Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Exciton dynamics in photovoltaic materials
(USC Thesis Other)
Exciton dynamics in photovoltaic materials
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
Exciton Dynamics in Photovoltaic Materials
Jimmy Joy
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)
May 2020
ii
Dedicated to family, especially my chosen one
iii
TABLE OF CONTENTS
Acknowledgments
viii
List of Figures
x
List of Tables
xvi
Chapter 1 An Introduction to Molecules, Processes, Techniques, and
Instruments
1
Chapter 2 Charge Transfer Dynamics in CdSe:P3HT Bulk
Heterojunctions
16
2.1 Introduction
16
2.2 Experimental Section
25
2.2.1 General Considerations and Synthetic Procedure 18
2.2.2 Ligand Exchanges
19
2.2.3 Device Fabrication
19
2.2.4 Characterization
20
2.3 Results and Discussion
23
2.3.1 Colloidal Ligand Exchange
23
2.3.2 Effect of Ligand on Device Performance
24
2.3.3 Effect of ligand type on J SC
27
2.3.4 Effect of ligand type on V OC
34
2.4 Conclusions
38
Chapter 3 Polaron Dynamics in PbS:Si-PCPDTBT Bulk Heterojunctions
40
iv
3.1 Introduction
40
3.2 Experimental Section
43
3.2.1 Materials, synthesis and ligand exchange
43
3.2.2 Characterization
44
3.2.3 Photoluminescence lifetime studies
44
3.2.4 Transient absorption studies
46
3.2.5 Chemical doping studies
46
3.2.6 Hybrid solar cell fabrication
47
3.3 Results and Discussion
48
3.3.1 Colloidal Iodide Ligand Exchange
48
3.3.2 Hybrid Solar Cells
50
3.3.3 Charge Separation Dynamics at the Hybrid Interface
52
3.4 Conclusions
60
Postscript 3.A Reevaluation of Transient Absorption Spectra of PbS:Si-
PCPDTBT Hybrid Bulk Heterojunction Films
62
3.S.1. Introduction
62
3.A.2. Results and discussion
64
3.A.2. Conclusions
64
Chapter 4 Linker Dependent Singlet Fission in Tetracene Dimers
76
4.1 Introduction
76
4.2 Experimental Section
83
v
4.2.1 Synthesis of meta- and para- dimers
82
4.2.2 Preparation of samples for spectroscopy
82
4.2.3 Femtosecond Broadband Transient Absorption
84
4.2.4 Time-correlated Single Photon Counting
85
4.2.5. Compartmental models and kinetic equations
85
4.3 Results and Discussion
85
4.3.1 Electrochemistry
86
4.3.2 Steady State Spectra
88
4.3.3 Transient Absorption of m-BETB in THF
91
4.3.4 Transient Absorption of p-BETB-ehex and p-BETB-
ohex in THF
95
4.3.5 Transient Absorption of p-BETB-ehex in PMMA
104
4.4: Conclusions
109
Chapter 5 Decoupling Intra- and Inter-Dimer Singlet Fission
112
5.1 Introduction
112
5.2 Experimental Section
115
5.2.1 Synthesis of ethynyltetracene dimers and sample
preparation
115
5.2.2 Femtosecond Transient Absorption
115
5.2.3 Time-Correlated Single Photon Counting
116
5.3 Results and Discussion
116
vi
5.3.1 Steady – state absorption spectroscopy of meta-
and para-BETB
116
5.3.2 Time-resolved photoluminescence of meta and
paraBETB
118
5.3.3 Transient absorption of meta- and para-BETB neat
films
119
5.4 Conclusions
124
Chapter 6 Fluorescence Upconversion Spectroscopy
126
6.1 Introduction
126
6.2 Materials and Methods
134
6.2.1 Ti:sapphire Oscillator and Regenerative Amplifier
135
6.2.2 Translation Stage
137
6.2.3 Sample Holder
137
6.2.4 Nonlinear crystal
137
6.2.5 Monochromator
138
6.2.6. Photomultiplier Tube
138
6.2.7 Photon Counter
139
6.2.8 Setting up the nonlinear crystal to achieve
frequency mixing
140
6.2.9 Labview Interface and Operation Schema
143
6.2.10 Aligning the upconverted beam on the detection
apparatus
143
6.2.11 Measuring the upconversion spectra
144
vii
6.3 Results and Discussion
150
6.3.1 Frequency Resolved Optical Gating with 800 nm
150
6.3.2 Measuring the instrument response
151
6.4 Conclusions
152
Bibliography
154
viii
Acknowledgments
To my Mom, Dad, Sis and family – thank you all for letting me choose my own path
and follow it. I haven’t seen you all in over 6 years, but I think of you everyday and
your love and support gave me strength to see graduate school through.
To my teachers from school – Priya, Mini, Reny, Cissy, Sheela, Daphne, and others –
I wouldn’t be here if I did not have such inspiring role models growing up. Thank you
for being such caring educators, for engaging my creativity and curiosity. I remember
the day I fell in love with chemistry. We had a pop quiz with Priya and the whole class
was puzzled on this one question - on why phosphoric acid is syrupy in consistency.
I remember seeing molecules in my head forming intermolecular hydrogen bonds
and coming to the right answer. That singular moment inspired so much of my life so
far – thank you for that.
To my professors from the Indian Institute of Science Education and Research -
Mahesh. Sureshan, Swathi, Hema, Anil, and others - thank for all the great classes,
conversations and helping me find my path in the world.
Friends from IISER Ajai, Rijo, Shinaj, Subhila, Reshmi, Anoop, Sarang, Soma,
Sreeganesh, Chitra, Salini, Preethi, Ajeesh, Vrindha – thank you for the
companionship through some really tough times, I couldn’t have made it through
without your love and support.
Donn – thank you for all the good times, and for bringing me here to Los Angeles.
Jose Ricardo and Drystan - the last six years on campus would have been so bleak
without our afternoon tea and conversations. Love you both, so glad we were all here
at the same time.
Mathew, David, Xichen, Kevin, Olivia and Tim – thank you for keeping me grounded in
the real world while my mind wandered with molecules and photons. I’m lucky to
have found such good people here in this city.
Dave, thank you for being such a great advisor and guide out of the campus setting.
I’m grateful to have you as a friend.
To the Fellowship, to Kevin, Anton, David, Ruaraidh, Gavin, Chris, Ray, Brian, Adam,
Robert, Dixon, Ray, and Greg - I would not have been able to finish graduate school if
I did had not found you all in the rooms. The last year has been one of toughest,
thank you for helping me through it. I’m grateful to have found family here with you
all.
ix
To the Group, Saptaparna, Anirban, Robert, Gaurav, Dhritiman, Laura, Michael, Ryan,
Matt and Tillman – thank you for a fun and exciting six years, for the late night
experiments and early morning clean ups, for being patient with me, and for all the
help and guidance.
Steve, I am so blessed to have you as my advisor. Besides being an excellent
educator, research advisor and extraordinary experimentalist, you have also been a
loving friend, a patient guide and an excellent role model. Thank you for being so
kind and patient with me. I’m immensely proud to call myself one of your students.
To Richard, Mark, Haipeng, Nadia, Matt - thank you for all the exciting
collaborations. It was an honor to work with such clever minds and you are all an
inspiration to me.
To USC, and to the agencies that funded the research – the Department of Energy
and the Office of Naval Research – thank you for making this dissertation possible.
x
List of Figures
Figure 1.1. Annual net additions to world solar capacity in gigawatts.
Historical data is shown in red, the estimate from the IEA for 2019 is shown
in black while those for years past are shown in shades of blue.
Reproduced from reference 3.
2
Figure 1.2. Chart (reproduced from the NREL website) of best research
cell efficiencies published by the National Renewable Energy Laboratory,
Golden CO – official verifier of efficiency records for research cells.
4
Figure 1.3. Photoactive polymers, molecules, nanoparticles discussed in
this thesis.
6
Figure 1.4. Excited state processes discussed in this thesis. 9
Figure 1.5. The internal structure of a photovoltaic cell (left) and schematic
of charge migration in the bulk heterojunction active layer in the cell
(right).
10
Figure 1.6. Layout of the femtosecond transient absorption setup (top) and
schematic showing how the measurement is made (bottom).
12
Figure 1.7. Layout of the Time-Correlated Single Photon Counting Setup
used to measure photoluminescence lifetimes.
14
Figure 2.1. TEM micrograph of the as-prepared nanocrystals with the
corresponding size histogram (left panel). The histogram displays the size
distribution of 100 CdSe quantum dots found to have an average size
diameter of 5.44 ± 0.28 nm, which agrees with the empirical sizing analysis
by UV-vis (right panel).
30
Figure 2.2. Absorption spectra of ligand exchanged neat quantum dots (a),
that of bulk heterojunctions made of ligand exchanged quantum dots and
P3HT polymer (b), emission spectra of ligand exchanged neat quantum
dots c) and that of corresponding bulk heterojunctions (d).
32
Figure 2.3. Photoluminescence decay profiles of neat ligand exchanged
quantum dots. Grey dots is the data obtained from TCSPC and colored
lines are multiexponential fits.
36
Figure 2.4. Photoluminescence decays from P3HT films annealed at
different temperatures.
37
Figure 2.5. Photoluminescence decays from bulk heterojunctions made
from CdSe quantum dots and P3HT polymer.
39
Figure 2.6. (a) Open circuit potentials vs ΔEDA for all the optimized
P3HT:CdSe quantum dot BHJ devices with a general device structure
ITO/PEDOT:PSS (30 nm)/P3HT:CdSe BHJ (70-100 nm)/ZnO (40 nm)/Al (100
nm) (slope = 0.6 V eV-1 and y-intercept = -0.14 V). (b) Dark I-V
characteristics for each optimized P3HT:CdSe quantum dot BHJ device.
(c) Device short circuit currents vs average PL lifetime of the hybrid films
measure using time correlated single photon counting.
40
xi
Figure 2.7. CV traces and corresponding energy level diagram for CdSe
quantum dots ligand exchanged with PA, Py, BA, tBT, TP, and THT ligands
relative to the as-prepared CdSe quantum dots with native ligands (NL).
Data were collected with a scan rate of 20 mV s
–1
in 0.1 M TBAP with a Pt
wire counter electrode and Ag wire pseudo-reference electrode calibrated
against the Fc/Fc
+
redox couple. All potentials are shown relative to NHE.
44
Figure 3.1. TEM image of PbS nanocrystals (left) schematic of the ligand
exchnage process (center), and the polymer used in the bulk
heterjunctions Si-PCPDTBT (right).
59
Figure 3.2. Steady state photophysical properties of PbS quantum dots
and Si-PCPDTBT polymer.
60
Figure 3.3. Schematic illustration of hybrid solar cells based on
ITO/PEDOT:PSS/Si-PCPDTBT:PbS nanocrystal/ZnO/Al device structures
(left panel) and state diagram representing the excitation and charge
transfer processes in the bulk heterojunctions.
62
Figure 3.4. Time-resolved PL traces of the pristine Si-PCPDTBT polymer
and Si-PCPDTBT:nanocrystal BHJs employing NH 4I- and PbI 2-exchanged
PbS acceptors ( λ ex = 500 nm; λ em = 700 nm). The instrument response
function (IRF) is shown in black.
64
Figure 3.5 Transient absorption spectra of (a) PbS(NH 4I):Si-PCPDTBT and
(b) PbS(PbI 2):Si-PCPDTBT film excited at 920 nm and probed in the 900-
1200 nm range.
66
Figure 3.6. Femtosecond TA spectra of three films as labeled. Hybrid BHJ
films were pumped at 920 nm and probed at 1200 nm; the neat Si-PCPDTBT
film was pumped at 660 nm and probed at 1200 nm.
67
Figure 3.7 Transient absorption spectra (at 4 ps, pumped at 920 nm)
overlapping with the absorption and emission spectra of the PbS
nanocrystals. (a) Hybrid Si-PCPDTBT:nanocrystal BHJ film with PbI2-
exchanged PbS, and (b) hybrid Si-PCPDTBT:nanocrystal BHJ film with
NH4I-exchanged PbS.
69
Figure 3.8 (a) Steady-state absorption spectra of the Si-PCPDTBT thin film
before doping (black) and after a 3 min dip in a 20 ppm solution of SbCl 5 in
acetonitrile (red), shown together with the difference spectrum ( A =
Adoped − Aundoped, given in blue). (b) Transient absorption spectra of
neat Si-PCPDTBT film (at 0.1 ps, pumped at 660 nm) overlapping with
chemically doped steady-state spectrum of the oxidized Si-PCPDTBT film.
Steady-state absorption spectra were taken in an integrating sphere and
thus account for scattering and reflections.
71
Figure 3.A.1 Transient absorption spectra of (a) neat Si-PCPDTBT, (b)
PbS(NH 4I):Si-PCPDTBT, and (c) PbS(PbI 2):Si-PCPDTBT films excited at 660
nm.
89
xii
Figure 3.A.2 Comparison of decay dynamics of the excited species in neat
Si-PCPDTBT, PbS(NH 4I):Si-PCPDTBT, and PbS(PbI 2):Si-PCPDTBT films
excited at 660 nm. Time traces were obtained by averaging over the
following wavelength ranges: Polaron 2 (965 nm -1015 nm), Polaron 1 (1150
nm -1200 nm), CT state (1050 nm – 1110 nm) and Bleach (395 nm – 455
nm).
90
Figure 3.A.3. Schematic illustration of hybrid solar cells based on
ITO/PEDOT:PSS/Si-PCPDTBT:PbS nanocrystal/ZnO/Al device structures
(left panel) and state diagram representing the excitation and charge
transfer processes in the bulk heterojunctions.
91
Figure 3.A.4. The transient absorption spectra of the polymer and hybrid
films at 1 ps are show with filled circles while the open circle traces are
the transients of hybrid films with the polymer signal subtracted out. The
green trace is the absorbance of the PbS(PbI 2):Si-PCPDTBT hybrid (x5).
93
Figure 3.A.5. The decay traces of the hybrid films at 1075 nm with the
contribution from the polymer subtracted away.
93
Figure 4.1. Chemical (top) and space-filling structures of o-BETB, m-BETB,
and p-BETB. Two para-BETB dimers with different substituents on the
linker were synthesized (p-BETB-ohex: R = OC 6H 13 and p-BETB-ehex: R =
2-ethyl-hexyl).
99
Figure 4.2. Extinction spectra of m-BETB, p-BETB-ethex, and
p-BETB-ohex compared to o-BETB and PET in a THF solution.
109
Figure 4.3. a) Steady state absorption and emission of m-BETB in THF. b)
Emission decay of m-BETB in THF (excitation at 405 nm, emission at
570 nm), with noticeable delayed fluorescence on the ~200 ns timescale.
111
Figure 4.4. Femtosecond (a) and nanosecond (b) transient absorption of
m-BETB, with the sensitized T n T 1 absorption shown in black. Inset to b)
shows an overlay of the transient spectrum of m-BETB at 135 ns and the
sensitized T n T 1 spectrum.
113
Figure 4.5. The femtosecond transient absorption (a), the species
associated difference spectra (b) and concentration curves (c) from the
target analysis of the TA data of p-BETB-ohex in THF.
115
Figure 4.6. The femtosecond transient absorption (a), the species
associated difference spectra (b) and concentration curves (c) from the
target analysis of the TA data of p-BETB-ehex in THF.
117
Figure 4.7. The plot shows the spectral signature of state 3 from target
analysis of p-BETB-ehex in THF compared to the normalized sensitized T n
T 1 absorption of ET-TMS (red-shifted, 20 nm) and sensitized T n T 1
absorption of p-BETB-ehex in THF (blue-shifted, 9 nm).
121
xiii
Figure 4.8. The proposed decoupling of the correlated triplet state in
p-BETB-ehex (alkyl chains have been removed for clarity) to give rise to
two independent triplets.
123
Figure 4.9. The femtosecond transient absorption of p-BETB-ehex in
PMMA.
125
Figure 4.10. The state diagrams showing (a) population 1 and (b) population
2 used to model the TA data from p-BETB-ehex in PMMA.
127
Figure 4.11. Dynamics of excited state absorption at 425 nm of
p-BETB-ehex in PMMA (black trace) overlaid with the time slices at 425 nm
of transient absorption spectra obtained by combining modelled transient
absorption of population 1 and population 2 in different fractions.
Numbers in the legend show the fraction of population 1 in that model. The
rate constants that gave the best fit are: k r = 5.5 × 10
7
, k nr = 2.9 × 10
6
, k 12 =
5 × 10
11
, k 21 = 4 × 10
12
, k 22 = 2 × 10
9
, k 12’ = 1.6 × 10
12
, k 23 = 1.2 × 10
10
, k 33 =
1.2 × 10
9
.
128
Figure 4.12. A diagram showing the closer lying orbitals of the linker and
the tetracenes in a) p-BETB-ehex compared to b) p-BETB-ohex. l a/b and h a/b
are the frontier molecular orbitals of tetracene.
129
Figure 5.1. Structures of o-BETB (a), m-BETB (b) and p-BETB-ehex (c)
dimers of ethynyltetracene
144
Figure 5.2. Steady state absorption spectra of m-BETB and p-BETB-ehex
in THF solution and neat films. The difference in the first excitonic peak
between solution and film for m-BETB is 0.086 eV and 0.146 eV for p-BETB-
ehex.
147
Figure 5.3. Photoluminescence decay of m-BETB in THF solution and neat
films as measured by Time-Correlated Single Photon Counting (TCSPC).
The decay can be fit with a single exponential with time constant 14.7 ns
for m-BETB in THF solution. Inset shows the decay of photoluminescence
from a m-BETB neat film is limited by the instrument response (22 ps).
148
Figure 5.4. Transient absorption of para-BETB-ehex neat film (a), excited
at 564 nm. Legend shows the pump probe delays at which each spectrum
was taken and (b) time slices of the transient absorption data at probe
wavelengths 425 nm, 565 nm and 660 nm overlaid with tri-exponential fits.
149
Figure 5.5. Sequential 2-state model used in the target analysis of the
transient absorption of para-BETB-ehex neat film and the time constants
(a), corresponding species associated spectra (SAS) (b), and the
concentration profiles (c).
150
Figure 5.6. Sequential 3-state model used in the target analysis of the
transient absorption of para-BETB-ehex in THF solution (a), the
151
xiv
corresponding species associated spectra (SAS) (b), and concentration
curves (c).
Figure 5.7. 2-state model used in the target analysis of the transient
absorption of m-BETB in THF solution (a), the corresponding species
associated spectra (SAS) (b), and concentration curves (c).
152
Figure 5.8. Sequential 2-state model used in the target analysis of the
transient absorption meta-BETB neat film (a), the corresponding species
associated spectra (SAS) (b), and concentration curves (c).
153
Figure 6.1. Schematic of a typical fluorescence upconversion setup. 159
Figure 6.2. The spectral response curves of the R3809U-50 series of
photomultiplier tubes from Hamamatsu (reproduced from the product
datasheet). The specific PMT used here is boxed in red in the figure.
165
Figure 6.3. Schematic of the experimental setup used to measure
frequency resolved autocorrelation of the gate beam. L1 and L2 are plano-
convex lenses (focal length 10 cm and 5 cm) used to focus the 800 nm at
the nonlinear crystal. L3 represents an 800 nm dielectric mirror used to
remove the 800 nm beam letting only the 400 nm pass and a 5 cm plano-
convex lens that focusses the 400 nm beam at the entrance slit of the
monochromator.
166
Figure 6.4. Schematic of the experimental setup used to measure
fluorescence upconversion. L1 – convex lens, 2.5 cm focal length, NLO1 –
1 mm thick type-1 BBO crystal cut at 32.5º, L2 – plano-convex lens focal
length 5 cm, L3 – convex lens focal length 5 cm, L4 – plano-convex lens
focal length 5 cm, L5 – system of collection optics formed by two plano-
convex lenses of focal lengths 2.5 cm and 10 cm, NLO2 – 0.25 mm type 1
BBO crystal cut at 28.9º, L6 – combination of 400 and 800 nm dielectric
mirrors to block the gate and excitation beams and only let the 266 nm
pass and a 5 cm plano-convex lens to focus the 266 nm at the entrance slit
of the monochromator.
168
Figure 6.5. The variation in number of photons detected by the PMT as a
function of the high voltage supplied to it. Note the plateauing around 1.0
kV.
171
Figure 6.6. The real time (RT) tab in the Labview interface that can be
used to align the upconverted signal to the entrance slit of the
monochromator.
179
Figure 6.7. Settings for the translation stage in the DELAY tab. 179
Figure 6.8. Settings for the translation stage in the COUNTER tab. 180
Figure 6.9. Settings for the translation stage in the MONO tab. 180
Figure 6.10. Settings for the translation stage in the PLOT tab. 181
Figure 6.11. Spectra of the 400 nm generated by mixing the 800 nm beam
with itself, measured as function of delay between the two pulses.
182
xv
Figure 6.12. Instrument response profile of the upconversion setup
characterized by measuring the spectra of the 266 nm generated by mixing
the 800 nm gate beam with the 400 nm excitation beam, measured as
function of delay between the two pulses.
183
Figure 6.13. Time slices from time-resolved spectra obtained by (a)
mixing the 800 nm with itself (taken at 393 nm) and (b) mixing the 800 nm
with 400 nm (taken at 263 nm). Pulse width in the former is 130 fs and in
the latter is 80 fs.
183
xvi
List of Tables and Schemes
Table 2.1. Photovoltaic Device Parameters for Optimized Hybrid
P3HT:CdSe BHJ Solar Cells with Different Ligands
28
Table 2.2. Lifetimes and amplitudes obtained by fitting
photoluminescence decays with exponentials and the calculated average
lifetime.
35
Table 2.3. Electrochemically measured Ered and ELUMO levels for the
CdSe quantum dots as a function of ligand type, and the corresponding
ΔEDA and optimized P3HT:CdSe device VOC.
42
Table 3.A.1. The amplitudes and time constants obtained from fitting the
transient spectra with two exponentials.
92
Scheme 4.1. Synthetic scheme for meta, para-BETB-ohex and para-
BETB-ehex.
102
Table 4.1. Electrochemical properties of the four dimers compared to the
monomer, and the optical gaps of these molecules. The
ferrocenium/ferrocene redox couple in dichloromethane was used as an
internal standard for electrochemical measurements.
108
Table 4.2. The time constants (in ps) of the decays used to fit the TA data
of the dimers in THF.
119
Table 4.3. The RAS-2SF-CI computed excited state energies (in eV) for p-
BETB-ehex as a function of the dihedral angle between the tetracenes.
124
1
CHAPTER 1
An Introduction to Molecules, Processes,
Techniques, and Instruments
The World Energy Outlook recently published by the International Energy Agency (IEA)
makes the dire prediction that global CO 2 levels will continue rising despite profound
shifts in the functioning of the global energy system.
1
The rise in these levels is
attributed to rising demand for oil and gas, despite the demand for coal plateauing
and the surge in wind and solar power meeting the majority of increases in global
energy demand. The IEA calls for a significant reallocation of investment away from
fossil fuels towards renewables, and the decommissioning half of the world’s coal-
fired power stations – in order to meet the limiting warming set out by the Paris
Agreement of 1.65C. Among such renewables, solar photovoltaics are becoming the
most competitive source of electricity in 2020 in China and India while in the European
Union and the United States it is predicted to largely close the gap with other sources
like coal and natural gas by 2030. Interestingly, the IEA has also consistently come
under fire for failing to predict the growth of renewables.
2
As seen in Figure 1.1, the
actual net additions to solar capacity has outperformed IEA outlooks since 2006 –
which the IEA attributes to shifts in government policy over time.
3
The agency has
called for a renewed acceleration in annual solar PV deployment, along with enhanced
2
efforts to ensure smooth integration of the resultant solar generation into existing
power systems to reach climate targets and other sustainable goals.
The power received by the earth from the sun – 120,000 trillion Watts is just one
billionth of the sun’s total power output – and that is more energy in an hour than all
the energy consumed on the earth in an entire year.
4
Scientists have spent
decades in designing and optimizing photovoltaic devices that can capture this
energy in form of sunlight and convert it to electricity. Although 90% of these devices
so far have been made of silicon, research in academia and the industry are focused
more on other emerging photovoltaic materials including dye-sensitized solar cells,
Figure 1.1. Annual net additions to world solar capacity in gigawatts. Historical
data is shown in red, the estimate from the IEA for 2019 is shown in black while
those for years past are shown in shades of blue. Reproduced from reference 3.
3
organic small molecule and polymer photovoltaics, perovskites, inorganic quantum
dots and hybrid photovoltaics made of combinations of these.
5-15
These emerging
technologies promise to be less expensive, thinner, flexible, and in some cases more
efficient at harvesting solar energy. Silicon photovoltaics also suffer from drawbacks
which these emerging technologies do not. Si-based solar cells require high purity
silicon on
the order of 99.999%, the cells are made via energy intensive crystal growth methods
and silicon cells use 1000 times more material to absorb the same amount of light as
dye-sensitized cells. Silicon solar cells are often made with thick layers of the
crystalline material because it does not absorb sunlight strongly and such thick layers
tend to be brittle and need secondary support architecture. Emerging photovoltaic
Figure 1.2. Chart (reproduced from the NREL website) of best research cell
efficiencies published by the National Renewable Energy Laboratory, Golden CO
– official verifier of efficiency records for research cells.
4
technologies on the other hand have a lower price tag but lower performance. Dye-
sensitized cells, organic photovoltaics, and quantum dot cells started off with few
percent efficiencies but have climbed slowly over the years as seen in Figure 1.2. The
Figure 1.3. Photoactive polymers, molecules, nanoparticles discussed in this
thesis.
5
research presented in this thesis focusses on two types of emerging photovoltaic
materials: solar cells made of organic molecules and cells that are made of a hybrid
mixture of organic polymers and inorganic nanoparticles.
16-21
Organic photovoltaics currently demonstrate light conversion efficiencies that
average at 17%. Cells of this type use photoactive polymers, small molecules,
fullerene derivatives or mixtures thereof to absorb light and generate electricity
(Figure 1.3). At the nanoscale these compounds maybe coarsely mixed in a network
often called a bulk heterojunction (BHJ). A BHJ provides a large area of contact
between the electron donor and electron acceptor thus mediating efficient charge
separation. Organic photovoltaics have the advantage of being thin, lightweight, and
flexible – unlike cells made of crystalline silicon. Moreover, organic thin films have
high absorption coefficients, their bandgaps can be predictably engineered using
organic synthesis, and have high charge carrier mobilities.
22
Hybrid photovoltaics
made of organic polymers and inorganic nanoparticles currently have similar
efficiencies as their all organic counterparts but have distinct chemical properties
that make them especially suited for light harvesting. In these cells, nanocrystals of
semiconducting metal chalcogenides like CdS, CdSe, PbS, and PbSe act as the
electron acceptor while organic polymers like P3HT, PCPDTBT, and Si-PCPDTBT act
as the electron donor. Unlike in all organic BHJs, the exciton generated in the active
layer of BHJ having an inorganic component has less Columbic force of attraction
holding the electron and hole together because inorganic materials have a higher
6
dielectric constant than organic molecules.
23
Chapters 2 and 3 of this thesis explore
charge carrier dynamics in BHJs made of organic polymers and inorganic
nanoparticles while Chapters 4 and 5 contain research on the excited state processes
in photoactive small molecule organics.
Figure 1.4 is a simplified energy level diagram that shows some of the excited state
pathways possible when a photoactive entity is excited by light. The right side of
Figure 1.5 depicts the generation of free charge carriers in a bulk heterojunction
containing a donor-acceptor blend. Depending on the alignment of their energy levels,
when either the donor or the acceptor is excited by light, an electron or hole maybe
injected into one layer from the other. This initial electron transfer step may generate
an interfacial charge-transfer (CT) state.
24-33
This CT state will in general form with
excess thermal energy to subsequently thermally relax. Due to relatively weak
electronic coupling, the CT state can undergo geminate recombination (GR) to form
the ground state or the CT state can undergo full charge separation (CS) to form
dissociated charge carriers (Figure 1.5). These separated charges can diffuse to their
respective electrodes to generate electricity – when such a donor acceptor BHJ is
used as the active layer in a solar cell. The separated charges can also run into charge
carriers of the opposite nature during this diffusion process resulting in bimolecular
recombination (BR) back to the ground state. Chapters 2 and 3 explore charge carrier
dynamics in bulk heterojunctions made of organic polymer donors and inorganic
nanoparticle acceptors. This combination of donor – acceptor pairs is interesting
7
because the inorganic nanoparticles have small molecule ligands on their surface
that are used to usually stabilize them in a solvent via dispersion. These ligands have
also been shown to not only influence the electronic properties of the nanoparticle
itself, but also act as a coupling agent that dictates the nature of the interface
between the nanoparticle and the polymer. In this manner the choice of ligand can
have profound influence on the charge carrier dynamics when this donor-acceptor
blend is photoexcited. Ultimately these carrier dynamics and energetics influence the
device properties of the solar cells like short circuit current (J sc), open circuit voltage
(V oc) and power conversion efficiency (PCE). In chapter 2, the charge carrier dynamics
in CdSe:P3HT BHJs is discussed in the context of different surface ligands that cap
the nanoparticle – a clear correlation is seen between the J sc and the
photoluminescence lifetimes as measured using a time correlated singlet photon
Figure 1.4. Excited state processes discussed in this thesis.
8
counting apparatus. In chapter 3 the charge carrier dynamics in PbS:Si-PCPDTBT
BHJs is investigated using transient absorption. Among the two kinds of surface
ligands capping the PbS nanoparticle, the inorganic PbI 2 ligand is found to mediate
charge transfer more efficiently than the organic NH 4I ligand – as seen in the power
conversion efficiencies of the solar cells made using these BHJs as the active layer.
Using transient absorption spectroscopy, the excited state dynamics of the BHJs can
be directly monitored and the observations correlate with the higher PCE of the
PbS(PbI 2):Si-PCPDTBT hybrid.
The left side of the figure depicts Singlet Fission (SF): a process by which a
molecule in the singlet excited state (S 1) shares energy with an adjacent molecule that
is in the ground state (S 0) to result in a triplet state on both molecules (T 1 – T 1).
34-38
The
transition from two molecules in a singlet and ground states to two molecules in
triplet states is mediated by a multiexciton (ME) state (
1/3/5
(T 1T 1)) in which the triplet
states on both molecules are electronically coupled and their spins are coherently
linked. Often the final two independent triplets undergo triplet-triplet annihilation
(TTA) to come back to the ground state.
39-42
If the resultant two triplet states can be
extracted out of the photoactive layer in a solar cell that contains a SF material, it has
the potential to deliver twice the number of charge carriers from the absorption of a
single photon when compared to the normal case when a photoexcited molecule
generates a single pair of charge carriers for every photon absorbed. In chapter 4
singlet fission in covalently linked tetracene dimer molecules in solution phase are
9
investigated using femtosecond transient absorption and photoluminescence
measurements. The three dimers presented in this chapter are linked in ortho-, meta-
and para-conformations. Such covalently linked dimers turn out of be ideal to study
the effect of molecular conformations on the electronic coupling that mediates
singlet fission.
43-50
The results from this chapters provide insight into design rules for
making molecules that singlet fission efficiently. Chapter 5 presents the excited state
properties of tetracene dimers in the film phase. By using transient absorption and
kinetic modeling, the effects of intra-dimer versus inter-dimer coupling on singlet
fission is observed. For example, the meta-tetracene dimer does not under singlet
fission in solution but does so in neat thin films owing to the fact that the intra-dimer
coupling (which is the only SF mediating force in solution) is weak while the inter-
dimer coupling (which is present only in thin film) is strong.
Figure 1.5. The internal structure of a photovoltaic cell (left) and schematic of
charge migration in the bulk heterojunction active layer in the cell (right).
10
Investigating the excited state properties of these molecular systems was made
possible using state of the art femtosecond spectroscopy tools. The first of these is a
home built broadband transient absorption system as seen in top half of Figure 1.6.
51
A Ti:sapphire oscillator (Micra, Coherent) operating at 76 MHz was used to seed a
Ti:sapphire amplifier (Legend, Coherent) which generated pulses of 800 nm light at 1
kHz with a width of ~ 35 fs. A fifth of the total amplifier output (3.2 W) was taken
through a retroreflector mounted on a programmable 30 cm translation state and then
focused on either a rotating 2 mm thick CaF 2 disc to generate a continuum in the 400
– 700 nm range or focused on a piece of sapphire to generate continuum in the 1000-
1500 nm range. The generated continuum beam was the probe for the transient
absorption experiment and was focused using off axis parabolic mirrors on the
sample stage which was moved in a direction perpendicular to the beam to avoid
photobleaching. The remaining 800 nm from the amplifier was used to pump an
optical parametric amplifier (OPA-800C, SpectraPhysics) that generated the pump
beam for the experiment. By tuning the phase matching angle of the nonlinear crystal
inside the OPA and adjusting its optical delays, it was possible to generate pump
beams centered any where from 300 nm to 1500 nm. The pump beam was then
chopped to half the frequency (500 Hz) using a programmable mechanical chopper
and then focused using an achromatic lens on the sample such that the pump beam
spot on the sample completely covered the probe beam spot. After passing through
the sample, the probe beam was focused on the entrance slit of a grating
11
polychromator using a pair of lenses while the pump beam was blocked. The grating
inside the monochromator could be replaced depending on whether the visible or the
near infrared continuum was being used and came with the option of either 150
lines/mm or 300 lines/mm. Since the polychromator was setup such that the grating
dispersed thee probe beam on to a 256-pixel Si photodiode array (for the visible probe)
or an InGaAs photodiode array (for the near infrared probe), the resolution of the
Figure 1.6. Layout of the femtosecond transient absorption setup (top) and
schematic showing how the measurement is made (bottom).
12
detection apparatus could be changed by using the appropriate grating. The output of
the photodiode arrays was carried to a breakout box that contained home built
circuitry that ran the sequence of events that collected data from the arrays by using
trigger signals from the laser and chopper to sequence the experiment correctly.
52
Data from the electronics box was then sent to the PC where a Labview interface was
used to visualize it in real time. As shown in the bottom half of figure 1.6, one cycle of
the transient absorption experiment involved the sample being probed by the
continuum, excited by the pump beam and then probed by the continuum beam again
after a variable delay. The first probe beam measures the absorption of the sample
when all the molecules are in the ground state while the second probe beam measures
the absorption of the sample when a fraction of the molecules are in the excited state.
Figure 1.7. Layout of the Time-Correlated Single Photon Counting Setup used to
measure photoluminescence lifetimes.
13
The second probe beam therefore sees less molecules in the ground state, and the
rest of the molecules in various excited states. A transient spectrum is the difference
spectrum between the probe beam spectrum after the sample is excited with the
pump and the probe beam spectrum before the sample is excited. Positive signals in
the transient spectrum therefore reflect excited state absorptions while negative
signals reflect bleaching of the ground state. By variably delaying the arrival of the
second probe beam after the pump, it is possible to capture the changes in the
features in the transient spectrum with time which then reflect excited state
dynamics of the sample.
The second of the time resolved spectroscopies is a Time-Correlated Single
Photon Counting Apparatus (TCSPC) that allowed for the measurement of
photoluminescence lifetimes (Figure 1.7).
53
5W of the output of the 534 nm continuous
wave beam from a Verdi V-18 (Coherent) was used to pump a Ti:sapphire oscillator
(Mira Seed, Coherent) producing a 76 MHz pulse train of width ~80 fs. The output of
the oscillator was expanded in time using a pair of gratings inside the
Expander/Compressor compartment and then used to seed a Ti:sapphire amplifier
(RegA 9050, Coherent) which was pumped using the remaining 13 W from the Verdi V-
18. The amplifier beam was then compressed to generate 1.3 W 800 nm pulsed at 250
kHz. 20% of the amplifier output was taken to a fast photodiode to generate a 250 kHz
SYNC signal that synchronizes the TCSPC experiment. The remaining output of the
amplifier was used to pump an optical parametric amplifier (OPA9450, Coherent)
14
which could be tuned to generate excitation beams centered anywhere from 400 to
750 nm. For the TCSPC experiment, the output of the OPA is used to excite the sample.
Fluorescence from the sample is collected in a perpendicular geometry using a pair
of plano-convex lenses and focused at the entrance slit of an additive double
monochromator (Digikrom CM112, Spectral Products). The monochromator is
controlled using its proprietary software to direct the wavelength of interest at the
detector which is a head-on PMT from Hamamatsu. The outputs of both the SYNC
signal from the fast photodiode and the signal from the PMT being triggered by the
fluorescence photon is carried to constant fraction discriminators (CFD) which help
eliminate signals from noise sources like thermal noise. The two signals are then
taken to the time amplitude converter circuit (TAC) in which the signal from the PMT
is used as a start trigger and the signal from the fast photodiode is used as the stop
trigger in a circuit that ramps a voltage linearly between the arrival of the two trigger
signals. The amplitude of this voltage is then a proxy for the time difference between
the two triggers. By varying the electronic delay, the fast photodiode signal passes
through before reaching the TAC, it is possible to match that event with excitation of
the sample. Thus, the time difference between the two signals represents the time
spent by the sample molecule in the singlet excited state before relaxing back to the
ground by releasing the fluorescent photon. For each pulse of the laser, one such
timing event is collected and by repeating for a few million times, the measured time
is built into a histogram by software on the PC. The distribution of time differences
15
then represents the decay profile of the singlet state, and by deconvoluting the
instrument response and fitting the curve with exponentials, it is possible to derive
the lifetime of the singlet state of the molecule being observed. While the TCSPC
apparatus proved useful in measuring the singlet lifetimes of many organic molecules
and inorganic nanoparticles, it suffered from two major drawbacks. First, its
instrument response was limited to 22 ps and frequently we encountered molecules
that displayed faster singlet kinetics that could not resolved with this setup. Often,
inorganic nanoparticles and polymers were encountered which had emission > 700 nm
that could not be measured using this setup. The second limitation of this setup was
the spectral response of the detector PMT dropped after 700 nm. To overcome both
these limitations, it was decided to build a fluorescence upconversion spectrometer.
The principles and design of this setup is explained in Chapter 6 bringing the contents
of this thesis to a conclusion.
16
CHAPTER 2
Charge Transfer Dynamics in
CdSe:P3HT Bulk Heterojunctions
2.1 INTRODUCTION
Surface chemistry of quantum dots dictates their photovoltaic properties owing to
their inherently large surface-area-to-volume ratios. The surface chemistry of these
solids is generally affected by their exposed crystal facets,
13, 54
surface
stoichiometry,
55, 56
and organic ligands that satisfy the coordination requirements of
their surface atoms.
57-59
The native ligands like oleic acid imprinted on quantum dot
surfaces through the solution phase synthesis serve two purposes - they provide
dispersibility in non-polar organic solvents and passivate electronic trap states.
These long-chain aliphatic nature of these native ligands (e.g., C14-C18 fatty acids
such as myristic acid or oleic acid) make them electrically insulating and, therefore,
they negatively affect charge mobility and conductivity in quantum dot solids.
60
Exchanging native ligands with smaller incoming ligands that still impart solution
dispersibility and allow for subsequent exciton management is therefore a hot area of
research into quantum dot chemistry.
61-67
Hybrid polymer:quantum dot bulk heterojunction (BHJ) solar cells require the
formation and separation of excitonic states and the creation of mobile charge
17
carriers across a donor/acceptor interface; these processes are highly dependent
upon the spatial and electronic coupling between the quantum dots and polymer
phase.
21, 68
These processes are intimately tied to the ligand shell separating the
polymer donor from the inorganic quantum dot acceptor. Exchanging the insulating
native ligands with smaller carboxylic acids (e.g., hexanoic acid),
69
amines (e.g.,
butylamine and pyridine),
19, 70-72
or sulfur-containing (e.g., tert-butylthiol)
20, 73
ligands
has been successful in mitigating against poor charge transport, trap states, and poor
electronic coupling in the acceptor phase of these hybrid solar cells.
74-76
Mitigation of
interfacial effects by engineering the quantum dot ligand shell can have a significant
influence on the overall device performance.
77-79
While the previously mentioned
ligand types are successful in having an overall positive effect on device performance,
it is difficult to directly compare and benchmark such ligand effects on specific device
parameters like short circuit current density (J SC) or open circuit voltage (V OC) across
various reports because of differences in polymer molecular weights and
polydispersity, quantum dot size, polydispersity and purification, ligand exchange
procedures, and device design, processing and testing conditions. For example,
device performance for hybrid poly(3-hexythiophene-2,5-diyl) (P3HT):CdSe quantum
dot BHJ solar cells where the quantum dots were ligand exchanged with the
ubiquitous pyridine ligand have been reported to range from 0.4-1.9% depending on
the quantum dot size, ligand exchange protocols, and nanocrystal purification
procedures.
80, 81
There has been no direct comparison of ligand effects on the
18
photophysics and device performance of hybrid polymer:CdSe quantum dot solar
cells made from the same parent batch of quantum dots that have been colloidally
exchanged using different ligand types.
This chapter is a systematic study of the ligands: pyridine, butylamine, tert-
butylthiol, thiophenol, pivalic acid, and tetrahydrothiophene for CdSe quantum dots.
The effect these different surface ligands have on the photophysics of the quantum
dots and hence on the device parameters in hybrid P3HT:CdSe quantum dot BHJ solar
is investigated using time-correlated single photon counting. Comparing the
photoluminescence decays from neat films of the ligand-capped quantum dot films
and decays from P3HT:CdSe BHJ films sheds light on the charge transfer between the
P3HT polymer and CdSe quantum dot, and how this process is influenced by the
chemistry of the surface ligand on the quantum dot. The photoluminescence decays
are also found to be correlated with the short circuit current J sc of photovoltaic devices
fabricated using the P3HT:CdSe BHJs as the active layer – providing insight into how
surface ligands on the quantum dots can influence the overall photoconversion
efficiencies of these devices.
2.2 EXPERIMENTAL SECTION
2.2.1 General Considerations and Synthetic Procedure
Chemicals for the synthesis of quantum dots, ligands, and P3HT were purchases
from vendors like Sigma Aldrich, Alfa Aesar and were all used as received. The
synthesis is based on literature methods
82
(Figure 2.1).
19
2.2.2 Ligand Exchanges
Ligand exchanges generally involved mixing a stock solution of the ligand to be
introduced on the quantum dot surface to a dispersion of as-prepared CdSe quantum
dots in toluene, followed by flocculation with methanol and centrifugation. The ligand
exchanged CdSe quantum dots were redispersed in the ligand stock solution and the
ligand exchange process was repeated 4 times. After the final wash, the ligand-
exchanged CdSe quantum dots were dispersed in 1,2-dichlorobenzene, filtered
through a 0.45 µm PTFE filter, and stored in the dark at 10 °C. Each ligand exchange
was individually optimized and confirmed by TGA, FT-IR, XPS, and
1
H NMR
spectroscopies. All concentrations of CdSe quantum dots were determined
gravimetrically.
2.2.3 Device Fabrication
Indium tin oxide (ITO) coated glass substrates were cleaned by sequential
sonication in tetrachloroethylene, acetone, and isopropyl alcohol. Finally, the ITO was
cleaned by UV-ozone for 30 min. A layer of poly(3,4-ethylenedioxythiophene)–
poly(styrene sulfonic acid, PEDOT:PSS) was spun-cast onto the ITO. The substrate
was then heated to under vacuum for 1 h. Solutions of P3HT in 1,2-dichlorobenzene
were prepared by heating at 50 ˚C until all the P3HT was dissolved. To make active
layer solutions of the desired loading ratios (wt/wt P3HT:CdSe), the required amount
of CdSe quantum dots was mixed with P3HT. These solutions were spun-cast onto the
PEDOT:PSS covered ITO forming BHJ films with a thickness between 70−100 nm, as
20
determined by ellipsometry (see Table 2.1). These films were dried in the dark under
flowing nitrogen and once dried, a solution of 40-nm ZnO nanocrystals in ethanol was
spun-cast and dried once again in the dark under nitrogen. The dried film was
annealed at different temperatures under flowing nitrogen on top of temperature-
controlled aluminum heating block, and then allowed to cool in the dark under flowing
nitrogen. Al deposition in a thermal deposition chamber under high vacuum.
2.2.4 Characterization
UV-vis spectra were acquired using a quartz cuvette for liquid samples on a Shimadzu
UV-1800 spectrophotometer. The absorption spectra of thin films on glass substrates
were collected in a 60-mm integrating sphere using a Perkin-Elmer Lambda 950
UV/Vis/NIR spectrophotometer. Photoluminescence spectra were conducted on a
Horiba NanoLog Spectrofluorometer System for all films. Neat P3HT films were cast
using the same conditions as the devices with the same amount of P3HT in its
solution. Thermogravimetric analysis (TGA) traces were obtained using dried CdSe
quantum dot sample in an alumina pan under flowing nitrogen with a TA Instruments
TGA Q50 model.
Cyclic voltammetry (CV) measurements were performed using a BASi Epsilon-EC
potentiostat. The thin film absorption spectra in these experiments were collected in
transmission mode with the substrate mounted inside of an electrolyte solution-filled
1 cm quartz cuvette. A bare ITO slide immersed in an electrolyte solution was used as
the reference blank for these measurements. A 0.1 M electrolyte solution containing
21
tetra-n-butylammonium hexafluorophosphate (TBAP, 98%, Sigma Aldrich) and dry
and degassed acetonitrile was put in a clean 25 mL three neck round bottom flask. A
Table 2.1 Photovoltaic Device Parameters for Optimized Hybrid P3HT:CdSe BHJ
Solar Cells with Different Ligands
Ligand
P3HT:CdSe wt/wt
(mg)
Film Thickness
(nm)
Annealing
Temperature (°C)
J SC (mA cm
–2
)
V OC (V)
FF
PCE (%)
Champion PCE
NL
Native
Ligand
18:3 98 175
1.02
± 0.03
0.564
± 0.046
0.45
± 0.02
0.26
± 0.03
0.29
PA
Pivalic Acid
18:3 89 200
6.67
± 0.16
0.625
± 0.004
0.53
± 0.01
2.23
0.12
2.35
Py
Pyridine
18:3 94 175
6.59
± 0.34
0.600
± 0.005
0.54
± 0.01
2.14
± 0.10
2.26
BA
Butylamine
18:3 97 200
4.71
± 0.23
0.710
± 0.004
0.52
± 0.01
1.72
± 0.11
1.89
tBT
tert-
Butylthiol
12:3 84 175
5.73
± 0.08
0.819
± 0.003
0.47
± 0.01
2.21
± 0.04
2.25
TP
Thiophenol
9:3 77 225
6.78
± 0.35
0.685
± 0.004
0.49
± 0.01
2.26
± 0.14
2.49
THT
Tetrahydroth
iophene
6:3 74 250
5.41
± 0.04
0.625
± 0.005
0.54
± 0.01
1.82
± 0.05
1.86
22
dilute CdSe quantum dot suspension was drop cast onto a glassy carbon (GC)
electrode and allowed to dry. A Pt wire counter-electrode and Ag wire pseudo-
reference electrode were placed in the flask along with the GC electrode.
Measurements were conducted under flowing nitrogen. The Ag electrode was
calibrated against a ferrocene/ferrocenium (Fc/Fc
+
) redox couple and all potentials
were reported relative to the normal hydrogen electrode (NHE). The electrolyte
solution was replaced and the GC electrode was cleaned between samples.
Time correlated single photon counting (TCSPC) measurements samples were
spun cast onto glass in a glovebox with optical densities between 0.1-0.2 at the
excitation wavelength of 500 nm. To avoid oxidative damage, the films also had an
additional glass window on the top surface, and the outer edges were sealed with
epoxy under a nitrogen atmosphere. Photoluminescence lifetimes were measured by
detecting the emission at 650 nm for the CdSe quantum dot neat films, and at 725 nm
for P3HT:CdSe hybrids, and neat P3HT films. TCSPC measurements were performed
using an R3809U-50 Hamamatsu PMT with a B&H SPC-630 module (time resolution of
22 ps). The grating placed in the monochromator was blazed at 600 nm with 1200
g/mm. The lifetime data were measured using an excitation fluence of 0.28 μJ cm
–2
for the CdSe quantum dot neat films and P3HT:CdSe hybrid films. For neat P3HT films,
an excitation density of 3.34 μJ cm
–2
was used.
Current vs voltage curves for the hybrid P3HT:CdSe quantum dot devices were
obtained under ambient conditions using a Keithley 2420 SourceMeter in the dark and
23
under ASTM G173-03 spectral mismatch corrected 1000 W m
−2
white light
illumination from an AM 1.5G filtered 300 W Xe arc lamp (Newport Oriel). Chopped and
filtered monochromatic light (250 Hz, 10 nm fwhm) was used from a Cornerstone 260
1/4 M double grating monochromater (Newport 74125) along with a lock-in amplifier
EG&G 7220 to perform all EQE measurements.
2.3 RESULTS AND DISCUSSION
2.3.1 Colloidal Ligand Exchange
The CdSe quantum dots were prepared according to literature methods.
32
The
purified quantum dots were 5.44 ± 0.28 nm in diameter, as determined by TEM
analysis, which is in agreement with the diameter calculated from the first exciton
peak at 627 nm in the absorption spectrum by using the empirical sizing equation of
Figure 2.1. TEM micrograph of the as-prepared nanocrystals with the
corresponding size histogram (left panel). The histogram displays the size
distribution of 100 CdSe quantum dots found to have an average size diameter of
5.44 ± 0.28 nm, which agrees with the empirical sizing analysis by UV-vis (right
panel).
24
Jasieniak et al.
33
Ligand exchange reactions were performed using a 70-80 mg mL
–1
stock solution of the as-prepared CdSe quantum dots in toluene, which were then
refluxed or stirred with a ligand for a given time and number of repetitions based on
ligand type. See Figure 2.1a for absorption spectra of ligand exchanged quantum dots.
In each case, the resulting ligand-exchanged CdSe quantum dots retain the position
of the first exciton peak in their respective UV-vis absorption spectra at ca. 627 nm,
suggesting that the quantum dot size remains largely unchanged.
2.3.2 Effect of Ligand on Device Performance
The solution deposition of BHJ active layers requires iterative progressions
through donor/acceptor loading ratio, active layer thickness, processing solvents,
etc. Without individually optimizing these processing conditions for every
donor/acceptor couple, it is impossible to realize the optimum device performance,
and thus develop an understanding of what donor/acceptor combinations work best.
There has not been a methodical study that explores a series of chemically distinct
ligand frameworks and individually optimizes the device processing conditions for
each quantum dot-ligand acceptor type in order to adequately disentangle the effects
that each ligand has on P3HT:CdSe quantum dot BHJ solar cell performance. Active
layers for the hybrid solar cells used in this study were incorporated into the general
device structure ITO/PEDOT:PSS (30 nm)/P3HT:CdSe BHJ (70-100 nm)/ZnO (40 nm)/Al
(100 nm) and were deposited by spin-casting blends of P3HT with the ligand
exchanged CdSe quantum dots dispersed in 1,2-dichlorobenzene. Two device
25
fabrication parameters were found to significantly influence PCE; that is, the loading
ratio of CdSe quantum dots to P3HT (varied here between 3:3 to 24:3 mg CdSe/mg
P3HT) and the device annealing temperature (varied here between 150-275 °C). The
effect of loading ratio and annealing temperature on short circuit current density (J SC),
open circuit potential (V OC), fill factor (FF), and PCE for each ligand has been published
previously. The optimized devices were then fabricated three times for each ligand to
Figure 2.2. Absorption spectra of ligand exchanged neat quantum dots (a), that of
bulk heterojunctions made of ligand exchanged quantum dots and P3HT polymer
(b), emission spectra of ligand exchanged neat quantum dots c) and that of
corresponding bulk heterojunctions (d).
26
statistically confirm the results. Table 2.I gives the average device parameters for
each device fabricated under optimized conditions.
Devices made from CdSe quantum dots with PA, Py, tBT, and TP ligand shells all
yield mean power conversion efficiencies > 2.1%. It is interesting to note that previous
work that directly compared the Py and tBT ligands in hybrid P3HT:CdSe quantum dot
devices that were identically processed (i.e., 8:1 wt/wt ratio of CdSe to P3HT and no
thermal anneal) gave significantly different device performance; that is, PCEs of 1.0%
(J SC = 3.69 mA cm
–2
, V OC = 0.57 V, FF = 0.47) for the Py ligand shell and 1.9% (J SC = 5.62
mA cm
–2
, V OC = 0.80 V, FF = 0.43) for the tBT ligand shell.
20
Here, with independent
optimization being performed for each ligand, the PCE of the hybrid P3HT:CdSe
quantum dot devices with Py and tBT ligands are nearly identical. The hybrid
P3HT:CdSe quantum dot device with Py ligands gave an average PCE of 2.1% (J SC =
6.59 mA cm
–2
, V OC = 0.60 V, FF = 0.54) while that with tBT ligands gave an average PCE
of 2.2% (J SC = 5.73 mA cm
–2
, V OC = 0.82 V, FF = 0.47). This example comparing Py and
tBT demonstrates the importance of independent optimization when comparing the
effects that different ligands have on their device efficiencies.
The optimization process allows for better comparisons to be made across all the
ligand frameworks being screened and their respective device parameters. In the
optimized devices, the TP ligand gave the highest J SC (6.78 mA cm
–2
), with PA and Py
close behind with J SC values of 6.67 and 6.59 mA cm
–2
, respectively. Although tBT
ligands gave lower J SC in the hybrid P3HT:CdSe quantum dot devices, it gave the
27
highest V OC (0.82 V), followed by TP at 0.69 V. The FF values for these two devices
utilizing tBT and TP were 0.47 and 0.49, respectively, with similar PCEs of 2.2% and
2.3%. The Py and PA ligands both gave lower V OC (0.60 and 0.63 V, respectively), but
higher FF values of 0.54 and 0.53. The higher FF allowed for device efficiencies of 2.1%
and 2.2% for Py and PA, respectively. These results demonstrate that different
ligands affect different device parameters (J SC, V OC, FF), but through device
optimization, the average PCE achieved in devices utilizing these four different
ligands (i.e., Py, PA, tBT, TP) can be made to be quite similar. The BA and THT ligands
both resulted in slightly lower PCEs (1.7 and 1.8%, respectively) for hybrid P3HT:CdSe
quantum dot devices, mainly resulting from lower photocurrent densities collected in
devices utilizing these two ligands. As expected, the hybrid P3HT:CdSe quantum dot
solar cells with the native ligands gives the lowest PCE of 0.3%, resulting from a low
short circuit current density (1.02 mA cm
–2
) caused by inefficient charge transfer and
collection from the insulating ligand shell on the CdSe quantum dots.
2.3.3 Effect of ligand type on J SC
Steady-state and time-resolved photoluminescence (PL) spectroscopies were
used to rationalize the short circuit current densities achieved for each ligand under
optimized device conditions. Figure 2.1d compares the normalized (at 725 nm)
steady-state PL of all the hybrid P3HT:CdSe quantum dot film to a neat P3HT film. The
emission from the hybrid BHJ films with ligands THT, tBT, BA, PA and TP closely
resemble that of a neat P3HT film with a maximum at 725 nm. The hybrid films with
28
native ligand and Py as the surface ligand exhibit emission spectra that have more
contribution from the CdSe component centered around 650 nm. All films were
prepared under the same conditions as the optimized working devices.
Although both
the quantum dot and the P3HT hybrid film components absorb at 550 nm, the
absorption cross section of the P3HT at this wavelength is significantly greater than
the nanocrystals. In fact, under the optimized nanocrystal:P3HT loadings utilized, we
estimate that the P3HT absorbs at least 70% of the incident light at 550 nm, with that
Table 2.2. Lifetimes and amplitudes obtained by fitting photoluminescence decays
with exponentials and the calculated average lifetime.
Ligand
Neat CdSe (ligand) films CdSe(ligand):P3HT BHJ films
Lifetimes (ns) and
amplitudes (%)
Average
lifetime (ns)
Lifetimes (ns) and
amplitudes (%)
Average
lifetime (ns)
NL
1.38 (75), 6.96
(22), 29.63 (3)
11.12 0.059(91), 0.262(9) 0.121
PA
0.85 (82), 4.80
(17), 22.04 (2)
7.280 0.048(97), 0.254(3) 0.077
Py
1.14 (93), 5.67 (6),
28.07 (0.2)
3.236 0.03(93), 0.187(7) 0.080
tBT
0.85 (61), 4.59
(35), 23.47 (5)
10.71
0.025(87),
0.153(13)
0.086
TP
1.38 (83), 7.69
(15),
33.86 (2)
7.209 0.042(95), 0.202(5) 0.074
THT
0.85 (77), 5.23
(20), 24.25 (2)
8.139 0.046(91), 0.239(9) 0.111
29
Figure 2.3. Photoluminescence decay profiles of neat ligand exchanged quantum
dots. Grey dots is the data obtained from TCSPC and colored lines are
multiexponential fits.
30
percentage increasing for the systems with lower nanocrystal loadings (i.e., tBT, TP,
THT hybrid ligand sets).
Interpretation of steady-state PL quenching data can be complicated by
differences in sample thickness and sample orientation within the instrument;
however, time-resolved photoluminescence measurements can be used to avoid
these complications.
19
Although both CdSe and P3HT components in the hybrid films
absorb at 550 nm, the PL lifetime measurements on hybrid films at 725 nm largely
monitor the decay of the emission from the donor polymer since the emission from
the CdSe reaches a maximum at 650 nm and is negligible at 725 nm. It is important to
measure the PL lifetimes of both the neat donors and acceptors, and the hybrid BHJ
films for comparison. The emission at 650 nm from neat CdSe quantum dot films
Figure 2.4. Photoluminescence decays from P3HT films annealed at different
temperatures.
31
excited at 550 nm measured over 50 ns is shown in Figure 2.3. The decay profiles of
the emission could not fit with a single exponential, but it is most useful to
characterize weighted average lifetimes. Average lifetimes of neat CdSe quantum dot
films (see Table 2.2) varied from 3.24 ns for CdSe(Py) up to 11.1 ns for CdSe(NL). The
PL lifetimes of neat P3HT films depended on their thermal annealing history but were
always between 60-150 ps for annealing temperatures relevant to device fabrication
(i.e., 150-250 °C) (see Figure 2.4). Weighted average lifetime was calculated using 𝜏 =
∑ 𝑎 𝑖 𝑡 𝑖 2
𝑖 ∑ 𝑎 𝑖 𝑡 𝑖 𝑖 , where a i and t i correspond to the amplitudes and time constants of individual
exponents, respectively.
83
Decreases in PL lifetimes of the neat samples upon incorporation into a BHJ imply
charge or energy transfer between the donor and acceptor. Hybrid P3HT:CdSe
quantum dot BHJ films prepared identically to optimized devices were excited with
an ultrafast 550 nm pulse, and the PL decays were probed at 725 nm for up to 50 ns.
Figure 2.5 shows the multi-exponential decays for each hybrid P3HT:CdSe quantum
dot BHJ film after excitation. We observed ligand-dependent average lifetimes
ranging from 74 ps for P3HT:CdSe(TP) hybrid films up to 121 ps for P3HT:CdSe(NL)
hybrid films. The time constants and their respective amplitudes obtained by fitting
the decays to exponential functions are given in Table 2.2 and the weighted average
lifetime was calculated using the same method described earlier for the neat
quantum dot films. Not surprisingly, the longest and shortest average lifetimes were
observed for the systems with smallest (P3HT:CdSe(NL)) and largest values
32
Figure 2.5. Photoluminescence decays from bulk heterojunctions made from CdSe
quantum dots and P3HT polymer.
33
(P3HT:CdSe(TP)) of J SC, respectively. The order of decreasing average PL lifetime of
the BHJ films from P3HT:CdSe(NL) to P3HT:CdSe(TP) (Table 2.2) is reflected in the
increasing order of their respective J SC values (Table 2.I). Fig 2.6c shows this negative
correlation between the J SC values and the average photoluminescence lifetime of the
hybrid films. The short lifetimes suggest fast charge transfer, which contribute to the
large J SC values measured in devices. The unusually long lifetime observed for the
P3HT:CdSe(NL) hybrid film may be a result of the well-passivated, but electronically
insulating, CdSe(NL) surface. If electron transfer from P3HT to CdSe(NL) occurs,
albeit with low efficiency due to the insulating nature of the NL shell, then it is
possible that radiative recombination may occur between the excited hole on P3HT
and the transferred electron on CdSe(NL). The NL shell is believed to provide a well-
passivated, relatively trap-free surface based on the long lifetimes observed in TCSPC
measurements on neat CdSe(NL). It is important to note that all three of the hybrid
Figure 2.6. (a) Open circuit potentials vs ΔEDA for all the optimized P3HT:CdSe
quantum dot BHJ devices with a general device structure ITO/PEDOT:PSS (30
nm)/P3HT:CdSe BHJ (70-100 nm)/ZnO (40 nm)/Al (100 nm) (slope = 0.6 V eV-1 and
y-intercept = -0.14 V). (b) Dark I-V characteristics for each optimized P3HT:CdSe
quantum dot BHJ device. (c) Device short circuit currents vs average PL lifetime
of the hybrid films measure using time correlated single photon counting.
34
BHJ samples made with sulfur-containing ligands (TP, THT and tBT) exhibit lifetimes
that are shorter than both neat P3HT and the respective neat CdSe quantum dot films.
This suggests that electron transfer from P3HT to CdSe, as well as hole transfer from
CdSe to either P3HT or the ligand, may be occurring both fast and efficiently. In the
case of PA and Py ligands, the hybrid BHJ lifetimes were 77 and 80 ps, respectively.
These lifetimes are shorter than those measured in neat CdSe(PA) (7.3 ns) and
CdSe(Py) (3.2 ns) quantum dot films. Indeed, pyridine has been previously shown to
act as a hole trap on CdSe and inhibit hole transfer to P3HT.
22
It was difficult to
accurately characterize the ultrafast dynamics of the BA-treated nanocrystals as
they consistently exhibited excited state dynamics that were unstable or not
reproducible even over short periods of time, likely because of the ligand volatility.
2.3.4 Effect of ligand type on V OC
It has been shown that organic ligands affect the frontier orbital energies of CdSe
quantum dots.
84-88
In organic photovoltaics, it is generally understood that the
maximum theoretical V OC is limited by the energy difference ( ΔE DA) between the
highest occupied molecular orbital (HOMO) of the donor and the lowest unoccupied
molecular orbital (LUMO) of the acceptor.
17
Since P3HT was used as the donor polymer
in all of these devices (with a HOMO energy of -5.2 eV), the relative differences in
optimized device V OC should correlate with ΔE DA and the CdSe quantum dot LUMO
values. Spectroelectrochemistry (SEC) suggested differences in CdSe quantum dot
LUMO values, and cyclic voltammetry (CV) was used to obtain the relative LUMO
35
energies (E LUMO) from the onset potential of the CdSe quantum dot reduction waves.
89
The energies of the LUMO level relative to vacuum were calculated from the CV onset
potential according to equation 1, where E red is the onset value of the reduction wave
vs NHE.
90, 91
E LUMO = − (E red + 4.5) eV (1)
While the optical band gaps of the ligand exchanged CdSe quantum dots do not
change significantly with ligand type, SEC measurements suggested a difference in
LUMO energy as a function of ligand framework. Absorption spectra of the CdSe
Table 2.3. Electrochemically measured E red and E LUMO levels for the CdSe quantum
dots as a function of ligand type, and the corresponding ΔE DA and optimized
P3HT:CdSe device V OC.
Ligand E red (V)
a
E LUMO (eV)
b
ΔE DA (eV)
c
V OC (V)
NL -0.479 -4.02 1.18 0.564±0.046
PA -0.604 -3.90 1.30 0.625±0.004
Py -0.549 -3.95 1.25 0.600±0.005
BA -0.689 -3.81 1.39 0.710±0.004
tBT -0.924 -3.58 1.62 0.819±0.003
TP -0.630 -3.87 1.34 0.685±0.004
THT -0.605 -3.89 1.31 0.625±0.005
a
Onset value of the reduction wave by CV vs the normal hydrogen electrode;
b
given
relative to vacuum;
c
using a E HOMO of -5.2 eV for P3HT.
36
quantum dot first exciton peak ( λ 1st exciton = 627 nm) tracked under reducing conditions
in a non-aqueous electrochemical cell have been previously published. When the
applied bias approaches the onset of the reduction potential determined from CV
measurements, injection of electrons into the CdSe LUMO causes the first exciton
peak at 627 nm to electrochemically bleach. CdSe(Py) quantum dot films were found
to have the energetically lowest (i.e., most negative) LUMO level by SEC, which
undergo bleaching at the least negative potential between -1000 and -1100 mV vs Ag
wire pseudo-reference electrode. The highest LUMO level by SEC is observed for
CdSe(tBT) quantum dot films, which undergo bleaching at the most negative applied
potential between -1350 and -1400 mV vs the Ag wire pseudo-reference electrode.
The trends observed in SEC experiments directly align with the trend in ligand-
induced LUMO level shifts observed by CV measurements (Table 2.3 and Figure 2.7).
The change in E LUMO with ligand substitution causes a change in ΔE DA because of
the fixed E HOMO for P3HT, which in turn affects the device V OC. The trend in ΔE DA from CV
and SEC correlates linearly with the V OC measured in optimized devices, as expected,
with a slope of 0.6 V eV
-1
and a y-intercept of -0.14 V (Fig. 2.6a).
92
As in typical OPV
systems, the V OC of the hybrid systems investigated can never reach the full value of
ΔE DA at room temperature due to losses originating from bimolecular charge carrier
recombination as well as dielectric losses from charge-transfer state exciton binding
energies. Additionally, trap states at the donor-acceptor interface should contribute
to further losses in V OC. Absent these losses, we would expect the slope of the V OC vs.
37
ΔE DA plot shown in Fig. 2.6a to approach a value of 1. The tBT ligand gives the highest
ΔE DA from the highest lying E LUMO, and correspondingly, gives the highest V OC (0.82 V).
In contrast, Py gave the lowest ΔE DA from the ligands screened, resulting in the lowest
lying E LUMO and the lowest V OC (0.60 V). To confirm that the observed range of 220 mV in
device V OC resulted from changes in ΔE DA and not to differences in the reverse
saturation currents (J O), the dark I-V curves were compared for all of the optimized
devices (Fig. 2.6b). Since all the devices were found to give J O within an order of
magnitude (10
–4
–10
–3
A cm
–2
), the effect that these differences have on V OC is not
significant enough to alter the overall trend measured by the device I-V curves. This
suggests that the correlation between the electrochemically measured ΔE DA and the
device V OC is robust.
Figure 2.7. CV traces and corresponding energy level diagram for CdSe quantum
dots ligand exchanged with PA, Py, BA, tBT, TP, and THT ligands relative to the as-
prepared CdSe quantum dots with native ligands (NL). Data were collected with a
scan rate of 20 mV s
–1
in 0.1 M TBAP with a Pt wire counter electrode and Ag wire
pseudo-reference electrode calibrated against the Fc/Fc
+
redox couple. All
potentials are shown relative to NHE.
38
2.4 CONCLUSIONS
While a great deal of effort has gone into hybrid solar cell device optimization by
changing the donor polymer and/or the quantum dot acceptor, much less effort has
gone into optimization of the interfacial ligands on the quantum dots. While it is well
understood that these interfacial ligands affect trap states on the acceptors, as well
as the quantum dot-polymer and quantum dot-quantum dot coupling, it has
heretofore been difficult to directly compare the effects of ligand type on device
parameters across multiple reports. This represents the first systematic study of the
effects of different interfacial ligands on the performance of hybrid polymer:CdSe
quantum dot solar cells, which (importantly) were made from the same parent batch
of CdSe quantum dots and independently optimized for each hybrid P3HT:CdSe
quantum dot BHJ device.
It was found that TP, Py, and PA gave the highest mean short circuit current
densities ( ≥ 6.6 mA cm
–2
), while BA and tBT gave the highest mean open circuit
potentials (> 0.7 V). A good correlation was found between P3HT:CdSe(ligand) BHJs
that possess short PL lifetimes and corresponding devices with the largest short
circuit current densities, suggesting that fast charge transfer across the
donor/acceptor interface is a large contributor to high J SC. The highest open circuit
potentials associated with the devices utilizing tBT and BA ligands is consistent with
those quantum dot-ligand constructs possessing elevated LUMO energies relative to
the HOMO of P3HT, thereby increasing ΔE DA and device V OC. In fact, a linear correlation
39
between ΔE DA and optimized device V OC was observed for the seven different ligand
types studied. The results from this chapter can be then used to derive design
principles to maximize photoconversion efficiencies of bulk heterojunction devices.
40
CHAPTER 3
Charge Transfer Dynamics in PbS:Si-PCPDTBT
Bulk Heterojunctions
3.1 INTRODUCTION
Organic polymer:fullerene bulk heterojunction (BHJ) solar cells offer the
opportunity for low-cost device fabrication as a result of the solution processability
of both active components, and in addition, employs a thin absorber layer that
translates to low materials costs.
33, 93
While organic photovoltaics have now achieved
power conversion efficiencies in excess of ~10%
94-96
, optimization of the organic
donor phase may be nearing its limit to further increase device performance. This has
led to increased emphasis on exploring new acceptor types.
97, 98
Semiconductor
nanocrystals have been explored as alternatives to the more well-established
fullerene acceptors, and they possess several attractive attributes in this regard: (i)
size- and composition-tunable absorption from the visible to near infrared (NIR), (ii)
intrinsically higher electron mobilities, (iii) the potential for multiple exciton
generation (MEG), and (iv) and higher dielectric constants to help overcome the strong
exciton binding energy of organic materials.
31, 99, 100
To date, hybrid solar cells based
on polymer:CdSe nanocrystal BHJs have been extensively explored and optimized
with respect to ligand engineering,
20, 77, 79, 101-103
size and shape effects,
73, 104-106
charge
41
transfer dynamics,
74107-110
and device architectures,
111-114
leading to power
conversion efficiencies (PCEs) of 4-5%. However, the best performing polymer:CdSe
nanocrystal BHJs only absorb visible light out to wavelengths that are limited by the
donor polymer, which prevents them from harvesting light into the NIR part of the
solar spectrum. For example, CdSe nanocrystals are intrinsically limited by the bulk
band gap (E g = 1.7 eV) to an absorption edge of ~730 nm. Moreover, the dielectric
constant of bulk CdSe (~6
115
) is not significantly different from that of the organic
polymer phase ( ~3-4
116
), whereas lower band gap semiconductors generally possess
high dielectric constants (e.g.,~17 for PbS
117
). Because of these attributes, lead
chalcogenide nanocrystals are being increasingly investigated as electron acceptors
in hybrid polymer:nanocrystal BHJ solar cells.
118-121
Although the photovoltaic performance of hybrid polymer:PbE nanocrystal BHJ
solar cells are promising, less known is about the effects of ligand engineering on
device performance when compared to CdSe acceptors.
109, 122
All of the high-
performing, state-of-the-art hybrid polymer:PbE (where E = S, Se, Te) nanocrystal
BHJ devices to date mandate a post-deposition, thin film ligand treatment to remove
insulating native ligands from the nanocrystal surface.
78–83
Such ligand exchange
processes are extremely critical to provide the interparticle electronic coupling that
is required for efficient excitonic processes to occur (i.e., charge transfer, charge
separation). Common bidentate ligands (e.g., 1,2-ethanedithiol (EDT), 3-
mercaptoproprionic acid)
109
and atomic halide ligands (e.g., iodide) have been applied
42
via thin film ligand exchange; however, thin film ligand exchange has several inherent
drawbacks: (i) ligand exchange can be impeded by slow and/or incomplete solid-state
ligand diffusion through the BHJ film, (ii) reduction of the film volume through
exchange of small ligands for large ones can lead to film cracking, and (iii) the dip-
washing process used for thin film ligand exchange can lead to a low ligand exchange
efficiency (yield) and may also be incompatible with large scale solution
processing.
123
Quantitative colloidal ligand exchange prior to BHJ deposition would
be beneficial, but it still remains a challenge to generate a stable suspension of
colloidal lead chalcogenide nanocrystals exchanged with small organic ligands
because of their tendency to agglomerate,
124
etch, and/or oxidize after ligand
exchange. On the other hand, atomic halide ligands have proven successful in
achieving high efficiency colloidal quantum dot solar cells, and in addition to allowing
for efficient interparticle coupling,
125-128
have been shown to effectively passivate
surface trap states in PbS nanocrystals.
125
This chapter reports the excited state dynamics of colloidal ligand exchanged PbS
nanocrystals using iodide ligands (either NH 4I or PbI 2) that allows them to be
dispersed in 1,2-dichlorobenzene. This facilitates, for the first time, the generation
of photovoltaic devices with halide-passivated nanocrystal acceptors (Figure 3.1)
that exhibit a broad spectral response into the NIR, leading to a PCE up to 4.8%. This
device performance is among the best for hybrid polymer:nanocrystal solar cells and
is the highest for a one-step BHJ deposition performed without any post-deposition
43
ligand exchanges. Time-resolved photoluminescence lifetimes using Time Correlated
Single Photon Counting (TCSPC) indicate efficient charge transfer from the excited
state of the polymer in hybrid films with PbI 2 ligand treatment when compared to
hybrid films with NH 4I surface ligands. Excited state dynamics measured using
femtosecond broadband transient absorption (TA) spectroscopy reveals that the
charge separated state in the PbS(PbI 2):Si-PCPDTBT hybrid bulk heterojunction has a
longer lifetime than in PbS(NH 4I):Si-PCPDTBT. Faster charge separation as seen in
TCPSC and slower recombination as seen in TA is directly correlated with the higher
PCE of the bulk heterojunction with PbI 2 ligand treatment.
3.2 EXPERIMENTAL SECTION
3.2.1 Materials, synthesis and ligand exchange
Lead oxide , lead iodide, and ammonium iodide, n-butylamine were purchased
from Alfa Aesar and used as received without further purification. The synthesis of E g
= 1.3 eV (first exciton peak ~920 nm) PbS nanocrystals was based on the procedure
developed by Hines and Scholes (Figure 3.1).
129
The final suspension was made in
toluene with a concentration of 50 mg mL
–1
. All ligand exchange processes are done
in air under ambient conditions (Figure 3.1). A solution of PbI 2 in a mixture of DMF and
methanol was prepared first. The suspension of as-prepared PbS nanocrystals was
then added into the PbI 2 solution and shaken for 1 −2 min, leading to an immediate
precipitation of the PbS nanocrystals. After centrifugation, the PbS nanocrystals
44
were isolated from the clear supernatant, followed by redispersing them in a mixture
of 1,2-dichlorobenzene and BA. For ligand exchange with ammonium iodide, a
solution of NH 4I in methanol (8 mL) was prepared. The suspension of as-prepared PbS
nanocrystals was then added into the NH 4I solution, leading to an immediate
precipitation of the PbS nanocrystals. After centrifugation, the PbS nanocrystals
were isolated from the clear supernatant, followed by redispersing them in a mixture
of 1,2-dichlorobenzene and BA. The ligand-exchanged PbS nanocrystals were then
filtered through a 0.45 m PTFE syringe filter for device fabrication.
3.2.2 Characterization
UV-Vis-NIR absorption spectra were acquired on a Perkin-Elmer Lambda 950
spectrophotometer equipped with a 150 mm integrating sphere, using a quartz
cuvette for liquid samples or a borosilicate glass microscope slide substrate for films.
Film thicknesses were determined using a J. A. Woollam variable angle spectroscopic
ellipsometer equipped with a 150 W Xe arc lamp. TEM images of BHJ films supported
on copper (Ted Pella, Inc.) were obtained on a JEOL JEM-2100F microscope at an
operating voltage of 200 kV, equipped with a Gatan Orius CCD camera.
3.2.3 Photoluminescence lifetime studies
For time correlated single photon counting (TCSPC) measurements, the samples
were spun cast onto borosilicate glass in a glovebox with optical densities between
0.1 −0.2 at the excitation wavelength of 500 nm. To avoid any oxidative damage, the
45
film also had an additional glass window placed on the top surface and the outer
edges were sealed with epoxy under a nitrogen atmosphere. Lifetime measurements
were carried out using the output of a Coherent RegA 9050 regenerative amplifier
operating at 250 kHz. The amplified 800 nm pulse was then used to pump an optical
parametric amplifier, Coherent OPA 9450, which produced the 500 nm excitation
beam used in the experiment. The lifetime data were measured using pump pulses of
power 0.5, 0.67, and 1.7 mW and with a spot size of 250 μm at the sample, resulting in
excitation fluences of 4.0, 5.4 and 14 μJ cm
–2
for the neat Si-PCPDTBT and hybrid Si-
PCPDTBT:nanocrystal BHJ films with NH 4I- and PbI 2-exchanged PbS acceptors,
respectively. PL lifetimes were measured by detecting the emission at 700 nm for the
neat polymer and hybrid films. The film samples were not moved for the duration of
the experiment and remained static with respect to the excitation beam. TCSPC
measurements were performed using a R3809U-50 Hamamatsu PMT with a B&H SPC-
630 module (time resolution of 22 ps). The monochromator grating was blazed at 600
nm with 1200 g mm
–1
and a slit width of 1.2 mm was used, giving a spectral bandpass
of 4.17 nm. The lifetime measurements were limited by the response of the detector
(22 ps), which was longer than the pulse width of the excitation beam from the OPA
(<200 fs).
46
3.2.4 Transient absorption studies
Broadband femtosecond TA experiments were carried out using the output of a
Coherent Legend Ti:sapphire amplifier (1 kHz, 3.5 mJ, 35 fs). The 900- 1200 nm
supercontinuum probe pulses were generated in a sapphire window driven by the 800
nm amplifier fundamental, polarized perpendicular to the pump pulses and detected
with an InGaAs Hamamatsu G9213-256S photodiode array after being dispersed by
an Oriel MS1271 spectrograph. The 920 nm pump was generated by doubling 1840
nm generated using ~10% of the amplifier’s output to seed a type II collinear OPA
(Spectra Physics OPA-800C). To minimize probe dispersion, a pair of off-axis
aluminum parabolic mirrors was used to collimate the probe and focus it into the
sample, while a CaF 2 lens focused the pump. Samples were translated perpendicular
to the path of the pump and probe to prevent photodamage. TA spectra were
measured with a pump energy 0.14 μJ and a spot size of 120 μm resulting in a fluence
of 1300 μJ cm
–2
. Spectra were acquired by averaging each time slice over 500 laser
shots, and delay traces were scanned 5 times.
3.2.5 Chemical doping studies
In order to characterize the steady state polaron spectrum of Si-PCPDTBT, thin
films were spun cast on quartz slides from a 2 mg mL
–1
solution in 1,2-
dichlorobenzene. Si-PCPDTBT films were subsequently oxidized by dipping them in a
47
20 ppm solution of SbCl 5 in acetonitrile.
101
The state-state absorption spectra were
acquired on a Perkin-Elmer Lamba 950 spectrophotometer equipped with a 150 mm
integrating sphere to account for scattering and reflections. The difference spectra
were acquired by subtracting the spectrum of the oxidized polymer from that of the
neat polymer.
3.2.6 Hybrid solar cell fabrication
All devices were fabricated in air and tested under nitrogen atmosphere.
Aluminum shot was purchased and used as received. Patterned ITO-coated glass
substrates were sequentially cleaned by sonication in tetrachloroethylene, acetone,
and isopropanol, followed by 20 min of UV-ozone treatment. A 35 nm hole
transporting layer of PEDOT:PSS was spun-cast onto the clean ITO and heated at 120
˚C for 30 min under vacuum. A Si-PCPDTBT solution of 15 mg mL
−1
was prepared in
1,2-dichlorobenzene by dissolution under mild heating for 1 h, followed by filtering
through a 0.45 m PTFE syringe filter. The ligand-exchanged PbS nanocrystals (in 1,2-
dichlorobenzene and BA) were then mixed for 1 h at room temperature with the
filtered Si-PCPDTBT solution to a final concentration of 2:30 mg mL
–1
(Si-
PCPDTBT:PbS). This solution was spun cast onto the dried PEDOT:PSS layer to get a
48 nm thick active layer, and it was aged in the dark under flowing nitrogen for 20-25
min. For the EDT-exchanged BHJ as control devices, the BHJ was then dipped into a
1% (vol/vol) solution of EDT in dry acetonitrile for 1 min, followed by washing with dry
acetonitrile twice to remove residue organic ligands. ZnO nanocrystals (synthesized
48
from a sol-gel method with diameters of 3-4 nm dispersed in ethanol (20 mg mL
−1
)
were spun cast on the active layer to produce a 40 nm thick electron transport layer.
The devices were then annealed at 130 ˚C under flowing nitrogen for 10 min, followed
by loading into a high vacuum (~2 Torr) thermal deposition for deposition of 100-nm
thick Al cathodes.
3.3 RESULTS AND DISCUSSION
3.3.1 Colloidal Iodide Ligand Exchange
It has been previously shown that PbS nanocrystals with a 1.3 eV band gap result
in the best charge separation efficiency in Si-PCPDTBT:PbS nanocrystal BHJs.
124
PbS
nanocrystals with an average diameter of 3.2 nm and a band gap of 1.3 eV were
synthesized from a modified hot-injection approach.
129
The as-synthesized PbS
nanocrystals are initially capped with oleate native ligands, which provide excellent
colloidal stability in nonpolar organic solvents.
For ligand exchange, the inorganic ligands are dissolved in methanol (for NH 4I), or
a mixture of methanol and DMF (1:2 vol/vol for PbI 2), followed by addition of a PbS
Figure 3.1. TEM image of PbS nanocrystals (left) schematic of the ligand exchnage
process (center), and the polymer used in the bulk heterjunctions Si-PCPDTBT
(right).
49
nanocrystal suspension. The integrity of the PbS nanocrystals is retained after this
iodide ligand exchange, as confirmed by solution UV-vis-NIR spectroscopy (Figure
3.2). The absorption spectra of the iodide-exchanged PbS nanocrystals show the
persistence of sharp excitonic features, albeit slightly red-shifted (~37 meV) when
compared to as-prepared PbS nanocrystals with oleate ligands.
130
This slight red
shift may be attributed to a change in surface dipole moments or solvent dielectric
constants.
127
Thermogravimetric analysis (TGA) data was used to measure the mass
loss upon heating from room temperature to 400 C under flowing nitrogen. TGA
provides information on the total mass of ligands through their
decomposition/volatilization temperature relative to that of the inorganic nanocrystal
core.
131
It is observed that after ligand exchange, the amount of organic content is
reduced. This is evidenced by a lower mass loss percentage at 400 °C and is indicative
of the efficient displacement of long-chain oleate native ligands during the ligand
Figure 3.2. Steady state photophysical properties of PbS quantum dots and Si-
PCPDTBT polymer.
50
exchange process.
64, 132, 133
The organic ligand content decreases from 30% for oleate-
capped PbS nanocrystals to 4.5 and 3.6% after NH 4I and PbI 2 ligand exchange,
respectively. Additionally, the hypsochromic shift of excitonic features in the UV-vis-
NIR absorption spectrum also suggests that ligand exchange with BA results in
significant etching of the PbS nanocrystal surface.
134
3.3.2 Hybrid Solar Cells
A Si-PCPDTBT:PbS nanocrystal BHJ absorber layer is expected to possess a type-
II band alignment between the donor and acceptor phases (Figure 3.3).
135
This active
layer was then incorporated into hybrid solar cells with the general device structure
ITO/PEDOT:PSS (35 nm)/Si-PCPDTBT:PbS nanocrystal BHJ/ZnO (40 nm)/Al (100 nm)
via spin-casting blends of Si-PCPDTBT with PbS nanocrystals dispersed in 1,2-
dichlorobenzene that were filtered through a 0.45 µm filter. The devices were
Figure 3.3. Schematic illustration of hybrid solar cells based on
ITO/PEDOT:PSS/Si-PCPDTBT:PbS nanocrystal/ZnO/Al device structures (left
panel) and state diagram representing the excitation and charge transfer
processes in the bulk heterojunctions.
PL-QD
PL *-QD
PL-QD*
PL
·+
-QD
·-
λ
ex
λ
ex
51
optimized with respect to the polymer/nanocrystal loading ratio, followed by the
active layer thickness, and finally the iodide ligand concentration used during ligand
exchange.
136
It was determined that the Si-PCPDTBT:PbS nanocrystal BHJ that
performs best has a loading ratio of 30:2 mg mL
–1
Si-PCPDTBT:PbS, an active layer
thickness of 48 nm, and a PbI 2 concentration of 0.02 M (18 mL) for exchanging 100
mg of PbS nanocrystals in 2 mL of toluene.
The highest performing hybrid solar cells under optimized conditions come from
PbI 2-capped PbS nanocrystals. The champion device gave J SC = 18.2 mA cm
–2
, V OC =
0.48 V, FF = 0.55, and PCE = 4.78%. This champion device performance is among the
best for hybrid solar cells to date. Average device parameters were calculated from
up to 16 devices over four separate substrates for hybrid Si-PCPDTBT:PbS
nanocrystal BHJ solar cells where the acceptor had been ligand exchanged with PbI 2,
giving J SC = 16.6 (1.6) mA cm
–2
, V OC = 0.48 (0.01) V, FF = 0.50 (0.04), and PCE = 3.98
(0.80)%. Strong photo-response from both Si-PCPDTBT and the PbI 2-exchanged PbS
nanocrystals can be observed in the external quantum efficiency (EQE) and internal
quantum efficiency (IQE) spectra,
111
both of which reveal a broad spectral response
from the UV into the NIR (i.e., 300−1100 nm). This BHJ combination gives a peak EQE
(at 650 nm) of 60%, with a NIR EQE (at 980 nm) of 10% originating from the PbS
nanocrystal acceptors. In comparison, hybrid solar cells based on BHJs of Si-
PCPDTBT:PbS nanocrystals ligand exchanged with NH 4I gave considerably lower
photocurrent density and fill factor (FF), with average device parameters of J SC = 6.4
52
(0.9) mA cm
–2
, V OC = 0.43 (0.04) V, FF = 0.31 (0.01), and PCE = 0.87 (0.18)%. In this
device, a lower peak EQE (at 500 nm) of 46% was observed, in addition to a much
weaker EQE response of 2% in the NIR at 980 nm. To further investigate the origin of
the performance differences between these two iodide ligand treatments, time-
resolved PL and ultrafast TA spectroscopic studies were conducted.
3.3.3 Charge Separation Dynamics at the Hybrid Interface
To probe the dynamics of charge transfer or energy transfer between the Si-
PCPDTBT donor and the PbS nanocrystal acceptors that have been ligand exchanged
with PbI 2 and NH 4I, time-resolved PL studies were performed. Time-resolved PL
spectra were collected for a neat Si-PCPDTBT thin film and Si-PCPDTBT:nanocrystal
Figure 3.4. Time-resolved PL traces of the pristine Si-PCPDTBT polymer and Si-
PCPDTBT:nanocrystal BHJs employing NH 4I- and PbI 2-exchanged PbS acceptors
( λ ex = 500 nm; λ em = 700 nm). The instrument response function (IRF) is shown in
black.
53
BHJ thin films containing PbI 2- and NH 4I-exchanged PbS (and prepared under
identical processing conditions to the working devices). A significant reduction of PL
lifetime for Si-PCPDTBT at 700 nm is observed for hybrid BHJs containing both NH 4I-
and PbI 2-exchanged PbS acceptors when the films were excited at 500 nm (Figure
3.4). At this excitation wavelength both the polymer and nanocrystals are excited.
Energy transfer from the excited state of the polymer can lead to an excited state of
the nanocrystals, while electron transfer from the excited polymer to nanocrystals
can result in a charge separated state. The same charge separated state can also be
formed by hole transfer from the excited nanocrystal to the polymer. The measured
PL lifetime of Si-PCPDTBT ( = 1.25 ns) is longer than a previous literature report of
250 ps.
121
The extended PL lifetime observed here might result from extra precautions
taken to avoid exposure of the polymer thin film to air. For the hybrid BHJ films, a bi-
exponential decay dynamic is observed for the Si-PCPDTBT:nanocrystal BHJ with
NH 4I-exchanged PbS (Figure 3.4). The lifetimes obtained by fitting the decay traces
to a bi-exponential decay function were 0.387 ns (9%) and 0.026 ns (91%), while the
excited state for the Si-PCPDTBT:nanocrystal BHJ with PbI 2-exchange PbS decays
faster than our instrument response function (<22 ps). These results suggest that the
PbI 2 ligand exchange enables a more efficient charge transfer or energy transfer than
NH 4I exchange based on time-resolved PL measurements, which correlates with J SC
measured in the hybrid devices.
54
Ultrafast TA spectroscopy was used in a complementary way to gain deeper insight
into the charge transfer dynamics. Previous TA studies for hybrid polymer:PbS
systems were based on excitation at wavelengths where both the polymer and PbS
nanocrystals possess considerable absorbance.
14, 121
As a result, both energy
transfer and charge transfer between the polymer and PbS nanocrystals can occur
simultaneously, as in our time-resolved PL quenching experiments, thereby
complicating the interpretation of the spectral dynamics. In our TA measurements,
the PbS nanocrystals were selectively excited by pumping at 920 nm, where Si-
PCPDTBT does not absorb, and the polaron dynamics of Si-PCPDTBT were probed at
1200 nm. This wavelength was chosen based on an assignment to a delocalized
polymer hole polaron at 1200 nm for its carbon-bridged analogue C-PCPDTBT in a
previous literature report.
137
To confirm this assignment for the particular Si-
PCPDTBT polymer used in this study, we carried out chemical doping experiments by
Figure 3.5 Transient absorption spectra of (a) PbS(NH 4I):Si-PCPDTBT and (b)
PbS(PbI 2):Si-PCPDTBT film excited at 920 nm and probed in the 900-1200 nm
range.
55
oxidizing the polymer with SbCl 5.
138
Chemical treatment of Si-PCPDTBT with SbCl 5
results in oxidation of the polymer, similar to the charge separated state we expect to
observe in the TA experiments after hole transfer from the PbS nanocrystals to the
polymer upon excitation at 920 nm. The steady-state absorption spectra of the
chemically oxidized Si-PCPDTBT displays a reduction in intensity of the 400 and 660
nm bands, concomitant with the appearance of a broad positive band around 1200 nm
(Figure 3.8), which is very similar to what has been previously observed for C-
PCPDTBT.
138
The steady-state absorption spectrum of the chemically oxidized
polymer does not exactly overlap in the NIR range with TA spectra obtained for the
neat Si-PCPDTBT polymer (Figure 3.8) and the hybrid BHJ films because of (i) a
(positive) photoinduced absorption at 1400 nm for the singlet exciton of the polymer
also contributes to the TA spectra of the neat polymer and, (ii) to the spectra of hybrid
Figure 3.6. Femtosecond TA spectra of three films as labeled. Hybrid BHJ films
were pumped at 920 nm and probed at 1200 nm; the neat Si-PCPDTBT film was
pumped at 660 nm and probed at 1200 nm.
56
films, a bleach at 920 nm from the PbS nanocrystals, and, specifically for NH 4I-
exchanged nanocrystals, a strong stimulated emission band at 1100 nm are clearly
contributing negative-going signals (Figure 3.7).
138
The hole polaron does contribute
to the positive 1200 nm signal of the TA spectra at the 0.1 ps decay time scale (Figure
3.5) based on the overlapping spectral feature from chemical oxidation. Therefore,
we probed the positive band at 1200 nm to gain insight into the dynamics of the
delocalized polymer hole polaron. The formation of polymer hole polarons is direct
evidence of a hole transfer process from the ligand-exchanged PbS nanocrystals to
Si-PCPDTBT, without any interference from energy transfer.
Figure 3.5 shows the transient absorption spectra in the near-infrared for the bulk
heterojunction films when excited at 920 nm and figure 3.6 shows the evolution of the
band at 1200 nm with pump-probe delay for the neat polymer film (pumped at 660
nm), and for the hybrid Si-PCPDTBT:nanocrystal BHJ films. In both cases, the polaron
signal appears within the instrument response ( ≤ 120 fs), suggesting good coupling
between the PbS nanocrystals and the Si-PCPDTBT polymer for either ligand
treatment.
111
While the 1200 nm induced absorption in the neat polymer shows a bi-
exponential decay with lifetimes of = 0.6 ps and 53 ps, the induced absorption
decays much faster in the Si-PCPDTBT:nanocrystal BHJ with NH 4I-exchanged PbS
acceptors, but persists for a considerably longer time in the BHJ film with PbI 2-
exchanged PbS acceptors. The 1200 nm kinetics in the hybrid BHJ films are clearly
affected by the presence of bleach from the PbS nanocrystals as well as the induced
57
absorption from the polaron in the same spectral window (Figure 3.8, 3.10). In the
case of the NH 4I ligand exchange, a stimulated emission signal is also clearly
observed in the complex TA trace for the hybrid BHJ film. This suggests that a fraction
of the NH 4I-exchanged PbS nanocrystals do not charge separate in the hybrid BHJ
film, perhaps because of large domains of the NH 4I-exchanged PbS acceptors (vide
infra) or poor coupling; however, the fast polaron rise time seen for the NH 4I-
exchanged nanocrystals that do charge separate suggest a heterogeneous model. A
detailed kinetic analysis of the various signals seen in the broadband TA experiment
using global analysis is underway and will be the subject of future work. However, it
is evident from Figure 3.9 that the lifetime of the polaron is much longer in hybrid BHJ
films with PbI 2-exchanged PbS nanocrystal acceptors as compared to the neat
Figure 3.7 Transient absorption spectra (at 4 ps, pumped at 920 nm) overlapping
with the absorption and emission spectra of the PbS nanocrystals. (a) Hybrid Si-
PCPDTBT:nanocrystal BHJ film with PbI2-exchanged PbS, and (b) hybrid Si-
PCPDTBT:nanocrystal BHJ film with NH4I-exchanged PbS.
58
polymer film, and clearly different from the hybrid BHJ films with NH 4I-exchanged
PbS nanocrystal acceptors. Such a long-lived hole polaron on the polymer in hybrid
BHJ films with PbI 2-exchanged PbS nanocrystal acceptors would thus allow the holes
to better percolate toward the electrode.
The differences in the lifetimes of the polaron in hybrid BHJs with NH 4I- and PbI 2-
exchanged PbS nanocrystals may be attributed to secondary hole transfer processes
(e.g., nanocrystal surface traps or reverse hole transfer from polaron to nanocrystals).
To test these hypotheses, PL spectra of PbS nanocrystal suspensions and the
morphologies of the hybrid BHJ films were investigated. It was determined that the
concentration-normalized PL intensity of the NH 4I- and PbI 2-exchanged PbS
nanocrystal suspensions were both lower than the as-prepared PbS nanocrystals
with oleate ligands, suggesting that both iodide exchanges result in more surface trap
states than the oleate-passivated PbS nanocrystals.
139
Such induced surface trap
states might account for the relatively low V OC in hybrid devices. Additionally, both
the NH 4I- and PbI 2-exchanged PbS nanocrystal suspensions exhibit similar
concentration normalized PL intensities, suggesting qualitatively similar densities of
surface trap states. Therefore, we cannot specifically ascribe the significant
differences in excited state lifetimes to surface trap states on the nanocrystal
acceptors. A considerable difference in the BHJ morphology between the two ligand
treatments is observed, however, by atomic force microscopy (AFM) and transmission
electron microscopy (TEM). The hybrid Si-PCPDTBT:nanocrystal BHJ with NH 4I-
59
exchanged PbS possesses a rough film surface (rms roughness of 72 nm) with a large
degree of nanocrystal aggregates, while that with PbI 2-exchanged PbS exhibits a
much smoother film morphology (rms roughness of 16 nm). Such morphology
variation can also be observed in the corresponding TEM images. The nanocrystal
aggregates observed in the hybrid BHJs with NH 4I-exchanged PbS nanocrystals might
provide a pathway for secondary reverse hole transfer from polaron to nanocrystals,
14
leading to a lower lifetime of the polaron, and/or result in a failure to charge separate.
These processes can explain the enhanced stimulated emission observed in the TA
spectra from the NH 4I-exchanged PbS nanocrystals in the hybrid film (Figure 3.10),
which is caused by an enhanced excited state population of the nanocrystals
compared to PbI 2-exchanged nanocrystals. The tendency of NH 4I-exchanged PbS
Figure 3.8 (a) Steady-state absorption spectra of the Si-PCPDTBT thin film before
doping (black) and after a 3 min dip in a 20 ppm solution of SbCl 5 in acetonitrile
(red), shown together with the difference spectrum ( A = Adoped − Aundoped,
given in blue). (b) Transient absorption spectra of neat Si-PCPDTBT film (at 0.1 ps,
pumped at 660 nm) overlapping with chemically doped steady-state spectrum of
the oxidized Si-PCPDTBT film. Steady-state absorption spectra were taken in an
integrating sphere and thus account for scattering and reflections.
60
nanocrystals to aggregate in the hybrid BHJ films is also consistent with their
behavior in the solution phase, where NH 4I-exchanged PbS nanocrystals aggregate
much more easily than the PbI 2-exchanged nanocrystals. These observations are
consistent with the EQE spectra of the hybrid BHJ devices in the NIR region, where no
distinct NIR exciton peak from the PbS nanocrystal acceptors is generally observed,
except for the hybrid Si-PCPDTBT:nanocrystal BHJ with PbI 2-exchanged PbS
nanocrystals. Therefore, nanocrystal aggregation in the hybrid Si-
PCPDTBT:nanocrystal BHJ with NH 4I-exchanged PbS may account for the poor charge
separation dynamics.
3.4 CONCLUSIONS
Hybrid solar cells based on BHJs of Si-PCPDTBT and PbS nanocrystals were
fabricated from the direct solution deposition of polymer and nanocrystal mixtures
without further solid-state ligand exchange. Si-PCPDTBT:nanocrystal BHJ devices
with PbI 2-exchanged PbS acceptors achieved a PCE of 4.8%; on the other hand, NH 4I-
exchanged PbS nanocrystals exhibit significantly lower photovoltaic performance
when blended with Si-PCPDTBT under the same conditions. Time-resolved PL
spectroscopy indicates a more efficient energy transfer or charge transfer process
occurs when PbI 2-exchanged PbS nanocrystals are blended with Si-PCPDTBT, which
is consistent with the measured short circuit current density and integrated EQE for
these devices. To further elucidate the exciton dynamics at the hybrid interface, we
were able to selectively probe the hole transfer dynamics from the PbS nanocrystals
61
to Si-PCPDTBT by ultrafast TA spectroscopy via the selective excitation of the
nanocrystal phase. A much longer hole polaron lifetime is qualitatively observed in
the hybrid Si-PCPDTBT:nanocrystal BHJ with PbI 2-exchanged PbS acceptors. This
taken together with excellent colloidal and air stability, facile one-step device
fabrication, and superior photovoltaic performance render the PbI 2 ligand system
attractive for future hybrid solar cells applications with lead chalcogenide acceptors.
62
Postscript 3.A
Reevaluation of Transient Absorption Spectra of
PbS:Si-PCPDTBT Hybrid Bulk Heterojunction
Films
3.A.1. INTRODUCTION
Chapter 3 correlated the photoconversion efficiencies (PCE) of bulk
heterojunctions made of lead sulfide nanoparticles with PbI 2 and NH 4I surface ligands
with the organic polymer Si-PCPDTBT. The bulk heterojunction PbS(PbI 2):Si-PCPDTBT
with a power conversion efficiency (PCE) of 4.8% exhibited shorter polymer singlet
excited state lifetime than PbS(NH 4I):Si-PCPDTBT with PCE 0.87%. The lifetime of the
polaron state as observed in the near-infrared using transient absorption
spectroscopy (with the PbS excited at 900 nm) was longer for PbS(PbI 2):Si-PCPDTBT
than PbS(NH 4I):Si-PCPDTBT – implying longer lived charged separated states and
therefore higher power conversion efficiencies. This addendum discusses the
transient absorption signals from Si-PCPDTBT, PbS(PbI 2):Si-PCPDTBT and
PbS(NH 4I):Si-PCPDTBT excited at 660 nm and probed in the visible region from 400-
700 nm and in the near-infrared region from 900-1200 nm.
63
Previous work from our group on the related CdSe:PCPDTBT hybrids established
the yield of charge transfer η CT ~ 82 % and the charge separation efficiency η CS ~ 30
%. The charge transfer and charge separation efficiencies showed that the limitation
of such hybrid systems is a rapid and measurable geminate recombination due to only
a small degree of separation of the initial electron hole pair. Although the favorable
energy level alignment between the CdSe quantum dot and the PCPDTBT polymer in
that hybrid helped isolate electron transfer from other excited state processes, the
effect of changing the quantum dot surface ligand on the coupling with the polymer
and hence the effect on charge transfer, charge separation efficiencies and on the
separation of the initial charge pair has not been explored. In this postscript, the
charge transfer band in the transient absorption spectra of the PbS:Si-PCPDTBT
hybrid is used to calculate the charge transfer and charge separation efficiencies and
compare these between hybrids with different quantum dot surface ligands. This
provides insight into the role that the surface ligand plays in the separation of the
initial charge pair and hence on the recombination rates in this hybrid bulk
heterojunction.
The spectral features for the neat polymer and the bulk heterojunctions with PbI 2
and NH 4I ligand treatments are compared in Figure 3.A.1 for different time delays
between the pump and the probe. The spectra of the neat polymer Si-PCPDTBT film
serves as a control experiment to disentangle the spectra of the transient species in
the hybrid films. The transient spectra of all samples were collected with a pump
64
excitation fluence of 150 µJ/cm
2
, the data was recorded in the same experimental run.
At 660 nm, the optical densities of Si-PCPDTBT, PbS(NH 4I):Si-PCPDTBT and
PbS(PbI 2):Si-PCPDTBT were ~ 0.1. The fluence of 150 µJ/cm
2
corresponds to an
incident pump photon density of 6.041 x 10
14
absorbed 660 nm photons per cm
2
. The
absorbance at this excitation wavelength arises 77% from the polymer component of
the bulk heterojunction and the remnant 23% from the quantum dot. All other
experimental parameters were the same as those reported in Chapter 3. The
dynamics of the different excited state populations are extracted and summarized in
Figure 3.A.2.
3.A.2. RESULTS AND DISCUSSION
In the transient absorption spectra of the neat polymer film (Figure 3.A.1a), the
photoinduced absorption signal seen in the 1150 nm – 1200 nm region (red in the
figure) has been previously associated with dissociated/delocalized polarons
(labelled Polaron I).
140, 141
The positive features in the 965 nm – 1015 nm range (blue
region in the figure) have been observed in the absorption spectra of chemically
oxidized Si-PCPDTBT
142, 143
and therefore attributed to polarons as well. In the TA
measurements this feature shows a significant long-lived component in the neat
polymer (Figure 3.A.2) that distinguishes it from the Polaron I band. This band may
arise from trapped/localized polarons
144, 145
caused by morphological defects and has
been labelled Polarons II. The bleach observed in the 395 nm – 455 nm and 620 nm –
65
690 nm ranges clearly reflect the absorption spectrum of the polymer and
corresponds to depopulation of the ground state species upon excitation by the 660
nm pump.
The striking difference in the transient spectra of the hybrid films is the presence
of the positive feature in the 1050 nm -1110 nm region that is absent in spectra of the
neat polymer. This band therefore is associated with the polymer in association with
the quantum dot in the blend. Previous work from our group on CdSe:PCPDTBT hybrid
films
74
have also observed similar bands and assigned these to charge transfer states
in the polymer-quantum dot blend. This band could arise from electronic transitions
in the reduced PbS quantum dot after electron transfer (refer to Figure 3.A.3) from the
polymer or hole transfer to the polymer following excitation energy transfer, the band
could also originate from transitions similar to charge transfer (CT) states that involve
positive polarons on the polymer chains and electrons on the nanoparticle.
146, 147
The
excited state of the PbS quantum dot is not known to have transient absorption in this
region.
148, 149
Like the neat polymer film, the transient absorption spectra of hybrid films show
clear signs of Polarons I at 1150 nm -1200 nm that corresponds to delocalized
Polarons I. There is a 2x increase in this signal from the hybrid films compared to the
neat polymer (Figure 3.A.2) that can be attributed to the increase in the number of
Polarons I on the polymer in the presence of the nanoparticle. The PbS provides an
66
additional channel to the formation of positive charges on the polymer chain thus
leading to an increase in the number of polarons and therefore an increase in the
positive signal at 1150 nm – 1200 nm. Comparing the two hybrid bulk heterojunctions
with PbI 2 and NH 4I ligands on the nanoparticle surface, the PbS(PbI 2):Si-
PCPDTBT hybrid has more intensity in this band than PbS(NH 4I):Si-PCPDTBT. The
difference in the intensity of the bands corresponding to Polarons I can be used to
measure the difference in the charge transfer efficiencies between the two hybrid
films. The PbI 2 ligand has been previously shown to provide better electronic coupling
and hence more charge transfer in the hybrid films than the NH 4I ligand. The decay
dynamics of the Polaron I band (Figure 3.A.2) can be fit with two exponentials with the
faster of the two components being similar between the polymer, PbI 2 hybrid and NH 4I
hybrid (0.64 ps (26 %); 0.54 ps (33 %) and 0.66 ps (41 %)). The second component is
faster in the PbI 2 and NH 4I hybrid films (6.8 ps (43 %) and 8.9 ps (41 %)) when compared
to the neat polymer with 45 ps (16 %). The faster lifetime of the second component
can be due to the recombination channels in the hybrid that are absent in the neat
polymer.
In both hybrid films, the positive spectral signatures corresponding to Polarons II
has been reduced by a bleach from the depopulation of the ground state of the PbS
nanoparticles. This depopulation occurs either via excitation energy transfer from the
singlet excited state of the Si-PCPDTBT polymer to PbS nanoparticle which would
result in the nanoparticle in the excited state (Figure 3.A.3) or via electron transfer
67
from the excited polymer to the quantum dot. The extent to which the positive feature
of Polarons II at 950 nm has been extinguished by the negative bleach signal from the
nanoparticle is indicative of the efficiency of excitation energy transfer and the
electron transfer from the polymer to the nanoparticle. From comparing the transient
spectra at 1 ps delay in Figures 3.A.1b and 3.A.1c it is evident that the in the
PbS(NH 4I):Si-PCPDTBT the positive feature at 950 nm has been reduced to 0.5 mOD
from 3 mOD in the neat polymer while in PbS(PbI 2):Si-PCPDTBT this value is
significantly lower at – 2 mOD at 1 ps. Concomitant with the observations in Chapter
3, it can be concluded that PbI 2 ligands provide efficient electronic coupling and
therefore, efficient energy and electron transfer between the Si-PCPDTBT polymer
and the PbS nanoparticle, in contrast to the NH 4I ligand. Figure 3.A.2 also compares
the difference in the intensity of the bleach signal at 395–455 nm between the neat
polymer and the two hybrid films. There is a 6x enhancement in the bleach signal
corresponding to the polymer in the hybrids when compared to the neat polymer
films. This is indicative of the depopulation of the polymer ground state as a result of
the formation of charge transfer states in the hybrid films. The slower bleach recovery
rates in the hybrids (Table 3.A.1) is also indicative of the formation of the charge
transfer states in the hybrid films that is absent in the neat polymer.
Comparing the intensities of the CT band in the PbI 2 and NH 4I hybrid films can be
used to quantify the extent of charge transfer in the hybrid films. As seen in Figure
3.A.2, there is a 3.2x enhancement in the signal at 1150– 1110 nm in the hybrid films
68
compared to the neat polymer film. The formation of this new band seen only in the
hybrid films and could arise from either electrons on the quantum dot as in a charge
separated state or transitions in a charge transfer state involving the acceptor
quantum dot and donor polymer. The excited state polymer can undergo excitation
energy transfer (followed by hole transfer from the nanoparticle) or electron transfer
to the nanoparticle (Figure 3.A.3) resulting in the formation of positive charges on the
polymer chain and negative charges on the nanoparticle. To explore the charge
transfer and recombination dynamics in the hybrid films, the contributions to the
transient absorption signal at 1150 nm – 1110 nm from the polymer in the hybrid film
was subtracted out and the resulting spectra is shown in Figure 3.A.4. Essentially
working under the assumption that the spectral signature of the hybrid film can be
decomposed as a linear combination of contributions from the polymer component
and from of the quantum dot component:
∆𝐴 𝐻𝑦𝑏𝑟𝑖𝑑 = 𝑎 × ∆𝐴 𝑝𝑜𝑙𝑦𝑚𝑒𝑟 + 𝑏 × ∆𝐴 𝑞𝑢𝑎𝑛𝑡𝑢𝑚 𝑑𝑜 𝑡
As seen in the Figure 3.A.4, taking the contribution of the polymer away from the
transient spectra of the hybrids results in the negative spectral feature in the 950 nm
– 1015 nm range. This feature, which is the contribution of the quantum dot in the
hybrid films looks similar to the absorption spectra of the hybrid and can be attributed
to ground state bleaching. The signal from the CT band at 1075 nm can be used as a
proxy for the number of electrons on the quantum dot because the contributions from
69
Figure 3.A.1 Transient absorption spectra of (a) neat Si-PCPDTBT, (b) PbS(NH 4I):Si-
PCPDTBT, and (c) PbS(PbI 2):Si-PCPDTBT films excited at 660 nm.
70
the polymer have been subtracted out. Figure 3.A.5 shows the evolution of the
transient signal at 1075 nm in both the PbS(PbI 2):Si-PCPDTBT and PbS(NH 4I):Si-
PCPDTBT hybrid films. The transient spectra of all samples were collected with a
pump excitation fluence of 150 µJ/cm
2
, the data was recorded in the same
experimental run. At 660 nm, the optical densities of Si-PCPDTBT, PbS(NH 4I):Si-
PCPDTBT and PbS(PbI 2):Si-PCPDTBT were ~ 0.1. The decay traces are distinctly
bimodal in both the hybrids with a fast decay that occurs within 50 fs and a slower
decay that occurs over 100 ps. This indicates two distinct pathways for charge
Figure 3.A.2 Comparison of decay dynamics of the excited species in neat Si-
PCPDTBT, PbS(NH 4I):Si-PCPDTBT, and PbS(PbI 2):Si-PCPDTBT films excited at 660
nm. Time traces were obtained by averaging over the following wavelength ranges:
Polaron 2 (965 nm -1015 nm), Polaron 1 (1150 nm -1200 nm), CT state (1050 nm –
1110 nm) and Bleach (395 nm – 455 nm).
71
recombination – the prompt decay can be attributed to recombination that happens
at the interface of the quantum dot and polymer in < 50 fs (geminate recombination)
while the delayed response can be attributed to the charges migrating in the bulk
heterojunction film and the recombining with charges of the opposite nature that they
might run into (non-geminate recombination). As observed in CdSe:PCPDTBT
hybrids,
74
the fraction of dissociated excitons that remains after 1 ps (i.e. after fast
recombination) is called the charge separation efficiency. Normalizing the zero time
signal at 1075 nm and comparing to the fraction that remains after 1 ps, in the
PbS(PbI 2):Si-PCPDTBT hybrid there is 80% of the signal remaining while in the
PbS(NH 4I):Si-PCPDTBT hybrid there is only 50%. This difference in the intensity of the
CT band in the hybrids and hence in the number of electrons on the quantum dot holds
enormous importance to the charge carrier dynamics and surface chemistry of
Figure 3.A.3. Schematic illustration of hybrid solar cells based on
ITO/PEDOT:PSS/Si-PCPDTBT:PbS nanocrystal/ZnO/Al device structures (left
panel) and state diagram representing the excitation and charge transfer
processes in the bulk heterojunctions.
PL-QD
PL *-QD
PL-QD*
PL
·+
-QD
·-
λ
ex
λ
ex
72
polymer:quantum dot hybrid films. In PbS(PbI 2):Si-PCPDTBT, after the formation of
the exciton, the hole on the polymer and the electron on the quantum dot are
separated by an inorganic ligand layer, while in PbS(NH 4I):Si-PCPDTBT, the charge
particles are separated by an organic ligand. Since the dielectric constant of the
material separating the charges is higher and the Coulombic force is weaker than in
PbS(NH 4I):Si-PCPDTBT where the charge particles are separated by an organic layer
with lower dielectric constant. The resultant weaker attractive force in PbS(PbI 2):Si-
Table 3.A.1. The amplitudes and time constants obtained from fitting the
transient spectra with two exponentials.
A 1 τ 1 (ps) A 2 τ 2 (ps)
Ground State Bleach
Neat polymer 1.8 0.12 0.35 1.8
NH 4I hybrid 3.4 1.0 1.2 146
PbI 2 hybrid 1.7 2.7 1.5 72
Charge Transfer State
Neat polymer 2.7 0.32 1.1 6.1
NH 4I hybrid 7.9 0.54 2.3 6.9
PbI 2 hybrid 9.9 0.65 1.4 7.9
Polaron 1
Neat polymer 5.1 0.64 0.96 45
NH 4I hybrid 6.6 0.54 2.6 6.8
PbI 2 hybrid 8.1 0.66 2.5 8.9
73
PCPDTBT results in lower charge recombination as the hole on the polymer and the
Figure 3.A.4. The transient absorption spectra of the polymer and hybrid films at 1
ps are show with filled circles while the open circle traces are the transients of
hybrid films with the polymer signal subtracted out. The green trace is the
absorbance of the PbS(PbI 2):Si-PCPDTBT hybrid (x5).
Figure 3.A.5. The decay traces of the hybrid films at 1075 nm with the contribution
from the polymer subtracted away.
74
electron on the quantum dot can separate at the interface and migrate away. In
PbS(NH 4I):Si-PCPDTBT however, the stronger attractive forces result in more charge
recombination and lower efficiency of charge separation. The higher efficiency of
charge separation in PbS(PbI 2):Si-PCPDTBT is reflected in the higher power
conversion efficiency of the device made of the corresponding bulk heterojunction
device (PCE = 4.8%) compared to a device made with PbS(NH 4I):Si-PCPDTBT with PCE
= 0.89%.
3.A.2. CONCLUSIONS
This post script presents the transient absorption spectra of the neat polymer Si-
PCPDTBT, and those of the hybrids PbS(NH 4I):Si-PCPDTBT and PbS(PbI 2):Si-PCPDTBT
when excited at 660 nm. Spectral bands corresponding to the ground state bleaching
and two types of polarons (Polaron 1 and Polaron II) were clearly seen when the
polymer is excited, whereas in the transient spectra of both the hybrids, an additional
band corresponding to the charge transfer state is also present. The slower recovery
of the band corresponding to the ground state bleach in the hybrid films when
compared to the polymer indicates that the presence of intermediate excited state in
the hybrids that is absent in the polymer. The intensity of the bands corresponding to
Polaron 1 in the hybrid films is 2x the intensity of the same band in the neat polymer
film corresponding to an increase in the number of polarons on the polymer chain in
the presence of the nanoparticle. While the spectral signature corresponding to
75
Polaron II is masked in the hybrids by the signal from the ground state bleaching of
the quantum dots, it is possible to separate the contribution from the polymer to the
transient spectra of the hybrid films. Once the contribution of the quantum dots to
the hybrid transient spectra was obtained, comparing the dynamics at the charge
transfer band led to the conclusion that the charge separation efficiency in the
PbS(PbI 2):Si-PCPDTBT hybrid film is 80% while in the PbS(NH 4I):Si-PCPDTBT hybrid
film, the efficiency is only 50%. This drastic difference in the separation efficiency of
charges formed after photoexcitation is further reflected in the photoconversion
efficiencies of the devices made of the two hybrid films: the device made of
PbS(PbI 2):Si-PCPDTBT had a PCE of 4.8% while that made of PbS(NH 4I):Si-PCPDTBT
had a PCE of 0.89%.
76
CHAPTER 4
Linker Dependent Singlet Fission
in Tetracene Dimers
4.1 INTRODUCTION
Singlet fission is a process in which a singlet (S 1) exciton on a chromophore shares
energy with a ground state chromophore to result in a pair of triplet (T 1) excitons.
38, 150
Incorporation of such materials in organic solar cells could improve the power
conversion efficiencies by doubling the number of excitons generated by high energy
photons, which would otherwise be lost to thermalization to the band edge. The
Shockley-Queisser limit – the theoretical power conversion efficiency limit of 33% for
a single junction device – could be increased to 46% by combining an efficient singlet
fission material with a red absorber in a device.
70
Congreve et al. have demonstrated
an internal quantum efficiency of 160% in a pentacene/poly(3-hexylthiophene)
device
151
. The power conversion efficiencies of most singlet fission devices are
however sub-optimal, leaving much room for improvement in regard to molecular
design.
An important criterion to use singlet fission materials for solar energy is the fast
rate of fission, to ensure that charge injection occurs from the triplet state, rather the
singlet state - singlet fission rates faster than 100 ps are desirable.
152
The rate of
77
singlet fission is proportional to the coupling:⟨𝑆 1
𝑆 0
|𝐻 𝑒𝑙
| (𝑇 1
𝑇 1
)
1
⟩ between the [S 1S 0]
state and the biexciton state, often described as correlated triplet pair [
1
(T 1T 1)], and is
also dependent on the energy gap between these states.
153
In the early models by
Smith and Michl, the ⟨𝑆 1
𝑆 0
|𝐻 𝑒𝑙
| (𝑇 1
𝑇 1
)
1
⟩ is maximized in slip-stacked, π-overlapping
configurations of the chromophores;
38, 154
however, more sophisticated theoretical
treatments show that coupling can also be significant in sandwich configurations and
even between coplanar chromophores.
155, 156
The effect of this electronic coupling on singlet fission is best demonstrated using
covalently-coupled chromophores like molecular dimers and larger arrays.
27, 157-162
Examining the excited state dynamics of dimers in solution is advantageous because
it allows for the separation of the effects of coupling from the solid-state effects on
the rates of singlet fission. There has been enormous progress made toward the
synthesis and photophysical characterization of acene dimers and larger arrays in
recent years, as well as in covalently linked systems of other singlet fission
chromophores.
7, 43, 46, 163-174
Efficient singlet fission has been reported in dimers where
the constituent chromophores exhibit substantial π orbital overlap, or are linked
together in a conjugated manner whereas dimers in which the chromophores are
cross-conjugated exhibit slower rates of singlet fission. Guldi et al. have reported fast
and efficient singlet fission in ortho-, meta- and para-TIPS-pentacene dimers, in
which the favorable Energy(Singlet) ≥ 2 x Energy(Triplet) alignment provides the
enthalpic driving force for fission, such that even the weakly coupled meta-linked
78
pentacene dimers exhibited singlet fission on picosecond timescales.
175, 176
In dimers
of tetracene and diphenylisobenzofuran, in which the chromophores are without any
significant π orbital overlap or conjugation, and are out of plane with respect to each
other, singlet fission was reported to be slow and inefficient, likely due to weak
coupling and the Energy(Singlet) ≈ 2 x Energy(Triplet) energy alignment.
135, 177, 178
The currently accepted mechanistic picture of singlet fission involves an
intermediate biexciton state
179
, which can be described as a correlated triplet pair
[
1
(T 1T 1)], connecting the singlet excited state [S 0S 1] with the two independent triplet
states [
3
T 1] .
38
In previous work from our group, we reported fast and efficient
formation of the biexciton state in ortho-bis-ethynyltetracenylbenzene (o-BETB), in
which the two tetracene units are oriented in a twisted configuration with π-overlap
and appear to be strongly electronically coupled – i.e. there is good coupling between
Figure 4.1. Chemical (top) and space-filling structures of o-BETB, m-BETB, and
p-BETB. Two para-BETB dimers with different substituents on the linker were
synthesized (p-BETB-ohex: R = OC 6H 13 and p-BETB-ehex: R = 2-ethyl-hexyl).
79
the S 0S 1 and the
1
(T 1T 1) states. Although the π orbital overlap between the tetracenes
provided significant through-space coupling, ab initio calculations showed that the
conjugating, ortho-diethynylbenzene linker was also responsible for some portion of
the coupling (i.e. through-bond) in the excited state between the tetracenes in o-BETB
just as in the ground state.
37
This chapter contains the solution state photophysical properties and excited state
dynamics of the meta- and the para- analogs of ortho-bis-ethynyltetracenylbenzene
(m-BETB, p-BETB-ehex and p-BETB-ohex, Figure 4.1). The position of the alkyne
substitution on the benzene ring heavily influences the ground state electronic
structure of molecules containing these diethynylbenzene units.
177
The ortho- and the
para- arrangements can be thought of as conjugating and having through-bond
contributions to the coupling, while the meta- does not facilitate conjugation between
the tetracene units that are pendant on the alkynes and has only small through-space
coupling. The conjugation provided by these linkers can be understood in terms of the
cumulene-like resonance form of the ortho- and para-diethynylbenzenes, which
strongly resembles the excited state electronic structure in molecules containing
these structural units.
180
The para-diethynylbenzene particularly facilitates efficient
electronic communication between the chromophores it bridges; as a result,
compounds based on this chemical motif have been incorporated into ultrasensitive
sensors,
181
energy harvesting materials,
134
and molecular electronic devices.
182
The
meta-diethynylbenzene is cross-conjugating, and has been demonstrated to provide
80
only weak coupling between the chromophores tethered to it.
30, 183
Calculations of
tetracene-based dimers and model structures have shown that the linker contributes
significantly to the coupling, via through-bond interaction, which affects the
character of the states involved in singlet fission.
184-186
Theoretical work by Abraham
and Mayhall provides a model for determining the effect of the covalent linker on the
boundedness of the
1
(T 1T 1) state based on Ovchinnikov rule. They estimate that the
ortho-, and para-benzene dimers of ethynylpentacene, and ethynyltetracene produce
bound
1
(T 1T 1) states, while meta- coupled dimers should have un-bound
1
(T 1T 1)
states.
32
In contrast to Guldi et al.’s pentacene dimers in which the favorable
Energy(Singlet) ≥ 2 x Energy(Triplet) ensures fast singlet fission which was
unresolvable with the instrument response of their transient absorption
spectrometer in ortho-, and para- dimers, the isoenergetic Energy(Singlet) ≈ 2 x
Energy(Triplet) conditions in our bis-ethynyltetracenylbenzene dimers result in
singlet fission rates that are fast but are still resolvable with our instrument response.
In comparison to the tetracene dimers reported by Bardeen et al., in which the acenes
were directly attached to the phenyl linker and displayed slow singlet fission, the
tetracene units in our dimers are connected to the phenyl linker via ethynyl groups.
This ensures conjugation between the tetracenes (via the cumulene resonance
structure) and less steric hindrance within a dimer, which could result in faster singlet
fission.
135
The presence of ethynyl groups in our ethynyltetracene dimers also
81
perturbs the S 1 and T 1 state energies in a favorable way, as evidenced by faster fission
rates.
37
Using ultrafast spectroscopy, we observe that the
para-bis-ethynyltetracenylbenzene dimers (p-BETB-ohex: R = OC 6H 13 and
p-BETB-ehex: R = 2-ethyl-hexyl in Figure 4.1) do undergo singlet fission while no
significant singlet fission in the meta-bis-ethynyltetracenylbenzene dimer (m-BETB)
is observed. To understand the differences in the excited state behavior of the
ortho- vs. the para- dimers, we present the transient spectroscopy of p-BETB-ehex in
THF and p-BETB-ehex suspended in a rigid polymer matrix
(poly-(methylmethacrylate), PMMA) and model the kinetics using two populations
that exhibit distinct excited state dynamics. From modeling and calculating
electronic coupling as a function of the angle between the ethynyltetracenes, we
suggest that the ethynyltetracenes must be in a perpendicular orientation to
complete separation of the triplets (i.e., for productive singlet fission). To study the
Scheme 4.1. Synthetic scheme for meta, para-BETB-ohex and para-BETB-ehex.
82
role of the linker in facilitating singlet fission between the tetracenes, we have also
synthesized the para-bis-ethynyltetracenylbenzene dimers p-BETB-ethex and
p-BETB-ohex (Scheme 4.1) with different linker units and observe that singlet fission
is faster in p-BETB-ohex.
4.2 EXPERIMENTAL SECTION
4.2.1 Synthesis of meta- and para- dimers
The meta- and the para- dimers were synthesized as shown in Scheme 4.1. The
alkynyl motif in these dimers allows them to be constructed via the highly efficient
Sonogashira reaction.
187
m-BETB was constructed from 5-bromotetracene and the
commercially available meta-diethynylbenzene. The para-dimers were synthesized
by Sonogashira reaction between the appropriate diethynylbenzene linker unit and
5-bromotetracene. The synthetic details and characterization data have been
published elsewhere.
182, 188
4.2.2 Preparation of samples for spectroscopy
All steady state solution measurements were performed in 1 cm quartz cuvettes.
The solutions were prepared such that the optical density in the visible region was
approximately 0.1. All solutions were sparged with nitrogen gas for 5 minutes and
sealed prior to collecting emission, lifetime and quantum yield measurements. All
transient absorption solution measurements were performed in 1 mm quartz
cuvettes. De-aerated solutions with an optical density of 0.2-0.4 were loaded into the
1 mm cuvettes and sealed inside of the glove box with a nitrogen atmosphere. All thin
83
film samples were made on quartz substrates. The neat films were prepared by
dissolving 0.005 g of acene in 0.4 mL of freshly distilled THF, filtering the resulting
solutions through a 0.22 µm syringe filter, and spin casting the solutions onto quartz
substrates at 3500 rpm. The thickness of the films were between 60 nm and 100 nm,
with an optical density of 0.3 to 0.2 at 500 nm. Emission quantum yield measurements
made using optically matched solutions of acene sampled and standard dyes inside
of an integrating sphere using either 1 cm cuvettes or un-sealed thin films, under a
stream of nitrogen gas, excited at 480 nm. The PMMA samples were prepared by
dissolving 0.001 or fewer g of acene and ≥0.100 g of PMMA in 0.8 mL of toluene inside
of the N2 filled glove box. The vials containing the toluene suspensions were sealed
tightly and wrapped with parafilm inside of the glove box. The suspensions were then
sonicated for more than an hour. The resultant viscous solutions were filtered through
a 0.22 µm syringe filter and spun cast onto quartz substrates at 700 rpm under yellow
lights or under an inert atmosphere. All thin films were protected from photo-
oxidation by quartz window sealed with epoxy along the edges, inside the glove box
under a nitrogen atmosphere. Emission quantum yield measurements of thin films
made inside of an integrating sphere using un-sealed thin films, under a stream of
84
nitrogen gas. All transient absorption solution measurements were performed in 1
mm quartz cuvettes.
4.2.3 Femtosecond Broadband Transient Absorption
A Ti:sapphire regenerative amplifier (Coherent Legend, 1 kHz, 800 nm, 4 mJ, 35 fs)
was used to generate the pump and probe pulses. Excitation pulses were obtained by
pumping a type II-OPA (Spectra S16 Physics OPA-800C). White light supercontinuum
probe pulses (375-675 nm) were obtained by focusing the amplifier output onto a
rotating CaF2 disk (2 mm thick). Nanosecond transient absorption was performed
using the same continuum probe and pump pulses obtained from an Alphalas
Pulselas-A-266-300 nanosecond laser. The 532 nm harmonic from the nanosecond
laser was used to excite the meta-dimer. The polarization of the pump and probe
pulses were oriented at magic angle to avoid any contribution from orientation
dynamics to the transient signals. Off-axis parabolic mirrors were used to collimate
and focus the probe beam onto the sample while the pump beam was focused using
a CaF2 lens. The cross correlation of the pump and probe beams on a quartz cell
containing neat solvent alone had a FWHM of 180 fs. After passing through the
sample, the probe beam was dispersed using an Oriel MS1271 spectrograph onto a
256-pixel silicon diode array (Hamamatsu). All transient spectra were obtained using
85
pump fluences of 146 µJ/cm
2
. Samples were translated perpendicular to the path of
the pump and probe beams using a linear stage to prevent photodamage.
4.2.4 Time-correlated Single Photon Counting
Fluorescence lifetimes at of the dimers were determined by the time-correlated
single-photon counting (TCSPC) technique, using IBH Fluorocube with a 405 nm LED
excitation source, and an IRF value of 0.4 ns..
182, 189
4.2.5. Compartmental models and kinetic equations
To describe the transient absorption data from PMMA, populations 1 and 2 were
modelled using compartmental models.
190
Population 1 was described using the
kinetic equations:
𝑑𝑆 1
𝑑𝑡 = 𝑘 𝐴 𝑆 0
− 𝑘 12
𝑆 1
+ 𝑘 21
𝑇𝑇 − 𝑘 𝑟 𝑆 1
− 𝑘 𝑛𝑟
𝑆 1
𝑑𝑇𝑇 𝑑𝑡 = 𝑘 12
𝑆 1
− 𝑘 21
𝑇𝑇 − 𝑘 22
𝑇𝑇
Where S 1 is the population of the singlet excited state, S 0 that of the ground state
and TT that of the correlated triplet pair state. Population 2 was described using the
kinetic equations:
𝑑𝑆 1
𝑑𝑡 = 𝑘 𝐴 𝑆 0
− 𝑘 ′
12
𝑆 1
𝑑𝑇𝑇 𝑑𝑡 = 𝑘 ′
12
𝑆 0
− 𝑘 23
𝑇𝑇
𝑑𝑇 𝑑𝑡 = 𝑘 23
𝑇𝑇 − 𝑘 33
𝑇
4.3 RESULTS ANND DISCUSSION
86
4.3.1 Electrochemistry
The extent to which the ethynyltetracene units are conjugated in the ground state
in ortho-, meta- and para- bis-ethynyltetracenylbenzene dimers can be qualitatively
estimated from cyclic voltametry (CV). It is expected that the first oxidation will be
localized on one of the tetracenes in the dimer, and that the second oxidation will take
place on the second tetracene in the dimer. Conjugation between the
ethynyltetracenes in a dimer will give rise to a large difference between the first and
second oxidation potentials, since the two ethynyltetracene units are not acting
independently. Conversely, if the ethynyltetracene units do not interact, both units in
a dimer will oxidize at the same potential and the difference between the first and
second oxidation potentials would be close to zero. The cyclic voltamograms of the
ortho-, meta-, and para- bis-ethynyltetracenylbenzene dimers, as well as
phenylethynyltetracene (PET, monomer) were collected in dichloromethane, with
ferrocene as the internal standard. Two oxidation peaks were observed in all dimers;
however, the peaks were poorly resolved in m-BETB, due to their proximity.
191
The
oxidation potentials and the values of the splitting between the oxidation peaks are
given in Table 4.1. Based on the magnitude of the splitting of the oxidation potentials,
the chromophores that are the most strongly conjugated are o-BETB and
p-BETB-ohex, with (Ox 1-Ox 2) values of 0.22 V, m-BETB at 0.07 V is the least conjugated
and p-BETB-ethex is intermediate with a difference of 0.13 V between the two
oxidations. Furthermore, the electrochemical properties of the dimers provide
87
information about the energetics of intramolecular charge resonance (CR) states,
which have been deemed important in the singlet fission mechanism.
37, 192
The
approximate energy of the intramolecular CR state (E redox, Table 4.1) can be roughly
estimated from the difference between the oxidation (E ox) and the reduction (E red)
potentials and a Coulombic term. The values of E ox, E red, and the electrochemical
energy gap for the dimers and the monomer obtained by CV are given in Table 4.1. The
values of the first excited state (E gap,optical), estimated from the absorption spectra of
these dimers, are also supplied for comparison. In all molecules the electrochemical
gap is energetically comparable to the optical gap, suggesting that CR states lie close
enough in energy to the S 1 to potentially influence singlet fission kinetics, as
suggested in the case of the series of TIPS-pentacene dimers.
178, 193
Table 4.1. Electrochemical properties of the four dimers compared to the
monomer, and the optical gaps of these molecules. The ferrocenium/ferrocene
redox couple in dichloromethane was used as an internal standard for
electrochemical measurements.
molecule
First
oxidation
(Ox 1,V)
Second
oxidation
(Ox 2,V)
Reduction
(Red, V)
Ox 1-Ox 2
(V)
E redox
(eV)
E optical
(eV)
E optical -
E redox
(eV)
PET 0.46 --- -1.92 --- 2.38 2.42 0.04
o-BETB 0.37 0.59 -1.95 0.22 2.32 2.37 0.05
m-BETB 0.41 0.48 -1.89 0.07 2.30 2.40 0.10
p-BETB-
ehex
0.46 0.59 -1.92 0.13 2.38 2.34 -0.04
p-BETB-
ohex
0.33 0.55 -1.88 0.22 2.21 2.28 0.07
88
4.3.2 Steady state spectra
The intensities and line shapes of the steady state extinction spectra are
representative of the magnitude of the transition dipole moments in the molecules
and can reflect the extent of the S 1S 0- S 0S 1 excited state coupling between the
chromophores within the dimers. In our previous work, we showed that the
ethynyltetracene moieties in o-BETB are aligned such that the transition dipole
moments of the lower energy transition (along the short axis of the acene) are coupled
in an H-aggregate-like manner, as evidenced by the enhanced ν 0- 1 and diminished ν 0- 0
absorption compared to PET (Figure 4.2).
37
The absorption is redistributed to longer
wavelengths, but the integrated intensity in the visible region is less than twice that
Figure 4.2. Extinction spectra of m-BETB, p-BETB-ethex, and p-BETB-ohex
compared to o-BETB and PET in a THF solution.
89
of the monomer. The absorption spectrum of m-BETB is nearly identical to that of PET,
but with a small red-shift of 4 nm (0.02 eV). The relative intensities of the ν 0- 0 and the
ν 0- 1 vibronic features of the PET and the m-BETB are identical. The most notable
difference between PET and m-BETB absorption properties is in their molar
absorptivity spectra: the molar absorptivity of m-BETB is twice that of PET in the
visible part of the spectrum (the ratio of integrated areas is 2.0). This suggests that
tetracenes in m-BETB are effectively independent of each other and the electronic
coupling between them is weak.
The absorption properties of the p-BETB-ohex and p-BETB-ethex are different from
those of o-BETB, m-BETB and the monomer PET. The molar absorptivity of
p-BETB-ethex in the visible region is more than twice that of PET, with the area under
the curve greater than twice that of PET by a factor of 2.6). The red shift and significant
difference between the integrated absorptivity of the dimer versus twice that of the
monomer suggest strong exciton coupling between the two ethynyltetracene units in
p-BETB-ethex.
37, 194
The increase in molar absorptivity of p-BETB-ohex over PET is
even greater, with p-BETB-ohex having an absorptivity more than three times greater
than that of PET in the visible region (the ratio of the integrated areas of the
absorption bands is 4.3), and is accompanied by a substantial red-shift in the visible
absorption. This further demonstrates that the electron donating alkoxy substituents
on the bridging benzene ring significantly alter the electronic structure of the dimers
and in this case appear to enhance excitons coupling.
90
Analogously to o-BETB, the para-dimers are very weakly emissive ( Φ fl < 1%),
suggesting that the process which quenches the fluorescence in these systems is
orders of magnitude faster than ~10 ns (viz. the excited state lifetime of the monomer
PET).
191, 195
The analogous para-bisethynylanthracenyl-benzene dimers have been
reported to emit with high quantum efficiencies: 97% for
para-bisethynylanthracenyl-benzene,
195
and 60% for the analogous hexyloxy
substituted para-bisethynylanthracenyl-benzene. Similar to the comparison of
o-BETB to its anthracene analog, the high quantum efficiency of emission of the
anthracene dimers compared to the tetracene dimers suggests that a rapid excited
state decay channel is available in the para-tetracene dimers, which is not available
in the anthracene analogs. Conversely, m-BETB is strongly emissive in solution
(Figure 4.3), with a quantum yield of 62%, suggesting that the rate and yield of
Figure 4.3. a) Steady state absorption and emission of m-BETB in THF. b) Emission
decay of m-BETB in THF (excitation at 405 nm, emission at 570 nm), with
noticeable delayed fluorescence on the ~200 ns timescale.
91
non-radiative excited state decay processes are not significant in this system. The
emission decay, shown in Figure 4.3b, has a small delayed component (~190 ns),
analogous to those observed for weakly-coupled tetracene dimers by Bardeen
196,
197
and Damrauer could be suggestive of triplet-triplet recombination. Presuming that
the delayed fluorescence is a result of recombination of two geminate triplets
generated by singlet fission, we used a kinetic model analogous to the one reported
by Muller et al. and Cook et al. to estimate the yield of formation of triplets to be
approximately 1% in m-BETB, and the rate of the decay of the delayed fluorescence
to be solvent dependent.
191
The photoluminescence lifetimes of all solutions of m-
BETB in THF with concentrations varying from 1.5 μM to 21 μM is the same (~170 ns),
indicating the absence of any intermolecular reactions between two dimers in
solution. The triplets on m-BETB therefore recombine via a geminate pathway.
4.3.3 Transient absorption of m-BETB
The femtosecond transient absorption spectra of m-BETB in THF after excitation
at 515 nm are shown in Figure 4.4(a). The characteristic acene S n S 1 absorption
peak is observed at 400 nm, along with ground state bleach at 516 nm and 480 nm –
similar features were observed in the case of the ethynyltetracene monomer (ET-TMS)
in our earlier studies.
37
Between 10 ps and 975 ps, there is no significant evolution of
these spectral features. The S n S 1 absorption is observed to decay after 5 ns when
excited at 532 nm in a nanosecond pump broadband transient absorption experiment
(Figure 4.4b), with a concomitant rise of a positive feature at ~510 nm. A peak in this
92
spectral region has been previously observed for T n T 1 absorption in acenes like
5,12-diphenyltetracene (DPT) and ET-TMS.
135, 176, 198
More importantly, the spectral
Figure 4.4. Femtosecond (a) and nanosecond (b) transient absorption of m-BETB,
with the sensitized T n T 1 absorption shown in black. Inset to b) shows an overlay
of the transient spectrum of m-BETB at 135 ns and the sensitized T n T 1 spectrum.
93
shape at 135 ns is identical to the sensitized T n T 1 absorption of m-BETB (inset to
Figure 4.4b).
The rise of the absorption feature at 510 nm is slow and suggests that the triplets
could be generated in this system by intersystem crossing, with similar intersystem
crossing rates as other tetracene derivatives.
197
The transient absorption data was fit
with a two-state sequential model with a time constant of 15 ns for the rise of the
triplet feature and a triplet lifetime of 4.6 μs (see Chapter 5). The triplets could also
arise from a parallel channel in which very slow (and therefore inefficient) singlet
fission results in the delayed fluorescence from m-BETB in solution (~190 ns, Figure
4.3b). It is reasonable to assign the negligible singlet fission and slow formation of
triplets in m-BETB in solution to the weak non-adiabatic coupling, similar to the
meta-dimer in a series of tetracene dimers reported by Muller et al.
135
Ab initio
calculations further support this notion. The coupling parameter, || γ||
2
, between the
S 1 and the
1
(T 1T 1) states in m-BETB is calculated to be 4.6×10
-5
, a value that is about
three orders of magnitude smaller than for o-BETB (0.04).
186
Additionally, the
computed binding energy (E b) of the triplets (calculated as the difference between the
energy of the biexciton state and the two separated triplets) indicates how strongly
the T 1 excitons are coupled in the
1
(T 1T 1) state within the dimer and is reflective of the
intramolecular chromophore coupling. The value of E b in m-BETB was calculated to
94
Figure 4.5. The femtosecond transient absorption (a), the species associated
difference spectra (b) and concentration curves (c) from the target analysis of the
TA data of p-BETB-ohex in THF.
95
be <0.001 eV (compared to 0.15 eV for o-BETB), suggesting that the tetracenes are
weakly coupled in this dimer. Although singlet fission was observed to proceed on a
200 ps timescale in the TIPS-pentacene analog of m-BETB, the rate of singlet fission
in the meta- dimer in the series of the reported TIPS-pentacene dimers was slower by
at least three orders of magnitude compared to the ortho- and para- dimers.
137, 176
4.3.4 Transient absorption of p-BETB-ohex and p-BETB-ehex in THF
The TA spectra of p-BETB-ohex and p-BETB-ethex in THF, excited at 545 nm, and
530 nm, respectively, are shown in Figure 4.5(a) and 4.6(a). At early times (< 25 ps),
both compounds display the characteristic excited state absorption band at 425 nm,
and ground state bleach at 520-540 nm. The vibronic structure associated with
ground state absorption and excited state emission (through contributions from
ground state bleaching and stimulated emission) are superimposed on the broader
excited state absorption bands in the overall transient absorption spectra as seen in
Figures 4.5(b) and 4.6(b). The broad absorption feature at 425 nm decays within 100
ps, giving rise to an induced absorption peak at 555 nm in p-BETB-ohex and 540 nm
in p-BETB-ethex. This feature of the p-BETB dimers decays much more rapidly than
that of m-BETB, and most importantly the spectral shape of the induced absorption
of p-BETB at ~100 ps is also different from that of o-BETB reported previously.
37
In order to get an idea of the excited state dynamics in the p-BETB dimers, the TA
data was subjected to target analysis.
47, 199
The data was best fit using a 3-state
sequential model ([state 1]→[state 2]→[state 3]). The species associated difference
96
Figure 4.6. The femtosecond transient absorption (a), the species associated
difference spectra (b) and concentration curves (c) from the target analysis of
the TA data of p-BETB-ehex in THF.
97
spectra (SADS) and corresponding concentration curves obtained from target
analysis are shown in Figures 4.5 and 4.6, respectively. The shapes of the SADS of
p-BETB-ohex and p-BETB-ethex in THF are qualitatively similar (Figure 4.5(b) and
4.6(b)). The spectral signature of state 1 in both systems resembles the typical acene
S n S 1 absorption
37
, with a sharp peak in the 375-470 nm region with a larger
(negative) amplitude from ground state bleach and stimulated emission. In both
systems, state 1 rapidly decays to state 2 with a broader band in the 375-470 nm
region, a net absorption from 575 – 650 nm and similar ground state bleach features
to state 1. The spectral shape of state 2 of p-BETB-ohex and p-BETB-ethex in THF
resembles that of the
1
(T 1T 1) state absorption in o-BETB. The difference in the SADS
corresponding to state 2 in p-BETB-ohex and p-BETB-ethex suggests that the
spectrum originating from this state is dependent upon the structure of the bridging
linker, and therefore the coupling between acenes.
Recall that based on the splitting in the oxidation potential peaks, the conjugation
between the tetracenes was the strongest in o-BETB and p-BETB-ohex (0.22 V), and
significantly weaker (0.13 V) in p-BETB-ethex. The rate of formation of state 2 in the
p-BETB dimers, however, is much faster than the formation of
1
(T 1T 1) in o-BETB. The
target analysis gives values of 0.1 ps (essentially on the timescale of the instrument
response) for p-BETB-ohex and 0.4 ps for p-BETB-ethex (Table 4.2), while in o-BETB,
1
(T 1T 1) was formed in 2 ps. The faster rate of formation in the para- dimers than in the
ortho- dimer leads us to conclude that the predominantly through-bond coupling
98
operating in both p-BETB-ohex and p-BETB-ehex is stronger than the combined
through-bond and through-space non-adiabatic coupling in o-BETB, although the
through-bond coupling in o-BETB is likely diminished due to the acenes being slightly
out of plane with the linker.
Unlike in o-BETB, the
1
(T 1T 1) state in p-BETB-ohex and p-BETB-ethex decays to a
third state with an identifiably different spectral signature, with prominent induced
absorption peaks at 555 nm and 540 nm in p-BETB-ohex and p-BETB-ethex,
respectively. The spectral shape of state 3 in p-BETB-ohex and p-BETB-ethex in
solution bears a strong resemblance to that of the sensitized T n T 1 induced
absorption of the monomer, ET-TMS.
81
The absorption of state 3 could correspond to
either one or two triplets per dimer. There are two factors which suggest that state 3
in fact corresponds to absorption of two triplets on the same dimer: (i) the rate of
formation of state 3 is fast – 19 ps in p-BETB-ohex and 48 ps in p-BETB-ehex (Table
4.2). These rates of T 1 formation are too fast for intersystem crossing in acenes
197
, but
are in the range of the measured singlet fission rates in tetracene
176, 200-202
. (ii) The
Table 4.2. The time constants (in ps) of the decays used to fit the TA data of the
dimers in THF.
o-BETB
1
m-BETB
2
p-BETB-ohex p-BETB-ehex
τ[S 1-
1
(T 1T 1)] 2.0
1.5 x 10
4
0.1 0.4
τ[
1
(T 1T 1)-
3
T 1] 19 48
τ[
3
T 1] 4.6 x 10
6
20 76
1
Taken from previously published work, reference 17.
2
Time constants corresponding to the intersystem crossing rate and the decay of
that triplet.
99
rate of the decay of state 3 is very fast – 20 ps in p-BETB-ohex and 76 ps in
p-BETB-ehex. Typically, a single triplet exciton localized on an acene dimer decays on
the timescale of microseconds.
197
However, the recent studies of pentacene dimers
showed that when two triplet excitons are located on the same dimer, the T n T 1
absorption decays on the 10-100 ps timescale, due to the favorable T 1-T 1 fusion
pathway to the ground state.
154
Therefore, the SADS for state 3 in the p-BETB dimers
in THF is the spectral signature of two triplets in close proximity to each other. The
most likely origin of these two triplets is via singlet fission.
In Figure 4.7, an overlay of the SADS for state 3 of p-BETB-ethex is made with T n
T 1 absorption spectra for a single triplet sensitized onto p-BETB-ethex and the model
compound, ET-TMS. The details regarding sensitization experiments have been
previously reported.
37
For both the sensitized spectra, a spectral shift is required to
line peaks up with the state 3 SADS. Despite small discrepancies in the intensities in
the 400 - 500 nm region, the relative peak positions and relative intensities overlay
well, strongly suggesting that state 3 corresponds to T n T 1 induced absorption in the
p-BETB dimers. It is in fact likely that the SADS of state 3 represents a scenario in
which the two T 1 excitons localize on their respective acenes without any
delocalization onto the benzene linker (presumably in 90° twisted orientation). This is
supported by the closer spectral similarity of state 3 to the sensitized triplet
spectrum of ET-TMS, which does not contain a phenyl group at the other end of the
alkyne. We estimated the triplet yield in these compounds by scaling the target
100
analysis populations by the extinction coefficients of the singlet and the triplet
absorptions and obtained 58% for p-BETB-ethex and 42% for p-BETB-ohex,
respectively. The triplet yields are fairly low in these p-BETB dimers because the rate
of recombination of the triplets is fast relative to the rate of their formation.
Surprisingly the results obtained from our ortho-,
37
meta- and para-
ethynyltetracene dimers are different from those obtained by Tykwinski, Guldi and
coworkers on isomorphic dimers of TIPS-pentacene.
176
All of the TIPS-pentacene
dimers undergo singlet fission, with only the meta-TIPS-pentacene dimer undergoing
fission slowly enough to be resolved by the instrument.
In our previous work on o-
Figure 4.7. The plot shows the spectral signature of state 3 from target analysis of
p-BETB-ehex in THF compared to the normalized sensitized T n T 1 absorption of
ET-TMS (red-shifted, 20 nm) and sensitized T n T 1 absorption of p-BETB-ehex in
THF (blue-shifted, 9 nm).
101
BETB, the formation of only the biexciton state was observed while here we find
evidence of predominantly radiative decay in m-BETB using femtosecond transient
absorption. The para-ethynyltetracene dimer is observed to form two independent
triplet states via the biexciton state. We observe using TCSPC (with more dynamic
range) a small fraction (1%) of m-BETB molecules gives rise to delayed fluorescence
which were not captured in transient absorption. The delayed fluorescence in m-BETB
is solvent dependent, suggesting that a charge resonance state is involved, similar to
the meta-TIPS pentacene dimer. The singlet excited state (1.88 eV) lies close to the
CR state (1.87 eV) in meta-TIPS pentacene dimer while in our m-BETB, the singlet
state (2.40 eV) is less likely to interact with the CR state (2.3 eV) and could be why we
do not see CR state mediated singlet fission in m-BETB using trasnsient absorption
while in the meta-TIPS-pentacene dimer, Tykwinski, Guldi and coworkers do observe
CR state mediated singlet fission. In the TIPS-pentacene dimers, the energy of 2 x
E(T 1) at 1.54 eV is far below the energy of the S 1 state at 1.88 eV. This energetic driving
force for singlet fission in the TIPS-pentacene systems results in ultrafast kinetics
from which mechanistic information is not easily obtained. In the case of
ethynyltetracene dimers presented here, the Energy(Singlet) ≈ 2 x Energy(Triplet)
isoenergetic conditions make singlet fission only slightly favorable. As a result of
which the subtle differences in the effect of through-bond and through space
coupling can be investigated.
102
The ortho- and the para- bis-ethynylbenzene dimers exhibit distinct excited state
processes upon excitation, the S 1 state in para-BETB gives way to the
1
(T 1T 1) state
which decays to form independent triplets while the
1
(T 1T 1) state in o-BETB only
relaxes to the ground state without producing independent triplet excitons. According
to the model proposed by Abraham and Mayhall, both ortho- and para- dimers should
produce bound
1
(T 1T 1) states and non-interacting triplet excitons should not be
observed in both of these cases, however our experiments suggest otherwise.
32
The
difference in the excited state processes in the ortho- and para- dimers could be
reconciled by considering the effect of molecular structure on the possible rotational
configurations of the chromophores in these dimers. In o-BETB the tetracenes exhibit
free rotation, however they always retain some amount of π overlap. Therefore, once
the
1
(T 1T 1) state is formed in o-BETB, the π overlap enforces the triplets to always be
bound to each other. In p-BETB, on the other hand, the only source of electronic
coupling (|| γ||
2
) is through-bond. Upon twisting of one of the tetracenes perpendicular
to the plane of the rest of the molecule, the electronic coupling between the
Figure 4.8. The proposed decoupling of the correlated triplet state in
p-BETB-ehex (alkyl chains have been removed for clarity) to give rise to two
independent triplets.
103
chromophores could be broken, and the triplets from the correlated triplet pair state
become independent of each other; the Abraham and Mayhall model may not apply to
a twisted configuration of the para- dimer. Figure 4.8 shows the possible molecular
orientations corresponding to the correlated and independent triplet pair states in
p-BETB-ehex.
Further insight into the effect of rotation on the excited state energies of p-BETB
can be obtained from theory. The energies of the S 1,
1
(T 1T 1), and
5
(T 1T 1) states were
computed as a function of the dihedral angle between the tetracenes in p-BETB-ehex
using the restricted active space configuration interaction with double spin-flip
method (RAS-2SF-CI, Table 4.3). The minimum of the S 1 corresponds to a planar
geometry and the
1
(T 1T 1), and
5
(T 1T 1) states are energetically downhill from the S 1 state
by about 0.4 eV at all rotational angles. The coupling between the S 1 and the
1
(T 1T 1)
states (|| γ||
2
) was calculated to be largest at a planar geometry (0.020) and is half that
at 45 (0.010) and drops to 0.005 at 90 . The
1
(T 1T 1) state has a minimum at 60 degrees,
Table 4.3. The RAS-2SF-CI computed excited state energies (in eV) for
p-BETB-ehex as a function of the dihedral angle between the tetracenes.
Angle E(S 1) E(
1
(T 1T 1)) E(
5
(T 1T 1))
0 3.839 3.381 3.422
30 3.907 3.449 3.490
60 3.856 3.370 3.380
90 3.874 3.391 3.391
104
while the
5
(T 1T 1) state has a minimum at 90 degrees, at which the
1
(T 1T 1) and the
5
(T 1T 1)
states become degenerate (see Table 4.3). These values indicate that the triplet
excitons from the correlated triplet pair state in p-BETB can become non-interacting
in the
5
(T 1T 1) state when the tetracenes rotate perpendicular to each other within the
dimer.
4.3.5 Transient absorption of p-BETB-ehex in PMMA
If rotational motion of the tetracene in para- dimer in solution provides the
mechanism for generating free triplets, then restricting that motion should reduce the
rate and the yield of the formation of free triplets. To test this hypothesis, a solution
of p-BETB-ethex in a rigid polymer matrix, poly(methylmethacrylate) (PMMA) was
prepared. p-BETB-ethex in PMMA produces a substantial amount of emission at room
Figure 4.9. The femtosecond transient absorption of p-BETB-ehex in PMMA.
105
temperature
188
, with a quantum efficiency of ~10%, which means that a significant
fraction of the molecules are confined to a geometry in which the radiative decay is
the predominant relaxation pathway from the singlet excited state. From a
comparison of the transient absorption spectra of p-BETB-ethex in THF and in PMMA
(Figure 4.9), it is evident that in PMMA the excited state absorption corresponding to
S 1→S n absorption centered at 420 nm decays much more slowly. The characteristic
T 1→T n absorption peak at 550 nm still forms in this media; however, the rate of its
formation is much slower than for p-BETB-ethex in THF. Unlike p-BETB-ethex in THF,
the transient signals observed from p-BETB-ethex in PMMA cannot be fit using a
3-state sequential model.
PMMA being a more rigid medium than THF, raises the barrier to rotation for the
tetracene units in p-BETB-ethex along the diethynylbenzene linker. This results in
p-BETB-ethex dimers that are conformationally locked in the polymer matrix in a
distribution of rotational angles. Upon excitation in this rigid medium not all the
p-BETB-ethex molecules which end up in the
1
(T 1T 1) state will be able to rotate and
form independent triplets. The TA spectra from p-BETB-ethex in PMMA with the
tetracene units locked in all dihedral angles can be modeled as a sum of signals from
two model populations with extreme conformations of the para-dimer: population 1
with the tetracenes coplanar with each other and population 2 with tetracenes
perpendicular (Figure 4.10). Population 1 represents the dimer molecules in PMMA
which cannot undergo singlet fission completely, and the
1
(T 1T 1) remains in
106
equilibrium with the S 1 state, giving rise to the fluorescence signal from the PMMA
sample. Population 2 is representative of dimer molecules that can fully undergo
singlet fission to two independent triplets and gives rise to the triplet band at 550 nm
in the TA spectra at 957 ns. Using compartmental models to represent the excited
state processes in the distinct populations, it is possible to obtain concentration
curves that describe the dynamics of the excited states.
190
The transient absorption
spectra corresponding to both model populations 1 and 2 can be derived by
multiplying the concentration curves with the Species Associated Difference Spectra
corresponding to the S 1,
1
(T 1T 1), and T 1-T 1 states. The fractions of the model population
1 and population 2 that best represent p-BETB-ethex in PMMA is obtained by
comparing the time slice of the transient absorption data at 425 nm with the time
slices at the same wavelength of the modeled transient absorption data; the latter is
Figure 4.10. The state diagrams showing (a) population 1 and (b) population 2 used
to model the TA data from p-BETB-ehex in PMMA.
2(T
1
)
Energy
S
1
1
(T
1
T
1
)
S
0
S
1
1
(T
1
T
1
)
⇌
S
0
k
’
12
k
23
k
33
k
12
k
21
k
22
k
R
k
NR
(a) (b)
107
obtained by combining the TA spectra of the population 1 and population 2 in different
fractions and optimizing the time constants for the various excited state processes.
As seen in Figure 4.11, the best representation of the excited state dynamics is
obtained by combining the dynamics of populations 1 and 2 in equal fractions and
reproduces the 10% quantum yield measured from the PMMA film. This suggests that
in a more realistic model, the tetracene units in p-BETB-ehex in the polymer matrix
would be locked at all dihedral angles with equal probability. This is quite likely the
case since the monomer units in PMMA are small and would not hold the tetracenes
at any one angle preferentially by non-covalent interactions. The TA spectra of
Figure 4.11. Dynamics of excited state absorption at 425 nm of p-BETB-ehex in
PMMA (black trace) overlaid with the time slices at 425 nm of transient absorption
spectra obtained by combining modelled transient absorption of population 1 and
population 2 in different fractions. Numbers in the legend show the fraction of
population 1 in that model. The rate constants that gave the best fit are: k r = 5.5
× 10
7
, k nr = 2.9 × 10
6
, k 12 = 5 × 10
11
, k 21 = 4 × 10
12
, k 22 = 2 × 10
9
, k 12’ = 1.6 × 10
12
, k 23
= 1.2 × 10
10
, k 33 = 1.2 × 10
9
.
108
p-BETB-ehex in PMMA and its modeling using conformationally restricted
populations emphasizes the importance of the rotation of the tetracene units in
allowing the
1
(T 1T 1) state to separate to independent triplets in p-BETB-ethex and not
in o-BETB – despite the similarities in electronic coupling.
Of additional interest are the differences in the excited state kinetics of
p-BETB-ohex and p-BETB-ethex. All three of the processes: relaxation of S 1 into
1
(T 1T 1), conversion of
1
(T 1T 1) to independent triplets, and the decay of the resultant
triplets are faster in p-BETB-ohex compared to p-BETB-ethex (Table 4.2). The
difference in the coupling of the tetracene units in p-BETB-ohex and p-BETB-ethex
can be understood in terms of the relative orbital energies of the tetracenes and the
two bis-ethynylbenzene linkers. A qualitative diagram showing the relative orbital
energies of these molecules is provided in Figure 4.12. In the alkyl-substituted
Figure 4.12. A diagram showing the closer lying orbitals of the linker and the
tetracenes in a) p-BETB-ehex compared to b) p-BETB-ohex. l a/b and h a/b are the
frontier molecular orbitals of tetracene.
(a) (b)
109
benzene linker, the orbitals are expected to lie energetically far from those of the
tetracenes and the coupling is not expected to be very large. However, introducing the
electron donating hexyloxy substituents onto the benzene linker raises the energy of
the HOMO, and therefore brings the orbital energies of the linker closer to those of the
tetracenes, resulting in stronger coupling between the acenes and faster singlet
fission in p-BETB-ohex.
4.4 CONCLUSIONS
In this chapter the excited state dynamics of meta- and para- analogues of the
previously reported ortho-bis-ethynyltetracenylbenzene are presented in order to
study the role of through-bond vs. through-space coupling in facilitating singlet
fission between the tetracene units. The ortho- and para- ethynylbenzene linkers
provide conjugation between the tetracene units while the meta- ethynylbenzene
linker is cross-conjugating in the ground state. Using broadband femtosecond
transient absorption spectroscopy, we established that the
meta- bis-ethynyltetracenylbenzene predominantly undergoes radiative decay in the
excited state. Surprisingly, the para- bis-ethynyltetracenylbenzene dimer undergoes
complete singlet fission, forming independent triplet states. As reported previously,
the ortho- bis-ethynyltetracenylbenzene dimer only forms the intermediate
correlated triplet pair state.
Comparing the different results from ortho-, meta- and
para- bis-ethynyltetracenylbenzene dimers provides crucial insight into the role of
110
conjugation and through-bond coupling in singlet fission in dimers. The ortho- and
para- dimers in which the tetracene units are conjugated through the ethynylbenzene
linker exhibit singlet fission while the cross-conjugated meta-dimer predominantly
undergoes radiative decay. Furthermore the [S 1]→[
1
(T 1T 1)] process is faster in
p-BETB-ohex and p-BETB-ehex than in o-BETB, suggesting that the through-bond
coupling operating in the para- dimers is stronger than the combined through-space
and through-bond coupling operating in the ortho- dimer. Unlike in the case of a series
of pentacene dimers, where the energetics are so favorable for singlet fission that the
ortho-, meta- and para- pentacene dimers all displayed triplet absorption signatures
in picosecond timescales, in our case, only the para-dimer forms independent triplets
on similar timescales. This provides an insightful comparison with the meta- dimer
which does not form significant number of triplets and the ortho-dimer which only
forms the intermediate correlated triplet pair state. We attribute the difference in the
excited state dynamics of the ortho- and para- dimers to the rotational flexibility of
the acenes in the para- dimer, which breaks the coupling between the chromophores
and allows the triplets to separate from the correlated triplet pair state. Such rotation
of the tetracene units in the ortho-dimer is hindered because they are confined to a
cofacial orientation. The role of tetracenes rotating to form separated triplets is clear
when comparing the transient absorption of the para-dimer in PMMA to that in THF.
PMMA being a more rigid medium prevents the rotation and significantly hinders the
formation of the two non-interacting triplets.
111
A comparison of the two para-bis-ethynyltetracenylbenzene dimers (p-BETB-ohex
vs. p-BETB-ehex) provides insight into the role of the energetics of the linker in singlet
fission in covalent dimers. The bis-hexyloxy-bis-ethynylbenzene linker in
p-BETB-ohex has molecular orbitals that are closer in energy to orbitals of tetracenes
than the bis-(2-ethyl)-hexyl-bis-ethynylbenzene linker in p-BETB-ehex. This results
in better mixing of the linker orbitals with those of the tetracenes and thus better
coupling in p-BETB-ohex than in p-BETB-ehex. Better coupling between the tetracene
units results in faster formation of the correlated triplet pair state and the separated
triplets in p-BETB-ohex than in p-BETB-ehex – highlighting the role of linker
energetics in singlet fission.
112
CHAPTER 5
Decoupling Inter- and Intra-Dimer Singlet Fission
5.1. INTRODUCTION
Singlet fission is a process in which the energy of a singlet exciton (S 1) on a
chromophore is shared with a chromophore in the ground state to create two triplet
excitons (T 1) on each chromophore.
38, 150, 203
Using materials that can singlet fission in
solar cells could theoretically improve the power conversion efficiency by doubling
the number of excitons produced from high energy photons.
151
The first of two steps
in singlet fission involves the formation of the correlated triplet pair or the
multiexciton state (ME) from the singlet state. The second step is separation of the
correlated triplet pair into two different triplet excitons. First observed in crystals of
anthracene,
191
singlet fission has been more recently studied in polyacenes
37, 204
,
thiophene polymers, and carotenoids, amongst others. In tetracene, singlet fission is
thermally activated (2 x energy of triplet – energy of singlet = 0.19-0.24 eV) and occurs
in 40-90 ps while in pentacene it is exergonic (2 x energy of triplet – energy of singlet
= -0.11 eV) and happens in 80 fs. However, the presence of substituents on the
tetracene ring make singlet fission exergonic with rate constants on the order of 1 ps
with some approaching the rates observed with pentacene systems.
182
Singlet fission
is not only dependent on the energy difference between the singlet and triplet states,
113
but also on the electronic coupling between the chromophores.
197
To understand the
role of coupling, there has been a great deal of interest in studying singlet fission in
dimers consisting of covalently linked chromophores. Such dimers provide the means
to vary the degree of coupling between the chromophores by synthetic means and
study the rate of singlet fission as a function of the coupling.
178, 205
Our group has studied singlet fission in amorphous films of 5,12-
diphenyltetracene
185
and in solutions of ortho-, meta- and para-ethynyltetracene
dimers as described in the last chapter.
182
It was found that in an amorphous film,
most singlet exciton need to diffuse to find optimal locations in the film where
adjacent 5,12-diphenyltetracene molecules are in the correct geometric
arrangement facilitating efficient coupling for singlet fission. The ortho-
ethynyltetracene dimer was designed to hold the ethynyltetracene units in close
contact and provide the coupling required to facilitate singlet fission between the
ethynyltetracene units. It was found through ultrafast transient absorption
experiments that in solution the photoexcited ortho-ethynyltetracene dimer does not
fully singlet fission – it is trapped in the intermediate multiexciton state. The meta-
ethynyltetracene dimer in solution does not undergo singlet fission at all – instead,
upon excitation it produces a single triplet exciton via intersystem crossing. The para-
ethynyltetracene dimer in solution forms independent triplet excitons via the
intermediate multiexciton state as observed via transient absorption spectroscopy.
114
In this chapter the results from a similar study on amorphous film samples of
meta- and para-ethynyltetracene (m-BETB and p-BETB-ehex) dimers (Figure 5.1) are
described. A comparison of the results from solution and film phase dimers of
ethynyltetracene from chapters 4 and 5 provides insight into the role of (1) intra-
dimer electronic coupling between chromophores and (2) inter-dimer interactions
have on singlet fission. Wavefunction analysis and calculations based on an adiabatic
framework have shown that through-space inter-dimer chromophore interactions
are not significant in comparison to intra-dimer interactions facilitated by through-
bond coupling.
130
We use broadband femtosecond transient absorption spectroscopy
and time-correlated single photon counting to reveal the excited state properties of
the meta- and para- ethynyltetracene dimers in neat film phases and use target
analysis to extract the species associated spectra and corresponding concentration
profiles of the excited states. The case of the meta-linked chromophores proves
interesting as the intra-dimer through bond coupling is unfavorable and the effect of
inter-dimer coupling is revealed.
Figure 5.1. Structures of o-BETB (a), m-BETB (b) and p-BETB-ehex (c) dimers of
ethynyltetracene
115
5.2 EXPERIMENTAL SECTION
5.2.1 Synthesis of ethynyltetracene dimers and sample preparation
Meta-, and para-ethynyltetracene dimers were synthesized using Sonogashira
coupling and purified using techniques as described previously for the ortho-
dimer.
188
Films samples of dimers were prepared by dissolving 0.0005 g of the acene
in 0.4 ml of freshly distilled THF, filtering that solution through a 2 μm syringe filter
and spin casting the solution on a 1 mm quartz substrate at 3500 rpm. Spin cast films
had an optical density between 0.2 and 0.3.
5.2.2 Femtosecond Transient Absorption
Pump and probe pulses for the TA experiment were derived from the output of a
Coherent Legend Ti:sapphire regenerative amplifier (800 nm, 1 kHz, 4 mJ, 35 fs).
Pump pulses were obtained using an OPA-800C optical parametric amplifier from
Spectra Physics. Excitation wavelengths used for the film samples were 535 nm and
564 nm respectively. Spot size of the pump on the samples was 300 μm in diameter
and the pulse energy was 0.2 μJ, resulting in an excitation fluence of 280 μJ cm
-2
.
Probe pulses of white light supercontinuum (350-650 nm) were generated by focusing
the amplifier output into a rotating CaF 2 disk. The polarization of the probe pulse was
set at magic angle to the pump pulse to avoid orientation effects to the dynamics. The
probe pulse was collimated and focused on the sample using two off-axis parabolic
mirrors while the pump pulse was focused using a CaF 2
lens. The instrument response
from this set up was previously characterized to be 150 fs for film samples and 180
116
fs for solution samples.
138
After passing through the sample, the supercontinuum was
dispersed using an Oriel MS1271 spectrograph onto a 256 pixel Hamamatsu silicon
diode array. Samples were slowly translated perpendicular to the beam path using a
linear stage to avoid photodamage.
5.2.3 Time-Correlated Single Photon Counting
Photoluminescence lifetimes were measured using single photon counting
apparatus. Excitation pulses were derived from a Coherent RegA 9050 250 kHz
regenerative amplifier and a Coherent OPA 9450 optical parametric amplifier.
Lifetime data was measured using a pump pulse power of 0.5 mW and a spot size of
250 μm at the sample, resulting in an excitation fluence of 4.0 μJ cm
-2
. TCSPC
measurements were performed using a Digikröm CM112 double monochromator and
a R3809U-50 Hamamatsu PMT with B&H SPC-630 module with a time resolution of
22 ps.
5.3 RESULTS AND DISCUSSION
5.3.1 Steady – state absorption spectroscopy of m- and p-BETB
The steady-state absorption spectra of meta- and para- ethynyltetracene dimers
in THF solutions and neat films are shown in Figure 5.2. The absorption spectra show
the presence of a strong vibrational progression that is a characteristic feature of
polyacenes. The red shift in the absorption spectra of the neat films compared to the
corresponding solution absorption spectra is indicative of inter-dimer electronic
coupling. The first excitonic peak of the meta- dimer changes from 516 nm in THF
117
solution to 535 nm in neat film – resulting in a shift corresponding to 0.086 eV. In the
para- dimer, this peak shifts from 529 nm in solution to 564 nm in neat film
corresponding to a shift of 0.146 eV. The larger shift in energy of the para- dimer
indicates more significant inter-dimer interaction compared to the meta- dimer. The
vibronic bands in solutions of both meta- and para- ethynyltetracene dimers are
observed to broaden in the corresponding neat films indicating the amorphous nature
of the spin coated films as observed in the case of 5, 12-diphenyltetracene and the
ortho-ethynyltetracene dimer.
182, 185
Figure 5.2. Steady state absorption spectra of m-BETB and p-BETB-ehex in THF
solution and neat films. The difference in the first excitonic peak between solution
and film for m-BETB is 0.086 eV and 0.146 eV for p-BETB-ehex.
118
5.3.2 Time-resolved photoluminescence of m-and p-BETB-ehex
The photoluminescence decays from the BETB dimers were measured using a
Time-Correlated Single Photon Counting (TCPSC) apparatus. Data for m-BETB is
shown in Figure 5.3. The photoluminescence decay from THF solution of meta- dimer
is well fit with a single exponential with time constant 14.7 ns. In contrast, the
photoluminescence from the neat film sample was observed to decay within the
instrument response of the apparatus i.e. < 22 ps (inset to Figure 5.3). No long lived
minor components are observed in the decay, in contrast to the o-BETB dimer
138
where the long-lived component signals the presence of delayed fluorescence. The
decrease in the photoluminescence lifetime of the neat film of m-BETB compared to
Figure 5.3. Photoluminescence decay of m-BETB in THF solution and neat films as
measured by Time-Correlated Single Photon Counting (TCSPC). The decay can be
fit with a single exponential with time constant 14.7 ns for m-BETB in THF solution.
Inset shows the decay of photoluminescence from a m-BETB neat film is limited
by the instrument response (22 ps).
119
isolated dimers in solution indicates the presence of a new channel that deactivates
the singlet state in neat film.
5.3.3 Transient absorption of m- and p-BETB-ehex neat films
Figure 5.4a shows the transient absorption spectra, obtained by exciting a neat
film of p-BETB-ehex at 564 nm. The broad, positive features seen at 400-450 nm and
at 600-650 nm are characteristic of the intermediate multiexciton state in singlet
fission, previously observed in the o-BETB dimer.
182
The negative signal at 500-600
nm is the bleach corresponding to the depopulation of the ground state by the pump
pulse. The evolution of these signals with change in pump-probe delay is shown in
Figure 5.4b. The transient absorption signals at 425 nm, 565 nm and 660 nm were
Figure 5.4. Transient absorption of para-BETB-ehex neat film (a), excited at 564
nm. Legend shows the pump probe delays at which each spectrum was taken and
(b) time slices of the transient absorption data at probe wavelengths 425 nm, 565
nm and 660 nm overlaid with tri-exponential fits.
120
observed to decay at similar rates as observed by the near-parallel time slices. The
time slices can all be fit with three decaying exponentials ~190 ps (25%), ~20 ps (35%)
and ~2 ps (40%). To obtain more insight into the excited state properties of p-BETB-
ehex in film, the transient absorption data was subjected to target analysis
190
using
different compartmental models. The best fit to the data was obtained using a
sequential two-state model as shown in Figure 5.5a. Unlike in the neat film, the
transient absorption spectra of p-BETB-ehex in solution (Figure 5.6) showed a sharp
positive feature
187, 188
at 410 nm that is characteristic of the singlet excited state,
similar to o-BETB.
182
The singlet state gives way to the multiexciton state with a
broader band at 410 nm, similar to what was observed in o-BETB. Using target
analysis, the best fit to the transient absorption data was obtained using a three-
state sequential model as shown in Figure 5.6a. The corresponding species
associated spectra (SAS) and concentration profiles of the three states are shown in
Figure 5.5. Sequential 2-state model used in the target analysis of the transient
absorption of para-BETB-ehex neat film and the time constants (a),
corresponding species associated spectra (SAS) (b), and the concentration
profiles (c).
121
Figures 5.6b and 5.6c. The difference in the excited state dynamics of p-BETB-ehex
in neat film as compared to the THF solution can be assigned to the presence of new
singlet fission channels in the film that were absent in the solution. The close packing
of p-BETB-ehex molecules in the neat film provides opportunities for inter-dimer
singlet fission between ethynyltetracene chromophores on different dimers as well
as intra-dimer singlet fission between ethynyltetracene units on the same dimer.
While in solution, the p-BETB-ehex units remain isolated and can only undergo intra-
molecular singlet fission. It is plausible that in the film, both the inter- and intra-
dimer singlet fission pathways quickly depopulate the singlet state of p-BETB-ehex
upon excitation such that the transfer to the multiexciton state happens within the
instrument resolution (150 fs). The first state observed upon excitation in the
transient absorption spectra of p-BETB-ehex in neat film is then the multiexciton
state which gives way to the triplet state in 5.7 ps (Figure 5.5a).
Figure 5.6. Sequential 3-state model used in the target analysis of the transient
absorption of para-BETB-ehex in THF solution (a), the corresponding species
associated spectra (SAS) (b), and concentration curves (c).
122
The transient absorption spectra of m-BETB in THF solution (Figure 5.7) had a
positive feature at 425 nm with a sharp peak that is characteristic of the S 1
state.
188
This state persisted for several hundred nanoseconds such that for the
experiment we needed to switch over to a nanosecond laser to keep measuring the
transient spectra at such long pump-probe delays. Thereafter, a positive band at 500
nm and 550 nm emerged, that is similar to the triplet state spectrum observed for o-
BETB.
182
The transient absorption spectra was best fit with a 2-state model, as shown
in Figure 5.7. This slow formation of the triplet state from the singlet state is
indicative of intersystem crossing in the THF solution of m-BETB.
206
In sharp contrast,
in a neat film of m-BETB, a sharp positive feature similar to that assigned to the
singlet state discussed previously is observed at 425 nm for only a few picoseconds
initially. Thereafter, this feature broadens and has a spectral shape reminiscent of a
multiexciton state that persists for a few hundreds of picoseconds.
188
The transient
absorption data is found to be best fit to a 2-state sequential model when subject to
Figure 5.7. 2-state model used in the target analysis of the transient absorption
of m-BETB in THF solution (a), the corresponding species associated spectra
(SAS) (b), and concentration curves (c).
123
target analysis (Figure 5.8a). The species associated spectra (Figure 5.8b) shows
states with sharp and broad positive features at 425 nm that can be assigned to the
singlet and multiexciton states, respectively. The concentration curves (Figure 5.8c)
reveal that while the singlet state gives way to the multiexciton state with a time
constant of 5.6 ps, the latter decays to the ground state much slower, with a time
constant of 456 ps. Thus, while m-BETB only undergoes intersystem crossing in
solution, in neat film it is does singlet fission to give rise to a multiexciton state. This
dramatic change in the excited state processes in m-BETB resulting in the singlet
state persisting for several hundreds of nanoseconds in solution and decaying in a
few picoseconds in film can be attributed to the inter-dimer chromophore
interactions between ethynyltetracene units present in the neat film but absent in
solution. The inter-dimer interactions are significant enough to dominate over the
weak intra-dimer meta-conjugation coupling, a motif that has been previously shown
to retard singlet fission in acenes.
207
Figure 5.8. Sequential 2-state model used in the target analysis of the transient
absorption meta-BETB neat film (a), the corresponding species associated
spectra (SAS) (b), and concentration curves (c).
124
5.4 CONCLUSIONS
The excited state processes involved in singlet fission in meta- and para-
ethynyltetracene dimers have been probed using transient absorption spectroscopy
and time-correlated single photon counting. The m-BETB dimer does not undergo
singlet fission isolated in solution, owing to the meta-conjugated, intra-dimer
through-bond coupling that is known to retard singlet fission. m-BETB instead simply
undergoes intersystem crossing to yield a single triplet, with a time constant of 16 ns.
In neat film, singlet fission in m-BETB is turned on, facilitated by the inter-dimer
through-space coupling present between ethynyltetracene units on adjacent dimers.
In the neat film, the singlet state of m-BETB gives way to the multiexciton state in 5.6
ps, an increase of four orders of magnitude in the rates of the processes depopulating
the singlet state. The p-BETB-ehex dimer does singlet fission when isolated in
solution, with the singlet state decay giving rise to the multiexciton state in 0.3 ps,
facilitated by the favorable para-conjugated intra-dimer, through-bond coupling. In
neat film, p-BETB-ehex not only has favorable intra-dimer through-bond coupling,
but also inter-dimer through-space coupling. Both these pathways to singlet fission,
i.e. intra-dimer as well as inter-dimer, work to depopulate the singlet state of p-
BETB-ehex in film so fast that it is not seen with the instrument resolution of out
setup (<150 fs) – thus the first state captured upon exciting the p-BETB-ehex film is
the multiexciton state. Thus, by studying a meta-conjugated dimer which does not
singlet fission when isolated in solution, it is possible to observe singlet fission
125
facilitated entirely by inter-dimer coupling in the neat film. It is also interesting to
note that singlet fission as facilitated by inter-dimer coupling depopulated the
singlet state in 5.6 ps (m-BETB film) while singlet fission facilitated by intra-dimer
coupling depopulated the singlet state in 0.3 ps (p-BETB-ehex solution) indicating
that intra-dimer, through-bond coupling is stronger in facilitating singlet fission than
inter-dimer, through-space coupling. A similar conclusion is arrived at for o-BETB
dimers via calculations on adiabatic surfaces and wavefunction analysis.
130
126
CHAPTER 6
Building a Fluorescence Upconversion
Spectrometer
6.1 INTRODUCTION
Time resolved fluorescence techniques
208
are a set of powerful tools for the
investigation of matter and living systems at a molecular level.
209-213
Generally, these
techniques involve monitoring the time-resolved fluorescence response after
exciting the fluorescent sample by high frequency modulated light (phase-modulated
fluorometry) or by short pulses of light (pulse fluorometry). Phase-modulated
fluorometry works in the frequency domain while pulse fluorometry works in the time
domain. Various time domain techniques exist: direct emission decay measurements
(> 100 ps), Time-Correlated Single Photon Counting (TCSPC, or now called Single
Photon Timing SPT, few ps to 100’s μs), streak cameras (1 ps to 10’s ns), and
fluorescence upconversion (10’s fs to 1 ns).
Of the various pulse modulated techniques, fluorescence upconversion is the
most advanced – it allows recording of fluorescence with very high time resolution,
basically only limited by the pulse widths of the lasers being used. Upconversion is
based of the principle of sum-frequency generation (SFG). It involves mixing a gate
127
pulse at frequency 𝜐 𝐺 with the fluorescence at frequency 𝜐 𝐹 in a nonlinear crystal to
create a short pulse at the sum frequency 𝜐 𝑈 = 𝜐 𝐺 + 𝜐 𝐹 . The gate pulse effectively acts
as a time window during which the fluorescence is upconverted. The intensity of the
upconverted light is proportional to the fluorescence intensity and by controlling the
delay between the gate pulse and fluorescence at the nonlinear crystal, a kinetic trace
of the fluorescence can be obtained (Figure 6.1).
Mahr and Hirsch in 1975 reported for the first time using upconversion to measure
the duration of sub-nanosecond pulses.
214
Mixing a gate pulse from a mode-locked
argon ion laser with the luminescence from Rhodamine 6G at an ammonium
dihydrogen phosphate (ADP) crystal, the authors were able to observe the decay of
the dye luminescence over few ns. The output of the upconversion was detected using
a Bausch and Lomb monochromator and EMI 6552S PMT assembly which afforded
measuring single-color measurement with respect to optical delay between the gate
Figure 6.1. Schematic of a typical fluorescence upconversion setup.
128
and luminescence pulses. Block and Shah followed up with an experiment in 1986 to
study the relaxation of photoexcited carriers InGaAs using upconversion.
215, 216
Briefly, the output of a cavity dumped Rh 6G dye laser was divided and one beam was
used to excite the InGaAs sample while the other beam was variably delayed. Both
beams were them focused on a 2 mm thick LiIO 3 crystal which was angle tuned to
achieve noncollinear phase matching and generate sum frequency photons. The
instrument resolution of this setup was found to be 380 fs by measuring the
upconversion of scattered light from the sample. For detection, the upconverted
photon were dispersed by a spectrometer and detected using a photomultiplier tube
followed by a photon counter.
The first experiment to study solvation dynamics using fluorescence
upconversion came out of Graham Fleming’s group in 1987 at the University of
Chicago.
217
The authors measured the time-resolved Stokes shift of the dye LDS-750
in various solvents in sub-picosecond timescales by upconverting the fluorescence
from the dye molecule using one half of the output of a mode locked Nd
3+
YAG laser
while the other half was used to excite the sample. The fluorescence and the gate
beams were focused collinearly into a 2 mm LiIO 3
crystal which produced UV sum
frequency photons which were then filtered (Schott UG11) and frequency selected
using an ISA DH-20 monochromator. The detector was an EMI 9789QA PMT whose
signal was fed into an EG&G 124A lock-in amplifier, then linearly digitized by an
Analog Devices voltage-to-frequency converter (VFC) and counted using a Tracor
129
Northern 7200 multichannel scaler (MCS). The instrument resolution of this setup was
~ 250 fs. Further improvements were made on this setup by the authors using
reflective optics to improve the instrument resolution to 100 fs.
218, 219
Bradforth et al.
in 1995 reported the rate of fluorescence depolarization in LH-1 photosynthetic
systems using fluorescence upconversion.
220, 221
The setup used the 870 nm output of
a 250 kHz regenerative amplifier and a reflective fluorescence upconversion setup.
The fluorescence was upconverted by mixing with the laser wavelength in a
noncolinear geometry at a 0.5 mm β-BaB 2O 4 (BBO) crystal. The sum frequency photons
were then directed into a double monochromator and detected by a PMT and
accumulated using a photon counter. The instrument resolution measured by tuning
the BBO to mix the gate beam and excitation beam was measured to be 160 fs.
Haacke et al. in 1998 reported first using a charge-coupled device detector (CCD)
for fluorescence upconversion.
222
The authors were able to use either a 1 mm thick
LiIO 3 or BBO crystal to achieve upconversion bandwidths as much as 150 meV in a
single CCD exposure. Low-frequency laser noise acts on the whole spectrum and can
be averaged out using multichannel CCD detection. A part of the 800 nm output of a
Ti:sapphire laser was fed into an optical parametric oscillator (OPO) to generate the
signal beam at 1.5 µm which acted as the gate beam while the remnant 800 nm was
used to excite the sample. The fluorescence was collected using parabolic mirrors
and focused on to either the BBO or LiIO 3 crystal. The detection system consisted of
the CCD (LN/CCD-1100-PB/UVAR, Princeton Instruments) mounted behind a 25-cm
130
spectrograph (250 IS, Chromex). The spectrograph was able to resolve 3-4 meV (8-9
pixels) with a 1200 lines/mm grating a slit width of 150 µm. The authors measured the
upconverted signal in one of two ways: either the nonlinear crystal is continuously
rotated while the CCD acquired spectra till phase matching for the whole spectra
range is achieved or a 1-mm thick BBO crystal which a large spectral bandwidth is
used, and the crystal need not be rotated. With the former method, the spectra
collected have to corrected by just the transfer function of the spectrometer and the
CCD while in the latter, correction using the spectral response of the BBO crystal has
be to applied to obtain the real fluorescence spectrum. Using a CCD detector provided
numerous advantages to this upconversion setup: high quantum efficiency, low
readout noise, and negligible dark counts. The multichannel detection with long
exposure times also eliminated laser noise and drift.
Significant advances in broadband fluorescence upconversion came from the
group of N. P. Ernsting at the Humboldt University of Berlin.
208, 223-227
The setup
consisted of 30 fs, 794 nm pulses at 1 kHz were split 6:1 with the former portion used
to pump an optical parametric amplifier (OPA) to deliver gate pulses at 13`00 nm. The
latter portion was frequency doubled to obtain pump pulses which were focused on
the sample – which consisted of a solution of coumarin C153 dye in methanol flowing
out of the interaction region after each laser shot. To maximize the upconversion,
fluorescence was collected over a solid angle of 1.8 sr and was imaged onto a
potassium dihydrogen phosphate crystal (KDP) with 10: 1 magnification. The gate
131
beam interacted with the fluorescence in this set up at a small angle of 3º. The
upconverted signal was focused on to the 0.4 mm input pinhole of a grating
polychromator with a dispersion of 12 nm/mm or 0.3 nm/pixel providing a resolution
of 3.5 nm. Spectra were collected in the 320-420 nm range by a CCD array (V420-BU,
Andor Technology). Research from the Ernsting group also pioneered the use of tilted
gate pulses to achieve instrument resolutions as short as 80 fs. In a typical
fluorescence upconversion set up, with appropriate geometry it is possible to
spatially separate the sum frequency photons from the ungated fluorescence
photons. When the angle between the gate and fluorescence beams at the nonlinear
crystal is increased for better spatial resolution, time resolution is lost because of the
difference in arrival times of the two outer gate rays. Pulse front tilting had been
proven to be effective at correcting this effect when laser pulses are added, or for
parametric amplification. In the Ernsting setup, the 1300 nm gate pulse is generated
by pumping a TOPAS OPA with 810 nm beam from a Ti:sapphire laser, passed through
a delay stage and then pulse front tilting is achieved using a prism. The sample
solution is excited with by a 400 nm generated by doubling the laser fundamental. An
off-axis Schwarzschild objective is used to refocus the generated fluorescence with
7-fold magnification and focused using subsequent reflective optics at the KDP
crystal. The upconverted signal is passed through a Glan polarizer to suppress the
direct fluorescence, pump scatter and gate harmonics. A concave mirror focuses the
signal light on the entrance of fiber bundle which carries the signal photon into a slit
132
at the spectrograph. Inside the spectrograph a ruled plane grating maps the spectrum
onto a CCD detector camera (DV420-BU, Andor Technology). Ernsting and coworkers
have also reported using the complementary downconversion of fluorescence along
with the upconversion to jointly achieve a 26000 cm
-1
multiplex window with 100 fs
resolution.
Chergui and coworkers implemented a broadband detection scheme for
fluorescence upconversion using a spectrometer (SpectraPro 500i, Acton Research)
equipped with a CCD camera (SDS 9000, Photometrics).
210, 228
In contrast to the
experimental setup used by Ernsting and coworkers where the KDP crystal was
stationary, here the nonlinear crystal is rotated with a constant angular speed to
achieve phase matching for the entire spectral range of interest. Using the Raman
peak of pure water, the authors were able to characterize the temporal resolution of
this upconversion setup to be 90 fs.
As described in detail in the previous chapters, the TCSPC setup in our lab consists
of an R3809U-50 Hamamatsu photomultiplier tube (PMT) with a B&H SPC-630 module
- providing a time resolution of 22 ps. Frequently samples were encountered which
have excited state dynamics faster than 22 ps that could not be resolved with the
TCSPC setup. Additionally, measuring the fluorescence lifetime of certain samples
like near-infra red absorbing dyes and inorganic quantum dots also poised a problem
because the spectral response curve of the detector PMT drops precipitously for
wavelengths > 700 nm (Figure 6.2).
133
This chapter describes an effort to circumvent the constraints of both the time-
resolution and spectral response, by building a fluorescence upconversion setup in
house. Such an apparatus built using femtosecond lasers would provide a time
resolution of ~100 fs - theoretically limited only by the cross-correlation of the gate
and the excitation pulses in an ideal setup. Realistically, the time resolution would be
also affected by the upconversion bandwidth of the nonlinear crystal and dispersion
from the optical elements in the setup like lenses. At the same time, an upconversion
setup is also a workaround using a near-infrared (NIR) detector to measure
fluorescent emission > 700 nm. Detectors that operate in the NIR tend to be expensive
Figure 6.2. The spectral response curves of the R3809U-50 series of
photomultiplier tubes from Hamamatsu (reproduced from the product datasheet).
The specific PMT used here is boxed in red in the figure.
134
to purchase and laborious to maintain because they require an external
heatsink/coolant to bring down dark counts. By upconverting a potential near-
infrared fluorescence from 800 – 1200 nm using the gate pulse (here 800 nm) to visible
wavelengths (400 – 500 nm), it is possible to use a visible detector instead of a near-
infrared detector. Another key advantage of a fluorescence upconversion over other
time-resolved fluorescence techniques is that by measuring only the upconverted
photons, upconversion is a background free technique.
6.2 MATERIALS AND METHODS
In setting up the fluorescence upconversion, it was chosen to build what is
essentially a frequency resolved optical gating setup first (Figure 6.3). This allowed us
to set up the detection apparatus, find time zero in the optical path and write Labview
Figure 6.3. Schematic of the experimental setup used to measure frequency
resolved autocorrelation of the gate beam. L1 and L2 are plano-convex lenses
(focal length 10 cm and 5 cm) used to focus the 800 nm at the nonlinear crystal. L3
represents an 800 nm dielectric mirror used to remove the 800 nm beam letting
only the 400 nm pass and a 5 cm plano-convex lens that focusses the 400 nm beam
at the entrance slit of the monochromator.
Expander/
compressor
Mira 900 RegA 9000
Verdi V18
λ/2
telescope
translation stage
BBO type 1
HV
photon counter
L1
L2
L3
135
code to automate many aspects of the experiment before moving on to the harder
challenge of carrying out sum frequency generation using fluorescence photons and
the gate beam. After time zero had been found, we replaced one arm of this set up to
generate a 400 nm excitation beam by doubling the laser fundamental (Figure 6.4).
The 400 nm is used to excite the sample (Fluorescein dye in basic ethanol) following
which the fluorescence is collected and focused on the nonlinear crystal – following
the same optical path that 800 nm beam did in the initial setup. Fluorescein was
chosen to benchmark this setup because its excited state dynamics in a variety of
solvents had been well characterized. Aligning the gate beam with the fluorescence
at the nonlinear crystal was made easier using this dye with a large quantum yield of
fluorescence.
6.2.1 Ti:sapphire Oscillator and Regenerative Amplifier
5 W of 532 nm cw beam from a Verdi V-18 (Coherent) was used to pump the
Ti:sapphire oscillator (Mira Seed, Coherent) while the remaining 13 W was used to
pump the regenerative amplifier (RegA 9050, Coherent). The 76 MHz 800 nm beam
from the oscillator was first expanded in a separate compartment and then used to
seed the amplifier. The output of which was compressed to generate 250 kHz pulsed
800 nm beam with a pulse width ~ 70 fs and bandwidth of 25 nm. This beam coming
out of the compressor cavity was slightly diverging, so a plano-concave lens and a
convex lens of short focal length were mounted on a translation stage to make a
Galilean telescope. The distance between the two lenses were adjusted to collimate
136
the beam by looking at the far field image. This 800 nm beam lands on a 50:50 beam
splitter to generate two arms. The first of which hit a corner cube gold retroreflector
(Plx Inc., beam deviation = 0.5 arc sec.) mounted on a linear translation stage that
provides a computer controlled variable delay between the two 800 nm pulses. In the
first experiment this transmitted 800 nm beam is taken after the delay stage to the
nonlinear crystal to be combined with the reflected portion (Figure 6.3). In the second
experiment the transmitted 800 nm is focused after the delay stage on a 1 mm type-
Figure 6.4. Schematic of the experimental setup used to measure fluorescence
upconversion. L1 – convex lens, 2.5 cm focal length, NLO1 – 1 mm thick type-1
BBO crystal cut at 32.5º, L2 – plano-convex lens focal length 5 cm, L3 – convex
lens focal length 5 cm, L4 – plano-convex lens focal length 5 cm, L5 – system of
collection optics formed by two plano-convex lenses of focal lengths 2.5 cm and
10 cm, NLO2 – 0.25 mm type 1 BBO crystal cut at 28.9º, L6 – combination of 400
and 800 nm dielectric mirrors to block the gate and excitation beams and only let
the 266 nm pass and a 5 cm plano-convex lens to focus the 266 nm at the entrance
slit of the monochromator.
Verdi V18
Expander/
compressor
Mira 900
RegA 9000
λ/2
telescope
translation stage
NLO2
HV
photon counter L2
L1
NLO1
L3
L4
L5
L6
sample
137
1 BBO crystal to generate the 400 nm that will be used to excite the sample (Figure
6.4).
6.2.2 Translation Stage
The translation stage used to provide the variable delay in this setup was the
refurbished UTM150PP.1 15 cm stepper drive linear translation stage from Newport
Corporation. The stage was connected via a 25-pin Sub-D connector to an ESP300
controller. The controller was connected to a PC via a RS-232 to USB connector which
allowed the stage to be programmatically moved through the Labview interface.
6.2.3 Sample Holder
The sample holder used in the experiment was a fluorimeter flow cell from Starna
Cells 3-1.30/TC which was made of quartz. The cell was connected to a peristaltic
pump to have the sample flowing throughout the volume of the cell. For measuring
the instrument resolution, the sample cell was flowed with methanol, so the gate 800
nm beam was mixed with the transmitted excitation 400 nm beam at the non linear
crystal. For the upconversion experiment the sample cell was flowed with fluorescein
in basic ethanol, so the gate 800 nm beam was mixed fluorescent photons. In both
cases, a convex lens (focal length 5 cm) focused the 400 nm beam at the center of the
flow cell while a pair of plano-convex lenses (focal lengths 2.5 cm and 10 cm) were
used to collect and parallelize the excitation/fluorescent photon emanating from the
center of the flow cell and refocus it at the nonlinear crystal.
6.2.4 Nonlinear crystal
138
A type-1 0.25 mm BBO crystal (Shalom EO, China) cut at 28.9º was used to achieve
both the doubling of the 800 nm in the first experiment and to find the instrument
resolution of the setup in the second experiment. A type-1 1 mm BBO crystal (Inrad
Optics) was used to double the 800 nm in the second experiment to generate the 400
nm excitation beam. The crystal was mounted with its optical axis vertical on a
circular stage that allowed for the phase matching angle to be varied to achieve both
the frequency doubling of the 800nm as well as the 800 + 400 → 266 nm to find the
instrument resolution. After upconversion, the remaining 800 and 400 nm photon
were blocked using dielectric mirrors while the 266 nm was focused at the entrance
slit of the monochromator using a 5 cm convex lens. All the lenses used in these
setups were made of fused silica.
6.2.5 Monochromator
A Digikrom CM112 monochromator was used to disperse the upconverted photon
in frequency before they hit the PMT. Three slits of 0.2 mm width were used at the
entrance, exit and a third position between the two grating elements to frequency
select as well as prevent stray light from entering the detection apparatus. The
monochromator was connected though an RS-232 to USC connector to the PC which
allowed for it to be controlled so as to automate the spectra collection process.
6.2.6. Photomultiplier Tube
A R928P (pre-amplified output) photomultiplier tube (Hamamatsu) was used to
detect the photons after they had been dispersed by the monochromator. It was
139
housed in a lightproof chamber (Products for Research, resourced from an old
fluorimeter) and supplied with a variable high voltage (Bertan). The optimum high
voltage to be supplied to the PMT was found by measuring the number of photons
detected at the PMT as a function of the high voltage supplied. As seen in Figure 6.5,
above 1.0 kV, the photon counts started plateauing and this was chosen to be the
operating high voltage for all the experiments.
6.2.7 Photon Counter
An SR400 gated photon counter (Stanford Research Systems) was used to count
the number of photons striking the PMT. The SR400 was setup to communicate to the
PC via a GPIB-Labview interface. The signal from the PMT is directly fed into one of
the channels of the photon counter which is set to operate in the internally triggered
Figure 6.5. The variation in number of photons detected by the PMT as a function
of the high voltage supplied to it. Note the plateauing around 1.0 kV.
140
mode using the 10 MHz internal clock. When the integration time is set to 10
-7
seconds, the counter integrates all the pulses emanating from the PMT for 1 second.
The Labview interface for the upconversion experiment is setup such that the
monochromator scans through the desired wavelength range, and at each
wavelength interval the counter counts all the signals arising from photons striking
the PMT for 1 second. In this manner, the number of photons at each wavelength is
measured and tabulated to give the spectrum of the light falling on the
monochromator entrance slit. For the high signal experiments considered here, it was
not necessary to externally trigger the photon counter. Such a provision does exist in
the SR400 which would allow for all the noise photons arriving outside of a pre-
determined time interval to be discarded. Using attenuation filters, the counts as
measured by the SR400 and that falling on the PMT were found to be in the linear
regime from 10 – 100,000 counts. It is possible that the intensity of the light or the
PMT used is such that when increasing the number of photon let through to the PMT
(by using OD filters) does not correlate linearly with the count rate measured by the
SR400 – and measurements cannot be made in this count rate regime.
6.2.8 Setting up the nonlinear crystal to achieve frequency mixing
To measure either fluorescence upconversion, instrument response (Figure 6.4),
or the frequency doubling of the 800 nm beam (Figure 6.3), the desired frequency
mixing is set up at the nonlinear crystal by finding the correct temporal, spatial
alignment and phase matching angle. To find the right temporal alignment (finding
141
time zero) of the pulses reaching the nonlinear crystal, the optical pathlength of the
two beams after they are split by the beam splitter is measured and made to be
congruent. Thereafter the beams are made to match spatially on the crystal by looking
at the beam overlap on the crystal using a microscope objective lens (watchmakers
magnifying glasses) and adjusting the turning mirrors that direct the beams on to the
crystal. The beams are also made to fall on the crystal with a small noncollinear angle
of 10º. By using apertures at near and far field, the beam delayed by the stage was
made to pass through the same spot at the crystal along the travel length of the stage.
The phase matching angle for the three frequency mixing experiments under
consideration here were found as follows:
(i) For the doubling of the 800 nm beams, it was necessary and sufficient just
to achieve spatial and temporal overlap of the two 800 nm beams at the
crystal because the crystal was cut at 28.9º to its optic axis – which is the
phase matching angle for 800 nm doubling.
(ii) For the upconversion experiment as well as for finding the instrument
resolution, first the optic axis of the crystal was found with respect to the
lab frame coordinates i.e. with respect to the circular rotating mount (with
polar angles marked on it) on which the crystal is mounted. This was found
by letting only the gate 800 nm beam fall the crystal and rotating just the
crystal by hand (but not rotating the mount) to achieve doubling just the 800
nm by itself. This resulted in the crystal being perpendicular to the 800 nm
142
beam because it is cut at 28.9º to its optics axis. Once the frequency
doubling is achieved, the position of the optic axis with respect to the lab
frame coordinates i.e. markings on the circular mount is known – it will be
at 28.9º from the beam direction. Once we know the position of the crystal
optic axis, the circular mount is rotated such that the beam is falling on the
crystal at that angle to the optic axis which corresponds to the phase
matching angle for either the fluorescence upconversion or measuring the
instrument response – as measured using the SNLO software.
229
Following
this, both the gate and fluorescence/excitation beams are allowed to hit the
crystal at the right phase matching angle. To optimize the frequency mixing
coarsely, the light coming out of the crystal is analyzed a modular
spectrometer (Ocean Optics) and the upconversion signal is maximized
adjusting the phase matching angle, beam focusing, crystal positioning and
the optical delay.
After optimizing the frequency mixing, for final data collection the beam
emanating from the crystal was collected using a 5 cm plano-convex lens. In
measuring the instrument response (Figure 6.4), this lens was followed by a 800
nm dielectric mirror and 400 nm dielectric mirror to remove the remnant gate and
excitation beams while the frequency mixed 266 nm passed through and was
focused on the entrance slit of the monochromator using a 5 cm convex lens. In
measuring the doubling of the two 800 nm beams, only the 800 dielectric mirror
143
was used (Figure 6.3) while the frequency mixed 400 nm could pass through to the
detection apparatus.
6.2.9 Labview Interface and Operation Schema
Measuring the upconversion spectra or the instrument response using this
setup first requires that correct spatial overlap, temporal overlap (time-zero), and
nonlinear crystal phase matching angle be found manually and optimized. Then the
Labview program can be used to automate the process of collecting spectra at each
delay point. The following images (Figures 6.6, 6.7, 6.8, 6.9) will be a guide through the
different step involved in running the Labview program correctly.
6.2.10 Aligning the upconverted beam on the detection apparatus
To begin, ensure that the PC is turned on, along with the high voltage supply to
the PMT (but set to 0.0 kV), the ESP controller to the stage, the SR400 photon counter
and the monochromator. Check the respective communication ports that connect the
devices to the PC: the GPIB cable for the SR400, and the RS-232 to USB cables for the
monochromator and the ESP controller. Once the upconversion signal has been found
and optimized, switch all the room lights off – this can over load the PMT if the HV is
turned on while the room lights are on. Use a turning mirror to direct the upconverted
signal to the entrance slit of the monochromator. Set the high voltage supply to the
PMT to be 1.0 kV – the photon counter should start to pick up dark counts in the order
of 5 – 10 now. Click to open the Labview program – currently labelled
“STAGE+mono+counter_edit26.vi” and run it (there are no dependent custom written
144
sub-vi’s for this program, but the vi’s that are part of the various instruments like the
monochromator, photon counter and the stage must be present in the PC). On the RT
(real time data) tab (Figure 6.6), click on “Show Real Time Data” button to switch it on.
This function was incorporated into the program to show the real time counts read by
the photon counter from the PMT. Set the wavelength to either 400 (for measuring the
800+800 nm doubling) or 266 (to measure the instrument response 400+ 800 nm) and
click enter. The Photon Counts dial and display should start showing the photons
falling on the detector in real time. Now the turning mirror can be used to align the
upconverted signal on the entrance slit of the monochromator by looking for a
maximum in the number of photons detected. After aligning using this function, click
“Show Real Time Data” again to turn it off – failure to do so will cause the program to
crash when attempting to use other functions.
Please read this note before proceeding further.
230
6.2.11 Measuring the upconversion spectra
Start by setting the required parameters in the set of the tabs on the left side.
The first of these tabs is labelled “DELAY” (Figure 6.7) and contains all parameters
necessary for the translation stage.
- Select the correct axis on the ESP controller to which the stage is attached
in Axis.
- Make sure the motor is on by clicking on Motor ON.
- The Target[mm], Move and “Set this as t0” options have not be incorporated
145
correctly so far.
- Set the Initial position [mm] of the stage as zero. Manually move the
retroreflector on the stage using the ESP controller to find out when the
upconverted signal disappears – this is the final position.
- Set the number of Data Points and calculate and set the Step Side [µm] by
taking the difference in initial and final positions and dividing it by the
number of data points.
- In the Data File Path entry, click on the folder icon, navigate to the
appropriate folder and create and empty text file. The Labview is set to
update this file with the photon counts at each wavelength point for each
delay point immediately after is recorded so that even if the program
crashes in between an experiment, the data will not be lost.
In the COUNTER tab (Figure 6.8), set the parameters for the photon counter as follows:
- Count Mode: A, B for Preset T
- Trigger mode: 10 MHz
- Channel A Input: Input 1 - this should be the signal from the PMT
- Number of measurements to take at one point: 1
- Dwell Time: 0.002
- Gate mode: CW
- Integrating time: 10
7
– this is measured in units of 10
-7
seconds, so
integrating time comes out to be 1 second
146
- Discriminator Slope: Fall
- Discriminator Level (mV) – can be used to exclude dark counts, this value
would depend on how light proof the room is
In the MONO tab (Figure 6.9), set the parameters for the monochromator as follows:
- Mono port: COMxxx – set this value to read the right communication serial
port number
- Time delay: 0
- NEW UNIT: choose the required units for the wavelength between
centimicrons, Angstroms or nanometers
- Scan from: set to start wavelength
-
Figure 6.6. The real time (RT) tab in the Labview interface that can be used to
align the upconverted signal to the entrance slit of the monochromator.
147
- Scan to: set to end wavelength
- Step size: set to required value
- Current wavelength: displays the current position of the grating
Figure 6.7. Settings for the translation stage in the DELAY tab.
Figure 6.8. Settings for the translation stage in the COUNTER tab.
148
Figure 6.9. Settings for the translation stage in the MONO tab.
Figure 6.10. Settings for the translation stage in the PLOT tab.
149
- Groves/mm, Blaze, Grating #: displays properties of the current grating
- Change Grating: helps switch between the two gratings in the CM112
The PLOT tab as shown in the Figure 6.10 can be used to input target wavelengths to
monitor while running the program. The photon counts at these wavelengths at each
delay point will be used to build time slices in the graph titled Decay in the Scan Data
tab on the right side. The graph title Spectrum in the same tab is programed to show
the spectra currently being measured. In the Scans tab on the right side, all the
previously accumulated spectra will be displayed. Once all the parameters have been
set, in the Scan Data tab click on the Collect Data! button to start data collection.
Figure 6.11. Spectra of the 400 nm generated by mixing the 800 nm beam with
itself, measured as function of delay between the two pulses.
150
6.3 RESULTS AND DISCUSSION
6.3.1 Frequency Resolved Optical Gating with 800 nm
The first setup built as seen in Figure 6.3 involved mixing the 800 nm with itself
after being delayed using the translation stage. The resultant 400 nm spectra were
collected using the detector apparatus as described above. Setting up essentially
what is a Χ
(2)
frequency resolved optical gating experiment with the 800 nm was
performed to match optical path lengths, optimize detection apparatus, and program
and test the Labview code for the upconversion experiment. As seen in Figure 6.4, the
Figure 6.12. Instrument response profile of the upconversion setup characterized
by measuring the spectra of the 266 nm generated by mixing the 800 nm gate beam
with the 400 nm excitation beam, measured as function of delay between the two
pulses.
151
pulse has a positive chirp – which can be attributed to the material dispersion caused
by the lenses the beam passes through before the nonlinear crystal. Although running
the complete phase retrieval algorithm
231, 232
that is used on FROG traces here is
beyond the scope of this experiment, by fitting the time slice at 393 nmz to a gaussian,
the temporal width of the 800 nm pulse is found to 130 fs.
6.3.2 Measuring the instrument response
To characterize the instrument response of the upconversion setup, the 400 nm
excitation beam was made to pass through the flow cell containing just the solvent
methanol and then mix with the 800 nm gate beam at the nonlinear crystal. The
resultant 266 nm beam is directed toward the monochromator and its spectra is
measured as function of the time delay between the two pulses (Figure 6.6). The
spectra at maximum temporal overlap was fit with a gaussian, according to methods
established in the literature.
225
The time resolution of the instrument so obtained is
Figure 6.13. Time slices from time-resolved spectra obtained by (a) mixing the
800 nm with itself (taken at 393 nm) and (b) mixing the 800 nm with 400 nm
(taken at 263 nm). Pulse width in the former is 130 fs and in the latter is 80 fs.
152
80 fs (Figure 6.13) which is comparable to that of instruments reported in the
literature.
223, 227, 228
6.4 CONCLUSIONS
In conclusion, chapter 6 describes how a fluorescence upconversion setup was
built in house. The time resolved fluorescence apparatus that was used for the
experiments described in the previous chapters was a time-correlated single photon
counting device which suffered from two major setbacks: (i) the time resolution of the
TCSPC device was 22 ps which kept faster singlet state kinetics from being resolved
and (ii) the detector efficiency dropped after 700 nm which prevented near-infrared
emitting species from being studied. Routinely chemical entities like organic
molecules undergoing sub-picosecond processes like singlet fission and symmetry
breaking charge transfer were encountered whose kinetics could not be studied using
the TCSPC setup. Furthermore, nanoparticles like those made of lead selenide and
cadmium sulfide exhibit size-dependent band gaps – larger particles often had
emission spectra that crossed over in to the NIR region and could not be probed using
the TCSPC setup either.
The femtosecond fluorescence upconversion setup described in this chapter
overcomes both these shortcomings. The instrument resolution of the setup was
measured to be 80 fs, which would be short enough to probe fast processes like
singlet fission and charge transfer. Moreover, being based on frequency mixing in a
nonlinear medium, upconversion can take NIR emission like those from large
153
nanoparticles and move them to the UV-Vis range where we have detection
apparatuses readily available. Although this chapter does not contain experimental
data from molecules we were interested in to study using upconversion, it describes
how to build the experimental setup from scratch, and also contains detailed
instructions on how to use the Labview program that was written to automate the
whole experiment – imparting the setup broadband capabilities.
154
BIBILIOGRAPHY
1. (IEA), I. E. A. World Energy Outlook; 2019.
2. Rueter, G., Is the IEA underestimating renewables? Deutsche Welle: 2018.
3. Evans, S., ‘Profound shifts’ underway in energy system, says IEA World Energy
Outlook. Carbon Brief: 2018.
4. C&EN Global Enterprise 2016, 94 (18), 30-35.
5. Beery, D.; Wheeler, J. P.; Arcidiacono, A.; Hanson, K., ACS Applied Energy
Materials 2019.
6. C&EN Global Enterprise 2019, 97 (28), 3-3.
7. Hetzer, C.; Guldi Dirk, M.; Tykwinski Rik, R., Chemistry - A European Journal
2018, 24, 1-14.
8. C&EN Global Enterprise 2018, 96 (33), 8-8.
9. Ueno, K.; Oshikiri, T.; Sun, Q.; Shi, X.; Misawa, H., Chemical Reviews 2018, 118
(6), 2955-2993.
10. Trinh, M. T.; Pinkard, A.; Pun, A. B.; Sanders, S. N.; Kumarasamy, E.; Sfeir, M. Y.;
Campos, L. M.; Roy, X.; Zhu, X. Y., Science Advances 2017, 3 (7).
11. Green, M. A.; Emery, K.; Hishikawa, Y.; Warta, W.; Dunlop, E. D., Progress in
Photovoltaics: Research and Applications 2016, 24 (NREL/JA-5J00-65643).
12. Ponseca, C. S.; Chábera, P.; Uhlig, J.; Persson, P.; Sundström, V., Chemical
Reviews 2017, 117 (16), 10940-11024.
13. Choi, H.; Kim, S.; Luther, J. M.; Kim, S.-W.; Shin, D.; Beard, M. C.; Jeong, S.,
Chemistry – A European Journal 2017, 23 (70), 17625-17625-17625.
14. Kahmann, S.; Mura, A.; Protesescu, L.; Kovalenko, M. V.; Brabec, C. J.; Loi, M. A.,
J. Mater. Chem. C 2015, 3 (21), 5499-5505-5505.
15. Schrauben, J. N.; Zhao, Y.; Mercado, C.; Dron, P. I.; Ryerson, J. L.; Michl, J.; Zhu,
K.; Johnson, J. C., ACS Applied Materials & Interfaces 2015, 7 (4), 2286-2293-
2293.
16. Couderc, E.; Greaney, M. J.; Brutchey, R. L.; Bradforth, S. E., Journal of the
American Chemical Society 2013, 135 (49), 18418-26-26.
17. Dennler, G.; Scharber, M. C.; Brabec, C. J., Adv Mater 21 2009 SRC -
BaiduScholar, 1323-1338.
18. Dennler, G.; Scharber, M. C.; Brabec, C. J., Adv Mater 21 2009 SRC -
BaiduScholar, 1323-1338-1338.
19. Radychev, N.; Lokteva, I.; Witt, F.; Kolny-Olesiak, J.; Borchert, H.; Parisi, J.; T.,
H., J Phys Chem C 115 2011 SRC - BaiduScholar, 14111-14122.
20. Greaney, M. J.; Das, S.; Webber, D. H.; Bradforth, S. E.; Brutchey, R. L.; S., A. C.,
6 2012 SRC - BaiduScholar, 4222-4230.
21. Cate, S.; Schins, J. M.; Siebbeles, L. D. A.; Acs, 6 2012 SRC - BaiduScholar, 8983-
8988-8988.
155
22. Günes, S.; Neugebauer, H.; Sariciftci, N. S., Chemical Reviews 2007, 107 (4),
1324-1338.
23. Tamai, Y.; Ohkita, H.; Benten, H.; Ito, S., The Journal of Physical Chemistry
Letters 2015, 6 (17), 3417-3428.
24. Basel, B. S.; Zirzlmeier, J.; Hetzer, C.; Reddy, S. R.; Phelan, B. T.; Krzyaniak, M.
D.; Volland, M. K.; Coto, P. B.; Young, R. M.; Clark, T.; Thoss, M.; Tykwinski, R. R.;
Wasielewski, M. R.; Guldi, D. M., Chem 2018, 4 (5), 1092-1111.
25. Hart, S.; Silva, W. R.; Frontiera, R. R., Chemical Science 2018, 9 (5), 1242-1250.
26. Mauck Catherine, M.; Bae Youn, J.; Chen, M.; Powers ‐Riggs, N.; Wu, Y. L.;
Wasielewski Michael, R., ChemPhotoChem 2017, 2 (3), 223-233.
27. Sung, J.; Nowak-Kro ́l, A.; Schlosser, F.; Fimmel, B.; Kim, W.; Kim, D.; Wu ̈rthner,
F., Journal of the American Chemical Society 2016, 138 (29), 9029-9032-9032.
28. Zimmerman, P. M.; Bell, F.; Casanova, D.; Head-Gordon, M., Journal of the
American Chemical Society 2011, 133 (49), 19944-19952-19952.
29. Mataga, N.; Yao, H.; Okada, T.; Rettig, W., The Journal of Physical Chemistry
1989, 93 (9), 3383-3386.
30. Pensack, R. D.; Tilley, A. J.; Parkin, S. R.; Lee, T. S.; Payne, M. M.; Gao, D.; Jahnke,
A. A.; Oblinsky, D. G.; Li, P.-F.; Anthony, J. E.; Seferos, D. S.; Scholes, G. D.,
Journal of the American Chemical Society 137 (21), 6790-6803-6803.
31. McGehee, M. D., MRS Bulletin 2009, 34 (2), 95-100-100.
32. Rolczynski, B. S.; Szarko, J. M.; Son, H. J.; Liang, Y.; Yu, L.; Chen, L. X., Journal
of the American Chemical Society 2012, 134 (9), 4142-4152-4152.
33. Thompson, B. C.; Fréchet, J. M. J., Angewandte Chemie International Edition
2008, 47 (1), 58-77-77.
34. Geacintov, N.; Pope, M.; Vogel, F., Physical Review Letters 1969, 22 (12), 593.
35. Moller, W. M.; Pope, M., The Journal of Chemical Physics 1973, 59 (5), 2760-
2761.
36. Müller, A. M.; Avlasevich, Y. S.; Müllen, K.; Bardeen, C. J., Chemical physics
letters 2006, 421 (4), 518-522.
37. Müller, A. M.; Avlasevich, Y. S.; Schoeller, W. W.; Müllen, K.; Bardeen, C. J.,
Journal of the American Chemical Society 2007, 129 (46), 14240-14250.
38. Smith, M. B.; Michl, J., Chemical Reviews 2010, 110 (11), 6891-6936.
39. Felter, K. M.; Grozema, F. C., The Journal of Physical Chemistry Letters 2019,
7208-7214.
40. Abraham, V.; Mayhall, N. J., The Journal of Physical Chemistry Letters 2017.
41. Breen, I.; Tempelaar, R.; Bizimana, L. A.; Kloss, B.; Reichman, D. R.; Turner, D.
B., Journal of the American Chemical Society 2017, 139 (34), 11745-11751.
42. Tayebjee, M. J. Y.; Sanders, S. N.; Kumarasamy, E.; Campos, L. M.; Sfeir, M. Y.;
McCamey, D. R., Nature Physics 2016, 13, 182.
43. Wang, X.; Wang, R.; Shen, L.; Tang, Z.; Wen, C.; Dong, B.; Liu, H.; Zhang, C.; Li, X.,
Physical Chemistry Chemical Physics 2018, 20 (9), 6330-6336.
156
44. Yamakado, T.; Takahashi, S.; Watanabe, K.; Matsumoto, Y.; Osuka, A.; Saito, S.,
Angewandte Chemie International Edition 2018, 57 (19), 5438-5443-5443.
45. Joy, J.; Korovina, N. V.; Thompson, M. E.; Bradforth, S. E., Proc. SPIE 10348,
Physical Chemistry of Semiconductor Materials and Interfaces XVI 2017.
46. Bettinger, H. F.; Einholz, R.; Gottler, A.; Junge, M.; Sattele, M.-S.; Schnepf, A.;
Schrenk, C.; Schundelmeier, S.; Speiser, B., Organic Chemistry Frontiers 2017,
4 (5), 853-860.
47. Berkelbach, T. C.; Hybertsen, M. S.; Reichman, D. R., The Journal of Chemical
Physics 2013, 138 (11), 114103.
48. Johnson, J. C.; Nozik, A. J.; Michl, J., Accounts of Chemical Research 2013, 46
(6), 1290-1299-1299.
49. Monahan, N.; Zhu, X. Y., Annual Review of Physical Chemistry 2015, 66 (1), 601-
618.
50. Schrauben, J. N.; Akdag, A.; Wen, J.; Havlas, Z.; Ryerson, J. L.; Smith, M. B.;
Michl, J.; Johnson, J. C., The Journal of Physical Chemistry A 2016, 120 (20),
3473-3483-3483.
51. Chen, X.; Larsen, D. S.; Bradforth, S. E.; van Stokkum, I. H. M., The Journal of
Physical Chemistry A 2011, 115 (16), 3807-3819.
52. Zhang, Y. TWO-PHOTON ABSORPTION SPECTROSCOPY AND EXCITED STATE
PHOTOCHEMISTRY OF SMALL MOLECULES University of Southern California,
Los Angeles, 2012.
53. Kloepfer, J. A.; Bradforth, S. E.; Nadeau, J. L., The Journal of Physical Chemistry
B 2005, 109 (20), 9996-10003.
54. Hewavitharana, I. K.; Brock, S. L., ACS nano 2017, 11 (11), 11217-11224.
55. Morris-Cohen, A. J.; Malicki, M.; Peterson, M. D.; Slavin, J. W. J.; Weiss, E. A.,
Chem Mater 25 2013 SRC - BaiduScholar, 1155-1165.
56. Moreels, I.; Lambert, K.; Muynck, D.; Vanhaecke, F.; Poelman, D.; Martins, J. C.;
Allan, G.; Hens, Z., Chem Mater 19 2007 SRC - BaiduScholar, 6101-6106-6106.
57. Roo, J.; Ibanez, M.; Geiregat, P.; Nedelcu, G.; Walravens, W.; Maes, J.; Martins,
J. C.; Driessche, I.; Kovalenko, M. V.; Henz, Z.; S., A. C., 10 2016 SRC -
BaiduScholar, 2071-2081.
58. Smock, S. R.; Williams, T. J.; Brutchey, R. L., Quantifying the Thermodynamics
of Ligand Binding to CsPbBr3 Quantum Dots. Vol. 2018 SRC - BaiduScholar, p
11711-11715-11715.
59. Bullen, C.; Mulvaney, P., Langmuir 22 2006 SRC - BaiduScholar, 3007-3013.
60. Zabet-Khosousi, A.; Dhirani, A.-A., Chemical reviews 2008, 108 (10), 4072-124.
61. Frederick, M. T.; Weiss, E. A., ACS nano 2010, 4 (6), 3195-200.
62. Frederick, M. T.; Amin, V. A.; Swenson, N. K.; Ho, A. Y.; Weiss, E. A., Nano Lett 13
2013 SRC - BaiduScholar, 287-292.
63. Buckley, J. J.; Couderc, E.; Greaney, M. J.; Munteanu, J.; Riche, C. T.; Bradforth,
S. E.; Brutchey, R. L., ACS Nano 2014, 8 (3), 2512-2521-2521.
157
64. Greaney, M. J.; Couderc, E.; Zhao, J.; Nail, B. A.; Mecklenburg, M.; Thornbury,
W.; Osterloh, F. E.; Bradforth, S. E.; Brutchey, R. L., Chem Mater 27 2015 SRC -
BaiduScholar, 744-756.
65. Nag, A.; Kovalenko, M. V.; Lee, J.-S.; Liu, W.; Spokoyny, B.; Talapin, D. V., J. Am.
Chem. Soc. 2011, 133 (27), 10612-10620-10620.
66. Tang, J.; Kemp, K. W.; Hoogland, S.; Jeong, K. S.; Liu, H.; Levina, L.; Furukawa,
M.; Wang, X.; Debnath, R.; Cha, D.; Chou, K. W.; Fischer, A.; Amassian, A.; Asbury,
J. B.; Sargent, E. H., Nature Materials 2011, 10 (10), 765.
67. Kroupa, D. M.; Aria, D. H.; Blackburn, J. L.; Carroll, G. M.; Granger, D. B.; Anthony,
J. E.; Beard, M. C.; Johnson, J. C.; S., Nano Lett 18 2018 SRC - BaiduScholar,
865-873.
68. Greenham; Peng; Alivisatos, Physical review. B, Condensed matter 1996, 54 (24),
17628-17637.
69. Zhou, Y.; Riehle, F. S.; Yuan, Y.; Schleiermacher, H.-F.; Niggemann, M.; Urban,
G. A.; Krüger, M., Appl. Phys. Lett. 2010, 96 (1), 013304.
70. Casanova, D., Chemical Reviews 2018.
71. Lokteva, I.; Radychev, N.; Witt, F.; Borchert, H.; Parisi, J.; Kolny-Olesiak, J., J
Phys Chem C 114 2010 SRC - BaiduScholar, 12784-12791.
72. Olson, J. D.; Gray, G. P.; Carter, S. A., Solar Energy Materials and Solar Cells
2009, 93 (4), 519-523-523.
73. Greaney, M. J.; Araujo, J.; Burkhart, B.; Thompson, B. C.; Brutchey, R. L.,
Chemical communications (Cambridge, England) 2013, 49 (77), 8602-4.
74. Couderc, E.; Greaney, M. J.; Brutchey, R. L.; Bradforth, S. E., Journal of the
American Chemical Society 2013, 135 (49), 18418-26.
75. Morgenstern, F. S. F.; Rao, A.; Böhm, M. L.; Kist, R. J. P.; Vaynzof, Y.; Greenham,
N. C., ACS nano 2014, 8 (2), 1647-54-54.
76. Li, L.; Greaney, M. J.; Li, K.; Sachdeva, A.; Brutchey, R. L.; Campbell, J. C.,
physica status solidi (b) 2015, 252 (6), 1325-1333-1333.
77. Greaney, M. J.; Brutchey, R. L., 2014 SRC - BaiduScholar, 31-38.
78. Fu, W.; Wang, L.; Ling, J.; Li, H.; Shi, M.; Xue, J.; Chen, H., Nanoscale 2014, 6 (18),
10545-10550-10550.
79. Lek, J. Y.; Xi, L.; Kardynal, B. E.; Wong, L. H.; Lam, Y. M., ACS applied materials
& interfaces 2011, 3 (2), 287-92.
80. Lokteva, I.; Radychev, N.; Witt, F.; Borchert, H.; Parisi, J.; Kolny-Olesiak, J., J
Phys Chem C 114 2010 SRC - BaiduScholar, 12784-12791-12791.
81. Zhou, Y.; Eck, M.; Krüger, M., Energy Environ. Sci. 2010, 3 (12), 1851-1864-1864.
82. Chen, O.; Chen, X.; Yang, Y.; Lynch, J.; Wu, H.; Zhuang, J.; Cao, C., Synthesis of
Metal-Selenide Nanocrystals Using Selenium Dioxide as the Selenium
Precursor. Vol. 2008 SRC - BaiduScholar, p 8638-8641-8641.
83. Berberan-Santos, M. N.; Bodunov, E. N.; Valeur, B., Chem Phys 317 2005 SRC -
BaiduScholar, 57-62.
158
84. Bloom, B. P.; Zhao, L.-B.; Wang, Y.; Waldeck, D. H.; Liu, R.; Zhang, P.; Beratan,
D. N., J. Phys. Chem. C 2013, 117 (43), 22401-22411-22411.
85. Albero, J.; Martínez-Ferrero, E.; Iacopino, D.; Vidal-Ferran, A.; Palomares, E.,
Physical chemistry chemical physics : PCCP 2010, 12 (40), 13047-51.
86. Jasieniak, J.; Califano, M.; Watkins, S. E., ACS nano 2011, 5 (7), 5888-902.
87. Munro, A. M.; Zacher, B.; Graham, A.; Armstrong, N. R., ACS applied materials &
interfaces 2010, 2 (3), 863-9.
88. Soreni-Harari, M.; Yaacobi-Gross, N.; Steiner, D.; Aharoni, A.; Banin, U.; Millo,
O.; Tessler, N., Nano letters 2008, 8 (2), 678-84.
89. Li, Y. C.; Zhong, H. Z.; Li, R.; Zhou, Y.; Yang, C. H.; Li, Y. F., Advanced Functional
Materials 2006, 16 (13), 1705-1716-1716.
90. Bard, A. J.; Faulkner, L. R.; John, A., 2001.
91. Cardona, C. M.; Li, W.; Kaifer, A. E.; Stockdale, D.; Bazan, G. C., Advanced
Materials 2011, 23 (20), 2367-2371-2371.
92. Street, R. A.; Hawks, S. A.; Khlyabich, P. P.; Li, G.; Schwartz, B. J.; Thompson, B.
C.; Yang, Y., J Phys Chem C 118 2014 SRC - BaiduScholar, 21873-21883.
93. Roncali, J., Acc. Chem. Res. 2009, 42 (11), 1719-1730-1730.
94. Liu, C.; Yi, C.; Wang, K.; Yang, Y.; Bhatta, R. S.; Tsige, M.; Xiao, S.; Gong, X., ACS
Appl. Mater. Interfaces 2015, 7 (8), 4928-4935-4935.
95. Chen, J.-D.; Cui, C.; Li, Y.-Q.; Zhou, L.; Ou, Q.-D.; Li, C.; Li, Y.; Tang, J.-X.,
Advanced Materials 2015, 27 (6), 1035-1041-1041.
96. He, Z.; Xiao, B.; Liu, F.; Wu, H.; Yang, Y.; Xiao, S.; Wang, C.; Russell, T. P.; Cao, Y.,
Nature Photonics 2015, 9 (3), 174.
97. Zhou, Y.; Ding, L.; Shi, K.; Dai, Y.-Z.; Ai, N.; Wang, J.; Pei, J., Advanced Materials
2012, 24 (7), 957-961-961.
98. Facchetti, A., Materials Today 2013, 16 (4), 123-132-132.
99. Milliron, D. J.; Gur, I.; Alivisatos, A. P., MRS Bulletin 2005, 30 (1), 41-44-44.
100. Borchert, H., Energy Environ. Sci. 2010, 3 (11), 1682-1694-1694.
101. Lokteva, I.; Radychev, N.; Witt, F.; Borchert, H.; Parisi, J.; Kolny-Olesiak, J., J.
Phys. Chem. C 2010, 114 (29), 12784-12791-12791.
102. Fu, W.; Wang, L.; Zhang, Y.; Ma, R.; Zuo, L.; Mai, J.; Lau, T.-K.; Du, S.; Lu, X.; Shi,
M.; Li, H.; Chen, H., ACS Appl. Mater. Interfaces 2014, 6 (21), 19154-19160-
19160.
103. Zhou, R.; Stalder, R.; Xie, D.; Cao, W.; Zheng, Y.; Yang, Y.; Plaisant, M.; Holloway,
P. H.; Schanze, K. S.; Reynolds, J. R.; Xue, J., ACS Nano 2013, 7 (6), 4846-4854-
4854.
104. Huynh, W. U.; Dittmer, J. J.; Alivisatos, A. P., Science 2002, 295 (5564), 2425.
105. Jeltsch, K. F.; Schädel, M.; Bonekamp, J.-B.; Niyamakom, P.; Rauscher, F.;
Lademann, H. W. A.; Dumsch, I.; Allard, S.; Scherf, U.; Meerholz, K., Advanced
Functional Materials 2012, 22 (2).
159
106. Ren, S.; Chang, L.-Y.; Lim, S.-K.; Zhao, J.; Smith, M.; Zhao, N.; Bulovi ć, V.;
Bawendi, M.; Grade čak, S., Nano Lett. 2011, 11 (9), 3998-4002-4002.
107. Grancini, G.; Biasiucci, M.; Mastria, R.; Scotognella, F.; Tassone, F.; Polli, D.;
Gigli, G.; Lanzani, G., J. Phys. Chem. Lett. 2012, 3 (4), 517-523-523.
108. Cappel, U. B.; Dowland, S. A.; Reynolds, L. X.; Dimitrov, S.; Haque, S. A., J. Phys.
Chem. Lett. 2013, 4 (24), 4253-4257-4257.
109. Colbert, A. E.; Wu, W.; Janke, E. M.; Ma, F.; Ginger, D. S., J. Phys. Chem. C 2015,
119 (44), 24733-24739-24739.
110. Strein, E.; Colbert, A.; Subramaniyan, S.; Nagaoka, H.; Schlenker, C. W.; Janke,
E.; Jenekhe, S. A.; Ginger, D. S., Energy Environ. Sci. 2013, 6 (3), 769-775-775.
111. Lu, H.; Bartynski, A. N.; Greaney, M. J.; Thompson, M. E.; Brutchey, R. L., ACS
Appl. Mater. Interfaces 2014, 6 (20), 18306-18311-18311.
112. Chen, Z.; Zhang, H.; Yu, W.; Li, Z.; Hou, J.; Wei, H.; Yang, B., Advanced Energy
Materials 2013, 3 (4), 433-437-437.
113. Chen, Z.; Zhang, H.; Du, X.; Cheng, X.; Chen, X.; Jiang, Y.; Yang, B., Energy
Environ. Sci. 2013, 6 (5), 1597-1603-1603.
114. Yang, J.; Qian, L.; Zhou, R.; Zheng, Y.; Tang, A.; Holloway, P. H.; Xue, J., Journal
of Applied Physics 2012, 111 (4), 044323.
115. Krauss, T. D.; Brus, L. E., Materials Science and Engineering: B 2000, 69-70, 289-
294-294.
116. Gregg, B. A.; Hanna, M. C., Journal of Applied Physics 2003, 93 (6), 3605-3614-
3614.
117. Moreels, I.; Lambert, K.; Smeets, D.; Muynck, D. D.; Nollet, T.; Martins, J. C.;
Vanhaecke, F.; Vantomme, A.; Delerue, C.; Allan, G.; Hens, Z., ACS Nano 2009, 3
(10), 3023-3030-3030.
118. Noone, K. M.; Strein, E.; Anderson, N. C.; Wu, P.-T.; Jenekhe, S. A.; Ginger, D. S.,
Nano Lett. 2010, 10 (7), 2635-2639-2639.
119. Seo, J.; Cho, M. J.; Lee, D.; Cartwright, A. N.; Prasad, P. N., Advanced Materials
2011, 23 (34), 3984-3988-3988.
120. Liu, Z.; Sun, Y.; Yuan, J.; Wei, H.; Huang, X.; Han, L.; Wang, W.; Wang, H.; Ma, W.,
Advanced Materials 2013, 25 (40), 5772-5778-5778.
121. Itskos, G.; Papagiorgis, P.; Tsokkou, D.; Othonos, A.; Hermerschmidt, F.;
Economopoulos, S. P.; Yarema, M.; Heiss, W.; Choulis, S., Advanced Energy
Materials 2013, 3 (11), 1490-1499-1499.
122. Giansante, C.; Mastria, R.; Lerario, G.; Moretti, L.; Kriegel, I.; Scotognella, F.;
Lanzani, G.; Carallo, S.; Esposito, M.; Biasiucci, M.; Rizzo, A.; Gigli, G., Advanced
Functional Materials 2015, 25 (1), 111-119-119.
123. Giansante, C.; Infante, I.; Fabiano, E.; Grisorio, R.; Suranna, G. P.; Gigli, G., J. Am.
Chem. Soc. 2015, 137 (5), 1875-1886-1886.
124. Müller, A. M.; Avlasevich, Y. S.; Schoeller, W. W.; Müllen, K.; Bardeen, C. J.,
Journal of the American Chemical Society 2007, 129 (46), 14240-14250.
160
125. Thon, S. M.; Ip, A. H.; Voznyy, O.; Levina, L.; Kemp, K. W.; Carey, G. H.; Masala,
S.; Sargent, E. H., ACS Nano 2013, 7 (9), 7680-7688-7688.
126. Chuang, C.-H. M.; Brown, P. R.; Bulovi ć, V.; Bawendi, M. G., Nature Materials
2014, 13 (8), 796.
127. Kim, S.; Noh, J.; Choi, H.; Ha, H.; Song, J. H.; Shim, H. C.; Jang, J.; Beard, M. C.;
Jeong, S., J. Phys. Chem. Lett. 2014, 5 (22), 4002-4007-4007.
128. Ip, A. H.; Thon, S. M.; Hoogland, S.; Voznyy, O.; Zhitomirsky, D.; Debnath, R.;
Levina, L.; Rollny, L. R.; Carey, G. H.; Fischer, A.; Kemp, K. W.; Kramer, I. J.; Ning,
Z.; Labelle, A. J.; Chou, K. W.; Amassian, A.; Sargent, E. H., Nature
Nanotechnology 2012, 7 (9), 577.
129. Hines, M. A.; Scholes, G. D., Advanced Materials 2003, 15 (21), 1844-1849-1849.
130. Feng, X.; Casanova, D.; Krylov, A. I., The Journal of Physical Chemistry C 2016,
120 (34), 19070-19077.
131. Porter, V. J.; Geyer, S.; Halpert, J. E.; Kastner, M. A.; Bawendi, M. G., J. Phys.
Chem. C 2008, 112 (7), 2308-2316-2316.
132. Morris-Cohen, A. J.; Malicki, M.; Peterson, M. D.; Slavin, J. W. J.; Weiss, E. A.,
Chem. Mater. 2013, 25 (8), 1155-1165-1165.
133. Webber, D. H.; Brutchey, R. L., J. Am. Chem. Soc. 2012, 134 (2), 1085-1092-1092.
134. Feng, X.; Luzanov, A. V.; Krylov, A. I., The Journal of Physical Chemistry Letters
2013, 4 (22), 3845-3852-3852.
135. Müller, A. M.; Avlasevich, Y. S.; Müllen, K.; Bardeen, C. J., Chemical Physics
Letters 2006, 421 (4), 518-522-522.
136. Lu, H.; Joy, J.; Gaspar, R. L.; Bradforth, S. E.; Brutchey, R. L., Chemistry of
Materials 2016, 28 (6), 1897-1906-1906.
137. Grabowski, Z. R.; Rotkiewicz, K.; Rettig, W., Chemical Reviews 2003, 103 (10),
3899-4032-4032.
138. Couderc, E.; Greaney, M. J.; Thornbury, W.; Brutchey, R. L.; Bradforth, S. E.,
Journal of Photonics for Energy 2015, 5 (1), 1-10-10.
139. Porter, V. J.; Geyer, S.; Halper, J. E.; Kastner, M. A.; Bawendi, M. G., J Phys Chem
C 112 2008 SRC - BaiduScholar, 2308-2316-2316.
140. Yamamoto, S.; Ohkita, H.; Benten, H.; Ito, S., The Journal of Physical Chemistry
C 2012, 116 (28), 14804-14810.
141. Hwang, I. W.; Soci, C.; Moses, D.; Zhu, Z.; Waller, D.; Gaudiana, R.; Brabec, C. J.;
Heeger, A. J., Advanced Materials 2007, 19 (17), 2307-2312.
142. Tautz, R.; Da Como, E.; Limmer, T.; Feldmann, J.; Egelhaaf, H.-J.; von Hauff, E.;
Lemaur, V.; Beljonne, D.; Yilmaz, S.; Dumsch, I.; Allard, S.; Scherf, U., Nature
Communications 2012, 3 (1), 970.
143. Wiebeler, C.; Tautz, R.; Feldmann, J.; von Hauff, E.; Da Como, E.; Schumacher,
S., The Journal of Physical Chemistry B 2013, 117 (16), 4454-4460.
161
144. Kaake, L. G.; Jasieniak, J. J.; Bakus, R. C.; Welch, G. C.; Moses, D.; Bazan, G. C.;
Heeger, A. J., Journal of the American Chemical Society 2012, 134 (48), 19828-
19838.
145. Etzold, F.; Howard, I. A.; Forler, N.; Cho, D. M.; Meister, M.; Mangold, H.; Shu, J.;
Hansen, M. R.; Müllen, K.; Laquai, F., Journal of the American Chemical Society
2012, 134 (25), 10569-10583.
146. McArthur, E. A.; Morris-Cohen, A. J.; Knowles, K. E.; Weiss, E. A., The journal of
physical chemistry. B 2010, 114 (45), 14514-20.
147. Knowles, K. E.; McArthur, E. A.; Weiss, E. A., ACS Nano 2011, 5 (3), 2026-2035.
148. Gesuele, F.; Sfeir, M. Y.; Koh, W. K.; Murray, C. B.; Heinz, T. F.; Wong, C. W., Nano
Letters 2012, 12 (6), 2658-2664.
149. El-Ballouli, A. a. O.; Alarousu, E.; Usman, A.; Pan, J.; Bakr, O. M.; Mohammed, O.
F., ACS Photonics 2014, 1 (3), 285-292.
150. Piland, G. B.; Burdett, J. J.; Dillon, R. J.; Bardeen, C. J., The Journal of Physical
Chemistry Letters 2014, 5 (13), 2312-2319.
151. Hanna, M. C.; Nozik, A. J., Journal of Applied Physics 2006, 100 (7), 074510.
152. Tayebjee, M. J. Y.; Gray-Weale, A. A.; Schmidt, T. W., The Journal of Physical
Chemistry Letters 2012, 3 (19), 2749-2754.
153. Congreve, D. N.; Lee, J.; Thompson, N. J.; Hontz, E.; Yost, S. R.; Reusswig, P. D.;
Bahlke, M. E.; Reineke, S.; Voorhis, T. V.; Baldo, M. A., Science 2013, 340 (6130),
334-337-337.
154. Grabowski, Z. R.; Rotkiewicz, K.; Rettig, W., Chemical Reviews 2003, 103 (10),
3899-4032.
155. Schrauben, J. N.; Zhao, Y.; Mercado, C.; Dron, P. I.; Ryerson, J. L.; Michl, J.; Zhu,
K.; Johnson, J. C., ACS Applied Materials & Interfaces 2015, 7 (4), 2286-2293.
156. Kolomeisky, A. B.; Feng, X.; Krylov, A. I., The Journal of Physical Chemistry C
2014, 118 (10), 5188-5195.
157. Feng, X.; Kolomeisky, A. B.; Krylov, A. I., The Journal of Physical Chemistry C
2014, 118 (34), 19608-19617-19617.
158. Matsika, S.; Feng, X.; Luzanov, A. V.; Krylov, A. I., The Journal of Physical
Chemistry A 2014, 118 (51), 11943-11955.
159. Margulies, E. A.; Miller, C. E.; Wu, Y.; Ma, L.; Schatz, G. C.; Young, R. M.;
Wasielewski, M. R., Nat Chem 2016, advance online publication.
160. Schrauben, J. N.; Akdag, A.; Wen, J.; Havlas, Z.; Ryerson, J. L.; Smith, M. B.;
Michl, J.; Johnson, J. C., The Journal of Physical Chemistry A 2016, 120 (20),
3473-3483.
161. Freeman, D. M. E.; Musser, A. J.; Frost, J. M.; Stern, H. L.; Forster, A. K.; Fallon,
K. J.; Rapidis, A. G.; Cacialli, F.; McCulloch, I.; Clarke, T. M.; Friend, R. H.;
Bronstein, H., Journal of the American Chemical Society 2017, 139 (32), 11073-
11080.
162
162. Wu, Y.; Wang, Y.; Chen, J.; Zhang, G.; Yao, J.; Zhang, D.; Fu, H., Angewandte
Chemie International Edition 2017, 56 (32), 9400-9404.
163. Mauck, C. M.; Bae, Y. J.; Chen, M.; Powers ‐Riggs, N.; Wu, Y. L.; Wasielewski, M.
R., ChemPhotoChem 2018, 2 (3), 223-233-233.
164. Pace, N. A.; Zhang, W.; Arias, D. H.; McCulloch, I.; Rumbles, G.; Johnson, J. C.,
The Journal of Physical Chemistry Letters 2017, 8 (24), 6086-6091.
165. Sakuma, T.; Sakai, H.; Araki, Y.; Mori, T.; Wada, T.; Tkachenko, N. V.; Hasobe, T.,
The Journal of Physical Chemistry A 2016, 120 (11), 1867-1875.
166. Carey, T. J.; Snyder, J. L.; Miller, E. G.; Sammakia, T.; Damrauer, N. H., The
Journal of Organic Chemistry 2017, 82 (9), 4866-4874.
167. Lehnherr, D.; Alzola, J. M.; Mulzer, C. R.; Dichtel, W. R., The Journal of Organic
Chemistry 2017.
168. Liu, H.; Wang, R.; Shen, L.; Xu, Y.; Xiao, M.; Zhang, C.; Li, X., Organic Letters 2017,
19 (3), 580-583.
169. Kumarasamy, E.; Sanders, S. N.; Tayebjee, M. J. Y.; Asadpoordarvish, A.; Hele,
T. J. H.; Fuemmeler, E. G.; Pun, A. B.; Yablon, L. M.; Low, J. Z.; Paley, D. W.; Dean,
J. C.; Choi, B.; Scholes, G. D.; Steigerwald, M. L.; Ananth, N.; McCamey, D. R.;
Sfeir, M. Y.; Campos, L. M., Journal of the American Chemical Society 2017.
170. Pun, A. B.; Sanders, S. N.; Kumarasamy, E.; Sfeir, M. Y.; Congreve, D. N.;
Campos, L. M., Advanced Materials 2017, 29 (41), 1701416.
171. Dean, J. C.; Zhang, R.; Hallani, R. K.; Pensack, R. D.; Sanders, S. N.; Oblinsky, D.
G.; Parkin, S. R.; Campos, L. M.; Anthony, J. E.; Scholes, G. D., Physical
Chemistry Chemical Physics 2017, 19 (34), 23162-23175.
172. Cook, J. D.; Carey, T. J.; Damrauer, N. H., The Journal of Physical Chemistry A
2016, 120 (26), 4473-4481.
173. Carey, T. J.; Miller, E. G.; Gilligan, A. T.; Sammakia, T.; Damrauer, N. H., Organic
Letters 2018, 20 (2), 457-460.
174. Wang, L.; Wu, Y.; Chen, J.; Wang, L.; Liu, Y.; Yu, Z.; Yao, J.; Fu, H., The Journal of
Physical Chemistry Letters 2017, 8 (22), 5609-5615.
175. Pace, N. A.; Zhang, W.; Arias, D. H.; McCulloch, I.; Rumbles, G.; Johnson, J. C.,
The Journal of Physical Chemistry Letters 2017, 8 (24), 6086-6091-6091.
176. Liu, H.; Wang, X.; Pan, L.; Shen, L.; Wang, X.; Chen, Q.; Li, X., Journal of
Photochemistry and Photobiology A: Chemistry 2017, 340, 21-28.
177. Yamakado, T.; Takahashi, S.; Watanabe, K.; Matsumoto, Y.; Osuka, A.; Saito, S.,
Angewandte Chemie International Edition 2018, 57 (19), 5438-5443.
178. Zirzlmeier, J.; Lehnherr, D.; Coto, P. B.; Chernick, E. T.; Casillas, R.; Basel, B. S.;
Thoss, M.; Tykwinski, R. R.; Guldi, D. M., Proceedings of the National Academy
of Sciences 2015, 112 (17), 5325-5330.
179. Johnson, J. C.; Akdag, A.; Zamadar, M.; Chen, X.; Schwerin, A. F.; Paci, I.; Smith,
M. B.; Havlas, Z.; Miller, J. R.; Ratner, M. A., The Journal of Physical Chemistry B
2013, 117 (16), 4680-4695.
163
180. Grieco, C.; Kennehan, E. R.; Kim, H.; Pensack, R. D.; Brigeman, A. N.; Rimshaw,
A.; Payne, M. M.; Anthony, J. E.; Giebink, N. C.; Scholes, G. D.; Asbury, J. B., The
Journal of Physical Chemistry C 2018, 122 (4), 2012-2022.
181. Zeng, T., The Journal of Physical Chemistry Letters 2016, 7 (21), 4405-4412.
182. Korovina, N. V.; Das, S.; Nett, Z.; Feng, X.; Joy, J.; Haiges, R.; Krylov, A. I.;
Bradforth, S. E.; Thompson, M. E., Journal of the American Chemical Society
2016, 138 (2), 617-627.
183. Hug, S. L. M.; Ian, C.; Gordon, L., Handbook of Photochemistry, Second Edition.
184. Sanders, S. N.; Kumarasamy, E.; Pun, A. B.; Trinh, M. T.; Choi, B.; Xia, J.; Taffet,
E. J.; Low, J. Z.; Miller, J. R.; Roy, X.; Zhu, X. Y.; Steigerwald, M. L.; Sfeir, M. Y.;
Campos, L. M., Journal of the American Chemical Society 2015, 137 (28), 8965-
8972.
185. Roberts, S. T.; McAnally, R. E.; Mastron, J. N.; Webber, D. H.; Whited, M. T.;
Brutchey, R. L.; Thompson, M. E.; Bradforth, S. E., Journal of the American
Chemical Society 2012, 134 (14), 6388-6400.
186. Reusswig, P. D.; Congreve, D. N.; Thompson, N. J.; Baldo, M. A., Applied Physics
Letters 2012, 101 (11), 113304.
187. Roberts, S. T.; Schlenker, C. W.; Barlier, V.; McAnally, R. E.; Zhang, Y.; Mastron,
J. N.; Thompson, M. E.; Bradforth, S. E., The Journal of Physical Chemistry
Letters 2010, 2 (2), 48-54-54.
188. Korovina, N. K.; Joy, J.; Bradforth, S. E.; Thompson, M. E., manuscript in
preparation.
189. Korovina, N. V.; Joy, J.; Feng, X.; Feltenberger, C.; Krylov, A. I.; Bradforth, S. E.;
Thompson, M. E., Journal of the American Chemical Society 2018, 140 (32),
10179-10190-10190.
190. van Stokkum, I. H. M.; Larsen, D. S.; van Grondelle, R., Biochimica et Biophysica
Acta (BBA) - Bioenergetics 2004, 1657 (2), 82-104.
191. Singh, S.; Jones, W. J.; Siebrand, W.; Stoicheff, B. P.; Schneider, W. G., The
Journal of Chemical Physics 1965, 42 (1), 330-342.
192. Ryerson, J. L.; Schrauben, J. N.; Ferguson, A. J.; Sahoo, S. C.; Naumov, P. e.;
Havlas, Z. k.; Michl, J.; Nozik, A. J.; Johnson, J. C., The Journal of Physical
Chemistry C 2014, 118 (23), 12121-12132.
193. 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.; Bulovi ć, V.;
Bawendi, M. G.; Baldo, Marc A., Nature Materials 2014, 13 (11), nmat4097.
194. Thorsmølle, V. K.; Averitt, R. D.; Demsar, J.; Smith, D. L.; Tretiak, S.; Martin, R.
L.; Chi, X.; Crone, B. K.; Ramirez, A. P.; Taylor, A. J., Physical Review Letters 2009,
102 (1), 017401.
195. Folie, B. D.; Haber, J. B.; Refaely-Abramson, S.; Neaton, J. B.; Ginsberg, N. S.,
Journal of the American Chemical Society 2018, 140 (6), 2326-2335.
164
196. Burdett, J. J.; Bardeen, C. J., Accounts of chemical research 2013, 46 (6), 1312-
1320.
197. Yost, S. R.; Lee, J.; Wilson, M. W. B.; Wu, T.; McMahon, D. P.; Parkhurst, R. R.;
Thompson, N. J.; Congreve, D. N.; Rao, A.; Johnson, K.; Sfeir, M. Y.; Bawendi, M.
G.; Swager, T. M.; Friend, R. H.; Baldo, M. A.; Van Voorhis, T., Nat Chem 2014, 6
(6), 492-497.
198. Sonogashira, K., Journal of Organometallic Chemistry 2002, 653 (1), 46-49.
199. Berkelbach, T. C.; Hybertsen, M. S.; Reichman, D. R., The Journal of Chemical
Physics 2013, 138 (11), 114102.
200. Congreve, D. N.; Lee, J.; Thompson, N. J.; Hontz, E.; Yost, S. R.; Reusswig, P. D.;
Bahlke, M. E.; Reineke, S.; Van Voorhis, T.; Baldo, M. A., Science 2013, 340
(6130), 334-337.
201. Lukman, S.; Musser, A. J.; Chen, K.; Athanasopoulos, S.; Yong, C. K.; Zeng, Z.;
Ye, Q.; Chi, C.; Hodgkiss, J. M.; Wu, J.; Friend, R. H.; Greenham, N. C., Advanced
Functional Materials 2015, 25 (34), 5452-5461-5461.
202. Lippitz, M.; Hübner, C. G.; Christ, T.; Eichner, H.; Bordat, P.; Herrmann, A.;
Müllen, K.; Basché, T., Physical Review Letters 2004, 92 (10), 103001.
203. Johnson, J. C.; Nozik, A. J.; Michl, J., Accounts of Chemical Research 2013, 46
(6), 1290-1299.
204. Ramanan, C.; Smeigh, A. L.; Anthony, J. E.; Marks, T. J.; Wasielewski, M. R.,
Journal of the American Chemical Society 2012, 134 (1), 386-397.
205. Sanders, S. N.; Kumarasamy, E.; Pun, A. B.; Steigerwald, M. L.; Sfeir, M. Y.;
Campos, L. M., Angewandte Chemie 2016, 128 (10), 3434-3438.
206. Bryce-Smith, D., Photochemistry. 1973; Vol. 4, p X001-X004-X004.
207. Zirzlmeier, J.; Casillas, R.; Reddy, S. R.; Coto, P. B.; Lehnherr, D.; Chernick, E. T.;
Papadopoulos, I.; Thoss, M.; Tykwinski, R. R.; Guldi, D. M., Nanoscale 2016, 8
(19), 10113-10123.
208. Lemmetyinen, H.; Tkachenko, N.; Valeur, B.; Hotta, J.-i.; Ameloot, M.; Ernsting,
N.; Gustavsson, T.; Boens, N., Pure and Applied Chemistry 2014, 86, 1969-1998.
209. Chosrowjan, H.; Taniguchi, S.; Tanaka, F., The FEBS Journal 2015, 282 (16),
3003-3015.
210. Messina, F.; Bräm, O.; Cannizzo, A.; Chergui, M., Nature Communications 2013,
4 (1), 2119.
211. Kwok, W.-M.; Ma, C.; Phillips, D. L., The Journal of Physical Chemistry B 2009,
113 (33), 11527-11534.
212. Yoon, M.-C.; Jeong, D. H.; Cho, S.; Kim, D.; Rhee, H.; Joo, T., The Journal of
Chemical Physics 2002, 118 (1), 164-171.
213. Maffeis, V.; Brisse, R.; Labet, V.; Jousselme, B.; Gustavsson, T., The Journal of
Physical Chemistry A 2018, 122 (25), 5533-5544.
214. Mahr, H.; Hirsch, M. D., Optics Communications 1975, 13 (2), 96-99.
165
215. Block, D.; Shah, J.; Gossard, A. C., Solid State Communications 1986, 59 (8), 527-
531.
216. Shah, J., IEEE Journal of Quantum Electronics 1988, 24 (2), 276-288.
217. Castner, E. W.; Maroncelli, M.; Fleming, G. R., The Journal of Chemical Physics
1987, 86 (3), 1090-1097.
218. Xie, S.; Du, M.; Mets, L.; Fleming, G. R., Femtosecond fluorescence
depolarization study of photosynthetic antenna proteins: observation of
ultrafast energy transfer in trimeric C-phycocyanin and allophycocyanin.
Proceedings Volume 1640, Time-Resolved Laser Spectroscopy in Biochemistry
III: 1992.
219. Rosenthal, S. J.; Jimenez, R.; Fleming, G. R.; Kumar, P. V.; Maroncelli, M.,
Journal of Molecular Liquids 1994, 60 (1), 25-56.
220. Bradforth, S. E.; Jimenez, R.; van Mourik, F.; van Grondelle, R.; Fleming, G. R.,
The Journal of Physical Chemistry 1995, 99 (43), 16179-16191.
221. Ricci, M.; Bradforth, S. E.; Jimenez, R.; Fleming, G. R., Chemical Physics Letters
1996, 259 (3), 381-390.
222. Haacke, S.; Taylor, R. A.; Bar-Joseph, I.; Brasil, M. J. S. P.; Hartig, M.; Deveaud,
B., Journal of the Optical Society of America B 1998, 15 (4), 1410-1417.
223. Schanz, R.; Kovalenko, S. A.; Kharlanov, V.; Ernsting, N. P., Applied Physics
Letters 2001, 79 (5), 566-568.
224. Zhao, L.; Luis Pérez Lustres, J.; Farztdinov, V.; Ernsting, N. P., Physical
Chemistry Chemical Physics 2005, 7 (8), 1716-1725.
225. Zhang, X. X.; Würth, C.; Zhao, L.; Resch-Genger, U.; Ernsting, N. P.; Sajadi, M.,
Review of Scientific Instruments 2011, 82 (6), 063108.
226. Sajadi, M.; Quick, M.; Ernsting, N. P., Applied Physics Letters 2013, 103 (17),
173514.
227. Gerecke, M.; Bierhance, G.; Gutmann, M.; Ernsting, N. P.; Rosspeintner, A.,
Review of Scientific Instruments 2016, 87 (5), 053115.
228. Cannizzo, A.; Bräm, O.; Zgrablic, G.; Tortschanoff, A.; Oskouei, A. A.; van Mourik,
F.; Chergui, M., Optics Letters 2007, 32 (24), 3555-3557.
229. Smith, A. V., How to use SNLO nonlinear optics software to select nonlinear
crystals and model their performance. 2003.
230. Joy, J., Note to User: At this point, I have not figured out how to
programmatically incorporate the optimized position of the translation stage as
time-zero for the experiment within this Labview interface – my apologies! This
was my less than elegant workaround: move the translation stage manually
using the ESP controller to optimize the upconversion signal. There will be a
position displayed on the controller at this time, in millimeters. Switch the ESP
controller off and on again. Doing so resets its memory and the display should
now read 0.0000 mm – which correctly sets up this position as time zero for the
rest of the experiment. I wish you good luck on figuring this glitch out! . 2019.
166
231. Trebino, R.; Kane, D. J., Journal of the Optical Society of America A 1993, 10 (5),
1101-1111.
232. DeLong, K. W.; Fittinghoff, D. N.; Trebino, R., IEEE Journal of Quantum
Electronics 1996, 32 (7), 1253-1264.
Abstract (if available)
Abstract
The World Energy Outlook recently published by the International Energy Agency (IEA) makes the dire prediction that global CO₂ levels will continue rising despite profound shifts in the functioning of the global energy system. The rise in these levels is attributed to rising demand for oil and gas, despite the demand for coal plateauing and the surge in wind and solar power meeting the majority of increases in global energy demand. The IEA calls for a significant reallocation of investment away from fossil fuels towards renewables, and the decommissioning half of the world’s coal fired power stations—in order to meet the limiting warming set out by the Paris Agreement of 1.65C. Among such renewables, solar photovoltaics are becoming the most competitive source of electricity in 2020 in China and India while in the European Union and the United States it is predicted to largely close the gap with other sources like coal and natural gas by 2030. Interestingly, the IEA has also consistently come under fire for failing to predict the growth of renewables. As seen in the actual net additions to solar capacity has outperformed IEA outlooks since 2006—which the IEA attributes to shifts in government policy over time.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
First steps of solar energy conversion; primary charge generation processes in condensed phase coupled chromophores
PDF
Singlet fission in disordered acene films: from photophysics to organic photovoltaics
PDF
Singlet fission in covalently-linked tetracenes
PDF
Interfacial vibrational and electronic excited states of photovoltaic materials probed using ultrafast nonlinear spectroscopy
PDF
Nanocrystal surface engineering as a route towards improved photovoltaics
PDF
Sustainable continuous flow syntheses of colloidal inorganic nanoparticle catalysts
PDF
Growth, characterization of gallium arsenide based nanowires and application in photovoltaic cells
PDF
Cathode and anode materials for sodium ion batteries
PDF
Molecular design for organic photovoltaics: tuning state energies and charge transfer processes in heteroaromatic chromophores
PDF
Organic photovoltaics: from materials development to device application
PDF
Development of electronic design automation tools for large-scale single flux quantum circuits
PDF
Healing of defects in random antiferromagnetic spin chains
PDF
Calculating solar photovoltaic potential on residential rooftops in Kailua Kona, Hawaii
PDF
Light Emission from Carbon Nanotubes and Two-Dimensional Materials
PDF
Multilayer grown ultrathin nanostructured GaAs solar cells towards high-efficiency, cost-competitive III-V photovoltaics
PDF
Photoexcitation and nonradiative relaxation in molecular systems: methodology, optical properties, and dynamics
PDF
Disordered quantum spin chains with long-range antiferromagnetic interactions
PDF
Temperature-dependent photoionization and electron pairing in metal nanoclusters
PDF
Topological protection of quantum coherence in a dissipative, disordered environment
PDF
Transient pulsed plasma for pollution remediation and energy conversion applications
Asset Metadata
Creator
Joy, Jimmy
(author)
Core Title
Exciton dynamics in photovoltaic materials
School
College of Letters, Arts and Sciences
Degree
Doctor of Philosophy
Degree Program
Chemistry
Publication Date
02/14/2020
Defense Date
02/14/2020
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
femtosecond,lasers,nanoparticle,OAI-PMH Harvest,photovoltaic,quantum dot,singlet fission,solar,spectroscopy
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Bradforth, Stephen (
committee chair
), Brutchey, Richard (
committee member
), Nakano, Aiichiro (
committee member
)
Creator Email
jimmyjoy@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c89-268532
Unique identifier
UC11674139
Identifier
etd-JoyJimmy-8175.pdf (filename),usctheses-c89-268532 (legacy record id)
Legacy Identifier
etd-JoyJimmy-8175.pdf
Dmrecord
268532
Document Type
Dissertation
Rights
Joy, Jimmy
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
Tags
femtosecond
lasers
nanoparticle
photovoltaic
quantum dot
singlet fission
solar
spectroscopy