University of Southern California Dissertations and Theses

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

(USC Thesis Other)

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

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Molecular Design for Organic
Photovoltaics

Tuning State Energies and Charge Transfer Processes
in Heteroaromatic Chromophores

by
Jessica H. Golden
__________________________________________________________________
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)

December 2018
ii

for
Mark Joseph Davey

&

a future of life on Earth

iii

The crucified planet Earth,
should it find a voice
and a sense of irony,
might now well say
of our abuse of it,
“Forgive them, Father,
They know not what they do.”

The irony would be
that we know what
we are doing.

When the last living thing
has died on account of us,
how poetical it would be
if Earth could say,
in a voice floating up
perhaps
from the floor
of the Grand Canyon,
“It is done.”
People did not like it here.

Credit: Kurt Vonnegut, “The Crucified Planet Earth” from A Man Without a Country.
Copyright © 2010 by Kurt Vonnegut. Used with the permission of The Permissions Company,
Inc., on behalf of Seven Stories Press, www.sevenstories.com.
iv
Acknowledgements
I owe my PhD to many, many people. I always knew I wanted to work to protect the planet in
some meaningful way. I was just not always sure of the best way to do it. In my formative
years, my mother and father nurtured a passion for learning, giving me new workbooks, texts,
and scientific tools in lieu of Barbies, Bratz, or Easy-Bake Ovens; they took me camping,
hiking, out to explore coastal tide-pools and alpine riverbeds. When I was at home and not
occupied collecting bugs outside or looking at water samples under my microscope, I spent
many hours on the couch, watching and learning about the natural world from people like Steve
Irwin and David Attenborough. I am constantly inspired and awed by the exploration of nature,
and, naturally, I at first thought a career as an explorer or a biologist would suit me best.
Unfortunately, with the advent of high resolution satellite imaging and lidar mapping, there
are no longer many positions open for explorers these days, and I found biology tedious, messy,
and lacking in rigor. Please don’t tell my biologist friends I said this. In my sophomore year of
high school, my chemistry teacher, Mr. Hillier, hooked me on chemistry the day he climbed
atop his desk and contorted his body into various shapes to illustrate VSEPR theory. This was
a science that bridged my love of the physical world with my pseudo-obsessive need for
precision. He taught me about chemistry’s role in the world; how it can be used to destroy or
to create, to harm or to repair, and that it took responsible science and responsible government
to favor the latter alternatives. He made it clear to me that the world needed saving, and to do
that, it needed chemists. From my undergraduate studies, several professors deserve
mentioning, as each of them taught me a portion of what was at the time a very piecemeal
conception of chemistry but, over time, would eventually become fully-fledged, uniform, and
self-supporting. Professors Scott Oliver (inorganic), Glenn Millhauser (statistical mechanics),
v
Bakthan Singaram (physical organic), and Rebecca Braslau (organic) taught me the
fundamentals which bridged my understanding of chemistry into something that could be used
creatively, fluidly, and rigorously in my graduate research. Professors Singaram and Braslau
are especially dear to me. Rebecca took me on as an undergraduate researcher in the beginning
of my junior year. There, she treated me like a graduate student, letting me make my own
mistakes and develop my wet chemistry skills without fear or intimidation. In this, I also have
Dr. Aruna Earla to thank. She was my graduate student mentor; she taught me to use my hands
efficiently and carefully in the laboratory, how to fix my mistakes, and how to limit them in
the future. She did this all with never-wavering patience and kindness and I owe her a great
debt of gratitude for that. I have tried to emulate her methods in my own mentorship of
undergraduate students. Both Professors Braslau and Singaram are, I feel, entirely responsible
for my ending up in the Department of Chemistry at the University of Southern California, as
they leveraged their connections within the department to make sure I had guidance through
the admissions process and support immediately upon walking through the gates. I am also
endlessly grateful to my PhD advisor, Professor Mark Thompson. I have constantly felt both
emotionally and intellectually supported under his advisement. He nurtures a group
environment of mutual respect, kindness, and intellectual rigor which I have not seen matched
anywhere else. I like coming to work; even when my projects are failing to produce results, the
commute is long, and the weather is bad (read: hot), I have looked forward to my time in the
MET Lab.

I hope that this dissertation does these people justice. Their support, their guidance, and their
kindness made it possible.
vi
Table of Contents

Dedication ii

Acknowledgements iv

Chapter 1. Introduction to Organic Photovoltaics: A Molecular Design Perspective 1
1.1 Global Climate Change and the Demand for a Clean Energy Future 1
1.2 Solar Energy Conversion Technologies 7
1.3 Photocurrent Generation in Organic Photovoltaics 13
1.3.1 Photon Absorption 13
1.3.2 Exciton Migration 20
1.3.3 Charge Transfer 22
1.3.4 Charge Separation 24
1.3.5 Charge Collection 24
1.4 State of the Art in Organic Photovoltaic Active Layer Materials 26
1.5 Molecular Design Principles for Organic Photovoltaics 27
1.6 Bibliography 28

Chapter 2. Dipyridylmethene (DIPYR) Dyes: Shedding Light on Pyridine-based
Chromophores 32
2.1 Introduction to Dipyridylmethene Dyes 32
2.2 Dipyridylmethane and its Quinoline and Isoquinoline Analogues 36
2.3 Boron Difluoride Dipyridylmethene 40
2.3.1 Borylation of Dipyridylmethene Dyes 41
2.3.2 Structure of DIPYR Dyes 43
2.3.3 Electrochemical Properties of DIPYR Dyes 46
2.3.4 Photophysical Properties of DIPYR Dyes 47
2.3.5 Computational Study of DIPYR Dyes 50
2.3.6 A Unified Explanation of the Photophysical Properties of DIPYR
Dyes 56
2.4 Carbon-Chelated Dipyridylmethene Salts 60
2.5 The Bright Future of DIPYR Dyes 65
2.6 Experimental Methods 67
2.7 Bibliography 71

Chapter 3. Near-Infrared Absorption in Dipyridylmethene Dyes 74
3.1 Introduction to Near Infrared Absorbing Dyes for Organic Photovoltaics 74
3.2 meso-Substituted DIPYR Dyes 77
vii
3.3 Linearly -extended DIPYRs 82
3.4 Double DIPYRs 85
3.5 Fused DIPYRs: littleDIPYR and bigDIPYR 93
3.6 Experimental Methods 96
3.7 Bibliography 97

Chapter 4. Symmetry-Breaking Charge Transfer in Boron Dipyridylmethene Dimers 101
4.1 Introduction to SBCT in DIPYR Dimers 101
4.2 Synthesis and Structural Properties of bis-DIPYRs 102
4.3 Steady-State Photophysical Properties 106
4.4 Electrochemical Properties and Spectroelectrochemistry 112
4.5 Time-Resolved Photophysics of bis-DIPYRs 117
4.6 Theoretical Modeling 124
4.7 Shifting Paradigms in Symmetry-Breaking Charge Transfer 138
4.8 Experimental Methods 130
4.9 Bibliography 132

Chapter 5. Symmetry-Breaking Charge Transfer in Dipyrrin Dimers: Zinc Dipyrrins
and bis-BODIPYs 135
5.1 Symmetry-Breaking Charge Transfer in bis-BODIPYs and Zinc Dipyrrins 135
5.2 Synthesis of bis-BODIPYs 137
5.3 Structural and Photophysical Characterization of bis-BODIPYs 139
5.4 Synthesis of Zinc Dipyrrins 140
5.5 Structural and Photophysical Characterization of Zinc Dipyrrins 143
5.6 Excited State Decay Dynamics in Dipyrrin Dimers 147
5.7 Aggregation Induced Emission (AIE) in Zinc Dipyrrins 154
5.8 Comparative Study of Dipyrrin Dimers – The Future of SBCT and its
Application to Organic Photovoltaics 156
5.9 Isoindole Condensation on Oxalyl Chloride – Formation of a Diradicaloid
Fluorophore 160
5.9 Experimental Methods 165
5.10 Bibliography 167

Chapter 6. Energy and Charge Transfer Processes in Dipyrrin and Porphyrin Materials for
Organic Photovoltaics, Photochemical Upconversion, and Photocatalytic Water
Splitting 171
6.1 Introduction to Multichromophoric Arrays in Organic Photovoltaics 171
6.1.1 Absorptivity and Energy Transfer in BDP-Por 172
6.1.2 Synthesis of BDP-Por 176
6.1.3 Performance of BDP-Por / C60 Organic Photovoltaics 176
6.1.4 Morphological Instability in the Ternary Blend Por+BDP2 Device 181
viii
6.1.5 Implications of Multichromophoric Arrays in Organic Photovoltaics 183
6.2 Photochemical Upconversion 184
6.2.1 Triplet-Triplet Annihilation for Photochemical Upconversion in Air 186
6.2.2 Oxygen Scavenging in PtTPBP PEG Solutions 188
6.2.3 Photochemical Upconversion in Air using PtTPBP as a Triplet
Sensitizer of BPEA 193
6.2.4 Rapid Photophysical Screening of TTA-UC Systems in Air 202
6.3 Introduction to Photocatalytic Water Splitting 203
6.3.1 Zeolite Nanoarchitectures for Photocatalytic Water Splitting 204
6.3.1 Synthesis of BODIPY-Pyridinium Plugs 205
6.3.4 Energy and Charge Transfer Control in Zeolite Nanoarchitectures
for Photocatalytic Water Splitting 208
6.4 Experimental Methods 213
6.5 Bibliography 216

Chapter 7. Tuning State Energies for Narrow Blue Emission in Tetradendate Pyridyl-
carbazole Platinum Complexes 221
7.1 Introduction to Organic Light-Emitting Diodes 221
7.2 Materials Design Considerations to Improve Device Lifetimes and
Efficiencies 222
7.3 Tetradentade Platinum Phosphors in Organic Light Emitting Diodes 223
7.4 Tuning State Energies in PtNON Derivatives 226
7.5 PtNON Suitability for Lighting and Display Blue OLEDs 234
7.6 Experimental Methods 236
7.7 Bibliography 240

Appendix I. Instrumentation & Experimental Methods 243

Appendix II. Nuclear Magnetic Resonance Spectra 248

Molecular Design for Organic Photovoltaics
Chapter 1 | 1
Chapter 1

Introduction to Organic Photovoltaics: A Molecular
Design Perspective

1.1 Global Climate Change and the Demand for a Clean Energy Future
The Earth’s climate is undergoing accelerating anthropogenic change largely stemming
from the burning of fossil fuels to meet global energy demand.
1
The scientific evidence for
anthropogenic climate change is so overwhelming, and the projections for its triggering of
economic instability and agricultural and ecological devastation are so compelling, that in
December 2015, the United Nations Framework Convention on Climate Change completed an
historic treaty, the Paris Agreement, to which all 195 nations worldwide have since signed,
agreed to, or ratified.
2
This treaty outlines the responsibility of nations to mitigate climate
change with the goal of keeping global warming below 2º C above pre-industrial levels. This
threshold of 2º C represents a generally accepted rallying cry to mitigate (but not to eliminate
or reverse) the deleterious effects of global warming on the environment, agriculture, and the
global economy. The idea of a discrete “tipping point” for irreversible climate change is better
understood, however, as a political motivator than as a scientific fact, and while this 2º C figure
represents a reasonable political goal embodying historically observed abrupt changes to local
environs, the goal of the scientific community should be to halt anthropogenic climate change
completely and as quickly as possible.
3

Although global warming has been linked with high confidence to an increase in
greenhouse gas (GHG) emissions, and despite the call to mitigate these emissions by nations
Golden
2
worldwide, combined GHG emissions grew by an average of 1.0 x 10
12
kg (2.2%) carbon
dioxide equivalents annually between 2000 and 2010, compared to 0.4 x 10
12
kg (1.3%)
annually between 1970 and 2000,
4
and annual global CO2 emissions continue to rise from the
continued burning of solid, liquid, and gaseous fossil fuels (coal, petroleum derivatives, and
natural gas) (Figure 1.1).
5
It has been shown that, given current mitigation efforts, the
statistical likelihood that the Paris Agreement’s 2º C threshold will be met is only 5%, and the
1850 1900 1950 2000
0
5
10
15
20
25
30
35
40

Annual CO
2
Emissions (x 10
12
kg)
Year
gas
liquid fuel consumption
solid fuel consumption
cement production
gas flaring

Figure 1.1. Annual CO 2 emissions in trillions of
kilograms as a function of source and year. Figure
generated from published historical data.
5

1900 1925 1950 1975 2000
-0.4
0.0
0.4
0.8
1.2

Temperature Rise Above 1951-1980 Average
Year
Annual Mean
Smoothed

Figure 1.2. Mean global surface temperature
increase based upon 1951-1980 average. 2016 was
the warmest year on record, and 2017 represents an
increase of 0.9º C.
8

Figure 1.3. Global surface temperatures in 2017 compared the 1951-1980 average.
8
The arctic has suffered the
highest temperature increase, causing glacial melting and contributing to sea level rise.

Molecular Design for Organic Photovoltaics
Chapter 1 | 3
probability of meeting the Agreement’s ideal goal of a 1.5º C global temperature increase
represents a 1% probability.
6
The 90% confidence interval for the global temperature increase
above the 1951-1980 average encompasses a range between 2.0 and 4.9º C by 2100. Further
data has shown that even if global emissions cease immediately and in their entirety, past
emissions have already committed the climate to an inevitable 1.5º C rise; at the time of writing
this dissertation, climate change has warmed the earth by 0.9º C above its 1951-1980 average
(Figure 1.2), with an average local surface temperature increase affecting almost every part
of the world (Figure 1.3).
7-8

The consequences of global warming have been summarized in the IPCC report on
Impacts, Adaptation, and Vulnerability.
9
Some highlights from the report are summarized as
follows, and Figure 1.4 contains an infographic from the report illustrating impact regions and
levels worldwide. The warming climate has caused glaciers worldwide to shrink, contributing
to sea level rise. Extinction events have begun and are likely to worsen, and some species have
shifted their geographical ranges and migration patterns in response to the changing climate.
Crop yields have been negatively affected in most regions, with wheat and maize showing
particular vulnerability. Human morbidity from heat-related causes has increased and the
distribution of disease vectors has shown change, with unknown consequences. Finally, heat
waves, droughts, floods, cyclones, and wildfires have become more frequent and more
detrimental to both ecological and human systems.
Global warming is not the only consequence of increased CO2 concentrations in the
Earth’s atmosphere. Approximately 30 to 40% of atmospheric CO2 dissolves into the planet’s
surface waters, including the oceans, and increasing levels of CO2 in the ocean are responsible
Golden
4

Figure 1.4. (A) Global patterns of impacts in recent decades attributed to climate change. Symbols indicate
categories of attributed impacts, the relative contribution of climate change (major or minor) to the observed
impact, and confidence in attribution. (B) Average rates of change in distribution (km per decade) for marine
taxonomic groups based on observations over 1900–2010. Positive distribution changes are consistent with
warming (moving into previously cooler waters, generally poleward). The number of responses analyzed is given
within parentheses for each category. (C) Summary of estimated impacts of observed climate changes on yields
over 1960–2013 for four major crops in temperate and tropical regions, with the number of data points analyzed
given within parentheses for each category. Figure and text duplicated from Ref. 9.
Molecular Design for Organic Photovoltaics
Chapter 1 | 5
for ocean acidification.
10
The oceans,
pre-industrially equilibrated at a pH of
8.25, had acidified to a pH of 8.14 by
2014 and are forecasted to reach a pH
of 7.85 by 2100, representing a factor
of 2.5 increase in H
+
ions over historical
averages.
1, 11
Excess carbonic acid in
the oceans depletes the concentration of
carbonate ions which calcifying
organisms such as corals, plankton, and shellfish require to build their exoskeletal structures.
These organisms are fundamental to marine ecosystem stability and as such, ocean
acidification places the entire marine ecosystem and all economic processes stemming from it
at risk of failure. The double-edged sword of ocean warming and acidification from increased
atmospheric CO2 concentrations have combined already to induce massive coral die-offs,
termed “coral bleaching” which will continue to worsen as CO2 concentrations rise.
1, 12

Despite the threat of catastrophic climate change from GHG emissions, global energy
demand continues to rise, with the forecasted energy demand in 2040 exceeding today’s value
by 33% (Figure 1.5), and fossil fuels remain the predominant energy source.
13
It is clear that
there is an urgent and ever increasing need for renewable, GHG emissions-free energy sources
to meet increasing global demand without furthering the deleterious effects of climate change.
A comparison of world energy reserves, including total fossil fuel reserves and total nuclear
fissionable material reserves as well as renewable sources including wind, hydroelectric, ocean
thermal, biomass, geothermal, tidal, and solar energy sources, shows that the total available
2015 2020 2025 2030 2035 2040
0
200
400
600
800
1000

Non-OECD
Europe
Asia
Middle East
Africa
Americas
Projected Energy Consumption
(quadrillion Btu)
Year
OECD
Americas
Europe
Asia

Figure 1.5. Historical and projected world energy
consumption by geographical region and OECD
classification from 2012 through 2040.
13

Golden
6
solar energy is orders of magnitude greater than all other individual energy resources (Figure
1.6) and many times greater than all other energy resources combined.
14-15

Although solar energy is by far the Earth’s most abundant energy resource, there are
several barriers left to overcome before the cost per watt of solar-derived electricity reaches
parity with that of fossil fuels.
16
In Los Angeles, the cost per watt of residential electricity in
April 2018 was $0.131, which was much higher than the United States average of $0.098 in
the same month;
17
grid parity is met only by systems which provide energy at this rate or lower.
Even given the consequences of continued CO2 emissions from the utilization of fossil fuels,
it is unlikely that solar will be adopted as the predominant energy source until cost parity with
non-renewables is met. It is the goal of this dissertation to present design strategies for the
discovery of materials which can be used in the development of inexpensive and efficient solar
photovoltaic technologies which can be used to decentralize energy production from this
abundant, GHG emissions-free resource. In so doing, it is the sincerest hope of this researcher
that the worst of the projected climate change effects may be avoided.

Figure 1.6. Total non-renewable nuclear and fossil fuel reserves compared to annual energy supply from
renewable sources and world energy demand.
14-15
Solar represents the largest available energy resource on
Earth. Total petroleum reserves are predicted to be completely depleted within 56 years.
Molecular Design for Organic Photovoltaics
Chapter 1 | 7
1.2 Solar Energy Conversion Technologies
There are two categories of solar energy conversion technologies; the first is solar
photovoltaics, which convert photons directly into electrical current. The second category
comprises concentrated solar, also called solar thermal, wherein solar energy is concentrated
from a wide area using mirrors and directed to a single point or a line, where it is used to heat
a fluid. The heated fluid, typically molten salt, goes on to produce steam which does work on
a turbine generator, converting steam energy into electrical current. Concentrated solar requires
enormous scales to approach theoretical efficiency limits, on the order of tens of megawatts
and larger; as such, concentrated solar installations are large and require significant capital
investment, and the cost per watt of this technology is high.
16
One benefit of concentrated solar
is that it may be combined in hybrid fossil fuel / concentrated solar plants, where turbine
generators are ultimately responsible for the generation of electrical current; this configuration
would allow the solar concentrator to do the bulk of current generation during the daylight
hours, while at night natural gas can be used to keep up with demand. The practical limitations
of concentrated solar, including scale, cost, and its reliance on direct solar irradiation, however,
make this technology feasible only in select locations.
To take advantage of the ubiquitous solar energy resource globally, decentralized
deployment of solar energy conversion technologies will be necessary. Solar photovoltaics can
be deployed at any scale without affecting the efficiency of the unit, the engineering cost for
their deployment and maintenance is far less, and they work well under ambient light
conditions.
16
There are, however, certain technological limitations to state-of-the-art
photovoltaics which have thus far prevented them from becoming the predominant energy
resource. First, power conversion efficiencies of single-junction cells have not yet met the
Golden
8
theoretical efficiency limit (33.7%).
18
Second, the best performing photovoltaics require a
large amount of material and the material and costs increase with the best performers.
4, 16
Third,
photovoltaics can be complicated to manufacture and the fabrication conditions for the cells
themselves require significant energy expenditures. Overcoming these limitations is the
primary concern in photovoltaics research; in the following paragraphs, different photovoltaics
technologies are compared in the light of these specific limitations. For the sake of consistency,
only single-junction devices that do not make use of solar concentrators are compared.
Building photovoltaic stacks with multiple junctions and including solar concentrators has
been shown to increase the efficiency of the overall device,
19
but these engineering techniques
do not address the materials-specific limitations of photovoltaics which this dissertation
attempts to parse.
The most common photovoltaics are based on silicon semiconductors. In these devices,
a layer of silicon absorbs a portion of the photons which are incident upon the surface of the
device and which have energies that are greater than the bandgap of the semiconducting silicon.
The active, light absorbing layer may vary from single crystalline to polycrystalline to
amorphous silicon, and doping to vary the bandgap energy is also possible; crystalline and
polycrystalline cells represent the bulk of the photovoltaics market, comprising over 90% of
commercial modules.
16
Crystalline silicon, however, is an indirect bandgap semiconductor and
as such has a low molar absorptivity; cells based on this material must be fabricated with a
large layer thickness, on the order of micrometers ( m), to absorb as much of the incident
photon flux as possible. Further, single crystalline cells are costly to manufacture, so the vast
majority of silicon photovoltaic cells are based on polycrystalline silicon. In polycrystalline
silicon, grain boundaries between crystals limit charge extraction compared to single
Molecular Design for Organic Photovoltaics
Chapter 1 | 9
crystalline devices, such that the record lab performance of a polycrystalline device is 22.3%,
while that of the single crystalline device is 26.7% (Figure 1.7).
19-20
Commercial modules tend
to perform at much lower efficiencies, with records of 19.9% and 24.4% for large-area
polycrystalline and single crystalline modules, respectively.
20
Record efficiencies are recorded
at normal photon incidence angle to the surface of the cell; oblique incidences result in lower
efficiencies due to a high degree of reflection from the poorly absorbing silicon devices.
21
This
results in much lower operating efficiencies for stationary silicon photovoltaic modules at the
early and late hours of the day, when the angle of the sun is not optimized to the surface of the
device. Including a mechanized rotor which tunes the angle of the module according to the
angle of the sun can improve the efficiency of the device throughout the day, but this is at
significant added installation and maintenance expense, bringing the cost per watt higher.
Although cheaper to produce than single crystalline silicon cells, polycrystalline silicon
photovoltaics still require relatively expensive fabrication conditions and a large amount of

Figure 1.7. Certified record photovoltaic efficiencies in various technological categories.
19

Golden
10
material in order to reach these efficiencies. It is only due to an economy of scale, given the
strength and maturity of the computer processing industry, that the cost of crystalline silicon
cell fabrication is as low as it currently is. Further inhibiting their wider deployment, these
brittle materials require glass backing for mechanical stability and hermetic sealing, usually
with fluorocarbon plastics, to provide long-term stability under operating conditions. The
resulting large, bulky, and expensive modules must be placed carefully to avoid damage or
obstruction of sight lines.
Alternative semiconductor materials, such as gallium arsenide (GaAs) and other group
III/V semiconductors, have shown device efficiencies of 28.8%, much closer to the theoretical
limit, due to bandgap energies closer to the ideal.
18-19
However, these materials are both
extremely costly to fabricate and have high toxicities requiring extensive encapsulation,
preventing them from being widely dispersed for commercial and residential solar
applications. Instead, these top-performing devices are suited well to aerospace applications
where increased watt per gram, rather than decreased dollar per watt, is the motivating
technological goal.
In order to reduce materials costs, devices based on thin film architectures have been
under development with limited commercial success (about 10% of the photovoltaics market).
Commercialized thin film photovoltaics include amorphous silicon (10.2% efficiency),
cadmium telluride (21%), and copper indium gallium diselenide (CIGS) (21.7%) based cells.
16,
20
Light absorption in these materials is one to two orders of magnitude greater than that of
crystalline silicon, so the active layer may be much thinner – only a few micrometers is
required for efficient light absorption. In addition, thin film technologies allow for simpler and
lower-energy fabrication conditions, leading to lower manufacturing costs and accelerated
Molecular Design for Organic Photovoltaics
Chapter 1 | 11
energy payback time (the operating time required before a solar cell has produced more energy
than it required to fabricate and install the device). Combined, the reduced materials and
fabrication costs have the potential to bring the cost per watt of thin film based photovoltaics
below that of crystalline silicon devices. However, these thin film semiconductor materials
tend to form polycrystalline domains which limit the efficiency of large area solar cells, while
reducing defect sites requires more complex manufacturing resulting in increased fabrication
costs. Large area CdTe and CIGS cells therefore have markedly lower efficiencies (18.6% and
15.7%, respectively) compared to laboratory scale devices. Concerns about the toxicity of
cadmium and the scarcity of tellurium have also stood as an impediment to their commercial
success. Nevertheless, CdTe photovoltaics are currently used in some of the world’s largest
solar farms including the Topaz Solar Farm (550 MW) in California and the Agua Caliente
Solar Project (290 MW) in Arizona, the panels for each of these solar farms having been
manufactured by First Solar, a CdTe photovoltaics manufacturing company in California.
Strategies for the recycling of CdTe photovoltaics have helped to mitigate concerns about the
scarcity of tellurium; it is estimated that recycled CdTe photovoltaics will be able to meet the
entirety of tellurium demand for the photovoltaics industry by 2038.
22

Emerging thin film photovoltaics based upon nanostructured inorganic/hybrid
materials such as perovskites, quantum dots, and dye-sensitized solar cells as well as others
based upon amorphous organic active layer materials are in various stages of materials R&D
and device optimization studies. Emerging thin film photovoltaics have the potential to
increase watt production per gram, reduce module costs from both a materials and a fabrication
perspective, introduce flexible and novel form factors to commercial modules, and be designed
with visual transparency thereby allowing novel deployment opportunities that the large,
Golden
12
heavy, and opaque wafer based technologies cannot offer. Technological limitations to
commercialization vary between the different thin film technologies, but generally include
device degradation, lowered theoretical conversion efficiencies,
23
and morphological
instability. Despite these limitations, emerging thin film photovoltaics have the best likelihood
of reaching grid parity in the near future due to their low materials and manufacturing costs,
provided device degradation is mitigated, module efficiencies approach thermodynamic limits,
and production methods are designed to provide morphological stability under operating
conditions. Materials design will play a role in overcoming each of these obstacles to
commercialization.
The research disseminated throughout the rest of this work encompasses materials
design specifications and the methodologies used to synthesize target structures for application
to organic photovoltaics (OPVs). OPVs are a subset of the emerging class of thin film
photovoltaics with active layers composed of non-toxic and atom-abundant materials with high
light absorptivity; exceptional absorptivity in OPVs allows for thin film fabrication on flexible
substrates with inexpensive and energy-efficient fabrication techniques such as roll-to-roll
processing and physical vapor deposition. In an analogy to the growth of the silicon
photovoltaics industry on the heels of the computer processing industry, OPV manufacturing
will benefit greatly from the growth of the organic light-emitting diode (OLED) industry,
which over the past decade has developed processing conditions capable of producing
hundreds of millions of commercial OLED displays – “modules” consisting of many individual
OLED devices – per year. The device manufacturing specifications for OLEDs share many
commonalities with those of OPVs, so development of these industrial processes may serve to
kick-start growth in the OPV manufacturing industry. The following Section 1.3 details the
Molecular Design for Organic Photovoltaics
Chapter 1 | 13
mechanism of photocurrent generation in organic photovoltaics and the sources of
photovoltage losses which will be used to develop design principles for the synthesis of active
layer materials capable of increasing the power conversion efficiency of OPVs.
1.3 Photocurrent Generation in Organic Photovoltaics
Photocurrent is an electrical current produced by the conversion of light energy into
electronic kinetic energy. For photocurrent generation to occur, a material (the active layer)
must absorb light, promoting an electron from the ground state to an excited state, and that
electron must migrate away from the point of initial absorption, forming two charge carriers:
a hole and an electron, which are coulombically independent of each other and thus deemed
“free carriers.” Finally, these carriers must migrate to and make electrical contact with the
anode and cathode of the device, respectively to complete an electrical circuit. In
semiconductors, this process is observed in its simplest form, as the initially formed excited
state (exciton) has a negligible excitonic binding energy, such that it rapidly forms free carriers
which may be collected by the corresponding electrodes. A schematic of a simple
semiconducting photovoltaic is depicted in Figure 1.8.

Figure 1.8. Mechanism of photocurrent generation in a semiconductor photovoltaic. In the first fundamental
process, the active layer absorbs photons with energy greater than or equal to the bandgap of the semiconductor,
promoting an electron from the valence band to the conduction band. The resulting exciton has negligible binding
energy, rapidly forming free carriers in the second fundamental process. Finally, the free carriers migrate to their
corresponding electrodes.

Golden
14
There are physical and materials-specific constraints limiting the probability of each of
the three fundamental processes (photon absorption, carrier formation, and charge collection)
necessary for photocurrent generation. Shockley and Queisser analyzed the physical
limitations alone, finding that the theoretical maximum power conversion efficiency for any
single-junction photovoltaic at the Earth’s surface is 33.7%.
18
Physical limitations on the
power conversion efficiency of a photovoltaic include geometric and atmospheric limitations
on the solar photon flux at the Earth’s surface, blackbody radiation from the cell itself,
recombination losses, and thermalization losses. Accounting for these factors, an ideal bandgap
energy, which accounts for the balance between current and voltage, can be extrapolated from
the efficiency limit; the value for this bandgap energy is materials independent and constrained
to 1.34 eV.
In an organic photovoltaic, the same physical considerations must be accounted for,
but because OPVs rely on organic materials for active layer absorption, additional fundamental
processes are necessarily involved in photocurrent generation and these processes again
constrain the theoretical efficiency of the cell. Molecular absorbers, unlike semiconductor
absorbers, have a significant exciton binding energy, meaning that after photon absorption, the
resulting carrier is a bound electron-hole pair, a Frenkel exciton, with a large binding energy
(between 0.3 and 1 eV).
24
This exciton is an energy carrier, rather than a charge carrier, and it
must migrate to a heterojunction with a correspondingly large energetic loss built in,
whereupon it may undergo charge transfer between a donor and an acceptor molecule, forming
a hole on the donor and an electron on the acceptor. At this stage, the electron and hole still
experience significant coulombic interaction, leading to a reduced probability of charge
separation. If charge separation occurs, then charge migration to the electrodes must follow to
Molecular Design for Organic Photovoltaics
Chapter 1 | 15
produce photocurrent. The mechanism of photocurrent generation in an OPV is depicted in
Figure 1.9. Instead of three fundamental process required for photocurrent generation in a
semiconducting photovoltaic, and OPV requires five fundamental processes: photon
absorption, exciton migration, charge transfer, charge separation, and charge
migration/collection.
Each process required for photocurrent generation imposes a limitation on the
maximum theoretical power conversion efficiency limit for an OPV, where the general
equation for power conversion efficiency in a photovoltaic is given by Equation 1.1, by

Figure 1.9. (a) A schematic of a lamellar heterojunction organic photovoltaic, (b) singlet exciton formation in a
molecular absorber, and (c) the mechanism of photocurrent generation in an OPV. The mechanism involves five
fundamental steps: (1) absorption of a photon resulting in an exciton, (2) exciton migration, (3) charge transfer,
(4) charge separation, and (5) charge migration followed by charge collection at the corresponding electrodes.

Golden
16
limiting one of more of the governing variables (open-circuit voltage, short circuit current, or
fill factor) in the power conversion efficiency equation. The additional mechanistic steps
compared to photocurrent generation in a semiconducting cell, namely exciton migration,
charge transfer, and charge separation, restrict the theoretical efficiency somewhat further than
that of a semiconductor-based photovoltaic; i.e. the theoretical efficiency of an excitonic
photovoltaic is less than that of a semiconducting photovoltaic.
24-27
The theoretical maximum
power conversion efficiency in an OPV has been estimated to be between 23 and 27%
depending on the magnitude of the driving force for charge transfer E built into the
donor/acceptor interface.
23-24, 26, 28
In order to approach this theoretical efficiency, active layer
materials must be designed to meet the assumptions leading to maximum efficiency which
underlie each fundamental step. These processes are outlined in the following subsections.
1.3.1 Photon Absorption
One of the major assumptions underlying the Shockley-Queisser limit is that all
photons incident upon the surface of the device with energies above the bandgap are absorbed.
The average incident solar flux at the Earth’s surface is defined in the AM1.5 spectrum (Figure
1.10). This spectrum accounts for atmospheric and geometric considerations which limit the
number and energy of photons which strike the Earth’s surface. An efficient active layer
absorber will have a sufficiently high probability for photon absorption such that all incident
photons are absorbed before passing through the device. With respect to the solar flux, the
active layer material must therefore be highly absorbing in the visible region and moderately
𝜂 =
𝑉 𝑂𝐶
𝐽 𝑆𝐶
𝐹𝐹
𝑃 𝑖𝑛

Equation 1.1. Power conversion efficiency in a photovoltaic is defined as the product of the open circuit
voltage (V OC), the short circuit current (J SC), the fill factor (FF), and the inverse of the incident photon flux P in.
Molecular Design for Organic Photovoltaics
Chapter 1 | 17
absorbing in the near infrared. The
intensity of photon absorption is
defined by a molecule’s molar
absorptivity (formally, the molar
attenuation coefficient, ), a
materials-specific property which is
a function of a set of selection rules
defining the probability for
electronic excitation. The first of
these rules is the spin selection rule,
which states that allowed transitions
result in the excitation of electrons
without a change in their spin
quantum number. The result of this
rule is that, for closed-shell
molecular species with low spin-orbit coupling, the most probable transitions are singlet
transitions. The second rule is the Laporte or orbital rule, which states that the symmetry of the
electronic distribution must change for a transition to be allowed. Thus, molecular species must
be deigned with singlet transitions in the visible and near infrared that are defined by a change
in electronic symmetry. Third, the probability for an electronic excitation increases with the
extent of wavefunction overlap between the involved orbitals. This results in the intensities of
allowed transitions in organic photovoltaics typically being much higher than those of
semiconducting materials, where wavefunction delocalization decreases overlap.

Figure 1.10. Incident photon flux (top) per photon energy and
spectral irradiance per photon wavelength (bottom) at the
Earth’s surface.

Golden
18
The energy of the absorption of a molecular species trends as a function of the extent
of electron delocalization, with localized electronic states generally resulting in higher-energy
(lower wavelength) transitions, while delocalized, highly conjugated electronic states result in
low-energy excitations. For molecular species with discrete orbitals, absorptions tend to be
narrow; this is especially true in comparison with ordered solid-state materials, where the
periodicity of the atomic lattice gives rise to orbital sets with closely matched energies which
comprise a continuum, thus resulting in a wide distribution of probable electron energies and,
therefore, allowed transition energies. The width of an absorption manifold may be tuned by
incorporating disorder, where a more disordered population of molecular absorbers will give
rise to a wider distribution of allowed absorption energies (albeit at the expense of absorption
intensity compared to a uniform population), or by incorporating absorbers with multiple
highly allowed singlet transitions, i.e. S0 to S1, S2, S3, Sn....
The absorption of a photon in both a condensed matter solid and a molecular absorber
results in the formation of an exciton – a coulombically bound electron-hole pair. In the case
of a solid-state structure such as semiconducting silicon or cadmium telluride, for example, the
continuous band structure results in a delocalized Wannier-Mott type excitonic state with a
low exciton binding energy, on the order of the ambient thermal energy kT, such that the
formation of free carriers upon photon absorption is a spontaneous process.
26
By contrast, in a
molecular absorber, the exciton formed is a Frenkel exciton, defined by a high degree of
confinement, typically to a single molecule even in a bulk film, and a high exciton binding
energy between 0.3 and 1 eV.
24

The initially formed excitonic state has sufficient time to relax via vibrational and
rotational modes to a Frank-Condon excited state; that is, the lowest energy excited state which
Molecular Design for Organic Photovoltaics
Chapter 1 | 19
defines the bandgap energy. This thermalization process results in the loss of photon energy in
the form of heat to the surroundings and is a significant source of energy loss in single-junction
photovoltaics. There is a balance between the number of photons which may be collected, then,
and the thermalization energy loss. High bandgap absorbers result in less thermalization losses,
but fail to absorb low energy photons at all. Low bandgap absorbers may collect all incident
photons but a large amount of energy is lost in the form of heat as the resultant excitons relax.
There is a defined bandgap energy where the trade-off between absorption and thermalization
is optimized. This optimal bandgap energy Eg varies slightly depending on the heterojunction
architecture, with values between 1.35 and 1.5 eV leading to efficiencies between 27% and
22% in organic photovoltaics.
23
In a semiconducting photovoltaic, the bandgap energy is
simply defined as the energy difference between the valence and conduction bands. In an
organic photovoltaic, where there are two absorbing species in the active layer, the E g is taken
to be the optical bandgap E00, often equal to the lowest singlet excitation energy, of the lowest
energy absorber.
26
The maximum possible photocurrent JSC in an ideal device is equal to the
number of solar flux of photons with energies equal to or above the bandgap energy.
From a materials design perspective, the photon absorption step can be optimized by
incorporating active layer materials with E00 energies between 1.34 and 1.5 eV, where the
lowest energy excitation is a singlet excitation with a high oscillator strength (probability for
excitation). Likely this will involve the design of highly conjugated aromatic, heteroaromatic,
or pseudo-aromatic structures with extended -systems. Red-shifting the E00 of a molecular
absorber can also be accomplished by incorporating donor-acceptor moieties into the molecule,
where appending donors at sites of HOMO density will destabilize the HOMO, and appending
acceptors at sites of LUMO density with stabilize the LUMO, the net effect being a decrease
Golden
20
in the HOMO-LUMO energy gap which generally correlates with a decrease in the optical gap.
Materials designed with broad absorptions or with multiple high-probability singlet
absorptions will give rise to more complete absorption of the entire photon flux.
1.3.2 Exciton Migration
Upon photoexcitation of a molecular absorber, a Frenkel exciton is formed. This
electron-hole pair occupies closely associated orbitals and consequently there is a large
associated coulombic binding energy preventing the formation of free carriers. In a molecular
absorber, the exciton itself is often localized onto a single chromophore, though exciton
delocalization can be increased by increasing electronic coupling between neighboring
chromophores. The energy carrier in an organic photovoltaic, then, begins the process of
photocurrent generation as an exciton. This exciton must migrate to an interface wherein a
donor and acceptor moiety are closely electronically coupled, and the heterojunction must be
designed such that the driving force E associated with charge transfer is greater than the
exciton binding energy for charge carriers to form. Fundamental to this process is migration of
the exciton from the point of photoexcitation to an appropriate interface at the donor-acceptor
heterojunction.
The mechanism of exciton migration can be described by either Förster-resonance
energy transfer (FRET) or Dexter energy transfer (DET), where FRET is the dominant
mechanism in the transfer of singlet excitons in a molecular absorber. FRET is a coulombic
energy transfer process wherein the relaxation of an excited electron to the ground state induces
a resonant excitation of a ground state electron into an excited state in a nearby chromophore
with an equivalent or smaller E00. This effect attenuates with distance as a factor of 1/(r)
6
where
r is the distance between excited and ground state. FRET and takes the form of a random hop
Molecular Design for Organic Photovoltaics
Chapter 1 | 21
and does not necessarily lead to migration toward a polar interface. FRET efficiency, due to
its dependence on electronic transitions, can be increased by increasing the molar absorptivity
and by increasing the photoluminescent quantum yield; further, greater overlap between the
absorption and emission spectrum yields a higher probability for resonant energy transfer, so
materials with small Stokes shifts are ideal.
Dexter energy transfer is another possible energy transfer mechanism. In DET, energy
is transferred due to direct wavefunction overlap between neighboring chromophores, wherein
simultaneous exchange of an excited state electron from a photoexcited chromophore and a
ground state electron from a ground state chromophore are formally exchanged. This
mechanism attenuates exponentially with distance due to its dependence on strong
interchromophoric wavefunction overlap for electronic exchange to occur. Thus it is by far the
minority mechanism for energy transfer in OPVs designed for the formation of charge carriers
from singlet excitons; in triplet-only devices, however, such as those based on the principles
of singlet fission (formation of two low-energy triplet excitons from a single photoexcited
singlet exciton), FRET is not spin-allowed, and DET becomes the primary mechanism for
photocurrent generation.
In a lamellar OPV such as the one depicted in Figure 1.9a, the distance between
absorption site and heterojunction is often on the order of tens of nanometers. The average
distance for exciton migration LD, defined as the distance migrated by the time 1/e or
approximately 35% of the excited state population has decayed to the ground state, is often as
low as 10 nm.
24
Efforts have been made to engineer bulk heterojunctions which maximize
donor/acceptor interfacial area by incorporating interdigitating layers of donor and acceptor.
Precise control of these interfaces, however, is difficult to maintain and is highly dependent on
Golden
22
material properties and processing conditions. Materials design strategies to increase the
probability of exciton migration to the donor/acceptor interface include increasing the size of
the photogenerated exciton through the synthesis of covalent chromophoric networks,
increasing the lifetime of the excited state by restricting geometric distortion, and increasing
molar absorptivity (therefore allowing for the deposition of thinner films).
1.3.3 Charge Transfer
The charge transfer step is closely correlated with the magnitude of the open circuit
voltage VOC.
29-32
In an ideal case where the donor-acceptor CT state has an absorptivity 
equal to zero, recombination is minimized and the required driving force - E for charge
transfer may approach zero and still result in effective exciton disassociation.
26
In this limiting
case, the efficiency limit of an OPV is nearly identical with that of an inorganic semiconducting
device. In practice, however, the absorptivity of molecular donor-acceptor CT states is finite
and, even for small values of a between 10
-5
and 10
-3
, correlates to significant energy losses
via recombination due to the direct relationship between absorption efficiency and
recombination efficiency. Further, due to the excitonic nature of OPVs and low dielectric
screening in organic thin films, a driving force E must be engineered into the donor/acceptor
heterojunction to efficiently separate charges and the greater the magnitude of this driving
force, the lower the VOC.
VOC may be maximized by reducing the CT absorptivity CT and and/or by reducing
the magnitude of the driving force built in to initiate charge transfer. Predicting CT is not
simple, however, as it is highly dependent on the wavefunction overlap between donor and
acceptor at the interface, which varies significantly with local interfacial morphology. The
required driving force for charge transfer from donor to acceptor may be estimated given
Molecular Design for Organic Photovoltaics
Chapter 1 | 23
enthalpic comparison between the optical LUMO of the donor, measured by adding the E00
energy to the HOMO energy, and the electron affinity of the acceptor. If the optical LUMO
energy of the donor is greater than the electron affinity of the acceptor, charge transfer is
exothermic.
24, 33
To reduce thermalization losses, the donor and acceptor should be judiciously
chosen such that energy loss at the donor-acceptor charge transfer step does not exceed the
minimum driving force required for CT.
One mechanism which has been successfully used to decrease the required driving
force for charge transfer is symmetry breaking charge transfer (SBCT).
34
In SBCT, a locally
excited state formed on one of two closely associated and identical ligands decays by an
intramolecular charge transfer process to form a hole on one (oxidized) chromophore and an
electron on the opposite (reduced) chromophore.
35-42
This electron hole pair is defined by
minimal coupling due to minimal wavefunction overlap. Typically this process is only
exothermic in polar media, for example at a local interface between donor and acceptor in an
OPV. SBCT-type materials may be used in either (or both) the donor and acceptor layers of an
OPV, wherein the nature of the excited state in the bulk of the film is a locally excited state

Figure 1.11. Photoexcitation of an SBCT-type chromophoric dyad in a polar environment induces spontaneous
symmetry breaking charge transfer, wherein charge is transferred from one chromophore in the dyad to the other.
This intramolecular charge transfer state facilitates charge transfer to an adjacent acceptor material. The resulting
electron-hole distance is larger than if the donor-acceptor CT state were formed directly from the photoexcited
state, and therefore back electron recombination is minimized.

Golden
24
which may migrate toward the interface by FRET, whereupon SBCT occurs, facilitating
formation of the donor acceptor charge-transfer state.
1.3.4 Charge Separation
Upon formation of the donor-acceptor CT state, the electron and hole remain closely
associated in space and the probability of back-electron recombination is high. Inefficient
charge separation due to back electron recombination is another major loss mechanism in
OPVs and has been shown to limit the VOC compared to semiconducting photovoltaics. In
addition to facilitating the forward charge transfer process in the previous mechanistic step,
SBCT has been shown to retard back electron recombination. It is thought that this is a product
of the through-space separation of hole and electron upon donor acceptor CT from an SBCT
complex (Figure 1.11). Whereas recombination losses typically limit OPVs to V OC values up
to 0.6 eV lower than the CT state energy, open circuit voltages in OPVs designed with SBCT
materials in the active layer have been shown to have recombination rates equal to those of
highly efficient semiconducting photovoltaics, leading to losses of only 0.37 V.
34
The
incorporation of symmetry-breaking charge transfer materials into the donor or acceptor layers
of OPVs represents a significant achievement in limiting energetic losses associated with the
charge transfer and charge separation steps of photocurrent generation in organic
photovoltaics.
1.3.5 Charge Collection
Upon formation of the charge separated state, the charges, now efficiently screened
from one another by distance, must migrate toward their corresponding electrodes to complete
the photocurrent generation cycle. This process follows a localized hopping mechanism
described by the nonadiabatic Marcus theory for electron transfer.
43-44
In this description, the
Molecular Design for Organic Photovoltaics
Chapter 1 | 25
radical ion induces a polarization in its neighboring chromophores which is followed by an
internal reorganization required for the ion and the accepting ground-state chromophore to
reach the same geometry. After this reorganization step, charge transfer may occur and the
process repeats itself until the charge reaches the electrode interface, encounters a trap site, or
migrates back toward the donor-acceptor interface and is lost through recombination.
Reorganization, trap sites, and back-electron recombination each represent probable energy
loss mechanisms.
Both reorganization energy loss and loss of carriers due to trap sites can be reduced by
increasing the crystallinity of the active layer;
24, 45
however, in practice, increasing organic thin
film crystallinity often results in the formation of “islands” – that is, microcrystalline domains
which protrude vertically from the film and can induce major defects leading to shirt-circuit
conditions in the device. Efforts to increase molecular order in organic thin films are ongoing
and not limited to organic photovoltaics; in OLEDs as well, molecular order has the potential
to increase external quantum efficiencies. Predicting the order or orientation of individual
molecules in a vapor-deposited film is non-trivial, but long-range order can generally be
induced by using a combination of thoughtfully engineered substrates with inherently
anisotropic molecular materials.
46

The high carrier mobilities and electric field effects inherent to a semiconducting
photovoltaic such as single-crystal silicon approach the ideal physical situation to minimize
back-electron recombination. In such systems, recombination losses are limited to about
0.36 eV, whereas a detailed balance analysis suggests the absolute minimum energy loss from
recombination is 0.3 eV.
18, 25
This value represents a situation in which non-radiative
recombination (through formation of triplet states, trap states, etc.) is negated entirely, and
Golden
26
radiative recombination is the only allowed recombination mechanism. It has been shown that
utilizing SBCT materials in either the donor or acceptor layer in combination with the
introduction of buffer materials to individually facilitate hole and electron conductivities to the
anode and cathode, respectively, limits recombination losses in OPVs to 0.37 eV,
34
which is
the same value as observed in semiconducting photovoltaics and approaches that of the detailed
balance limit.
1.4 State of the Art in Organic Photovoltaic Active Layer Materials
With few exceptions,
47-49
significant achievements in efficiencies for OPVs have lately
been consistently met by donor polymer / fullerene acceptor architectures in which the donor
material is a low-bandgap donor-acceptor type material and the fullerene is a C70 derivative.
19-
20, 50
The benefit of the donor-acceptor architecture is exceptional bandgap tunability, as the
HOMO and LUMO can be independently modified by altering the donor and acceptor
moieties, respectively. Although polymer OPVs are readily tunable, they limit device
fabrication techniques to solution processing methods. Further, simple fullerenes are cost-
prohibitive for large-sale manufacturing due to major challenges in purification, while the best-
performing fullerene derivatives, such as PC70BM (Figure 1.12) compound this problem due
to added synthetic steps and reach in some cases over $2000/gram from major suppliers. If
OPVs are to meet grid parity, materials and manufacturing costs must be reduced.
To reduce processing costs and maintain high batch-to-batch uniformity, vacuum
thermal evaporation (VTE) is the most efficient processing technique because it allows for
highly controlled deposition of thin films across a large surface area with low energy costs.
VTE requires materials which are stable to sublimation at moderate temperatures >100 ºC
while remaining in the solid state at room temperature under high vacuum (10
-7
torr).
Molecular Design for Organic Photovoltaics
Chapter 1 | 27
Generally, materials design considerations to meet the requirements for VTE include:
moderate molecular weights (200-900 a.m.u.), zero net charge, and moderate thermal stability.
In order to reduce the cost per watt in OPVs, novel donor and acceptor materials must
be designed; specifically, low-bandgap small molecule donors and molecular acceptors which
can be processed by VTE, which exhibit charge transfer and carrier mobility behavior which
is the same as or better than that of fullerenes must be developed. Barring these developments,
it is unlikely that OPVs will meet the cost requirements for commercial implementation.
1.5 Molecular Design Principles for Organic Photovoltaics
This dissertation focuses on the development of molecular semiconducting materials
designed to meet the specific challenges in organic photovoltaics. Materials design
considerations to accomplish intense and broad-band photon absorption throughout the solar
spectrum, efficient exciton migration, and efficient charge transfer, separation, and collection
are developed throughout the following chapters. Small molecule heteroaromatic
chromophores such as dipyridylmethenes (DIPYRs) (Chapters 2-4), dipyrromethenes
(Chapters 5-6), and porphyrins (Chapter 6) are the primary focus of this work. The role of
molecular and electronic structure is discussed in the context of tuning state energies to modify

Figure 1.12. Typical fullerene acceptors for OPVs: (left to right) C 60, PC 60BM, and PC 70BM.

Golden
28
absorption, internal conversion, intersystem crossing, and intramolecular charge transfer
processes. In Chapter 7, these same principles are applied to solving specific challenges in
organic light-emitting diodes.
1.6 Bibliography
1. Stocker, T. F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels,
Y. Xia, V. Bex and P.M. Midgley (Eds.) Climate Change 2013: The Physical Science Basis.
Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental
Panel on Climate Change; Cambridge University Press, Cambridge, United Kingdom and
New York, NY, USA, 2013; p 1535.
2. Paris Agreement: 21st Conference of the Parties of the UNFCCC. United Nations
Framework Convention on Climate Change: Paris, 2015.
3. Crowley, T. J.; North, G. R., Abrupt Climate Change and Extinction Events in Earth
History. Science 1988, 240 (4855), 996.
4. Edenhofer, O., R. Pichs-Madruga, Y. Sokona, E. Farahani, S. Kadner, K. Seyboth, A.
Adler, I. Baum, S. Brunner, P. Eickemeier, B. Kriemann, J. Savolainen, S. Schlömer, C. von
Stechow, T. Zwickel and J.C. Minx (Eds.) Climate Change 2014: Mitigation of Climate
Change. Contribution of Working Group III to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change; Cambridge University Press, Cambridge, United
Kingdom and New York, NY, USA, 2014; p 1436.
5. Boden, T. A., Andres, R.J., Global, Regional, and National Fossil-Fuel CO2 Emissions.
Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, U.S.
Department of Energy, Oak Ridge, Tenn., U.S.A 2017.
6. Raftery, A. E.; Zimmer, A.; Frierson, D. M. W.; Startz, R.; Liu, P., Less than 2 °C
warming by 2100 unlikely. Nat. Clim. Change 2017, 7, 637.
7. Mauritsen, T.; Pincus, R., Committed warming inferred from observations. Nat. Clim.
Change 2017, 7, 652.
8. NASA/GISS Global Climate Change: Vital Signs of the Planet. (accessed May 5 2018).
9. Field, C. B., Barros, V. R., Dokken, D. J., Mach, K. J., Mastrandrea, M. D., Bilir, T.
E., Chatterjee, M., Ebi, K. L., Estrada, Y. O., Genova, R. C., Girma, B., Kissel, E. S., Levy, A.
N., MacCracken, S., Mastrandrea, P. R., and White, L. L. (Eds) Climate Change 2014 –
Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fifth
Assessment Report of the Intergovernmental Panel on Climate Change; Cambridge University
Press, Cambridge, United Kingdom and New York, NY, USA, 2014; p 1820.
10. Millero, F. J., Thermodynamics of the carbon dioxide system in the oceans. Geochim.
Cosmochim. Acta 1995, 59 (4), 661-677.
11. Jacobson Mark, Z., Studying ocean acidification with conservative, stable numerical
schemes for nonequilibrium air‐ocean exchange and ocean equilibrium chemistry. J. Geophys.
Res. Atmos. 2005, 110 (D7).
12. Prada, F.; Caroselli, E.; Mengoli, S.; Brizi, L.; Fantazzini, P.; Capaccioni, B.; Pasquini,
L.; Fabricius, K. E.; Dubinsky, Z.; Falini, G.; Goffredo, S., Ocean warming and acidification
synergistically increase coral mortality. Sci. Rep. 2017, 7, 40842.
Molecular Design for Organic Photovoltaics
Chapter 1 | 29
13. International Energy Outlook 2016. U.S. Energy Information Administration, U.S.
Department of Energy: Washington, DC, 2016; p 290.
14. Perez, M., Perez, R. UPDATE 2015 -- A Fundamental Look at Supply Side Energy
Reserves For the Planet; IEA-SHCP-Newsletter: 2015.
15. BP Statistical Review of World Energy; 2017.
16. Amatya, R. B., F.; Campanella, A.; Kavlak, G., Macko, J.; Maruano, A.; McNerney,
J.; Osedach, T.; Rodilla, P.; Rose, A.; Sakti, A.; Steinfeld, E.; Trancik, J.; Tuller H., The Future
of Solar Energy: An Interdisciplinary MIT Study. Massachusetts Institute of Technology: 2015.
17. Average Energy Prices, Los Angeles-Long Beach-Anaheim – April 2018 : Western
Information Office : U.S. Bureau of Labor Statistics. 2018.
18. Shockley, W.; Queisser, H. J., Detailed Balance Limit of Efficiency of p‐n Junction
Solar Cells. J. Appl. Phys. 1961, 32 (3), 510-519.
19. Best Research-Cell Efficiencies. https://www.nrel.gov/pv/assets/images/efficiency-
chart.png (accessed May 15, 2018).
20. Green Martin, A.; Hishikawa, Y.; Dunlop Ewan, D.; Levi Dean, H.; Hohl‐Ebinger, J.;
Ho‐Baillie Anita, W. Y., Solar cell efficiency tables (version 51). Prog. Photovolt. Res. Appl.
2017, 26 (1), 3-12.
21. Yen-Jen, H.; Shao-Sun, H.; Yu-Ting, W.; Chun-Hsiang, C.; Liang-Yi, C.; Liang-Yu,
S.; Jian-Jang, H., Polarization dependent solar cell conversion efficiency at oblique incident
angles and the corresponding improvement using surface nanoparticle coating.
Nanotechnology 2011, 22 (48), 485202.
22. Marwede, M.; Reller, A., Future recycling flows of tellurium from cadmium telluride
photovoltaic waste. Resour., Conserv. Recycl. 2012, 69, 35-49.
23. Giebink, N. C.; Wiederrecht, G. P.; Wasielewski, M. R.; Forrest, S. R., Thermodynamic
efficiency limit of excitonic solar cells. Phys. Rev. B 2011, 83 (19), 195326.
24. Schlenker, C. W.; Thompson, M. E., The molecular nature of photovoltage losses in
organic solar cells. Chem. Commun. 2011, 47 (13), 3702-3716.
25. Menke, S. M.; Ran, N. A.; Bazan, G. C.; Friend, R. H., Understanding Energy Loss in
Organic Solar Cells: Toward a New Efficiency Regime. Joule 2018, 2 (1), 25-35.
26. Gruber, M.; Wagner, J.; Klein, K.; Hörmann, U.; Opitz, A.; Stutzmann, M.; Brütting,
W., Thermodynamic Efficiency Limit of Molecular Donor‐Acceptor Solar Cells and its
Application to Diindenoperylene/C60‐Based Planar Heterojunction Devices. Adv. Energy
Mater. 2012, 2 (9), 1100-1108.
27. Elumalai, N. K.; Uddin, A., Open circuit voltage of organic solar cells: an in-depth
review. Energ. Environ. Sci. 2016, 9 (2), 391-410.
28. Janssen René, A. J.; Nelson, J., Factors Limiting Device Efficiency in Organic
Photovoltaics. Adv. Mater. 2012, 25 (13), 1847-1858.
29. Zou, Y.; Holmes, R. J., Correlation between the Open-Circuit Voltage and Charge
Transfer State Energy in Organic Photovoltaic Cells. ACS Appl. Mater. Interfaces 2015, 7 (33),
18306-18311.
30. Guan, Z.; Li, H.-W.; Cheng, Y.; Yang, Q.; Lo, M.-F.; Ng, T.-W.; Tsang, S.-W.; Lee,
C.-S., Charge-Transfer State Energy and Its Relationship with Open-Circuit Voltage in an
Organic Photovoltaic Device. J. Phys. Chem. C 2016, 120 (26), 14059-14068.
31. Vandewal, K.; Tvingstedt, K.; Manca, J. V.; Inganäs, O., Charge-Transfer States and
Upper Limit of the Open-Circuit Voltage in Polymer:Fullerene Organic Solar Cells. IEEE J.
Sel. Top. Quantum Electron. 2010, 16 (6), 1676-1684.
Golden
30
32. Hörmann, U.; Kraus, J.; Gruber, M.; Schuhmair, C.; Linderl, T.; Grob, S.; Kapfinger,
S.; Klein, K.; Stutzman, M.; Krenner, H. J.; Brütting, W., Quantification of energy losses in
organic solar cells from temperature-dependent device characteristics. Phys. Rev. B 2013, 88
(23), 235307.
33. Ward Alexander, J.; Ruseckas, A.; Kareem Mohanad, M.; Ebenhoch, B.; Serrano Luis,
A.; Al‐Eid, M.; Fitzpatrick, B.; Rotello Vincent, M.; Cooke, G.; Samuel Ifor, D. W., The
Impact of Driving Force on Electron Transfer Rates in Photovoltaic Donor–Acceptor Blends.
Adv. Mater. 2015, 27 (15), 2496-2500.
34. Bartynski, A. N.; Gruber, M.; Das, S.; Rangan, S.; Mollinger, S.; Trinh, C.; Bradforth,
S. E.; Vandewal, K.; Salleo, A.; Bartynski, R. A.; Bruetting, W.; Thompson, M. E., Symmetry-
Breaking Charge Transfer in a Zinc Chlorodipyrrin Acceptor for High Open Circuit Voltage
Organic Photovoltaics. J. Am. Chem. Soc. 2015, 137 (16), 5397-5405.
35. Golden, J. H.; Estergreen, L.; Porter, T.; Tadle, A. C.; Sylvinson M. R, D.; Facendola,
J. W.; Kubiak, C. P.; Bradforth, S. E.; Thompson, M. E., Symmetry-Breaking Charge Transfer
in Boron Dipyridylmethene (DIPYR) Dimers. ACS Appl. Energy Mater. 2018.
36. Trinh, C.; Kirlikovali, K.; Das, S.; Ener, M. E.; Gray, H. B.; Djurovich, P.; Bradforth,
S. E.; Thompson, M. E., Symmetry-Breaking Charge Transfer of Visible Light Absorbing
Systems: Zinc Dipyrrins. J. Phys. Chem. C 2014, 118 (38), 21834-21845.
37. Vauthey, E., Photoinduced Symmetry-Breaking Charge Separation. ChemPhysChem
2012, 13 (8), 2001-2011.
38. Whited, M. T.; Patel, N. M.; Roberts, S. T.; Allen, K.; Djurovich, P. I.; Bradforth, S.
E.; Thompson, M. E., Symmetry-breaking intramolecular charge transfer in the excited state
of meso-linked BODIPY dyads. Chem. Commun. 2012, 48 (2), 284-286.
39. Wu, Y.; Young, R. M.; Frasconi, M.; Schneebeli, S. T.; Spenst, P.; Gardner, D. M.;
Brown, K. E.; Würthner, F.; Stoddart, J. F.; Wasielewski, M. R., Ultrafast Photoinduced
Symmetry-Breaking Charge Separation and Electron Sharing in Perylenediimide Molecular
Triangles. J. Am. Chem. Soc. 2015, 137 (41), 13236-13239.
40. Ivanov, A. I.; Dereka, B.; Vauthey, E., A simple model of solvent-induced symmetry-
breaking charge transfer in excited quadrupolar molecules. J. Chem. Phys. 2017, 146 (16),
164306.
41. Piet, J. J.; Schuddeboom, W.; Wegewijs, B. R.; Grozema, F. C.; Warman, J. M.,
Symmetry Breaking in the Relaxed S1 Excited State of Bianthryl Derivatives in Weakly Polar
Solvents. J. Am. Chem. Soc. 2001, 123 (22), 5337-5347.
42. Cook, R. E.; Phelan, B. T.; Kamire, R. J.; Majewski, M. B.; Young, R. M.;
Wasielewski, M. R., Excimer Formation and Symmetry-Breaking Charge Transfer in Cofacial
Perylene Dimers. J. Phys. Chem. A 2017, 121 (8), 1607-1615.
43. Coropceanu, V.; Cornil, J.; da Silva Filho, D. A.; Olivier, Y.; Silbey, R.; Brédas, J.-L.,
Charge Transport in Organic Semiconductors. Chem. Rev. 2007, 107 (4), 926-952.
44. Nelson, J.; Kwiatkowski, J. J.; Kirkpatrick, J.; Frost, J. M., Modeling Charge Transport
in Organic Photovoltaic Materials. Acc. Chem. Res. 2009, 42 (11), 1768-1778.
45. Mazzio, K. A.; Luscombe, C. K., The future of organic photovoltaics. Chem. Soc. Rev.
2015, 44 (1), 78-90.
46. Schmidt, T. D.; Lampe, T.; Sylvinson M. R, D.; Djurovich, P. I.; Thompson, M. E.;
Brütting, W., Emitter Orientation as a Key Parameter in Organic Light-Emitting Diodes. Phys.
Rev. Appl. 2017, 8 (3), 037001.
Molecular Design for Organic Photovoltaics
Chapter 1 | 31
47. Zhao, W.; Li, S.; Yao, H.; Zhang, S.; Zhang, Y.; Yang, B.; Hou, J., Molecular
Optimization Enables over 13% Efficiency in Organic Solar Cells. J. Am. Chem. Soc. 2017,
139 (21), 7148-7151.
48. Zhang, Q.; Kan, B.; Liu, F.; Long, G.; Wan, X.; Chen, X.; Zuo, Y.; Ni, W.; Zhang, H.;
Li, M.; Hu, Z.; Huang, F.; Cao, Y.; Liang, Z.; Zhang, M.; Russell, T. P.; Chen, Y., Small-
molecule solar cells with efficiency over 9%. Nat. Photon. 2014, 9, 35.
49. Che, X.; Chung, C. L.; Hsu, C. C.; Liu, F.; Wong, K. T.; Forrest Stephen, R., Donor–
Acceptor–Acceptor's Molecules for Vacuum‐Deposited Organic Photovoltaics with Efficiency
Exceeding 9%. Adv. Energy Mater. 2018, 0 (0), 1703603.
50. Mori, S.; Oh-oka, H.; Nakao, H.; Gotanda, T.; Nakano, Y.; Jung, H.; Iida, A.; Hayase,
R.; Shida, N.; Saito, M.; Todori, K.; Asakura, T.; Matsui, A.; Hosoya, M., Organic photovoltaic
module development with inverted device structure. Mater. Res. Soc. Symp. Proc. 2015, 1737.

Golden
32
Chapter 2

Dipyridylmethene (DIPYR) Dyes: Shedding Light on
Pyridine-based Chromophores

2.1 Introduction to Dipyridylmethene Dyes
One of the most ubiquitous chromophores in modern dye chemistry is boron difluoride
dipyrromethene (4,4-difluoro-4-bora-3a,4a-diaza-s-indacene by IUPAC convention and
BODIPY in common use) (Figure 2.1). Over 20,000 papers and
patents report the properties of these materials and speak to their
diverse utility. They are used as biosensors, as fluorescent switches
and labels, as laser dyes, in energy conversion applications, as
active layer materials in photovoltaics,
1
and their utility continues
to grow each year as they are adapted for further applications. There
are BODIPY dyes reported with emission energies spanning the
visible region of the spectrum, from blue-emitting meso amino-substituted BODIPYs, to the
green-emissive parent structure, to yellow, orange, and red emitters involving increasingly
complex levels of alkyl and aryl substitutions about the core chromophore.
BODIPYs are derived from dipyrromethenes, which are commonly synthesized by
condensing two alpha-unsubstituted pyrroles onto a benzaldehyde or derivative thereof
(mesitaldehyde, for example), followed by treatment of the ligand with a quinone oxidizing
agent, DDQ, to fully conjugate the system (Scheme 2.1).
2
An alternative dipyrromethene
synthesis involves coupling the pyrroles directly onto an acid chloride, eliminating the need

Figure 2.1. 4,4-difluoro-
4-bora-3a,4a-diaza-s-
indacene (BODIPY)
labeled according to
convention

Molecular Design for Organic Photovoltaics
Chapter 2 | 33
for a subsequent oxidation step. In either case, the simplest synthetic methods result in the
formation of a dipyrromethene which is substituted with an aryl or alkyl group at the meso
position. Another substitution common to most dipyrromethenes is a high degree of alkylation,
especially at the alpha-positions of the pyrroles. This is necessary because the alpha-
unsubstituted pyrrole is susceptible to polymerization and formation of porphyrins both in the
bottle and during the dipyrromethene forming reaction. Further aliphatic substitutions at the
pyrroles prevent nucleophilic attack at reactive pyrrolic sites and impart useful qualities to the
ultimate BODIPY, such as red-shifting the absorption and emission energies and increasing
the photoluminescent quantum yield by blocking rotation of meso-aryl groups. The limited
range of synthetic methods to access the fully unsubstituted dipyrromethene and the instability
of this material makes it difficult to synthesize the parent BODIPY. Indeed, it was not until
2009, 41 years after the first characterization of a BODIPY dye by Treibs and Kreuzer,
3
that
three individual groups independently published methods describing the synthesis and
properties of the fully unsubstituted parent BODIPY.
4-6

Scheme 2.1. BODIPY is commonly synthesized by condensation of two pyrroles onto an aldehyde followed by
oxidation with a quinone (usually 2,3-Dichloro-5,6-Dicyanobenzoquinone (DDQ)) or by reaction of pyrroles with
an acid chloride. The resulting dipyrromethene is then treated with a tertiary amine and BF 3OEt 2 to yield the final
product.

Golden
34
The volume of work on BODIPYs is considerable, and due to this, structure-property
relationships in these materials are well-defined. Even so, they present particular obstacles
which make them imperfect choices for application as active layer materials in photovoltaics.
For example, there is only one site available for -substitution to red-shift their absorption and
emission profiles, so further shifts require increasingly complex substitution of conjugating
systems to the pyrroles;
1, 7
the cost of the pyrroles are high and the yield of BODIPY reactions
tend to be low (30% for simple BODIPYs); they are prone to sublimation at very low
temperatures, making control of the physical vapor deposition process difficult; and they have
exceptionally shallow oxidation and reduction potentials, making them poorly matched for
charge transfer to common fullerene-based acceptor materials.
1
We hoped to mitigate some of
these problems by replacing the pyrroles in dipyrromethenes with pyridines, giving the
dipyridylmethenes (DIPYRs). Pyridines are less reactive and often less expensive than
pyrroles, they are shelf-stable, and they offer an extra site for alkylation and benzannulation.
At first inspection, DIPYRs appeared to be more amenable to tuning of their HOMO and
LUMO energies, and thus of state energies by benzannulation using quinoline and isoquinoline
in lieu of pyridine, and there seemed to be no obvious reason why a pyridine-based BODIPY
analogue - which would also serve as a structural analogue of another ubiquitous chromophore,
anthracene - would not be at least a passable chromophore, with appreciable molar
absorptivities and photoluminescent quantum yields. Further evidence for the possibility of
strong fluorescent properties in pyridine-based analogues of BODIPY existed in the reports of
the blue-emissive meso-cyano substituted boron difluoride dipyridylmethene dye (  = 450 nm,
PL = 0.2)
8
and of nitrogen-bridged boron difluoride dipyridylamines which display UV-violet
emission (  em = 362–416 nm) and have even higher luminescent quantum yields (  PL = 0.4).
9

Molecular Design for Organic Photovoltaics
Chapter 2 | 35

Some reports existed describing the synthesis of 2,2-dipyridylmethane,
10
the
unconjugated precursor to dipyridylmethene, and a complete synthesis of the boron difluoro
dipyridylmethene conjugate was also available in the literature.
11
It was odd, however, that
other than this first report characterizing the structural properties and the absorption
characteristics of this material, there existed only a small handful of papers describing
pyridine-based BODIPY analogues and none which characterized the emissive properties of
the parent structure. The source of this lack of investigation appears to be a report which
erroneously described the parent boron difluoro DIPYR complex (hereafter referred to simply
as DIPYR) as non-emissive.
9
This mistake was repeated in the otherwise incredibly thorough
and insightful 2007 Burgess BODIPY review.
2
One benefit of being a young graduate student,
however, is naiveté; although I was familiar with BODIPYs and with this review in particular,
it wouldn’t be incorrect to say that I had skimmed the section on pyridine-based BODIPYs and
was therefore unburdened by the common wisdom, which supposed them to be entirely useless.
The syntheses of DIPYR as well as its quinoline- and isoquinoline-based derivatives -DIPYR
and -DIPYR were therefore carried out without prejudice. The characterization of the
structural, electronic, and photophysical properties of the DIPYR family of chomophores

Figure 2.2. DIPYR, a pyridine-based BODIPY analogue, with site-labeling. -DIPYR, and -DIPYR are
quinoline- and isoquinoline derivatives, respectively, of the parent complex DIPYR.

Golden
36
(Figure 2.2) comprise the rest of this chapter and indeed inform much of the rest of this
dissertation. Particular attention is focused on deriving the structure-property relationships
which define these materials in order to inform future derivatization.
*12-13

2.2 Dipyridylmethane and its Quinoline and Isoquinoline Analogues
The synthetic methods used to obtain the ligands 2,2’-dipyridylmethane, 2,2’-
diquinolylmethane, and 1,1’-diisoquinolylmethane were obtained from the literature and are
outlined in Scheme 2.2, Methods I-IV. Method I was utilized for the synthesis of 2,2’-
dipyridylmethane; this procedure involves asymmetric starting materials to produce a
symmetric product and is therefore inherently atom uneconomical.
10
However, the procedure
works well at large scale (>20g) as well as at test scales, and the product is isolated in high
purity by distillation. Therefore it was the method of choice for producing 2,2’-
dipyridylmethane. In it, 2-picoline is deprotonated with a strong base (both nBuLi and LDA
yield similar results), and the 2-picolinic anion is then treated with 2-fluoropyridine, resulting
in nucleophilic aromatic substitution (SNAr) to form the symmetric dipyridylmethane. This
method could not be used for the quinoline or isoquinoline derivatives, however, because the
fluorine-substituted heterocycles were not available and the chlorine derivatives were not
sufficiently electrophilic for SNAr. Three alternative synthetic strategies (Scheme 2.2, Methods
II-IV) to access the benzannulated derivatives were used.
14-16

Methods II and III were utilized to access the quinoline and isoquinoline derivatives,
respectively. Method II involves lithiation of the 2-bromoquiniline starting material, then
coupling onto a carbonate to give the 2,2’-diquinolylketone. This intermediate is then reduced

*
Much of the work in this chapter is published in the literature in Ref. 12; the ultrafast transient absorption spectra
were acquired with the collaboration of Laura Estergreen from the Bradforth lab at the University of Southern
California, and was itself published in Ref 13.
Molecular Design for Organic Photovoltaics
Chapter 2 | 37
to the methane-bridged product via a Wolff-Kishner reduction. The procedure works in fair
yields (68%, two steps) and provides a relatively pure product after two steps with purification
of the ketone intermediate. The starting material, however, is expensive, at nearly $100/gram,
and the procedure is also atom uneconomical, as only a single carbon atom from the
diethylcarbonate starting reagent is carried through to the final product. The 1-
bromoisoquinoline starting material required to use Method II for the synthesis of 1,1’-
diisoquinolylmethane ligand, however, is unreasonably expensive, at over $120 per gram and
the chloro- derivative, though much less expensive, was not sufficiently reactive. Method II
was also used for the attempted synthesis of 6,6’-diphenanthridinylmethane. The resulting
material is bright orange in color and mostly insoluble in all solvents. It was purified by
precipitation from acetone and identified by MALDI-TOF but never characterized by NMR.
There is literature evidence for this species existing predominantly as its conjugate base,
wherein the meso proton has been relinquished and the nitrogen lone pairs form a N-H···N
bridge.
17

Method III was chosen initially as the best alternative reaction to obtain 1,1’-
diisoquinolylmethane with reasonably priced starting materials. It is highly atom economical,
as it entails direct coupling of the 1-methyl and 1-chloroisoqunoline starting materials under
high temperature, high pressure conditions without the use of a solvent, yielding the ligand
product and HCl as the only major products. With both starting materials at about $35/gram,
it is not a terribly expensive procedure. It is, however, sensitive to impurities in the starting
materials and yields a dark brown gummy paste which can be difficult to extract and requires
multiple columns to yield the pure product.
Golden
38
Method IV was published after both the quinoline and isoquinoline derivatives had
been initially synthesized characterized by Methods II and III.
16
It requires no functional
handles to be installed in the heteroaromatic starting material; instead, quinoline or
isoquinoline are added to a pressure flask under nitrogen, with dry toluene, catalytic
tetramethylethylenediamine, and 3 equivalents of methyl Grignard reagent. The Grignard
reagent selectively installs the methyl group into the 2- and 1- positions of the quinoline or
isoquinoline starting materials, respectively, which is then deprotonated in situ and attacks
another quinoline or isoquinoline to yield the ligand in four hours at 140 ºC. It is a superior
reaction due to its scalability, the price of reagents, and its rapidity. It was utilized for the
scaling of both the quinoline and isoquinoline products.
Scheme 2.2. Methods I-IV are used to obtain various diheteroarylmethane ligands.

Molecular Design for Organic Photovoltaics
Chapter 2 | 39
2,2’-Dipyridylmethane exists in a single tautomeric form, while both the quinoline and
isoquinoline derivatives exist in either of two tautomers a or b (Figure 2.3); the proportion of
tautomer a to tautomer b depends on the pH of the solution, with more acidic solutions favoring
tautomer b. The absorption and emission spectra of each ligand were acquired in dilute and
concentrated solutions. In dichloromethane solution, the major tautomer a predominates, and
the resulting absorption profiles are dominated by UV transitions. In the quinoline and

0.0
0.5
1.0
200 250 300 350 400 450 500 550 600
0.0
0.5
1.0
200 250 300 350 400 450 500 550 600
0.0
0.5
1.0
2,2'-diquinolylmethane
1,1'-diisoquinolylmethane

2,2'-dipyridylmethane
Absorption / Emission (a.u.)

Wavelength (nm)
0.0
0.5
1.0
400 450 500 550 600 650
400 450 500 550 600 650
0.0
0.5
1.0
Absorption / Emission (a.u.)
2,2'-diquinolylmethane

1,1'-diisoquinolylmethane

Wavelength (nm)

Figure 2.3. 2,2’-diquinolylmethane and 1,1’-diisoquinolylmethane, the benzannulated derivatives of DIPYR,
exist in two tautomeric forms a (left) and b (right). UV-visible absorption spectra and fluorescence emission
spectra of 2,2’-dipyridylmethane, 2,2’-diquinolylmethane, and 1,1’-diisoquinolylmethane (a) dilute in
dichloromethane and (b) concentrated in dichloromethane. Concentrated spectra highlight the effect of
electron delocalization across the ligand.

Golden
40
isoquinoline derivatives, another much weaker absorption can be seen from 400-500 nm.
Concentrated solutions of each were prepared and the absorption and fluorescence emission
spectra of the b tautomers were measured. The absorption and emission features of the b
tautomers are significantly red-shifted compared to tatutomer a, and are similar in lineshape to
those of anthracene, indicating that the b tautomers exhibit significant -   electronic
transitions that are delocalized across the entire ligand.
2.3 Boron Difluoride Dipyridylmethenes
A dye is, simply, a visible light absorbing compound with a high molar absorptivity. In
order to have this particular property, dyes are characterized by conjugated systems which
exhibit electronic resonance. Small aromatic systems such as benzene rings have high energy
absorptions in the UV; as the -system is expanded, the energy of the absorption decreases,
pushing the absorption manifold into the visible region. The dipyridylmethanes are not, then,
dyes on their own. They have small conjugated systems localized on the heteroaromatic
systems, relegating absorption into the UV. By eye, they appear as a light yellow color, due to
their absorption manifolds tailing into the blue region of the spectrum. In order to make good
dyes of dipyridylmethanes, they must be fully conjugated such as is observed in Figure 2.3
tautomer b. This tautomer can be stabilized by chelation; in this work, boron difluoride is
chosen as the chelating group, as it is known to form remarkably stable compounds with
dipyrromethenes (BODIPY dyes). In addition to adding a chelating group, a base must be
applied to deprotonate the meso position and complete the conjugation of the system. The result
is a formally zwitterionic boron difluoride dipyridylmethene (DIPYR) complex with intense
absorption in the visible region – a DIPYR dye.
Molecular Design for Organic Photovoltaics
Chapter 2 | 41
In addition to high molar absorptivity in the visible region, a further design criterion
for organic photovoltaics is a high photoluminescent quantum yield and a low rate of non-
radiative decay. If a material has a high non-radiative decay rate after photoexcitation, it can
be reasonably assumed that the exciton thereby generated will not have sufficient time to
migrate to the donor/acceptor interface to generate a pair of charge carriers. To avoid this
problem, which is severely detrimental to device efficiency, materials are usually designed
with structural rigidity in mind. More rigid materials tend to have fewer mechanisms for non-
radiative decay and, consequently, lower non-radiative decay rates. Borylation of the
dipyridylmethenes, then, served two purposes: it stabilizes the system into the visible-light
absorbing tautomer, and it locks it into a planar geometry wherein there are few mechanisms
for energy loss outside of re-emission of a photon.
2.3.1 Borylation of Dipyridylmethene Dyes
The synthesis of boron difluoride dipyridylmethenes involves first borylating the
diheteroarylmethane ligand by refluxing in 1,2-dichloroethane in the presence of two
equivalents of boron trifluoride diethyl etherate (Scheme 2.3). Almost immediately after
treatment with BF3OEt2. a precipitate begins to form. After two hours of refluxing, the mixture
is cooled, whereupon the precipitate settles and the solution is slightly blue-emissive. The
yellow precipitate is isolable and is assigned as the fluoride salt of the borylated ligand. It is
very easily converted to the final product in the presence of weak base (sodium bicarbonate
solution is sufficient, for example), and the optimized reaction does not require isolation of the
intermediate salt. Instead, it is deprotonated in situ with an excess of Hünig’s base to produce
DIPYR in 38% yield. This method was also used to prepare the quinoline and isoquinoline
derivatives -DIPYR and -DIPYR in 39% and 36% yields, respectively.
Golden
42
This borylation procedure is similar to the synthesis reported for meso cyano DIPYR,
a blue-emissive fluorophore, while it differs from the syntheses of dipyridylamines, which do
not require the addition of base to obtain the deprotonated and fully conjugated product.
8
There
is an early paper characterizing nitrogen bridged boron difluoro dipyridylamines which
compares them with the parent carbon-bridged DIPYR.
9
This report, which outlines a synthesis
for DIPYR which does not include a deprotonation step, characterizes DIPYR as non-emissive;
it is possible that the intermediate fluoride salt was isolated by this preparation in lieu of
DIPYR itself, and that this is the source of confusion in the literature regarding the emissive
properties of DIPYR dyes.
This procedure is also quite different from the synthesis of BODIPY, which first
involves the in situ oxidation of dipyrromethane to form dipyrromethene, followed by
borylation which typically occurs at room temperature and within 15 minutes after treatment
with BF3OEt2 and a tertiary amine.
2
In the DIPYR synthesis, there is no need for an oxidation
step, as the ligand is already in the same oxidation state as the final product. Further
differentiating the two procedures, simultaneous treatment of the dipyridylmethane ligand with
BF3OEt2 and base is observed to preclude any reaction from occurring. It seems that in the case
of the dipyridylmethane borylation reaction, formation of the intermediate salt is a prerequisite
for formation of the final, fully conjugated product. In the case of dipyrromethene, the pyrrole
Scheme 2.3. The borylation of dipyridylmethanes is a two-step reaction with an isolable intermediate fluoride
salt.

Molecular Design for Organic Photovoltaics
Chapter 2 | 43
must first be deprotonated before it is able to bind to boron, as the nitrogen lone pair is tied up
in the aromaticity of the ring.
2.3.2 Structure of DIPYR Dyes
In addition to its similarity to BODIPY, DIPYR can be thought of as a diazaboro
derivative of anthrancene. Indeed, the molecular structure of DIPYR obtained by single-crystal
X-ray diffraction (Figure 2.4) shows peripheral carbon–carbon bonds lengths which mimic a
pattern of alternating short (C1–C2 and C3–C4) and long (C2–C3 and C4–C10) distances
similar to those found in anthracene.
18
This bond length alternation pattern is preserved in the
benzannulated derivatives -DIPYR and -DIPYR. The carbon–nitrogen bond distances (C1–
N1 and C10–N1) in DIPYR are elongated relative to the values found in pyridine (C–N =
1.337 Å), possibly as a result of the reduction in the aromaticity of the pyridine ring in DIPYR
compared to free pyridine.
Similarities in the bond lengths of the boron–nitrogen (1.544–1.566 Å) bonds and boron–
fluorine (1.390–1.404 Å) bonds in all three derivatives suggest that borylation is equally stable
in each of the three derivatives. In all three compounds, the central six-membered borocyclic
ring adopts a puckered conformation, wherein the BF2 moiety lies above the plane of the central
ring (Figure 2.7d). The distance to which the BF2 moiety is out-of-plane (dop) is larger in -
DIPYR (dop = 0.23 Å) than it is in DIPYR (dop = 0.09 Å) or -DIPYR (dop = 0.04 Å). In all
three cases, however, this deviation from planarity is small and it is not certain whether it is a
quality intrinsic to the structure of the molecule or is a result of crystal packing forces.
Comparing the structure of DIPYR with BODIPY, it can be seen that the variation in C–C
bond lengths in the five-membered ring of BODIPY is smaller (1.370–1.410 Å)
6
than it is in
Golden
44
the pyridine of DIPYR (1.360–1.430 Å), suggesting a greater degree of cyanine-like
interaction between the canonical resonance structures of BODIPY than of DIPYR. Bond
distances between B–N and B–F are comparable in both types of chromophores, and both
exhibit slight puckering of the boron atom out of the plane of the borocyclic ring (d op = 0.146
Å in BODIPY and 0.09 Å in DIPYR). One significant difference between BODIPY and the
DIPYR derivatives is the average distance between the fluorine atoms and the nearest hydrogen
atoms (Figure 2.1 BODIPY -positions and Figure 2.4 starred positions in DIPYRs). The
closest intramolecular H···F distances in the parent BODIPY (3.0 Å) fall outside the Van der
Waal radii of the two atoms (2.67 Å).
19
In contrast, the equivalent protons in DIPYRs are
positioned much closer to the fluorine atoms; the shortest intramolecular H···F distances found
in DIPYR and -DIPYR are 2.55 Å and 2.43 Å, respectively, whereas the benzannulated rings

Figure 2.4. Molecular structures of DIPYR (a), -DIPYR (b), and -DIPYR (c) are presented with bond
lengths labeled. Carbon (grey), nitrogen (blue), boron (pink), and fluorine (yellow) residues are depicted as
thermal ellipsoids at 50% probability; hydrogen atoms (light grey) are drawn as spheres with radius of 0.15
Å. The three DIPYR (green), -DIPYR (blue), and -DIPYR (red) dyes are overlaid in (d) to show the subtle
variation in central ring puckering of the BF 2 moiety.

Molecular Design for Organic Photovoltaics
Chapter 2 | 45
of -DIPYR place the atoms in even closer proximity (H···F = 2.34 to 2.49 Å). The effect of
the short H···F distances is manifested in the coupling of the
1
H NMR resonances at these
positions. The doublets in DIPYR (  = 7.90 ppm,
3
JHH = 5.5 Hz) and -DIPYR (  = 7.88 ppm,
3
JHH = 7.0 Hz) are broadened due to through-space
coupling to the
19
F atoms,
20
and the corresponding
protons in -DIPYR (  = 8.57 ppm) are clearly
resolved into a doublet of triplets (
3
JHH = 8.8 Hz,
4
J FH = 3.5 Hz). These close interatomic distances
provide an energetic barrier to flip-flopping of the
puckered BF2 group. Such structural rigidity is
promising for narrow line-widths and small Stokes
shifts in absorption and emission spectra and
reduced non-radiative decay rates.
The crystal packing of each of the three
derivatives is portrayed in Figure 2.5. The unit cell
of DIPYR is comprised of eight molecules in a
monoclinic Pbca space group, while the unit cells
of -DIPYR and -DIPYR are comprised of four
molecules each in monoclinic P21/c and Cc space
groups, respectively. DIPYR packs in a
herringbone fashion, preventing any -stacking interactions between adjacent chromophores,
while γ-DIPYR packs in a slipped, pseudo-herringbone unit cell. -DIPYR differs
significantly in packing from the other two compounds, exhibiting significant π-stacking
a)

b)

c)

Figure 2.5. (a) DIPYR, (b) α-DIPYR, and (c)
γ-DIPYR unit cells, depicting differing
packing. γ-DIPYR packs non-
centrosymmetrically.

Golden
46
interactions wherein the quinoline moiety
of one molecule lies antiparallel to the
quinoline moiety of the adjacent molecule
at a distance of 3.59 Å. One distinguishing
feature of the γ-DIPYR packing is that its
unit cell is non-centrosymmetric, differing
from the centrosymmetric unit cells of the
previous two compounds. This feature may
lend them to use in nonlinear optical
applications.
2.3.3 Electrochemical Properties of
DIPYR dyes
The electrochemical properties of the
dyes were examined using cyclic
voltammetry (Figure 2.6). The redox
potentials, referenced to the oxidation of
ferrocene (Fc
0/+
), are enumerated in Table
2.1. As can be seen in the cyclic
voltammograms, the oxidation waves in all
three materials are irreversible. The
reduction wave in DIPYR is quasi-
reversible, while the corresponding
features in both -DIPYR and -DIPYR are
-3 -2 -1 0 1
-20
-10
0
10
20
Current (uA)
Potential (V)
DIPYR

-2 -1 0 1
-20
-10
0
10
20
Current (uA)
Potential (V)
-DIPYR

-2 -1 0 1
-80
-40
0
40
80
120
Current (uA)
Potential (V)
-DIPYR

Figure 2.6. Cyclic voltammograms of DIPYR with
internal decamethyl ferrocene reference (top), and
-DIPYR (middle) and -DIPYR (bottom) with
internal ferrocene references. Spectra were acquired in
acetonitrile with 0.1 M tetrabutylammonium
hexafluorophosphate as the electrolyte under anaerobic
conditions and referenced to the Fc
0/+
redox couple.

Molecular Design for Organic Photovoltaics
Chapter 2 | 47
reversible. The irreversible oxidation in DIPYR dyes leads to the growth of a new, broad
feature around -0.5 V vs. Fc
0/+
, and may be the result of the coupling of two oxidized molecules,
a phenomenon which has been observed during the electrochemical oxidation of both
BODIPYs and cyanines.
21-24
Oxidation of DIPYR occurs at a much lower potential
(Eox = 0.14 V) than in BODIPY (Eox = 1.35 V),
25
whereas the reduction potential is
correspondingly more negative (DIPYR, Ered = -2.32 V; BODIPY, Ered = -1.05 V) such that
the redox gaps between DIPYR (2.48 V) and BODIPY (2.40 V) are remarkably similar.
24, 26

The redox potentials of both -DIPYR and -DIPYR are anodically shifted relative to the
parent compound, indicating stabilization of both valence MOs upon benzannulation, with the
LUMO stabilized to a slightly greater degree (~0.1 V) than the HOMO.
27-28

2.3.3 Photophysical Properties of DIPYR Dyes
The photophysical properties of the three DIPYR derivatives were characterized using UV-
visible absorption, photoluminescence emission, and time-correlated single photon counting
spectroscopies. The absorption and emission spectra of each of the three dyes in
methylcyclohexane are plotted in Figure 2.7 and the photophysical data are summarized in
Table 2.2. Each of the DIPYR dyes display intense (  > 10
4
M
-1
cm
-1
), vibronically structured
absorption bands in the UV-visible spectrum. The vibronic manifolds are characterized by
major progressions of 1475 cm
-1
, similar to the spacings found in anthracene.
29
The lineshape
of the DIPYR absorbance band is similar to the profile reported for meso-cyano DIPYR,
8

Table 2.1. Redox potentials (V vs. Fc
0/+
)
a
and calculated HOMO and LUMO energies (eV)
b

E ox

E red

E redox      LUMO
DIPYR 0.14 -2.32 2.48 -4.80 -2.01
-DIPYR
0.40 -1.95
2.35 -5.16 -2.46
-DIPYR 0.27 -2.07
2.34 -4.98 -2.31
a
In acetonitrile with 0.1 M TBAF.
b
HOMO and LUMO energies derived from redox potentials.

Golden
48
though it differs somewhat from the
benzannulated derivatives. First, the
absorption energies of the benzannulated
derivatives are slightly bathochromically
shifted compared to DIPYR, in accordance
with the smaller redox gaps observed
electrochemically. More notably, there is an
intense high-energy absorption in DIPYR
the benzannulated derivatives appear to have
a different progression of line intensities
than that observed in DIPYR, with DIPYR
showing broader absorptive features than
either -DIPYR or -DIPYR. This is
particularly evident when contrasting the
linewidth for the 0-0 transition of DIPYR, which has a full width half-maximum (fwhm) of
840 cm
-1
, with those of -DIPYR (fwhm = 250 cm
-1
) and -DIPYR (fwhm = 700 cm
-1
). The
0-0 linewidth of -DIPYR is exceptionally narrow, even more so than that of anthracene
(fwhm = 440 cm
-1
). The narrow linewidth for the 0-0 transition in -DIPYR, along with the
large ratio between the 0-0 and 0-1 transitions, indicate that there is very little structural
distortion in the excited state. This is consistent with the exceptionally close H···F interatomic
distances observed in the X-ray structure and confirmed by NMR.
Also differing from the quinoline and isoquinoline derivatives, the absorption lineshape of
DIPYR is solvent dependent; the 0-1 peak of the vibronic progression gradually increases in
0
10k
20k
30k
0
40k
80k
300 400 500 600 700 800
0
50k
100k

DIPYR
DIPYR

DIPYR
Molar Absorption (M
-1
cm
-1
)

Wavelength (nm)
0
1
0
1
Emission (a.u.)
0
1

Figure 2.7. Absorption (solid) and normalized
fluorescence (dashed) and phosphorescence (blue)
spectra recorded in methylcyclohexane; left axis is
molar absorbance; right axis is normalized emission
intensity.
Molecular Design for Organic Photovoltaics
Chapter 2 | 49
intensity with increasing solvent polarity, becoming more intense than the 0-0 transition in
acetonitrile solution (Figure 2.8). The change in the lineshape with solvent polarity for
DIPYR suggests the presence of a second, weaker transition lying at a slightly higher energy
than the S0-S1 transition that is polarized along the short, permanent dipole-containing axis of
DIPYR. This supposition is supported by TD-DFT calculations of the excited states of DIPYR,
which were performed to compare with experimental data and are presented in Section 2.3.6.
All of the DIPYR derivatives are highly luminescent in solution and are characterized by
almost negligible Stokes shifts (< 5 nm) in non-polar solvents. A noticeable asymmetry is
present in the mirror image relationship
between absorption and emission in
DIPYR, consistent with the absorption
profile being distorted by an S2 state lying
at slightly higher energy than the S1 state
and therefore overlapping the S1
absorption feature. In contrast, the
absorption and emission spectra of - and
-DIPYR display a near perfect mirror
symmetry. DIPYR has a
400 500 600 700
0.00
0.25
0.50
0.75
1.00

Intensity (a.u.)
Wavelength (nm)
Acetonitrile
DCM
THF
Toluene
MeCyHex

Figure 2.8. Normalized absorption (solid) and emission
(dashed) spectra of DIPYR in solvents of varying
polarity. Note that the relative intensity of the 0-1
vibronic feature increases with increasing polarity,
suggesting the presence of a second, underlying
absorptive feature which is solvent-dependent.

Table 2.2. Photophysical properties of DIPYR dyes

absorption
 max / nm
 / M
-1
cm
-1
emission
 max / nm
 PL  / ns k r / 10
8
s
-1
k nr / 10
8
s
-1

DIPYR 481 (476) 2.9×10
4
484 (496)
482
a
, 577
b

0.17 (0.085) 1.9 (1.6)
2.0
a
0.91(0.53) 4.5 (5.7)
-DIPYR 520 (515) 1.1×10
5
520 (521)
520
a
, 638
b

0.77 (0.77) 5.7 (5.0)
4.7
a
1.4 (1.5) 0.41 (0.46)
-DIPYR 500 (493) 8.7×10
4
504 (504)
504
a
, 626
b

0.80 (0.75) 3.9 (4.1)
3.6
a
2.0 (1.8) 0.51 (0.61)
Measurements acquired at room temperature in methylcyclohexane (acetonitrile).
a
Fluorescence and
b
phosphorescence at 77 K in methylcyclohexane.
Golden
50
photoluminescent quantum yield (  PL) of 17% in methylcyclohexane and 8.5% in acetonitrile,
whereas the values for - and -DIPYR are much higher (  PL = 77% and 80%, respectively)
and largely solvent independent. The radiative rate constants (kr) are similar in value among
all three derivatives (kr = 0.91–2.0 x 10
8
s
-1
), whereas the rate for non-radiative decay (knr) of
DIPYR in methylcyclohexane (knr = 4.5 x 10
8
s
-1
) is nearly an order of magnitude larger than
in - or -DIPYR (knr ≈ 4.5 x 10
7
s
-1
). The notably faster rate for non-radiative decay in DIPYR
is not the result of structural distortions in the excited state, as no major change in the
luminescent lifetimes ( ) occurs when comparing spectra acquired in fluid solution at room
temperature (  = 1.9 ns) or rigid media at 77 K (  = 2.0 ns). A more plausible explanation for
the large knr value of DIPYR is a fast rate for intersystem crossing (ISC) between singlet and
triplet states. Support for this mechanistic hypothesis is based on the fact that phosphorescent
emission (E0-0 = 577 nm) is readily observed in the parent DIPYR in frozen methylcyclohexane
at 77 K, whereas none is observed under similar conditions for - and -DIPYR. It is possible,
however, to record phosphorescence from the benzannulated compounds in frozen solutions
using more rigorous conditions of gated detection, along with the addition of iodomethane to
promote ISC through an external heavy atom effect (Figure 2.7).
2.3.5 Computational Study of DIPYR Dyes
The electronic characteristics of DIPYR dyes in the gas phase were examined using both
time‑dependent density functional (TD-DFT) and extended multi-configurational quasi-
degenerate second order perturbation (XMCQDPT2)
30
theoretical calculations. The
XMCQDPT2 calculations were employed after a series of computational studies reported that
optical transitions in meso-cyano DIPYR, BODIPY, and related cyanine-type dyes are
multireference in nature and are accompanied by a significant double excitation character in
Molecular Design for Organic Photovoltaics
Chapter 2 | 51
their excited states.
31-32
Singlet excitation energies calculated for these pseudo-aromatic dyes
using single reference methods such as TD-DFT are found to have errors greater than 0.3 eV.
However, despite the large absolute errors, these reports found that calculated TD-DFT values
for this class of molecules can be scaled linearly with experimental excitation energies.
32

Multiconfigurational approaches like CASSCF (complete active space self-consistent field)
are necessary to accurately describe the excited states of such systems without invoking a
correction factor. CASSCF calculations alone, however, do not account for dynamic
correlation effects, leading again to significant errors in calculated state energies, though
dynamic correlation can be incorporated into the CASSCF calculations using perturbative
treatments. For these systems, the extended multiconfigurational quasi-degenerate second
order perturbation theory
30
(XMCQDPT2) method was employed to incorporate dynamic
correlation into the CASSCF calculations. For DIPYR, an active space of 12 electrons in 12
orbitals (12,12) was selected, while for both - and -DIPYR, a larger active space containing
14 electrons in 13 orbitals (14,13) was used. These active space combinations were found to
strike a good balance between computational effort and accuracy. CASSCF calculations were
performed with state-averaging over six roots. The 6-31G(d) basis set was used for all state
averaged (SA)-CASSCF/XMCQDPT2 calculations. The XMCQDPT2 calculations were
performed on B3LYP/6-31G** optimized geometries.
Table 2.3. Experimental and calculated state energies in DIPYR dyes.
S 3 (nm, eV, ƒ) S 2 (nm, eV, ƒ) S 1 (nm, eV, ƒ) T 2 (nm, eV) T 1 (nm, eV)
DIPYR (exp)
TD-DFT
XMCQDPT2

312, 3.98, 0.449

389, 3.19, 0.029

482, 2.57
396, 3.13, 0.232
466, 2.66
--
511, 2.43
462, 2.69
576, 2.15
587, 2.11
547, 2.27
-DIPYR (exp)
TD-DFT
XMCQDPT2

319, 3.88, 0.067

401, 3.09, 0.030

524, 2.37
448, 2.77, 0.464
537, 2.31
--
480, 2.58
415, 2.99
638, 1.94
668, 1.86
589, 2.11
-DIPYR (exp)
TD-DFT
XMCQDPT2

346, 3.59, 0.22

361, 3.43, 0.030

504, 2.46
435, 2.85, 0.590
522, 2.37
--
450, 2.76
406, 3.05
626, 1.98
658, 1.89
659, 1.88

Golden
52
Table 2.3 compares experimental excited state energies with those calculated by TD-DFT
and by XMCQDPT2; it can be seen that TD-DFT indeed overestimates the energy of singlet
excited state transitions in all three DIPYR derivatives due to significant double-excitation
character (>4%) as indicated by the CASSCF results tabulated in Table 2.4. In the case of
DIPYR, -DIPYR, and -DIPYR, the linear correction factor was found to be -0.44 eV by
comparing the mean difference between calculated singlet (S1, S2, and S3) energies against
those observed experimentally; i.e. TD-DFT overestimates singlet energies in DIPYRs by a
wider margin than it does for BODIPYs and cyanines. The S1 and T1 excitation energies
Table 2.4. Dominant configurations (>1%)* involved in the singlet and triplet states of the DIPYR, -DIPYR, and -
DIPYR along with their corresponding percent contributions calculated using the state averaged CASSCF method.
DIPYR (12,12)
S 0 % S 1 % T 1 % T 2 %
222222 000000 77.16 222221 100000 71.24 222221 100000 75.94 222221 010000 64.46
222221 010000 2.74 222220 110000 1.84 222212 010000 1.26 222212 001000 8.20
222212 100000 0.74 222211 101000 1.36 221222 100000 1.16 222212 100000 2.24
222202 002000 0.43 212221 100100 0.94 222220 110000 0.82 221222 010000 1.50
222202 020000 0.36 222122 010000 0.86 221222 001000 0.66 222220 101000 1.14
212222 000100 1.04
-DIPYR (14,13)
S 0 % S 1 % T 1 % T 2 %
2222222 000000 80.44 2222221 100000 71.04 2222221 100000 60.00 2222221 010000 23.78
2221222 010000 2.08 2222122 100000 5.36 2222212 100000 23.46
2222220 110000 1.12 2222212 010000 4.76 2222122 010000 15.00
2222211 200000 0.90 2221222 010000 4.38 2221222 100000 5.12
2222121 110000 0.86 2222221 000010 2.38 2222221 000100 2.74
2122222 010000 1.04 2221222 000010 1.94
2122222 100000 1.84
1222222 000100 1.26
2122222 001000 1.06
2122222 000010 1.04
2222212 001000 1.02
-DIPYR (14,13)
S 0 % S 1 % T 1 % T 2 %
2222222 000000 81.73 2222221 100000 70.10 2222221 100000 74.96 2222212 100000 47.76
2222202 200000 0.32 2222121 200000 2.18 2222122 100000 2.02 2222122 000100 15.52
2222022 000200 0.32 2222211 200000 1.44 2222221 001000 1.78 2122222 000010 3.92
1222221 100001 1.10 2222221 000010 1.40 2221222 100000 3.60
2222121 100100 1.08 1222222 100000 0.50 2222221 000100 2.62
2122222 100000 1.78
2222212 000010 1.58
1222222 000100 1.28
*In cases where the number of configurations with more than 1% contribution is less than five, configurations with
smaller contributions are also listed. Configurations with double excitations are indicated in bold fonts.
Molecular Design for Organic Photovoltaics
Chapter 2 | 53
calculated at the SA-
CASSCF/XMCQDPT2/6-31G(d) level are
found to be in good agreement with the
experimental values for all three dyes.
However, XMCQDPT2 calculations showed
the triplet transitions to have no significant
(<1%) double-excitation character, and as
such they are well estimated (within
150 meV) by TD-DFT.
DFT calculations of the ground state
structure of DIPYR replicates the pattern of
C–C and C–N bond length alternation found
in the X-ray structures of the molecules,
including an out-of-plane distortion (dop =
0.42 Å) of the BF2 group in the six-membered borocyclic ring in DIPYR.
Both DIPYR and BODIPY can be idealized as having C2v symmetry. When compared this
way, the nodal characteristics for the frontier orbitals of DIPYR can be seen to share distinct
similarities and differences with those of BODIPY; the three pertinent orbitals for both
structures are compared in Figure 2.9. In DIPYR, the HOMO-1 and LUMO of the idealized
system are a2 symmetric, whereas the HOMO is b1 symmetric. The same symmetric
configuration exists for BODIPY except that the orbital relationship is shifted such that the
HOMO and LUMO+1 are described by a2 symmetry, whereas the LUMO is b1 symmetrized.
This difference results in the HOMO of DIPYR being characterized by significant orbital
BODIPY DIPYR

LUMO+1 a 2
0.73 eV

LUMO a2
-1.42 eV

LUMO b1
-2.86 eV

HOMO b1
-4.87 eV

HOMO a2
-5.98 eV

HOMO-1 a 2
-7.37 eV

C7H9 cation

HOMO b1

LUMO a2
Figure 2.9. Orbital contributions to the frontier
molecular orbitals in BODIPY (left) compared with
DIPYR (right).
Golden
54
density at the meso position, whereas in BODIPY, it is the LUMO which is largely localized
at the meso position. This relative shift of orbital densities suggests that where the substitution
of the meso -H atom in DIPYR with an electron donating group would bathochromically shift
the absorption and emission (by destabilizing the HOMO without affecting the LUMO)
energies, the same substitution in BODIPY would induce a hypsochromic shift due to
destabilization of the LUMO without affecting the HOMO. By the same logic, substitution at
the meso position with an electron withdrawing group would induce a hypsochromic shift in
DIPYR and a bathochromic shift in BODIPY. In addition, it is worth noting that the nodal
characteristics of the top half of DIPYR are identical to those of the C7H9 cation. This orbital
pattern, where nodes bisect every other carbon atom in the HOMO and LUMO, is a feature
common to non-alternant hydrocarbons and leads to intense, narrow absorption bands in
chromophores such as cyanine dyes.
33

The relative energies calculated for the frontier molecular orbitals are in good agreement
with the electrochemical oxidation and reduction potentials reported in Section 2.3.4. The
calculated energies for the HOMO and LUMO in DIPYR are destabilized by approximately
1.1 eV and 1.3 eV, respectively, compared to the energies for the corresponding orbitals in
BODIPY. Likewise, HOMO and LUMO energies calculated for - and -DIPYR reflect the
relative stabilization of these orbitals determined from their corresponding redox potentials
(for -DIPYR: HOMO = -5.00 eV, LUMO = -1.96 eV; for -DIPYR: HOMO = -5.01 eV,
LUMO = -1.91 eV). Although extension of the π-system in chromophores by benzannulation
is typically understood to induce a bathochromic shift of the absorption and emission bands,
the location of benzannulation plays a key role in modulating this effect.
34
In the case of both
-DIPYR and -DIPYR, benzannulation induces a stabilization of both the HOMO and to a
Molecular Design for Organic Photovoltaics
Chapter 2 | 55
slightly lesser extent of the LUMO,
corresponding to an overall reduction of the
HOMO-LUMO energy gap which is fairly small
(0.31 eV and 0.35 eV, respectively). This
difference corresponds well to the respective
bathochromic shifts of 0.20 eV (34 nm) and 0.10
eV (19 nm) observed when comparing the E0-0
energies for the S1 state of the three derivatives.
The orbital contributions for the lowest
singlet and triplet excitations, as determined by
TD-DFT, are summarized in Table 2.5 and the
involved orbitals are portrayed in Figure 2.10.
Both S0-S1 and S0-T1 transitions are primarily
between the HOMO and LUMO, with transition
dipole moments polarized along the long
molecular axis, perpendicular to the permanent
dipole oriented along the short (C2) axis. This
same orthogonal arrangement of (long)
transition and (short) permanent dipole
moments leads to non-solvatochromic absorption and emission features in BODIPY.
35
The S0-
S2 and S0-T2 transitions in DIPYR, however, are predominantly between the HOMO and
LUMO+1, and are polarized along the short axis of the chromophores such that they lie parallel
to the molecular dipole. As such, the S0-S2 and S0-T2 transitions are stabilized in polar
Table 2.5. Orbital contributions, energies and
oscillator strengths calculated for S 0 → S n, T n
transitions.
DIPYR
state transition %  (nm, eV) ƒ
S 1 56 → 57
56 → 59
91
9
396, 3.13 0.232
S 2 56 → 58 100 389, 3.19 0.029
S 3 56 → 57
56 → 59
9
90
311, 3.98 0.449
T 1 56 → 57 100 587, 2.11 0.000
T 2 56 → 58 97 511, 2.43 0.000
-DIPYR
state transition %  (nm, eV) ƒ
S 1 82 → 83 99 448, 2.77 0.463
S 2 82 → 84 98 384, 3.23 0.030
S 3 82 → 85
80 → 83
61
39
319, 3.88 0.067
S 4 81 → 83 96 312, 3.97 0.015
S 5 82 → 85
80 → 83
37
61
304, 4.08 0.349
T 1 82 → 83 100 668, 1.86 0.000
T 2 82 → 84 90 480, 2.58 0.000
-DIPYR
state
transition %
 (nm,
eV)
ƒ
S 1 82 → 83 99 435, 2.85 0.590
S 2 82 → 84 98 362, 3.43 0.030
S 3 82 → 85 100 345, 3.59 0.022
S 4 82 → 86 100 336, 3.69 0.0046
S 5 81 → 83 96 309, 4.01 0.033
T 1 82 → 83 99 658, 1.89 0.000
T 2

82 → 84
81 → 83
81 → 85
69
12
8
450, 2.76

0.000

Golden
56
environments, while the S0-S1 and S0-T2 transitions are not affected by solvent polarity.
2.3.6 A Unified Explanation of the Photophysical Properties of DIPYR Dyes
Compiling all of the observed data with the computational studies allows for precise
assignment of the spectral characteristics for the DIPYR chromophore. DIPYR displays three
DIPYR
57 (-1.42 eV) a 2 58 (-1.05 eV) b 1 59 (-0.43 eV) a 2

56 (-4.87 eV) b 1 55 (-7.37 eV) a 2

-DIPYR
83 (-1.96 eV) a2 84 (-1.20 eV) b1 85 (-0.47 eV) a2

82 (-5.00 eV) b1 81 (-6.42 eV) a2 80 (-6.50 eV) b1

-DIPYR
83 (-1.91 eV) a2 84 (-0.97 eV) b1 85 (-0.80 eV) a2 86 (-0.79 eV) b1

82 (-5.01 eV) b1 81 (-6.37 eV) a2

Figure 2.10. Orbitals of (top) DIPYR, (middle) -DIPYR, and (bottom) -DIPYR which contribute to the lowest
singlet and triplet transitions.

Molecular Design for Organic Photovoltaics
Chapter 2 | 57
strong absorption bands in the UV-visible spectrum, corresponding to transitions from the S0
ground state to the S1, S2 and S3 states (Table 2.6). The S1 and S2 states in DIPYR are close
enough in energy to have overlapping transition manifolds, with the oscillator strength (f) of
the S0-S1 transition being roughly an order of magnitude larger than the S0-S2 transition. Thus,
the vibronic progression of the S0-S1 absorption band is distorted by the underlying S0-S2
transition. The S0-S2 transition dipole moment is polarized along the short, permanent-dipole
containing axis of the dye. Therefore, the energy of the S2 state is affected by the polarity of
the surrounding medium. In contrast, the S0-S1 transition dipole moment lies along the long
axis of the molecule, which contains no net dipole and is largely unaffected by solvent polarity.
The close juxtaposition in energy between the S1 and S2 states induces a polarity-dependent
relationship in the lineshape of the lowest energy absorption bands in solution, as the S0-S2
transition will bathochromically shift in energy with increasing polarity whereas the S0-S1
transition will remain relatively unchanged. Consequently, the S0-S2 transition is mixed into
the lowest excited state to a greater extent in polar media, evidenced by the relative increase in
the 0-1 vibronic feature in the absorption spectrum of DIPYR in polar solvents. Further support
for the state-mixing hypothesis is in the observation of a lineshape change in DIPYR emission
with increasing solvent polarity; in addition to the expected increase in inhomogenous
broadening, there is an increase in the 1-0 vibronic feature of DIPYR fluorescence which is
Table 2.6. Calculated (TD-DFT: B3LYP/6-31G**) and experimental state energies in DIPYR dyes in nm (eV).

S 2 calc* S 1 calc* S 1 exp T 2 calc T 1 calc T 1 exp
DIPYR 451 (2.75)
f = 0.029
461 (2.69)
f = 0.232
482 (2.57) 511 (2.43) 587 (2.11) 577 (2.15)
-DIPYR 444 (2.79)
f = 0.030
533 (2.33)
f = 0.463
524 (2.37) 480 (2.58) 668 (1.86) 638 (1.94)
-DIPYR 415 (2.99)
f = 0.030
515 (2.41)
f = 0.590
504 (2.46) 450 (2.76) 658 (1.89) 626 (1.98)
*Calculated singlet energies were corrected by subtracting 0.44 eV from the computational output.
Golden
58
difficult to explain by any other means. This
competition between transitions is reflected in
the radiative rate constants of DIPYR in
methylcyclohexane (kr = 9.1 x 10
7
s
-1
) versus
acetonitrile (kr = 5.3 x 10
7
s
-1
), where the
stabilizing effect of polar solvent increases the
contribution from the weaker, dipole-
containing S2 state. In the case of both - and
-DIPYR, the S0-S2 transition does not
overlap with the S0-S1 transition, and thus in these systems the lineshape of the absorption and
emission bands and the magnitude of the radiative rate constants are unchanged upon an
increase in solvent polarity.
Another important change brought about by benzannulation of DIPYR is a relative
reordering of the S1 and T2 state energies (Figure 2.11). The energy of the T2 state in DIPYR
is calculated to be lower than that of the experimental S1 state, whereas the energy of the T2
states calculated for - and -DIPYR are greater than that of the experimental S1 states. This
difference in ordering between the S1 and T2 state energies is due to a stabilization of the
LUMO upon benzannulation (ca. -0.50 eV) while the LUMO+1 remains mostly unchanged (-
0.15 eV in α-DIPYR and +0.08 eV in -DIPYR). The relative ordering of state energies in
DIPYR is what leads to the low  PL that is further suppressed by an increase in solvent polarity,
since ISC from S1 to T2 is symmetry allowed and favored by the close proximity in energy
between the two states.
36
The net result in DIPYR is a fast rate for ISC (kISC > 10
9
s
-1
) that
outcompetes radiative decay from the S1 state. In anthracene, this same energetic situation

Figure 2.11. Intersystem crossing from S 1 to T 2
occurs rapidly in DIPYR, as the process is both
exergonic and symmetry-allowed. In - and
-DIPYR, however, the only exergonic ISC
mechanism is symmetry forbidden, and thus occurs
too slowly to compete with fluorescence.

Molecular Design for Organic Photovoltaics
Chapter 2 | 59
(T2 < S1) is considered the source of a high rate of intersystem crossing and relatively low
luminescent efficiency (kISC = 10
8
s
-1
, PL = 0.24).
37-38
In contrast, the higher energy of the T2
states compared to the S1 states in - and -DIPYR make ISC to T2 an endothermic,
unfavorable process. Therefore, the only pathway for ISC is directly from S1 to T1, which is
symmetry forbidden, and thus a slow process (kISC < 10
9
s
-1
) that results in high  PL
independent of solvent polarity. Support for this hypothesis is provided by phosphorescence
emission experiments. Phosphorescence from DIPYR is readily observed in frozen
350 400 450 500 550 600 650
-10
-8
-6
-4
-2
0
2
4
Abs. (mOD)
Wavelength (nm)
1 ps
10 ps
100 ps
400 ps
600 ps
800 ps
900 ps
DIPYR in Methylcyclohexane

350 400 450 500 550 600 650
-5
-4
-3
-2
-1
0
1
2
Abs. (mOD)
Wavelength (nm)
1 ps
10 ps
100 ps
400 ps
600 ps
800 ps
900 ps
DIPYR in Acetonitrile

400 450 500 550 600 650
-6
-4
-2
0
Abs. (mOD)
Wavelength (nm)
0.5 ps
1 ps
10 ps
100 ps
500 ps
900 ps
-DIPYR in Methylcyclohexane
400 450 500 550 600 650
-5
-4
-3
-2
-1
0
1
2
Abs. (mOD)
Wavelength (nm)
0.5 ps
1 ps
10 ps
100 ps
500 ps
900 ps
-DIPYR in Acetonitrile

Figure 2.12. Femtosecond transient absorption spectra of DIPYR and -DIPYR. The transient of DIPYR
shows a new absorptive feature to the red of the ground state bleach which grows throughout the duration of
the experiment. This feature is assigned to growth of the triplet state due to symmetry-allowed intersystem
crossing from the S 1 to the T 2 state. This transition is endothermic in -DIPYR and is consequently not
observed.

Golden
60
methylcyclohexane, whereas more rigorous methods of gated detection and treatment with a
heavy atom are needed to observe phosphorescence emission in the benzannulated derivatives,
consistent with slow ISC and thus a low yield of T1 for the latter compounds.
This state juxtaposition theory was probed experimentally by femtosecond transient
absorption (TA) spectroscopy (Figure 2.12).
13
Samples of DIPYR and -DIPYR were
prepared in both methylcyclohexane and acetonitrile. In both solvents, the TA spectra of -
DIPYR show simple two-state S1 to S0 kinetics, with all features in the observed spectrum
relaxing at the same rate. In DIPYR, however the TA spectra are characterized by an increase
in population around the ground-state bleach (GSB), with this feature being more pronounced
in acetonitrile. This growth is attributed to the population of the triplet state, which is occurs
in 2 ps in acetonitrile and 30 ps in methylcyclohexane. This lends further credence to the
hypothesis that the T2 state, which has a transition dipole moment parallel to the permanent
dipole of the chromophore, is stabilized in polar media. It is not clear from the fsTA alone
whether the observed triplet corresponds to the T1 or T2 state; although the T2 state is populated
first, it may be that internal conversion from the T2 to the T1 state occurs faster than can be
resolved by this experiment. Both triplet sensitization studies of the TA signature of the DIPYR
T1 state and further time resolution into the nanosecond to millisecond regime can provide
evidence to answer this question; while triplet sensitization will provide a definitive picture of
the T1 absorptive feature, following complete decay of the triplet at longer times will show
whether a new, lower energy triplet is formed before DIPYR relaxes to the ground state.
2.4 Carbon-Chelated Dipyridylmethene Salts
It was shown in Section 2.2 that the b tautomers of the ligands 2,2’-diquinolylmethane
and 1,1’-diisoquinolylmethane have absorption and emission spectra which are similar to those
Molecular Design for Organic Photovoltaics
Chapter 2 | 61
of the borylated derivatives -DIPYR and -DIPYR, indicating that chelation with BF2 plays
a role in stabilization of the fully conjugated ligand and structural rigidification leading to high
luminescence efficiency, but is mostly innocent with regard to the energy of the radiative
transition. In dipyrromethenes as well, it has been shown that chelation with non-BF2 groups,
such as zinc (II) ions, leads to the formation of brightly emissive chromophores with similar
absorption and emission features compared to their BODIPY counterparts.
Chelation of the diheteroarylmethane bases 2,2’-dipyridylmethane, 2,2’-
diquinolylmethane, and 1,1’-diisoquinolylmethane with -CH2-, followed by deprotonation to
form the coordinated dye, forms a carbocyanine dye salt (Scheme 2.4).
†
In contrast to the
zwitterionic boron difluoride DIPYR series, these species are cationic and require a balancing
anion which can be interchanged. Methylene-chelated 2,2’-dipyridylmethene is known in the
literature as pyridine red, and quinoline and phenanthridine derivatives thereof are known as
quinoline red and phenanthridine red; these names refer to the vivid red color of the salt in its
solid form.
17
The emission characteristics of pyridine red (heretofore referred to as Me-

†
The synthesis and initial characterization of Me-carDIPYR and Et-carDIPYR was performed by Stewart Sawyer
as part of his Bachelor’s thesis during a summer research internship in the Thompson lab. He has named Me-
carDIPYR “Trojan Red”, a very fitting description of its vivid red hue in the solid state.
Scheme 2.4. Synthesis of carDIPYR dyes.

Golden
62
carDIPYR) are described qualitatively in the primary literature as vivid green – tellingly, this
is the same color as the emission of DIPYR – but the radiative rate and quantum yield of this
material appear to be unpublished.
39-40

2,2’-Dipyridylmethane was conjugated with dibromomethane and with 1,2-
dibromoethane according to literature preparation,
39
forming the bromide salts of methyl
carDIPYR (Me-carDIPYR) and ethyl carDIPYR (Et-carDIPYR) which were purified by
converting to the hexafluorophosphate salts and precipitated from water in 16% and 7% yields,
respectively. The yields of the reactions are higher as revealed by LCMS than is evident after
purification because of the tendency of these reactions to form a sticky red paste of charged
materials which were purified instead by sequential recrystallizations. 2,2’-diquinolylmethane
was conjugated with dibromomethane using a method adapted from the synthesis of
Me-carDIPYR. Purification by recrystallization proved insufficient to isolate the pure product
-carDIPYR, so the deep magenta solid was purified by flash chromatography on alumina
with a gradient eluent from 0-5% MeOH in CH2Cl2. This purification method was quick and
simple, and future preparations of carDIPYRs should check for separability on alumina.
The absorption and emission spectra of each of the three carDIPYRs are provided in
Figure 2.13 and the photophysical parameters are compared with their borylated counterparts
Table 2.7. Photophysical parameters of carDIPYRs compared to boron difluoro DIPYRs in acetonitrile.
 max abs.
(nm)
 max em.
(nm)
 PL
(%)
  
(ns)
k r
(x10
8
s)
k nr
(x10
8
s)
DIPYR 476 496 9.5 1.95 0.49 4.6
Me-carDIPYR 493 550 10 2.17 0.46 4.1
Et-carDIPYR 483 504 3.2 0.32 (95%) /
3.91 (5%)
0.64 19.4
-DIPYR 515 521 77 5.0 1.5 0.46
-carDIPYR 522* 568 72 7.4 0.97 0.39
*  max of a-carDIPYR reported as most intense vibronic peak, which is not the lowest energy absorption band.

Molecular Design for Organic Photovoltaics
Chapter 2 | 63
in Table 2.7. The absorption maxima of all of
the carDIPYRs are slightly bathochromically
shifted (~10 nm) relative to the boron difluoro
chelated DIPYRs. In acetonitrile solution at
room temperature, the absorption of
Me-carDIPYR is broadened, appearing almost
gaussian, and lacks the fine vibronic structure
that is seen in observed in DIPYR. In contrast,
Et-carDIPYR has vibronic features which
much more closely resemble the absorption
spectrum of DIPYR. Like the borylated DIPYR
compound, both carDIPYRs appear to have
multiple excitations with significant oscillator
strengths; likely the S1, S2, and S3 states are juxtaposed in much the same way as they appear
in DIPYR, with the S1 and S2 manifolds overlapping in the 400-500 nm region and the S3
transition dominating the UV. -CarDIPYR has an absorption spectrum which differs
significantly from -DIPYR; while the latter has narrow, intense vibronic features wherein the
0-0 absorption peak is higher than the 0-1, the former is significantly broader, and the lowest
energy vibronic absorption peak is not as intense as the higher energy vibronic peak. The
specific vibrational energy levels to which these peaks belong have not yet been assigned.
The emission spectra of each of the three carDIPYRs are significantly
bathochromically shifted relative to the borylated derivatives. The emission of Me-carDIPYR
in acetonitrile at room temperature is Stokes shifted by 2100 cm
-1
and characterized by an
0.5
1.0
300 400 500 600 700
0.5
1.0
300 400 500 600 700
0.0
0.5
1.0

Me-carDIPYR

Et-carDIPYR

-carDIPYR
Wavelength (nm)

Figure 2.13. Normalized absorption (solid black)
and fluorescence emission of room temperature
acetonitrile (solid red) and 77 K 2-methylTHF
(dashed blue) solutions of the PF 6 salts of
Me-carDIPYR, Et-carDIPYR, and -carDIPYR.

Golden
64
emission manifold which broadened relative to DIPYR, though vibronic features are
discernable. At 77 K, these features sharpen significantly and the Stokes shift decreases (1500
cm
-1
), indicating that there is structural distortion in the excited state and that this distortion
can be minimized by freezing. In Et-carDIPYR, the room temperature emission has a much
lower Stokes shift (860 cm
-1
) which sharpens and becomes less shifted (10 nm / 410 cm
-1
) at
77 K. The radiative and non-radiative rates (Table 2.7) of Me-carDIPYR are slightly decreased
compared to DIPYR, though the difference is within experimental error, as the PL and  are
close. In Et-carDIPYR, the  PL is much lower than in Me-carDIPYR or DIPYR (0.03 vs. 0.095
and 0.10, respectively), and the fluorescence lifetime is biexponential with a significant short-
lived component (0.32 ns, 95%) and a small longer-lived component which is more similar to
that observed in Me-carDIPYR and DIPYR (3.91 ns, 5% vs 1.95 ns and 2.17 ns). While the
radiative rate of Et-carDIPYR is somewhat faster than that observed in Me-carDIPYR and
DIPYR, the non-radiative rate is approximately five times faster, leading to the much lower
quantum yield. The ethyl bridge is likely the source of the fast non-radiative rate, as the seven-
membered central ring of the fluorophore is more prone to conformational changes than the
six-membered rings of DIPYR and Me-carDIPYR.
The emission of -carDIPYR is similar to that of -DIPYR, though it is somewhat
more Stokes shifted and significantly broader. This broadening is likely due to an ion pairing
effect; alternate counter-anions should be considered to affect complete dissociation in
solution. At 77 K in 2-MeTHF, the emission spectrum blue-shifts and narrows, becoming more
similar to that of the borylated derivative. Despite these differences, the photoluminescent
quantum yield is similar to that of -DIPYR; the lifetime, however, is much longer (7.4 ns
versus 5.0 ns), corresponding to both a slower radiative rate and a slower non-radiative rate.
Molecular Design for Organic Photovoltaics
Chapter 2 | 65
This indicates that the S1-S0 transition oscillator strength is less than that of -DIPYR and that
distortion in the excited state is not severe.
Altogether, the photophysics of the carDIPYR series faithfully mirror those of the
boron difluoro DIPYR dye series, with some important caveats: the radiative rates and
nonradiative rates of the methylene-chelated carDIPYRs are slower (the ethyl-chelated
carDIPYR has a marginally faster radiative rate and a much faster non-radiative rate, leading
to a lower overall photoluminescent quantum yield), the absorption and emission spectra are
broadened relative to the DIPYRs, and both the absorption and emission spectra are
bathochromically shifted relative to the DIPYRs. The carDIPYRs, being charged complexes,
cannot be vapor deposited as easily as the zwitterionic boron difluoro DIPYRs. They are,
however, good candidates for biological studies. Their planarity make them possible
intercalants for DNA, they can be made water soluble by exchanging the counter-anion, and
their red emission has greater tissue transparency than the green emission of the boron-chelated
DIPYR dyes.
2.5 The Bright Future of DIPYR Dyes
With the insights gleaned from the preceding sections in this chapter, one can begin to
follow a set of design rules for the application of dipyridylmethene dyes to photovoltaics.
Design criteria for the application of DIPYRs as donor materials include: a long-lived exciton,
broad spectral sensitivity, and an S1 energy in the NIR – between 1.5 and 1.1 eV, ideally. In
order to have a long-lived exciton wherein non-radiative deactivation rates are minimized, the
T2 energy must be sufficiently destabilized so as not to provide an exothermic, symmetry-
allowed pathway from the singlet to the triplet manifold. This is accomplished in -DIPYR
and -DIPYR by stabilizing the HOMO and LUMO without changing the energy of the
Golden
66
LUMO+1. In fact, the HOMO energy need not be
changed at all to induce this effect; it is only necessary
to stabilize the LUMO relative to the LUMO+1. Using
the calculated DIPYR orbital densities as a guide
(Figure 2.14), it is not difficult to imagine further
substitutions which may induce this effect. The
LUMO+1, for example, can be destabilized by
substitution with electron donating groups at the 4 and 6
positions, where the LUMO has a no orbital density.
This substitution would simultaneously destabilize the
HOMO, resulting in an overall bathochromic shift of the
S1 energy, helping to meet another design criterion at the
same time. Broad spectral sensitivity is accomplished in
DIPYR where there are high oscillator strengths for
three electronic transitions: from the S0 to the S1, S2, and
S3 states. Although more difficult to predict, this can be
easily checked by calculation; whereas the absolute
energies of the state transitions are miscalculated by TD-
DFT, the relative energies are well represented and can
be corrected to better approximate experimental values
by applying the correction factor -0.44 eV to calculated
values. Finally, shifting the energy of the S1 state to the NIR will require either (or both)
stabilization of the LUMO or destabilization of the HOMO, as the S0-S1 transition is HOMO-

LUMO +1

LUMO

HOMO
Figure 2.14. HOMO / LUMO /
LUMO +1 densities and labeling
paradigm for DIPYR.

N
B
F
2
N
1
2
3
4 5 6
7
8
9
10
meso



DIPYR
11
Molecular Design for Organic Photovoltaics
Chapter 2 | 67
LUMO in all three DIPYR dyes. Again, this can be performed using the orbital densities as a
guide; for example, the HOMO has significant orbital density at the meso position whereas the
LUMO has a node in that position. Consequently, the substitution of an electron donating
group at the meso position would destabilize the HOMO without significantly effecting the
LUMO, causing a red-shift of the S1 energy relative to DIPYR. Such substitutions and the
calculated energies of the resulting state transitions are the subject of Chapter 3 in this text.
2.6 Experimental Methods
All reagents were purchased from Sigma Aldrich and used without purification. Anhydrous
1,2-dichloroethane was purchased from EMD Millipore.
2,2’-dipyridylmethane: A reported procedure was followed.
10
2-Methylpyridine (1.98g, 10
mmol) was dissolved in 20 mL of dry THF in a 100 mL schlenk flask equipped with a stir
bar and reflux condenser. The solution was cooled to -78 ˚C under a nitrogen atmosphere and
2.5 M nBuLi (8 mL, 20 mmol) was added dropwise. The solution was stirred for one hour,
then warmed to -20 ˚C whereupon 2-fluoropyridine (0.86 g, 10 mmol) was added dropwise.
The reaction was subsequently heated to reflux for 25 minutes, then hydrolyzed with ice. The
aqueous layer was separated and extracted three times with 20 mL CH2Cl2 and the organic
layers were combined and dried over Na2SO4, then reduced to a brown oil by rotary
evaporation. The product was purified by vacuum distillation at 120 ˚C, and collected as a
yellow oil (91% yield).
1
H NMR (400 MHz, Chloroform-d) δ 8.50 (d, J = 4.9 Hz, 1H), 7.67 –
7.43 (m, 2H), 7.21 (d, J = 7.8 Hz, 2H), 7.10 – 7.00 (m, 2H), 4.29 (s, 2H).
13
C NMR (101
MHz, Chloroform-d) δ 159.40, 149.36, 136.43, 123.44, 121.45, 47.28.
Golden
68
2,2’-diquinolylmethane: Procedure from the literature.
16
Quinoline (2.58 g, 20 mmol), 3 M
methylmagnesium bromide (20 mL, 60 mmol), and N,N,N’,N’-tetramethylethylenediamine
(0.3 mL, 2.0 mmol) were added to an oven dried pressure flask under a nitrogen atmosphere,
then treated with 15 mL toluene. The flask was sealed and heated to 140 ˚C overnight. The
reaction was then cooled to room temperature and carefully quenched by the addition of a
saturated sodium sulfite solution, then extracted three times into ethyl acetate, washed with
brine, dried, and condensed to a dark orange oil by rotary evaporation. The product was
purified by silica gel column chromatography with the eluent 1:1 ethyl acetate/hexanes,
yielding an orange wax. Further purification by recrystallization from CH2Cl2 layered with
hexanes yielded peach-orange crystals (30% yield).
1
H NMR (400 MHz, Chloroform-d) δ
8.14 – 8.07 (8.5 br. d, 2H), 8.03 (dd, J = 8.4, 0.8 Hz, 2H), 7.75 (dd, J = 8.2, 1.4 Hz, 2H), 7.69
(ddd, J = 8.5, 6.9, 1.5 Hz, 2H), 7.49 (ddd, J = 7.9, 6.9, 1.1 Hz, 2H), 7.41 (d, J = 8.5 Hz, 1H),
4.72 (s, 2H).
13
C NMR (101 MHz, Chloroform-d) δ 159.45, 147.87, 136.60, 129.52, 128.99,
127.52, 126.90, 126.13, 121.89, 49.00.
1,1’-diisoquinolylmethane: Procedure adapted from the literature.
16
Isoquinoline (2.58 g, 20
mmol), 3 M methylmagnesium bromide (20 mL, 60 mmol), and N,N,N’,N’-
tetramethylethylenediamine (0.3 mL, 2.0 mmol) were added to an oven dried pressure flask
under a nitrogen atmosphere, then treated with 15 mL toluene. The flask was sealed and
heated to 140 ˚C overnight. The reaction was then cooled to room temperature and carefully
quenched by the addition of a saturated sodium sulfite solution, then extracted three times
into ethyl acetate, washed with brine, dried, and condensed to a dark orange oil by rotary
evaporation. The product was purified by silica gel column chromatography with the eluent
1:1 ethyl acetate/hexanes, yielding an orange wax, which was further purified by
Molecular Design for Organic Photovoltaics
Chapter 2 | 69
precipitation from CH2Cl2 with hexanes (68% yield).
1
H NMR (400 MHz, Chloroform-d) δ
8.59 (d, J = 8.9 Hz, 2H), 8.47 (d, J = 5.7 Hz, 2H), 7.81 – 7.75 (m, 2H), 7.62 (ddd, J = 8.2,
6.8, 1.2 Hz, 2H), 7.58 – 7.49 (m, 4H), 5.38 (s, 2H).
13
C NMR (101 MHz, dmso) δ 162.30,
138.30, 132.75, 129.31, 127.06, 126.75, 126.62, 126.43, 105.15, 40.06.
General synthesis for borylation of DIPYRs: A 15 mM solution of diheteroarylmethane ligand
in dry 1,2-dichloroethane was prepared in an N2-purged schlenk flask equipped with a
magnetic stir bar and fitted with a reflux condenser. The flask was submerged in a preheated
oil bath and brought to reflux, at which time 2.0 eq. boron trifluoride diethyl etherate were
added dropwise, causing the color to change from deep yellow to an opaque, pale yellow,
corresponding to the rapid formation of precipitate. The solution was stirred for 2 hours at
reflux, then cooled to room temperature and treated with 5 eq. N,N-diisopropylethylamine,
causing the precipitate to dissolve and the solution to turn deep yellow-orange with a bright
green fluorescence. The solution was washed with water and the aqueous layer was separated
and extracted three times with dichloromethane. The organic layers were combined, dried over
sodium sulfate, filtered, and reduced to a bright orange polycrystalline solid by rotary
evaporation. The products were purified by silica gel flash chromatography with the eluent
30% dichloromethane in hexanes, followed by recrystallization. Single crystal X-ray quality
crystals were obtained for DIPYR and -DIPYR by recrystallization from the slow diffusion
of hexanes into a concentrated dichloromethane solution. Single crystals of -DIPYR were
obtained by very slow cooling (in a dewar) of a concentrated hot methanol solution to -48 ˚C
for 72 hours.
DIPYR: 2.54g (38%) yield. Orange crystals with green reflectance.
1
H NMR (400 MHz,
CDCl3) δ 7.94–7.86 (br. d, 2H, J = 5.5 Hz), 7.25 (ddd, J = 8.9, 6.7, 1.6, 2H), 6.82 (d, J = 8.8
Golden
70
Hz, 2H), 6.50 (dd, J = 6.7, 1.2 Hz, 2H), 5.36 (s, 1H).
13
C{
1
H} NMR (101 MHz, CDCl3) δ
149.3, 136.8, 135.6, 121.2, 111.4, 111.4, 85.3. Spectra match those reported in the literature.
11

MS (MALDI) m/z calcd for C11H9BF2N2
+
218.083 [M]
+
; found 217.906.
 ‑ DIPY R : 0.27 g (39%) yield. Orange crystals with green reflectance.
1
H NMR (400 MHz,
CDCl3) δ 8.57 (dt, J = 8.8, 3.5 Hz, 2H), 7.64–7.54 (m, 4H), 7.51 (dd, J = 7.8, 1.5 Hz, 2H),
7.32–7.26 (m, 2H), 6.85 (d, J = 9.0 Hz, 2H), 5.48 (s, 1H).
13
C{
1
H} NMR (101 MHz, CDCl3) δ
149.7, 139.8, 136.5, 130.8, 128.1, 124.8, 123.8, 121.9, 121.4 (t, J = 8.3 Hz), 91.5. Anal. Calcd
for C19H13BF2N2: C, 71.73; H, 4.12; N, 8.81. Found: C, 71.35; H, 4.24; N, 8.55. A melting
temperature determination was attempted, but sublimation was observed at atmospheric
pressure in the capillary tube at 217 ˚C. MS (MALDI) m/z calcd for C19H13BF2N2
+
318.114
[M]
+
; found 317.999.
 ‑D IPYR: 0.84 g (36%) yield.

Red-orange crystals with green reflectance
.
1
H NMR (400 MHz,
CDCl3) δ 8.37 (d, J = 7.8 Hz, 2H), 7.91–7.85 (br. d, J = 7.0 Hz, 2H), 7.68 (ddd, J = 8.0, 6.8,
1.2 Hz, 2H), 7.64–7.56 (m, 4H), 7.16 (s, 1H), 6.97–6.92 (br. d, J = 7.1 Hz, 2H).
13
C{
1
H} NMR
(101 MHz, CDCl3) δ 148.6, 135.1, 131.7, 130.9, 127.6, 127.0, 124.9, 124.7, 112.1, 81.4. Anal.
Calcd for C19H13BF2N2: C, 71.73; H, 4.12; N, 8.81. Found: C, 71.57; H, 4.16; N, 8.57. A
melting temperature determination was attempted, but sublimation was observed at
atmospheric pressure in the capillary tube at 258 ˚C. MS (MALDI) m/z calcd for C19H13BF2N2
+

318.114 [M]
+
; found 318.010.
-carDIPYR: 2,2’-diquinolylmethane (0.470 g, 1.74 mmol, 1 eq.) was dissolved in 20 mL
dry acetonitrile under a nitrogen atmosphere. Dibromomethane (0.611 mL, 8.69 mmol, 5 eq.)
was added to the solution, and after 1.5 hours, diisopropylethylamine (0.567 mL, 3.48 mmol,
Molecular Design for Organic Photovoltaics
Chapter 2 | 71
2 eq.) was added. The solution was stirred at reflux for 4 days, after which only a small
amount of product was observed. An additional 3 mL of dibromomethane was added to the
solution, and the reaction was refluxed for a further 3 days, after which the solvent was
evaporated off and the red solid was re-dissolved in dichloromethane. The organic solution
was rinsed with 100 mL water followed by 100 mL saturated ammonium chloride solution.
The organics were then dried over sodium sulfate and the solvent was removed by rotary
evaporation, producing a red solid which was recrystallized from DCM layered with hexanes.
The solid residue was filtered off, mixed into a rapidly stirring aqueous solution of potassuim
hexafluorophosphate, then filtered again. The red residue was finally loaded onto basic
alumina and purified by flash chromatography on alumina with a gradient eluent 100%
dichloromethane to 5% methanol in dichloromethane, producing the pure product as a
brightly orange-emissive magenta solid. Product was characterized by LCMS; due to ion
pairing effect, NMR spectra could not be resolved. For NMR characterization, counter-anion
should be varied.
2.7 Bibliography
1. Chen, J. J.; Conron, S. M.; Erwin, P.; Dimitriou, M.; McAlahney, K.; Thompson, M.
E., High-Efficiency BODIPY-Based Organic Photovoltaics. ACS Appl. Mater Interfaces 2015,
7 (1), 662-669.
2. Loudet, A.; Burgess, K., BODIPY Dyes and Their Derivatives: Syntheses and
Spectroscopic Properties. Chem. Rev. 2007, 107 (11), 4891-4932.
3. Treibs, A.; Kreuzer, F.-H., Difluorboryl-Komplexe von Di- und Tripyrrylmethenen.
Justus Liebigs Annalen der Chemie 1968, 718 (1), 208-223.
4. Schmitt, A.; Hinkeldey, B.; Wild, M.; Jung, G., Synthesis of the Core Compound of
the BODIPY Dye Class: 4,4′-Difluoro-4-bora-(3a,4a)-diaza-s-indacene. J. Fluoresc. 2009, 19
(4), 755-758.
5. Tram, K.; Yan, H.; Jenkins, H. A.; Vassiliev, S.; Bruce, D., The synthesis and crystal
structure of unsubstituted 4,4-difluoro-4-bora-3a,4a-diaza-s-indacene (BODIPY). Dyes Pigm.
2009, 82 (3), 392-395.
6. Arroyo, I. J.; Hu, R.; Merino, G.; Tang, B. Z.; Peña-Cabrera, E., The Smallest and One
of the Brightest. Efficient Preparation and Optical Description of the Parent
Borondipyrromethene System. J. Org. Chem. 2009, 74 (15), 5719-5722.
Golden
72
7. 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. Chem. Eur. J. 2004,
10 (19), 4853-4871.
8. Kubota, Y.; Tsuzuki, T.; Funabiki, K.; Ebihara, M.; Matsui, M., Synthesis and
Fluorescence Properties of a Pyridomethene−BF2 Complex. Org. Lett. 2010, 12 (18), 4010-
4013.
9. Sathyamoorthi, G.; Soong, M.-L.; Ross, T. W.; Boyer, J. H., Fluorescent tricyclic β-
azavinamidine–BF2 complexes. Heteroat. Chem 1993, 4 (6), 603-608.
10. Dyker, G.; Muth, O., Synthesis of Methylene- and Methine-Bridged Oligopyridines.
Eur. J. Org. Chem. 2004, 2004 (21), 4319-4322.
11. Douglass, J. E.; Barelski, P. M.; Blankenship, R. M., Diazaboracyclic cations. III. A
homomorph of 9,10-dihydroanthracene. J. Heterocycl. Chem. 1973, 10 (2), 255-257.
12. Golden, J. H.; Facendola, J. W.; Sylvinson M. R, D.; Baez, C. Q.; Djurovich, P. I.;
Thompson, M. E., Boron Dipyridylmethene (DIPYR) Dyes: Shedding Light on Pyridine-
Based Chromophores. J. Org. Chem. 2017, 82 (14), 7215-7222.
13. Golden, J. H.; Estergreen, L.; Porter, T.; Tadle, A. C.; Sylvinson M. R, D.; Facendola,
J. W.; Kubiak, C. P.; Bradforth, S. E.; Thompson, M. E., Symmetry-Breaking Charge Transfer
in Boron Dipyridylmethene (DIPYR) Dimers. ACS Appl. Energy Mater. 2018.
14. Burns, C. T.; Jordan, R. F., Ethylene Polymerization by Sterically Crowded
Palladium(II) Complexes that Contain Bis(heterocycle)methane Ligands. Organometallics
2007, 26 (27), 6737-6749.
15. Engl, R. B.; Ingraham, L. L., Derivatives of 1-Methylisoquinoline. J. Org. Chem. 1961,
26 (12), 4933-4936.
16. Stephens, D. E.; Nguyen, V. T.; Chhetri, B.; Clark, E. R.; Arman, H. D.; Larionov, O.
V., Organocatalytic Synthesis of Methylene-Bridged N-Heterobiaryls. Org. Lett. 2016, 18
(22), 5808-5811.
17. Scheibe, G.; Daltrozzo, E., Diquinolylmethane and Its Analogs. In Adv. Heterocycl.
Chem., Katritzky, A. R.; Boulton, A. J., Eds. Academic Press: 1967; Vol. 7, pp 153-181.
18. Cruickshank, D. W. J.; Sparks, R. A., Experimental and Theoretical Determinations of
Bond Lengths in Naphthalene, Anthracene and Other Hydrocarbons. Proc. R. Soc. London,
Ser. A 1960, 258 (1293), 270.
19. Bondi, A., van der Waals Volumes and Radii. J. Phys. Chem. 1964, 68 (3), 441-451.
20. Hierso, J. C., Indirect Nonbonded Nuclear Spin-Spin Coupling: A Guide for the
Recognition and Understanding of "Through-Space" NMR J Constants in Small Organic,
Organometallic, and Coordination Compounds. Chem. Rev. 2014, 114 (9), 4838-4867.
21. Parton, R. L.; Lenhard, J. R., Dimerization reactions of cyanine radical dications. J.
Org. Chem. 1990, 55 (1), 49-57.
22. Lenhard, J. R.; Parton, R. L., Electrochemical and spectroscopic analyses of the
thermodynamics of the reversible dimerization of cyanine radical dications. J. Am. Chem. Soc.
1987, 109 (19), 5808-5813.
23. Lenhard, J. R.; Cameron, A. D., Electrochemistry and electronic spectra of cyanine dye
radicals in acetonitrile. J. Phys. Chem. 1993, 97 (19), 4916-4925.
24. Nepomnyashchii, A. B.; Brö ring, M.; Ahrens, J.; Bard, A. J., Chemical and
electrochemical dimerization of BODIPY compounds: electrogenerated chemiluminescent
detection of dimer formation. J. Am. Chem. Soc. 2011, 133 (48), 19498-19504.
Molecular Design for Organic Photovoltaics
Chapter 2 | 73
25. Benniston, A. C.; Harriman, A.; Whittle, V. L.; Zelzer, M.; Harrington, R. W.; Clegg,
W., Exciplex-like emission from a closely-spaced, orthogonally-sited anthracenyl-boron
dipyrromethene (Bodipy) molecular dyad. Photochem. Photobiol. Sci. 2010, 9 (7), 1009-1017.
26. Nepomnyashchii, A. B.; Bard, A. J., Electrochemistry and Electrogenerated
Chemiluminescence of BODIPY Dyes. Acc. Chem. Res. 2012, 45 (11), 1844-1853.
27. W. D’Andrade, B.; Datta, S.; R. Forrest, S.; Djurovich, P.; Polikarpov, E.; Thompson,
M., Relationship between the ionization and oxidation potentials of molecular organic
semiconductors. Org. Electron. 2005, 6 (1), 11-20.
28. Djurovich, P. I.; Mayo, E. I.; Forrest, S. R.; Thompson, M. E., Measurement of the
lowest unoccupied molecular orbital energies of molecular organic semiconductors. Org.
Electron. 2009, 10 (3), 515-520.
29. Sidman, J. W., Electronic and Vibrational States of Anthracene. J. Chem. Phys. 1956,
25 (1), 115-121.
30. Granovsky, A. A., Extended multi-configuration quasi-degenerate perturbation theory:
The new approach to multi-state multi-reference perturbation theory. J. Chem. Phys. 2011, 134
(21), 214113.
31. Momeni, M. R.; Brown, A., Why Do TD-DFT Excitation Energies of BODIPY/Aza-
BODIPY Families Largely Deviate from Experiment? Answers from Electron Correlated and
Multireference Methods. J. Chem. Theory Comput. 2015, 11 (6), 2619-2632.
32. Charaf-Eddin, A.; Le Guennic, B.; Jacquemin, D., Excited-states of BODIPY-
cyanines: ultimate TD-DFT challenges? RSC Adv. 2014, 4 (90), 49449-49456.
33. Dähne, S., Color and Constitution: One Hundred Years of Research. Science 1978, 199
(4334), 1163.
34. Hanson, K.; Roskop, L.; Djurovich, P. I.; Zahariev, F.; Gordon, M. S.; Thompson, M.
E., A Paradigm for Blue- or Red-Shifted Absorption of Small Molecules Depending on the
Site of π-Extension. J. Am. Chem. Soc. 2010, 132 (45), 16247-16255.
35. Bergström, F.; Mikhalyov, I.; Hägglöf, P.; Wortmann, R.; Ny, T.; Johansson, L. B. Å.,
Dimers of Dipyrrometheneboron Difluoride (BODIPY) with Light Spectroscopic Applications
in Chemistry and Biology. J. Am. Chem. Soc. 2002, 124 (2), 196-204.
36. Nijegorodov, N.; Winkoun, D. P., Dependence of the fluorescence parameters and the
intersystem crossing rate constant on the orbital nature of the s1 state of catacondensed
aromatics. Spectrochim. Acta, Part A 1997, 53 (12), 2013-2022.
37. Turro, N. J. R., V.; Scaiano, J. C., Principles of Molecular Photochemistry: An
Introduction. University Science Books: Sausalito, California, USA, 2009.
38. Kellogg, R. E., Second Triplet State of Anthracene. J. Chem. Phys. 1966, 44 (1), 411-
412.
39. Leubner, I. H., Synthesis and properties of pyrido- and azapyridocyanines. J. Org.
Chem. 1973, 38 (6), 1098-1102.
40. Leubner, I. H., Nuclear magnetic resonance spectra of pyrido‐ and azapyridocyanines.
Org. Magn. Reson. 1974, 6 (5), 253-258.

Golden
74
Chapter 3

Near-Infrared Absorption in Dipyridylmethene Dyes

3.1 Introduction to Near Infrared Absorbing Dyes for Organic Photovoltaics
Poor power conversion efficiencies in organic photovoltaics stem largely from (1)
imperfect control of the donor/acceptor charge transfer state and from (2) transmission losses
in the red and near infrared regions of the solar spectrum due to absorption from active layer
materials in this region. While the former challenge represents a significant source of energy
loss affecting the VOC,
1-4
the latter directly affects the magnitude of photo-generated charge
carriers by limiting JSC.
3-6
Red and near infrared (NIR) molecular absorbers with high
photoluminescent quantum yields  PL are in high demand in many fields,
7-12
including but not
limited to solar energy conversion.
13-16
Materials with these properties, however, are few and
far between, and while there have been major research efforts to develop small molecule
absorbers with efficient NIR absorption and emission, few publications in recent decades have
shown successful development of such compounds.
8, 12, 17-18
The substrate scope comprising
the NIR absorbing materials class has remained largely unchanged over the past several
decades; generally, these materials involve extensive -conjugation and/or charge resonance
effects and are based upon one of the following chromophore classes: polymethines (such as
cyanines and squaraines), pentacene and larger acenes, annulenes, pthalocyanines, metal
ditholenes, quinones, and azo dyes.
7
The most intensely absorbing of these materials are those
in which a -  * transition characterizes the natural transition orbitals of the excited state. -
Molecular Design for Organic Photovoltaics
Chapter 3 | 75
 * transitions are, by virtue of inherent orbital overlap, characterized by high oscillator
strengths. In order to bathochromically shift the -  * transition to the NIR, however, extensive
 conjugation is required (as is represented by the large and highly -conjugated molecular
structures which encompass this small class of chromophores). Each of the above materials
exhibiting high molar absorptivities in the NIR exhibits one or more of the following problems
limiting their applicability to OPVs: namely, low PL due to high spin-orbit coupling
promoting high intersystem crossing rates or high non-radiative relaxation rates due to
distortion, limited processability due to insolubility or molecular charge, photobleaching under
the oxidizing/reducing conditions inherent to OPVs, or limited scalability due to either
synthesis or purification.
Alternative strategies to induce bathochromic shifts include the design of materials with
n-  * transitions or radical character, but these strategies tend to produce weak, tailing
absorption manifolds and negligible fluorescence quantum yields due to rapid intersystem
crossing rates.
7
Likewise, incorporation of significant intramolecular charge transfer character
into the lowest energy excited state has been shown to shift absorption and emission profiles
to the NIR, but this method also sacrifices molar absorptivity and  PL by limiting HOMO-
LUMO overlap.
7, 17
Both intense molar absorptivity (>10
4
M
-1
cm
-1
) and  PL values which
approach unity are critical properties in the development of active layer materials for efficient
OPVs, as both these properties directly affect the photon absorption and exciton migration
steps of photocurrent generation (Chapter 1). The lack of available NIR absorbing materials
which meet the requirements for application to OPVs is perhaps the most significant
impediment to the commercialization of this technology.
Golden
76
Further affecting  PL in the NIR is the energy gap law. This law states that internal
conversion from the lowest energy excited state to the ground state via vibrational and
rotational modes become increasingly allowed as the difference in energy between the excited
and ground states becomes smaller. The law results simply from increased overlap between
state energy manifolds that facilitates non-radiative vibronic relaxation processes. As a
consequence, excited state lifetimes become shorter and PL efficiencies decrease as the S0-S1
energy gap is decreased. Therefore, as the Eg of active layer materials is pushed into the NIR,
where the Shockley-Queisser limit holds that the theoretical efficiency is maximized, the
energy gap law becomes increasingly problematic and it becomes more important to minimize
other sources of non-radiative relaxation such as distortion and intersystem crossing in order
to reduce thermalization losses.
3-5, 19

Novel red and NIR absorbing materials in recent literature have largely been based
upon donor-acceptor-donor or acceptor-donor-acceptor architectures leading to intramolecular
charge transfer excitons
20-21
or are derivatives of previously studied chromophore classes
including those mentioned above, while the most recent novel NIR absorbing architectures
have been based upon -extended and aza-BODIPY dyes.
12, 22-24
These structures however,
often involve a high degree of material complexity leading to difficulties of scale, purification,
and/or thin film processing.
The discovery of a new intensely absorbing (  = 10
4
to 10
5
M
-1
cm
-1
) and emissive
(  PL = 0.8) materials class, the dipyridylmethene (DIPYR) dyes (Chapter 2),
25
represents an
exciting opportunity to expand the substrate scope of NIR absorbing materials beyond the few
known structure classes which have pervaded the literature for decades. DIPYRs present many
opportunities for design of OPV-friendly materials via site-specific functionalization to tune
Molecular Design for Organic Photovoltaics
Chapter 3 | 77
the highly allowed - *, S0-S1 transition in a rigid, neutral material which is both intensely
emissive and easily synthesized and processed. Toward this goal, several NIR-absorbing
DIPYR derivatives have been modeled using the physical-theoretical characterization methods
discussed in the previous chapter. In silico studies of representative DIPYR structural
modifications and synthetic progress toward the development of NIR absorbing and emitting
dipyridylmethene derivatives are described in the following sections.
3.2 meso-Substituted DIPYR Dyes
In Chapter 2 of this dissertation, the orbital characteristics leading to the observed
absorption and emission properties in DIPYR, -DIPYR, and -DIPYR were parsed in order
to develop a set of design rules which could be applied to the synthesis of NIR-absorbing
dipyridylmethene dyes. The dominant orbitals contributing to the S0-S1 excited state transition
were found to be the HOMO and LUMO orbitals, and the positions in which one of these
orbitals is defined by significant density while the other is defined by a node were identified
as ideal for site-specific modifications to affect a change in the energy of the S 1 excited state.
Keeping in mind one of the other lessons learned from the exploration of the photophysics of

HOMO LUMO meso-substitution
Figure 3.1. The HOMO of DIPYR is characterized by a large lobe of orbital density at the meso position, while the
LUMO is characterized by a node at this position. Thus, substitution of the meso-H with an electron donating group
will destabilize the HOMO without significantly affecting the LUMO, reducing the energy of the HOMO-LUMO
S 1 transition and causing a bathochromic shift in the absorption and emission.
Golden
78
these materials, stabilization of the
LUMO with respect to the LUMO+1 was
found to be necessary to avoid rapid
intersystem crossing to a low-energy T2
excited state, which results in a low
photoluminescent quantum yield and a
high non-radiative decay rate.
Computational screening of target compounds was performed using the parent DIPYR
as a model complex. The meso position was identified as an ideal position for modification
both for reasons of synthetic feasibility and due to its being both the locus of a large degree of
HOMO orbital density and a of node in the LUMO density (Figure 3.1). Increasingly electron
donating groups were appended to the meso position in silico (Table 3.1) and the effect on the
HOMO energy was calculated. Predictably, the HOMO gradually increases in energy due to
the presence of the destabilizing electron donating groups, with -NH2 causing the greatest
effect: a 0.86 eV destabilization compared to the parent DIPYR. The LUMO in meso-amino
DIPYR, on the other hand, is only marginally destabilized (0.04 eV), leading to a bathochromic
shift in the calculated singlet energy of 0.68 eV – a 200 nm red-shift compared to DIPYR.
Notably, the even more electron-donating substituent dimethylamine induced no such
destabilization in the HOMO energy. This is due to the fact that the adjacent 4,6 hydrogens in
DIPYR cause the bulky dimethylamino group to twist out of the plane of the chromophore,
negating its electron donating effect entirely. Theoretically this can be remedied by fusing a
tertiary amine to the 4 and 6 positions of DIPYR, replacing the offending hydrogens with a
carbon bridge to the amine and locking the amine into the plane of the chromophore. Although
Table 3.1. Calculated HOMO, LUMO, and S 1 state
energies for meso-substituted DIPYR.
-R HOMO
(eV)
LUMO (eV) S 1 (nm, eV)*
-H -4.87 -1.42 461, 2.57
-CH 3 -4.75 -1.41
-OH -4.76 -1.47
-OCH 3 -4.74 -1.42
-NH 2 -4.01 -1.38 660, 1.89
-N(CH 3) 2 -4.79 -1.37
fused -NEt 2 -4.15 -1.30 599, 2.07
*S 1 energies have been corrected by -0.44 eV.
Molecular Design for Organic Photovoltaics
Chapter 3 | 79
fusion of a meso-diethylamino substituent to the 4 and 6
positions did repair the amine to its former electron-
donating role, the overall effect was less destabilizing to
the HOMO and more destabilizing to the LUMO than was
observed the primary amine, resulting in a less extensive
bathochromic shift compared to the meso-amino DIPYR.
In Chapter 2 it was shown that when the
LUMO and LUMO+1 are close in energy, the S1-T2
intersystem crossing rate is high, which leads to rapid
internal conversion to the T1 and represented a major
source of thermalization loss and a reduction in quantum
efficiency.
25
It was also shown that benzannulation at
either the  or  positions resulted in a stabilization of the HOMO and LUMO energies without
affecting the LUMO+1, thereby increasing the LUMO-LUMO+1 energy gap and making the
S1-T2 intersystem crossing process endothermic. DFT showed that in the case of meso-amino
DIPYR, the LUMO and LUMO+1 energies remained close, so the meso-amino quinoline
DIPYR was also calculated. In the case of the quinoline derivative, meso-amino- -DIPYR, the
calculated HOMO and LUMO energies are -4.20 and -1.96 eV, respectively (Figure 3.2),
resulting in a calculated S1 absorption energy of 790 nm (1.57 eV). This considerable
bathochromic shift compared to -DIPYR, which has an S1 energy of 524 nm (2.37 eV), comes
not only from the predicted destabilization of the HOMO due to the presence of the meso-
amino group, but also from a significant stabilization of the LUMO. These in silico data
suggested meso-amino- -DIPYR is an ideal candidate to affect a significant shift to a deep red

LUMO
-1.96 eV

HOMO
-4.30 eV
Figure 3.2. HOMO and LUMO of
meso-amino- -DIPYR.
Golden
80
absorption energy with a single simple structural modification. Indeed, the meso-amino
substitution in -DIPYR requires only two additional synthetic steps from the
diquinolylketone intermediate generated by ligand synthesis Method II (Figure 2.4).
Two proposed retrosyntheses of meso-amino- -DIPYR from 2-bromoquinoline
starting materials are outlined in Scheme 3.1. Route I involves the formation of 2,2’-
diquinolylketone by nucleophilic attack of 2-lithioquinoline on a carbonate. This procedure is
known and was used in the synthesis of -DIPYR.
26
After formation of the ketone, the oxime
is formed in the presence of hydroxylamine hydrochloride and strong base, then is reduced to
the primary amine. This reactivity has been demonstrated in the analogous
dipyridylmethaneamine system.
27
The diquinolylmethanamine complex was synthesized as
described and characterized by LCMS in relatively high purity, but was found not to run on
silica gel. The crude material was subjected to borylation conditions with an added equivalent
Scheme 3.1. Retrosyntheses of meso-amino- -DIPYR from reaction of 2-bromoquinoline with (I) diethyl
carbonate or (II) thiophosgene.

Molecular Design for Organic Photovoltaics
Chapter 3 | 81
of BF3OEt2, in order to chelate the amine group. These conditions proved too acidic, however,
and deamination of the base resulted, forming -DIPYR as the major product. It is evident that
an amine protecting group is necessary. Common protecting groups for primary amines are
Boc (tert-butyloxycarbonyl) and Fmoc (fluorenylmethyloxycarbonyl). Although Boc is
usually the protecting group of choice, it is acid labile and will not withstand the highly Lewis
acidic conditions caused by the two-hour reflux in BF3OEt2. Fmoc is acid stable and base
labile; this is an optimal situation, as the protected ligand can be borylated in the presence of
BF3OEt2, then conjugated and deprotected in situ to yield the product meso-amino- -DIPYR.
An alternative to this synthesis is outlined in Scheme 3.1, retrosynthesis II. In this
procedure, the 2-bromoquinoline starting material is again lithiated, then reacted with the hot
electrophile thiophosgene. Upon formation of the thioketone, reduction to the thiomethane can
be affected by reaction with methyl Grignard, or with lithium diisopropymamide and
iodomethane by analogy to the reactivity of thiobenzophenone.
28-29
The thiomethane is
sufficiently protected to endure the conditions of borylation, forming the meso-thiomethyl- -
DIPYR, a deep blue emissive fluorophore due to significant stabilization of the HOMO by
substitution with the electron withdrawing thiomethyl group. The borylated dye should in this
case be easily purified and identified. Functional group substitution of the thiomethane to
amine can be accomplished thereafter by reaction with ammonia in polar aprotic solvent or by
reaction with ammonium acetate.
30-31
Alternative functional group substitutions to make a
library of meso-substituted -DIPYR dyes should also be possible via this method. This
research is ongoing at time of submission.

Golden
82
3.3 Linearly -extended DIPYRs
In addition to push-pull substitutions, increasing the extent of -conjugation is a tried-
and-true method used, with few exceptions,
24, 32
to affect a bathochromic shift in absorption
and emission by reducing the HOMO-LUMO gap. DIPYR is a heterocyclic analogue of
anthracene and has been shown to share many structural and photophysical qualities with
anthracene.
25
Thus, the photophysical properties of the acene group can be used to provide a
first approximation for the properties of -extended DIPYRs. Whereas benzene and
naphthalene are UV absorbing materials, anthracene is bathochromically shifted to the deep
blue portion of the visible spectrum. Extending the -system by one more ring, to tetracene,
shifts the absorption spectrum into the blue-green. Pentacene, meanwhile, is red absorbing.
With increasing linear benzannulation, the HOMO-LUMO gap decreases, reducing the S1 state
energy. Due to the heteroaromatic rings which comprise the DIPYR system, any even-
numbered ring system will provide an inherently asymmetric electronic symmetry in the
ground state; i.e. an even-numbered DIPYR annulation will result in a permanent dipole along
both the long and short axes in the ground state (whereas the parent three-membered ring
DIPYR has a permanent dipole along the short -BF2 containing axis only). An odd-numbered
annulation, e.g. a pentacene or heptacene derivative, will result in an electronic system which
is a direct analogue of DIPYR, albeit with a lower HOMO-LUMO energy gap. The geometry
optimized structure of the symmetric pentacene derivative boron difluoro 1,1’-
diisoquinolylmethene ( -DIPYR) was calculated at the B3LYP / 6-311G** level of theory,
and TD-DFT was performed to determine the singlet and triplet state energies.
The symmetric diisoquinolylmethene analogue of pentacene, -DIPYR, has a
calculated lowest singlet excitation energy of 1.51 eV, corresponding to an absorption at
Molecular Design for Organic Photovoltaics
Chapter 3 | 83
820 nm and representing a 1.06 eV
bathochromic shift relative to DIPYR.
However, the lessons learned from the
juxtaposition of state energies in DIPYR
very much apply to -DIPYR. In -DIPYR,
much like in DIPYR, the T2 state (1.49 eV)
is slightly lower in energy than the S1 state.
This fact alone is not enough to exclude -
DIPYR as a candidate for high quantum
efficiency fluorescence emission (a desired materials screening quality as it is an indicator of
limited thermalization energy losses and high FRET efficiency), but the symmetry of the T2
state (B2) is different than that of the S1 state (A1), which means that while intersystem crossing
from the singlet to the triplet manifold remains spin forbidden, it is both exothermic and
symmetry-allowed. Furthermore, the oscillator strength for the S1 excitation is very low
(f = 0.02), while the same transition in DIPYR is much more highly allowed (f = 0.232),
meaning that the S1 absorption is expected to be quite weak in -DIPYR. In -DIPYR as well
as in DIPYR, the S2 state energy (1.67 eV) is close to that of the S1, but again the oscillator
strength corresponding to this transition is low (f = 0.01). Summarizing these results, the
expected photophysical behavior of -DIPYR will involve low-energy absorptions with low
molar extinction coefficients and which lead to intersystem crossing rates that are competitive
with fluorescence emission; i.e. -DIPYR will most likely appear to be a predominately blue
absorber (corresponding to S0→S3 and higher excitations) which thermalizes energy rapidly
through internal conversion to the S1 state, then further via intersystem crossing to the T2 state

Figure 3.3. Jablonski diagram depicting state energies
in -DIPYR. Low oscillator strengths calculated by
TD-DFT for the S 1 and S 2 states correspond to
predicted weak absorption in the NIR. Furthermore,
endothermic and symmetry-allowed ISC to the T 2 will
result in a low  PL.
Golden
84
followed by internal conversion to a low energy T1 state (1.29 eV) (Figure 3.3). Due to a lack
of spin-orbit coupling which might otherwise facilitate phosphorescence emission from this
triplet, decay of this state will likely take the form of non-radiative vibronic relaxation. Such a
system is non-ideal for application to OPVs due to inefficient absorption, extensive
thermalization energy losses through the conversion of high energy singlet excitons to low
energy triplet excitons, and poor exciton migration resulting both from the low S0→S1
oscillator strength and from the formation of triplet excitons which are constrained to the
Dexter energy transfer mechanism.

Figure 3.4. Linearly -extended DIPYR dyes.

Molecular Design for Organic Photovoltaics
Chapter 3 | 85
3.4. DoubleDIPYRs
An alternative strategy to
affect a linear -extension in DIPYR
is to fuse two DIPYR moieties
together. Synthesis of a pentacene
analogue of DIPYR can thereby be
accomplished using either a
pyrimidine or a pyrazine central
heteroaromatic ring to form a ligand
with two coordination sites. Boron
difluoro DIPYR dyes are illustrated
in Figure 3.4, but coordination of
alternative groups (carbon, metals),
which need not be identical, is also
possible. A symmetric tetracene analogue built upon an azo group can also be envisioned. Each
of these doubleDIPYR compounds was modeled computationally in an effort to screen for
DIPYR dyes capable of serving as efficient red-absorbing donor materials in OPVs.
Both the pyrimidine and pyrazine doubleDIPYR derivatives were found to have
predominantly planar geometries, with out-of-plane puckering of the BF2 moieties on the same
order as is observed in DIPYR, a fact that bodes well for low non-radiative decay rates in these
materials. TD-DFT was used to estimate the energies of state transitions in the doubleDIPYRs
(Table 3.2) and it was observed that pyrimidine-DIPYR2, like -DIPYR, has a very low S0-S1
oscillator strength (0.02). Interestingly the S1 state energy in pyrimidine-DIPYR2 (2.19 eV,
Table 3.2. Orbital contributions, energies, and oscillator
strengths of major S 0 → S n and S 0 → T n transitions in the
double DIPYR series (corrected).
pyrimidine-DIPYR2
state transition %  (nm, eV) ƒ sym.
S 1 91 → 93 96 566, 2.19 0.02 B 2
S 2 91 → 92 94 556, 2.23 0.77 A 1

pyrimidine- -DIPYR2
state transition %  (nm, eV) ƒ sym.
S 1 117 → 118 99 617, 2.01 1.12 B 2

pyrazine-DIPYR2
state transition %  (nm, eV) ƒ sym.
S 1 91 → 92 98 1060, 1.17 0.12 B u
S 2 91 → 93 90 574, 2.16 0.63 B u
T 1 91 → 92 1810, 0.69 0.00 B u
T 2 91 → 93 727, 1.70 0.00 B u

pyrazine- -DIPYR2
state transition %  (nm, eV) ƒ sym.
S 1 117 → 118 97 1078, 1.15 0.20 B u
S 2 117 → 119 93 620, 2.0 1.01 B u
T 1 117 → 118 1653, 0.75 0.00 B u
T 2 117 → 119 775, 1.59 0.00 B u

Golden
86
566 nm) is also only marginally red-shifted from that of DIPYR, much less so than is observed
in -DIPYR. Its poor oscillator strength also appears to stem from the natural transition orbitals
comprising the S1 state (Figure 3.5), which show HOMO-derived hole density delocalized
across the cyanine-like carbon backbone of the chromophore while the particle density is
confined largely to the central ring (LUMO+1). Thus, this transition involves charge transfer
from the outer rings to the central ring. The transition dipole moment for a transition of this

pyrimidine-DIPYR2 (C2v)
92 (-1.83 eV) a 2
LUMO
93 (-1.57 eV) b 1
LUMO +1

91 (-4.69 eV) b 1
HOMO

pyrimidine- -DIPYR2 (C2v)
118 (-2.20 eV) a 2
LUMO

117 (-4.86 eV) b 1
HOMO

Figure 3.5. Natural transition orbitals in pyrimidine-DIPYR 2 derivatives.

Molecular Design for Organic Photovoltaics
Chapter 3 | 87
type is small, as the two charge transfer vectors (from outside rings to central ring) point in
opposing directions, resulting in a net cancellation of magnitude. The transition, then, is
formally forbidden since the ground and excited states have the same electronic symmetry (b1).
This fact coupled with limited orbital overlap results in the very low calculated oscillator
strength. While the S2 state energy (a HOMO to LUMO transition) is very close to that of the
calculated S1 and is characterized by a much higher oscillator strength (f = 0.77), orange-red
absorption is not the goal of this work. A marginally increased bathochromic shift (to 2.01 eV,
617 nm) and marked increase in oscillator strength (f = 1.12)
*
for the S0-S1 transition may be
accomplished by substituting quinolines for the outer pyridine rings in pyrimidine-DIPYR2.
Although the pyrimidine-DIPYR2 structures are interesting from a physical-theoretical
perspective, their synthetic complexity compared to the monomer DIPYR dyes and lack of
significant bathochromic shifts in absorption energies exclude them from this particular
materials screen.
Substituting the central ring for pyrazine rather than pyrimidine yields a fundamentally
different electronic structure (Figure 3.6). The symmetry of the pyrazineDIPYR2 dyes is C2h,
rather than C2v as is observed in the pyrimidineDIPYR2 dyes and DIPYR monomers, and the
lowest energy singlet excitation S1 is significantly bathochromically shifted, to 1.17 eV
(1060 nm). This, incidentally, is almost identical to the bandgap energy of silicon (1.1 eV at
300 K) and represents a full 1.5 eV bathochromic shift relative to DIPYR. The S1 transition is
characterized by a HOMO to LUMO transition which is, like pyrimidine-DIPYR2, largely a
symmetric charge transfer. In contrast to pyrimidineDIPYR2, however, charge is moved from

*
Although oscillator strengths greater than 1 appear erroneous, the Thomas-Reiche-Kuhn sum rule states that the
sum of all oscillator strengths must be equal to the number of electrons. Thus for materials with many forbidden
transitions, as will be the case in these highly symmetric doubleDIPYRs, oscillator strengths for individual
transitions may indeed be greater than 1.
Golden
88
the boron-containing rings to the central ring (rather than the peripheral rings to the central
ring); further, the symmetry of the HOMO is bu while that of the LUMO is ag, resulting in this
transition being symmetry allowed. Due to the allowedness of this transition and the fair orbital
overlap between the HOMO and LUMO, the oscillator strength f is found to be 0.12, 6-fold
higher than that of pyrimidineDIPYR2 and about half that of the same transition in DIPYR).
Whereas pyrimidineDIPYR2 and DIPYR have very low S1 / S2 energy gaps, the energy gap
pyrazine-DIPYR2 (C2h)
92 (-2.48 eV) a g
LUMO
93 (-1.58 eV) a g
LUMO+1

91 (-4.42 eV) b u
HOMO

pyrazine- -DIPYR2 (C2h)
118 (-2.47 eV) a g
LUMO
119 (-1.91 eV) a g
LUMO+1

117 (-4.49 eV) b u
HOMO

Figure 3.6. Natural transition orbitals in pyrazine-DIPYR 2 derivatives.

Molecular Design for Organic Photovoltaics
Chapter 3 | 89
between these states in pyrazine-DIPYR2 is 1.0 eV. The S2 state is a HOMO to LUMO+1
transition, and because the LUMO and LUMO+1 have the same orbital symmetry (ag), the
overall symmetry of both the S1 and S2 state is Bu. This fact means that while the S0-S1 and S0-
S2 transitions are characterized by high oscillator strengths, internal conversion from S2 to S1
is symmetry forbidden. Furthermore, the large energy difference between S2 and S1 states
provides an energetic barrier which raises the possibility that non-Kasha fluorescence emission
from the S2 may be competitive with internal conversion to the S1 state, i.e. in this material it
is possible that emission can occur from both the S2 and S1 states. From a purely physical-
theoretical perspective this material represents an exciting possibility for a paradigm shift; that
a small molecule organic material might exhibit NIR absorption and emission and non-Kasha
emission.
The rate of intersystem crossing in pyrazine-DIPYR2 can be estimated again from the
symmetry and juxtaposition of its state energies. The T1 and T2 states are both Bu symmetric,
as were the S1 and S2 states, making internal conversion both a spin and symmetry forbidden
process. Thus it is possible that the pyrazine-DIPYR2 dye might be capable of independently
transporting long-lived high energy and low energy singlet excitons in the active layer of an
OPV, reducing both transmission and thermalization losses in a thoughtfully engineered device
which makes judicious use of both energy states.
The quinoline derivative pyrazine- -DIPYR2 was also modeled and found to have the
same symmetric and energetic situation as pyrazine-DIPYR2. Contrasting the -derivative
with pyrazine-DIPYR2, it can be seen that the quinolines provide an extra bathochromic shift
in both the S1 and S2 states (to 1.15 and 2.0 eV, respectively). In addition, the S1 transition is
more highly allowed in pyrazine- -DIPYR2 than it is in pyrazine-DIPYR2, indicating that it
Golden
90
may be a more highly absorbing NIR chromophore. Both the pyrazine-DIPYR2 and pyrazine-
-DIPYR2 structures are highly promising for application to organic photovoltaics.
Finally, a tetracene derivative, azo-DIPYR2 was modeled. This structure dispenses with
a central ring in order to generate a symmetric tetra-annular heterocyclic system, instead
making use of an azo bridge. The azo-DIPYR2 structure was found to be highly curved from
planarity and therefore symmetry arguments are more difficult to assign than in the planar
pentacene derivatives discussed above. This structure was found to have a S1 excitation energy
equal to 1.90 eV (652 nm) with a high oscillator strength (f = 0.69). The T1 energy (0.69 eV)
was determined by the delta self-consistent field method,
33
whereby the triplet geometry was
optimized and the free energy of the ground state geometry was subtracted from that of the
triplet.
Retrosynthetic strategies (Scheme 3.2) to access the pyrazine-DIPYR2 structures from
pyrazine stating materials were derived from the syntheses used to obtain the DIPYR monomer
ligands. As the 2,5-dimethylpyrazine and 2,5-dibromopyrazine materials are available for
Scheme 3.2. Synthetic strategies to access the pyrazineDIPYR 2 ligand from 2,5-symmetrically disubstituted
pyrazine starting materials

Molecular Design for Organic Photovoltaics
Chapter 3 | 91
purchase, Strategies I and III were attempted
first. Strategy I involves the double
deprotonation of 2,5-dimethylpyrazine with a
strong base (LDA and nBuLiwere both
screened). The dibasic intermediate is then
dripped slowly into a 2-fluoropyridine solution
in an attempt to induce SNAr at the 2-position
of pyridine. A high concentration of 2-
fluoropyridine relative to dimethylpyrazine is
required in order to prevent the formation of
branching products, as the intermediate 2-pyrazinyl-2-pyridylmethane is more basic at the
diarylmethane (proto-meso) position than at the benzylic position. Even after careful addition
and a screening of several reaction conditions, progress thus far has been limited to the hemi
pyrazine-DIPYR2 ligand, indicating that the 2,5-dimethylpyrazine dianion is either not formed
or is not stable enough to undergo the relatively slow SNAr reaction with 2-fluroropyridine.
Nevertheless, this ligand was borylated as a model for future borylation of the pyrazine-
DIPYR2 ligand systems, and the absorption and emission are depicted in Figure 3.7.
Strategy III was attempted with both 2-bromo pyridine and quinoline substrates. The
pyrazine-2,5-dimethylester and pyrazine-2,5-diacidchloride were each synthesized according
to literature procedures by oxidation of 2,5-dimethylpyrazine to the dicarboxylic acid, followed
by esterification or chlorination, respectively.
34
Lithiation of the 2-bromopyridyl substrates
was accomplished with either n-BuLi at -78 K or t-BuLi at room temperature in dry
tetrahydrofuran. Subsequent reaction with the either of the pyrazine-2,5-dicarboxyicl

300 400 500 600 700 800
0.0
0.2
0.4
0.6
0.8
1.0

Absorbance / Emission (a.u.)
Wavelength (nm)
Absorption
Emission

Figure 3.7. Procedure for the borylation of hemi
pyrazineDIPYR 2 and experimental absorption and
emission spectra.
Golden
92
derivatives resulted in the formation of no product; rather, protonation of the lithiopyridine
species was observed, indicating that the pyrazine-2,5-dicarboxylic species are not sufficiently
reactive toward nucleophilic attack.
Alternatives to these strategies are presented in Strategy II and Strategy IV, which are
directly analogous to Strategies I and III, respectively, with an inversion of the chosen
functional handles between heteroaromatic systems. Strategy II relies upon a 2,5-
difluoropyrazine starting material which is not yet known in the literature. Despite this
drawback, the substrate, if formed, has a good chance of providing the desired reactivity, as
pyrazine has a very similar charge density (+0.07) at its 2- and 5-positions as does pyridine at
its 2-position (+0.08), where the 2-fluoropyridine has been shown in the synthesis of DIPYR
to be highly amenable to nucleophilic attack by lithiopicoline.
Strategy IV makes use of a 2,5-dilithiopyrazine acquired from the lithiation of 2,5-
dibromopyrazine. This material may be used to perform a nucleophilic attack at a picolinic
ester or acid chloride to yield pyrazine-2,5-dipyridylketone, which is expected to be a solid
that can be readily purified by precipitation from the crude reaction mixture. Although this
procedure and Strategy III might facilitate purification, the added synthetic step required to
reduce the diketone to the ligand makes this process far less atom economical than Strategies
I or II. Design of appropriate syntheses of the pyrazine-DIPYR2 materials are currently
ongoing;
†
in addition to facilitating progress in OPVs, this work has the potential to facilitate
the synthesis of medically relevant structures, as many pyrazines are known to have
pharmaceutical activity in humans.
35

†
This work is being performed in collaboration with Konstantin Mallon in fulfillment of his master’s thesis
requirements in the Thompson Lab.
Molecular Design for Organic Photovoltaics
Chapter 3 | 93
3.5 Fused DIPYRs: littleDIPYR and bigDIPYR
Expansion of the -system in DIPYRs can also be affected by planarization of a meso-
coupled DIPYR. Triply fused dimeric systems such as those shown in the tetrapyridyl DIPYR
dimer littleDIPYR and the tetraquinolyl DIPYR dimer bigDIPYR delocalize the excited state
symmetrically and in two dimensions; this strategy has been shown to affect bathochromic
shifts by as much at 700 nm in triply fused porphyrins relative to the monomeric species.
36-40

The geometry optimized structures of littleDIPYR and bigDIPYR were calculated in the gas
phase at the B3LYP / 6-311G** level of theory and TD-DFT was used to determine the
excitation energies of both materials (Table 3.3). It was found that the S1 energy in littleDIPYR
corresponded to a low-oscillator strength (f = 0.08) NIR transition (1333 nm, 0.93 eV) after
correction (subtracting 0.44 eV from calculated singlet energies), corresponding to a very large

Figure 3.8. Molecular structures (top) of triply fused DIPYR dimers, littleDIPYR and DIPYR and (bottom)
geometry optimized structure of bigDIPYR shown from 3,3’-biquinoline edge (left), front (center), and BF 2-edge
(right). Edge views show significant twisting of the geometry optimized structure in the gas phase.

Golden
94
bathochromic shift of 1.64 eV relative to the monomer
DIPYR.
25
This value is red-shifted beyond the ideal bandgap
energy of 1.3 eV as estimated by the detailed-balance limit.
3, 19

Further, pyridines have very poor reactivity at the 3-positions,
which must be oxidized in order to form the triply fused
product. Reported synthesis for the 3,3’-coupling of quinolines
has been demonstrated in the literature, so the quinoline
derivative bigDIPYR was chosen as the target structure.
41

The geometry optimized structure of bigDIPYR (Figure
3.8) (D2h) shows significant twisting of the quinoline moieties
in the ground state. Orbital analysis, however, shows that the
frontier molecular orbital densities are confined primarily to the
octacyclic central core (Figure 3.9), with small contributions to
the LUMO and LUMO+1 on the outer quinoline rings.
The S1 state (1.04 eV, 1192 nm) in bigDIPYR is
bathochromically shifted by 1.33 eV (668 nm) compared to the
monomer -DIPYR. It is predominately a HOMO to LUMO
transition; however, the symmetries of these orbital sets are
both b1u, making this a symmetry-forbidden transition with a
correspondingly low oscillator strength (f = 0.01). The S0→S2 transition (1.19 eV), on the
other hand, is HOMO to LUMO+1, where the LUMO+1 is characterized by b1u symmetry, and
therefore symmetry allowed (f = 0.16). The symmetry of the S1 and S2 states are Ag and B2u,
respectively, making internal conversion from S1 to S2 a symmetry-allowed process and

LUMO+1
163 (-2.20) a 2g

LUMO
162 (-2.46 eV) b 1u

HOMO
161 (-4.27 eV) b 1u

Figure 3.9. Frontier molecular
orbital densities, energies, and
symmetries in bigDIPYR
Molecular Design for Organic Photovoltaics
Chapter 3 | 95
indicating that Kasha’s rule will
be obeyed in this system. The
corresponding energy loss
(0.15 eV) resulting from this
internal conversion process is
quite small compared to the -
DIPYR system discussed above
wherein the S0→S1 transition is also formally forbidden. Higher energy state transitions are
also predicted with high oscillator strengths in bigDIPYR (Table 3.3), indicating that this
structure represents a low-bandgap absorber with broad spectral sensitivity. Notably, there is
a gap in absorptivity predicted in the 460-650 nm regime wherein a judiciously selected
acceptor material could be utilized to affect complete spectral absorption in a heterojunction
OPV.
The symmetry relationship between the S0 and S1 states indicate that fluorescence will
be slow. The implications for the effect this will have on exciton migration in a bulk film are
not immediately obvious, so the successful synthesis of this structure will represent a unique
opportunity to probe novel energy transfer effects in a low-bandgap, highly absorptive
molecular material with a symmetry-forbidden S1-S0 transition. The comparative rate between
energy transfer and non-radiative decay will be key to determining whether this material will
function as desired in an OPV. Non-radiative decay by intersystem crossing is forbidden by
both symmetry and by spin, so will likely not compete with either fluorescence or vibronic
relaxation processes.
Table 3.3. Orbital contributions, energies, and oscillator strengths
of major S 0 → S n (corrected) and S 0 → T n transitions in
bigDIPYR.
bigDIPYR
state transition %  (nm, eV) ƒ sym.
S 1 161 → 162 100 1192, 1.04 0.01 A g
S 2 161 → 163 0.99 1042, 1.19 0.16 B 2u
S 3 161 → 164 .99 742, 1.67 0.44 B 1u
S 4 161 → 165 .99 656, 1.89 0.00 A g
S 5 160 → 162 98 455, 2.72 0.14 B 2g
T 1 161 → 162 100 1436, 0.86 0.00 A g
T 2 161 → 163 0.96 1066, 1.16 0.00 B 2u

Golden
96
As in the above sections, this work is ongoing at the time of submission. The meso-
coupling of both DIPYR and -DIPYR to form the orthogonal dimeric species is described in
Chapter 4. Although both the 3,3’-biquinoline and the 2,2’,2’’,2’’’-tetraquinolylethane (meso
dimer) structures have each been synthesized as part of this research effort, affecting both of
these transformations on the same substrate has proven challenging. In either case, the first
bond formation (either beginning with the meso-coupling or beginning with the 3,3’-coupling)
limits the reactivity required for the second bond formation. However, pyridine, pyrazine,
quinoline, and isoquinoline chemistries are actively under development in the literature due to
favorable bioactivity from derivatives of these heteroaromatic structures. The combination of
novel chemistries and an increase in available substrates will prove key to the successful
synthesis of these exciting NIR-absorbing chromophores.
3.6 Experimental Methods
2,2’-diquinolylketoxime: 2,2’-diquinolylketone (0.500 g, 1.76 mmol) was dissolved with
hydroxylamine hydrochloride (0.183 g, 2.64 mmol) in 1.5 mL ethanol. Sodium hydroxide
(0.347 g, 8.68 mmol) was slowly added over 90 minutes. The paste turned from off-white to
bright yellow in color upon continued addition of sodium hydroxide. Additional ethanol
(4 mL) was added to facilitate stirring, and the solution was brought to reflux for 15 minutes.
After this time, the reaction was cooled to room temperature, suspended in water, and
neutralized with concentrated hydrochloric acid (0.75 mL), forming a yellow precipitate. The
solvent was removed in vacuo and the solid residue was treated with 5 mL saturated aqueous
sodium carbonate solution and filtered. The filter cake was washed with water and dried in
vacuo to produce the ketone oxime (0.459 g, 87% yield).
1
H NMR (400 MHz, CDCl3) δ 8.36
(br. d), 8.16 (br. d) 7.92 (br. d), 7.90 (br. t), 7.85, 7.68 (br. t), 7.67 (br. d), 7.57 (br. d). N.B.
Molecular Design for Organic Photovoltaics
Chapter 3 | 97
The oxime has poor solubility in chloroform, two isomeric forms, and an acidic -OH group
that make characterization by NMR in chloroform difficult. The oxime may be soluble in
DMSO, which would allow the resolution of both isomeric forms.
2,2’-diquinolylmethanamine: 2,2’-diquinolylketoxime (0.40 g, 1.34 mmol), ammonium
acetate (0.227 g, 2.94 mmol), and concentrated aqueous ammonia (1.8 mL, 68 mmol) were
dissolved in a 1:1 mixture of ethanol and water (2.4 mL). The slurry was heated to reflux and
zinc powder was added over two hours, causing a color change from light yellow to dark green.
It was heated a further three hours, forming a solid which was removed by filtration. Two
equivalents of sodium hydroxide were added to the filtrate and the solution was extracted three
times with diethyl ether. The organic layers were combined and the solvent was removed via
rotary evaporation, forming a green solid. The solid was characterized by LCMS as the crude
product. It did not run on silica gel and was used without further purification.
3.7 Bibliography
1. Guan, Z.; Li, H.-W.; Cheng, Y.; Yang, Q.; Lo, M.-F.; Ng, T.-W.; Tsang, S.-W.; Lee,
C.-S., Charge-Transfer State Energy and Its Relationship with Open-Circuit Voltage in an
Organic Photovoltaic Device. J. Phys. Chem. C 2016, 120 (26), 14059-14068.
2. Vandewal, K.; Tvingstedt, K.; Manca, J. V.; Inganäs, O., Charge-Transfer States and
Upper Limit of the Open-Circuit Voltage in Polymer:Fullerene Organic Solar Cells. IEEE J.
Sel. Top. Quantum Electron. 2010, 16 (6), 1676-1684.
3. Gruber, M.; Wagner, J.; Klein, K.; Hörmann, U.; Opitz, A.; Stutzmann, M.; Brütting,
W., Thermodynamic Efficiency Limit of Molecular Donor‐Acceptor Solar Cells and its
Application to Diindenoperylene/C60‐Based Planar Heterojunction Devices. Adv. Energy
Mater. 2012, 2 (9), 1100-1108.
4. Schlenker, C. W.; Thompson, M. E., The molecular nature of photovoltage losses in
organic solar cells. Chem. Commun. 2011, 47 (13), 3702-3716.
5. Janssen René, A. J.; Nelson, J., Factors Limiting Device Efficiency in Organic
Photovoltaics. Adv. Mater. 2012, 25 (13), 1847-1858.
6. Giebink, N. C.; Wiederrecht, G. P.; Wasielewski, M. R.; Forrest, S. R., Thermodynamic
efficiency limit of excitonic solar cells. Phys. Rev. B 2011, 83 (19), 195326.
7. Fabian, J.; Nakazumi, H.; Matsuoka, M., Near-infrared absorbing dyes. Chem. Rev.
1992, 92 (6), 1197-1226.
Golden
98
8. Quan, L.; Liu, S.; Sun, T.; Guan, X.; Lin, W.; Xie, Z.; Huang, Y.; Wang, Y.; Jing, X.,
Near-Infrared Emitting Fluorescent BODIPY Nanovesicles for in Vivo Molecular Imaging and
Drug Delivery. ACS Appl. Mater. Interfaces 2014, 6 (18), 16166-16173.
9. Zhang, H.; Jenatsch, S.; De Jonghe, J.; Nüesch, F.; Steim, R.; Véron, A. C.; Hany, R.,
Transparent Organic Photodetector using a Near-Infrared Absorbing Cyanine Dye. Sci. Rep.
2015, 5, 9439.
10. Sreejith, S.; Joseph, J.; Lin, M.; Menon, N. V.; Borah, P.; Ng, H. J.; Loong, Y. X.;
Kang, Y.; Yu, S. W.-K.; Zhao, Y., Near-Infrared Squaraine Dye Encapsulated Micelles for in
Vivo Fluorescence and Photoacoustic Bimodal Imaging. ACS Nano 2015, 9 (6), 5695-5704.
11. Carr, J. A.; Franke, D.; Caram, J. R.; Perkinson, C. F.; Saif, M.; Askoxylakis, V.; Datta,
M.; Fukumura, D.; Jain, R. K.; Bawendi, M. G.; Bruns, O. T., Shortwave infrared fluorescence
imaging with the clinically approved near-infrared dye indocyanine green. Proc. Natl. Acad.
Sci. 2018.
12. Ni, Y.; Wu, J., Far-red and near infrared BODIPY dyes: synthesis and applications for
fluorescent pH probes and bio-imaging. Org. Biomol. Chem. 2014, 12 (23), 3774-3791.
13. Thompson, N. J.; Wilson, M. W. B.; Congreve, D. N.; Brown, P. R.; Scherer, J. M.;
Bischof, Thomas S.; Wu, M.; Geva, N.; Welborn, M.; Voorhis, T. V.; Bulović, V.; Bawendi,
M. G.; Baldo, Marc A., Energy harvesting of non-emissive triplet excitons in tetracene by
emissive PbS nanocrystals. Nat Mater 2014, 13 (11), 1039-1043.
14. Saccone, D.; Galliano, S.; Barbero, N.; Quagliotto, P.; Viscardi, G.; Barolo, C.,
Polymethine Dyes in Hybrid Photovoltaics: Structure–Properties Relationships. Eur. J. Org.
Chem. 2016, 2016 (13), 2244-2259.
15. Lunt, R. R.; Bulovic, V., Transparent, near-infrared organic photovoltaic solar cells for
window and energy-scavenging applications. Appl. Phys. Lett. 2011, 98 (11), 113305.
16. Sargent, E. H., Infrared photovoltaics made by solution processing. Nat. Photon. 2009,
3, 325.
17. Bricks, J. L.; Kachkovskii, A. D.; Slominskii, Y. L.; Gerasov, A. O.; Popov, S. V.,
Molecular design of near infrared polymethine dyes: A review. Dyes Pigm. 2015, 121, 238-
255.
18. Jiao, L.; Wu, Y.; Wang, S.; Hu, X.; Zhang, P.; Yu, C.; Cong, K.; Meng, Q.; Hao, E.;
Vicente, M. G. H., Accessing Near-Infrared-Absorbing BF2-Azadipyrromethenes via a Push–
Pull Effect. J. Org. Chem. 2014, 79 (4), 1830-1835.
19. Shockley, W.; Queisser, H. J., Detailed Balance Limit of Efficiency of p‐n Junction
Solar Cells. J. Appl. Phys. 1961, 32 (3), 510-519.
20. Chen, Y.; Wan, X.; Long, G., High Performance Photovoltaic Applications Using
Solution-Processed Small Molecules. Acc. Chem. Res. 2013, 46 (11), 2645-2655.
21. Qian, G.; Dai, B.; Luo, M.; Yu, D.; Zhan, J.; Zhang, Z.; Ma, D.; Wang, Z. Y., Band
Gap Tunable, Donor−Acceptor−Donor Charge-Transfer Heteroquinoid-Based Chromophores:
Near Infrared Photoluminescence and Electroluminescence. Chem. Mater. 2008, 20 (19),
6208-6216.
22. Chen, J. J.; Conron, S. M.; Erwin, P.; Dimitriou, M.; McAlahney, K.; Thompson, M.
E., High-Efficiency BODIPY-Based Organic Photovoltaics. ACS Appl. Mater. Interfaces
2015, 7 (1), 662-669.
23. Erten-Ela, S.; Yilmaz, M. D.; Icli, B.; Dede, Y.; Icli, S.; Akkaya, E. U., A Panchromatic
Boradiazaindacene (BODIPY) Sensitizer for Dye-Sensitized Solar Cells. Org. Lett. 2008, 10
(15), 3299-3302.
Molecular Design for Organic Photovoltaics
Chapter 3 | 99
24. Ni, Y.; Zeng, W.; Huang, K.-W.; Wu, J., Benzene-fused BODIPYs: synthesis and the
impact of fusion mode. Chem. Commun. 2013, 49 (12), 1217-1219.
25. Golden, J. H.; Facendola, J. W.; Sylvinson M. R, D.; Baez, C. Q.; Djurovich, P. I.;
Thompson, M. E., Boron Dipyridylmethene (DIPYR) Dyes: Shedding Light on Pyridine-
Based Chromophores. J. Org. Chem. 2017, 82 (14), 7215-7222.
26. Burns, C. T.; Jordan, R. F., Ethylene Polymerization by Sterically Crowded
Palladium(II) Complexes that Contain Bis(heterocycle)methane Ligands. Organometallics
2007, 26 (27), 6737-6749.
27. Chang, J.; Plummer, S.; Berman, E. S. F.; Striplin, D.; Blauch, D., Synthesis and
Characterization of Bis(di-2-pyridylmethanamine)ruthenium(II). Inorg. Chem. 2004, 43 (5),
1735-1742.
28. Biellmann, J., Ducep, J., Allylic and Benzylic Carbanions Substituted by Heteroatoms.
In Organic Reactions, John Wiley and Sons, Inc.: 1982.
29. Gebert, A.; Jasiński, M.; Mlostoń, G.; Heimgartner, H., Reduction of Thiocarbonyl
Compounds with Lithium Diisopropylamide. Helv. Chim. Acta 2014, 97 (7), 931-938.
30. Hojo, M.; Masuda, R.; Okada, E.; Sakaguchi, S.; Narumiya, H.; Morimoto, K., O-N,
S-N and N-N exchange reactions at olefinic carbon atoms: Facile synthetic method for β-
trifluoroacetylvinylamines. Tetrahedron Lett. 1989, 30 (45), 6173-6176.
31. Eiden, F.; Meinel, F., Über 2,3‐Dihydro‐1,4‐benzothiazin‐3‐one. Arch. Pharm. 1979,
312 (4), 302-312.
32. Hanson, K.; Roskop, L.; Djurovich, P. I.; Zahariev, F.; Gordon, M. S.; Thompson, M.
E., A Paradigm for Blue- or Red-Shifted Absorption of Small Molecules Depending on the
Site of π-Extension. J. Am. Chem. Soc. 2010, 132 (45), 16247-16255.
33. Gavnholt, J.; Olsen, T.; Engelund, M.; Schiøtz, J., Delta self-consistent field method to
obtain potential energy surfaces of excited molecules on surfaces. Phys. Rev. B 2008, 78 (7),
075441.
34. Coufal, R.; Prusková, M.; Císařová, I.; Drahoňovský, D.; Vohlídal, J., Simple and
efficient access to pyrazine-2,5- and -2,6-dicarbaldehydes. Synth. Commun. 2016, 46 (4), 348-
354.
35. Dolezal, M.; Zitko, J., Pyrazine derivatives: a patent review (June 2012 – present).
Expert Opin. Ther. Pat. 2015, 25 (1), 33-47.
36. Diev, V. V.; Hanson, K.; Zimmerman, J. D.; Forrest, S. R.; Thompson, M. E., Fused
Pyrene-Diporphyrins: Shifting Near-Infrared Absorption to 1.5 mu m and Beyond. Angew.
Chem., Int. Ed. 2010, 49 (32), 5523-5526.
37. Zimmerman, J. D.; Diev, V. V.; Hanson, K.; Lunt, R. R.; Yu, E. K.; Thompson, M. E.;
Forrest, S. R., Porphyrin-Tape/C-60 Organic Photodetectors with 6.5% External Quantum
Efficiency in the Near Infrared. Adv. Mater. 2010, 22 (25), 2780-+.
38. Zimmerman, J. D.; Yu, E. K.; Diev, V. V.; Hanson, K.; Thompson, M. E.; Forrest, S.
R., Use of Additives in Porphyrin-Tape/C60 Near-Infrared Photodetectors. Org. Electron.
2011, 12, 869.
39. Cho, H. S.; Jeong, D. H.; Cho, S.; Kim, D.; Matsuzaki, Y.; Tanaka, K.; Tsuda, A.;
Osuka, A., Photophysical Properties of Porphyrin Tapes. J. Am. Chem. Soc. 2002, 124 (49),
14642-14654.
40. Tsuda, A.; Osuka, A., Fully Conjugated Porphyrin Tapes with Electronic Absorption
Bands That Reach into Infrared. Science 2001, 293 (5527), 79-82.
Golden
100
41. Nelson, T. D., Crouch, R. D., Cu, Ni, and Pd Mediated Homocoupling Reactions in
Biaryl Syntheses: The Ullmann Reaction. In Organic Reactions, 2004.
Molecular Design for Organic Photovoltaics
Chapter 4 | 101
Chapter 4

Symmetry-Breaking Charge Transfer in Boron
Dipyridylmethene Dimers

4.1 Introduction to SBCT in DIPYR Dimers
A potential solution to mitigate photovoltage losses in organic photovoltaics involves
judicious use of the symmetry-breaking charge transfer (SBCT) materials in the active layer.
SBCT complexes can be applied to OPVs either as donor materials, as acceptor materials, or
both, in order to decrease the energy loss intrinsic to forming charge separated species from
excitonic materials. The SBCT phenomenon itself can be generally defined as an excited state
process wherein a locally excited state in a symmetrical dimer undergoes electron transfer
between two electronically degenerate states, producing a desymmetrized CT state. This
process can occur either intermolecularly, as is the case in natural photosynthetic systems, or
intramolecularly, as has been studied in an array of chromophoric dyads. Materials capable of
SBCT have been applied to artificial photosynthesis,
1-2
photovoltaics,
3-4
and photonics
5
where
rapid formation of photo-oxidized and photo-reduced products are necessary. The rate of
SBCT varies depending on the strength of chromophoric coupling and the solvent
environment.
2, 6
In dyads with weak electronic coupling between adjacent chromophores, such
as orthogonal or spatially separated acene,
7-9
perylene,
10-12
and BODIPY dimers,
13
as well as
homoleptic metallodipyrrins,
3, 14
SBCT occurs on the order of hundreds of femtoseconds to
tens of picoseconds in polar solvents, and is generally disfavored in nonpolar media. In the
opposite case, if the ground-state geometry involves significant wavefunction overlap between
Golden
102
chromophores (i.e. if there is strong electronic coupling), exciton coupling predominates and
the relaxed excited state lacks CT character.
15
Significant population of the SBCT state in
nonpolar media is exceptional, only occurring in systems where the electronic coupling
between chromophores is strong enough to induce ultrafast charge transfer but weak enough
to stabilize a symmetry-broken state.
16

Chapter 2 of this work describes the photophysical behavior of dipyridylmethene
(DIPYR) monomers.
17
This class of intensely fluorescent (PLQY = 0.2 – 0.8) pyridine-based
chromophores is structurally analogous to both BODIPYs and acenes, the dimers of which are
known to form symmetry-broken charge transfer states. In this chapter, meso-linked DIPYR
dimers (bis-DIPYRs), are described. This work represents the first study of DIPYR dimers and
has been published in the literature.
18
The rate of formation of the SBCT state in polar and
nonpolar solvents and its decay pathways are probed by a combination of steady state
spectroscopies as well as femtosecond- and nanosecond transient absorption spectroscopies.
4.2 Synthesis and Structural Properties of bis-DIPYRs
The DIPYR dimers bis-DIPYR and bis- -DIPYR were prepared by a modified two-
step procedure wherein the diheteroarylmethane ligand was coupled via direct C-C bond
coupling using copper (II) acetate as a sacrificial oxidant, and the resulting
tetraheteroarylethane ligand was borylated by the same method developed for synthesis of the
monomers (Scheme 1). Both tetraheteroarylethane ligands were previously reported in the
literature, but it appears that this work represents the first preparation of tetraquinolylethane
by this method.
19-20
The synthesis of bis-DIPYR from tetrapyridylethane is fairly tolerant of
impurities in the ligand, whereas the synthesis of bis- -DIPYR from tetraisoqiunolylethane is
Molecular Design for Organic Photovoltaics
Chapter 4 | 103
not. Achieving a yield over 30% for bis- -DIPYR requires multiple recrystallizations of the
tetraquinolylethane ligand. In cases where the starting material is not pure, the monoborylated
bis- -DIPYR predominates. The precise reason for this is unclear; modification of the
procedure by using chlorobenzene as solvent, refluxing at higher temperatures, refluxing for
longer durations and/or adding extra equivalents of BF3OEt2 resulted in either no reaction,
decomposition, or further production of the monoborylated product.
The monoborylated bis- -DIPYR and the fully borylated bis- -DIPYR tend to be well-
resolved on silica gel TLC as a yellow spot and a red spot, respectively, which run near the
solvent front. However, the two materials tend to co-elute in a column. They can be best
resolved by first eluting the monoborylated product, then eluting bis- -DIPYR using a gradient
solvent system from 30% dichloromethane in hexanes to 50% dichloromethane in hexanes.

Scheme 4.1. (a) 2 eq. Cu(OAc) 2·H 2O, dimethylacetamide, 120 ˚C, 24 h. (b) 4 eq. BF 3OEt 2, 1,2-dichloroethane,
reflux 2h, N,N-diisopropylethylamine

Golden
104
The dimers were analyzed
by single-crystal X-ray
diffraction using crystals grown
from the slow diffusion of
hexanes into a saturated
dichloromethane solution
(Figure 4.1). Bis-DIPYR
crystallized readily into P 21/c
crystals with four bis-DIPYR
molecules comprising the unit
cell, whereas bis- -DIPYR crystallized into small red needles packed into a C 2/c space group
with n-hexane and two bis- -DIPYR molecules incorporated into the asymmetric unit (Figure
4.2). The measured structure of bis- -DIPYR is inherently more disordered due to the presence
of hexane in the crystal structure, resulting in a poor R-factor of 0.12. Thus, while the structure
of bis-DIPYR gives bond distances and angles with low estimated standard deviation values,
the structure of bis- -DIPYR can be used to verify connectivity and gross structure, but the
bond lengths and angles should not be considered accurate.
Bis-DIPYR exhibits greater structural distortion of the DIPYR chromophoric units than
is observed in the crystal structure of the monomer. This quality was quantified by measuring
the angle between the planes formed by the carbons in each of the two flanking pyridine rings
of the DIPYR moiety. In bis-DIPYR, the angle between pyridyl planes, which would be 0° in
a perfectly flat chromophoric unit, is 21°; in the monomer DIPYR, this value is 4.7°,
considerably closer to the ideal. In contrast, the structure of bis- -DIPYR shows an angle

Figure 4.1. ORTEP structures of bis-DIPYR (left) and bis- -DIPYR
(right) with hydrogen atoms displayed as white spheres with atomic
radius set to 0.15 A, and thermal ellipsoids for carbon (grey),
nitrogen (blue), boron (pink), and fluorine (yellow) atoms set to 50%
probability. Note that bis- -DIPYR packs into two distinct
conformers in the asymmetric unit. The conformer depicted is the
one packed closest to hexane.
Molecular Design for Organic Photovoltaics
Chapter 4 | 105
between the pyridyl planes of 13º, close to the value of 7º seen for -DIPYR. The extra
puckering in bis-DIPYR is likely due to intermolecular crystal packing forces in the dimer,
which exhibits considerable -  stacking between adjacent pyridine rings. The monomer,
which packs in a herringbone fashion, does not experience the same degree of steric repulsion
as is observed in the crystal structures of the dimer.
17
Of considerable importance to the
photophysical characterization of dilute solutions is the degree wavefunction overlap between
coupled chromophores. The dihedral angles between coupled chromophores in both bis-
DIPYR dimers, measured as the angle between the planes defined by the meso carbons and the
two flanking carbons on each chromophoric unit, are 89.9  and 86° in bis-DIPYR and bis- -

Figure 4.2. Crystal packing structures of bis-DIPYR (a) and bis- -DIPYR (c). Note that bis- -DIPYR packs
with hexanes in the asymmetric unit (b).

a)
b)
c)
Golden
106
DIPYR, respectively. Meso-bound dimers with orthogonal DIPYR units are expected to have
minimal wavefunction overlap and thus negligible excitonic coupling between the DIPYR
units.
4.3 Steady-State Photophysical Properties
The absorption and emission of the dimers in a range of solvents of varying polarity
are shown in Figure 4.3 and the photophysical properties are summarized in Table 4.1. A
bathochromic shift in the absorption maxima of bis-DIPYR and bis- -DIPYR relative to their
respective monomer models is observed (Δṽ = 465 and 815 cm
-1
in methylcyclohexane),
possibly due to an acceptor effect exerted by
the polar BF2 group para to the meso-bridge.
The molar absorptivities (   of bis-DIPYR
and bis- -DIPYR at max in
methylcyclohexane are 4.1x10
4
and 2.6x10
4

L·mol
-1
·cm
-1
while the molar absorptivities
of their respective monomers are 2.9x10
4

and 1.1x10
5
. The absorption spectra of the
dimers are broadened relative to the
monomers (FWHM = 3148 and 844 cm
-1
for
bis-DIPYR and bis- -DIPYR, compared to
2057 and 222 cm
-1
in the monomers).
17
Even
when integrating over the entire molar
absorptive cross-section of bis- -DIPYR,
300 400 500 600 700
0.0
0.2
0.4
0.6
0.8
1.0

Absorbance/Emission (a.u.)
Wavelength (nm)
MeCyHex
Toluene
THF
DCM
MeCN
bis-DIPYR
300 400 500 600 700
0.0
0.2
0.4
0.6
0.8
1.0

Absorbance/Emission (a.u.)
Wavelength (nm)
MeCyHex
Toluene
THF
DCM
MeCN
bis- -DIPYR

Figure 4.3. UV-vis absorption (solid) and
fluorescence emission (dashed) spectra of DIPYR
dimers in various solvents.
Molecular Design for Organic Photovoltaics
Chapter 4 | 107
the dimer is shown to be considerably less absorbing per chromophoric unit than its monomeric
analogue -DIPYR. This comparison is non-trivial for bis-DIPYR where there are overlapping
state transitions. The reason for the decreased absorptivity is bis- -DIPYR is not clear and
merits further investigation.
While the absorption spectrum for DIPYR is solvent dependent, that of bis-DIPYR is
not; instead, it has a similar lineshape in all solvents to the absorption spectra of DIPYR in
polar media. The reason for this discrepancy is rooted in the parentage of the lowest energy
absorption bands. The visible absorption spectra of both bis-DIPYR and its parent monomer
DIPYR are characterized by two nearly degenerate transitions: S 0 S1 and S0 S2.
17
In
DIPYR, the lowest energy absorption maximum (S0-S1) corresponds to an electronic transition
along the long-axis of the DIPYR chromophore, which contains no net permanent dipole in
the ground state. The S0 S2 electronic transition, which is approximately degenerate with the
S0 v=0  S1 v=1 transition, occurs along the short, permanent dipole-containing axis of the
chromophore and is therefore enhanced in polar media. In the dimer, which is symmetric about
both axes, the S0-S2 transition involves no net permanent dipole and therefore is not affected
by a change in solvent polarity. Consequently, the S0-S2 transition in the dimer is equally
stabilized in polar and nonpolar solvents; i.e. the relative population of the S1 and S2 states are
not changed by varying solvent polarity in bis-DIPYR. The result is an absorption spectrum
wherein the first- and second-lowest energy maxima do not change in intensity with a change
in solvent polarity, which contrasts with the monomer DIPYR which exhibits significant
changes in the shape of the absorption spectrum with changes in solvent.
The emission spectra of both bis-DIPYR and bis- -DIPYR are broad and featureless,
exhibiting large Stokes shifts (1290 cm
-1
and 887 cm
-1
, respectively). This is in contrast with
Golden
108
the monomeric species, which have structured emission spectra and almost no Stokes shift
(130 and 40 cm
-1
).
17
It is instructive to note that the monoborylated bis- -DIPYR exhibits
absorption and emission features which are nearly identical with the monomer -DIPYR,
indicating that the broadening and bathochromic shifts of absorption and emission spectra in
the dimers is not simply correlated with the substitution of a C-H bond at the meso position
with a C-C bond. Rather, it appears that the source of emission broadening is related to the
dimerization of the chromophores, indicating either that there is significant distortion of the S1
excited state relative to the ground state or that emission occurs from another state, the nature
of which could be either a symmetry-broken state or an exciton coupled state, the latter of
which would require twisting in the excited state while the former would involve a geometry
in which the chromophores remain mutually orthogonal.
In both dimers, the photoluminescent quantum yields (  PL) and radiative rates (kr)
decrease commensurate with an increase in the non-radiative decay rate (knr) as the solvent
polarity is increased (Table 4.1). The radiative rates decrease linearly with solvent polarity, as
estimated by the ET(30) solvent polarity index
21
(Figure 4.4). However, plotting radiative
Table 4.1. Solvent dependent steady-state photophysical properties of DIPYR dimers

Solvent E T(30) absorption
 max (nm)
emission
 max (nm)
 PL  (ns) k r (10
8
s
-1
) k nr (10
8
s
-1
)
bis-DIPYR MeCyHex 30.9 492 523 0.23 3.1 0.75 2.5
Toluene 33.9 494 528 0.17 2.8 0.61 3.0

THF 37.4 492 532 0.12 2.5 0.49 3.5
DCM 40.7 492 530 0.08 1.9 0.42 4.8
MeCN 45.6 489 537 0.03 0.96 0.31 10.
Polystyrene - 492 524 0.16 2.3 0.70 3.6
bis- -DIPYR MeCyHex 30.9 543 570 0.57 4.6 1.2 0.97
Toluene 33.9 543 576 0.45 4.0 1.1 1.4
THF 37.4 541 573 0.44 4.5 0.98 1.2
DCM 40.7 541 574 0.37 3.9 0.95 1.6
MeCN 45.6 537 572 0.31 4.1 0.78 1.7
Polystyrene - 541 567 0.47 4.8 0.98 1.1

Molecular Design for Organic Photovoltaics
Chapter 4 | 109
rates versus dielectric constant yields a
poor fit, indicating that solvation power
(i.e. the specific solvation sphere of the
excited state) has a greater influence on the
emissive state than does the bulk dielectric
of the solvent.
21-23
Importantly, the
linearity of the trend with solvent polarity,
coupled with the observation that there is
no significant change in emission
lineshape, suggests that the same state is responsible for emission in all solvents.
As mentioned above, there are three probable parentages of the emissive state. In case
one, emission occurs from a locally excited state wherein there is significant distortion relative
to the ground state. In the second case, the dihedral angle is reduced and increased
wavefunction overlap leads to excitonic coupling. In the last case, the excited state geometry
remains the same as in the ground state, where the chromophores comprising the dimer are
mutually orthogonal, and charge (either a full charge or partial charge) is transferred to the
adjacent chromophore, inducing a symmetry-broken state. In order to differentiate between
these three possibilities, dilute thin films (1% in polystyrene) of both dimers were prepared
from toluene solution on quartz to examine the effect of rigidification on excited state decay
dynamics. The absorption features of the thin films show no significant differences from
absorption in solution. Likewise, the emission spectra of the films are very similar to emission
from methylcyclohexane solution (Figure 4.5), showing a slight narrowing at the red edge of
the emission band and no significant difference in the emission onset or maximum. These data
30 35 40 45
0.0
0.2
0.4
0.6
0.8
1.0
1.2

MeCyHex
Toluene
THF
DCM
MeCN
Radiative Rates (10
8
S
-1
)
Solvent Polarity Index

Figure 4.4. The relationship between radiative rate and
solvent polarity index in both bis-DIPYR (black) bis- -
DIPYR (red) is linear, which is characteristic of
emission from excited states with large dipole moments.

Golden
110
suggest that the geometry of the relaxed
excited state is similar to that of the ground
state, i.e. that the chromophoric units remain
orthogonal in the emissive state, providing
evidence that neither case one nor case two,
which both hinge on significant distortion in
the excited state, result in the observed
emission.
In frozen solution
(methylcyclohexane at 77 K), the
rigidochromic effect is enhanced even
further than in the polymer films. In this experiment, the emission spectra of the dimers are
narrowed (Figure 4.5) and display discernable vibronic features which are nevertheless
significantly broader than are observed in the monomer models. The presence of vibronic
features at 77 K which are not observed at room temperature indicate that the emissive state is
slightly distorted relative to the ground state, though the polymer film study above suggests
that this distortion likely involves minor changes in bond lengths rather than any significant
change in bond angles or in the magnitude of the dihedral angle. The fluorescence lifetimes at
77 K (3.25 ns for bis-DIPYR, 4.44 ns for bis- -DIPYR), taken to be equal to the radiative
lifetimes, are much faster than the radiative lifetimes in methylcyclohexane at room
temperature (13 ns and 8 ns, respectively), suggesting that distortion in the excited state
reduces the radiative rate, which is supported by the observation of decreasing radiative rate
with increasing solvent polarity.
0.5
1.0
300 400 500 600 700
300 400 500 600 700
0.0
0.5
1.0

bis- -DIPYR

bisDIPYR
Wavelength (nm)

Figure 4.5. DIPYR dimer absorption (dashed) and
emission (solid) in methylcyclohexane at room
temperature (black) and 77 K (blue) and in a dilute
polystyrene thin film (red).

Molecular Design for Organic Photovoltaics
Chapter 4 | 111
Interestingly, new bands at 593 nm and 657 nm in the 77 K emission spectrum of bis-
DIPYR and bis- -DIPYR, respectively, are observed. Gated detection enhances these new
bands relative to fluorescence features, and they are thus assigned as phosphorescence
emission from the T1 state. Phosphorescence of the same magnitude is observed in the DIPYR
monomer under similarly mild conditions (77 K in methylcyclohexane), as the triplet is readily
accessed in DIPYR via intersystem crossing from the S1 to the T2 state. However,
phosphorescence is not readily observed in -DIPYR; in order to discern the phosphorescence
features in the monomer, it was necessary to treat with methyl iodide to facilitate intersystem
crossing. Even then, phosphorescence was not discernable from the fluorescence features and
the more rigorous method of gated detection was required to resolve it. It was therefore
surprising to observe phosphorescence from bis- -DIPYR at 15% of the fluorescence intensity
at 77 K in methylcyclohexane without the need for an external heavy atom or for gated
detection. Intersystem crossing (ISC) from the S1 to the T1 state in bis- -DIPYR is both a spin-
and a symmetry-forbidden transition, and ISC to the T2 is intensely endothermic. While the
observation of a phosphorescence band in bis-DIPYR is expected due to the juxtaposition of
state energies facilitating ISC, the observation in bis- -DIPYR is striking and suggests the
presence of an intermediate state at or below the energy of the S1 state which facilitates ISC to
the T1. In light of the previous steady-state experiments, this suggests that the emissive species
belongs to the third possible case: a symmetry-broken state which is formed rapidly upon
photoexcitation. Direct evidence for the existence of such a state is presented in the following
sections.
Golden
112
4.4 Electrochemical Properties and Spectroelectrochemistry
The electrochemical properties of the dimers were explored using a combination of
cyclic voltammetry (CV) and differential pulse voltammetry (DPV).
*
Electrochemistry was
performed in acetonitrile solutions using a three-electrode set-up consisting of a polished 3 mm
glassy carbon working electrode, a Pt wire auxiliary electrode and a Ag/AgCl pseudo-reference
electrode. The CV of bis- -DIPYR (Figure 4.6) shows two fully reversible, one-electron

*
These experiments were performed at the University of California, San Diego, with Tyler Porter in the Kubiak
lab. Many thanks are owed to Tyler for his patience while instructing me in the intricacies of electrochemical and
spectroelectrochemical characterization methods.
1 0 -1 -2
-80
-40
0
40

Current (uA)
Potential (V)
bis-DIPYR

1 0 -1 -2
-40
0
40

Current (uA)
Potential (V)
Oxidations
Reductions
Oxidation DMFc
Reductions DMFc

1 0 -1 -2
-8
-4
0
4

Current (uA)
Potential (V)
bis- -DIPYR
1 0 -1 -2
-8
-4
0
4
8

Current (uA)
Potential (V)
Ox vs Fc
Red vs Fc

Figure 4.6. Cyclic voltammograms at 100 mV/s (left) and differential pulse voltammograms (right) of bis-
DIPYR (top) and bis- -DIPYR (bottom) in acetonitrile with 0.1 M tetrabutylammonium
hexafluorophosphate as electrolyte. The DPV of bis-DIPYR shows the spectra with and without
deccamethylferrocene, and that of bis-a-DIPYR is represented with ferrocene.

Molecular Design for Organic Photovoltaics
Chapter 4 | 113
reductions at -1.81 and -2.02 V vs Fc
+/0
corresponding to the reduction of each chromophore
in the dimer. Upon sweeping to oxidative potentials, a quasi-reversible oxidation is observed
at 0.45 V vs Fc
+/0
that becomes more reversible at faster scan rates. The CV of bis-DIPYR is
characterized by an irreversible first reduction and a quasi-reversible second reduction at 2.31
and 2.55 V vs Fc
+/0
respectively. Analogous to bis- -DIPYR, bis-DIPYR displays an
irreversible oxidation at 0.19 V vs Fc
+/0
that gains reversibility at faster scan rates. The CVs of
both bis-DIPYR and bis- -DIPYR display a third set of very broad, shallow reduction waves
centered at 0.52 V and 0.83 V vs Fc
+/0
, respectively, which grow in as a function of the number
of scans past the oxidation potential. These features are assigned to the reduction of a
decomposition product formed by the partially irreversible oxidation of bis-DIPYRs.
Differential pulse voltammetry was used to obtain the redox potentials enumerated in
Table 4.2, and these values were used to estimate the HOMO and LUMO energies of both
dimers.
24-25

The redox gap of bis-DIPYR is only 10 mV smaller than that of its parent monomer,
DIPYR, consistent with the small relative bathochromic shift observed in the absorption
spectrum of the dimer. A larger bathochromic shift is observed for the absorption of bis- -
DIPYR relative to the monomer -DIPYR, correlating well with a 100 mV decrease in the
redox gap between the two. Notably, the observed S1 absorption energies of both dimers
Table 4.2. Red./Ox. potentials and corresponding calculated HOMO/LUMO levels. Electrochemical values are
in V vs. ferrocene and HOMO/LUMO values are in eV.

Ox. Red. 1 Red. 2 (E
ox
– E
red
)

HOMO LUMO
DIPYR 0.14 -2.32 - 2.46 -4.80 -2.01
-DIPYR 0.40 -1.95 - 2.35 -5.16 -2.46
bis-DIPYR 0.16 -2.29 -2.56 2.45 -4.82 -2.05
bis- -DIPYR 0.45 -1.80 -2.01 2.25 -5.23 -2.64

Golden
114
(2.51 eV and 2.28 eV, respectively for bis-DIPYR and bis- -DIPYR) are slightly higher in
energy than their respective redox gaps (2.45 V and 2.25 V).
The SBCT state in a covalent dimer is formed when the lowest energy localized excited
state transfers an electron from one chromophore to another in the dimer. It is thus
characterized by a cation, localized on one (oxidized) chromophore, and an anion, localized on
the other (reduced) chromophore. The cation and anion are expected to have unique absorption
signatures which can be detected by transient absorption spectroscopy. The spectroscopic
signatures of the cationic and anionic forms of the dyes can be used to assign states observed
in transient absorption spectroscopy.
Several techniques were used to identify the spectroelectrochemical features
comprising the anionic and cationic forms of the dimers. The first method involved the use of
a customized H-cell with a 1 cm glass cuvette bridged through fritted glass to a reservoir tube.
This apparatus was filled with 0.1 M TBAF solution in acetonitrile also containing the dimer
at an optical density between 0.5 and 0.8. This solution was blanked in the UV-Vis
spectrometer, and a glassy carbon working electrode and a silver pseudo-reference electrode
were inserted above the observation window in the cuvette, while a platinum counter electrode
was inserted into the reservoir. The solution was degassed with nitrogen, and a steady stream
of nitrogen was bubbled into the solution throughout the duration of the experiment. Spectra
were acquired using chronoamperometry with the voltage set to 100 mV above the oxidation
potential or 100 mV below the reduction potential for the cation or anion spectra, respectively.
This method provided a strong signal, but deconvoluting the signal of the analyte versus that
of decomposition products was not possible.
Molecular Design for Organic Photovoltaics
Chapter 4 | 115
The second apparatus was custom-built by the Kubiak group and has been described in
the literature.
26-27
In this setup, a mirrored disk electrode coated with a thin film of acetonitrile
solution with 0.1 M TBAF and 0.005 M dimer in a rigorously air-free container is scanned
using a scanning UV-Vis spectrometer, with the beam aligned to pass twice through the
double-layer at the electrode surface, where, as in the first method, bulk electrolysis provides
either the cation or the anion. The benefits of this method include prevention of decomposition
from reaction with oxidation and the ability to isolate the spectra of decomposition products
by grounding the system and observing recovery of the neutral compound absorption . It
350 400 450 500 550 600 650
-8
-6
-4
-2
0
2

Absorption (a.u.)
Wavelength (nm)
Cation
Oxidation
Neutral
350 400 450 500 550 600 650
-4
-3
-2
-1
0
1
2

Absorption (a.u.)
Wavelength (nm)
Cation
Oxidation
Neutral

350 400 450 500 550 600 650
-2.0
-1.6
-1.2
-0.8
-0.4
0.0
0.4

Absorption (a.u.)
Wavelength (nm)
Anion
Reduction
Neutral
350 400 450 500 550 600 650
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0

Absorption (a.u.)
Wavelength (nm)
Anion
Reduction
Neutral

Figure 4.7. Spectroelectrochemistry traces for bis-DIPYR (left) and bis- -DIPYR (right). Oxidation and
reduction traces were obtained by acquiring a 10 s exposure UV/Vis spectrum through a platinum mesh
working electrode cycling at 3 V/s about the redox wave of interest. The cation and anion traces (red) are
generated from subtracting the normalized neutral absorption spectrum in acetonitrile (blue - · -) from the
oxidation and reduction traces (black), respectively.

Golden
116
suffers, however, from being considerably noisier than the first method and, because the
measurement was carried out on a scanning UV-vis spectrometer rather than a photodiode
array spectrometer, it takes several minutes to acquire each spectrum, allowing for
decomposition to occur.
The final method is performed in a 1 mm cuvette wherein a platinum mesh electrode is
placed in the viewing window of the UV-vis spectrometer, and a counter and reference
electrode are placed at the head of the solution. Instead of bulk electrolysis, which results in
the formation of a decomposed product of the irreversible oxidation, as shown in Figure 4.6,
this experiment is performed using CV at a rapid scan rate (3 V/s was suitable in this case).
This apparatus was suitable for the photodiode array spectrometer, which was set to an
exposure time of 10 s. CV was performed at +/- 100 mV about the oxidation and reduction
potentials to provide the cation and anion spectra. The rapidity of the scan rate mitigates
decomposition, and at the end of the experiment, the system can be grounded and a spectrum
of any remaining features belonging to decomposition products is easily obtained. This method
is superior in that it provides a spectrum with little noise, it is very fast, and decomposition
features can be identified. Figure 4.7 depicts the spectroelectrochemical traces for both dimers
acquired by this method and Figure 4.8 depicts the isolated decomposition features.
The spectroelectrochemical absorption spectra in Figure 7 are overlaid with the
inverted spectrum of the neutral compound, normalized to the neutral compound bleach; the
anion/cation spectra plotted therein were obtained by subtracting the neutral absorption
spectrum from the spectroelectrochemical absorption spectrum. The decomposition spectra
were used to identify features which should not be considered unique to the anion or cation
spectra. The anion of bis-DIPYR largely overlaps with the neutral dimer, albeit with a
Molecular Design for Organic Photovoltaics
Chapter 4 | 117
discernable bathochromic shift. The cation is also generally overlapped with the neutral
compound, with some absorptive features to the blue of the neutral species. In bis- -DIPYR,
the cation displays more pronounced features largely overlapping with, but slightly
hypsochromically shifted from, the neutral species, while the cation has few distinguishing
features in the visible spectrum. The decomposition features in both materials appear in the
deep blue / UV portion of the spectrum.
4.5 Time-Resolved Photophysics of bis-DIPYRs
The three proposed explanations for the differences between monomer and dimer
spectra (a distorted LE state, exciton coupling, or symmetry-breaking charge transfer) cannot
be conclusively distinguished given only the steady-state absorption spectra. The distortion
argument does not provide a likely mechanism for formation of the triplet state in bis- -
DIPYR, which is observed in cryogenic emission studies. It and the exciton coupling argument
also fail to account for the similarity between emission from frozen and fluid
methylcyclohexane solutions, which indicates that the emitting state geometry is largely
unchanged upon freezing into the Frank-Condon excited state compared to its relaxed excited
250 300 350 400 450 500 550 600 650
0
5
10

Absorption (a.u.)
Wavelength (nm)

bis-DIPYR oxidative
decomposition
250 300 350 400 450 500 550 600 650
0
1
2
3

Absorption (a.u.)
Wavelength (nm)
bis- -DIPYR oxidative
decomposition

Figure 4.8. The above spectra were recorded after recovery of the ground state bleach, and indicate an
oxidative decomposition product with a high molar absorptivity in the UV-violet region.

Golden
118
state. This observation is consistent with the dilute film emission, which even more closely
resembles solution state emission, since significant distortion (i.e. twisting) is prohibited in a
rigid matrix. The final explanation for the observed photophysics is symmetry-breaking charge
transfer, but experimental evidence for this mechanism requires time-resolved photophysics,
such that the formation of the SBCT state can be directly observed.
The SBCT state is formed via decay of the S1 excited state into a radical cation and a
radical anion, localized on either chromophore in the dimer. While strong emission, such as
that observed in the bis-DIPYRs, is rare from SBCT states, support for this hypothesis is
provided by comparing the energy of the S1 excited states of the dimers (2.51 eV and 2.28 eV,
respectively for bis-DIPYR and bis- -DIPYR) to their associated redox gaps (2.44 V and
2.25 V). The excited state energies of both dimers are slightly higher than their redox gaps,
making the possible formation of an SBCT state an exothermic process. Further, the close
proximity of the directly linked chromophores, and their mutual orthogonality as observed by
X-ray crystallography suggest that the wavefunction overlap between chromophores in the
dimers may be both strong enough to allow for charge transfer and weak enough to stabilize
the resulting symmetry-broken state. Further support for this hypothesis is given by the
observation of triplet phosphorescence at 77 K in bis- -DIPYR; it is well documented that CT
states facilitate intersystem crossing.
28-29
The time-correlated single photon counting (TCSPC)
fluorescent lifetime traces for both bis-DIPYR dimers, however, are fitted well with a mono-
exponential function independent of detected photon wavelength, suggesting that a single
radiative process is responsible for the observed emission reported in Table 4.1. The
instrument response function of the TCSPC spectrophotometer used to obtain this data does
not allow for resolution of kinetic processes occurring in under 1.2 ns, which suggested that if
Molecular Design for Organic Photovoltaics
Chapter 4 | 119
an intermediate CT state is responsible for the observed emission, it must be populated within
the first 1200 ps after excitation. To directly observe the excited state decay processes and
350 400 450 500 550 600 650
-1.0
-0.5
0.0

Abs
Wavelength (nm)
0.5 ps
1 ps
10 ps
100 ps
500 ps
900 ps
Abs.
Em.
bis-DIPYR in Cyclohexane

350 400 450 500 550 600 650
-1.0
-0.5
0.0
0.5

Abs
Wavelength (nm)
0.3 ps
1 ps
10 ps
100 ps
500 ps
900 ps
Abs.
Em.
bis- -DIPYR in Cyclohexane

350 400 450 500 550 600 650
-1.0
-0.5
0.0

Abs
Wavelength (nm)
0.5 ps
1 ps
10 ps
100 ps
500 ps
900 ps
Abs.
Em.
bis-DIPYR in THF
350 400 450 500 550 600 650
-1.0
-0.5
0.0
0.5

Abs
Wavelength (nm)
0.3 ps
1 ps
10 ps
100 ps
500 ps
900 ps
Abs.
Em.
bis- -DIPYR in THF

350 400 450 500 550 600 650
-1.0
-0.5
0.0

Abs.
Wavelength (nm)
0.3 ps
1 ps
10 ps
100 ps
500 ps
900 ps
Abs.
Em.
bis-DIPYR in Acetonitrile
350 400 450 500 550 600 650
-1.0
-0.5
0.0
0.5

Abs.
Wavelength (nm)
0.3 ps
1 ps
10 ps
100 ps
500 ps
900 ps
Abs.
Em.
bis- -DIPYR in Acetonitrile

Figure 4.9. Femtosecond absorption transient spectra of bis-DIPYR (left) and bis- -DIPYR (right), in cyclohexane
(top), THF (middle), and acetonitrile (right). The steady state absorption and emission spectra are overlaid, and
black arrows indicate features corresponding to the growth of the CT state. The above spectra were normalized to
their ground state bleach (GSB) minima so that the spectral features which are independent of S n → S 1 decay are
shown more prominently.

Golden
120
characterize the nature of the emissive state, it was necessary to use ultrafast transient
absorption spectroscopy to obtain kinetic resolution in the 150 fs to 1 ns range.
Transient absorption spectra of each dimer were obtained in cyclohexane, THF, and
acetonitrile solutions and in a dilute polymer matrix, and are reported in Figures 4.9 and 4.10.
†

The femtosecond DIPYR dimer and monomer transient absorption experiments were
performed by pumping the S1 state in each sample and probing with a white light continuum
set at magic angle with respect to the pump. The spectra were collected with an instrument
response of 150 fs between -2 ps and 1 ns. Within the first picosecond after excitation in both
nonpolar and polar solvents, the TA traces of bis-DIPYR show excited state absorption (350 –
450 nm) and stimulated emission (525 – 650 nm). Notably, the stimulated emission (SE) trace
at short time delays (from 150 fs to 1 ps) differ from the steady-state emission spectrum; there
is no significant Stokes shift and the SE appear to have vibronic features. Decay of these
excited state absorption and SE features follow in the first 10 ps after excitation in all fluid

†
The transient absorption spectra and fits were gathered and performed by Laura Estergreen, an eminently patient,
capable, and enthusiastic collaborator.
350 400 450 500 550 600 650
-1.0
-0.5
0.0
0.5

Abs.
Wavelength (nm)
0.3 ps
1 ps
10 ps
100 ps
500 ps
bis-DIPYR in PMMA

350 400 450 500 550 600 650
-1.0
-0.5
0.0
0.5

Abs.
Wavelength (nm)
0.3 ps
1 ps
10 ps
100 ps
500 ps
900 ps
bis- -DIPYR in Polystyrene

Figure 4.10. Femtosecond absorption transient spectra of bis-DIPYR (left) and bis- -DIPYR (right), in
poly(methylmethacrylate) and polystyrene, respectively. Films were spun-cast from with 2% w/w dimer in polymer,
dissolved in 75 mg/mL toluene.
Molecular Design for Organic Photovoltaics
Chapter 4 | 121
solvents. The decay in the excited state absorption corresponds to a growth in a new absorption
feature centered at 513, 523, and 506 nm in cyclohexane, THF, and acetonitrile respectively.
These new spectral features correspond to the cation features observed
spectroelectrochemically (Figure 4.7), and are assigned to growth of the CT state. The rate of
population of the CT state was determined by global analysis of the transients, and yielded
values of 3.3 ps in cyclohexane, 1.7 ps in THF, and 330 fs in acetonitrile (Table 3). The faster
rate of SBCT in the more polar solvents is consistent with known SBCT dimers.
2-3, 6, 8, 11, 13-16

The fsTA spectra of bis-DIPYR was also measured in a PMMA matrix to observe the decay
kinetics of species locked into the ground state geometry; the rate of SBCT in the polymer
matrix was slowed by two orders of magnitude relative to fluid solution (Table 4.3). These
measurements were attempted in less polar polystyrene films, but bis-DIPYR was observed to
photo-bleach in the presence of polystyrene, making it impossible to take a sufficient number
of scans before the signal decays completely. Even in poly(methylmethacrylate), the rate of
photobleaching in bis-DIPYR was rapid and the resulting spectra are plagued by significant
noise.
The TA of bis- -DIPYR likewise shows evidence for SBCT in both nonpolar and polar
media. Within the first picosecond after excitation, both excited state absorption and stimulated
emission are observed. In the next 10 ps, the intense excited state absorption feature at 510 nm
decays, while a new structured absorption feature with local maxima at ~460 and 480 nm in
all three solvents grows in. This new peak is assigned to the growth of the CT state, and
resembles features observed in the anion and cation spectra of bis- -DIPYR. Also resolved is
an apparent narrowing of the red edge of the ground-state bleach which occurs on the same
timescale as the decay of the 510 nm S1 band; this narrowing corresponds well to a growth in
Golden
122
the anion absorption as determined by spectroelectrochemistry, as well as decreased stimulated
emission from the S1, which is replaced by a red-shifted stimulated emission feature that better
resembles the steady-state emission spectrum. Fitting of these features by global analysis
yielded rates of population of the CT state equal to 2 ps, 2 ps, and 1.4 ps in cyclohexane, THF,
and acetonitrile, respectively. These rates are remarkable both for the fact that the CT state
appears to be nearly as stable in nonpolar solvent as it is in polar solvents. Femtosecond TA
traces of bis-α-DIPYR in a dilute polystyrene thin film are consistent with steady-state
photophysical measurements in dilute films, which indicate that the emitting species is the
same in the solid state as in solution. In the rigid matrix, SBCT is also observed, albeit with a
rate two orders of magnitude slower than in solution (Table 4.3).
The CT state persists well past the experimental timeframe of the femtosecond transient
absorption, so excited state decay processes were probed by nanosecond transient absorption
spectroscopy (Figure 4.11). Decay of the CT state in bis- -DIPYR was analyzed by
nanosecond TA in cyclohexane and in THF and fit via global analysis to a four-state model.
We were unable to acquire nanosecond TA spectra of bis-DIPYR, however, due to poor energy
Table 4.3. Rates of excited state transitions in DIPYR dimers in various media.
kSBCT (s
-1
) kISC (s
-1
) kIC (s
-1
) krec (s
-1
) knr (s
-1
)
bis-DIPYR PMMA 5x10
9
- - 2x10
8
-
cyclohexane 3x10
11
- - 3x10
8
-
THF 6x10
11
- - 4x10
8
-
acetonitrile 3x10
12
- - 1x10
9
-
bis- α-DIPYR polystyrene 2x10
10
- - 2x10
8
-
cyclohexane 5x10
11
7.7x10
6
1.2x10
8
2.1x10
8
3x10
5
THF 5x10
11
8.3x10
6
2.9x10
7
2.1x10
8
3x10
4
acetonitrile 7x10
11
- - 2.4x10
8
-

Rate of SBCT determined by global analysis of fsTA. Rates of ISC, IC, and non-radiative decay determined by a
global analysis of nsTA. Rate of recombination determined by TCSPC and applied to global analysis fitting of
nsTA.

Molecular Design for Organic Photovoltaics
Chapter 4 | 123
matching between the excitation source and the absorption manifold. In bis- -DIPYR in both
THF and cyclohexane, the CT state decays by two competing mechanisms. The first and most
predominant is recombination to the S0 ground state, occurring in 4.8 ns, 4.8 ns, and 4.2 ns in
cyclohexane, THF, and acetonitrile, respectively. These recombination rates are consistent
with the fluorescence emission lifetimes measured by TCSPC (Table 4.1), and thus
fluorescence emission in the dimers is assigned to direct CT→S0 emission. Although relatively
rare, fluorescence from an SBCT state is not unknown in the literature; two important structural
analogues of DIPYR dimers, 9,9’-bianthryl, and 10,10’-dicyano-9,9’-bianthryl, are also
characterized by bright SBCT states (0.22 and 0.29, respectively, in acetonitrile).
30
Notably,
350 400 450 500 550 600 650
-10
-5
0
5
50 ns
100 ns
510 ns
1 s
5 s
11 s

Abs. (mOD)
Wavelength (nm)
0.3 ps
1 ps
10 ps
100 ps
500 ps
1 ns
5 ns
bis- -DIPYR in Cyclohexane
350 400 450 500 550 600 650
-5
0
5
50 ns
150 ns
510 ns
1.05 s
21 s
51 s

Abs. (mOD)
Wavelength (nm)
0.3 ps
1 ps
10 ps
100 ps
500 ps
1.6 ns
5 ns
bis- -DIPYR in THF

10
-2
10
-1
10
0
10
1
10
2
10
3
10
4
10
5
10
6
10
7
0.0
0.2
0.4
0.6
0.8
1.0
Concentration
Time (ps)
S
1
(ICT)
1
(ICT)
3
T
1
bis- -DIPYR in Cyclohexane
10
-2
10
-1
10
0
10
1
10
2
10
3
10
4
10
5
10
6
10
7
0.0
0.2
0.4
0.6
0.8
1.0

Concentration
Time (ps)
S
1
(ICT)
1
(ICT)
3
T
1
bis- -DIPYR in THF

Figure 4.11. Combined femtosecond and nanosecond transient absorption of bis- -DIPYR in cyclohexane (top
left) and THF (top right), and (bottom) time-dependent concentration of the species-associated decay spectra.

Golden
124
10,10’-dicyano-9,9’-bianthryl also
displays significant CT character in
cyclohexane (55% of the excited state
population),
31
while 9,9’-bianthryl has
considerable less CT character in
nonpolar solvents, showing partial CT in
heptane.
32
In the case of DIPYR dimers,
it appears that 100% of S1 excitons
decay to an SBCT state. This appears to
be a unique phenomenon in nonpolar media and may be due to exceptional polarizability in
the DIPYR dimers.
As evidenced by the steady-state phosphorescence study, the second SBCT state decay
pathway involves the formation of a triplet state. According to the global analysis fitting, a
four-state model best describes the decay dynamics in the nanosecond decay traces, suggesting
that an intermediating state between the SBCT state and the ultimate triplet is formed. This
state is tentatively assigned as a triplet SBCT state pending further investigation. The assigned
states and their associated rate constants in THF are depicted in the Jablonski diagram in
Figure 4.12; the full set of rate constants for the observed excited state decay processes are
summarized in Table 4.3.
4.6 Theoretical Modeling
The monomers DIPYR and -DIPYR exhibit relatively simple excited state photophysics.
After photon absorption, radiative and non-radiative decay process occur directly from the S1
and T1 states to the S0 state (Figure 2.15). In contrast, the S1 state for the dimers rapidly decays
0.0
0.5
1.0
1.5
2.0
2.5
T
1
3
CT
S
1
k
nr
= 3 x 10
4
k
IC
= 2.9 x 10
7
k
SBCT
=
5 x 10
11
k
ISC
= 8.3 x 10
6
bis-DIPYR
bis- -DIPYR
Energy (eV)
Abs.
Emission
k
rec
= 2.1 x 10
8
1
CT

Figure 4.12. Jablonski diagram summarizing state energies
of bis-DIPYR (black) and bis- -DIPYR (red). Rate
constants are derived from femtosecond and nanosecond
transient absorption spectra of bis- -DIPYR in THF.

Molecular Design for Organic Photovoltaics
Chapter 4 | 125
into a symmetry-broken charge transfer state in both polar and nonpolar solvents. The SBCT
state then decays by either of two pathways: recombination to the ground state or formation of
a triplet state via an intermediate
3
SBCT state. Bis-DIPYRs are anomalous in that that the
SBCT state is formed nearly as rapidly in cyclohexane as it is in the much more polar
acetonitrile and that the SBCT is a bright state, responsible for the observed fluorescence
emission. Theoretical modeling of the orbital configurations of both dimers was performed
using densityfFunctional theory (DFT) at the B3LYP/6-31G** level to shed light upon the
novel photophysical behavior of the bis-DIPYRs.
The resulting geometry optimized structures and their molecular orbital configurations
were validated by comparison to X-ray crystal structures and measured electrochemical and
photophysical properties. The calculated geometry of bis-DIPYR exhibits dihedral angles
between the chromophores of 85.9 degrees, compared to 86.0 degrees according to the crystal
structure. The BF2 groups are puckered out of plane by an average of 0.42 Å in the calculated
structure, compared to 0.51 Å in the crystal. These small differences can be explained by
crystal packing forces, as optimized geometries were calculated for molecules in the gas phase.
In bis- -DIPYR, computational modeling results in a dihedral angle of 89.8 degrees between
the planes of the chromophores, compared to 89.9  from the crystal structure of bis- -DIPYR.
The uncorrected calculated singlet excited states (Table 4.4), determined by TD-DFT of the
optimized geometries, are 0.25 and 0.14 eV higher than experimental values in bis-DIPYR and
bis- -DIPYR, respectively; error margins of this magnitude are typical for cyanine-like dyes
due to significant multireference character in their excited singlet states.
33
These errors are
smaller than those obtained from TD-DFT for the parent structures, which differed from
experiment by an average of 0.44 eV.
17
As a further validation, excited states of bis-DIPYR
Golden
126
were computed using the multi-reference XMCQDPT2 (extended multi-configuration quasi-
degenerate second order perturbation theory) method, which was shown previously to more
accurately calculate the singlet excited state energies of the monomers, and the computed S1
energy (2.36 eV) is in good agreement with the experimental value (2.51 eV).
17, 34

XMCQDPT2 calculations for bis- -DIPYR, however, were found to be prohibitively
expensive.
‡
The triplet energy was determined by the delta self-consistent field approach
within DFT, wherein the triplet state geometry was optimized and the energy of the optimized
ground state was subtracted from the energy of the optimized triplet state.
35
This method
provided a very good estimation of the triplet energy, within 100 meV of experimental values.
The frontier molecular orbital configurations of both dimers are localized on the
dipyridylmethene portion of the dimers, with very little orbital density on the outer benzene
rings of bis- -DIPYR, similar to what is observed in the monomers (Figure 4.13). Both
dimers exhibit localization of the LUMO orbitals on each chromophore, while the HOMO
orbitals are characterized by significant electron density at the meso position, which bridges
the two chromophores. For this reason, it is reasonable to expect that there is some electronic
coupling between the two chromophores, even in the ground state, and that this is the source
of the bathochromic shifts observed when comparing the UV-Vis absorption spectra of the
dimers to their corresponding monomers. It is interesting to note that the orbital configuration

‡
USC’s HPC (high-performance computing) resource was used for the calculation of bis- -DIPYR, but the size
of the structure was too large to be computed accurately. The calculated orbitals did not make physical sense.
Table 4.4. Calculated and experimental (  max abs) excited state energies.
S 1 calc. S 1 exp. T 1 calc. T 1 exp.
bis-DIPYR 449 nm (2.76/2.36
a
eV) 493 nm (2.51 eV) 596 nm (2.08 eV) 593 nm (2.09 eV)
bis- -DIPYR 513 nm (2.42 eV) 543 nm (2.28 eV) 685 nm (1.81 eV) 657 nm (1.89 eV)

a
XMCQDPT2(14,13)/6-31G(d)
Molecular Design for Organic Photovoltaics
Chapter 4 | 127
of these bis-DIPYRs is opposite to that observed in bis-BODIPYs, which are characterized by
a node at the meso position of their HOMO orbitals and chromophoric bridging in the LUMOs.
The natural transition orbitals (SI) of the calculated S1 excited states in both dimers are
characterized by a hole at the HOMO, localized at the meso position, and an electron at the
LUMO, delocalized across the chromophore. The rate of formation of the SBCT state is
thought to be proportional to the
degree of wavefunction overlap
between coupled chromophores.
2

The rapidity and near solvent
independence of charge transfer
rates in both DIPYR dimers is
possibly due to the orbital
configurations between coupled
chromophores. The most possible
explanation for the unique
photophysical behavior or the
bis-DIPYR systems involves a
mechanism for formation of the
SBCT state in which the hole is
localized at the meso-carbon of one of the DIPYR moieties after photoexcitation, then spreads
to the adjacent chromophore but remains insulated from solvent near the meso C-C bridge,
while the LUMO remains localized on the DIPYR moiety in which it was originally formed;
in this mechanism, the resulting CT state may involve the transfer of partial charges, with some

Figure 4.13. bis-DIPYR (left) and bis- -DIPYR (right) HOMO
(bottom, solid) and LUMO (top, mesh) orbitals.
Golden
128
hole density on the originally excited DIPYR moiety, while a portion of the hole becomes
localized on the adjacent DIPYR.
32, 36
This process is likely facilitated by the highly polarizable
zwitterionic chromophores, which self-stabilize the induced dipole. In such a system, it is
possible that small changes in local solvent density could further facilitate stabilization of the
SBCT state; this is consistent with the observation of SBCT state formation in polymer
matrices at much slower rates than in solvent media.
31, 37
Insulation of charge density on the
interior of these relatively bulky systems could be an explanation for the lack of
solvatochromism in the fluorescence spectra. Further probing the mechanism of SBCT state
formation, perhaps by transient infrared spectroscopy, is necessary to develop a clearer picture
of this unique phenomenon.
4.7 Shifting Paradigms in Symmetry-Breaking Charge Transfer
In this study, two novel bis-DIPYR compounds, homoleptic covalent meso-linked dimers
of DIPYR and -DIPYR were synthesized, showing strong UV- and visible-light absorption
and yellow-green to orange fluorescence. The excited state photophysical properties of the
dimers were explored using a combination of steady-state and time-resolved spectroscopic
methods. It was discovered that the dimers couple such that the chromophore units are mutually
orthogonal, with the HOMO localized at the meso position and bridging the two chromophores,
and the LUMO delocalized across each chromophore. This coupling facilitates rapid and
efficient symmetry-breaking charge transfer in both polar acetonitrile (330 fs in bis-DIPYR
and 1.4 ps in bis- -DIPYR) and nonpolar cyclohexane (3 ps in bis-DIPYR and 2 ps in bis- -
DIPYR) solvents. The dimers were shown to exhibit stable, highly emissive ( ΦPL = 0.03 to
0.57) symmetry-broken states even in nonpolar solvents.
Molecular Design for Organic Photovoltaics
Chapter 4 | 129
The structural likeness between bis-DIPYRs, 9,9’-bianthryl,
6, 8, 30
and bis-BODIPY
13

merits a brief consideration of their photophysical similarities and differences. All three
structure classes have been shown to undergo symmetry breaking charge transfer, forming a
radical cation localized on one chromophore, and a radical anion localized on the other
chromophore shortly after photoexcitation. The SBCT state is formed in bis-BODIPY in 170 fs
in acetonitrile, in 4.5 ps in the considerably less polar toluene, and not at all in cyclohexane,
where locally excited BODIPY fluorescence is observed. The faster rate of SBCT in
bis-BODIPY may be due to the relative rearrangement of HOMO and LUMO orbital densities
in the BODIPY dimer compared to the DIPYR dimers. In bis-BODIPY, the HOMO is
delocalized across the chromophore, while the LUMO is localized at the meso position
bridging the two chromophores. In both DIPYR dimers, as discussed above, the opposite
arrangement is observed. The localization of excited state electron density at the meso position
in bis-BODIPY may be the source of the faster rate of SBCT in bis-BODIPY in acetonitrile.
Whereas the rate of SBCT is significantly slowed in bis-BODIPY in toluene relative to
acetonitrile, that of bis- -DIPYR remains nearly the same in nonpolar solvents as it is in
acetonitrile. This suggests that the mechanism for charge transfer in DIPYR dimers is different
than it is in bis-BODIPY, possibly involving migration of the hole in DIPYRs versus migration
of the electron in bis-BODIPY. Also of note, the SBCT state in bis-BODIPY is largely a dark
state, in contrast to that of the bis-DIPYRs, which is responsible for the observed fluorescence
emission.
In 9,9’-bianthryl as well, SBCT has been shown to occur efficiently in solvents of
strong polarity, within 390 fs in acetonitrile, and to a lesser extent (about 10% of the excited
state population) in 510 fs in cyclohexane.
31
The rate of SBCT is enhanced in the more
Golden
130
polarizable 10,10’-dicyano-9,9’-bianthryl relative to 9,9’-bianthryl in acetonitrile (150 fs) and
in cyclohexane (400 fs), where the equilibrium population of the CT state relative to the
localized excited state is also increased in cyclohexane from 10% in 9,9’-bianthryl to 55% in
the cyano- derivative. Given these contrasts, it is herein proposed that the rate of symmetry
breaking charge transfer in mutually orthogonal dimers is enhanced by (a) increasing the
polarizability of the dimer thereby allowing greater self-solvation effects to stabilize CT
transitions, (b) localizing HOMO or LUMO density at the bridging position between
chromophores, and (c) allowing for either rapid torsional (via steric decongestion) or solvent
(via an increase in intermolecular interaction) relaxation to induce the charge transfer event.
This work represents the discovery of a new structure class capable of rapid and
efficient charge transfer to form a stable SBCT state which persists for several nanoseconds.
Such a system is highly promising for application to new energy conversion technologies
including organic photovoltaics, solar fuels, artificial photosynthesis, and photonics. Many
questions remain to be answered, namely: what is the mechanism for SBCT state formation in
bis-DIPYRs and how does is compare to bis-BODIPYs; what are the rates for triplet formation
and decay in bis-DIPYR and how do they compare to those observed in bis- -DIPYR; and
will SBCT in bis-DIPYRs translate to higher open-circuit voltages in organic photovoltaics
compared to monomer models?
4.8 Experimental Methods
General Synthesis of bis-DIPYRs:
A 15 mM solution of 1,1,2,2-tetra(pyridin-2-yl)ethane or 1,1,2,2-tetra(quinolin-2-yl)ethane
ligand in dry 1,2-dichloroethane was prepared in an N2-purged schlenk flask equipped with a
magnetic stir bar and fitted with a reflux condenser. The flask was submerged in a preheated
Molecular Design for Organic Photovoltaics
Chapter 4 | 131
oil bath and brought to reflux, at which time 4.0 eq. boron trifluoride diethyl etherate were
added dropwise, at which time a precipitate was formed. The solution was stirred for 2 hours
at reflux, then cooled to room temperature and treated with 10 eq. N,N-diisopropylethylamine,
causing the precipitate to dissolve. The solution was washed with water and the aqueous layer
was separated and extracted three times with dichloromethane. The organic layers were
combined, dried over sodium sulfate, filtered, and reduced to a polycrystalline solid by rotary
evaporation. The products were purified by silica gel flash chromatography with the gradient
eluent 0-50% dichloromethane in hexanes, followed by recrystallization from a concentrated
dichloromethane solution layered with hexanes.
bis-DIPYR: 0.668 g (31% yield) bright orange crystals.
1
H NMR (500 MHz, Chloroform-
d) δ 8.11 (d, J = 6.1 Hz, 4H), 7.28 – 7.21 (m, 4H), 6.74 (d, J = 9.0 Hz, 4H), 6.63 (t, J = 6.8 Hz,
4H).
13
C NMR (101 MHz, Chloroform-d) δ 149.39, 137.43 (t J = 1.8 Hz), 136.92, 119.34,
112.07, 88.21. Anal. Calcd for C22H16B2F4N4: C, 60.88; H, 3.72; N, 12.91. Found: C, 60.99;
H, 3.81; N, 12.61. MS (MALDI-TOF) m/z calcd for C22H16B2F4N4
+
434.150 [M
+
]; found
633.892.
bis- -DIPYR: 0.105 g (32% yield) magenta needles.
1
H NMR (400 MHz, Chloroform-d)
δ 8.79 – 8.73 (m, 4H), 7.68 (ddd, J = 8.9, 7.1, 1.7 Hz, 4H), 7.55 (d, J = 9.1 Hz, 4H), 7.52 (dd,
J = 7.8, 1.6 Hz, 4H), 7.33 (ddd, J = 7.8, 7.0, 0.9 Hz, 4H), 7.01 (d, J = 9.3 Hz, 4H).
13
C NMR
(101 MHz, Chloroform-d) δ 150.53, 140.30, 138.23, 131.07, 128.23, 124.97, 124.22, 121.91
(t, J = 8.5 Hz), 119.72, 94.25. Anal. Calcd for C38H24B2F4N4: C, 71.96; H, 3.81; N, 8.83. Found:
C, 69.60; H, 4.81; N, 7.78. MS (MALDI-TOF) m/z calcd for C38H24B2F4N4
+
634.212 [M
+
];
found 634.031.
Golden
132
4.9 Bibliography
1. Rybtchinski, B.; Sinks, L. E.; Wasielewski, M. R., Combining Light-Harvesting and
Charge Separation in a Self-Assembled Artificial Photosynthetic System Based on
Perylenediimide Chromophores. J. Am. Chem. Soc. 2004, 126 (39), 12268-12269.
2. Vauthey, E., Photoinduced Symmetry-Breaking Charge Separation. ChemPhysChem
2012, 13 (8), 2001-2011.
3. Bartynski, A. N.; Gruber, M.; Das, S.; Rangan, S.; Mollinger, S.; Trinh, C.; Bradforth,
S. E.; Vandewal, K.; Salleo, A.; Bartynski, R. A.; Bruetting, W.; Thompson, M. E., Symmetry-
Breaking Charge Transfer in a Zinc Chlorodipyrrin Acceptor for High Open Circuit Voltage
Organic Photovoltaics. J. Am. Chem. Soc. 2015, 137 (16), 5397-5405.
4. Sisson, A. L.; Sakai, N.; Banerji, N.; Fürstenberg, A.; Vauthey, E.; Matile, S., Zipper
Assembly of Vectorial Rigid-Rod π-Stack Architectures with Red and Blue
Naphthalenediimides: Toward Supramolecular Cascade n/p-Heterojunctions. Angew. Chem.
Int. Ed. 2008, 47 (20), 3727-3729.
5. Collet, E.; Lemée-Cailleau, M.-H.; Buron-Le Cointe, M.; Cailleau, H.; Wulff, M.; Luty,
T.; Koshihara, S.-Y.; Meyer, M.; Toupet, L.; Rabiller, P.; Techert, S., Laser-Induced
Ferroelectric Structural Order in an Organic Charge-Transfer Crystal. Science 2003, 300
(5619), 612.
6. Mataga, N.; Yao, H.; Okada, T.; Rettig, W., Charge-transfer rates in symmetric and
symmetry-disturbed derivatives of 9,9'-bianthryl. J. Phys. Chem. 1989, 93 (9), 3383-3386.
7. Grozema, F. C.; Swart, M.; Zijlstra, R. W. J.; Piet, J. J.; Siebbeles, L. D. A.; van
Duijnen, P. T., QM/MM Study of the Role of the Solvent in the Formation of the Charge
Separated Excited State in 9,9‘-Bianthryl. J. Am. Chem. Soc. 2005, 127 (31), 11019-11028.
8. Rettig, W.; Zander, M., Broken Symmetry in Excited States of 9,9′-Bianthryl. Berichte
der Bunsengesellschaft für physikalische Chemie 1983, 87 (12), 1143-1149.
9. Lueck, H.; Windsor, M. W.; Rettig, W., Pressure dependence of the kinetics of
photoinduced intramolecular charge separation in 9,9'-bianthryl monitored by picosecond
transient absorption: comparison with electron transfer in photosynthesis. J. Phys. Chem. 1990,
94 (11), 4550-4559.
10. Holman, M. W.; Yan, P.; Adams, D. M.; Westenhoff, S.; Silva, C., Ultrafast
Spectroscopy of the Solvent Dependence of Electron Transfer in a Perylenebisimide Dimer. J.
Phys. Chem. A 2005, 109 (38), 8548-8552.
11. Wu, Y.; Young, R. M.; Frasconi, M.; Schneebeli, S. T.; Spenst, P.; Gardner, D. M.;
Brown, K. E.; Würthner, F.; Stoddart, J. F.; Wasielewski, M. R., Ultrafast Photoinduced
Symmetry-Breaking Charge Separation and Electron Sharing in Perylenediimide Molecular
Triangles. J. Am. Chem. Soc. 2015, 137 (41), 13236-13239.
12. Sung, J.; Nowak-Król, A.; Schlosser, F.; Fimmel, B.; Kim, W.; Kim, D.; Würthner, F.,
Direct Observation of Excimer-Mediated Intramolecular Electron Transfer in a Cofacially-
Stacked Perylene Bisimide Pair. J. Am. Chem. Soc. 2016, 138 (29), 9029-9032.
13. Whited, M. T.; Patel, N. M.; Roberts, S. T.; Allen, K.; Djurovich, P. I.; Bradforth, S.
E.; Thompson, M. E., Symmetry-breaking intramolecular charge transfer in the excited state
of meso-linked BODIPY dyads. Chem. Commun. 2012, 48 (2), 284-286.
14. Trinh, C.; Kirlikovali, K.; Das, S.; Ener, M. E.; Gray, H. B.; Djurovich, P.; Bradforth,
S. E.; Thompson, M. E., Symmetry-Breaking Charge Transfer of Visible Light Absorbing
Systems: Zinc Dipyrrins. J. Phys. Chem. C 2014, 118 (38), 21834-21845.
Molecular Design for Organic Photovoltaics
Chapter 4 | 133
15. Cook, R. E.; Phelan, B. T.; Kamire, R. J.; Majewski, M. B.; Young, R. M.;
Wasielewski, M. R., Excimer Formation and Symmetry-Breaking Charge Transfer in Cofacial
Perylene Dimers. J. Phys. Chem. A 2017, 121 (8), 1607-1615.
16. Giaimo, J. M.; Gusev, A. V.; Wasielewski, M. R., Excited-State Symmetry Breaking
in Cofacial and Linear Dimers of a Green Perylenediimide Chlorophyll Analogue Leading to
Ultrafast Charge Separation. J. Am. Chem. Soc. 2002, 124 (29), 8530-8531.
17. Golden, J. H.; Facendola, J. W.; Sylvinson M. R, D.; Baez, C. Q.; Djurovich, P. I.;
Thompson, M. E., Boron Dipyridylmethene (DIPYR) Dyes: Shedding Light on Pyridine-
Based Chromophores. J. Org. Chem. 2017, 82 (14), 7215-7222.
18. Golden, J. H.; Estergreen, L.; Porter, T.; Tadle, A. C.; Sylvinson M. R, D.; Facendola,
J. W.; Kubiak, C. P.; Bradforth, S. E.; Thompson, M. E., Symmetry-Breaking Charge Transfer
in Boron Dipyridylmethene (DIPYR) Dimers. ACS Appl. Energy Mater. 2018.
19. Itoh, M.; Hirano, K.; Satoh, T.; Miura, M., Copper-Catalyzed α-Methylenation of
Benzylpyridines Using Dimethylacetamide as One-Carbon Source. Org. Lett. 2014, 16 (7),
2050-2053.
20. Scheibe, G.; Friedrich, H. J., Synthese und Reaktionen der Chinolylmethane. Chem.
Ber. 1961, 94 (5), 1336-1345.
21. Reichardt, C., Solvatochromic Dyes as Solvent Polarity Indicators. Chem. Rev. 1994,
94 (8), 2319-2358.
22. Kumpulainen, T.; Lang, B.; Rosspeintner, A.; Vauthey, E., Ultrafast Elementary
Photochemical Processes of Organic Molecules in Liquid Solution. Chem. Rev. 2017, 117 (16),
10826-10939.
23. Wang, S.; Cai, J.; Sadygov, R.; Lim, E. C., Intramolecular Charge Transfer and
Solvent-Polarity Dependence of Radiative Decay Rate in Photoexcited Dinaphthylamines. J.
Phys. Chem. 1995, 99 (19), 7416-7420.
24. Djurovich, P. I.; Mayo, E. I.; Forrest, S. R.; Thompson, M. E., Measurement of the
lowest unoccupied molecular orbital energies of molecular organic semiconductors. Org.
Electron. 2009, 10 (3), 515-520.
25. W. D’Andrade, B.; Datta, S.; R. Forrest, S.; Djurovich, P.; Polikarpov, E.; Thompson,
M., Relationship between the ionization and oxidation potentials of molecular organic
semiconductors. Org. Electron. 2005, 6 (1), 11-20.
26. Zavarine, I. S.; Kubiak, C. P., A versatile variable temperature thin layer reflectance
spectroelectrochemical cell. J. Electroanal. Chem. 2001, 495 (2), 106-109.
27. Machan, C. W.; Sampson, M. D.; Chabolla, S. A.; Dang, T.; Kubiak, C. P., Developing
a Mechanistic Understanding of Molecular Electrocatalysts for CO2 Reduction using Infrared
Spectroelectrochemistry. Organometallics 2014, 33 (18), 4550-4559.
28. Dance, Z. E. X.; Mickley, S. M.; Wilson, T. M.; Ricks, A. B.; Scott, A. M.; Ratner, M.
A.; Wasielewski, M. R., Intersystem Crossing Mediated by Photoinduced Intramolecular
Charge Transfer: Julolidine−Anthracene Molecules with Perpendicular π Systems. J. Phys.
Chem. A 2008, 112 (18), 4194-4201.
29. Liu, Y.; Zhao, J.; Iagatti, A.; Bussotti, L.; Foggi, P.; Castellucci, E.; Di Donato, M.;
Han, K.-L., A Revisit to the Orthogonal Bodipy Dimers: Experimental Evidence for the
Symmetry Breaking Charge Transfer-Induced Intersystem Crossing. J. Phys. Chem. C 2018.
30. Piet, J. J.; Schuddeboom, W.; Wegewijs, B. R.; Grozema, F. C.; Warman, J. M.,
Symmetry Breaking in the Relaxed S1 Excited State of Bianthryl Derivatives in Weakly Polar
Solvents. J. Am. Chem. Soc. 2001, 123 (22), 5337-5347.
Golden
134
31. Kovalenko, S. A.; Pérez Lustres, J. L.; Ernsting, N. P.; Rettig, W., Photoinduced
Electron Transfer in Bianthryl and Cyanobianthryl in Solution: The Case for a High-
Frequency Intramolecular Reaction Coordinate. J. Phys. Chem. A 2003, 107 (48), 10228-
10232.
32. Asami, N.; Takaya, T.; Yabumoto, S.; Shigeto, S.; Hamaguchi, H.-o.; Iwata, K., Two
Different Charge Transfer States of Photoexcited 9,9′-Bianthryl in Polar and Nonpolar
Solvents Characterized by Nanosecond Time-Resolved Near-IR Spectroscopy in the
4500−10 500 cm−1 Region. J. Phys. Chem. A 2010, 114 (22), 6351-6355.
33. Charaf-Eddin, A.; Le Guennic, B.; Jacquemin, D., Excited-states of BODIPY-
cyanines: ultimate TD-DFT challenges? RSC Advances 2014, 4 (90), 49449-49456.
34. Granovsky, A. A., Extended multi-configuration quasi-degenerate perturbation theory:
The new approach to multi-state multi-reference perturbation theory. J. Chem. Phys. 2011, 134
(21), 214113.
35. Gavnholt, J.; Olsen, T.; Engelund, M.; Schiøtz, J., Delta self-consistent field method to
obtain potential energy surfaces of excited molecules on surfaces. Phys. Rev. B 2008, 78 (7),
075441.
36. Takaya, T.; Saha, S.; Hamaguchi, H.-o.; Sarkar, M.; Samanta, A.; Iwata, K., Charge
Resonance Character in the Charge Transfer State of Bianthryls: Effect of Symmetry Breaking
on Time-Resolved Near-IR Absorption Spectra. J. Phys. Chem. A 2006, 110 (13), 4291-4295.
37. Ivanov, A. I.; Dereka, B.; Vauthey, E., A simple model of solvent-induced symmetry-
breaking charge transfer in excited quadrupolar molecules. J. Chem. Phys. 2017, 146 (16),
164306.

Molecular Design for Organic Photovoltaics
Chapter 5 | 135
Chapter 5

Symmetry-Breaking Charge Transfer in Dipyrrin
Dimers: Zinc Dipyrrins and b i s-BODIPYs

5.1 Symmetry-Breaking Charge Transfer in bis-BODIPYs and Zinc Dipyrrins
Symmetry-breaking charge transfer, introduced in Chapter 4, has been shown to
increase the VOC in OPVs compared to systems with active layer materials which have
otherwise identical electrochemical potentials.
1
It is thought that this is due to a reduction in
the back-electron recombination rate following the donor/acceptor charge transfer step of
photocurrent generation when either the donor or acceptor material undergoes SBCT (Figure
1.11). This is fundamentally an argument for kinetic control over the through-space interaction
between a photogenerated electron and hole pair; instead of building large energetic
(thermodynamic) driving force for charge separation (EL = Eg - ECT), an intermediating charge
transfer step (SBCT) is invoked, wherein the hole and electron are spatially separated on either
the donor or acceptor material before charge is transferred to the acceptor or donor material.
Although the requirements for SBCT are qualitatively understood (S1 energy must be
greater than redox gap, and electronic coupling between chromophores must be limited to the
minimum overlap required for CT), a thorough quantitative description for the mechanism of
SBCT and its role in OPVs is lacking.
2-3
Material classes known to perform SBCT include
acene,
2, 4-10
perylene,
11-15
and BODIPY dimers,
16-17
and homoleptic metallodipyrrins,
1, 18
as
well as the famous in vivo and bio-inspired examples in porphyrins.
19-22
With the discovery of
reduced solvation effects in the formation of the SBCT state in bis-DIPYRs,
23
questions over
Golden
136
the role of through-space interaction between the natural transition orbitals comprising the S 1
and SBCT states have risen. Specifically, it has become clear that some of the accepted
paradigms in the understanding of SBCT – for example that high dielectric solvents are
required to stabilize the SBCT state, or that SBCT states are dark states – are not necessarily
true.
3, 9, 24
A better quantitative understanding of SBCT will require comparative study between
similar materials structures wherein orbital density and charge separation distances can be
minutely tuned.
Chapter 4 presented the bis-DIPYR orthogonal chromophoric dimer system, where it
was seen that SBCT could be affected with little to no stabilization from solvent reorganization
effects. In this material, the HOMO density is consolidated largely at the meso-carbon-carbon
bridge between the orthogonal chromophores, such that any small degree of bond rotation
would induce electronic coupling between the HOMOs of the two chromophores. It would be
instructive to compare this system against one in which electronic coupling via direct
wavefunction overlap was strictly forbidden yet the chromophores remain rigorously
orthogonal. Such a case could be designed in a tetrahedral metallo DIPYR dimer such as
Zn(II)DIPYR2. However, efforts to synthesize ZnDIPYR2 from either 2,2’-dipyridlylmethane
or 2,2’-diquinolylmethane produced highly unstable solutions. Although a chemical
transformation could be seen to occur in the glove box when either ligand was exposed to
diethyl zinc, and indeed even after the growth of vivid orange crystals from the 2,2’-
dipyridylmethane / diethyl zinc reaction, exposure to air induced a rapid transformation to an
intractable white solid (the crystals became colorless and transparent) within seconds to
minutes. The reason for this apparent instability is not immediately clear, but it was evident
that the Zn DIPYR system was not suitable for the rigorous photophysical studies required,
Molecular Design for Organic Photovoltaics
Chapter 5 | 137
which would necessitate the synthesis of a significant amount of material which is stable in
solution and under laser irradiation.
A simple alternative to the DIPYR system is the dipyrrin system. Initially the
inspiration for the development of dipyridylmethenes, dipyrromethenes are the brightly
emissive, photostable chromophores known for their intense absorptivities (  >10
5
) and high
photoluminescent quantum efficiencies  PL. Furthermore, both carbon-bridged and
metallodipyrin dimers have been reported in the literature and are known for their ability to
undergo SBCT.
1, 16-18, 25-26
This chapter compares syntheses of both meso-carbon-carbon
bridged BODIPY dimers (bis-BODIPYs) and directly analogous zinc dipyrrin (ZnDIPY2)
structures. Initial photophysical characterization of both dimer geometries, including some
excited state decay dynamics, are presented in order to lay the groundwork for a quantitative
understanding of the role of orbital density overlap and charge separation effects in SBCT
which can be used to establish materials design principles for SBCT materials in OPVs.
5.2 Synthesis of bis-BODIPYs
The original published synthesis for the meso-bridged orthogonal bis-BODIPY
architecture (B-1) was published by the Thompson lab in 2012.
16
This synthesis involved four
steps with independent workups and purification and resulted in a 10% overall yield (0.025 g)
from the deprotected 2-methylpyrrole. For the purposes of thorough structural,
electrochemical, photophysical, and device characterization, significantly more material will
be required, so a more efficient synthesis was necessary. Fortunately, in 2014, another bis-
BODIPY study was published which improved the synthesis significantly; this preparation
requires only one pot over a few hours, no workup, followed by purification with silica gel
flash column chromatography to yield the bis-BODIPY from pyrrole.
27
Using this method over
Golden
138
only a few days, two bis-BODIPY dimers, bis-2,4-dimethylBODIPY (B-2) and bis-3-ethyl-
2,4-dimethylBODIPY (B-3)
*
were synthesized (Scheme 5.1); this work represents the first
reported synthesis of B-3. An adaptation of this synthetic scheme was attempted using a 4,7-
dihydroisoindole precursor followed by oxidation with DDQ to obtain a -benzannulated
meso-bridged BODIPY dimer. However, the reaction resulted in only two substitutions at the
oxalyl chloride, forming an entirely unique structure B-rad which is discussed in Section 5.9.

*
This material was synthesized by Ali Akil as part of his graduate research in the Thompson Lab.
Scheme 1. Synthesis of bis-BODIPYs from deprotected pyrrole starting materials.

Molecular Design for Organic Photovoltaics
Chapter 5 | 139
5.3 Structural and Photophysical Characterization of bis-BODIPYs
Notably, while the structures of B-2 and B-3 preclude chromophore coupling between
the respective BODIPYs in the meso-coupled dimer due to steric interactions between the 4-
methyl substituents, the structure of B-1 allows for some orbital overlap.
16
The absorption and
emission spectra of B-1 have been reported previously, and it can be seen that the absorption
spectrum narrows and blue-shifts somewhat in dichloromethane compared to cyclohexane.
16

In B-2 and B-3 (Figure 5.2), the absorption spectra remain relatively constant in all solvents,
with the exception of toluene which is subtly red-shifted relative to all other solvents. This
difference may be due to -  interaction between the toluene solvent and the chromophores.
†

The absorption manifold differs significantly in lineshape from that of a BODIPY monomer,
first in that the 0-0 vibronic peak is not the max, and second in that the absorption band is
broadened and somewhat red-shifted compared to the monomer. This may indicate

†
Complete structural and photophysical characterization of B-2 and B-3 is ongoing at the time of writing and will
be performed by Ali Akil of the Thompson lab in collaboration with Laura Estergreen and Mike Kellogg of the
Bradforth group at USC.

B-2 B-3
Figure 5.1. Space-filling models of B-2 and B-3, which are seen to be conformationally restricted to a low-energy
mutually orthogonal geometry due to steric interactions between the 4-methyl substituents.

Golden
140
chromophoric coupling between the two chromophores in the dimer; alternatively, it may be
due to poor solubility causing the observation of nanocrystalline aggregates.
5.4 Synthesis of Zinc Dipyrrins
Zinc dipyrrins were synthesized using identical dipyrrin cores to those of the BODIPY
dimers above (Figure 5.3). For synthetic ease, the meso position of the dipyrrin was appended
with a phenyl ring in Z-2, Z-3, and Z-4. There has been some discussion in the literature about
the rotation of the phenyl ring on a zinc dipyrrin leading to non-radiative decay of the
photoexcited state.
25
In these systems, the 4-position on the pyrrole was substituted with a
methyl group or with a benzene ring, which effectively blocks phenyl ring rotation. In Z-1,
where there is no alkyl group to block meso aryl rotation, mesityl was used in lieu of phenyl to
limit rotational deactivation.
Previously published syntheses of zinc dipyrrins (ZnDIPY)2 materials resulted in
generally poor yields (>15%) of alkyl-pyrrole-derived zinc dipyrrin complexes,
18
while the
synthesis of Zn(benzoDIPY)2 proved impossible by this method due to instability of the zinc
complex toward any conditions for the oxidation of the protected isoindoles (benzopyrroles)
300 350 400 450 500 550
0.0
0.5
1.0
Absorption (a.u.)
Wavelength (nm)
MeCyHex
toluene
THF
DCM
B-2
300 350 400 450 500 550
0.0
0.5
1.0
MeCyHex
Toluene
THF
DCM

Absorption (a.u.)
Wavelength (nm)
B-3

Figure 5.2. Absorption spectra of B-2 and B-3.

Molecular Design for Organic Photovoltaics
Chapter 5 | 141
to their fully conjugated form in the final material.
28-31
BODIPYs, however, are relatively easy
to synthesize in decent yields (>35%) from pyrrolic starting materials, and recent publications
have shown that they can be quantitatively deborylated to yield the free base dipyrrin.
‡32-35

Dipyrrin free bases are unstable to oxidation, so this method provides not only a more efficient
method to yield the free base from the expensive starting pyrroles, but the BODIPY itself
serves as a valuable protecting group for the dipyrrin. Furthermore, purification of dipyrrin
from direct condensation to aldehyde followed by oxidation is difficult, as neither the
oligopyrrolic side products nor the dipyrrins themselves run on a silica gel column. Rather than
purify the material at this stage, borylation is completed in situ to generate the BODIPY, which
is readily purified by silica gel flash column chromatography. This shelf stable material can
be kept for years without decomposition and deborylated when the dipyrrin is needed. The
dipyrrin, quantitatively produced by one of several literature methods,
32-35
is purified by
solvent phase separation followed by recrystallization. For the purposes of this work,
deborylation with the Lewis acid ZrCl4 was utilized.
32
Coordination to zinc may then proceed,

‡
Ming Hau synthesized many of the BODIPYs used for this work as part of his undergraduate research in the
Thompson Lab.

Figure 5.3. Structures of zinc dipyrrins synthesized for this work.

Golden
142
again at quantitative yields, under mild, room temperature conditions with the addition of an
equivalent of non-nucleophilic base such as diisopropylethylamine in the presence of zinc
acetate dihydrate in a few minutes. The zinc dipyrrin is itself unstable to acidic conditions, and
silica gel is sufficiently acidic to dezincate the dipyrrin. Purification of the zinc dipyrrin may
be accomplished by solvent phase separation, followed by running the organic residues through
a short, basic alumina plug, and recrystallization with aprotic solvents. Though lacking
somewhat in atom economy, this method provides exceptionally pure zinc dipyrrins with a
reduction in purification complexity compared to direct synthetic methods, in yields (>35%)
which are better than double those of direct methods (Scheme 5.2). The BODIPY-protected
pathway was the only method capable of producing the benzannulated derivative Z-4; notably,
the metalation step in the synthesis of Z-4 proceeds in quantitative yield without the need for
Scheme 5.2. BODIPY is used as a protected intermediate for the synthesis of zinc dipyrrins, as it is both more
stable and more easily purified than the dipyrrin ligand. Yields are after purification. In the direct synthesis, the
intermediate dipyrromethane and dipyrromethene are not purified.

Molecular Design for Organic Photovoltaics
Chapter 5 | 143
added base. Due to the increased acidity of the N-H bond in the benzannulated dipyrrin relative
to the alkyl dipyrrins, it seems that the acetate ion is sufficiently basic to affect deprotonation
of the ligand and facilitation of zinc coordination to form the Zn(benzoDIPY)2 complex Z-4.
This work represents the first reported synthesis of Z-4. A fifth Zn(DIPY) derivative,
Zn(MesDIPY) Z-0, wherein the pyrroles are unsubstituted, was also synthesized by this
method.
5.5 Structural and Photophysical Characterization of Zinc Dipyrrins
Initial photophysical characterization of the Zn(DIPY) systems was performed using a
combination of UV-vis absorption, fluorescence emission, photoluminescent quantum yield,
and time correlated single-photon counting spectroscopies. The UV-vis absorption and
emission spectra are seen to bathochromically shift down the series Z-2 < Z-3 < Z-4, with
increasing substitution on the pyrroles. The most pronounced bathochromic shift (80 nm
relative to Z-1) is observed in the benzannulated derivative Z-4. UV-vis absorption spectra in
all Zn(DIPY) systems remained generally consistent over a wide range of solvent polarities
(from cyclohexane to acetonitrile). In Figure 5.3, absorption and emission spectra in
(methyl)cyclohexane (black traces) and dichloromethane (red traces) are depicted; in the case
of Z-4, the polar solvent shown is 2-methyltetrahydrofuran (green trace). It can be seen
throughout the series that the more polar solvent induces a narrowing of the emission spectrum
compared to cyclohexane and a subtle hypsochromic shift of the emission maximum which is
most pronounced in Z-3. In both Z-2 and Z-3, the emission spectrum tails to the red edge of
the spectrum. Due to the very low  PL values (Table 5.1) for both of these complexes, it is
unclear whether this tailing is weak emission from a CT state or is an artefact.
Golden
144
Both the Z-1 and Z-4 complexes have intense fluorescence emission in nonpolar
cyclohexane solvents (66% and 24%, respectively) which is markedly decreased in the more
polar tetrahydrofuran solvents (9% and 4%, respectively) and in more polar dichloromethane
and acetonitrile solvents is beneath the limit of detection for the integrating sphere used to
acquire these measurements. Interestingly, both Z-2 and Z-3 are essentially non-emissive (  PL
<0.005) in all solvents. It has long been assumed that in phenyl-substituted zinc dipyrrins, aryl
rotation leads to rapid non-radiative deactivation from the S1 excited state.
25
However, in both
Z-2 and Z-3, phenyl rotation is blocked by methyl groups appended to the 4-positions of the
300 400 500 600 700
0.0
0.5
1.0

Absorption / Emission (a.u.)
Wavelength (nm)
Z-1
300 400 500 600 700
0.0
0.5
1.0

Absorption / Emission (a.u.)
Wavelength (nm)
Z-2
300 400 500 600 700
0.0
0.5
1.0
Absorption / Emission (a.u.)
Wavelength (nm)
Z-3
300 400 500 600 700
0.0
0.5
1.0

Absorption / Emission (a.u.)
Wavelength
Z-4

Figure 5.3. Absorption (solid) and emission (dashed) spectra of zinc dipyrrins in (methyl)cyclohexane (black),
dichloromethane (red), or 2-methyltetrahydrofuran (green). Emission spectra in polar solvents are narrowed and
blue-shifted relative to cyclohexane solvents.

Molecular Design for Organic Photovoltaics
Chapter 5 | 145
pyrroles (Figure 5.4). Further, the monomeric BODIPY analogues of Z-2 and Z-3 are highly
emissive in all solvents ( PL>0.8) despite the presence of the meso-phenyl group; it is unclear
why a change in the chelating group of otherwise identical chromophores would affect the
bond rotation dynamics to such a dramatic extent.
36

The geometry optimized gas phase structures of Z2, Z3, and Z4 were calculated at the
B3LYP / LACVP** level of theory and the dihedral angles between the meso phenyl rings and
the dipyrrin were measured; the dihedral angles between dipyrrins were measured as well, and
both are reported in Table 5.2. Z-4 behaves as what might be considered a typical SBCT
chromophore, in that it has a reasonably high PL and single-exponential fluorescent decay
dynamics in nonpolar solvents, while in increasingly polar solvents, the PL decreases and a
fast (sub-nanosecond) excited state decay process is observed in competition with S1→S0
decay (Table 5.1). As such, Z-4 can be used as a standard by which the effect of phenyl rotation
and interchromophoric coupling can be measured. The dihedral angles between phenyl and
dipyrrin  Ph-DIPY and between dipyrrin and dipyrrin DIPY-DIPY in Z-4 are close to ideal (95.5º

Z-2 Z-3 Z-4
Figure 5.4. Space-filling models calculated for Z-2, Z-3, and Z-4 in the gas phase (B3LYP / LACVP**)
Golden
146
and 96.5º, respectively), where mutual orthogonality (90º) is the preferred orientation. In Z-2
as well, these values are close to the ideal (95.6º and 90.7º). In Z-3, however, where the 3-
positions of the pyrroles are substituted with a bulky ethyl group, there is significant distortion
from mutual orthogonality with respect to both dihedral angles; first, the Ph-DIPY is twisted at
a 110.5º angle, while the dipyrrins themselves are twisted to an even more extreme degree
(115.8º). Ground-state geometry is not the only metric by which non-radiative decay pathways
can be estimated. The relative freedom of bond rotation should also be considered. To this end,
the interatomic distance between the phenyl ring and the nearest atom on the adjacent R-group
(carbon in the case of Z-2 and Z-3 and hydrogen in the case of Z-4) was measured. It is seen
that the Z-2 and Z-3 derivatives have a longer interatomic distance (3.02 and 3.03 Å) between
phenyl ring and alkyl group than does Z-4 (2.62 Å). These factors together suggest that meso-
Table 5.1. Solvent dependent steady-state photophysical properties of zinc dipyrrins. Lifetimes were fit to the
minimum number of exponentials to provide 
2
values under 1.3. Lifetimes faster than the instrument response
function (0.2 ns) of the TCPSC spectrophotometer are not reported (--). Values which were not recorded are
marked n.r.

Solvent absorption
 max (nm)
emission
 max (nm)
 PL
(%)
  
(ns)
Z-0 MeCyHex 37 2.43
MeTHF 4 n.r.
MeCN <1 1.03
Z-1 CyHex 493 506 66 4.8
Toluene 495 509 19 n.r.

THF 493 507 9 n.r.
DCM 493 508 / 650 <1 2.5 (18%) / <0.2 (82%)
Z-2 MeCyHex 486 511 <0.5 1.1 (20%) / 4.6 (81%)
Toluene 489 509 <0.5 0.1 (82%) / 3.6 ns (18%)
THF 485 507 <0.5 2.3 (23%) / 8.7 (44%)
DCM 485 512 <0.5 2.7 (50%) / 7.5 (37%) / 0.14 (13%)
Z-3 MeCyHex 504 531 <0.5 --
Toluene 507 537 <0.5 --
THF 503 541 <0.5 --
DCM 504 515 <0.5 0.52 (40%) / 4.3 (60%)
Z-4 MeCyHex 570 584 24 1.83
MeTHF 572 578 4 0.88 (20%) / 2.43 (80%)

Molecular Design for Organic Photovoltaics
Chapter 5 | 147
aryl bond rotation is not necessarily the
limiting factor in the emissivity of
Zn(DIPY)s; rather, it may be that bond
vibration is the cause of non-radiative decay
in these systems. In the structurally rigid benzannulated derivative Z-4, twisting modes are
largely prevented by steric interaction and the larger resonant -system.
5.6 Excited State Decay Dynamics in Dipyrrin Dimers
A thorough analysis of molecular dynamics in both the B and Z classes of dipyrrin
dimers structures should be performed to probe the effects of bond torsion and bond vibration
on the energy of conformers. It may be that in the zinc dipyrrins, bond vibrational modes are
responsible for the ultrafast excited state decay dynamics. In preparation for such an
examination, the orbital characteristics of the two dimer classes are compared (Figure 5.5). It
can be seen in the B-series that the consolidation of LUMO density at the meso-position leads
to chromophoric coupling for any dipyrrin-dipyrrin dihedral angle which is not strictly
orthogonal. Such coupling is strictly forbidden in the Z-series due spatial separation between
chromophores (through zinc), and due to an inversion of the relative configuration of the
chromophores (where in the B-series, nearest chromophoric interactions are between meso
positions, in the Z-series, nearest interactions are through methyl substituents, wherein there
is negligible frontier molecular orbital density. In Z-3, some LUMO density is observed on the
phenyl rings which is not observed in either of the other two Z-series dimers. These
distinctions each will lead to variable excited state decay processes, whereby the rates and
efficiency of symmetry-breaking charge transfer can be tuned based on chromophoric structure
and the degree of coupling between dimers. Experimental probing of the time-resolved
Table 5.2. Dihedral angles  between meso-phenyl
rings and dipyrrin core and between both dipyrrins,
and distance between 4-substituent (R) and meso-
phenyl ring (Ph) in Zn(Ph-DIPY) complexes.
 Ph-DIPY  DIPY-DIPY r R-Ph (Å)
Z-2 95.6 90.7 3.02
Z-3 110.5 115.8 3.03
Z-4 95.5 96.6 2.62

Golden
148
transient decay dynamics the Z-series was begun using a combination of
spectroelectrochemistry (Figure 5.6) and transient absorption spectroscopy (Figure 5.7).
These data are published for B-1 and, with the new procedure allowing for rapid and large-

Figure 5.5. HOMO (solid) and LUMO (mesh) densities in B-series and Z-series 2, 3, and 4 dipyrrin dimers.
Molecular Design for Organic Photovoltaics
Chapter 5 | 149
scale synthesis of bis-BODIPYs, comparative studies with B-2, B-3, and B-4 are underway.
16,
27
Preliminary data for the Z-series are depicted in Figures 5.6 and 5.7.
350 400 450 500 550 600 650
-1.0
-0.5
0.0
0.5
Abs. (a.u.)
Wavelength (nm)
0
5
10
15
20
Z-1 Spectroelectrochemical Oxidation
350 400 450 500 550 600 650
0.0
0.5
1.0
Abs. (a.u.)
Wavelength (nm)
0
5
10
15
20
Z-0 Spectroelectrochemical Reduction

350 400 450 500 550 600 650
0.0
0.5
1.0

Abs. (a.u.)
Wavelength (nm)
0
5
10
15
20
Z-2 Spectroelectrochemical Oxidation

350 400 450 500 550 600 650
-1.0
-0.5
0.0
Abs. (a.u.)
Wavelength (nm)
0
5
10
15
20
Z-2 Spectroelectrochemical Reduction

400 500 600 700
-1.0
-0.5
0.0
0.5

0
5
10
15
20
Abs. (a.u.)
Wavelength (nm)
Z-4 Spectroelectochemical Reduction

Figure 5.6. Specroelectrochemical traces for Z-1, Z-2, and Z-4 under oxidizing (left) and reducing (right)
conditions. Spectra are shown as a bleach from neutral compound absorption under chronoamerometric conditions
at 100 mV overpotentials to the respective redox features. Spectra were recorded at five-minute intervals over a
period of 20 minutes.

Golden
150
Ultrafast transient absorption spectra were acquired for solutions of Z-0, Z-1, and Z-4
in cyclohexane and a more polar solvent (THF in the case of Z-0 and Z-1, and toluene in the
case of Z-4). In both systems, cyclohexane shows standard A→B decay, consistent with direct
relaxation via radiative and non-radiative processes from the S1 to the S0 state. In the more
350 400 450 500 550 600 650
-1.0
-0.5
0.0

Norm. Abs. (a.u.)
Wavelength (nm)
1 ps
10 ps
100 ps
500 ps
900 ps
Steady State
Abs.
Z-0 in Cyclohexane
350 400 450 500 550 600 650
-1.0
-0.5
0.0
Norm. Abs. (a.u.)
Wavelength (nm)
1 ps
10 ps
100 ps
500 ps
900 ps
Steady State
Abs.
Z-0 in THF

350 400 450 500 550 600 650
-1.0
-0.5
0.0
Norm. Abs. (a.u.)
Wavelength (nm)
1 ps
10 ps
100 ps
500 ps
900 ps
Steady State
Abs.
Z-2 in Cyclohexane
350 400 450 500 550 600 650
-1.0
-0.5
0.0
Norm. Abs. (a.u.)
Wavelength (nm)
1 ps
10 ps
100 ps
500 ps
900 ps
Steady State
Abs.
Z-2 in THF

400 500 600 700
-1.0
-0.5
0.0

Abs. (a.u.)
Wavelength (nm)
0.15 ps
1 ps
10 ps
100 ps
500 ps
900 ps
Z-4 in Cyclohexane

400 500 600 700
-1.0
-0.5
0.0
Abs. (a.u.)
Wavelength (nm)
0.15 ps
1 ps
10 ps
100 ps
500 ps
900 ps
Z-4 in Toluene

Figure 5.7. Femtosecond transient absorption spectra of Z-0, Z-2, and Z-4 in various solvents. Spectral features
associated with SBCT are indicated with black arrows.

Molecular Design for Organic Photovoltaics
Chapter 5 | 151
polar THF, Z-0 shows distinctive features associated with formation of the SBCT state (black
arrows).
18
Z-4 is not stable in coordinating solvents such as tetrahydrofuran, so toluene was
utilized to examine the lower limit of the SBCT threshold in this material. It can be seen that
there is a small growth of an absorptive feature between 600 and 650 nm within a picosecond
of excitation, which overlaps with the stimulated emission feature. This feature is indicative of
SBCT state formation in toluene, and global analysis fitting suggest that in toluene, while
SBCT is overserved, it is slower than back-electron transfer. This observation makes Z-4 an
interesting candidate for a solvent-polarity dependent study in order to find the polarity at
which kSBCT = kBET.
§
The time constants for emission (  em), SBCT ( SBCT), back-electron
transfer ( BET) and recombination ( rec) in Z-0 and Z-4 were calculated by global analysis fitting
of the transient absorption spectra and are presented in Table 5.3.
As observed from the steady-state photophysical studies described above, Z-2 has an
exceptionally short transient absorption lifetime. In both cyclohexane and THF, complete
relaxation from the excited state to the ground state is observed within 1 ns. In both solvents
as well, this process is characterized by the growth and decay of multiple unique features, some

§
This suggestion was offered by Laura Estergreen and this work represents progress in an ongoing collaboration
between the Thompson and Bradforth groups at USC.
Table 5.3. Time constants for excited-state decay processes in zinc dipyrrins.

τ em τ ICT τ SBCT τ BET τ rec
Z-0 Cyclohexane 2.4 ns N/A -- -- --
THF 3 ns N/A 17 ps 27 ps --
Z-2 Cyclohexane 300 ps 700 fs 4 ps 1 ps 10 ps (1)
THF 1 ns 400 fs 3 ps -- 4 ps (1) / 110ps(2)
Z-4 Cyclohexane 2 ns N/A -- -- --
Toluene 2 ns N/A 11 ps 5 ps --

Golden
152
of which cross the baseline more than once. This complicated excited state decay behavior is
not consistent with simple S1→S0 radiative and vibronic decay processes. Global analysis
fitting indicates instead that there are three involved transitions, from an initial excited state
350 400 450 500 550 600 650
-100
-80
-60
-40
-20
0
20
Abs. (mOD)
Wavelength (nm)
0.3 ps
1 ps
5 ps
10 ps
20 ps
30 ps
40 ps
50 ps
60 ps
70 ps
80 ps
90 ps
100 ps
Z-2 in cyclohexane
a)

350 400 450 500 550 600 650
-120
-80
-40
0
40 c)

Amplitude (a.u.)
Wavelength (nm)
ES1
ES2
ES3
Z-2 in Cycloxexane
1E-3 0.01 0.1 1 10 100
0.0
0.2
0.4
0.6
0.8
1.0
Concentration
Time (ps)
ES1
ES2
ES3
Z-2 in Cyclohexane
d)

350 400 450 500 550 600 650
-120
-80
-40
0
40

Amplitude (a.u.)
Wavelength (nm)
ES1
ES2
ES3
Z-2 in THF
e)
1E-3 0.01 0.1 1 10 100
0.0
0.2
0.4
0.6
0.8
1.0
Concentration
Time (ps)
ES1
ES2
ES3
Z-2 in THF
f)

Figure 5.8. Z-2 (a) TA spectrum in cyclohexane over the first 100 ps after photoexcitation, (b) Jablonski diagram
depicting state transitions (energy axis is arbitrary), and (c-f) species-associated decay spectra.

Molecular Design for Organic Photovoltaics
Chapter 5 | 153
(ES1), through two other excited states (ES2 and ES3), then to the ground state (GS)
ES1→ES2→ES3→GS, and that this is true in both nonpolar cyclohexane and in toluene. This
process occurs largely within the first 100 ps upon photoexcitation; TA spectra and species-
associated decay spectra illustrating the excited state decay process in Z-2 within this
timeframe are depicted in Figure 5.8. The photoexcitation/excited state decay process in Z-2
is tentatively assigned as depicted in the Jablonski diagram in Figure 5.8: photoexcitation
results in the formation of ES1, a locally excited state on a single dipyrrin molecule,
characterized by a - * type electronic excitation. Upon formation of this state, an
intramolecular charge transfer state (ES2), which may involve transfer of partial charge density
from dipyrrin to phenyl. Evidence for this possibility is given by the observation of LUMO
density on the phenyl rings in the geometry-optimized structures of Z-2 and Z-3. Interestingly,
when mesityl is substituted in leiu of phenyl, no LUMO density is observed on the meso-aryl
ring. This indicates that meso phenyl substitution in zinc dipyrrins may induce electronic
deactivation of the ES1, rather than vibronic ralaxation, as has been suggested previously.
25

Upon formation of this partial ICT state, SBCT state formation (ES3) may follow rapidly,
involving charge transfer from dipyrrin to dipyrrin. The resulting charge separated species is
evidently unstable, and recombination follows rapidly (Table 5.3).
The respective roles of electronic transitions and geometric distortions in the excited
state decay mechanisms of Z-2 and Z-3 can be distinguished given the synthesis of a
thoughtfully constructed heteroleptic zinc dipyrrinato complex using an ancillary ligand which
will maintain the same geometric configuration about the zinc atom. Ideally this ligand will
have the same ligand field strength as the dipyrrin itself, and be photophysically innocent. One
potential ligand is bis(oxazoline), which is very similar in geometry to the dipyrrin. Indeed,
Golden
154
stable, highly emissive ( PL >70%) heteroleptic zinc dipyrrinato bis(oxazoline) complexes
have been recently reported in the literature.
37
This study utilized meso-mesityl dipyrrin
ligands; it would be instructive to compare meso-mesityl and meso-phenyl heteroleptic
derivatives of Z-2 and Z-3 in order to discriminate between the effects of SBCT and excited
state distortion.
**

5.7 Aggregation Induced Emission (AIE) in Zinc Dipyrrins
In the process of preparing solutions of Z-2 for photophysical analysis, it was observed
that crystalline Z-2 is emissive. This indicates that a rigidochromic effect prevents or reduces
the rapid excited state decay process which is observed in Z-2 in solution. A brief study of
aggregation-induced emission (AIE) in Z-2 was performed to observe the effects of
aggregation on the excited state decay of Z-2. Aggregation effects on emission in colloids and
thin films were observed (Figure 5.9). In the colloid study, a sample of Z-2 in acetonitrile was
prepared and compared against acetonitrile/water mixtures at the same analyte concentration.
In predominately aqueous solutions, the emission is observed to markedly red-shift, and a
narrow emission line (approximately 1.75 times the emission intensity compared to acetonitrile
solution) grows in. Note that the very narrow band at 625 nm in the 25:75 acetonitrile/water
mixture is assigned to raman scattering. Excitation spectra acquired from the observation of
both the 600 nm and the 700 nm emission features showed the same species is responsible for
both emission bands. The same emission features are observed in thin films spun-cast from
dichloromethane solution. The AIE emission features are markedly different from the emission
features of monomeric dipyrrin chromophores, which tend to be narrow and vibronic. It

**
To this end, dipyrrins 2 and 3 have been synthesized and a collaboration has been fostered with Savannah
Kapper of the Thompson Lab for the synthesis of a 2-unsubstituted bis(oxazoline) ligand.
Molecular Design for Organic Photovoltaics
Chapter 5 | 155
appears from these data that in the solid state, Z-2 forms J-aggregates.
38
The presence of two
distinctive AIE features suggests that J-aggregation may be efficient in two different
dimensions. The growth of X-ray quality crystals was attempted from hot acetone solution, but
solvolysis occurred during crystal growth. After this study was performed, another study
examining AIE effects of J-aggregation in zinc dipyrrins was published in the literature.
39

While that study reported  PL efficiencies up to 3%, the 6.8%  PL efficiency (Table 5.4)

500 550 600 650 700 750 800 850
0.0
0.5
1.0
1.5

Emission (a.u.)
Wavelength (nm)
100:0
85:15
65:35
45:55
25:75
5:95
Obs. 600
Obs. 700
Ex. 440
400 500 600 700 800
0.0
0.5
1.0

Excitation & Emission (a.u.)
Wavelength (nm)

Figure 5.9. Top: microscope photographs of Z-2 polycrystalline powder under visible light illumination (a),
and under monochromatic green-light excitation (b) and crystals under visible light illumination (c) and
monochromatic green-light excitation (d), (e). Photograph of Z-2 in hexanes solution under UV handlamp
illumination (f). Bottom: AIE emission spectra of Z-2 in solution (left) and thin films (right). Colloids were
prepared by increasing the water content in acetonitrile : water solutions with constant analyte concentrations.
Thin films were spun-cast from dichloromethane solution.

Golden
156
observed here appears to be a record for solid-state emission from a zinc dipyrrin; the narrow
700 nm emission band observed in this study appears to be unique to the Z-2 complex. Z-3 has
also shown AIE, though a quantitative analysis of emission from Z-3 aggregates has not yet
been performed.
5.8 Comparative Study of Dipyrrin Dimers – The Future of SBCT and its Application to
Organic Photovoltaics
The study of both the B and Z series of dipyrrin dimers has given rise to many unique
observations which are not yet fully described. Complete characterization via steady-state and
time-resolved spectroscopies, molecular dynamics, x-ray crystallography, electrochemistry,
and spectroelectrochemistry will be needed to establish a firm understanding of the interplay
between state energies, steric interaction, bond rotation and vibration, and wavefunction
overlap in these systems. With the new syntheses described for the preparation of both dipyrrin
dimer classes, many finely tuned derivatives can be synthesized easily and at large scales in
order to complete this complex study.
Although the molecular structures of dipyrrin dimers are eminently tunable, allowing
for precise control over steric factors and rotational freedom which are beneficial for the study
of SBCT mechanistic pathways, the electrochemical potentials in alkyl-substituted dipyrrins
Table 5.4. Lifetime and photoluminescent quantum yield measurements for colloidal preparations of Z-2.
% MeCN % H2O 
1
(ns) 
2
(ns) 
3
(ns) 
2

PL
(%)
100 0 1.6 (3%) 3.4 (10%) 7.4 (86%) 1.00 0.0
85 15 1.9 (3%) 4.4 (21%) 7.5 (76%) 1.02 0.1
65 35 1.8 (3%) 4.7 (30%) 7.7 (67%) 1.01 0.1
45 55 0.71 (1.2%) 4.1 (35%) 7.5 (64%) 1.01 0.1
25 75 0.70 (4%) 4.1 (38%) 7.8 (58%) 1.02 6.8
5 95 0.56 (6%) 3.5 (43%) 8.3 (51%) 1.01 6.8

Molecular Design for Organic Photovoltaics
Chapter 5 | 157
are shallow, making most dipyrrin derivatives
ill-matched for application to OPVs with
fullerene-type acceptors. Some adjustment of
electrochemical potentials in dipyrrins is
possible, however. In addition to
perchlorination, which was performed for the
synthesis of ZCl, a perchlrorodipyrrin
analogue of Z-0, the electrochemical
potentials of zinc dipyrrins can be deepened by meso-aza substitution. In addition to
eliminating problems stemming from meso-aryl rotation and twisting modes, this methodology
also decreases the LUMO energy, affecting a significant bathochromic shift in absorption and
emission. The development of zinc aza-dipyrrin structures is actively underway as part of an
ongoing research project.
††
One such structure, Zn(2-phenyl-4-methyl-azaDIPY)2, has been
synthesized as part of this ongoing effort (Figure 5.10). Next steps involve the synthesis of a
zinc aza-dipyrrin with alkyl substitution at the 2-pyrrolic positions, to control for aryl coupling
effects between adjacent dipyrrin ligands.
In the meantime, some work has been accomplished toward the fabrication of a series
of non-fullerene bulk-heterojunction small molecule/polymer OPVs containing zinc dipyrrin
donors or acceptors paired against an energy-matched polymer acceptor or donor. The device

††
This work has been ongoing in collaboration with Ming Hau as part of his undergraduate research experience.
300 400 500 600 700
0.0
0.5
1.0

Norm Abs / Em
Wavelength (nm)
Absorption
Emission

Figure 5.10. Absorption and emission spectra for a
zinc aza-dipyrrin.

Golden
158
structures are depicted in Figure 5.11. Black lines show approximate work functions of ITO
and aluminum. Colored bars indicate HOMO and LUMO orbitals of active layer materials. Z-
4 (purple) has relatively deep HOMO and LUMO energies relative to alkyl dipyrrins; even so,
is LUMO is 2 eV higher in energy than that of C60 fullerene. Thus, to mitigate massive
energetic losses upon charge transfer from donor to acceptor, a non-fullerene type acceptor,
PCTPTI (Figure 5.12), was selected. This polymeric material is better suited as an acceptor
for Z-4, although the ECT (1.4 eV) is 0.8 eV lower in energy than the Eg (the S1 energy of Z-4).
Perchlorination of the dipyrrin deepens the HOMO and LUMO levels by more than 2 eV
relative to alkyl dipyrrins, such that ZCl becomes better suited as an acceptor molecule than a
donor. It has already been shown that SBCT in ZCl significantly improves VOC in lamellar
OPVs.
1
In this study, the polymeric active layer necessitates solution processing, so BHJ
architectures using ZCl will be compared to the earlier lamellar devices. Whereas PCTPTI is
used as an acceptor for the shallow Z-4 SBCT donor, it can be used as a donor for the SBCT-
type acceptor ZCl. In this

Figure 5.12. Molecular structures of PCTPTI, ZCl, and PCDTBT-
DPP.
5.11

Figure 5.11. Zinc dipyrrin / polymer heterojunction
architectures utilizing Z-4 (purple) as a donor, PCTPTI (blue)
as either a donor or acceptor, ZCl (green) as a donor, and
PCDTBT-DPP (red) as an acceptor.

Molecular Design for Organic Photovoltaics
Chapter 5 | 159
architecture, the ECT, Eg, and driving force (Eg – ECT) are equal to those of the Z-4 device. The
primary difference between these two device architectures is SBCT occurs on the donor in the
Z-4 device, while it occurs on the acceptor in the ZCl device. In both of these architectures,
the lowest energy chromophore in the heterojunction is the dipyrrin; thus, FRET from polymer
to dipyrrin may occur faster than CT, such that the initially formed CT state will be an SBCT
state localized on the dipyrrin, followed by rapid intermolecular CT from dipyrrin to polymer.
The third architecture replaces the PCTPTI polymeric donor for a shallower polymeric donor,
PCDTBT-DPP. While the ECT in this architecture is the same as in both of the previously
described architectures, the Eg (2.0 eV) is slightly lower, as PCDTBT-DPP has a lower S1
energy than ZCl. Thus, excitons will be able to undergo FRET in this device from ZCl to
polymer. It may be that this architecture will mitigate the SBCT effect by controlling the
intermolecular CT event such that it originates from the polymer rather than the SBCT
material. These three architectures will allow for a highly controlled comparative study of
SBCT in the active layer of OPVs. Several device stacks have been prepared with the general
architecture ITO/PEDOT:PSS/BHJ/BCP/Al, with generally poor results. Devices with and
without the BCP buffer layer (14 nm) were prepared. The BHJ was alternatively prepared with
and without annealing at various temperatures. The highly crystalline zinc dipyrrins tend to
form islands when solution processed which lead to rapid device degradation and very low
power conversion efficiencies (under 0.01%), so the development of better solution processing
conditions will be necessary to complete this work.
‡‡

With the development of a large array of energy-tunable dipyrrin dimers, a thorough
and quantitative analysis of the mechanism of SBCT and its role in OPVs may be developed.

‡‡
This project has been performed in collaboration with Betsy Melenbrink in the Barry Thompson Lab at USC
as part of her PhD research and with the assistance of Ming Hau as part of his undergraduate research.
Golden
160
Molecular structure plays a key role in the energetic and electronic characteristics of these
materials, and with the new syntheses outlined in this chapter, it is hoped that the study of
dipyrrins will lead to breakthroughs in the development of SBCT-type materials for OPVs.
5.9 Isoindole Condensation on Oxalyl Chloride – Formation of a Diradicaloid
Fluorophore
The synthesis of a -benzannulated meso-bridged BODIPY dimer was attempted using
a 4,7-dihydroisoindole derivative as a protected substrate for condensation onto oxalyl
chloride, according to a modification of the synthesis outlined in Scheme 5.1. The reaction,
however, failed to produce the tetrapyrrolic derivative; instead, only two isoindoles reacted
with oxalyl chloride, producing a diketone intermediate as outlined in Scheme 5.3. Borylation
of this product in situ resulted in the formation of a 12-electron (4n) conjugated -system which
is likely stabilized by a diradicaloid resonance structure. Interestingly, both the oxygen atoms
and the nitrogen atoms chelate to BF2 groups, forming a di(boron difluoro) complex. Oxidation
Scheme 5.3. The synthesis of B-rad forms a 20 e
-
system which appears to be stabilized by a diradicaloid
resonance form.

Molecular Design for Organic Photovoltaics
Chapter 5 | 161
of the dihydroisoindole moieties to isoindole was performed without characterizing the
intermediate, producing B-rad. The conjugated heteroaromatic -system in B-rad lends itself
to any of three resonance structures. In one case, a positive charge is in resonance between the
two oxygen atoms, while another is in resonance between the two nitrogen atoms. In this case,
the permanent dipole moment of the structure is limited. However, the highly electronegative
oxygen atoms may also exert an electron with drawing effect, pushing both positive charges
onto the nitrogen atoms and forming an overall 20 e
-
(antiaromatic) system with a large
permanent dipole moment. In this situation, the antiaromaticity can be broken up by forming
a diradicaloid structure as shown in Scheme 5.3. Thus, two likely resonance forms can be
imagined for B-rad, which have a polarity dependent relationship. In nonpolar media, the 22 e
-

pseudo-aromatic system predominates, while in polar media, the diradicaloid form
predominates. Preliminary evidence for the diradical character of B-rad is given by the
1
H-NMR spectra (Figure 5.13) in DMSO-d6, which shows persistently broad features

Figure 5.13.
1
H-NMR spectra of B-rad in CDCl 3 (top) and DMSO-d 6 (bottom). In chloroform, the spectral features
are well resolved despite the very low solubility of B-rad in that solvent. In DMSO, however, where the solubility
of B-rad is high, features are broad. DMSO residual peak remains finely featured, indicating that B-rad in DMSO
may take on a paramagnetic form.

Golden
162
characteristic of a paramagnetic system. Filtration to remove any solid residues did not result
in better spectral resolution. Although the product is much less soluble in chloroform, NMR
spectra in CDCl3 were acquired which show well-resolved features, indicating that there may
be a solvent-dependent equilibrium between the antiaromatic and the diradicaloid resonance
forms.
Crystals grown from the diffusion of methanol into a saturated dichloromethane
solution of B-rad were analyzed by single-crystal X-ray diffraction (Figure 5.14) to elucidate
the structure of the high polarity
resonance form. Carbon-oxygen bond
lengths were measured at 1.33 Å,
supporting the hypothesis that this
structure is not the low-polarity
resonance form, which would have
C=O double bond character and
therefore much shorter bond lengths
(on the order of 1.21 Å). Bond length
and analysis of the X-ray structure
supports the hypothesis of diradical character in the ground state of B-rad.
40
It is first apparent
that there is a bond alternation pattern, indicating that the closed-shell, antiaromatic resonance
form contributes to the ground state structure. However, the C-C bond bridging the five-
membered boron dioxo ring and the seven-membered boron dinitro ring has a bond length
(1.49 Å) which is significantly lengthened relative to a typical C=C (1.34 Å), approaching that

Figure 5.14. ORTEP diagram of B-rad with bond lengths
labeled. Thermal ellipsoids shown for carbon (grey), nitrogen
(blue), oxygen (red), boron (pink), and fluorine (red) shown
at 50% probability. Protons are depicted as white spheres.
Molecular Design for Organic Photovoltaics
Chapter 5 | 163
of a single bond (1.54 Å). This bond lengthening indicates that the diradicaloid resonance form
significantly contributes to the ground state of B-rad.
The absorption and emission features of B-rad further indicate that there may be two
resonance forms in equilibrium (Figure 5.15). In both the absorption and emission spectra,
there is a distinctive growth of a new, red-shifted absorption feature in increasingly polar
solvents. In nonpolar toluene, the absorption manifold shows highly structured vibronic
features with a max at 550 nm. In increasingly polar media, a red shoulder is seen to grow in
at 610 nm. In dimethylformamide (DMF), the absorption band which predominates at 550 nm
in nonpolar media is entirely absent from the spectrum and is replaced by the red-shifted
absorption band at 610 nm. This trend is also followed in the emission spectra; further, the
emission spectrum is excitation wavelength dependent. When the blue feature is selectively
excited (ex. 480 nm), an emission band at 515-525 nm is observed in toluene, acetone, and
acetonitrile. In DMF, however, even at an excitation of 480 nm, the predominant emission
feature is blue-shifted to 556 nm, and an intense shoulder at 586 nm is observed, indicating
that only one form is present in this highly polar solvent. When the excitation wavelength is
400 500 600 700
0.0
0.5
1.0

Absorption (a.u.)
Wavelength (nm)
Toluene
DCM
Acetone
MeCN
DMF
B-4
500 550 600 650 700 750
0.0
0.5
1.0
Toluene 480
Acetone 480
MeCN 480
DMF 480
Acetone 520
MeCN 520
DMF 520
Emission (a.u.)

Wavelength (nm)
B-4

Figure 5.15. Emission spectra of B-4 were acquired using a monochromatic excitation source at 480 nm (solid)
and 520 nm (symbols). The combined absorption and emission spectra indicate that in this species there are two
distinct forms present in a solvent-dependent equilibrium.
Golden
164
changed to 520 nm, no emission is observed in toluene, and the predominant emission feature
in acetone, acetonitrile, and DMF is observed at 592 nm. These data suggest that there is a
solvent-dependent equilibrium between two species which have distinct absorption and
emission features. It is postulated here that the two forms are the closed-shell, 22 e
-
pseudo-
aromatic resonance form, and the diradical form depicted in Scheme 5.3. Toluene and DMF
represent opposite extremes of this equilibrium, wherein only one form is observed, while
acetone and acetonitrile show both conformers. Indeed, wavelength-dependent
photoluminescent lifetimes are also observed (Table 5.5), further supporting the hypothesis
that there is a solvent-dependent equilibrium between the two forms of B-rad. Another notable
and not yet understood anomaly in the photophysical behavior of B-rad is that in all solvents,
the max of the emission band is higher in energy than that of the absorption band. This structure
may represent a unique opportunity to examine the effects of solvent polarity on the
equilibrium between an antiaromatic and a diradical form of the same molecular structure.
Further analysis by EPR spectroscopy is recommended to confirm the hypothesis made based
on the NMR, X-ray, and photophysical spectroscopic characterization methods discussed here.
If the diradical form is indeed significantly contributing to the unique photophysical behavior
of B-rad, it may be an interesting candidate for analysis by magnetic field strength dependent
fluorescence emission spectroscopy.
Table 5.5. Integrated quantum yield and photoluminescent lifetime (ns) data of B-4 (observed at 510 nm and
600 nm).
510 nm 600 nm 

 1  2 
2
 1  2 
2
 PL
Toluene 1.97 (79) 4.3 (21) 1.1 0.66 (67) 2.39 (33) 1.0 6
Acetone 1.89 (77) 4.93 (23) 1.0 0.42 (57) 2.53 (43) 1.2 2
MeCN 2.11 (73) 5.81 (77) 1.0 0.50 (48) 3.17 (52) 1.3 2
DMF 0.50 (45) 3.23 (55) 1.3 0.64 (11) 3.38 (90) 1.0 8

Molecular Design for Organic Photovoltaics
Chapter 5 | 165
5.10 Experimental Methods
General Synthesis of bis-BODIPYs: A modified literature procedure was followed.
27
To an
oven-dried Schlenk flask filled with nitrogen was added dry dichloromethane (40 mL), the
starting pyrrole (20 mmol), and oxalyl chloride (5 mmol), turning the solution dark orange-
brown. After two hours, BF3OEt2 was added (6 mL) to the deep red-orange solution, followed
by dropwise addition of diisopropylethylamine (4 mL). The solution changed color to a vivid
red hue, and was stirred for a further three hours. After this time, the solution changed to a dark
yellow color, and solvent was removed by rotary evaporation. The solid residue was loaded
onto silica without further workup, then purified by silica gel flash column chromatography
with the gradient eluate 20%-60% dichloromethane in hexanes. This material was dissolved in
minimal dichloromethane and layered with hexanes to form yellow needles over two days
at -48 ºC.
B-2: 0.538 g (22% yield) vivid yellow-orange crystals.
1
H NMR (500 MHz, Chloroform-d)
δ 6.42 (s, 1H), 2.62 (s, 3 H), 2.58 (s, 4H).
13
C NMR (101 MHz, CDCl3) δ 166.31, 152.43,
149.32, 132.19, 124.88, 145.40, 15.30.
19
F NMR (376 MHz, CDCl3) δ -145.06 (dd, J = 26.6,
11.7 Hz).
11
B NMR (128 MHz, CDCl3) δ -0.28 (t, J = 13.9 Hz).
B-rad: To an oven-dried Schlenk flask filled with nitrogen was added dry dichloromethane
(20 mL), 2-methyl-4,7-dihydroisoindole (10 mmol), and oxalyl chloride (2.5 mmol). After two
hours, the solution was treated with 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (2.5 mmol)
and stirred a further 30 minutes. BF3OEt2 was added (3 mL), followed by dropwise addition of
diisopropylethylamine (2 mL). The reaction was quenched by the addition of 50 mL saturated
sodium sulfite solution, diluted with 100 mL dichloromethane, and extracted three times with
water. The organic layer was rinsed with brine, dried over sodium sulfate, and the solvent was
Golden
166
removed by rotary evaporation. The solid purple-red residue was loaded onto silica gel and
purified by silica gel flash column chromatography to yield the product as a vivid, deep red
polycrystalline powder, which was recrystallized from the diffusion of methanol into a
saturated dichloromethane solution, to yield the product as deep red needles.
1
H NMR (400
MHz, CDCl3) δ 8.59 (d, J = 8.1 Hz, 1H), 7.90 (d, J = 7.9 Hz 2H), 7.78 (ddd, J = 8.2, 7.2, 1.0
Hz, 2H), 7.56 (ddd, J = 8.1, 7.2, 1.0 Hz, 2H), 3.07 (t, J = 3.1 Hz, 6H).
13
C NMR (101 MHz,
CDCl3) δ 166.18, 154.95, 137.03, 134.45, 132.09, 128.45, 125.53, 124.31, 123.54, 29.68.
General deborylation procedure:
To an oven-dried Schlenk flask equipped with a reflux condenser was added BODIPY
(3.22 mmol) and 250 mL dry acetonitrile under a nitrogen atmosphere. Zirconium tetrachloride
was dissolved in 50 mL dry methanol under nitrogen and the zirconium solution was added
via cannula to the BODIPY solution. The solution was stirred at reflux for two hours, until
TLC (20% dichloromethane in hexanes) showed no remaining BODIPY. The solvent was
removed via rotary evaporation and the solid residue was dissolved in dichloromethane
solution (200 mL) and washed three times with water (150 mL), until the aqueous layer became
colorless. The organic layers were combined, dried over sodium sulfate, and the solvent was
evaporated by rotary evaporation to form a dull but deeply colored powder. The powder was
recrystallized by the slow diffusion of pentane into a saturated chloroform solution to form
deeply colored crystals at quantitative yields.
General synthesis of zinc dipyrrins:
The dipyrrin free base (0.54 mmol) was dissolved in 30 mL methanol and treated with
diisopropylethylamine (2 mL), forming a deeply colored dark solution. Zinc acetate dihydrate
(0.27 mmol) was added in one portion, whereupon brightly colored crystals began to
Molecular Design for Organic Photovoltaics
Chapter 5 | 167
precipitate from solution. The solvent was removed by rotary evaporation and the product was
loaded onto basic alumina and purified by passing through a basic alumina plug.
Z-0: 0.314 g, 99% yield, vivid yellow-orange crystals with green reflectance.
1
H NMR (400
MHz, CDCl3) δ 7.46 (d, J = 1.3 Hz, 4H), 6.97 – 6.89 (br. s, 4H), 6.57 (dt, J = 4.1, 1.2 Hz, 4H),
6.33 (dt, J = 4.2, 1.3 Hz, 4H), 2.37 (s, 6H), 2.21 – 2.13 (m, 12H).
Z-2: 0.345 g, 98% yield, vivid red-green crystals.
1
H NMR (400 MHz, CDCl3) δ 7.49 – 7.36
(m, 6H), 7.36 – 7.27 (m, 4H), 5.93 (d, J = 1.0 Hz, 4H), 2.03 (s, 12H), 1.28 (d, J = 0.8 Hz, 12H).
Z-3: 0.343 g, 63% yield, rose gold crystals.
1
H NMR (400 MHz, CDCl3) δ 7.43 (m, 6H), 7.36
– 7.31 (m, 4H), 2.25 (q, J = 7.6 Hz, 8H), 1.96 (s, 12H), 1.18 (s, 12H), 0.91 (t, J = 7.5 Hz, 12H).
13
C NMR (101 MHz, CDCl3) δ 155.74, 143.37, 140.98, 138.26, 135.37, 131.45, 129.88,
128.39, 127.76, 17.86, 14.99, 14.41, 12.74.
Z-4: The free base 2-methylisoindolylmethene (0.123 mmol) was dissolved in 30 mL acetone
in a vial. Zinc acetate dihydrate (0.617 mmol) was added in one portion and the vial was shaken
vigorously for three minutes, then stored at -48 C overnight. The residue was loaded onto basic
alumina and purified by passing through a basic alumina plug to form a violet powder. This
powder was recrystallized via the slow diffusion of hexanes into a saturated dichloromethane
solution, forming gold crystals (0.126 g, 100% yield).
1
H NMR (400 MHz, CDCl3) δ 7.76 –
7.64 (m, 4H), 7.64 – 7.57 (m, 4H), 7.56 – 7.48 (m, 4H), 7.01 (ddd, J = 7.8, 6.8, 1.0 Hz, 4H),
6.90 (ddd, J = 8.2, 6.8, 1.1 Hz,4), 5.84 (dt, J = 8.4, 0.9 Hz, 4H), 2.39 (s, 12H).
5.11 Bibliography
1. Bartynski, A. N.; Gruber, M.; Das, S.; Rangan, S.; Mollinger, S.; Trinh, C.; Bradforth,
S. E.; Vandewal, K.; Salleo, A.; Bartynski, R. A.; Bruetting, W.; Thompson, M. E., Symmetry-
Breaking Charge Transfer in a Zinc Chlorodipyrrin Acceptor for High Open Circuit Voltage
Organic Photovoltaics. J. Am. Chem. Soc. 2015, 137 (16), 5397-5405.
2. Mataga, N.; Yao, H.; Okada, T.; Rettig, W., Charge-transfer rates in symmetric and
symmetry-disturbed derivatives of 9,9'-bianthryl. J. Phys. Chem. 1989, 93 (9), 3383-3386.
Golden
168
3. Vauthey, E., Photoinduced Symmetry-Breaking Charge Separation. ChemPhysChem
2012, 13 (8), 2001-2011.
4. Grozema, F. C.; Swart, M.; Zijlstra, R. W. J.; Piet, J. J.; Siebbeles, L. D. A.; van
Duijnen, P. T., QM/MM Study of the Role of the Solvent in the Formation of the Charge
Separated Excited State in 9,9‘-Bianthryl. J. Am. Chem. Soc. 2005, 127 (31), 11019-11028.
5. Rettig, W.; Zander, M., Broken Symmetry in Excited States of 9,9′-Bianthryl. Ber.
Chem. 1983, 87 (12), 1143-1149.
6. Lueck, H.; Windsor, M. W.; Rettig, W., Pressure dependence of the kinetics of
photoinduced intramolecular charge separation in 9,9'-bianthryl monitored by picosecond
transient absorption: comparison with electron transfer in photosynthesis. J. Phys. Chem. 1990,
94 (11), 4550-4559.
7. Takaya, T.; Saha, S.; Hamaguchi, H.-o.; Sarkar, M.; Samanta, A.; Iwata, K., Charge
Resonance Character in the Charge Transfer State of Bianthryls: Effect of Symmetry Breaking
on Time-Resolved Near-IR Absorption Spectra. J. Phys. Chem. A 2006, 110 (13), 4291-4295.
8. Kovalenko, S. A.; Pérez Lustres, J. L.; Ernsting, N. P.; Rettig, W., Photoinduced
Electron Transfer in Bianthryl and Cyanobianthryl in Solution: The Case for a High-
Frequency Intramolecular Reaction Coordinate. J. Phys. Chem. A 2003, 107 (48), 10228-
10232.
9. Piet, J. J.; Schuddeboom, W.; Wegewijs, B. R.; Grozema, F. C.; Warman, J. M.,
Symmetry Breaking in the Relaxed S1 Excited State of Bianthryl Derivatives in Weakly Polar
Solvents. J. Am. Chem. Soc. 2001, 123 (22), 5337-5347.
10. Asami, N.; Takaya, T.; Yabumoto, S.; Shigeto, S.; Hamaguchi, H.-o.; Iwata, K., Two
Different Charge Transfer States of Photoexcited 9,9′-Bianthryl in Polar and Nonpolar
Solvents Characterized by Nanosecond Time-Resolved Near-IR Spectroscopy in the
4500−10 500 cm−1 Region. J. Phys. Chem. A 2010, 114 (22), 6351-6355.
11. Holman, M. W.; Yan, P.; Adams, D. M.; Westenhoff, S.; Silva, C., Ultrafast
Spectroscopy of the Solvent Dependence of Electron Transfer in a Perylenebisimide Dimer. J.
Phys. Chem. A 2005, 109 (38), 8548-8552.
12. Wu, Y.; Young, R. M.; Frasconi, M.; Schneebeli, S. T.; Spenst, P.; Gardner, D. M.;
Brown, K. E.; Würthner, F.; Stoddart, J. F.; Wasielewski, M. R., Ultrafast Photoinduced
Symmetry-Breaking Charge Separation and Electron Sharing in Perylenediimide Molecular
Triangles. J. Am. Chem. Soc. 2015, 137 (41), 13236-13239.
13. Sung, J.; Nowak-Król, A.; Schlosser, F.; Fimmel, B.; Kim, W.; Kim, D.; Würthner, F.,
Direct Observation of Excimer-Mediated Intramolecular Electron Transfer in a Cofacially-
Stacked Perylene Bisimide Pair. J. Am. Chem. Soc. 2016, 138 (29), 9029-9032.
14. Cook, R. E.; Phelan, B. T.; Kamire, R. J.; Majewski, M. B.; Young, R. M.;
Wasielewski, M. R., Excimer Formation and Symmetry-Breaking Charge Transfer in Cofacial
Perylene Dimers. J. Phys. Chem. A 2017, 121 (8), 1607-1615.
15. Giaimo, J. M.; Gusev, A. V.; Wasielewski, M. R., Excited-State Symmetry Breaking
in Cofacial and Linear Dimers of a Green Perylenediimide Chlorophyll Analogue Leading to
Ultrafast Charge Separation. J. Am. Chem. Soc. 2002, 124 (29), 8530-8531.
16. Whited, M. T.; Patel, N. M.; Roberts, S. T.; Allen, K.; Djurovich, P. I.; Bradforth, S.
E.; Thompson, M. E., Symmetry-breaking intramolecular charge transfer in the excited state
of meso-linked BODIPY dyads. Chem. Commun. 2012, 48 (2), 284-286.
Molecular Design for Organic Photovoltaics
Chapter 5 | 169
17. Liu, Y.; Zhao, J.; Iagatti, A.; Bussotti, L.; Foggi, P.; Castellucci, E.; Di Donato, M.;
Han, K.-L., A Revisit to the Orthogonal Bodipy Dimers: Experimental Evidence for the
Symmetry Breaking Charge Transfer-Induced Intersystem Crossing. J. Phys. Chem. C 2018.
18. Trinh, C.; Kirlikovali, K.; Das, S.; Ener, M. E.; Gray, H. B.; Djurovich, P.; Bradforth,
S. E.; Thompson, M. E., Symmetry-Breaking Charge Transfer of Visible Light Absorbing
Systems: Zinc Dipyrrins. J. Phys. Chem. C 2014, 118 (38), 21834-21845.
19. Moore, L. J.; Zhou, H.; Boxer, S. G., Excited-State Electronic Asymmetry of the
Special Pair in Photosynthetic Reaction Center Mutants: Absorption and Stark Spectroscopy.
Biochemistry 1999, 38 (37), 11949-11960.
20. Lathrop, E. J. P.; Friesner, R. A., Simulation of optical spectra from the reaction center
of Rb. sphaeroides. Effects of an internal charge-separated state of the special pair. The Journal
of Physical Chemistry 1994, 98 (11), 3056-3066.
21. Wasielewski, M. R., Energy, Charge, and Spin Transport in Molecules and Self-
Assembled Nanostructures Inspired by Photosynthesis. J. Org. Chem. 2006, 71 (14), 5051-
5066.
22. Yu, L.; Muthukumaran, K.; Sazanovich, I. V.; Kirmaier, C.; Hindin, E.; Diers, J. R.;
Boyle, P. D.; Bocian, D. F.; Holten, D.; Lindsey, J. S., Excited-State Energy-Transfer
Dynamics in Self-Assembled Triads Composed of Two Porphyrins and an Intervening
Bis(dipyrrinato)metal Complex. Inorg. Chem. 2003, 42 (21), 6629-6647.
23. Golden, J. H.; Estergreen, L.; Porter, T.; Tadle, A. C.; Sylvinson M. R, D.; Facendola,
J. W.; Kubiak, C. P.; Bradforth, S. E.; Thompson, M. E., Symmetry-Breaking Charge Transfer
in Boron Dipyridylmethene (DIPYR) Dimers. ACS Appl. Energy Mater. 2018.
24. Ivanov, A. I.; Dereka, B.; Vauthey, E., A simple model of solvent-induced symmetry-
breaking charge transfer in excited quadrupolar molecules. J. Chem. Phys. 2017, 146 (16),
164306.
25. Sazanovich, I. V.; Kirmaier, C.; Hindin, E.; Yu, L.; Bocian, D. F.; Lindsey, J. S.;
Holten, D., Structural Control of the Excited-State Dynamics of Bis(dipyrrinato)zinc
Complexes: Self-Assembling Chromophores for Light-Harvesting Architectures. J. Am.
Chem. Soc. 2004, 126 (9), 2664-2665.
26. Kusaka, S.; Sakamoto, R.; Nishihara, H., Luminescent Heteroleptic
Tris(dipyrrinato)indium(III) Complexes. Inorg. Chem. 2014, 53 (7), 3275-3277.
27. Wang, L.; Cao, J.; Wang, J.-w.; Chen, Q.; Cui, A.-j.; He, M.-y., Facile synthesis of
dimeric BODIPY and its catalytic activity for sulfide oxidation under visible light. RSC Adv.
2014, 4 (28), 14786-14790.
28. Chen, J. J.; Conron, S. M.; Erwin, P.; Dimitriou, M.; McAlahney, K.; Thompson, M.
E., High-Efficiency BODIPY-Based Organic Photovoltaics. ACS Appl. Mater. Interfaces
2015, 7 (1), 662-669.
29. Ni, Y.; Zeng, W.; Huang, K.-W.; Wu, J., Benzene-fused BODIPYs: synthesis and the
impact of fusion mode. Chem. Commun. 2013, 49 (12), 1217-1219.
30. 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. Chem. Eur. J. 2004,
10 (19), 4853-4871.
31. Tetsuo Okujima, G. J., Yusuke Hashimoto, Hiroko Yamada, Hidemitsu Uno, Noboru
Ono, Synthesis of 4,7-dihydro-2H-isoindole derivatives via diels-alder reaction of
tosylacetylene. Heterocycles 2006, 70 (1), 619-626.
Golden
170
32. Lakshmi, V.; Chatterjee, T.; Ravikanth, M., Lewis Acid Assisted Decomplexation of
F-BODIPYs to Dipyrrins. Eur. J. Org. Chem. 2014, 2014 (10), 2105-2110.
33. Lundrigan, T.; Cameron, T. S.; Thompson, A., Activation and deprotection of F-
BODIPYs using boron trihalides. Chem. Commun. 2014, 50 (53), 7028-7031.
34. Smithen, D. A.; Baker, A. E. G.; Offman, M.; Crawford, S. M.; Cameron, T. S.;
Thompson, A., Use of F-BODIPYs as a Protection Strategy for Dipyrrins: Optimization of
BF2 Removal. J. Org. Chem. 2012, 77 (7), 3439-3453.
35. Yu, M.; Wong, J. K. H.; Tang, C.; Turner, P.; Todd, M. H.; Rutledge, P. J., Efficient
deprotection of F-BODIPY derivatives: removal of BF(2) using Brønsted acids. Beilstein J.
Org. Chem. 2015, 11, 37-41.
36. Kuznetsova, R. T.; Aksenova, Y. V.; Tel’minov, E. N.; Samsonova, L. G.; Maier, G.
V.; Kopylova, T. N.; Yutanova, S. L.; Antina, E. V.; Berezin, M. B.; Guseva, G. B., Spectral,
Luminescent, Photochemical, and Laser Properties of a Series of Boron Fluoride Complexes
of Dipyrrolylmethenes in Solutions. Opt. Spectrosc. 2012, 112 (5), 746-754.
37. Kögel Julius, F.; Kusaka, S.; Sakamoto, R.; Iwashima, T.; Tsuchiya, M.; Toyoda, R.;
Matsuoka, R.; Tsukamoto, T.; Yuasa, J.; Kitagawa, Y.; Kawai, T.; Nishihara, H., Heteroleptic
[Bis(oxazoline)](dipyrrinato)zinc(II) Complexes: Bright and Circularly Polarized
Luminescence from an Originally Achiral Dipyrrinato Ligand. Angew. Chem. Int. Ed. 2015,
55 (4), 1377-1381.
38. Telfer, S. G.; McLean, T. M.; Waterland, M. R., Exciton coupling in coordination
compounds. Dalton Trans. 2011, 40 (13), 3097-3108.
39. Tsuchiya, M.; Sakamoto, R.; Shimada, M.; Yamanoi, Y.; Hattori, Y.; Sugimoto, K.;
Nishibori, E.; Nishihara, H., Bis(dipyrrinato)zinc(II) Complexes: Emission in the Solid State.
Inorg. Chem. 2016, 55 (12), 5732-5734.
40. Ni, Y.; Lee, S.; Son, M.; Aratani, N.; Ishida, M.; Samanta, A.; Yamada, H.; Chang, Y.-
T.; Furuta, H.; Kim, D.; Wu, J., A Diradical Approach towards BODIPY-Based Dyes with
Intense Near-Infrared Absorption around λ=1100 nm. Angew. Chem. Int. Ed. 2016, 55 (8),
2815-2819.

Molecular Design for Organic Photovoltaics
Chapter 6 | 171
Chapter 6

Energy and Charge Transfer Processes in Dipyrrin
and Porphyrin Materials for Organic Photovoltaics,
Photochemical Upconversion, and Photocatalytic
Water Splitting

6.1 Introduction to Multichromophoric Arrays in Organic Photovoltaics
Organic materials exhibit exceptional wavelength-specific absorptivity compared to
inorganic semiconductors; however, the absorption manifolds in small molecules tend to be
narrow, leading to significant losses in individual donor or acceptor contribution to JSC for
wavelengths outside of a 100-200 nm region of the UV-vis-NIR spectrum. There are two basic
approaches that have been reported for achieving broadband coverage of the solar spectrum
with an OPV. Multiple junction tandem OPVs that use complementary donor and/or acceptor
materials in separate active layers have been shown to provide broad spectral coverage and
high JSC values.
1
The tandem approach, however, significantly increases device engineering
complexity. A simpler approach is to build OPVs with multiple complimentary donors or
acceptors in a single heterojunction. Ternary OPVs have been demonstrated to give good
absorption from the UV into the NIR.
2
Ternary devices have been reported in which
polymeric
3-5
, molecular
6-9
or nanoparticle
10-12
sensitizers are incorporated into the donor or
acceptor layers of an OPV, leading to enhanced light absorption. In these ternary devices, a
sensitizer typically transfers energy via FRET or DET to the donor or acceptor, which then
Golden
172
initiates the charge transfer and separation processes at the donor-acceptor interface; in such a
system, the active wavelength range for the ternary device is the sum of those for the sensitizer,
donor and acceptor. An extension of the ternary approach to enhancing absorptivity is to
covalently link two or more small molecule chromophores with carefully matched state
energies, creating a multichromophoric array with broad spectral sensitivity which is capable
of efficiently transferring energy between chromophores. Such systems hold promise not only
in the field of organic photovoltaics, but also in photocatalytic water splitting and the
generation of solar fuels.
The following sections compare the use of a ternary blend photovoltaic comprising two
intensely absorbing chromophores, platinum tetraphenylbenzoporphyrin (PtTPBP) and boron
difluoro dipyrromethene (BODIPY), in the donor layer to an OPV incorporating a covalently
linked BODIPY-PtTPBP multichromophoric array as a donor material.
*13
This BODIPY-
PtTPBP multichromophoric array has been previously shown to rapidly equilibrate the exciton
between the porphyrin core and the pendant BODIPYs, and it is this feature which is expected
to lead to efficient exciton migration in bulk films, whereas many other multichromophoric
arrays tend to trap the exciton on the core, precluding exciton migration.
14

6.1.1 Absorptivity and Energy Transfer in BDP-Por
Porphyrins and closely related materials are ubiquitous in nature for light absorption
and have been incorporated into modern small molecule OPVs.
15-17
These materials have two
intense absorption peaks, a strong but narrow absorption in the blue region of the visible
spectrum, called the Soret band, and a second absorption in the red, called the Q band (Figure

*
This work was performed under the mentorship of Dr. Patrick Erwin as part of his dissertation research and is
published in Ref. 13. Portions of the text and some figures reported herein are duplicated from published materials.
Molecular Design for Organic Photovoltaics
Chapter 6 | 173
6.1). OPVs utilizing PtTPBP as donor
materials have been reported and
characteristic current-voltage (J-V)
and external quantum efficiency
(EQE) spectra of these devices are
depicted in Figure 6.1.
18
Porphyrin
OPVs have good photoelectric
performance characteristics such as
high FF when paired with C60
acceptors, but the absence of
significant absorptivity in the green to
orange portion of the visible spectrum,
where solar flux tends to be quite high,
negatively impacts the JSC. The EQE
of the illustrated PtTPBP/C60 device
architecture depicts that energy losses
stem from low power conversion
efficiencies between 500 and 600 nm.
Were it not for the strong C60
absorbance between 400-500 nm, the
spectral response would be even worse
than shown here.
a)
PtTPBP
R = Ph, BODIPY
BODIPY

400 500 600 700
0
10
20
30
40
0.0
0.2
0.4
0.6
0.8
1.0
EQE
External 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
b)
Absorption (a.u.)
PtTPBP Abs.
C
60
Abs.
k

300 400 500 600 700
0.0
0.5
1.0

BODIPY-PtTPBP
Blend
Normalized Absorption (a.u.)
Wavelength (nm)
c)
0.0
0.5
1.0
300 400 500 600 700

PtTPBP
BODIPY

Figure 6.1. (a) Molecular structures of PtTPBP and
BODIPY-PtTPBP. (b) External quantum efficiency and J-V
characteristics of a PtTPBP/C 60 OPV. (c) Top plot,
absorption spectra of PtTPBP (red) and BODIPY (blue)
shows that BODIPY and PtTPBP have complimentary
absorption manifolds and bottom plot, absorption of
multichromophoric array versus sum of normalized
BODIPY and PtTPBP spectra.

Golden
174
BODIPYs have been shown throughout this work to be particularly efficient absorbers
with high PLQYs; as a result of their photophysical properties, they have been incorporated
into OPVs with much success.
19-22
Fortuitously, the unsubstituted BODIPY absorption falls
directly between the Soret and Q-bands of the PtTPBP (Figure 6.1). The complementarity of
BODIPY and PtTPBP absorption provides an opportunity to design a multichromophoric array
with a broad absorption manifold which can be used as a donor material to increase the
photocurrent and power conversion efficiency relative to a simple PtTPBP based OPV. The
array itself is designed such that four BODIPY groups are appended at the meso position to a
single platinum porphyrin core at its four meso positions using a bridging phenyl group to
provide spatial separation between BODIPY and porphyrin.
14
The absorption spectrum of the
array is nearly identical to the arithmetic sum of the absorptions of the separate PtTPBP and
BDP chromophores, indicating that there is limited electronic coupling between the
chromophores, which is a product of the bridging phenyl group, which twists due to steric
factors such that it is orthogonal to both the porphyrin and the BODIPY groups. This new
molecule has a broadened absorption spectrum, with an AM1.5G photon absorption percentage
in solution 60% greater than that of PtTPBP alone.
14
The absorption spectra of the BODIPY-
PtTPBP film changes little from that in solution, with the film spectrum showing only a small
red-shift of both the BODIPY absorption (23 nm) peak and the Q-band of the porphyrin (13
nm) and a slight broadening of absorptive features relative to those in the solution spectrum,
which is attributed to increased -orbital overlap in the film.
In many multichromophoric arrays, energy is collected by peripheral (shell or
antennae) chromophores and funneled to a central core in a unidirectional manner,
23-28
thereby
isolating the exciton from its nearest neighbors and limiting both FRET and DET efficiency.
Molecular Design for Organic Photovoltaics
Chapter 6 | 175
BODIPY-PtTPBP has a molecular structure similar to other core-shell arrays, but the state
energy alignment is specifically tuned for bidirectional energy transfer. Consequently, after
excitation at the core, spin orbit coupling at the platinum porphryin leads to formation of a
triplet exciton on the picosecond time scale. This triplet then equilibrates between the
porphyrin (T1 = 1.62 eV) and the nearly degenerate BODIPY triplet (T1 = 1.64 eV) with a K eq
of 0.61.
14
Thus the triplet exciton is evenly distributed over the both the core and shell
chromophores. The singlet exciton formed at the BODIPY may either undergo FRET to the
lower energy Q-band of the porphyrin (whereupon formation of the triplet will rapidly follow),
or it may undergo FRET to an adjacent BODIPY moiety as a singlet exciton, where again,
close interaction with the porphyrin will lead to intersystem crossing to the porphyrin triplet
followed by exciton distribution over the core and shells moieties.
Scheme 6.1. A masked isoindole starting material was substituted in the present synthesis, allowing for ease
of purification of masked BDP-Por via column chromatography. This method obviates the need for a separate
oxidation step post-metalation, as oxidation to the final product BDP-Por is accomplished via a retro-Diels-
Alder mechanism during the platination step.

Golden
176
6.1.2 Synthesis of BDP-Por
The synthesis of BDP-Por was performed previously in the Thompson group, but was
prepared by a shorter synthetic route for the present study.
14
The 4,5,6,7-tetrahydroisoindole
starting material utilized in the previous method was replaced with a bicyclic “masked”
derivative, 4,7-dihydro-4,7-ethano-2H-isoindole, according to the simplified procedure for the
synthesis of benzoporphyrins proposed by Ono.
29
This synthetic route presents two
advantages; the first is that the masked porphyrin has increased solubility in organic solvents
and is more readily purified by column chromatography. The second is that oxidation of the
benzene rings occurs readily via a retro-Diels-Alder mechanism at 200°C, which is the
temperature of reflux during the platination step. The platination and oxidation steps,
previously separate, are now combined as shown in Scheme 1 into a single synthetic step, the
overall yield of which is four-fold higher than as obtained by the earlier two-step method.
6.1.3 Performance of BDP-Por / C60 Organic Photovoltaics
BDP-Por is not stable to vacuum thermal evaporation, so devices with the architecture
ITO/BDP-Por/C60(40 nm)/BCP(10 nm)/Al were fabricated with the BDP-Por layer deposited
at various thicknesses via spin coating from a chloroform solution at concentrations of 1, 2,
and 3 mg/mL. Donor films were annealed for 10 minutes at 80°C under nitrogen to drive off
residual solvent. These three concentrations produced films with thicknesses of 8.8 nm, 12 nm
and 23 nm as measured by ellipsometry. The C60, BCP, and Al layers were deposited
sequentially by vacuum thermal evaporation. The J-V curves from these devices are shown on
Figure 6.2. The device with the thinnest donor layer (8.8 nm) was the most efficient, with a
power conversion efficiency (PCE) of 1.42%. Increasing the BDP-Por layer thickness does not
increase the short circuit current (JSC) despite elevated light absorption, suggesting that the
Molecular Design for Organic Photovoltaics
Chapter 6 | 177
exciton diffusion length (LD) of BDP-Por is on the order of 8 nm, similar to that of PtTPBP.
18

The short exciton diffusion length can be understood in the context of the predominant exciton
migration mechanism; due to rapid formation of a triplet state in both PtTPBP and in BDP-
Por, Dexter energy transfer limits exciton migration to chromophores with Van der Waal’s
radii which overlap with the excited state chromophore. The BDP-Por OPV compares
favorably to the PtTPBP device (Figure 6.2), achieving the primary objective by increasing
absorption of the BDP-Por device relative to that of PtTPBP, increasing the photocurrent (JSC)
from 2.47 mA/cm
2
in the PtTPBP device to 3.84±0.81 mA/cm
2
in the array-based device. The
BDP-Por devices have higher error bars than Por devices due to limited morphological control
stemming from the spin coating process, compared with VTE deposition. The EQE for BDP-
Por devices shows that the enhanced current density is indeed due to an improved response in
spectral responsivity attributed to absorption from the pendant BODIPY. The fill factor is also
comparable between BDP-Por and PtTPBP, though it decreases as the BDP-Por thickness
increases. The VOC is unchanged in the BDP-Por device (VOC of 0.66±0.03 vs. 0.64±0.01 V),
-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
Voltage (V)

Current density (mA/cm
2
)
400 500 600 700
0
10
20
30
40
b)
8.8 nm BDP-Por
12 nm BDP-Por
22 nm BDP-Por
PtTPBP
External Quantum Efficiency (%)
Wavelength (nm)

Figure 6.2. (a) J-V curves for BDP-Por devices at several thicknesses against that of a solution processed 15 nm
PtTPBP device. It is clear that the increased absorption has led to an increase in J SC. This is reflected in the EQE plot
(b), where there is enhanced spectral response past 550 nm due to the BODIPY absorption.

Golden
178
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.
30-31

It is important to contrast the utility of
the more complicated multichromophoric array
material with a ternary blend architecture.
Keeping in mind the ultimate goal in OPV
performance is increased power conversion per
dollar spent on materials, manufacturing,
installation, and maintenance, the question of
whether increased materials complexity
provides enough of a benefit when contrasted
with the engineering complexity involved in the
fabrication of a ternary blend device raises
itself. In order to probe for an answer to this
question, ternary blend films with varying ratios
of PtTPBP and BODIPY
32
(Figure 6.3) were
prepared.
Unfortunately, the direct analogue of the
BODIPY pendant to BDP-Por (BDP1) is
exceptionally volatile and, as a consequence,
approximately 50% of the layer was lost from
the co-deposited film under the required
vacuum level for deposition of the C60, BCP and Al layers of the device. To address the
BDP1 BDP2
a)

-0.4 0.0 0.4 0.8
-4
-2
0
2
4
6
b)

Voltage (V)

Current Density (mA/cm
2
)
1:1 Por+BDP2
1:2 Por+BDP2
1:3 Por+BDP2
BDP2
BDP-Por
400 500 600 700
0
5
10
15
20
25
30 c)
1:1 Por+BDP2
1:2 Por+BDP2
1:3 Por+BDP2
External Quantum Efficiency
Wavelength (nm)

Figure 6.3. (a) Molecular structures of BDP1 and
BDP2. (b) J-V curves of devices made with donor
layers of Por+BDP2 show that the 1:2 ratio
generates the best performance. (c) The EQE
spectra of the same devices shows a shoulder at
550 nm, corresponding to added responsivity
from BDP2 absorption, grows in with increasing
BDP2 concentration.
Molecular Design for Organic Photovoltaics
Chapter 6 | 179
volatility problem of BDP1, a phenyl linked BODIPY dimer (BDP2, Figure 6.3) was
synthesized. BDP2 has nearly double the molecular weight and thus has a much lower volatility
compared to BDP1 while maintaining the same electronic and photophysical properties as the
monomer. Blended films of PtTPBP and BDP2 in ratios 1:1, 1:2, and 1:3 were prepared by
spin coating from chloroform solutions. As BDP2 contains two BODIPY units, its molar
absorptivity is effectively doubled compared to the monomer. The absorption spectra of the
Por+BDP2 films were indistinguishable from the Por+BDP1 films, given the corresponding
chromophore ratios. BDP2 has sufficiently high molecular weight such that there was no
detectable loss of BDP2 from the Por+BDP2 film even after extensive exposure to high
vacuum (10
-7
torr). Having prepared a stable blended film of similar composition to the
covalently bound chromophore array, the following task was to prepare OPVs with the same
structure as the BDP-Por OPVs. Devices with the three different blended donor layers and a
reference device with a solely BDP2 donor layer were prepared. The J-V curves for these
devices are depicted in Figure 6.3. OPVs with the 1:2 Por+BDP2 blend, i.e. the same
chromophoric ratio of BODIPY to porphyrin as is present in BDP-Por, gave the best device
performance (PCE of 1.33%). The performances amongst devices with varying porphyrin to
BODIPY ratios do not differ significantly, with the JSC and VOC remaining fairly constant with
the increasing BDP2 content (again suggesting that the PtTPBP/C60 CT state is the limiting
factor determining the VOC). The primary difference between devices made from different
Por:BDP2 ratios observed in the FF, which is at a maximum in 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.
Golden
180
When comparing these blended donor films to the devices made with the
multichromophoric array, the JSC, FF, and the general shape of the EQE curve are comparable,
as is to be expected given the similar absorption patterns between the blended film and that of
the array. However, the measured VOC values in the blended donor devices were markedly
lower than those of the array (0.57±0.03 V vs. 0.66±0.03 V, respectively), leading to a reduced
power conversion efficiency compared to the array device. The pure BDP2 donor device gives
a lower JSC and VOC than either the blended or BDP-Por donor layer devices. While a lower
JSC was expected for the BDP2 based OPV, due to the lack of the porphyrin absorber, the lower
VOC was not intuited. The oxidation potential for BDP2 is 0.68 V (vs. Fc
+/0
) while that of
BDP-Por is 0.45 V (vs. Fc
+/0
), similar to the oxidation potential reported for PtTPBP of 0.4 V
(vs. Fc
+/0
)
33
. Based on the 230 mV larger energy difference between the donor-HOMO and
C60-LUMO ( EDA) for the BDP2 OPV compared to the porphyrin devices, the logical
prediction was that the BDP2 device should have a larger VOC than the other device
architectures studied here. A similar situation has been observed when comparing copper-
phthalocyanine and PtTPBP based devices; in that report, steric interactions between the large,
planar PtTPBP which hinder close approach of the donor to the spherical C60 were identified
as the reason for the anomalous VOC behavior.
30
BDP2 can associate directly with the C60,
while the structure of BDP-Por allows only an edge-on interaction of the BODIPY moiety with
C60; it is likely that this difference limits back-electron recombination and/or the absorption
coefficient, CT, in the BDP-Por/C60 charge transfer step of photocurrent generation, which
enhances its VOC relative to the BDP2 device.
Molecular Design for Organic Photovoltaics
Chapter 6 | 181
6.1.4 Morphological Instability in the Ternary Blend Por+BDP2 Device
The major difference in performance between the multichromophoric array and the
ternary blend device stems from morphological instability in the blended device architecture.
J-V curves for a set of Por+BDP2 devices under illumination for increasing durations are
depicted in in Figure 6.4. In acquiring 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 VOC of the devices had
significantly decreased. The dark curve,
however, is seen to remain stable for
multiple voltage sweeps as long as no light
is cast upon 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, as is observed in the
multichromophoric array device
architecture.
34-36
Upon initial illumination, however, the Por+BDP2 device produced a J-V
curve with a VOC of 0.58 V which clearly does not converge with the dark curve, indicating
that some morphological change has already occurred in the two seconds it requires 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. Further illumination of the
device produces J-V curves with progressively lower VOC values, but leaves the JSC and FF
-0.4 0.0 0.4 0.8
-4
-2
0
2
4
6
Current Density (mA/cm
2
)
Voltage (V)

Dark (0 sec)
Light (0 sec)
Dark (15 sec)
Light (15 sec)
Light (1 min)
Dark (1 min)

Figure 6.4. The J-V curves of an
ITO/Por+BDP2/C60/BCP//Al device after varying
periods of illumination shows how the V OC starts around
0.64V and eventually stabilizes at 0.36 V.
Golden
182
parameters unchanged. These devices ultimately stabilize after about 1 minute of illumination
at a VOC of 0.36 V.
The identity of the donor and acceptor,
30-31, 37-40
as well as the local morphology at the
D/A interface contributing to the formation of the CT state,
17, 41-48
have been shown to be the
primary factors determining the VOC in OPVs. These two factors are manifested in a clear
correlation between the VOC and the energy of the charge transfer state (ECT) between the donor
and acceptor materials in the OPV.
39, 44-45
As the VOC is characteristic of a particular material
system, one can use the VOC of a device to give insight into the nature of the molecular D/A
charge transfer interaction at the interface of the device even when more than one donor or
acceptor is present.
47, 49-51
The BDP-Por devices and the initial scans of the Por+BDP2 devices
give VOC values very close to those observed for an OPV designed with a neat PtTPBP donor
layer, suggesting that, at least initially, the porphyrin containing devices have a similar CT
state involving a PtTPBP/C60 pair at the interface. To determine the characteristic VOC of a
device with a BDP2/C60 ECT, an OPV was prepared with a neat BDP2 donor layer
(ITO/BDP2(40nm)/C60(40 nm)/BCP(10nm)/Al). The J-V curve of this device is shown in
Figure 6.3, where it can be seen that the BDP2/C60 architecture leads to a VOC of 0.34 V. In
the ternary blend Por+BDP2 device, the VOC is observed to fall from a value of 0.58 V to
0.36 V, suggesting that under illumination, the donor/acceptor CT state in the ternary blend
rapidly changes from one which is largely PtTPBP/C 60 upon fabrication to one which is largely
BDP2/C60 after only one minute under one sun intensity. This instability of the Por+BDP2
devices under illumination persists even when the devices are tested under an inert atmosphere,
ruling out oxidative pathways as an explanation for the change in VOC. The films of Por+BDP2
were investigated before and after illumination by GIXRD but show no evidence of
Molecular Design for Organic Photovoltaics
Chapter 6 | 183
crystallization and no discernible difference between the array and the blended film, indicating
that both films are largely amorphous, and if any crystalline domains exist, they are too small
to be detected by diffraction.
6.1.5 Implications of Multichromophoric Arrays in Organic Photovoltaics
Broadening the spectral responsivity in OPV active layer materials is a primary
challenge that must be overcome if OPVs are to approach theoretical efficiency limits. In this
work (Section 6.1), it has been shown that the inclusion of multiple chromophores with
complementary absorptions in one or more of the photoactive layers can help broaden the
spectral responsivity and enhance the JSC, provided that the energetics are correctly designed
as to not introduce charge or exciton traps. This can be accomplished by either using a ternary
blend or by covalently binding complimentary chromophores into a multichromophoric array.
In the former case, there is no penalty in terms of materials complexity, but there is a cost
borne in terms of increased engineering complexity. In the latter, while engineering complexity
remains simple, the multi-step synthetic process to build and link the multichromophoric array
will lead to increased materials costs. It was found in this study, however, that the
multichromophoric array offers enhanced morphological stability relative to the ternary blend
photovoltaic which may offset the increased materials costs in analogous devices going
forward.
13

The specific materials used in these device architectures, namely platinum porphyrins,
introduce a large and unavoidable thermalization loss pathway in the form of rapid intersystem
crossing due to spin orbit coupling from the platinum atom, resulting in energy lost from a high
energy singlet to a low energy triplet. This thermalization loss is reflected in the relatively low
VOC of 0.66 V even in the best-performing multichromophoric array device, and may be the
Golden
184
reason for the rapid morphological change in the ternary blended device, where heat generation
resulting from vibronic relaxation to the triplet may have induced diffusion of the lighter BDP2
moieties to the D/A interface. In designing arrays for better performing OPVs in the future,
singlet-only arrays may provide a pathway to the desired increase in JSC without sacrificing
VOC in the process.
6.2 Photochemical Upconversion
Photochemical upconversion (UC), a phenomenon based on sensitized triplet-triplet
annihilation (TTA), is a non-coherent process wherein the energy of a photon absorbed by a
triplet sensitizer is eventually emitted from a judiciously selected acceptor/emitter, resulting in
an anti-Stokes energy shift.
1
In the TTA-based UC process, triplet sensitization between
photoexcited donors and ground state acceptors leads to the accumulation of long-lived excited
triplet acceptors. TTA eventually occurs between two triplet-excited acceptors resulting in the
formation of one ground state and one singlet excited state, the latter generating fluorescence
emission characteristic of the triplet acceptor (Figure 6.5). The TTA-UC process has been
proposed as a method to increase the efficiency of photovoltaics by decreasing sub-bandgap
transmission losses.
52
That is, it is
possible to design a ternary blend or
multichromophoric single junction
photovoltaic with a large bandgap Eg
(resulting in a large VOC) which also
collects photons below the bandgap
energy of the device in the form of
triplet excitons. These triplet excitons

Figure 6.5. Jablonski diagram depicting the processes
involved in triplet sensitized photochemical upconversion
(ISC = intersystem crossing, TTET = triplet−triplet energy
transfer, TTA = triplet−triplet annihilation, DF = delayed
fluorescence).

S
1
S
1
T
1
T
1
T
1
T
1
S
1
S
1
TTA
ISC ISC
TTET TTET
DF
Sensitizer Sensitizer Emitter Emitter
S
0
S
0
S
0
S
0
Molecular Design for Organic Photovoltaics
Chapter 6 | 185
can then annihilate onto a triplet-accepting electron donor or acceptor material to form a high
energy singlet exciton capable of initiating the CT process at the D/A heterojunction. It has
been proposed that the single-junction efficiency limit may be increased to 43% given a
bandgap energy Eg of 1.76 eV using the TTA-UC process.
53
Materials considerations for the
sensitizer species in an efficient TTA-UC photovoltaic include: high molar absorptivity in the
sub-bandgap region with respect to the donor and acceptor molecules which will comprise the
photocurrent generating CT state, high intersystem crossing yield (usually obtained from
incorporation of a heavy atom into the molecular structure), long triplet lifetime, and a
relatively small singlet-triplet energy gap in the photosensitizer to reduce thermalization losses
inherent to the intersystem crossing step. The triplet-accepting species (which may be either
the ultimate donor or acceptor material) must have a singlet state which is slightly lower in
energy than twice its triplet state energy, it must have a triplet which is slightly lower in energy
than the triplet of the photosensitizer, and the lifetime of its triplet must be very long (>100 us)
in order for enough time to pass for the bimolecular TTA process to occur. Materials classes
which are commonly used in the study of photochemical upconversion also happen to be
common in the study of organic photovoltaics; metalloporphyrins and heavy-atom
incorporating BODIPYs have been shown to be efficient triplet sensitizers, and acenes are
common triplet acceptors.
Given the complicated state energy dynamics involved in TTA UC, materials are
typically screened for UC efficiency using steady-state photophysical experiments in solution.
The study of TTA-UC has been largely confined to deoxygenated organic solvents, however,
that ensure solubility, facile diffusion, low polarity, and suppression of triplet state reactivity
with dioxygen;
54
O2 reactivity represents one of the most problematic issues in TTA-based UC
Golden
186
processes. It serves as an efficient quencher for both the donor and acceptor triplet excited
states, thus leading to a dramatic decrease of upconversion efficiencies as well as
photobleaching due to the strong oxidative properties of the main long-lived photosensitized
product, singlet oxygen (
1
O2
*
).
55
It is worthwhile to develop a system which facilitates this
screening method so that the more complicated process of device engineering can be
undertaken without long lag-times between materials development and characterization. In this
work, a polyethylene glycol solvent system capable of screening high efficiency TTA UC
processes in air is developed and its utililty is probed to ensure long-term stability of solutions
in air.
†54

6.2.1 Triplet-Triplet Annihilation for Photochemical Upconversion in Air
In this work, it is demonstrated that efficient upconversion can be achieved under
ambient conditions without prior deoxygenation by using a thoughtful combination of solvent
and antioxidant. Polyethylene glycol (PEG) is largely gas impermeable and serves to act as a
barrier to oxygen diffusion from air.
56-57
Moreover, PEGs are transparent across the UV, visible
and NIR spectral regions, photochemically stable, viscosity tunable, and inexpensive. The
concomitant utilization of a singlet oxygen scavenger in the PEG solutions would afford the
complete removal of dioxygen, thereby leading to a photochemically deoxygenated medium
in which to study photochemical upconversion. Three oxygen scavengers were selected here
from a vast number of potential candidates, namely, oleic acid (OA), 9,10-dimethylanthracene
(DMA), 2,5-dimethylfuran (DMF). The upconversion system was based on the red absorbing

†
This work was performed in collaboration with the Castellano group at NC State University as part of the post-
doctoral research performed by Dr. Cédric Mongin. Figures and portions of the text in Section 6.2 are duplicated
from material published in Ref. 54.
Molecular Design for Organic Photovoltaics
Chapter 6 | 187
triplet photosensitizer platinum(II) tetraphenyltetrabenzoporphyrin (PtTPBP), synthesized as
previously reported,
13
in concert with the highly fluorescent blue/green emitter 9,10-
bisphenylethynylanthracene (BPEA, ΦF = 0.94) (Figure 6.6). Relevant triplet energy transfer
parameters and upconversion metrics have been documented to determine the impact of the
various components on the ultimate upconversion performance. In particular, the influence of
PEG viscosity was thoroughly investigated using Stern-Volmer photoluminescence quenching
experiments in order to yield the most efficient compositions. The relevant bimolecular energy
transfer reactions were shown to be diffusion limited in these viscous PEG solvents, leading to
a dramatic decrease of the energy transfer and upconversion quantum efficiencies with
increasing viscosity. The influence of the oxygen scavenging compounds along with their
requisite antioxidant by-products on upconversion stability has also been evaluated by
monitoring the emission signal intensity over the course of (at least) 20 hours of continuous
irradiation.

Figure 6.6. Normalized absorption (solid lines) and photoluminescence (dashed lines) spectra of PtTPBP (red)
and BPEA (blue) in PEG-400 measured under ambient conditions and molecular structures of PtTPBP and
BPEA.

400 500 600 700 800
0.0
0.2
0.4
0.6
0.8
1.0
Normalized OD/Emission
Wavelength (nm)
Golden
188
6.2.2 Oxygen Scavenging in PtTPBP PEG Solutions
The PtTPBP sensitizer was initially investigated in the three different molecular weight
PEG solutions: 200 g/mol (PEG-200), 400 g/mol (PEG-400) and 600 g/mol (PEG-600) having
respective viscosities at 25°C (and melting points) of 55 cP (-45°C), 90 cP (4°C) and 140 cP
(20°C). The phosphorescence lifetimes were determined under deoxygenated conditions by
measuring the strong phosphorescent emission of PtTPBP (Φphos = 0.67 in PEG) centered at
770 nm (Figure 6.6) with pulsed laser excitation. The excited state lifetime of PtTPBP
increases slightly with solution viscosity 46 µs, 47.5 µs and 48.5 µs in PEG-200, PEG-400 and
PEG-600, respectively, due to the rigidochromic effect.
58
By comparison, the measured
PtTPBP lifetime in toluene (η = 0.59 cP) is 38 µs (Φphos = 0.51),
59
which represents a 25%
decrease with respect to the PEG-based solvents. The gas permeability of the three different
PEGs was evaluated using the phosphorescence lifetime of PtTPBP. Briefly, a solution of PEG

Figure 6.7. Oxygen insulation properties of the various PEG solutions. (Left) PtTPBP excited state lifetime
stability at room temperature in PEG solutions exposed to air after initial deaeration by nitrogen sparging.
(Right) Near-IR singlet oxygen emission from PtTPBP in PEG-600 under nitrogen (green line), air sparged
(blue line), and exposed to air for 24h after sparging with nitrogen (red line) at room temperature. Inset displays
the first 8 hours of singlet oxygen signal variation under air, with (red circles) and without (blue squares)
nitrogen sparging.

0 10 20 30 40 50
40
42
44
46
48
50
PEG-200
PEG-400
PEG-600
PtTPBP lifetime (µs)
Time (h)
1200 1250 1300 1350 1400 1450
0.0
0.2
0.4
0.6
0.8
1.0
1.2
0 2 4 6 8
0
2
4
6
Emission Intensity ( 10
5
cts)
Wavelength (nm)
N
2
then Air
N
2
Air
Air
N
2
then Air
Intensity ( 10
5
cts)
Time (h)
Molecular Design for Organic Photovoltaics
Chapter 6 | 189
(200, 400 or 600) containing PtTPBP (10 µM) was thoroughly bubble deaerated under nitrogen
and left under nitrogen atmosphere until no gas bubbles were present in the solution. The
lifetime and photoluminescence spectrum under nitrogen were recorded and the sample was
then opened to air and the lifetime was measured at regular time intervals. If oxygen dissolved
into these solutions, both the lifetime and the photoluminescence emission of PtTPBP would
readily respond through quenching. However, in every sample, even after being open to air for
48 hours, absolutely no signs of oxygen diffusion were observed in either experiment
(Figure 6.7). Moreover, no singlet oxygen emission at 1270 nm could be observed in any of
these samples under strong continuous wave laser excitation of PtTPBP. Control experiments
using toluene and polydimethylsiloxane (PDMS) as solvent were also performed and as
anticipated, dramatic emission intensity as well as lifetime decreases were clearly observed for
the PtTPBP phosphorescence within the first few minutes of air exposure (Figure 6.8). The
use of PEGs as solvent appears to be a wise choice to prevent oxygen re-dissolution and
subsequent quenching in reactions containing O2-sensitive components.
While PEG is a solvent of choice to
avoid oxygen dissolution, it still requires
deaeration prior to experimentation since
oxygen is initially present in the as-received
reagent. Several processes can be used to
remove dissolved oxygen from solvent. The
most efficient would be to perform a series of
Freeze/Pump/Thaw (FPT) degas cycles but
this process is extremely time consuming and

Figure 6.8. Evolution of PtTPBP emission intensity
at 770 nm over time at room temperature upon 614
nm excitation in PDMS under nitrogen and after
being opened to air.

Golden
190
resource intensive. Another option is to bubble deaerate the solution using nitrogen or argon,
but this again is time consuming, necessitates the use of inert gases, and is not suitable for
highly viscous solvents such as PEGs due to the generation of microbubbles in the sample that
represent a source of light scattering. Thus, in the present study, we focused on the exploitation
of the impressive oxygen insulating properties of PEGs but needed to remove the initially
dissolved oxygen by an unconventional method. Visible irradiation of an aerated PtTPBP
solution leads to the formation of singlet oxygen with a quantum yield approaching unity. The
singlet oxygen product that forms is a powerful oxidant and can react with a variety of
functional groups.
60
The chemical trapping of singlet oxygen by an oxygen scavenging species
would result in a continuous decrease of the original O2 concentration, eventually leading to
its complete removal from the initial solution. Fortunately, PtTPBP is not prone to
photobleaching and is not degraded by the singlet oxygen generated under these experimental
conditions, probably due to a low oxygen concentration in the initial aerated solution and slow
diffusion rates inherent to high viscosity PEG solvents. This study focused on three oxygen
scavenging species: OA, DMA and DMF. An aerated solution of PtTPBP (10 µM) in each
PEG solvent (200, 400 and 600) containing 10 mM of OA, DMA or DMF was irradiated using
622 nm (16 µJ, 5 Hz) 500 ps laser pulses. Oxygen scavenging occurs through oxidation of the
olefin-containing substrate by the PtTPBP-sensitized singlet oxygen, forming an endoperoxide
that eventually dismutates into a peroxide species.
61-63
Under air saturated conditions, the
PtTPBP excited state lifetime in PEG was determined to be near 10 µs, confirming the low
solubility of oxygen and slow diffusion compared to commonly utilized organic solvents where
the PtTPBP lifetime and triplet excited state lifetime in general are on the order of hundreds of
nanoseconds.
64
Under laser pulse irradiation at 622 nm, regardless of the oxygen quencher
Molecular Design for Organic Photovoltaics
Chapter 6 | 191
used, a net increase of the PtTPBP lifetime was observed, which reached a plateau at 46 µs, 47
µs and 48 µs in PEG-200, PEG-400 and PEG-600 solutions, respectively, in good agreement
with the lifetimes measured under nitrogen saturated conditions. Under air-saturated
conditions, the PtTPBP photoluminescence intensity decays measured in each PEG solution is
adequately modeled using single exponential decay kinetics. However, upon visible irradiation
in the presence of the O2 scavenger, a transitional regime was observed and these initial
intensity decays were adequately fit using a stretched exponential function, I(t) = A exp(-kt)
β

(Figure 6.8). This behavior likely indicates that the scavenging process is not homogeneous in
the sample due to the high viscosity of the PEGs utilized. Thus, domains are likely being
generated where there are distributions of excited state lifetimes. This distribution remains
relatively narrow (β values between 0.75 and 1) until complete oxygen removal (after extensive
photolysis) occurs in the sample, upon which time the intensity decay then reverts back to

Figure 6.8. Anti-oxidants applied as oxygen scavengers in PEG-600. (Left) PtTPBP lifetime increases (purple
to red) observed in the PEG-600 solution at room temperature in presence of OA (10 mM) upon prolonged
(t = 0-100 min) pulsed laser irradiation (622 nm, 16 µJ, 5 Hz). The inset highlights the transition regime, from
mono-exponential to stretched exponential then back to mono-exponential photoluminescence intensity decay
behavior observed during the photochemical scavenging process. (Right) PtTPBP lifetime increases with
photolysis time using pulsed laser irradiation (622 nm, 16 µJ, 5Hz) in PEG-600 solutions at room temperature
in the presence of DMF (red top triangles), DMA (blue down triangles), or OA (green diamonds).

0 50 100 150 200
0.0
0.2
0.4
0.6
0.8
1.0
0 50 100 150 200
0.1
1
Normalized Emission Intensity
Time (µs)
Normalized Emission
Time (µs)
Deoxygenated/N
2
t = 100 min
Transition
Air
t = 0
0 20 40 60 80 100
10
20
30
40
50
t
DMF
= 10 min
t
DMA
= 30 min
t
OA
= 90 min
Deoxygenated
N
2
DMF
DMA
OA
PtTPBP lifetime (µs)
Exposure (min)
Air
Golden
192
explicit single exponential behavior. Moreover, depending on the specific singlet oxygen
scavenger, this limit is achieved after different exposure times. DMF was the most efficient
scavenger with complete deoxygenation occurring in less than 10 minutes, followed by DMA
taking 30 minutes, and OA being complete in 90 minutes (Figure 6.8). These differences in
terms of O2 removal efficiency are easily explained by the previously determined reaction rate
constants (kr) for these compounds with singlet oxygen: kr = 2×10
8
, 3×10
7
and 5×10
4
L mol
−1

s
−1
, respectively, for DMF, DMA and OA, in addition to their attenuated diffusion coefficients
in the PEG solvents. As expected, the irradiation time required to perform complete
deoxygenation follows the same trend. The concentration of oxygen scavenger can be
considered constant since it is at least 100 times higher than the initially dissolved oxygen
concentration (< 0.1 mM in air equilibrated conditions). No degradation of PtTPBP was
observed during the oxygen trapping process and its photoluminescence as well as its
absorption spectrum remained unchanged as a result of this procedure. After being
photochemically deoxygenated, each system was maintained open to air, and the lifetime and
emission spectra of PtTPBP were recorded at regular intervals. As anticipated, no signs of
oxygen diffusion from air were observed and the lifetime remained constant for more than 48
hours in each instance. All the scavengers tested proved to be very efficient for the removal of
oxygen from these PEG solutions. However, for the remainder of the upconversion-focused
portion of this study, OA was used exclusively due to experimental convenience.
Dimethylfuran (DMF) was the most efficient scavenger, but it is volatile and its concentration
in solution decreased rapidly over time. DMA is a solid that is not very soluble in the PEGs
and tends to crystalize with time. OA is a highly viscous liquid (40 cP at 20°C) with a high
boiling point (360°C). These properties promote straightforward dispersion in PEG, limit
Molecular Design for Organic Photovoltaics
Chapter 6 | 193
potential viscosity variations, and avoid the loss of the scavenger with time, thereby markedly
increasing the system’s operational lifespan.
6.2.3 Photochemical Upconversion in Air using PtTPBP as a Triplet Sensitizer of BPEA
The energy transfer quenching kinetics between PtTPBP sensitizers and BPEA
acceptors were studied under oxygen free conditions (nitrogen sparging followed by bubble
removal) in PEG-200, 400 and 600 using time-resolved photoluminescence spectroscopy.
Stern-Volmer constants (KSV) were determined to be 4200, 2800 and 1900 M
-1
(Table 6.1,
Figure 6.9) in PEG-200, 400 and 600, respectively, corresponding to bimolecular quenching
rate constants (kq) of 9.4×10
7
, 6.1×10
7
and 4.0×10
7
M
−1
s
−1
. As anticipated, increasing viscosity
causes a decrease in the Stern-Volmer quenching constant due to limited molecular diffusion
in solution. In toluene, a Stern-Volmer constant of KSV = 110,000 M
-1
, corresponding to a

Figure 6.9. Study of PtTPBP photoluminescence quenching by BPEA in the PEG solutions. (Left) Stern-Volmer
plot in deaerated PEG-200 (red squares), PEG-400 (green circles) and PEG-600 (blue triangles) generated from
photoluminescence lifetime quenching of PtTPBP in presence of BPEA using single exponential fitting. (Left)
Correlation between the inverse of the quenching constant (k q) and PEG viscosity (red circles) compared with
correlation between the PEG molecular weight and solution viscosity (blue squares). Data have been fit using
linear equations.

0 1 2 3 4
0
5
10
15
PEG-200
PEG-400
PEG-600


  
[BPEA] (mM)
40 60 80 100 120 140
10
15
20
25
1/k
q
( 10
-9
M s)
Viscosity (cP)
800
600
400
200
Molecular Weight (g mol

)
KSV(PEG-200) = 4200 M
‒1
KSV(PEG-400) = 2800 M
‒1

KSV(PEG-600) = 1900 M
‒1

MW  
1/kq  
Golden
194
quenching rate constant of 2.9×10
9
M
−1
s
−1
has been previously determined, which approaches
the diffusion limit in toluene, kd = 1.1×10
10
M
−1
s
−1
. In PEGs, the high viscosity leads to
significantly lower calculated diffusion limit values, kd = 1.2×10
8
, 7.2×10
7
and 4.6×10
7
M
−1

s
−1
in PEG-200, PEG-400 and PEG-600, respectively. Based on these values, the energy
transfer quenching processes are indeed diffusion limited in all of these PEG solvents (kq ~ kd).
The experimental bimolecular energy transfer rate constant results are plotted versus the
viscosity of the medium and have been fit with a straight line using Stokes-Einstein theory
(Equation 6.1), where kd is the diffusion constant, R is the sum of the molecular radii, r, of
the two species A and B involved in the collisional quenching, N A is Avogadro’s number, kB
is the Boltzmann constant, D is the diffusion coefficient, T is the temperature, η is the viscosity
Table 6.1. Performance of sensitized photon upconversion compositions in PEG solvents.

K SV
a

(M
‒1
)
k q
(×10
7
M
‒1
s
‒1
)
k d
b

(×10
7
M
‒1
s
‒1
)
Φ UC
c
N 2
(%)
Φ UC
c
OA Air (%)
PEG-200 4200 ± 110 9.4 12 30.9 ± 1.2 (3 mM) 30.3 ± 1.5 (3 mM)
PEG-400 2800 ± 80 6.1 7.2 21.2 ± 0.5 (4 mM) 20.4 ± 0.7 (4 mM)
PEG-600 1900 ± 210 4.0 4.6 13.9 ± 0.9 (5 mM) 13.1 ± 1.2 (5 mM)
a,c
Reported values are the average of at least three independent measurements, and error margins represent the
standard deviation of those measurements.
b
Diffusion limits, k d’s, were calculated using the Stokes-Einstein
theory (Equation 6.1).
c
Acceptor concentrations used are given in parentheses.

Equation 6.1
→

Equation 6.1

 
1 1 3
10 4
 
   s M N D D R k
A B A d

r
T k
D
B
 6

3 3
10
3
8
10
6
2 2 4   








  
A
B
A
B
d
N
T k
N
r
T k
r k
 


1
 
d q
k k
Molecular Design for Organic Photovoltaics
Chapter 6 | 195
and r is the radius of one molecule. In order to simplify the equation,the radius of A and B
were assumed to be equal : rA = rB = 2r = R.

The quenching rate constant kq in a given PEG solvent is easily predicted to vary as a
function of the solvent viscosity (Figure 6.9). Moreover, it is interesting to note that the
viscosity of the PEG solution is directly proportional to its molecular weight and therefore to
the number of ethylene glycol subunits present. Thus, quenching rates can be readily predicted
from the molar mass of the PEG host (Figure 6.9).
The upconversion quantum efficiency was evaluated in all three PEG solutions under
deoxygenated conditions (nitrogen sparging followed by bubble removal). First, the
upconversion quantum yields (ΦUC) were determined in PEG solutions where the PtTPBP
lifetime was quenched by at least 90% using a 635 nm continuous wave laser diode focused
through a lens (spot size 1 mm
2
). The incident power was tuned using a set of neutral density
filters and a 550 nm long pass filter was used to remove any potential high-energy excitation
wavelengths. The upconversion quantum yield (ΦUC) measured under the specified
experimental conditions is the product of the quantum yields of each step in the reaction:
sensitizer intersystem crossing (ΦISC), triplet−triplet energy transfer (ΦTTET), triplet−triplet
annihilation (ΦTTA), and acceptor fluorescence (ΦF) according to Equation 6.2.

Equation 6.2

The methylene blue cation in water (MB, ΦF = 0.04) was used as an internal reference
fluorophore according to Equation 6.3, where Φstd is the emission quantum yield of the
standard, I is the integrated emission intensity of the standard or upconversion solutions, A is
F TTA TTET ISC UC
        
Golden
196
the absorbance of the standard and upconversion solutions at 635 nm, and η is the refractive
index of the solvents used. Two absorbed photons are needed to generate one emitted photon
through TTA; therefore; the factor of 2 is included in Equation 6.3 to make the theoretical
maximum quantum yield unity rather than 0.5. Quantum yields were measured under
optimized experimental conditions (linear response regime) and are reported in Table 6.1.

Equation 6.3

Due to the relatively low Stern-Volmer quenching constants, the concentration of
acceptor required to achieve at least 90% quenching of the PtTPBP triplet exciton was 3 mM
for PEG-200, 4 mM for PEG-400 and 5 mM for PEG-600. These conditions lead to strong
reabsorption of the upconverted photons (inner filter effects). In order to avoid this
phenomenon and to provide the most accurate photophysical characterization, the apparatus
was modified in a manner to minimize the thickness of the solution penetrated by the
upconverted emission. However, even under these experimental conditions evidence of some
reabsorption was observed in the BPEA emission profile and the reported spectra have been
corrected accordingly to account for this loss of detected photons. The upconversion quantum
yields increase with decreasing PEG viscosity, with ΦUC values of 0.14 (PEG-600), 0.21 (PEG-
400), and 0.31 (PEG-200) (Figure 6.10). This decrease in efficiency upon increasing viscosity
can be rationalized by the slower diffusion of triplet excited state acceptors with increasing
viscosity, favoring the pseudo-first order decay process over second-order annihilation,
Molecular Design for Organic Photovoltaics
Chapter 6 | 197
decreasing the upconversion efficency. Power dependence studies reveal typical behavior
expected for upconversion systems (Figure 6.10).
The population of singlet excited state acceptors (
1
A
*
), from which the upconverted
emission originates, results from two competing kinetic processes: kt which is the first order
decay of the acceptor triplet excited state (
3
A
*
), and ktt which is the second order triplet−triplet
annihilation (TTA) of the acceptor triplet excited state, ktt[
3
A*]
2
. TTA is the process
responsible for the upconverted photoluminescence observed. The first order rate constant
includes the inherent decay of the triplet excited state acceptor as well as the pseudo-first-order
decay resulting from quenching by the media (oxygen, solvent, other molecules). These two
processes produce two distinct kinetic regimes. In the weak annihilation limit (low power) kt
is dominant, resulting in a quadratic dependence of the upconverted emission versus the

Figure 3.10. Incident light power dependence study of photochemical upconversion occurring in PEG solutions.
(Left) Upconversion quantum yields measured for PtTPBP (20 µM) and BPEA (3, 4 or 5 mM depending on the
PEG utilized) in deoxygenated PEG solutions, PEG-200 (red squares), PEG-400 (green circles) and PEG-600
(blue triangles), as a function of excitation power density at 635 nm. Gray dashed lines represent the average
quantum yield in the plateau region. (Right) Upconverted emission spectra of BPEA (4 mM) sensitized by
PtTPBP (20 µM) in deoxygenated PEG-400 solution and the emission spectra of the MB standard in water with
635 nm excitation. Inset displays the double logarithmic plot of BPEA integrated emission intensity as a function
of 635 nm excitation power density. Solid lines illustrate a slope of 2 (red, quadratic) and a slope of 1 (blue,
linear).

0 500 1000 1500 2000 2500
0
10
20
30
PEG-200
PEG-400
PEG-600
UC Quantum Yield (%)
Power Density (mW/cm²)
500 600 700 800
0
2
4
6
8
1 10 100 1000
10
5
10
6
10
7
10
8
10
9
10
10
Emission Intensity ( 10
8
cts)
Wavelength (nm)
Integrated Emission
Power (mW/cm²)
Golden
198
excitation intensity. In the strong annihilation limit, ktt is dominant resulting in a linear
dependence of the upconverted emission versus the excitation intensity. In the strong
annihilation limit the system reaches a plateau where the upconversion quantum yield is
maximized. The power needed to reach this limit depends on the sensitizer absorbance and the
acceptor concentration. The transition region can be associated with the range where the two
processes, kt and ktt, have the same weight on acceptor triplet excited state decay.
The quadratic dependence of the upconverted photoluminescence was indeed observed
for low energy incident power density (< 100 mW cm
−2
) excitation in all of the different PEGs.
The linear region observed at higher incident power densities corresponds to the upconversion
maximum efficiency limit.
65
For all the three different PEGs used in this study, the transition
threshold between the quadratic and the linear regime occurs near 200 mW cm
-2
. Although this
value is not realistic for direct solar energy conversion applications under one-sun excitation,
other upconversion applications such as imaging are indeed viable and this solvent system
nevertheless provides a simple and reliable screening method for characterizing upconversion
behavior which can then be applied to OPVs.
The stabilities of the three different PEG systems were studied under identical
experimental conditions at 50 mW cm
-2
with 635 nm light while stirring for at least 20 h. Under
deoxygenated conditions, no signs of degradation were observed in any of the studied systems.
The upconversion signal intensity remains quasi-stable over extended time frames, indicating
the absence of photo-degradation of the emitter and sensitizer (Figure 6.11). These results
imply exemplary photochemical stability that is promising for future applications. Moreover,
if the stirring was ceased, interesting behavior was observed. In those instances, the
upconversion intensity rapidly increased to a plateau that remained stable for hours. This was
Molecular Design for Organic Photovoltaics
Chapter 6 | 199
attributed to local heating of the PEG solution resulting in a decrease of the local viscosity,
leading to more efficient upconversion as a result (increasing Stern-Volmer constant, TTET
and TTA efficiency). To model this process, a sample containing PtTPBP (10 µM) and BPEA
(3 mM) in PEG-600 (melting point 18-21°C) was frozen and studied at 20°C, where it
remained solid. The upconversion in this sample was subsequently monitored under high
power continuous laser diode excitation (635 nm, 1 W cm
−2
). Initially, no upconversion signal
was observed; only scattering from the laser and phosphorescence from the sensitizer were
detected. With continued laser irradiation, local heating became apparent and the PEG-600
began to melt, inducing the upconversion process. The upconversion signal continued to
steadily increase, eventually achieving a plateau that remained stable for several hours (Figure
6.11). After the laser diode was turned off, the system slowly started to freeze, thereby halting
molecular diffusion and arresting the photochemical upconversion process.

Figure 6.11. Stability of the upconverted emission intensity with prolonged light exposure. (Left) Upconversion
in PEG-400 monitored over 20 h while irradiating at 635 nm (50 mW/cm
2
). Inset displays the upconverted
emission spectra before (red) and after (green) 20 h of irradiation. (Right) Upconversion in a frozen PEG-600
solution monitored for 4 h while irradiating at 635 nm (1 W/cm
2
). Inset displays the first 60 s of photolysis.

0 2 4 6 8 10 12 14 16 18 20
0
1
2
3
4
480 520 560 600
0.0
0.5
1.0
1.5
2.0
2.5
Emission Intensity ( 10
7
cts)
Time (h)
Intensity ( 10
5
cts)
Wavelength (nm)
t = 0
t = 20 h
0 250 500 750 5k 10k
0
1
2
3
4
5
6
0 20 40 60
0
5
10
15
Emission Intensity ( 10
5
cts)
Time (s)
Intensity ( 10
4
cts)
Time (s)
Golden
200
It has been demonstrated that upconversion easily takes place in liquid PEGs of various
molecular weights (200, 400, 600 g mol
-1
) under deoxygenated conditions. It has also been
shown that PEGs are highly impermeable to oxygen and that an inexpensive singlet oxygen
scavenger, such as OA, can be used to completely remove dissolved oxygen from these
solutions. Therefore, photochemical upconversion under ambient conditions in the presence of
OA was also studied in the three different PEG solutions. Each was prepared under ambient
atmospheric conditions using PtTPBP (20 µM) as sensitizer, BPEA (3 mM in PEG-200, 4 mM
in PEG-400 and 5 mM in PEG-600) as acceptor/emitter and OA (30 mM) as oxygen scavenger.
No upconversion signal could be detected at the beginning of each experiment and only weak
phosphorescence emanating from PtTPBP at 770 nm was observed at this time. After a few
seconds, the upconversion signal appeared and continuously increased, ultimately achieving a
plateau after extended photolysis. At this point it can be stated that all of the dissolved oxygen
is consumed by the OA, and the signal remains stable over several hours of continuous
exposure to air (Figure 6.12). The oxygen consumption rate was faster in these experiments
in comparison to that observed in the photoluminescence lifetime study depicted in Figure 6.8
due to the high density power utilized in upconversion experiments (50 mW/cm²) compared to
the low flux needed for lifetime measurements (0.5 to 0.6 mW/cm²). Similar quantum
efficiencies were obtained using the antioxidant strategy compared to the analogous nitrogen
sparged deoxygenated systems (Table 6.1). No significant differences in terms of stability
were observed between the three different PEG solutions containing OA. No degradation of
PtTPBP or BPEA was observed as evidenced by UV-Vis absorption spectra before and after
each long-term photolysis experiment, thereby indicating that singlet oxygen was
quantitatively trapped by OA and was unavailable to react with BPEA or PtTPBP and
Molecular Design for Organic Photovoltaics
Chapter 6 | 201
furthermore that OA does not react with either the triplet acceptor or the sensitizer species.
After completion of the experiment, the samples were stored in the dark while remaining open
to air for several days and then measured for upconversion response. No time delay in the
appearance of the upconverted emission was observed, implying that the system remains O2
depleted even after several days of continuous air exposure (Figure 12), consistent with the
combined results described above. This furthermore indicates that the by-products of singlet
oxygen trapping, namely oleic acid peroxides, do not negatively impact the upconversion
system stability during long-term irradiation or storage.

Figure 6.12. Photochemical upconversion operating under aerated ambient conditions in the presence of OA.
(Left) Upconversion signal appearance and stability in PEG-400 solution composed by PtTPBP (20 µM), BPEA
(4 mM) and OA (30 mM) under 635 nm excitation (50 mW/cm
2
). Inset displays the early time of the process
revealing the delay in upconversion signal appearance due to oxygen consumption (635 nm at 50 mW/cm
2
).
(Right) Monitoring of oxygen diffusion through the signal delay of the upconversion appearance in PEG-400
solution kept in the dark and periodically exposed to irradiation (635 nm, 50 mW/cm
2
). Inset displays several
cycles of laser on/off experiments monitored over 60 hours.

0 100 3k 6k 9k 12k
0
1
2
3
4
0 10 20 30
0
1
2
3
4
5
Emission Intensity ( 10
5
cts)
Time (s)
Intensity ( 10
5
cts)
Time (s)
0 2 4 6 8 10
0
1
2
3
4
0 20 40 60
0
1
2
3
4
Emission Intensity ( 10
5
cts)
Time (h)
Intensity ( 10
5
cts)
Time (h)
Golden
202
6.2.4 Rapid Photophysical Screening of TTA-UC Systems in Air
::
Polyethylene glycol oligomer solvents have proven to be promising candidates as stable
media for hosting upconversion systems and possibly other oxygen-sensitive chemical and
photochemical reactions. Due to their high oxygen impermeability, the PEG solvents provide
the necessary environment to enable photophysical/photochemical processes involving triplet
excited states which would otherwise be quenched by dissolved oxygen (Figure 6.13). Once
paired with a suitable singlet oxygen scavenger, such as OA, the systems are rapidly depleted
of oxygen, permitting the extended photophysical evaluation of O2-sensitive photoreactions
under ambient conditions without the need for substantial preparative techniques. The oxygen
scavengers as well as their oxidation decomposition products were not observed to interact
with the PtTPBP or BPEA ground or excited states, thereby avoiding undesirable side
reactions. Triplet sensitized upconversion was readily achieved with this chromophoric pair,
exhibiting impressive photochemical stability over periods of several days and with high
quantum efficiency in comparison to previously established traditional organic solvent
compositions. This provided the necessary proof-of-principle that inexpensive, non-
flammable, safe and “green” polyethylene glycol polymers can be successfully utilized as
efficient media to host oxygen sensitive photochemical processes, thus opening the door to a
plethora of applications in solar
energy conversion, photocatalysis,
and imaging. Although the current
formulations focused on
photochemical upconversion
processes, the oxygen-depleted

Figure 6.13. Representation of the successive steps involved in
TTA based UC in PEG in presence of antioxidant in air.
1
[sens]
1
[sens]
*
3
[sens]
*
3
[O
2
]
1
[O
2
]
*
+ Anti-Ox Anti-Ox-O
2
1
[sens]
*
3
[sens]
*
1
[Em]
3
[Em]
*
h ν
ISC ISC
TTET
O
2
Trapping
TTA
1
[Em]
*
UC
Molecular Design for Organic Photovoltaics
Chapter 6 | 203
environments developed here can be utilized to study a variety of oxygen intolerant
photochemical reactivity.
6.3.1 Introduction to Photocatalytic Water Splitting
Photovoltaics convert solar energy directly into electrical energy, which must be used
immediately in a circuit or stored in a battery for future use. This poses many problems for
commercial implementation, as new battery technologies capable of high energy density
storage are currently lagging behind the ability of Generation 1 and Generation 2 photovoltaic
technologies to generate power. Grid-scale solar farms plague the grid with overloads during
daylight hours and no energy production during evening hours, when energy demand tends to
increase.
66
The development of high energy density batteries for grid-scale energy storage is
currently underway, but in the meantime, the development of alternative solar energy storage
methods is highly desirable.
67
The most ubiquitous method of solar energy storage is
photosynthesis; performed by plants and certain bacteria, photosynthesis converts solar energy
into chemical energy.
68-69
This process is so effective that it supports nearly all life on Earth
(indeed, until quite recently, it was thought to be the only energy supply for life).
70
Duplication
of highly evolved photosynthetic systems in an efficient and commercializable manner,
however, is a challenge.
71
Whereas photosynthesis stores solar energy in the form of sugars,
artificial photosynthesis processes can be designed to store energy in the form of more useful
fuels such as hydrogen gas. Photocatalytic water splitting is an artificial photosynthetic process
in which solar energy is used to dissociate water into molecular oxygen and molecular
hydrogen. This method of artificial photosynthesis is desirable, as it makes use of two naturally
abundant resources, solar energy and water, to generate a green (non-GHG production) fuel
Golden
204
(hydrogen gas). As with the development of next-generation organic photovoltaics, the
development of efficient photocatalytic water splitting hinges on molecular design.
In photocatalytic water splitting, a dye material absorbs incident photon flux,
generating an exciton which either serves as a donor or an acceptor, inducing charge transfer
(either intermolecularly or intramolecularly) to an acceptor or donor moiety. The separated
charges are then utilized by a catalyst (the catalyst can be the dye itself) to split water into H2
and O2. The electrochemical requirements for this process require photovoltages in excess of
1.23 eV at a pH of 0 at standard pressure and temperature, thus, molar absorptivity in excess
of this energy must be intense and the materials must be stable to aqueous environments under
operating photovoltages.
In this work, a zeolite-based nanoarchitecture is used to host BODIPY-based
photosensitizers.
‡
These photosensitizers are paired with a variety of hydrogen evolution
reaction (HER) catalysts to demonstrate efficient water splitting upon photoinduced electron
transfer from the BODIPY to the catalyst. Molecular design considerations for the BODIPY
photosensitizers are developed to increase fuel production in future designs.
6.3.1 Zeolite Nanoarchitectures for Photocatalytic Water Splitting
The architectures utilized in this study are based upon discotic zeolite nanoparticles
with vertical pore openings. The pores are loaded with an organic acceptor material and
plugged with a BODIPY-pyridinium donor material which also serves as a stopcock due to
strong ionic binding between the positively charged pyridinium moiety and the zeolite

‡
The architectures utilized in this study were designed and developed by Rebecca Wilson as part of her PhD
research in the Thompson Lab at USC; the synthesis and photocatalytic activity of these systems were studied by
Rebecca Wilson in collaboration with Professors Luisa De Cola at the Université de Strausbourg and Peter
Bruggeller at the Universität Innsbruck, Austria.
Molecular Design for Organic Photovoltaics
Chapter 6 | 205
structure. The fully assembled architecture
(Figure 6.14) is referred to as a light-
harvesting unit (LHU). BODIPY was
selected as a photosensitizer due to its high
molar absorptivity, high photoluminescent
quantum yields, demonstrated
photostability in aqueous media, and tunable energy and charge transfer kinetics.
72
The
acceptor moieties used to fill the pores of the zeolite were either anthracene or tetracene; the
acene type was varied in order to tune FRET efficiency in favor of photoinduced electron
transfer (PET). The BODIPY stopcock materials were synthesized based on either an alkyl- or
a benzoBODIPY, covalently bound at the meso position to a spacer group and a pyridinium,
where pyridinium serves both as a binding unit and as a PET-inducing moiety due to its
electron withdrawing effects. The architecture is designed and tuned such that FRET follows
an energy cascade from acene to BODIPY, and PET follows in the reverse direction, from
BODIPY through pyridinium to the acene.
6.3.2 Synthesis of BODIPY-Pyridinium Plugs
Three distinct plug architectures were synthesized for this study (Scheme 6.2). The
first, BODIPY-diPY, was synthesized from a meso-benzaldehyde derivative of 3-ethyl-2,4-
dimethylBODIPY. Although this synthesis was designed and executed using conditions known
to affect mono-functionalization with pyridinium on non-BODIPY substrates, no conditions
were found which yielded the mono-pyridinium BODIPY (BODIPY-Py) from these
substrates. Instead, functionalization occurred efficiently (48% yield), producing a BODIPY-
pyridinium structure in which through-space charge and energy transfer mechanisms are

Figure 6.14. Zeolite nanoarchitectures (yellow discs)
are loaded with acene acceptor materials (orange lines)
and capped with BODIPY-pyridinium photosensitizers
(blue spheres) to complete the light-harvesting unit
(LHU).

Golden
206
allowed but through-bond energy and charge transfer mechanisms are not. Though not the
intended consequence of this synthesis, this structure represented a unique opportunity to probe
energy and charge transfer mechanisms in conjugated versus non-conjugated BODIPY-
Scheme 6.2. Synthesis of BODIPY-pyridinium stopcock photosensitizers.

(a) piperidine, methanol, reflux 3h: 48% yield. (b) 2 eq. 2,4-dimethylpyridinium, 0.06 eq. piperidine, methanol,
reflux 6h. (c) 0.04 eq. V 2O 5, 4 eq. 30% H 2O 2 (aq), methanol, 0 ºC, 30 min: 94% yield. (d) [1] cat. TFA, 2 eq. 3-
ethyl, 2-4-dimethylpyrrole, DCM, 12h. [2] DDQ, 30 min. [3] diisoprolylethylamine, BF 3OEt 2, 3h. (e) [1] cat.
TFA, 2 eq. 4,4-dihydroisoindole, DCM, 12h. [2] 1 eq. DDQ, 30 min. [3] diisoprolylethylamine, BF 3OEt 2, 3h. [4]
2 eq. DDQ.
Molecular Design for Organic Photovoltaics
Chapter 6 | 207
pyridinium systems. The BODIPY-Py structure was synthesized from terepthaldehyde, which
was functionalized under the same reaction conditions as those used in the functionalization
the BODIPY aldehyde, though in this case, the mono-substituted, conjugated pyridinium-
phenyl-acetal intermediate was the only observed product. Deprotection of the acetal was
affected with vanadium(V) oxide, and the BODIPY-Py structure was synthesized under typical
pyrrole condensation reaction conditions at the aldehyde. Synthesis of benzoBODIPY-Py was
affected by the same mechanism.

Figure 6.15. (a) Diffuse reflectance spectra of LHUs with anthracene acceptors at various loadings: 0% (red), 25%
(light grey), 50% (grey), 75% (dark grey), and 100% (black) and capped with BODIPY-diPy; (b) emission spectra of
the same samples excited at 500 nm (BODIPY) and (c) excited at 360 nm (anthracene), and (d) TCSPC-lifetime decay
of the same samples, excited at 375 nm.

Golden
208
6.3.3 Energy and Charge Transfer Control in Zeolite Nanoarchitectures for
Photocatalytic Water Splitting
The nanoparticles were first loaded with acene, then capped with the photosensitizer.
Full BODIPY pore capping affected by calculating the number of pore channel openings per
sample and adding a stoichiometric quantity of the photosensizer to a solution of the acene-
loaded zeolite nanoparticles. Quantitative loading was determined by UV-Vis absorption of
the solution after incubation with the zeolite; when the solution showed no BODIPY
absorption, it was determined that all of the photosensitizer had adhered to the nanoparticle.
Upon washing with dichloromethane, no photosensitizer was observed to dissolve into
solution, indicating that the photosensitizers had migrated into the openings of the pores,
completing the LHU. A control sample of zeolite with no acene was also capped by this
method.
The energy transfer dynamics of the LHU were probes using diffuse reflectance,
photoluminescence, and time-correlated single-photon counting spectroscopies (Figure 6.15).
In the first experiment, various acene loadings with anthracene as the electron acceptor moiety
and 100% BODIPY-diPy capping was probed. It can be seen that both the anthracene and the
BODIPY efficiently absorb photon flux from 300 nm to 550 nm. When the BODIPY is
selectively excited, fluorescence exclusively from the BODIPY moiety is observed. This is
consistent with endothermic FRET from the low energy S1 state of the BODIPY to that of
anthracene. When anthracene is selectively excited, the predominant emission is from
anthracene itself, although there is some emission from the BODIPY which is observed,
indicating that while FRET efficiency is low, it is non-zero. FRET is further indicated by the
observation of a reduction in the photoluminescence lifetime of anthracene measured by
Molecular Design for Organic Photovoltaics
Chapter 6 | 209
TCSPC; it appears both from the photoluminescence and the lifetime spectra, that FRET is
made more efficient with increased anthracene loadings, possibly due to enhanced
wavefunction overlap between the BODIPY and the anthracene. benzoBODIPY-Py has a
markedly red-shifted absorption and emission profile relative to BODIPY-Py and is therefore
better matched to eliminate FRET from the acene moieties in the LHU (Figure 6.16). That
said, it is significantly more difficult to synthesize, as the isoindole must be prepared from a
protected intermediate, adding several steps to the overall synthesis.
19

Integrated quantum yields of the LHUs with no anthracene compared against those with
increased anthracene loadings are
observed to decrease. As FRET is
minimized in this system, this excited
state quenching appears to be due to
photoinduced electron transfer. As
FRET from BODIPY to anthracene is
endothermic and therefore not
observed, PET can be probed directly
by monitoring the  PL of the BODIPY photosensitizer relative to anthracene loadings.
However, although the PL of BODIPY-diPy is high in solution, (0.59, MeCN), it is upon
loading into the zeolite (0.02). This appeared to be due to self-quenching of the BODIPY,
which has a negligible Stokes shift and therefore strong photon reabsorption. To mediate this
self-quenching mechanism, a set of naphthalene-pyridinium stopcocks were synthesized to
dilute the BODIPY stopcocks. Naphthalene has a high S1 energy, so FRET from BODIPY to

Figure 6.16. BenzoBODIPY-Py absorption (solid dark red)
and emission (solid bright red) compared to benzoBODIPY
absorption (dashed) spectra.

Golden
210
naphthalene is highly endothermic, and its HOMO and LUMO energies are lower and higher
than those of BODIPY, so PET from BODIPY to naphthalene is also forbidden.
LHUs with variable naphthalene-pyridinium (N-Py) and BODIPY-diPY
concentrations were prepared. The absorption spectra were measured using diffuse reflectance
UV-visible spectroscopy, and the emission spectra from the BODIPY moieties were measured
by selective excitation at 490 nm. The absorption spectra feature two distinct peaks; one from
BODIPY (λmax = 520 nm) and another assigned to naphthalene (λmax = 375 nm). The BODIPY-
diPy was diluted from 100% to 1% and the quantum yield of the LHU was measured (Figure

Figure 6.17. (a) Diffuse reflectance spectra of N-Py: BODIPY:Py ratios 0:100 (black), 25:75 (dark red), 50:50
BODIPY-Py (red), 75:25 (orange) and 100:0 (blue); (b) emission spectra of the same samples, excited at 490
nm (note that no emission is observed in the pure N-Py sample); (c) quantum yield as a function of BODIPY-
Py concentration on the surface of the zeolite; and (d) lifetime decay of the same samples excited at 405 nm,
also including samples 80:20 (yellow) and 85:15 (light yellow) compared to that of a dilute solution of BODIPY-
diPy in acetonitrile (pink).
Molecular Design for Organic Photovoltaics
Chapter 6 | 211
6.17). Indeed, self-quenching of the BODIPY is observed; as the concentration of BODIPY
is increase, the quantum yield decreased. At high dilutions, however, the quantum yield
approaches that of solution. The lifetime decay, used as another indicator of self-quenching, as
show that decreasing the concentration of BODIPY on the surface result in lifetimes similar to
those of BODIPY-diPy in dilute solution.
Although an effective diluting agent to reduce self-quenching from BODIPY-diPy, the
naphthalene pyridinium complex absorbs in competition with BODIPY-diPy and therefore the
intermediate acetal (Scheme 7.2), which has an absorption spectrum in the UV has been
identified as a more suitable diluting agent to be paired with the BODIPY (Figure 7.18).
Diluting the concentration of photosensitizer, however, will negatively affect the
photocatalytic efficiency. It would be desirable to instead synthesize a stopcock structure
which features both a high  PL efficiency but also a large Stokes shift to mitigate self-
quenching effects without reducing the concentration of photosensitizer.
The photocatalytic activity of the LHU with 100% loading of anthracene and 100%
pore-stoppering with BODIPY-diPy was tested, as this system showed the greatest evidence

Figure 7.18. (left) Naphthalene-pyridinium absorption and emission (solid blue) spectra compared to
unfunctionalized naphthalene emission (dashed blue); (right) acetal-pyridinium absorption (green trace) – note
that no emission is observed.
Golden
212
for PET. Controls experiments in
which there was 0% BODIPY in the
LHU and one in which BODIPY-Py
was free in solution were also tested
to ensure that catalytic water
splitting is originating from the
BODIPY-anthracene charge transfer event (Table 6.3). Various hydrogen-evolution reaction
(HER) catalysts which had previously shown efficient water splitting when paired with
photosensitizers were prepared. It was found that the LHU performed best with [Ni(pyS) 3]
-

under basic conditions, demonstrating an excellent proof-of-principle, that zeolite
nanoarchitectures doped with judiciously selected photoactive donor and acceptor species are
capable of sensitizing catalysts for water splitting. When the complete LHU was used, the TON
was 465 after 20 hours, compared to the BODIPY-dipyridium which resulted in a low TON of
3.5; some photocatalytic activity in this system may have resulted from PET from BODIPY to
pyridinium. In the anthracene-only LHU, 0 TONs were recorded.
In an alternative experiment, palladium nanoparticles (Pd NP) to the surface of the
LHU. Pd NPs have been shown to catalyze water splitting upon photoexcitation, and this
system facilitates the PET process by adhering the catalyst directly to the photosensitizing
LHU system (Table 6.4). As a control experiment, heterogeneous solutions of LHUs and Pd
NPs were prepared to monitor photocatalytic efficiency
as a function of LHU-Pd NP interaction. From these data,
it is evident that homogeneous LHU-catalytic systems
result in better TONs and higher TOFs.
Table 6.3. Photocatalytic water
splitting with [Ni(pyS) 3]
-
HER catalyst
over 20h.
Photosensitizer TON

Z-A 100-BPy LHU 465
BODIPY-diPy 3.5
Z-A 100 0

Table 6.4. Photocatalytic water splitting efficiency with
palladium nanoparticle HER catalysts.

TON
20 h 40 h 60 h
Intramolecular
Z -Pd NP 0 - -
Z-A 100-Pd NP 90 103 -
Z-A 100-BD-Pd NP 122 163 180
Intermolecular
Z -Pd NP 0 - -
Z-A 100-Pd NP 0 - -
Z-A 100-BD-Pd NP 114 125 -

Molecular Design for Organic Photovoltaics
Chapter 6 | 213
This section represents part of an ongoing collaborative research effort; as such,
molecular design principles in the fabrication of LHUs for photocatalytic water splitting are
actively being developed. From these initial experiments, it appears that donor materials with
large Stokes shifts are desired to reduce self-quenching at the LHU surface. In addition, there
should be a large pseudo-Stokes shift between donor and acceptor, to minimize FRET between
these moieties, and PET should be favored by control of conjugation, polarization of the
exciton, and CT interaction between donor and acceptor. The molecular design principals
developed throughout the rest of this work can and should be applied to the development of
novel materials to solve problems related to the generation of non-GHG fuels from readily
available resources via artificial photosynthesis.
6.4 Experimental Methods
Synthesis
Platinum tetrapehnylbenzoporphyrin (PtTPBP) was synthesized as described in the
literature.
29
Synthesis of BODIPY-pyridinium stopcock materials are shown in Scheme 6.2.
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 N2. Four equivalents of 2,4-Dimethyl-3-ethylpyrrole (2.43 g,
19.7 mmol) were 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 one portion. After one hour, the reaction was quenched with saturated Na2S2O3
(50 mL), washed with saturated NaHCO3 (2  50 mL) and water (2  50 mL). The organic
layer was removed, dried over MgSO4, filtered and concentrated. The product was purified by
Golden
214
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 (400 MHz, CDCl3): 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 (400 MHz, CDCl3): 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. MALDI m/z for
C40H48B2F2N4 Calculated 682.4 Found 682.7.
Tetra(4,4-difluoro-3,5-dimethyl-4-bora-3a,4a-diaza-s-indacene) ethylene-
bridged hexenoporphyrin: The starting ethylene bridged pyrrole, 4,7-dihydro-4,7-ethano-
2H-isoindole, was synthesized according to literature procedure
73
. The ethylene bridged
tetraBODIPYhexenoporphyrin was made according to a modified literature procedure,
29

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, CDCl3), 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).
BDP-Por: Platinum (II) chloride (90 mg, 0.338 mmol) was added to dry, degassed benzonitrile
(100 mL), and the mixture was heated while stirring under N2 at 100 ˚C for 20 minutes, until
the platinum salts dissolved, turning the solution yellow. 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 2Cl2 and filtered to
remove solids and the filtrate was dried via rotavap to yield a bright green solid, which was
Molecular Design for Organic Photovoltaics
Chapter 6 | 215
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/CH2Cl2 to 97.5:2.5 CH2Cl2
/acetone, and the pure product was recovered as an olive green solution, which was further
purified by recrystallization from CH2Cl2/diethyl ether (36 mg, 41%). The
1
H and
13
C NMR
spectra matched literature spectra:
1
H-NMR (400 MHz, CDCl3)
14
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).
13
C NMR (400 MHz, CDCl3) δ 158.29, 143.33, 141.89, 137.46, 134.94,
134.70, 134.43, 131.11, 130.09, 126.19, 124.30, 119.91, 115.07, 15.07.
OPV Preparation and Testing. BDP1 was synthesized according to literature procedure.
74

C60 (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: C60 (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 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.
75

Golden
216
6.5 Bibliography
1. Ameri, T.; Dennler, G.; Lungenschmied, C.; Brabec, C. J., Organic tandem solar cells:
A review. Energ. Environ. Sci. 2009, 2 (4), 347-363.
2. Ameri, T.; Khoram, P.; Min, J.; Brabec, C. J., Organic Ternary Solar Cells: A Review.
Adv. Mater. 2013, 25 (31), 4245-4266.
3. Lu, L.; Xu, T.; Chen, W.; Landry, E. S.; Yu, L., Ternary blend polymer solar cells with
enhanced power conversion efficiency. Nat. Photon. 2014, 8, 716.
4. Hwang, I.; McNeill, C.; C. Greenham, N., Evolution of phase separation upon
annealing and the influence on photocurrent generation in ternary blend organic solar cells.
Synth. Met. 2014, 189, 63–68.
5. Yang, Y.; Chen, W.; Dou, L.; Chang, W.-H.; Duan, H.-S.; Bob, B.; Li, G.; Yang, Y.,
High-performance multiple-donor bulk heterojunction solar cells. Nat. Photon. 2015, 9, 190.
6. Rajesh, K. R.; Paudel, K.; Johnson, B.; Hallani, R.; Anthony, J.; Ostroverkhova, O. In
Design of organic ternary blends and small-molecule bulk heterojunctions: photophysical
considerations, SPIE: 2015; p 15.
7. Uchida, S.; Xue, J.; Rand, B. P.; Forrest, S. R., Organic small molecule solar cells with
a homogeneously mixed copper phthalocyanine: C60 active layer. Appl. Phys. Lett. 2004, 84
(21), 4218-4220.
8. Honda, S.; Ohkita, H.; Benten, H.; Ito, S., Multi-colored dye sensitization of
polymer/fullerene bulk heterojunction solar cells. Chem. Commun. 2010, 46 (35), 6596-6598.
9. Honda, S.; Nogami, T.; Ohkita, H.; Benten, H.; Ito, S., Improvement of the Light-
Harvesting Efficiency in Polymer/Fullerene Bulk Heterojunction Solar Cells by Interfacial
Dye Modification. ACS Appl. Mater. Interfaces 2009, 1 (4), 804-810.
10. Vijaya Kumar Naidu, B.; Su Park, J.; Chul Kim, S.; Park, S.-M.; Lee, E.-J.; Yoon, K.-
J.; Joon Lee, S.; Wook Lee, J.; Gal, Y.-S.; Jin, S.-H., Novel hybrid polymer photovoltaics made
by generating silver nanoparticles in polymer: fullerene bulk-heterojunction structures. Sol.
Energy Mater. Sol. Cells 2008, 92, 397-401.
11. Wang, C. C. D.; Choy, W. C. H.; Duan, C.; Fung, D. D. S.; Sha, W. E. I.; Xie, F.-X.;
Huang, F.; Cao, Y., Optical and electrical effects of gold nanoparticles in the active layer of
polymer solar cells. J. Mater. Chem. 2012, 22 (3), 1206-1211.
12. Fu, H.; Choi, M.; Luan, W.; Kim, Y.-S.; Tu, S.-T., Hybrid solar cells with an inverted
structure: Nanodots incorporated ternary system. Solid-State Electron. 2012, 69, 50-54.
13. Erwin, P.; Conron, S. M.; Golden, J. H.; Allen, K.; Thompson, M. E., Implications of
Multichromophoric Arrays in Organic Photovoltaics. Chem. Mater. 2015, 27 (15), 5386-5392.
14. 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.
J. Am. Chem. Soc. 2010, 133 (1), 88-96.
15. Lee, S.; Sarker, A. K.; Hong, J., A zinc porphyrin sensitizer modified with donor and
acceptor groups for dye-sensitized solar cells. Bull. Korean Chem. Soc. 2014, 35 (10), 3052-
3058.
16. 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. Phys. Chem. Chem. Phys. 2014, 16, 24755-24762.
Molecular Design for Organic Photovoltaics
Chapter 6 | 217
17. 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. Chem. Mater. 2012, 24 (13), 2583-2591.
18. 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. Adv. Mater. 2009, 21, 1517-1520.
19. Chen, J. J.; Conron, S. M.; Erwin, P.; Dimitriou, M.; McAlahney, K.; Thompson, M.
E., High-Efficiency BODIPY-Based Organic Photovoltaics. ACS Appl. Mater. Interfaces
2015, 7 (1), 662-669.
20. Kim, B.; Ma, B.; Donuru, V. R.; Liu, H.; Frechet, J. M. J., Bodipy-backboned polymers
as electron donor in bulk heterojunction solar cells. Chem. Commun. 2010, 46 (23), 4148-4150.
21. 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. Org. Lett. 2011, 13 (17), 4574-4577.
22. 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. Chem. Commun. 2010, 46 (28), 5082-
5084.
23. 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. J. Am. Chem. Soc. 1992, 114
(8), 2944-2950.
24. Sato, T.; Jiang, D.-L.; Aida, T., A Blue-Luminescent Dendritic Rod:
Poly(phenyleneethynylene) within a Light-Harvesting Dendritic Envelope. J. Am. Chem. Soc.
1999, 121 (45), 10658-10659.
25. Stewart, G. M.; Fox, M. A., Chromophore-Labeled Dendrons as Light Harvesting
Antennae. J. Am. Chem. Soc. 1996, 118 (18), 4354-4360.
26. 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. J. Phys. Chem. B 2004, 108 (34), 12821-12832.
27. Ziessel, R.; Ulrich, G.; Haefele, A.; Harriman, A., An Artificial Light-Harvesting Array
Constructed from Multiple Bodipy Dyes. J. Am. Chem. Soc. 2013, 135 (30), 11330-11344.
28. 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. Supramol. Chem. 2005, 17 (4), 283-289.
29. Ito, S.; Murashima, T.; Ono, N.; Uno, H., A new synthesis of benzoporphyrins using
4,7-dihydro-4,7-ethano-2H-isoindole as a synthon of isoindole. Chem. Commun. 1998, (16),
1661-1662.
30. Perez, M. D.; Borek, C.; Forrest, S. R.; Thompson, M. E., Molecular and
Morphological Influences on the Open Circuit Voltages of Organic Photovoltaic Devices. J.
Am. Chem. Soc. 2009, 131 (26), 9281-9286.
31. 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. Adv. Mater. 2013, 25 (42), 6076-6082.
Golden
218
32. Qin, W.; Baruah, M.; Van der Auweraer, M.; De Schryver, F. C.; Boens, N.,
Photophysical Properties of Borondipyrromethene Analogues in Solution. J. Phys. Chem. A
2005, 109 (33), 7371-7384.
33. Chen, P.; Finikova, O. S.; Ou, Z.; Vinogradov, S. A.; Kadish, K. M., Electrochemistry
of Platinum(II) Porphyrins: Effect of Substituents and π-Extension on Redox Potentials and
Site of Electron Transfer. Inorg. Chem. 2012, 51 (11), 6200-6210.
34. Renshaw, C. K.; Zimmerman, J. D.; Lassiter, B. E.; Forrest, S. R., Photoconductivity
in donor-acceptor heterojunction organic photovoltaics. Phys. Rev. B 2012, 86 (8), 085324.
35. Giebink, N. C.; Wiederrecht, G. P.; Wasielewski, M. R.; Forrest, S. R., Ideal diode
equation for organic heterojunctions. I. Derivation and application. Phys. Rev. B 2010, 82 (15),
155305.
36. 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. Adv. Funct. Mater. 2012, 22 (14), 3089-3094.
37. 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. Chem. Mater. 2011, 23 (18), 4132-
4140.
38. Khlyabich, P. P.; Burkhart, B.; Thompson, B. C., Efficient Ternary Blend Bulk
Heterojunction Solar Cells with Tunable Open-Circuit Voltage. J. Am. Chem. Soc. 2011, 133
(37), 14534-14537.
39. 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.
40. Bartynski, A. N.; Trinh, C.; Kirlikovali, K. O.; Thompson, M. E., Multichromophoric
energy sensitization of C60 for organic photovoltaics. Appl. Phys. Lett. 2014, 105 (11), 113305.
41. 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.
42. 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 Lett. 2012, 12 (8), 4366-4371.
43. 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. Org. Electron. 2015, 21, 160-170.
44. 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. Adv. Mater. 2014, 26 (23), 3839-3843.
45. 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. Appl. Phys. Lett. 2009, 95 (12), 123303.
46. Vandewal, K.; Tvingstedt, K.; Manca, J. V.; Inganäs, O., Charge-Transfer States and
Upper Limit of the Open-Circuit Voltage in Polymer:Fullerene Organic Solar Cells. IEEE J.
Sel. Top. Quantum Electron. 2010, 16 (6), 1676-1684.
Molecular Design for Organic Photovoltaics
Chapter 6 | 219
47. 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.
48. Zou, Y.; Holmes, R. J., Correlation between the Open-Circuit Voltage and Charge
Transfer State Energy in Organic Photovoltaic Cells. ACS Applied Materials & Interfaces
2015, 7 (33), 18306-18311.
49. 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.
50. 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.
51. 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.
52. Schulze, T. F.; Schmidt, T. W., Photochemical upconversion: present status and
prospects for its application to solar energy conversion. Energy & Environmental Science
2015, 8 (1), 103-125.
53. Auzel, F., Upconversion and Anti-Stokes Processes with f and d Ions in Solids. Chem.
Rev. 2004, 104 (1), 139-174.
54. Mongin, C.; Golden, J. H.; Castellano, F. N., Liquid PEG Polymers Containing
Antioxidants: A Versatile Platform for Studying Oxygen-Sensitive Photochemical Processes.
ACS Applied Materials & Interfaces 2016, 8 (36), 24038-24048.
55. DeRosa, M. C.; Crutchley, R. J., Photosensitized singlet oxygen and its applications.
Coord. Chem. Rev. 2002, 233-234, 351-371.
56. Mexal, J.; Fisher, J. T.; Osteryoung, J.; Reid, C. P. P., Oxygen Availability in
Polyethylene Glycol Solutions and Its Implications in Plant-Water Relations. Plant Physiology
1975, 55 (1), 20.
57. Lin, H.; Freeman, B. D., Gas solubility, diffusivity and permeability in poly(ethylene
oxide). Journal of Membrane Science 2004, 239 (1), 105-117.
58. Blatt, E.; Treloar, F. E.; Ghiggino, K. P.; Gilbert, R. G., Viscosity and temperature
dependence of fluorescence lifetimes of anthracene and 9-methylanthracene. The Journal of
Physical Chemistry 1981, 85 (19), 2810-2816.
59. Borisov, S. M.; Nuss, G.; Haas, W.; Saf, R.; Schmuck, M.; Klimant, I., New NIR-
emitting complexes of platinum(II) and palladium(II) with fluorinated benzoporphyrins.
Journal of Photochemistry and Photobiology A: Chemistry 2009, 201 (2), 128-135.
60. Marsico, F.; Turshatov, A.; Peköz, R.; Avlasevich, Y.; Wagner, M.; Weber, K.;
Donadio, D.; Landfester, K.; Baluschev, S.; Wurm, F. R., Hyperbranched Unsaturated
Polyphosphates as a Protective Matrix for Long-Term Photon Upconversion in Air. J. Am.
Chem. Soc. 2014, 136 (31), 11057-11064.
Golden
220
61. Min, D. B.; Boff, J. M., Chemistry and Reaction of Singlet Oxygen in Foods.
Comprehensive Reviews in Food Science and Food Safety 2006, 1 (2), 58-72.
62. Lissau, J. S.; Nauroozi, D.; Santoni, M.-P.; Edvinsson, T.; Ott, S.; Gardner, J. M.;
Morandeira, A., What Limits Photon Upconversion on Mesoporous Thin Films Sensitized by
Solution-Phase Absorbers? The Journal of Physical Chemistry C 2015, 119 (9), 4550-4564.
63. Bouas-Laurent, H.; Castellan, A.; Desvergne, J.-P.; Lapouyade, R., Photodimerization
of anthracenes in fluid solution: structural aspects. Chem. Soc. Rev. 2000, 29 (1), 43-55.
64. Gijzeman, O. L. J.; Kaufman, F.; Porter, G., Oxygen quenching of aromatic triplet
states in solution. Part 1. Journal of the Chemical Society, Faraday Transactions 2: Molecular
and Chemical Physics 1973, 69 (0), 708-720.
65. Haefele, A.; Blumhoff, J.; Khnayzer, R. S.; Castellano, F. N., Getting to the (Square)
Root of the Problem: How to Make Noncoherent Pumped Upconversion Linear. The Journal
of Physical Chemistry Letters 2012, 3 (3), 299-303.
66. Ipakchi, A.; Albuyeh, F., Grid of the future. IEEE Power and Energy Magazine 2009,
7 (2), 52-62.
67. Dunn, B.; Kamath, H.; Tarascon, J.-M., Electrical Energy Storage for the Grid: A
Battery of Choices. Science 2011, 334 (6058), 928.
68. Ross, R. T.; Calvin, M., Thermodynamics of Light Emission and Free-Energy Storage
in Photosynthesis. Biophys. J. 1967, 7 (5), 595-614.
69. Meech, S. R.; Hoff, A. J.; Wiersma, D. A., Evidence for a very early intermediate in
bacterial photosynthesis. A photon-echo and hole-burning study of the primary donor band in
Rhodopseudomonas sphaeroides. Chem. Phys. Lett. 1985, 121 (4), 287-292.
70. Beatty, J. T.; Overmann, J.; Lince, M. T.; Manske, A. K.; Lang, A. S.; Blankenship, R.
E.; Van Dover, C. L.; Martinson, T. A.; Plumley, F. G., An obligately photosynthetic bacterial
anaerobe from a deep-sea hydrothermal vent. Proceedings of the National Academy of Sciences
of the United States of America 2005, 102 (26), 9306.
71. Listorti, A.; Durrant, J.; Barber, J., Solar to fuel. Nature Materials 2009, 8, 929.
72. Xie, A.; Pan, Z.-H.; Yu, M.; Luo, G.-G.; Sun, D., Photocatalytic hydrogen production
from acidic aqueous solution in BODIPY-cobaloxime-ascorbic acid homogeneous system.
Chin. Chem. Lett. 2018.
73. 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.
74. 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.
75. 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.

Molecular Design for Organic Photovoltaics
Chapter 7 | 221
Chapter 7

Tuning State Energies for Narrow Blue Emission in
Tetradentate Pyridyl-Carbazole Platinum Complexes

7.1 Introduction to Light-Emitting Diodes
Organic light-emitting diodes (OLEDs) function according to principles which are
directly analogous to OPVs. Whereas in the case of an OPV, photon energy is converted into
electrical current, in an OLED, electrical current is converted into photon energy. In its
simplest form, an OLED involves the injection of charges into an emissive organic layer.
Where these charges recombine, an exciton is formed, and exciton decay via fluorescence or
phosphoresce results in the emission of a photon. Electrogenerated excitons are formed as
either a singlet (25%) or a triplet (75%) based upon spin statistics.
1
The internal quantum yield
in OLEDs has been pushed to its theoretical limit, 100%, due to the advent of phosphorescent
OLEDs, which utilize spin-orbit coupling to convert all excitons to triplets which phosphoresce
at unity quantum yield.
2
Due to this technological advancement, OLEDs hold much promise
as the premier technology for the next generation of display and lighting technologies. They
are highly energy efficient, color-tunable, and provide the option of novel form factors such as
curved displays and flat-panel lighting. The compatibility of amorphous organic materials with
a wide range of substrates and deposition techniques has enabled their adoption in
state-of-the-art displays and lighting for televisions, mobile devices, wearables, automotive
and aerospace applications.
3
Rapid research progress and commercial adoption in these fields
has been achieved over the past several years but major deficiencies still remain. In particular,
Golden
222
the absence of a stable and efficient blue phosphorescent OLED have stunted operational
lifetimes, energy efficiencies and costs for both display and lighting applications.
4

7.2 Materials Design Considerations to Improve Device Lifetimes and Efficiencies
Highly efficient blue OLEDs require phosphorescent dopants with high energy triplets
(2.8 eV), unity electroluminescent quantum yield, and high stability under operating
conditions. Materials design considerations to affect each of these qualities, however, are not
yet fully developed. Amongst the possible classes of phosphorescent organometallic materials,
those based on octahedral Ir(III) complexes have shown the most promise due to their ability
to harvest 100% of electrogenerated excitons.
2, 5
However, operational lifetimes of blue Ir(III)
devices remain much lower than green and red devices due to degradation pathways caused by
the high energy exciton, and materials capable of narrow, blue phosphorescence emission are
rare.
6-7
Pure organic thermally-activated delayed fluorescent (TADF) materials as well as
materials based on other metals such as Cu(I), Au(I), Au(III), Pd(II), Pt(II) have emerged over
the past several years with device efficiencies matching or exceeding those of Ir(III)
derivatives, though many of these new phosphors have yet to demonstrate comparable
operational stabilities.
8-12
This work makes an effort to establish design criteria, based upon
the same principles outlined in previous chapters for the development of OPV materials, for
the synthesis of platinum-based phosphorescent materials capable of meeting the challenges
inherent to highly efficient and stable blue OLEDs.
*

*
This work was performed in collaboration with Tyler Fleetham as part of his post-doctoral research in the
Thompson Group, and going forward this work remains a collaborative effort with him at Universal Display
Corporation.
Molecular Design for Organic Photovoltaics
Chapter 7 | 223
7.3 Tetradentade Platinum Phosphors in Organic Light Emitting Diodes
Recently, the success of Ir(III) complexes has been challenged by highly stable Pt(II)
complexes utilizing rigid tetradentate ligands.
13
Pt(II) complexes with tetradentate ligands,
particularly those containing pyridyl-carbazole moieties, have shown remarkable operational
lifetimes, highly saturated colors, and very high efficiencies.
14
However, the simultaneous
achievement of saturated blue emission, high efficiency, and long operational lifetime remains
elusive for this class of materials as it has been for more the more established family of blue
Ir(III) phosphors. The recent report of a tetradentate Pt complex, PtNON, with all 6-membered
chelates, however, demonstrated an
operational lifetime over 600 hours
at 1000 cd/m
2
for an emission with a
triplet energy in excess of 2.8 eV.
15

The absence of 5-membered azole
rings or fluorinated ligands, each
thought to be a source of device
degradation, as well as its rigid
design may be partially responsible for this extended operational lifetime. Unfortunately, the
charge transfer (CT) character of the lowest energy triplet excited state in PtNON at room
temperature (Figure 7.1) leads to phosphorescence with an emission maximum at 492 nm
(green), rendering it unsuitable for display-blue pixels or high color-rendering index (CRI)
white lighting.
15
It is highly desirable to develop emitters with similar structures (and therefore
operational lifetimes and quantum efficiencies) to PtNON, but with narrower emission
characteristics that can be tuned for display and lighting applications. In this work, the design,
synthesis, and characterization of a series of functionalized PtNON derivatives with variable
400 450 500 550 600 650 700
0.0
0.5
1.0
77K
PMMA
MeCyHex
2-MeTHF
DCM
MeCN

Normalized PL Intensity
Wavelength (nm)

Figure 7.1. Room Temperature emission spectrum of PtNON
in various media.
Golden
224
lineshapes and emission maxima ranging from 446 nm to 590 nm are presented and materials
design considerations for narrow, high efficiency emission from platinum phosphors are
considered.
Recently, a few reports have demonstrated narrow phosphorescence emission from
platinum organometallic complexes through a variety of ligand designs and substitutions.
16-18

However, a sufficient paradigm for blue-shifting and narrowing phosphorescent emission in
these materials has been lacking. Li et al. have shown that functionalization of phenyl-pyrazole
(ppz) based Pt complexes (PtON1) is a method which can be used to tune the lineshapes of
room temperature emission spectra.
18
The proposed explanation in that work was that narrow
emission can be achieved by suppressing the contribution of metal to ligand charge transfer
(MLCT) from the pyridyl-carbazole relative to the ligand centered (LC) triplet of the ppz
fragment. They demonstrated the utility of this strategy by substituting the pyridine with
various electron donating (stabilized MLCT) or electron withdrawing (destabilized MLCT)
moieties to make the phosphorescent emission broader (CT) or sharper (LC) respectively.
Although this work served to illuminate the utility of state energy tuning to achieve narrowed
emission features, the model reported therein proved insufficient to explain the 77 K emission
properties of other previously reported ppz emitters, such as PtOO1 which has very large first
and second vibronic progressions, or to explain the similarity in the 77 K emission spectra of
PtNON to that of PtON1.
19-20
In fact, many of the tetradentate complexes containing
carbazole-pyridine ligands, for example PtON1, PtON7-dtb, and PtNON, showed similar sharp
features in their cryogenic emission spectra.
15-16, 20
Thus, it appears that the narrow emission
in the PtON1 derivatives is not from the ppz portion of the ligand but rather from the
pyridyl-carbazole moiety that is common to each of these emitters.
Molecular Design for Organic Photovoltaics
Chapter 7 | 225
As can been seen in Figure 7.1, PtNON has a positively solvatochromic
phosphorescence emission at room temperature. In non-polar methylcyclohexane, the emission
max is at 480 nm, while in acetonitrile, the max is bathochromically shifted to 505 nm,
indicating that room-temperature emission is from a charge transfer state which is stabilized in
polar media. In 2-methyltetrahydrofuran (2-MeTHF) at 77 K, however, phosphorescence in
PtNON is characterized by a narrow, vibronically featured emission profile, and in room
temperature dilute films in poly(methylmethacrylate) (PMMA), a high energy shoulder distorts
the emission compared to fluid solution, indicating that there is also another, higher-energy
triplet state involved in the emission that is more favored in a rigid matrix. Ultimately, in a
device setting the charge transfer triplet character dominates and a broad, structureless
electroluminescence (EL) is observed. Ideally, an EL spectrum with a narrow character like
that observed in the 77 K emission spectrum of PtNON is highly desired in order to blue-shift
LUMO = -1.36 eV
LUMO = -1.34 eV
LUMO = -1.11 eV
HOMO = -5.01 eV HOMO = -5.01 eV
HOMO = -4.98 eV
E
T
= 2.80 eV E
T
= 2.87 eV
E
T
= 2.89 eV
LUMO = -2.0 eV
HOMO = -5.15 eV
E
T
= 2.32 eV

Figure 7.2. Top row: molecular structure of PtNON derivatives. Middle row: LUMO densities and energies.
Bottom row: HOMO densities and energies. DFT calculations were performed at the B3LYP / LACVP** level of
theory in a dichloromethane room temperature solvent continuum.
Golden
226
the emission color without increasing the triplet energy. Thus, in this work we have attempted
to tune the relative contributions of the competing triplet states at room temperature in order
to achieve efficient and narrow blue emission.
7.4 Tuning State Energies in PtNON Derivatives
To elucidate the origin of the narrow emission in pyridyl-carbazole based tetradentate
Pt complexes, four PtNON derivatives with various substitutions at the 4-position of the
pyridine were modeled (Figure 7.2) via DFT at the B3LYP / LACVP** level of theory. In the
rest of this text, the unsubstituted PtNON complex will be referred to as PtNON-H, the other
derivatives being PtNON-CF3, PtNON-Me, and PtNON-OMe. As can be seen from DFT
calculations, the lowest unoccupied molecular orbitals (LUMO) densities are localized on the
pyridine rings in all cases and the highest occupied molecular orbitals (HOMO) densities are
all localized predominantly on the platinum and the platinum-adjacent rings of the carbazole
moieties. Consequently, the HOMO energies are minimally influenced by the pyridine
substitutions, whereas the LUMO energies vary dramatically from -2.0 eV for PtNON-CF3
to -1.11 eV for PtNON-OMe.
The energies and natural transition orbitals of the two lowest-energy triplet states in all
four derivates were modeled using TD-DFT at the CAM-B3LYP / LACVP** level of theory
with the integral equation formalism variant polarizable continuum model (IEF-PCM) of
tetrahydrofuran, as this method most accurately described the energies for the previously
observed triplet states in PtNON-H. It can be seen in Figure 7.3 that the lowest energy triplet
T1 in PtNON-OMe, PtNON-Me, and PtNON-H is a locally excited (LE) state centered on the
carbazole moiety, where both the hole and the electron comprising the triplet state are localized
predominately on the carbazole, with some particle character extending onto the pyridine in
Molecular Design for Organic Photovoltaics
Chapter 7 | 227
the case of PtNON-H. In PtNON-CF3, where the LUMO density is pyridine-centered and
lower in energy than the other three complexes due to the strong electron withdrawing effect
of the trifluromethyl group, the T1 state is CT in character, wherein the hole is delocalized
across the carbazole and pyridine, and the particle is confined to the pyridine. In PtNON-OMe,
the difference in energy between the T1 (LE) and T2 (CT) energies was calculated to be 90
meV, nearly fourfold greater than kT at room temperature; in PtNON-Me and PtNON-H, the
difference between T1 and T2

was calculated to be much
smaller (40 and 50 meV,
respectively), closer to kT and
indicating that temperature and
solvation effects may have a
greater effect on the emitting
state in these species than in
either of the two more extreme
cases. Together, these data
suggest that destabilizing the
carbazole-pyridine CT state
relative to the ligand-centered carbazole LE state could be accomplished by a single site-
substitution affecting the pyridine-centered LUMO energy, leading to a narrowing and
blue-shifting of the emitting state.
Each of the four modeled complexes, PtNON-CF3, PtNON-H, PtNON-Me, and
PtNON-OMe were synthesized according to Scheme 7.1 and their electronic and

Figure 7.3. Natural transition orbitals for the T 1 (bottom) and T 2
(top) states in PtNON derivatives. In PtNON-CF 3, the T 1 state is CT

in character, whereas in the other three complexes the T 1 state at
equilibrium as modeled in tetrahydrofuran is a locally excited (LE)
state centered on the carbazole.
Golden
228
photophysical behaviors were compared with calculated values to probe the effects of tuning
the relative energies of the LE and CT states. In each of the four compounds, the oxidation
potentials, which can be used to approximate the HOMO energy, remained relatively constant
(Table 7.1).
21
Two distinct reduction potentials were observed in all of the complexes except
PtNON-OMe. The first reduction potential, shared by all the derivatives, remained relatively
constant as well (-1.5 eV). It has been shown in the literature that the reduction potential can
be used to approximate the LUMO energy, so this potential is assigned to reduction of the
carbazole moiety, the energy of which remains unchanged between the four derivatives.
22
The
second, deeper reduction potential varies by 0.7 V between PtNON-CF3 and PtNON-Me, and
is assigned to reduction of the pyridine moiety. This reduction potential varies significantly
between the derivatives, consistent with the trend observed in the calculated LUMO values.
Errors in accuracy between the DFT-derived LUMO levels and the calculated values may stem
from the fact that calculations were simulated for tetrahydrofuran solutions, whereas these
electrochemically derived orbital energies are calculated for acetonitrile solutions.
Scheme 1. Synthesis of pyridine-substituted PtNON derivatives

Molecular Design for Organic Photovoltaics
Chapter 7 | 229
Nevertheless, the trend in pyridine-centered, electrochemically derived LUMO energies
follows that observed from DFT calculations.
The absorption spectra of the four PtNON derivatives in 2-MeTHF solution are
depicted in Figure 7.4. Each of the four compounds is characterized by a strong absorption
feature between 300-350 nm, attributed to π-π* transitions on the ligands. There is a
characteristically broad charge transfer absorption band evident between 350 and 500 nm in
the absorption of PtNON-CF3, which is assigned to carbazole to pyridine charge transfer. The
CT absorption in PtNON-H is blue-shifted relative to PtNON-CF3, and that of PtNON-Me is
even further blue-shifted. In PtNON-OMe, there is no discernible CT absorption, although it
is possible that one exists which is blue-shifted relative to the methyl derivative and therefore
obscured by the π-π* absorption feature.
These absorption characteristics indicate
the expected tuning of the charge transfer
energies by pyridine substitution, with
CT energy for PtNON-OMe >
PtNON-Me > PtNON > PtNON-CF3.
Table 7.1. Redox potentials (V) acquired by DPV in acetonitrile vs Fc
+/0
, derived HOMO and LUMO levels (eV),
and triplet energies (eV) and assignments.
E ox. E red.
(pyr)
E red.
(cbz)
HOMO* LUMO
(pyr)*
LUMO
(cbz)*
T 1** T 2**
PtNON-CF3 0.51 -1.96 -2.76 -5.31 -2.45 -1.50 2.26 / CT --
PtNON-H 0.45 -2.41 -2.65 -5.36 -1.91 -1.63 2.76 / LE 2.81 / CT
PtNON-Me 0.41 -2.56 -2.84 -5.17 -1.73 -1.40 2.80 / LE 2.84 / CT
PtNON-OMe 0.36 -- -2.76 -5.10 -- -1.50 2.83 / LE 2.92 / CT
*HOMO and local LUMO levels were derived from redox potentials according to Refs 21-22 and assigned
according to DFT calculations. **T 1 and T 2 energies obtained from TD-DFT. (--) Not observed.
300 350 400 450 500 550
0
1
2
3
4

Extinction Coefficient (M
-1
cm
-1
x10
4
)
Wavelength (nm)
PtNON-OMe
PtNON-Me
PtNON-H
PtNON-CF
3

Figure 7.4. Absorption spectra of PtNON derivatives in
2-MeTHF

Golden
230
The emission spectra of the four complexes in 2-MeTHF at room temperature, in
frozen 2-MeTHF solution at 77 K, and in 1% doped PMMA thin films are shown in Figure
7.5 and their corresponding photophysical data are tabulated in Table 7.2. The 77 K frozen
solution, room temperature solution and doped PMMA film of PtNON-CF3 show broad,
featureless emission spectra in accordance with the strong charge transfer character that was
observed in their respective absorption spectra. The room temperature solution gave a weak,
broad orange-red emission band with a maximum at 590 nm which tailed past 700 nm. The
low quantum yield ( ) of 0.02 and the excited state decay time ( ) of 74 ns indicate a very
high non-radiative rate. The non-radiative decay rate is suppressed in PMMA (  = 0.20,  = 1
400 450 500 550 600 650 700 750
0.0
0.5
1.0
77K RT PMMA
PtNON-H

400 450 500 550 600 650 700 750
0.0
0.5
1.0
Wavelength (nm)
77K RT PMMA
PtNON-OMe

400 450 500 550 600 650 700 750
0.0
0.5
1.0

Normalized PL Intensity
Wavelength (nm)
77K RT PMMA
PtNON-Me
400 450 500 550 600 650 700 750
0.0
0.5
1.0
Normalized PL Intensity
77K RT PMMA
PtNON-CF
3

Figure 7.5. Photoluminescent spectra of substituted PtNON derivatives at 77 K (black traces), and room
temperature in a dilute PMMA film (blue trace) and in fluid 2-MeTHF solution.

Molecular Design for Organic Photovoltaics
Chapter 7 | 231
s), and is further suppressed in frozen solution at 77 K (  = 7.4 s). These data suggest that
the highly stabilized CT state, which delocalizes triplet spin density across the chromophore,
is more sensitive to structural distortions than the more carbazole-localized triplet state
observed in the parent structure PtNON.
PtNON-H, which shows a much higher energy CT absorption than the CF3 derivative,
still shows CT emission at room temperature  max = 492 nm in 2-MeTHF solution and 469
nm in PMMA). Unlike the CF3 derivative, PtNON-H exhibits highly efficient emission in both
solution (  = 0.60) and PMMA (  = 0.99) with relatively short excited state decay times of
3.1 s and 4.2 s, respectively. Moreover, in frozen solution PtNON-H forms a narrow and
highly structured emission at 77 K, accompanied by a much longer radiative lifetime of 11.5
s. The sharp, structured emission and longer radiative lifetime indicates a change in the nature
of the emissive state from mostly charge transfer in character to a ligand centered state in
frozen solution.
PtNON-Me, which has an even higher energy CT absorption band  max = 480 nm),
shows a blend of vibronically structured and CT emission depending on the matrix. In fluid
solution, PtNON-Me phosphorescence is characterized by an efficient (  = 0.79), mostly
structureless emission band  max = 478 nm,  = 4.5 s). In PMMA, however, the emission is
narrowed and vibronically structured  max = 446 nm,  = 0.95), and the excited state decay
Table 7.2. Summary of photophysical properties of PtNON derivatives
2-MeTHF
(r.t)
2-MeTHF
(77K)
PMMA
(r.t.)
λ max
(nm)
Φ τ
(μs)
k r
(x10
4
s
-1
)
k nr
(x10
4
s
-1
)
λ max
(nm)
τ
(μs)
λ max
(nm)
Φ τ (μs)
-OMe 446 0.26 13.1 2.0 5.6 438 39.9 446 0.89 27.1
-Me 478 0.79 4.5 17.5 4.7 438 25.9 446 0.95 3.21 (38%) + 9.59 (62%)
-H 492 0.60 3.1 19.3 12.9 440 11.5 469 0.99 4.16
-CF3 590 0.02 0.074 27.0 1324.3 514 7.9 539 0.20 0.45 (38%) + 1.28 (62%)

Golden
232
time is described by a two-component  with a similarly fast 3.2 s component (38%
contribution) and a slower 9.6 s component (62% contribution) compared to PtNON-H.
Cooling a dilute solution of PtNON-Me to 77 K again sharpens the emission relative to room
temperature and results in a significantly lengthened radiative lifetime (25.9 s). The increased
lifetime indicates that PtNON-Me phosphorescence has much more ligand-centered character
than the CF3 or unsubstituted derivatives. The biexponential excited state decay time at room
temperature in the polymer matrix of this species further indicates that there may be two
competing emission processes; i.e. both the LE
3
(T1) and CT
3
(T2) states emit efficiently at
room temperature. An alternative possibility is that the film contains two conformers, each of
which is observed to emit with equal intensity. However, this seems improbable, as neither the
PtNON-H nor the PtNON-OMe films exhibit biexponential decay.
In the fourth derivative, a very strong electron donating group (methoxy) was selected
to destabilize the LUMO energy of the pyridine ring. This substitution raises the energy of the
PtNON-OMe CT state so much so that it is no longer observed in the emission spectrum in any
media. The resultant emission is defined by a narrow, vibronically featured spectrum in all
matrices with max = 446 nm for the room temperature solution and film spectra and
max = 438 nm in the frozen solution at 77 K. This more ligand-centered emission resulted in
much longer excited state decay times of 13.1 s in 2-MeTHF solution (  = 0.26) and 27.1 s
in PMMA (  = 0.89). In this case, the highly destabilized CT state is not involved in the
emission and a long-lived ligand centered emission dominates.
The observation of identical max in PtNON-OMe, PtNON-Me, and PtNON-H in frozen
solutions suggest that phosphorescence in these species under rigid, cryogenic conditions, is
dominated by a state which they all share (Table 2). This is strong support for the hypothesis
Molecular Design for Organic Photovoltaics
Chapter 7 | 233
that narrow, structured emission in carbazole-containing tetradentate platinum(II) complexes
is a result of emission from a carbazole-centered triplet state. In rigid films at room
temperature, both PtNON-OMe and PtNON-Me again appear to share emission from states of
identical origin (Table 3). It is clear that there is some rigidochromic effect at play in these
systems which modulates the degree of CT and LE character in the emitting state, as can be
seen especially in PtNON-Me and PtNON-H, where the CT and LE states are closer in energy.
Whereas cryogenic solutions are characterized by minimal structural distortion, room
temperature solid films still allow for moderate structural distortion. It appears that the CT
state energy is highly dependent on the matrix in all PtNON derivatives where it is observed,
and it is thus likely characterized by a significant change in geometry. One possibility for a
structural change may involve bond stretching and possibly a twisting of the pseudo-square
planar Pt(II) ground-state geometry to a more helix-like excited state geometry approaching
that of a tetrahedron; if the CT state is indeed characterized by a carbazole-platinum to pyridine
charge transfer, the resulting oxidation of Pt(II) to Pt(III) would provide a driving force
favoring tetrahedral geometry. This distortion, which tends to be inherent to CT states, may
also lead to a significant increase in the non-radiative rate, and should be carefully mediated
whenever high PLQY is desired.
Likewise, it can be seen from the deleterious effect on the radiative rate and the PLQY
of PtNON-OMe compared to PtNON-Me that localizing the spin density too much onto the
carbazole reduces the degree of spin-orbit coupling and thus limits the efficiency of
phosphorescence relative to non-radiative decay. Thus, it appears that mediating the degree of
CT character in the T1 is of paramount importance when attempting to both narrow and
blue-shift phosphorescence emission while maintaining the high luminescent efficiencies and
Golden
234
fast radiative rates which are desirable for OLED and lighting applications. The tunability of
the CT state by simple para-substitution of the pyridine ring with electron-donating or
withdrawing groups is a promising paradigm for application-specific derivatization of related
complexes.
7.5 PtNON Suitability for Lighting and Display Blue OLEDs
The combination of high photoluminescent quantum efficiency and fast radiative rate
in PtNON-Me indicated that it could be an excellent representative for efficient, narrow blue
emission from a platinum-based phosphor in an organic light-emitting diode. As a preliminary
study for its suitability in an organic charge-transport host matrix, its radiative efficiency in
poly(vinylcarbazole) (PVK) films was measured (Table 7.3, Figure 7.6). It was found that its
radiative efficiency drops significantly in PVK relative to PMMA, and its emission profile red-
shifts to a turquoise green color (474 nm) which is not suitable for display or lighting
applications. It appears from these data that in the polar PVK matrix, the CT state is stabilized
relative to the LE state. The dramatic difference in emission profile between PMMA and PVK
films indicate that the nature of the T1 state can be controlled according to the nature of the
host and perhaps even the deposition conditions. Further studies of vapor deposited films and
OLED devices utilizing PtNON-Me as the dopant are underway.
In this work, molecular design principles were utilized to tune the relative energies of
two closely-lying triplet state energies. Like in the OPV research discussed in previous
Table 7.3. Photophysical properties of PtNON-Me doped films in PMMA and PVK
 max (nm)  PL  (us)
1 wt% PMMA 446 0.69 4.00
1 wt% PVK 474 0.19 0.343 (54%) + 1.20 (46%)
10 wt% PVK 479 0.28 0.486 (38%) + 1.45 (62%)

Molecular Design for Organic Photovoltaics
Chapter 7 | 235
sections, molecular design in OLEDs has played a key role in the rapid advancement of this
energy-efficient technology, and will continue to do so as more efficient lighting and displays
are developed. PtNON may not be the champion blue dopant that has been sought after in the
two decades since the discovery of phosphorescent OLEDs, but the lessons derived from its
development can be used to synthesize more chemically robust, narrow blue phosphors in the
future. Importantly, it was observed that structural distortion should be limited in order to affect
narrower, blue-shifted emission with short excited state lifetimes and high photoluminescent
and electroluminescent quantum yields. In PtNON, this is accomplished through forming a
tetradentate ligand; the structure of this ligand, however, provides some freedom for distortion
which is reflected in the rigidochromic effect on the CT state. In next-generation derivatives,
a more planar, structurally restricted ligand geometry may help eliminate this deleterious
effect. Further, it was seen that judicious selection of site-substitution can be used to modulate
the energy of one state without affecting the other. These principles, building upon a physical-
theoretical foundational understanding of frontier molecular orbital densities and their role in

Figure 7.6. PtNON-Me in poly(methylmethacrylate) (PMMA) and poly(vinylcarbazole) (PVK) films at varying
doping percentages.

Golden
236
affecting the juxtaposition of state energies, apply to all optoelectronic molecular design
efforts.
7.6 Experimental Methods
2a-d. A 0.25M solution of pyridine 1 (1.1 eq.) in 20 mL dry toluene was added to a dry 100
mL Schlenk flask equipped with a condenser and containing 2-bromocarbazole (1.0 eq.),
lithium t-butoxide (1.5 eq.), copper(I) chloride (0.05 eq.), and 1-methylimidazole (0.10 eq.).
The solution was sparged with nitrogen, then refluxed overnight with stirring. The reaction
was cooled, quenched with 100 mL water, and extracted into ethyl acetate (3x100 mL). The
organic layers were combined, dried over sodium sulfate, and condensed by rotary evaporation.
The product was purified by silica gel flash chromatography with eluent 30% ethyl acetate in
hexanes, followed by recrystallization from a saturated dichloromethane solution layered with
hexanes.
2c: 1.61g (94%) yield. Colorless powdery solid.
1
H NMR (400 MHz, CDCl3) δ 8.54 (dd, J =
5.8, 0.5 Hz, 1H), 8.07 (ddd, J = 7.8, 1.3, 0.7 Hz, 1H), 8.04 – 8.00 (m, 1H), 7.97 – 7.92 (m, 1H),
7.77 (dt, J = 8.3, 0.8 Hz, 1H), 7.45 (ddd, J = 8.4, 7.2, 1.3 Hz, 1H), 7.43 – 7.39 (m, 1H), 7.34 –
7.28 (m, 1H), 7.10 (dd, J = 2.3, 0.5 Hz, 1H), 6.87 (dd, J = 5.8, 2.3 Hz, 1H), 3.94 (s, 3H).
13
C
NMR (101 MHz, CDCl3) δ 167.71, 152.66, 150.70, 140.38, 139.64, 126.55, 123.98, 123.62,
123.09, 121.21, 121.20, 120.20, 119.77, 114.53, 111.19, 108.26, 105.13, 55.64.
2d: 2.79g (88%) yield.
1
H NMR (400 MHz, CDCl3) δ 8.90 (dt, J = 5.2, 0.7 Hz, 1H), 8.10 –
8.05 (m, 2H), 7.94 (d, J = 8.3 Hz, 1H), 7.85 (dt, J = 1.6, 0.8 Hz, 1H), 7.75 (dt, J = 8.3, 0.9 Hz,
1H), 7.55 – 7.43 (m, 3H), 7.35 (ddd, J = 8.0, 7.3, 1.0 Hz, 1H).
13
C NMR (101 MHz, CDCl3) δ
152.34, 150.90, 141.22, 140.88, 139.94, 139.06, 126.91, 124.86, 124.10, 123.48, 122.03,
121.32, 120.44, 120.09, 116.81 (q, J = 3.3 Hz), 114.65, 114.30 (q, J = 3.7 Hz), 110.83.
Molecular Design for Organic Photovoltaics
Chapter 7 | 237
3a-d. To a 50 mL Schlenk flask containing a 0.25 M solution of 2 in a 4:1 DMSO/H2O mixture
was added lithium hydroxide (2.1 eq.), copper(I) chloride (0.05 eq.), and N
1
,N
2
-bis(4-hydroxy-
2,6-dimethylphenyl)ethanediamide (0.05 eq.). The slurry was sparged with nitrogen for 30
min, then heated to 110 ˚C and stirred overnight. The flask was then cooled to room
temperature, quenched with 100 mL water, and extracted into ethyl acetate (3x100 mL). The
organic layers were combined, dried over sodium sulfate, and condensed to a light brown solid
which was purified by silica gel flash chromatography with eluent 30% ethyl acetate in
hexanes. The product eluted as the second major fraction; the first major fraction contained
compound 4.
3c: 0.50g (76%) yield. Colorless powdery solid.
1
H NMR (400 MHz, CDCl3) δ 8.50 (d, J = 5.8
Hz, 1H), 8.00 – 7.96 (m, 1H), 7.90 (d, J = 8.4 Hz, 1H), 7.72 (dt, J = 8.3, 0.8 Hz, 1H), 7.36 –
7.32 (m, 2H), 7.30 – 7.26 (m, 1H), 7.11 (d, J = 2.3 Hz, 1H), 6.84 (dd, J = 5.8, 2.3 Hz, 1H), 6.82
– 6.77 (m, 1H), 5.83 (s, 1H), 3.93 (s, 3H).
13
C NMR (101 MHz, CDCl3) δ 167.68, 155.12,
153.12, 150.37, 140.92, 139.55, 124.87, 124.51, 120.99, 120.94, 119.34, 118.02, 110.78,
109.90, 108.12, 104.91, 97.85, 55.61.
3d: 0.81g (75%) yield.
1
H NMR (400 MHz, CDCl3) δ 8.84 (dt, J = 5.1, 0.7 Hz, 1H), 7.98 (ddd,
J = 7.6, 1.4, 0.7 Hz, 1H), 7.90 (dd, J = 8.4, 0.5 Hz, 1H), 7.86 (dq, J = 1.6, 0.7 Hz, 1H), 7.78 –
7.73 (m, 1H), 7.46 (ddt, J = 5.1, 1.6, 0.6 Hz, 1H), 7.40 – 7.33 (m, 2H), 7.33 – 7.26 (m, 1H),
6.85 (dd, J = 8.4, 2.2 Hz, 1H), 5.69 (s, 1H).
13
C NMR (101 MHz, CDCl3) δ 13C 155.15, 152.79,
150.65, 140.89 (q, J = 34.3 Hz), 140.38, 139.08, 125.28, 124.91, 123.85, 121.48 (q, J = 3.4
Hz), 119.53, 118.53, 116.37 (q, J = 3.8 Hz), 114.26, 114.22, 110.70, 110.57, 97.95.
4a-d. To a 50 mL Schlenk flask containing a 0.15 M solution of 3 in DMSO was added 2 (1.1
eq.), potassium phosphate (2.0 eq.), copper(I) iodide (0.10 eq.), and picolinic acid (0.20 eq.).
Golden
238
The solution was sparged with nitrogen for 30 min, then heated to 110 ˚C and stirred for three
days. The reaction was then cooled to room temperature, quenched with 100 mL water, and
extracted into ethyl acetate (3x100 mL). The organic layers were combined, dried over sodium
sulfate, and condensed to powdery solid which was purified by silica gel flash chromatography
with eluent 30% ethyl acetate in hexanes. The product was further purified by recrystallization
from a concentrated dichloromethane solution layered with hexanes.
4b: Xg (X%) yield. Eggshell white powder.
1
H NMR (400 MHz, CDCl3) δ 8.50 (dd, J = 5.2,
2.0 Hz, 2H), 8.05 (td, J = 8.3, 7.8, 2.3 Hz, 4H), 7.80 (dd, J = 8.3, 2.4 Hz, 2H), 7.57 (t, J = 2.5
Hz, 2H), 7.45 – 7.36 (m, 4H), 7.31 (td, J = 7.5, 2.1 Hz, 2H), 7.06 (ddt, J = 9.0, 6.3, 3.1 Hz,
4H), 2.39 (d, J = 3.4 Hz, 6H).
13
C NMR (101 MHz, CDCl3) δ 156.78, 151.55, 150.05, 149.23,
140.69, 140.04, 125.46, 124.09, 122.61, 120.96, 120.91, 120.03, 119.67, 119.53, 112.64,
111.10, 101.85, 21.14.
4c: 0.50g (43%) yield. Light brown powder.
1
H NMR (400 MHz, CDCl3) δ 8.44 (d, J = 5.8
Hz, 2H), 8.03 (t, J = 8.3 Hz, 4H), 7.81 (d, J = 8.2 Hz, 2H), 7.53 (d, J = 2.2 Hz, 2H), 7.44 – 7.36
(m, 1H), 7.30 (t, J = 7.5 Hz, 2H), 7.10 – 7.02 (m, 4H), 6.78 (dd, J = 5.8, 2.3 Hz, 2H), 3.81 (s,
6H).
13
C NMR (101 MHz, CDCl3) δ 167.57, 156.89, 152.94, 150.45, 140.58, 139.96, 125.51,
124.08, 121.02, 120.93, 120.04, 119.64, 112.81, 111.23, 108.34, 104.58, 101.95, 55.48.
4d: 0.95g (53%) yield. Yellow powder/pale yellow to colorless crystals.
1
H NMR (400 MHz,
CDCl3) δ 8.79 (d, J = 5.0 Hz, 2H), 8.04 (m, 4H), 7.83 (m, 4H), 7.79 (d, J = 8.2 Hz, 2H), 7.61
(d, J = 2.1 Hz, 2H), 7.46 – 7.40 (m, 4H), 7.34 (dt, J = 7.5, 0.94 Hz, 2H), 7.09 (dd, J = 8.5, 2.1
Hz, 2H).
13
C NMR (101 MHz, CDCl3) δ 156.88, 152.60, 150.73, 140.83 (q, J = 34.3 Hz),
140.21, 139.41, 125.86, 124.58, 123.78, 121.88, 121.12, 120.48, 119.93, 116.43 (q, J = 3.3
Hz), 114.17 (q, J = 3.7 Hz), 113.46, 110.85, 102.09.
Molecular Design for Organic Photovoltaics
Chapter 7 | 239
5a-d. To a 100 mL Schlenk flask containing a 0.02 M solution of 4 in glacial acetic acid was
added tetrabutylammonium bromide (0.10 eq.) and potassium tetrachloroplatinate(II) (1.1 eq.).
The solution was sparged with nitrogen for 30 min, then stirred overnight at room temperature.
After a night, the slurry, containing copious precipitated solids, was heated to 110 ˚C, re-
dissolving the solids. It was then stirred at 110 ˚C for three days. The reaction was then cooled
to room temperature and slowly poured into cold water. The precipitated solids were collected
by filtration and rinsed with water. The filter cake was dissolved into dichloromethane, dried
over sodium sulfate, and condensed by rotary evaporation to form a powdery solid which was
purified by silica gel flash chromatography with eluent 25% ethyl acetate in hexanes. The
product was further purified by recrystallization from a concentrated dichloromethane solution
layered with methanol.
PtNON-Me, 5b: Pale yellow crystals.
1
H NMR (400 MHz, Benzene-d6) δ 8.55 (d, J = 6.1 Hz,
2H), 7.98 (ddd, J = 7.6, 1.4, 0.6 Hz, 2H), 7.87 (d, J = 8.3 Hz, 2H), 7.80 (d, J = 8.3 Hz, 2H),
7.69 (dt, J = 8.0, 0.9 Hz, 2H), 7.31 – 7.25 (m, 4H), 7.24 – 7.18 (m, 2H), 5.96 (ddd, J = 6.2, 1.8,
0.7 Hz, 2H), 1.47 (s, 6H).
13
C NMR (101 MHz, Benzene-d6) δ 153.10, 150.70, 149.25, 148.53,
143.75, 138.13, 129.51, 123.12, 122.52, 119.87, 119.83, 116.06, 115.85, 113.38, 113.20,
94.82, 20.34.
PtNON-OMe, 5c: 0.36g (60%) yield. Light beige powder/pale yellow needle-like crystals.
1
H
NMR (400 MHz, CDCl3) δ 8.82 (d, J = 6.9 Hz, 2H), 8.08 – 8.01 (m, 2H), 8.00 – 7.93 (m, 2H),
7.82 (d, J = 8.3 Hz, 2H), 7.57 (d, J = 2.6 Hz, 2H), 7.43 – 7.34 (m, 4H), 7.34 – 7.30 (m, 2H),
6.63 (dd, J = 6.9, 2.6 Hz, 2H), 3.91 (s, 6H).
13
C NMR (101 MHz, CDCl3) δ 166.84, 152.49,
152.30, 150.37, 143.24, 138.09, 129.22, 123.51, 122.62, 119.86, 115.89, 115.73, 113.55,
112.63, 107.98, 99.99, 93.34, 55.95.
Golden
240
PtNON-CF3, 5d: 0.54g (62%) yield. Bright orange powder.
1
H NMR (400 MHz, Benzene-d6)
δ 8.37 (d, J = 6.3 Hz, 2H), 7.91 (dd, J = 7.7, 0.7 Hz, 2H), 7.87 – 7.77 (m, 6H), 7.61 (d, J = 8.1
Hz, 2H), 7.25 (td, J = 7.5, 0.9 Hz, 2H), 7.14 (t, J = 1.1 Hz, 2H), 6.26 (dd, J = 6.4, 1.8 Hz, 2H).
13
C NMR (101 MHz, Benzene-d6) δ 152.98, 152.58, 148.80, 142.95, 138.36 (q, J = 34.6 Hz),
137.60, 129.76, 124.00, 123.74, 121.08, 120.21, 116.66, 116.23, 113.82, 113.55 (d, J = 3.0
Hz), 112.79, 112.55 (d, J = 4.0 Hz), 93.60.
7.6 Bibliography
1. Baldo, M. A.; O’Brien, D. F.; Thompson, M. E.; Forrest, S. R., Excitonic singlet-triplet
ratio in a semiconducting organic thin film. Phys. Rev. B 1999, 60 (20), 14422-14428.
2. Baldo, M. A.; O'Brien, D. F.; You, Y.; Shoustikov, A.; Sibley, S.; Thompson, M. E.;
Forrest, S. R., Highly efficient phosphorescent emission from organic electroluminescent
devices. Nature 1998, 395, 151.
3. Forrest, S. R., The path to ubiquitous and low-cost organic electronic appliances on
plastic. Nature 2004, 428, 911.
4. Zhang, Y.; Lee, J.; Forrest, S. R., Tenfold increase in the lifetime of blue
phosphorescent organic light-emitting diodes. Nat. Commun. 2014, 5, 5008.
5. Xiao, L.; Chen, Z.; Qu, B.; Luo, J.; Kong, S.; Gong, Q.; Kido, J., Recent Progresses on
Materials for Electrophosphorescent Organic Light‐Emitting Devices. Adv. Mater. 2010, 23
(8), 926-952.
6. Kim, S.; Bae, H. J.; Park, S.; Kim, W.; Kim, J.; Kim, J. S.; Jung, Y.; Sul, S.; Ihn, S.-
G.; Noh, C.; Kim, S.; You, Y., Degradation of blue-phosphorescent organic light-emitting
devices involves exciton-induced generation of polaron pair within emitting layers. Nat.
Commun. 2018, 9 (1), 1211.
7. Lee, J.; Chen, H.-F.; Batagoda, T.; Coburn, C.; Djurovich, P. I.; Thompson, M. E.;
Forrest, S. R., Deep blue phosphorescent organic light-emitting diodes with very high
brightness and efficiency. Nat. Mater. 2015, 15, 92.
8. Uoyama, H.; Goushi, K.; Shizu, K.; Nomura, H.; Adachi, C., Highly efficient organic
light-emitting diodes from delayed fluorescence. Nature 2012, 492, 234.
9. Di, D.; Romanov, A. S.; Yang, L.; Richter, J. M.; Rivett, J. P. H.; Jones, S.; Thomas,
T. H.; Abdi Jalebi, M.; Friend, R. H.; Linnolahti, M.; Bochmann, M.; Credgington, D., High-
performance light-emitting diodes based on carbene-metal-amides. Science 2017, 356 (6334),
159.
10. To, W. P.; Zhou, D.; Tong Glenna So, M.; Cheng, G.; Yang, C.; Che, C. M., Highly
Luminescent Pincer Gold(III) Aryl Emitters: Thermally Activated Delayed Fluorescence and
Solution‐Processed OLEDs. Angew. Chem. Int. Ed. 2017, 56 (45), 14036-14041.
11. Fleetham, T.; Ji, Y.; Huang, L.; Fleetham, T. S.; Li, J., Efficient and stable single-doped
white OLEDs using a palladium-based phosphorescent excimer. Chem. Sci. 2017, 8 (12), 7983-
7990.
Molecular Design for Organic Photovoltaics
Chapter 7 | 241
12. Brooks, J.; Babayan, Y.; Lamansky, S.; Djurovich, P. I.; Tsyba, I.; Bau, R.; Thompson,
M. E., Synthesis and Characterization of Phosphorescent Cyclometalated Platinum
Complexes. Inorg. Chem. 2002, 41 (12), 3055-3066.
13. Li, K.; Ming Tong, G. S.; Wan, Q.; Cheng, G.; Tong, W.-Y.; Ang, W.-H.; Kwong, W.-
L.; Che, C.-M., Highly phosphorescent platinum(ii) emitters: photophysics, materials and
biological applications. Chem. Sci. 2016, 7 (3), 1653-1673.
14. Fleetham, T.; Li, G.; Li, J., Phosphorescent Pt(II) and Pd(II) Complexes for Efficient,
High‐Color‐Quality, and Stable OLEDs. Adv. Mater. 2016, 29 (5), 1601861.
15. Fleetham, T. B.; Huang, L.; Klimes, K.; Brooks, J.; Li, J., Tetradentate Pt(II)
Complexes with 6-Membered Chelate Rings: A New Route for Stable and Efficient Blue
Organic Light Emitting Diodes. Chem. Mater. 2016, 28 (10), 3276-3282.
16. Fleetham, T.; Li, G.; Wen, L.; Li, J., Efficient “Pure” Blue OLEDs Employing
Tetradentate Pt Complexes with a Narrow Spectral Bandwidth. Adv. Mater. 2014, 26 (41),
7116-7121.
17. Li, G.; Fleetham, T.; Turner, E.; Hang, X. C.; Li, J., Highly Efficient and Stable
Narrow‐Band Phosphorescent Emitters for OLED Applications. Adv. Opt. Mater. 2014, 3 (3),
390-397.
18. Li, G.; Wolfe, A.; Brooks, J.; Zhu, Z.-Q.; Li, J., Modifying Emission Spectral
Bandwidth of Phosphorescent Platinum(II) Complexes Through Synthetic Control. Inorg.
Chem. 2017, 56 (14), 8244-8256.
19. Turner, E.; Bakken, N.; Li, J., Cyclometalated Platinum Complexes with Luminescent
Quantum Yields Approaching 100%. Inorg. Chem. 2013, 52 (13), 7344-7351.
20. Hang, X. C.; Fleetham, T.; Turner, E.; Brooks, J.; Li, J., Highly Efficient Blue‐Emitting
Cyclometalated Platinum(II) Complexes by Judicious Molecular Design. Angew. Chem. Int.
Ed. 2013, 52 (26), 6753-6756.
21. W. D’Andrade, B.; Datta, S.; R. Forrest, S.; Djurovich, P.; Polikarpov, E.; Thompson,
M., Relationship between the ionization and oxidation potentials of molecular organic
semiconductors. Org. Electron. 2005, 6, 11-20.
22. Djurovich, P. I.; Mayo, E. I.; Forrest, S. R.; Thompson, M. E., Measurement of the
lowest unoccupied molecular orbital energies of molecular organic semiconductors. Org.
Electron. 2009, 10 (3), 515-520.

Golden
242

Molecular Design for Organic Photovoltaics
Appendix I | 243
Appendix I.
Instrumentation & Experimental Methods

Nuclear Magnetic Resonance
NMR spectra were recorded on a Varian 400 NMR spectrometer and referenced to the residual
proton resonance of chloroform (CDCl3) solvent at 7.26 ppm. Spectra were processed using
MestReNova.
Mass Spectroscopy
Matrix assisted laser desorption/ionization (MALDI) mass spectroscopy data was acquired on a
Bruker Autoflex Speed LRF.
Elemental Analysis
CHN analyses were performed using 2-3 mg samples in tin foil on a Thermo Scientific Flash 2000
CHNS Analyzer equipped with an 18 mm OD Empty Quartz Reactor. Samples were analyzed
against sulfanilamide and 2,5-Bis (5-tert-butyl-benzoxazol-2-yl) thiophene standards.
Electrochemistry
Cyclic Voltammetry and Differential Pulse Voltammetry
Cyclic voltammetry (CV) and differential pulse voltammetry (DPV) measurements were
performed using an EG&G Potentiostat/Galvanostat model 283. Samples were run in 0.1 M tetra-
n-butyl-ammonium hexafluorophosphate solution in acetonitrile and sparged with nitrogen. The
counter, pseudoreference, and working electrodes were platinum, silver, and glassy carbon,
respectively. Scans were performed at 100 mV/s, and oxidation/reduction values were calibrated
to ferrocene/ferrocenium internal references.x
Golden
244
Spcetroelectrochemistry
Spectroelectrochemical samples were prepared at an optical density of 0.6-0.8 in an acetonitrile
solution inside a glass H-cell with tetrabutylammonium hexafluorophosphate as the electrolyte.
One cuvette of the H-cell, containing the working (glassy carbon) and reference (silver) electrodes
was placed in the path of the UV-Vis excitation source, and the other, containing the counter
electrode (platinum) was separated by a medium frit designed to minimize mixing of oxidized and
reduced analyte between the two cells. The UV-Vis was blanked to the neutral solution, and then
chronoamperometry at 100 mV overpotentials for the desired redox process was begun.
Steady-State Photophysics
UV-visible spectra were recorded on a Hewlett‑Packard 4853 diode array spectrometer.
Photoluminescence spectra were measured using a QuantaMaster Photon Technology
International phosphorescence/fluorescence spectrofluorometer. Quantum yield measurements
were carried out using a Hamamatsu C9920 system equipped with a xenon lamp, calibrated
integrating sphere and model C10027 photonic multi-channel analyzer (PMA). Molar extinction
coefficients were obtained by plotting solutions at four concentrations between 0.1 and 0.9 on a
Beer’s Law plot, with y-intercept set to zero. Line fitting for all samples provided R
2
values greater
than 0.98. Excitation, emission, photoluminescence quantum yield, and lifetime measurements
were acquired from solutions at maximum optical densities between 0.1-0.2 to minimize the effects
of solute-solute interactions and inner filter effects. Room temperature photophysical
measurements were recorded in indicated solvents in 1 cm glass or quartz cuvettes, as needed, and
cryogenic photophysical measurements were carried out in an NMR tube in methylcyclohexane or
2-methyltetrahydrofuran, as indicated, at 77 K.
Molecular Design for Organic Photovoltaics
Appendix I | 245
Time-Resolved Photophysics
Time-Correlated Single Photon Counting
Photoluminescence lifetimes were measured by time-correlated single-photon counting using an
IBH Fluorocube instrument equipped with an LED excitation source.
Femtosecond Transient Absorption
Pump and probe pulses were obtained from the output of a Ti:Sapphire regenerative amplifier
(Coherent Legend Elite, 1 kHz, 3.2 mJ, 35 fs). The pump pulses were generated by pumping a type
II OPA (Spectra Physics OPA-800C) with 10% of the 800 nm amplifier output and mixing the
OPA signal with the 800 nm residual on a type II BBO to generate 500 nm and 485 nm. 500 nm
was used to pump the DIPYR dimers and α-DIPYR monomer while 485 nm was used to pump the
DIPYR monomer. A white light super continuum (320-950 nm) was generated by focusing a small
amount of the 800 nm on a rotating CaF2 disk. The supercontinuum probe was collimated and
focused by a pair of off-axis parabolic mirrors onto the sample, while the pump was focused before
the sample with a 25 cm CaF2 lens. The probe was set to magic angle (54.7 ̊) with respect to the
pump to avoid any contribution from reorientational motion. The supercontinuum probe was
dispersed using a spectrograph (Oriel MS127i) onto a 256-pixel silicon diode array (Hamamatsu)
for multiplexed detection of the probe. Samples containing bis-DIPYR and bis-α-DIPYR dissolved
in cyclohexane and THF were placed in a closed capped 1mm quartz cuvette. The DIPYR
monomer and α-DIPYR monomer were dissolved in methylcyclohexane and acetonitrile and also
placed in a 1mm capped quartz cuvette. The samples were made such that the optical densities
were between 0.2 and 0.4 at the pump wavelength. During data collection the samples were moved
on a motorized stage perpendicular to the pump to decrease photodamage caused by the pump.

Golden
246
Nanosecond Transient Absorption
Samples were prepared on a Schlenk line where the samples were purged with N 2 and dissolved
in dry solvents such that the maximum absorbance did not exceed OD=0.5. The samples were
sealed in a capped 1 mm quartz cuvette which was oscillated on a stage. Nanosecond pump
generation was performed using a 532 nm output from a Nd:YAG laser (Alphalas, 532 nm, 1 kHz,
700 ps) which is externally triggered and synchronized with the femtosecond amplifier. The pump
pulse is delayed with respect to the femtosecond supercontinuum probe using a delay generator
DG 645 (Stanford Research Systems).
Computational Methods
All calculations, unless otherwise noted, were performed using the Jaguar software package on the
Schrodinger Material Science Suite. Gas phase geometry optimizations were calculated using
CAM-B3LYP functional with the 6-31G** or LACVP** basis set as noted and implemented in
Jaguar. Also employed was extended multiconfigurational quasi-degenerate second order
perturbation theory (XMCQDPT2). Single point XMCQDPT2 calculations were performed on the
B3LYP/6-31G** optimized structure using the 6-31G(d) basis set. The XMCQDPT2 calculations
were performed using the Firefly quantum chemistry package.

Molecular Design for Organic Photovoltaics
Appendix I | 247

Golden
248
Appendix II.

Nuclear Magnetic Resonance Spectra

Pertaining to Chapter 2 – DIPYR ligands, boron difluoro DIPYRs, and carDIPYRs
*

2,2’-dipyridylmethane

*
All spectra are recorded in CDCl 3 unless otherwise noted.
Molecular Design for Organic Photovoltaics
Appendix II | 249
diquinolylketone

Golden
250
DIPYR

Molecular Design for Organic Photovoltaics
Appendix II | 251

Golden
252
-DIPYR

Molecular Design for Organic Photovoltaics
Appendix II | 253
-DIPYR

Golden
254
Pertaining to Chapter 3
boron difluroro methylpyrazine pyridine (DMSO-d6)

Molecular Design for Organic Photovoltaics
Appendix II | 255

Golden
256
Pertaining to Chapter 4 - DIPYR dimer ligands, DIPYR dimers

1,2-tetra-2,2’,2’’,2’’’-pyridylethane (benzene-d6)

Molecular Design for Organic Photovoltaics
Appendix II | 257

Golden
258
bis-DIPYR

Molecular Design for Organic Photovoltaics
Appendix II | 259

Golden
260
Pertaining to Chapter 5 –
bis-2,4-dimethylBODIPY (B-2)

Molecular Design for Organic Photovoltaics
Appendix II | 261

Golden
262
B-rad

Molecular Design for Organic Photovoltaics
Appendix II | 263

Golden
264

meso-mesityldipyrrin

Molecular Design for Organic Photovoltaics
Appendix II | 265

Golden
266
Zn(mesitylDIPY)2 (Z-0)

Molecular Design for Organic Photovoltaics
Appendix II | 267
2,4-dimethylBODIPY / 2,4-dimethyldipyrrin / Zn(2,4-dimethylDIPY)2 (Z-2)

Golden
268
Zn(3-ethyl-2,4-dimethylDIPY)2 (Z-3)

Molecular Design for Organic Photovoltaics
Appendix II | 269
2-methylbenzoBODIPY / 2-methylbenzodipyrrin / Zn(2-methylbenzoDIPY)2 (Z-4)

Golden
270
Pertaining to Chapter 6
dipyridinium-BODIPY

Molecular Design for Organic Photovoltaics
Appendix II | 271

Golden
272

Molecular Design for Organic Photovoltaics
Appendix II | 273
Pt(BDP-TPBP)

Golden
274
Pertaining to Chapter 7
PtNON-CF3

Molecular Design for Organic Photovoltaics
Appendix II | 275

Golden
276
PtNON-Me

Molecular Design for Organic Photovoltaics
Appendix II | 277

Abstract (if available)

Abstract This dissertation focuses on the development of molecular semiconducting materials designed to meet the specific challenges in organic photovoltaics. Materials design considerations to accomplish intense and broad-band photon absorption throughout the solar spectrum, efficient exciton migration, and efficient charge transfer, separation, and collection are developed throughout the following chapters. Small molecule heteroaromatic chromophores such as dipyridylmethenes (DIPYRs) (Chapters 2-4), dipyrromethenes (Chapters 5-6), and porphyrins (Chapter 6) are the primary focus of this work. The role of molecular and electronic structure is discussed in the context of tuning state energies to modify absorption, internal conversion, intersystem crossing, and intramolecular charge transfer processes. In Chapter 7, these same principles are applied to solving specific challenges in organic light-emitting diodes.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Molecular design of pyridine and quinoline based compounds for organic optoelectronics

PDF

Energy management in organic photovoltaics

PDF

Organic photovoltaics: from materials development to device application

PDF

Properties and applications of dipyrrin-based molecules in organic photovoltaics

PDF

Development of N-type chromophores for organic photovoltaics, and thermally activated delayed fluorescence NHC complexes for organic light-emitting diodes

PDF

Pyrrolic squaraines and energy management in organic photovoltaics

PDF

Novel pyrimidine-based hole blocking materials for long-lived and highly efficient organic light emitting diodes

PDF

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

PDF

First steps of solar energy conversion; primary charge generation processes in condensed phase coupled chromophores

PDF

Molecular design strategies for blue organic light emitting diodes

PDF

Studies of molecular orientation using iridium phosphors and integration of corannulene into organic light emitting diodes (OLEDs)

PDF

Molecular approaches to solve the blue problem in organic light emitting diode display and lighting applications

PDF

Molecular and polymeric donors for bulk heterojunction organic solar cells

PDF

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

PDF

Organic solar cells: molecular electronic processes and device development

PDF

Computational materials design for organic optoelectronics

PDF

Utilizing n-heterocyclic chromophores for solar energy harvesting.

PDF

Simple complexes: synthesis and photophysical studies of luminescent, monovalent, 2-coordinate carbene-coinage metal complexes and higher coordination geometries

PDF

Ir(III) and Pt(II) phosphorescent emitters in organic light emitting diodes: from materials development to light out-coupling

PDF

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

Asset Metadata

Creator Golden, Jessica H, (author)

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

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Chemistry

Publication Date 11/05/2018

Defense Date 06/26/2018

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

Tag BODIPY,boron,chemistry,DIPYR,dyes,OAI-PMH Harvest,OLED,OPV,organic chemistry,organometallics,photovoltiacs,platinum,Solar energy

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

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

Creator Email jhgolden@usc.edu,jhgolden16@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-100833

Unique identifier UC11675761

Identifier etd-GoldenJess-6928.pdf (filename),usctheses-c89-100833 (legacy record id)

Legacy Identifier etd-GoldenJess-6928.pdf

Dmrecord 100833

Document Type Dissertation

Format application/pdf (imt)

Rights Golden, Jessica H,

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