51 (m, 8H), 2.78 (s, 24H), 2.73–2.57 (m,
8H), 1.44 (d, 4H, J = 8 Hz), 0.85 (d, 4H, J = 8 Hz).

BODIPY Platinum Beznoporphyrin (BDP-Por) Platinum (II) chloride (90 mg, 0.338 mmol)
was added to dry, degassed benzonitrile (100 mL), and the mixture was heated with stirring
under N
2
at 100 ˚C for 20 minutes, until the platinum salts dissolved, turning the solution yellow.
117

The BODIPY ethylene bridged Hexenoporphyrin (90 mg, 0.047 mmol) was added as a solid and
the solution was heated to reflux with stirring for 3h 20 min, until product peaks ceased to
increase in intensity as monitored via UV-Vis. The reaction mixture was cooled to 0 ˚C, and the
solvent was removed by vacuum distillation at 70 ˚C. The solid residues were dissolved in
CH
2
Cl
2
and filtered to remove solids and the filtrate was dried via rotovap to yield a bright green
solid, which was filtered and washed with MeOH (3x 10 mL). The solids were purified by flash
column chromatography on silica, using a gradient from 1:1 Hexanes/CH
2
Cl
2
to 97.5:2.5 CH
2
Cl
2

/acetone, and the pure product was recovered as an olive green solution, which was further
purified by re-crystallization from CH
2
Cl
2
/diethyl (36 mg, 41%). The
1
H-NMR matched
literature spectra (400 MHz, CDCl
3
) δ 8.55 (d, 8H, J = 8 Hz), 8.05 (d, 8H, J = 8 Hz), 7.44–7.31
(m, 8H), 7.24-7.26 (m, 8H), 7.17 (d, 8H, J = 4 Hz), 6.5 (d, 8H, J = 4 Hz), 2.79 (s, 24H).

OPV Preparation and Testing. BDP1 was synthesized according to literature.
19
C
60
(MTR
Unlimited), 2,9-dimethyl-1-4,7-diphenyl-1,10-phenanthroline (BCP) (Aldrich), were purified by
thermal gradient sublimation in vacuum prior to use. Aluminum (99.999% pure, Alfa Aesar)
was used as received and evaporated through a shadow mask to form 2 mm width striped
cathodes. Photovoltaic cells were fabricated on patterned indium tin oxide (ITO)-coated glass
substrates that were solvent cleaned and baked with UV ozone for 10 minutes. Films of BDP-
Por, Por+BDP1, and Por+BDP2, were made using a spin-coater operated at 4000 rpm for 40
seconds. The remaining materials were grown by vacuum thermal evaporation at the following
rates: C
60
(2 Ås
-1
), BCP (1 Ås
-1
), and Al (2 Ås
-1
). Current-voltage characteristics of the cells
were measured in the dark and under simulated AM1.5G solar illumination conditions (Oriel
Instruments) using a Keithley 2420 3A Source Meter. Incident power was adjusted using a
118

calibrated Si photodiode to match 1 sun intensity (100 mWcm
-2
), and spectral response was
measured using a Newport-Oriel monochromatic light source. Spectral mismatch was calculated
and used to correct the measured efficiencies following standard procedures.
20

5.4 Results and Discussion
5.4.1 BDP-Por Devices
BDP-Por is not stable to vacuum thermal evaporation, so devices of composition
ITO/BDP-Por(X nm)/C
60
(40 nm)/BCP (10nm)/Al were made with the BDP-Por layer deposited
by spin coating from a chloroform solution in concentrations of 1, 2, and 3 mg/mL. Films were
annealed for 10 minutes at 80°C under nitrogen to drive off residual solvent. These three
concentrations produced films of thicknesses 8.8 nm, 12 nm and 23 nm, when measured with
spectroscopic ellipsometry. The C
60
, BCP, and Al layers were deposited by vacuum thermal
evaporation at a rate of 2 Å/s. The J-V curves from these devices are shown on figure 5.3(a).
The device with the thinnest donor layer (8.8 nm) was the most efficient, with a power
conversion efficiency of 1.42%. Increasing the thickness does not increase the short circuit
current (J
SC
) despite elevated light absorption, suggesting that the exciton diffusion length (L
D
)
of BDP-Por is on the order of 8 nm, similar to that of the pure PtTPBP.
6
The BDP-Por OPV
compares favorably to the PtTPBP device (figure 5.3(a)) achieving the primary objective by
increasing absorption of BDP-Por, consequently raising the photocurrent (J
SC
of 3.84±0.81 vs.
2.47±0.05 mA/cm
2
). The EQE for these devices shows that the enhanced current density is
indeed due to an improved response in the spectral region covered by the BODIPY unit which
absorbs to 550 nm (figure 5.3(b)). The fill factor is also comparable between BDP-Por and
PtTPBP, though it decreases as the BDP-Por thickness increases. It is notable that this increased
119

-0.8 -0.4 0.0 0.4 0.8
-6
-4
-2
0
2
4
6
a)
8.8 nm BDP-Por
12 nm BDP-Por
24 nm BDP-Por
15 nm PtTPBP
(mA/cm
2
)
Voltage (V)

Current density
400 500 600 700 800
0
10
20
30
40
50 b)
8.8 nm BDP-Por
12 nm BDP-Por
22 nm BDP-Por
Quantum Efficiency
(%)
Wavelength (nm)

Figure 5.3 (a) J-V curves for BDP-Por (X nm)/C
60
(40 nm) devices are shown against that of a
solution processed PtTPBP (15 nm)/C
60
(40 nm) device. It can be seen there that J
SC
has almost
doubled which can be attributed to the increased absorption. (b)The increased absorption can
be seen in the EQE plot where there is enhanced spectral response past 550 nm in the same
region as the BDP absorption.
120

current is achieved despite using considerably less material than in the PtTPBP based device
(9 nm vs. 15 nm). The V
OC
is unchanged in the BDP-Por device (V
OC
of 0.66±0.03 vs. 0.64±0.01
V), indicating that the porphyrin unit still acts as the electron donor in the CT process and that its
energy levels have not been shifted by the substitution.
21, 22

5.4.2 Blended Chromophore Systems
While BDP-Por offers some definite advantages over PtTPBP alone it involves a more
complicated synthetic process and cannot be deposited by thermal vacuum deposition methods.
Since the BDP and porphyrin units are not electronically coupled, an important question is if a
simple mixture of BODIPY and PtTPBP would give similar properties to BDP-Por in an OPV.
To answer this question blended films of varying ratios of PtTPBP and 5,5-difluoro-3,7-
dimethyl-10-phenyl-5H-dipyrrolo[1,2-c:2',1'-f][1,3,2]diazaborinin-4-ium-5-uide (BDP1 figure
3(a), blended thin films will be referred to as Por+BDP1) were prepared.
23
These materials were
chosen because they most closely match the chromophores used in the covalent array. Blended
films of Por+BDP1 ratios of 1:2, 1:4 and 1:6 were prepared by spin coating chloroform
solutions, in which the porphyrin content was held constant at the concentration of the 1mg/mL
BDP-Por solution and the BDP1 concentration was varied. The Por+BDP1 absorption spectra
are shown in figure 3(b). Here the BDP1 absorption at 515 nm can be seen growing in, without a
change in the PtTPBP absorption bands (see “as cast” in figure 5.4(b)). The spectrum of the 1:4
blend is nearly identical to that of BDP-Por, as expected for non-interacting chromophores.
During the processing of devices with these blended films, we observed that the
concentration of BDP1 in the blended film changed as a function of the time it was exposed to
121

BDP1 BDP2
a)

300 400 500 600 700
0.00
0.02
0.04
0.06
0.08
0.10
b)
1:2 As Cast
1:2 Post Vacuum
1:4 As Cast
1:4 Post Vacuum
1:6 As Cast
1:6 Post Vacuum

Absorbance
Wavelength (nm)

Figure 5.4 (a) The structures of BDP1 and BDP2. (b) The thin film absorptivities of three
different ratios of Por+BDP1 1:2 (red), 1:4 (blue), and 1:6 (green) before (squares) and after
(circles) being exposed to high vacuum for 2 hrs. It is clear in this plot that a substantial amount
of the BDP1 has been lost from the film under vacuum.
122

-0.8 -0.4 0.0 0.4 0.8
-6
-4
-2
0
2
4
6
a)
1:1 Por+BDP2
1:2 Por+BDP2
1:3 Por+BDP2

(mA/cm
2
)
Voltage (V)

Current density
400 500 600 700
0
5
10
15
20
25
30 b)
1:1 Por+BDP2
1:2 Por+BDP2
1:3 Por+BDP2
Quantum Efficiency
Wavelength (nm)

Figure 5.5 (a) J-V curves of devices made with donor layers of Por+BDP2 show that the 1:2 ratio
generates the best performance. (b) The EQE spectra of the same devices shows a shoulder grow
in with increasing BDP2 concentration.
123

high vacuum, which is required for deposition of the C
60
, BCP and Al layers of the device. Due
to BDP1’s low molecular weight it has a much higher volatility than that of the porphyrin.
Consequently, BDP1 is selectively sublimed out of the film when exposed to the high vacuum.
The absorption spectra of the mixed Por+BDP1 films are shown in figure 5.4(b) before and after
being exposed to high vacuum for 2 hrs. During that time more than half of the BDP1 is lost
from the film under vacuum leading to devices that were inconsistent and of poor quality overall.
To address the volatility problem of the BDP1 analog, we prepared a phenyl linked
BODIPY compound 10,10'-(1,3-phenylene)bis(2,8-diethyl-5,5-difluoro-1,3,7,9-tetramethyl-5H-
dipyrrolo[1,2-c:2',1'-f][1,3,2]diazaborinin-4-ium-5-uide) (BDP2, figure 5.4(a)), that has nearly
double the molecular weight and thus has a much lower volatility. The two BODIPY units of
BDP2 show no interaction in their ground states, leading to absorption spectra that are
unchanged from BDP1. In addition, the fluorescence spectra and redox properties are unchanged
between BDP1 and BDP2. Blended films of PtTPBP and BDP2 in ratios 1:1, 1:2, and 1:3 were
prepared by spin coating chloroform solutions. As BDP2 contains two BODIPY units its molar
absorptivity is doubled. The absorption spectra of the Por+BDP2 films were indistinguishable
from the Por+BdP1 films of the corresponding chromophore ratios. BDP2 has sufficiently high
molecular weight that there was no detectable loss of BDP2 from the Por+BDP2 film after
extensive exposure to high vacuum. Having made a stable blended film of similar composition
to the covalently bound chromophore array, we turned to make OPVs with the same structure as
the BDP-Por OPVs. Devices with the blended donor layer were prepared in the same fashion as
the BDP-Por devices and their J-V curves are shown in figure 5.5(a). OPVs with the 1:2
Por+BDP2 blend, i.e. the same chromophore ratio of BODIPY to porphyrin as in BDP-Por, gave
the best device performance (PCE of 1.33%). The performances of the devices with the different
124

-0.4 0.0 0.4 0.8
-4
-2
0
2
4
6
Current density (mA/cm
2
)
a)
Voltage (V)

Dark (0 sec)
Light (0 sec)
Dark (15 sec)
Light (15 sec)
Light (1 min)
Dark (1 min)
-0.8 -0.4 0.0 0.4 0.8
-4
-2
0
2
4
6
b)
Voltage (V) Current density (mA/cm
2
)

Dark
Light
400 500 600 700 800
0
10
20
30
40
50
Quantum Efficiency
Wavelength (nm)

Figure 5.6 (a) The J-V curves of an ITO/Por+BDP2/C
60
/BCP//Al device after varying periods
of illumination shows how the V
OC
starts around 0.64V and eventually stabilizes at 0.36 V. (b)
The J-V curve of a device with a neat BDP2 donor layer, i.e. ITO/BDP2/C
60
/BCP//Al, giving a
V
OC
of 0.34 V. (Inset) The EQE of the neat BDP2 based OPV.
125

ratios do not differ significantly, as the J
SC
and V
OC
remain fairly constant with the increasing
BDP2 content. The primary difference between devices made from different Por:BDP2 ratios is
seen in the FF of the devices, which maximize at the 1:2 Por to BDP2 ratio. The effect of the
increased BDP2 content on the quantum efficiency is seen in a proportional enhancement in the
response from the 500 to 550 nm. When comparing these blended donor films to the devices
made with the multi-chromophoric array, the J
SC
, FF, and the general shape of the EQE curve
are comparable as expected given their very similar absorption patterns. However the V
OC
values
measured in the devices with a blended donor layer where markedly lower than those with the
array (0.66±0.03 vs. 0.57±0.03 V, respectively).
The J-V curves for the Por+BDP2 devices are shown in figure 5.6 (a). For this set of data
the dark curve was measured before illumination. The first light scan (0 sec) was illuminated
under 1 sun intensity for ca. 2 seconds, the time it takes to complete the voltage sweep. When
this scan was repeated after further illumination the V
OC
of the devices decreased. The dark curve
remains stable for multiple voltage scans as long as no light is cast on the device. It is typically
the case that the light and dark curves will converge at high forward biases, assuming that there
isn’t substantial photoconductivity.
24-26
Upon initial illumination, the Por+BDP2 device
produced a J-V curve, with a V
OC
of 0.58 V, which clearly does not converge with the dark curve
at high potentials (figure 5.6(a) red curves), indicating that some change has already occurred in
the 2 seconds it takes to measure the J-V curve. Measuring a new dark scan after illumination
produces a curve that is consistent with the preceding light curve, converging at high forward
bias (compare blue dark and red light curves in figure 5.6(a)). Further illumination of the device
produces J-V curves with progressively lower V
OC
’s, but leaves the J
SC
and FF parameters
126

unchanged. These devices ultimately stabilize after about 1 minute of illumination at a V
OC
of
0.36 V.
The identity of the donor and acceptor,
21, 22, 27-30
as well as the structure at the D/A
interface,
5, 31-35
have been shown to be the primary factors determining the V
OC
values observed
for OPVs. These two factors are manifested in a clear correlation between the V
OC
and the
energy of the charge transfer state (E
CT
) between the donor and acceptor materials in the OPV.
29,
34, 35
As the V
OC
is characteristic of a particular material system, one can use the V
OC
of a device
to give insight into the molecular D/A pair at the interface of the device when more than one
donor or acceptor is present.
36-39
The BDP-Por devices and the initial scans of the Por+BDP2
devices give V
OC
values very close to the one observed for an OPV with a neat PtTPBP donor
layer, suggesting that the devices have a similar CT state, involving a PtTPBP-C
60
pair at the
interface. To determine the characteristic V
OC
of a device with a BDP2-C
60
E
CT
, an OPV was
prepared with a donor layer composed of only BDP2, i.e. ITO/BDP2(4nm)/C
60
(40 nm)/BCP
(10nm)/Al. The J-V curve of this device is shown in figure 5.6(b) which shows a V
OC
of 0.34 V.
After illumination the V
OC
values for the mixed donor Por+BDP2 devices stabilize at 0.36 V,
very close to that of the BDP2/C
60
OPV, suggesting that they both have the same ECT, namely
one based on a BDP2-C
60
CT state. This lets us conclude that illumination of the Por+BDP2/C
60

device leads to a morphological change where BDP2 concentrates at the D/A interface in such a
way that it starts to dominate the CT process. This instability of the Por+BDP2 devices under
illumination occurs when the devices are tested under an inert atmosphere, ruling out any
oxidation pathway as an explanation. The films of Por+BDP2 were investigated before and after
illumination by GIXRD but show no crystallization and no discernible difference between the
127

array and the blended film, indicating that both films are amorphous, and any domains are too
small to be detected by diffraction.
Summarizing these results, we have determined that when a film composed of both POR
and BDP is intimately intermixed, either as an array or blended film, the principal D/A pair that
contributes to the CT process at the D/A interface is Por/C
60
. This is not surprising as it mirrors
the emission of these films, where energy is dispersed over both chromophores but only a
porphyrin signal is detected due to the competition of emission rates.
10
However, after the
morphological changes that are induced during illumination, the blended films appear to phase
segregate in such a way that BDP2 concentrates at the D/A interface, giving rise to a V
OC
shift
based on the BDP2/C
60
CT state. However, despite the phorphyrin moiety being moved from the
interface, light is collected by both moieties, because the exciton is still in a dynamic equilibrium
between PtTPBP and BDP2. This leads to the reduced V
OC
but leaves the other device
parameters unchanged.

5.5 Conclusions
Broadening the spectral response of OPVs is a primary challenge that needs to be
overcome. It has been shown here that the inclusion of multiple chromophores with
complementary absorptions in one or more of the photoactive layers can help extend the spectral
response of devices, provided that the energetics are correctly designed as to not introduce
charge or exciton traps. This works by both creating a multi-chromophoric array or by simple
blending of two separate chromophores. It was found here that both methods can enhance the
performance over the neat device; however, the multi-chromophoric array here produces a more
stable device. The covalently bonded chromophores in the array form a stable, intimately
128

intermixed film, while the blended donor layer films are only intimately intermixed at casting
and appear to reorganize under illumination to give a high concentration of BDP at the D/A
interface, leading to a diminished V
OC
. The light induced phase segregation results in a V
OC
similar to that produced by a device with a neat layer of BDP2 as the donor layer indicating that
it has phase segregated to the interface. In this way the multi-chromophoric array can hold an
advantage to blended films by producing an inherently more morphologically stable film, though
it is more difficult to synthesize.

129

5.6 Chapter 5 References
1. Birks, J. B., Photophysics of aromatic molecules. Wiley-Interscience: London, 1970.
2. Turro, N. J., Modern Molecular Photochemistry. Universty Science Books: 1991.
3. Lee, S.; Sarker, A. K.; Hong, J., A zinc porphyrin sensitizer modified with donor and
acceptor groups for dye-sensitized solar cells. Bulletin of the Korean Chemical Society 2014, 35,
(10), 3052-3058.
4. Lu, J.; Zhang, B.; Liu, S.; Li, H.; Yaun, H.; Shen, Y.; Xu, J.; Cheng, Y.; Wang, M., A
cyclopenta[1,2-b:5,4-b']dithiophene-porphyrin conjugate for mesoscopic solar cells: a D-π-D-A
approach. Physical Chemistry Chemical Physics 2014, ASAP.
5. Trinh, C.; Whited, M. T.; Steiner, A.; Tassone, C. J.; Toney, M. F.; Thompson, M. E.,
Chemical Annealing of Zinc Tetraphenylporphyrin Films: Effects on Film Morphology and
Organic Photovoltaic Performance. Chemistry of Materials 2012, 24, (13), 2583-2591.
6. Perez, M. D.; Borek, C.; Djurovich, P. I.; Mayo, E. I.; Lunt, R. R.; Forrest, S. R.;
Thompson, M. E., Organic Photovoltaics Using Tetraphenylbenzoporphyrin Complexes as
Donor Layers. Advance Materials 2009, 21, 1517-1520.
7. Kim, B.; Ma, B.; Donuru, V. R.; Liu, H.; Frechet, J. M. J., Bodipy-backboned polymers
as electron donor in bulk heterojunction solar cells. Chemical Communications 2010, 46, (23),
4148-4150.
8. Kubo, Y.; Watanabe, K.; Nishiyabu, R.; Hata, R.; Murakami, A.; Shoda, T.; Ota, H.,
Near-Infrared Absorbing Boron-dibenzopyrromethenes that Serve As Light-Harvesting
Sensitizers for Polymeric Solar Cells. Organic Letters 2011, 13, (17), 4574-4577.
9. Rousseau, T.; Cravino, A.; Ripaud, E.; Leriche, P.; Rihn, S.; De Nicola, A.; Ziessel, R.;
Roncali, J., A tailored hybrid BODIPY-oligothiophene donor for molecular bulk heterojunction
solar cells with improved performances. Chemical Communications 2010, 46, (28), 5082-5084.
10. Whited, M. T.; Djurovich, P. I.; Roberts, S. T.; Durrell, A. C.; Schlenker, C. W.;
Bradforth, S. E.; Thompson, M. E., Singlet and Triplet Excitation Management in a
Bichromophoric Near-Infrared-Phosphorescent BODIPY-Benzoporphyrin Platinum Complex.
Journal of the American Chemical Society 2010, 133, (1), 88-96.
11. Denti, G.; Campagna, S.; Serroni, S.; Ciano, M.; Balzani, V., Decanuclear homo- and
heterometallic polypyridine complexes: syntheses, absorption spectra, luminescence,
electrochemical oxidation, and intercomponent energy transfer. Journal of the American
Chemical Society 1992, 114, (8), 2944-2950.
12. Sato, T.; Jiang, D.-L.; Aida, T., A Blue-Luminescent Dendritic Rod:
Poly(phenyleneethynylene) within a Light-Harvesting Dendritic Envelope. Journal of the
American Chemical Society 1999, 121, (45), 10658-10659.
130

13. Stewart, G. M.; Fox, M. A., Chromophore-Labeled Dendrons as Light Harvesting
Antennae. Journal of the American Chemical Society 1996, 118, (18), 4354-4360.
14. Hindin, E.; Forties, R. A.; Loewe, R. S.; Ambroise, A.; Kirmaier, C.; Bocian, D. F.;
Lindsey, J. S.; Holten, D.; Knox, R. S., Excited-State Energy Flow in Covalently Linked
Multiporphyrin Arrays: The Essential Contribution of Energy Transfer between Nonadjacent
Chromophores. The Journal of Physical Chemistry B 2004, 108, (34), 12821-12832.
15. Ziessel, R.; Ulrich, G.; Haefele, A.; Harriman, A., An Artificial Light-Harvesting Array
Constructed from Multiple Bodipy Dyes. Journal of the American Chemical Society 2013, 135,
(30), 11330-11344.
16. Ohya, Y.; Nakajima, A.; Ouchi, T., Time-resolved Laser Spectroscopic Analysis of
Multi-step Fluorescence Resonance Energy Transfer on Chromophore Array Constructed by
Oligo-DNA Assembly. Supramolecular Chemistry 2005, 17, (4), 283-289.
17. Okujima, T.; Jin, G.; Hashimoto, Y.; Yamada, H.; Uno, H.; Ono, N., Synthesis of
4,7Dihydro-2H-isoindole derivatives Via Diels-Alder Reaction of Tosylacetylene. Heterocycles
2006, 70, 619-626.
18. Ghosh, A., A Perspective of One-Pot Pyrrole–Aldehyde Condensations as Versatile Self-
Assembly Processes. Angewandte Chemie International Edition 2004, 43, (15), 1918-1931.
19. Shen, Z.; Röhr, H.; Rurack, K.; Uno, H.; Spieles, M.; Schulz, B.; Reck, G.; Ono, N.,
Boron–Diindomethene (BDI) Dyes and Their Tetrahydrobicyclo Precursors—en Route to a New
Class of Highly Emissive Fluorophores for the Red Spectral Range. Chemistry – A European
Journal 2004, 10, (19), 4853-4871.
20. Shrotriya, V.; Li, G.; Yao, Y.; Moriarty, T.; Emery, K.; Yang, Y., Accurate Measurement
and Characterization of Organic Solar Cells. Advanced Functional Materials 2006, 16, (15),
2016-2023.
21. Perez, M. D.; Borek, C.; Forrest, S. R.; Thompson, M. E., Molecular and Morphological
Influences on the Open Circuit Voltages of Organic Photovoltaic Devices. Journal of the
American Chemical Society 2009, 131, (26), 9281-9286.
22. Graham, K. R.; Erwin, P.; Nordlund, D.; Vandewal, K.; Li, R.; Ngongang Ndjawa, G. O.;
Hoke, E. T.; Salleo, A.; Thompson, M. E.; McGehee, M. D.; Amassian, A., Re-evaluating the
Role of Sterics and Electronic Coupling in Determining the Open-Circuit Voltage of Organic
Solar Cells. Advanced Materials 2013, 25, (42), 6076-6082.
23. Qin, W.; Baruah, M.; Van der Auweraer, M.; De Schryver, F. C.; Boens, N.,
Photophysical Properties of Borondipyrromethene Analogues in Solution. The Journal of
Physical Chemistry A 2005, 109, (33), 7371-7384.
24. Renshaw, C. K.; Zimmerman, J. D.; Lassiter, B. E.; Forrest, S. R., Photoconductivity in
donor-acceptor heterojunction organic photovoltaics. Physical Review B 2012, 86, (8), 085324.
131

25. Giebink, N. C.; Wiederrecht, G. P.; Wasielewski, M. R.; Forrest, S. R., Ideal diode
equation for organic heterojunctions. I. Derivation and application. Physical Review B 2010, 82,
(15), 155305.
26. Jeong, W.-I.; Lee, Y. E.; Shim, H.-S.; Kim, T.-M.; Kim, S.-Y.; Kim, J.-J.,
Photoconductivity of C60 as an Origin of Bias-Dependent Photocurrent in Organic
Photovoltaics. Advanced Functional Materials 2012, 22, (14), 3089-3094.
27. Schlenker, C. W.; Barlier, V. S.; Chin, S. W.; Whited, M. T.; McAnally, R. E.; Forrest, S.
R.; Thompson, M. E., Cascade Organic Solar Cells. Chemistry of Materials 2011, 23, (18), 4132-
4140.
28. Khlyabich, P. P.; Burkhart, B.; Thompson, B. C., Efficient Ternary Blend Bulk
Heterojunction Solar Cells with Tunable Open-Circuit Voltage. Journal of the American
Chemical Society 2011, 133, (37), 14534-14537.
29. Vandewal, K.; Tvingstedt, K.; Gadisa, A.; Inganas, O.; Manca, J. V., On the origin of the
open-circuit voltage of polymer-fullerene solar cells. Nat Mater 2009, 8, (11), 904-909.
30. Bartynski, A. N.; Trinh, C.; Kirlikovali, K. O.; Thompson, M. E., Multichromophoric
energy sensitization of C60 for organic photovoltaics. Applied Physics Letters 2014, 105, (11),
113305.
31. Zimmerman, J. D.; Lassiter, B. E.; Xiao, X.; Sun, K.; Dolocan, A.; Gearba, R.; Vanden
Bout, D. A.; Stevenson, K. J.; Wickramasinghe, P.; Thompson, M. E.; Forrest, S. R., Control of
Interface Order by Inverse Quasi-Epitaxial Growth of Squaraine/Fullerene Thin Film
Photovoltaics. ACS Nano 2013, 7, (10), 9268-9275.
32. Zimmerman, J. D.; Xiao, X.; Renshaw, C. K.; Wang, S.; Diev, V. V.; Thompson, M. E.;
Forrest, S. R., Independent Control of Bulk and Interfacial Morphologies of Small Molecular
Weight Organic Heterojunction Solar Cells. Nano Letters 2012, 12, (8), 4366-4371.
33. Spoltore, D.; Vangerven, T.; Verstappen, P.; Piersimoni, F.; Bertho, S.; Vandewal, K.;
Van den Brande, N.; Van Mele, B.; De Sio, A.; Parisi, J.; Lutsen, L.; Vanderzande, D.; Maes,
W.; Manca, J. V., Effect of molecular weight on morphology and photovoltaic properties in
P3HT:PCBM solar cells. Organic Electronics, (0).
34. Vandewal, K.; Widmer, J.; Heumueller, T.; Brabec, C. J.; McGehee, M. D.; Leo, K.;
Riede, M.; Salleo, A., Increased Open-Circuit Voltage of Organic Solar Cells by Reduced
Donor-Acceptor Interface Area. Advanced Materials 2014, 26, (23), 3839-3843.
35. Vandewal, K.; Oosterbaan, W. D.; Bertho, S.; Vrindts, V.; Gadisa, A.; Lutsen, L.;
Vanderzande, D.; Manca, J. V., Varying polymer crystallinity in nanofiber poly(3-
alkylthiophene): PCBM solar cells: Influence on charge-transfer state energy and open-circuit
voltage. Applied Physics Letters 2009, 95, (12), 123303.
132

36. Ran, N. A.; Kuik, M.; Love, J. A.; Proctor, C. M.; Nagao, I.; Bazan, G. C.; Nguyen, T.-
Q., Understanding the Charge-Transfer State and Singlet Exciton Emission from Solution-
Processed Small-Molecule Organic Solar Cells. Advanced Materials 2014, 26, (43), 7405-7412.
37. Kirchartz, T.; Bisquert, J.; Mora-Sero, I.; Garcia-Belmonte, G., Classification of solar
cells according to mechanisms of charge separation and charge collection. Physical Chemistry
Chemical Physics 2015, 17, (6), 4007-4014.
38. Hörmann, U.; Lorch, C.; Hinderhofer, A.; Gerlach, A.; Gruber, M.; Kraus, J.; Sykora, B.;
Grob, S.; Linderl, T.; Wilke, A.; Opitz, A.; Hansson, R.; Anselmo, A. S.; Ozawa, Y.; Nakayama,
Y.; Ishii, H.; Koch, N.; Moons, E.; Schreiber, F.; Brütting, W., Voc from a Morphology Point of
View: the Influence of Molecular Orientation on the Open Circuit Voltage of Organic Planar
Heterojunction Solar Cells. The Journal of Physical Chemistry C 2014, 118, (46), 26462-26470.
39. Ko, S.; Hoke, E. T.; Pandey, L.; Hong, S.; Mondal, R.; Risko, C.; Yi, Y.; Noriega, R.;
McGehee, M. D.; Brédas, J.-L.; Salleo, A.; Bao, Z., Controlled Conjugated Backbone Twisting
for an Increased Open-Circuit Voltage while Having a High Short-Circuit Current in
Poly(hexylthiophene) Derivatives. Journal of the American Chemical Society 2012, 134, (11),
5222-5232.

Abstract (if available)

Abstract The field of organic photovoltaics has received much attention in recent years and has made great strides toward market relevancy. The need for new, clean and environmentally friendly sources of energy has been made clear with the latest reports on global climate change. Solar-to-electric energy conversion is the ideal candidate to produce the large scale power needed with little impact to the planet, promising terawatt energy scale production with zero emissions. However, the development of solar energy is hamstrung with high production costs leading to prohibitively high capital investment necessary for implementation. In this, organic photovoltaics has an opportunity to make an impact in the market by providing a technology with inherently lower materials costs due to the promise of carbon-based dyes that can be manufactured with roll-to-roll technologies. Before this objective can be achieved however, conversion efficiencies need to be taken to economical levels. Key in the optimization of these devices will be the understanding of the links between materials properties and device parameters. ❧ It is the objective of this thesis to elucidate some relationships between materials properties of organic chromophores and the device parameters in the resulting solar cells. Much insight has been taken from the inorganic silicon solar cell industry in predicting how organic photovoltaics work, but there are many fundamental differences due to the molecular nature of the materials and these divergences are explored in this work. Different series of organic materials were developed with systematically altered chemical structures allowing conclusions to be drawn about, for example, how the resulting open circuit voltage of the device is impacted by the changes under study. Film morphologies created by the packing of these molecules and the impact in the resulting device will also be presented here. Finally, the efficacy of supramolecular chromophore films will be contrasted with blended films of the same chromophore units. It is the hope that through this study of structure property relationships that the next generation in organic photovoltaic materials.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Molecular aspects of photoconversion processes in organic solar cells

PDF

Organic solar cells: molecular electronic processes and device development

PDF

Molecular and polymeric donors for bulk heterojunction organic solar cells

PDF

Organic photovoltaics: from materials development to device application

PDF

Utilizing n-heterocyclic chromophores for solar energy harvesting.

PDF

Improving the efficiency and stability of organic solar cells through ternary strategies

PDF

Energy management in organic photovoltaics

PDF

Porphyrin based near infrared‐absorbing materials for organic photovoltaics

PDF

The influence of CdSe surface ligands on hybrid inorganic/organic solar cell performance

PDF

Molecular design of pyridine and quinoline based compounds for organic optoelectronics

PDF

Molecular design for organic photovoltaics: tuning state energies and charge transfer processes in heteroaromatic chromophores

PDF

Synthesis and characterization of 3-hexylesterthiophene based random and semi-random polymers and their use in ternary blend solar cells

PDF

Efficient ternary blend bulk heterojunction solar cells with tunable open-circuit voltage

PDF

Properties and applications of dipyrrin-based molecules in organic photovoltaics

PDF

Synthesis and characterization of novel donor-acceptor systems for solar water splitting

PDF

Development of a family of semi-random multichromophoric polymers for application in organic solar cells

PDF

Squaraines and their applications to organic photovoltaics

PDF

Singlet fission in disordered acene films: from photophysics to organic photovoltaics

PDF

Light management in nanostructures: nanowire solar cells and optical epitaxial growth

PDF

Expanding the library of surface ligands for semiconductor nanocrystal synthesis and photovoltaic applications

Asset Metadata

Creator Erwin, Patrick R. (author)

Core Title Molecular and morphological effects on the operational parameters of organic solar cells

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Chemistry

Publication Date 08/05/2016

Defense Date 05/22/2015

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag OAI-PMH Harvest,organic materials,organic photovoltaics,solar cell

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Thompson, Mark E. (committee chair), Armani, Andrea M. (committee member), Brutchey, Richard L. (committee member)

Creator Email perwin@usc.edu,perwin83@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-627204

Unique identifier UC11305798

Identifier etd-ErwinPatri-3811.pdf (filename),usctheses-c3-627204 (legacy record id)

Legacy Identifier etd-ErwinPatri-3811.pdf

Dmrecord 627204

Document Type Dissertation

Format application/pdf (imt)

Rights Erwin, Patrick R.

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA