Optical Studies of Colloidal Quantum Dots: Optical
Trapping with Plasmonic Nanoapertures and
Thermal Recovery from Photoinduced Dimming
by
Russell Andrew Jensen
LU
o
0
Doctor of Philosophy
at the
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
February 2015
@
Massachusetts Institute of Technology 2015. All rights reserved.
A uthor ..............................
Signature redacted
Department of Chemistry
January 15th, 2015
C ertified by .......................
Signature redacted
Moungi G. Bawendi
Lester Wolfe Professor of Chemistry
Thesis Supervisor
Accepted by .....................
Signature redacted
Robert W. Field
Chairman, Department Committee on Graduate Students
-
U,
-
Submitted to the Department of Chemistry
in partial fulfillment of the requirements for the degree of
2
Optical Studies of Colloidal Quantum Dots: Optical Trapping
with Plasmonic Nanoapertures and Thermal Recovery from
Photoinduced Dimming
by
Russell Andrew Jensen
Submitted to the Department of Chemistry
on January 15th, 2015, in partial fulfillment of the
requirements for the degree of
Doctor of Philosophy
Abstract
This doctoral research has been defined by two main goals. The first has been to
develop single colloidal quantum dot (QD) absorption as a new spectroscopic tool
for investigating single QD electronic properties, dynamics, and inhomogeneities. In
an important step towards achieving this goal, QDs were introduced into the field of
optical trapping. Silica coated QDs were optically trapped using bowtie apertures in
a thin silver film with low incident flux of 1.56 MW/cm 2 at 1064 nm. Additionally,
QDs emitted upon trapping via two-photon excitation from the trapping laser due to
strong field enhancement inside the aperture.
The second goal of this research has been to investigate processes involved in single
QD fluorescence intermittency, or blinking. Specifically, the transition from a nonemissive QD to an emissive QD was investigated using controlled amounts of thermal
energy to drive recovery from photoinduced dimming in QD ensembles. Nonlinear
thermal recovery was well described by a stretched exponential function, and further
analysis yielded an underlying probability distribution of rate constants. Casting the
rate constants as a collection of first-order activated processes provided an activation
barrier probability distribution with significant density at room temperature thermal
energy that peaks at 200 meV before decaying to zero.
Progress towards single QD absorption using alternative nanoscale structures, including slot waveguides and circular apertures in silver film, is also discussed. Lastly,
self-assembled cyanine-dye nanotubes were monitored during flash dilution with absorption spectroscopy at a high frame rate to separate spectroscopic contributions of
the outer layer in double walled and bundled nanotubes.
Thesis Supervisor: Moungi G. Bawendi
Title: Lester Wolfe Professor of Chemistry
3
4
To my
5
family
6
Acknowledgments
The completion of this degree was a team effort. I would not have been able to do it
without help and support from a lot of people, and this is by no means a complete
list. I would like to express my deepest gratitude to
" Areum for always believing in me.
" Mom and Dad for providing me the tools and support to accomplish whatever
I want.
" Tyler for all of the weekend alien and zombie killing.
* Billijo for being an awesome sister and putting up with all of my visits to CA!
" Shannon and Lee for introducing me to the beauty of Vermont and Heady
Topper.
" Grandma and Grandpa for being proud of me. It means a lot.
* Eric Victor, Kaz Yamanouchi, Peter Allen, and Dan Harris for being the best
bros a guy could ask for. I couldn't have remained sane without you guys.
" Jeff Eliason for all those hours we spent together in the paint and behind the
line of scrimmage.
* Jon Axtell for bringing Areum and me to the best lake on the east coast and
giving us the luxury suite.
* Eric Hontz for all of the trails through dirt and snow.
" Michael Trujillo for being a great role model and "inviting" me to Thanksgiving.
* Maggie Brown and Hanine Hajj for making my guys happy and being so generous.
" Wen Liu and Fiona Zhang for being fun house guests when terrorists were about.
" Ed de Courreges for always being down to roll. HELL YEAH!
" Larry Rich for introducing me to New England through the eyes of a local.
" Lisa Marshall for remaining a friend after getting me started in the lab and for
the great Japan adventure.
" Thomas Bischof for being far too helpful and one of the most capable people I
know.
" Andrew Beyler for letting me ask you math questions anytime, and helping me
get into trouble by discussing brewing and politics with me way too late.
7
"
Jian Qui being a great shooter and a happy drinker.
" Igor Coropceanu for stepping in to help at the perfect time.
* Ddrthe Eisele for all of those pep talks while torturing nanotubes.
* Daniel Franke for keeping the legacy going.
" Ou Chen for always being considerate, helpful, and generous.
" Mark Wilson for always being positive and enthusiastic. And for Feuerzangenbowle.
" David Strasfeld for teaching me how to eat chicken wings and showing everybody
how karaoke is done.
" He Wei for being hilarious.
* Jennifer Scherer for being helpful even though you didn't have to be. And a K1
fighter.
" Jennifer Choy for being a good friend and so patient when dealing with a plasmonics rookie.
" Qimin Quan for teaching me enough optics to get me though orals.
" Marko Loncar for always being excited and creative.
" Ben Ofori-Okai for your friendship and dedicated tutelage.
" Steph Teo for being a close friend and reteaching me calculus.
" the CGSC for all your great work but especially the retreats, cruises, BBQs,
Kraken fountains, "tapped" kegs, etc.
" the noon ball and city league guys for giving me something to look forward to
every week.
" Ddrcio Lira Jiu Jitsu for helping me to rebuild my confidence. OSS!
" Michael Grenier for being the glue that holds the Muddy Charles Pub together.
" Sylvia Ceyer for looking out for me and always trying to fix problems.
" Keith Nelson for always asking great questions.
* Jianshu Cao for being helpful and understanding.
* Li Miao for always getting stuff done!
" Moungi Bawendi for always providing the best possible guidance for me, whether
that meant applying pressure or giving me time to work things out. As a
scientist, your attention to detail and critical thinking has profoundly influenced
my thought process and I am a better person because of it.
8
This doctoral thesis has been examined by a Committee of the
Department of Chemistry as follows:
Signature redacted
C'
Professor Jianshu Cao
Thesis Committee Chairman
Signature redacted
Professor Moungi G. Bawendi
Thesis Supervisor
Signature redacted
Professor Keith A. Nelson
Thesis Committee Member
10
Contents
19
1.1
Colloidal Quantum Dot Synthesis
19
1.1.1
Monodisperse Cores . . .
20
1.1.2
Passivation
. . . . . . .
21
Optical Properties
. . . . . . .
21
.
.
.
1.2.2
Corrections
. . . . . . .
25
1.2.3
Fine Structure . . . . . .
25
Emission Dynamics . . . . . . .
26
1.3.1
Blinking Models . . . . .
27
1.3.2
Ensemble Behavior . . .
28
.
.
21
Optical Trapping and Two-Photon Excitation of Colloidal Quantum
Dots using Bowtie Apertures
31
. . . . . . . . . . .
31
2.2
Experimental Methods . . . . . . . . . . . .
. . . . . . . . . . .
34
2.2.1
Silica Coated Quantum Dot Synthesis a nd Characterization
34
2.2.2
Aperture Fabrication . . . . . . . . .
. . . . . . . . . . .
35
2.2.3
Simulations . . . . . . . . . . . . . .
. . . . . . . . . . .
36
2.2.4
Packaging and Instrumentation
. . .
. . . . . . . . . . .
38
Results . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . .
39
2.3.1
Single Particle Trapping . . . . . . .
. . . . . . . . . . .
39
2.3.2
Spectrally Resolved Trapping
. . . . . . . . . . .
41
11
. . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
2.3
.
Introduction . . . . . . . . . . . . . . . . . .
.
2.1
.
2
Initial Approximations
.
1.3
1.2.1
.
1.2
.
Introduction
.
1
42
C onclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
45
3
.
2.4
.
Multiple Particle Trapping and Emission Dynamics . . . . .
2.3.3
Thermal Recovery of Colloidal Quantum Dot Ensembles following
47
Photoinduced Dimming
47
. . . . . . . . . . .
48
3.2.1
Synthesis, Packaging, and Instrumentation . . . . . . . . . . .
48
3.2.2
Procedure . . . . . . . . . . . . . . . . .
. . . . . . . . . . .
49
Experimental Methods ...............
49
3.3.1
Scaling and Global Fit . . . . . . . . . .
49
3.3.2
M odeling
. . . . . . . . . . . . . . . . .
52
D iscussion . . . . . . . . . . . . . . . . . . . . .
55
.
.
.
Results . . . . . . . . . . . . . . . . . . . . . . .
.
3.4
.
3.2
. . . . . . .
.
M otivation . . . . . . . . . . . . ..
3.3
Progress Towards Single Quantum Dot Absorption Spectroscopy
57
Introduction . . . . . . . . . . . . . .
. . . . . . . . . .
57
4.2
Silica Coated Quantum Dots . . . . .
. . . . . . . . . .
59
4.3
Slot Waveguides . . . . . . . . . . . .
. . . . . . . . . .
60
. . . . . . . . . .
. . . . . . . . . .
61
.
.
4.3.1
Description
4.3.2
Transmission and Emission Measurements
. . . . . . . . . .
61
4.3.3
Placement of QDs . . . . . . . . . . . . .
. . . . . . . . . .
64
Circular Apertures in Silver Film . . . . . . . .
. . . . . . . . . .
67
. . . .
.
.
4.4.1
Fabrication and Characterization
. . . . . . . . . .
68
4.4.2
QD Absorption Spectrum through a Circular Aperture . . . .
68
.
4.4
.
4.1
.
4
. . . . . . . . . . .
3.1
A Fast-Acquisition Absorption Spectroscopy of Self-Assembled CyanineDye Nanotubes: A summary of contributions made to Eisele et al.,
2014
73
A.1 Self-Assembled Nanotubes . . . . . . . . . . . . . . . . . . . . . . . .
73
. . . . . . . . . . . . . . .
74
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
77
A.2 Fast-Acquisition Absorption Spectroscopy
A .3 Sum m ary
12
List of Figures
1-1
Commonly used precursors (top) cadmium oleate and (bottom) TOPSe.
20
1-2
1-3
Band alignment for QD cores (CdSe) and common shell materials (CdS,
ZnS). Reproduced from Reiss et al., 2009. . . . . . . . . . . . . . . .
22
Absorption and emission spectra for CdSe/CdS core/shell QDs. The
absorption spectrum shows discrete low energy absorption transitions
as evidence for quantization of the semiconductor band gap.
1-4
. . . . .
23
Illustration of a QD size series with accompanying band diagrams. The
band diagram for each QD compares the bulk semiconductor band
(solid line) to the approximated band (dashed line) with quantized
energies that vary with QD diameter. Adapted from Norris 2003.
1-5
. .
24
(a) An emission intensity trace from a single CdSe,/ZnS core/shell QD
exhibits blinking. (b) Probability distributions for ON/OFF duration
measured from single QD intensity traces follow a power law distribution (adapted from Shimizu et al., 2001).
1-6
. . . . . . . . . . . . . . .
26
Experimental ensemble emission intensity on a log-log plot. The cutoff
times, T, for the single QD ON/OFF time probability distributions are
labeled. Adapted from Chung and Bawendi, 2004.
13
. . . . . . . . . .
29
2-1
(a)(b) SEM images of the bowtie apertures used in the experiments,
overlapped with field intensity enhancement profiles at 1064 nm. The
confined gap mode is dominant when the polarization is across the
gap. (c) The simulated transmission spectra of the two apertures used
in trapping experiments, showing peak resonances are blue-shifted from
the 1064 nm trapping laser. (d) Transmission electron microscope image of the silica coated quantum dots used in trapping.
2-2
. . . . . . .
33
Normalized absorption (blue) and emission (red) spectra for scQDs.
Continuous wave 532 nm excitation was used as an excitation source
for the emission spectrum.
2-3
. . . . . . . . . . . . . . . . . . . . . . .
Size distribution of scQDs before (green) and after (blue) filtering measured with DLS. Sizes were calculated by volume. . . . . . . . . . . .
2-4
34
35
(a) The simulated field intensity distribution inside the aperture showing field enhancement on both faces of the aperture.
The scQD is
shown in its final position at the bottom of the aperture touching the
SiN membrane. (b) Potential energy calculation results showing that
scQDs of at least 25 nm will have a potential lower than - 1 kBT at
the bottom of the aperture.
2-5
........
38
..............................
Instrument schematic for simultaneous trapping with 1064 nm laser
(gray beam) and scQD emission detection at 640 nm (red beam).
2-7
37
The calculated potential for particle trapping with the 56 nm aperture
in figure 2-1b.
2-6
. . . . . . . . . . . . . . . . . . . . . . .
. .
40
The (a) emission and (b) 1064 nm transmission channels show a stepwise increase in signal at 50 seconds, suggesting individual scQD trapp in g .
2-8
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
QD emission spectra before (blue) and after (red) optical trapping and
two-photon excitation. . . . . . . . . . . . . . . . . . . . . . . . . . .
2-9
41
42
(a) Emission and (b) 1064 nm transmission channels for spectrally
resolved emission detection presented in figure 2-8.
14
. . . . . . . . . .
43
2-10 (a) Emission and (b) 1064 nm transmission for filtered scQDs in the
56 nm aperture shows evidence to QD blinking inside the optical trap.
Multiple trapping events are detected in the (c) emission and (d) 1064
nm transmission channels for filtered scQDs in the 38 nm aperture that
exhibit rapid quenching at 265 and 280 seconds in the emission channel
on ly. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3-1
Example of collected emission data for illustrative purposes.
44
Pho-
todimming and recovery for three separate excitation spots. Each spot
was dimmed for 30 seconds and probed for recovery only once at logarithmically spaced wait times.
3-2
. . . . . . . . . . . . . . . . . . . . .
Recovery probabilities collected at all of temperatures. Data from each
temperature is offset for clarity. . . . . . . . . . . . . . . . . . . . . .
3-3
50
51
(a) Temperature-dependent recovery with global stretched exponential
fit. (b) Data plotted on a time axis corrected by ko emphasizes the validity of the fitting procedure. Data falls along a stretched exponential
with 0 = 0.326 (dashed line).
. . . . . . . . . . . . . . . . . . . . . .
3-4
Temperature-independent rate constant probability distribution.
3-5
Arrhenius plot of ln(ko) as a function of inverse temperature (K-)
.
with linear fit yields Ea,o = 188.03 meV and AO = 0.0105 s-1. ....
3-6
53
54
Probability distribution of activation energy barriers for air-free CdSe
cores. ........
4-1
52
...................................
54
a)QD655 on glass substrate, b)QD655 on TiO 2 substrate, c)SiO 2 coated
QDs on glass substrate, and d)Si0 2 coated QDs on TiO 2 substrate.
Emission is quenched only in b), where the QD shell is in direct contact with the TiO 2 substrate.
. . . . . . . . . . . . . . . . . . . . . .
4-2
Slot waveguide fabrication procedure.
4-3
(Left) Scanning electron microscope image of a TiO 2 slot waveguide
. . . . . . . . . . . . . . . . .
60
62
with 60 nm wide slot. (Right) Illustration of slot waveguide structure.
Dimensions are 10 pm in length, a = 150 nm, and b = 200 nm.
15
. . .
62
4-4
Illustrations of the transmission (left) and emission (right) experimental schem es.
4-5
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
63
(Left) Reference (blue), transmitted broadband light (green), and calculated transmission (red) for a slot waveguide with a 20 nm slot width.
The reference spectrum was measured by reflecting the beam off a flat
surface on the sample. The calculated transmission has a FSR of 7.2
nm and was normalized by the reference spectrum. (Right) Illustration
of a slot waveguide structure and parameters used to calculate FSR. A
central transmission wavelength A0
=
620 nm, length 1 = 10 pm, index
of refraction n = 2.7, and a coupling angle of 0 = 0 were used. ....
4-6
64
(a) Overlay of emission channel (color) on laser scatter channel (grayscale)
for an array of slot waveguides. (b) Overlay image of slot waveguide
with emission at both ends as highlighted in (a). (c) The same 20 nm
gap slot waveguide excited at the top facet. Emission was collected
on a CCD camera and showed emission at both ends of the waveguide
with blinking individual QDs along the waveguide.
4-7
. . . . . . . . . .
65
Top-view SEM image of a 100 nm gap slow waveguide with 20 nm
diameter silica coated QDs. It was possible to image QDs inside the
slot with larger gap slot waveguides only.
4-8
. . . . . . . . . . . . . . .
66
AFM images of a slot waveguide spun cast with 20 nm silica coated
QDs (a) before and (b) after sweeping QDs into the slot with the AFM
tip in contact mode.
4-9
. . . . . . . . . . . . . . . . . . . . . . . . . . .
67
(top) Circular aperture fabrication schematic. (1) Spin-cast HSQ on
silicon nitride membrane supported by silicon scaffold. (2) Write posts
in HSQ. (3) Evaporate silver film. (4) Liftoff with HF to remove posts.
(bottom left) SEM image of post structures.
image of circular apertures.
(Bottom Right) SEM
. . . . . . . . . . . . . . . . . . . . . . .
69
4-10 (a) Image of circular apertures back lit by a broadband source. (b) Calculated (dashed) and measured (solid) transmission spectra for circular
apertures of varying radius.
. . . . . . . . . . . . . . . . . . . . . . .
16
69
4-11 QD absorption spectra measured for an evaporated film ensemble and
for a collection of QDs inside a 140 nm radius circular aperture.
. . .
70
4-12 SEM of a 100 nm radius circular aperture coated by a thin film of QDs.
The bottom and the inside walls are coated with QDs.
. . . . . . . .
71
A-1 (a) Illustration of a double-walled LHN. (b) Absorption spectra reveal changes in excitonic properties upon assembly from monomers
to double-walled LHNs and further assembly to bundled LHNs. The
symbols 11and _ indicate the polarization of each band. Adapted with
permission from Eisele et al., 2014. Copyright 2014 by the Proceedings
of the National Academy of Sciences. . . . . . . . . . . . . . . . . . .
75
A-2 Fast-acquisition absorption spectrometer schematic. . . . . . . . . . .
76
A-3 Flash dilution results. Reproduced with permission from Eisele et al.,
2014. Copyright 2014 by the Proceedings of the National Academy of
Sciences. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17
76
18
Chapter 1
Introduction
Colloidal quantum dots (QDs) are semiconductor nanocrystals that are simple to
synthesize and have interesting optical properties that make them potentially useful
for applications where light-matter interactions are emphasized. Applications include
light harvesting, 1,2,3,4,5 multiband detection, 6'7 biological imaging, 8'9' 0'1 and display technologies. 12 ,13 , 14 ,1 5 Additionally, hybrid particles that incorporate QDs have
been developed to extend the applicability of QDs. 1617 A brief review of QD synthesis, optoelectronic properties, and emission behavior at the single QD and ensemble
levels will illustrate how results presented in this doctoral thesis can help to improve
understanding of QD emissive inefficiencies and push QDs into new applications.
1.1
Colloidal Quantum Dot Synthesis
Colloidal QDs with varying optical properties can be synthesized from an array of
semiconductor materials, including indium phosphide,1 8 indium arsenide, 19 lead sulfide, 20 ,21 cadmium telluride,
22 , 23
and cadmium arsenide. 24 Cadmium selenide QDs,
however, have emerged as the model system for colloidal QDs due to their compact
size, monodispersity, photostability, and emission in the visible region.
19
Cadmium Oleate
0- -0
0
Cd*+
0
Tri-n-octylphosphine Selenide
P
II
Se
Figure 1-1: Commonly used precursors (top) cadmium oleate and (bottom) TOPSe.
1.1.1
Monodisperse Cores
A colloidal QD synthetic procedure that produced monodisperse particles was an
important step towards making particles that could be confidently investigated and
used for applications." Particle size plays an essential role in the optical properties of
QDs, so monodispersity is a requirement for consistent and narrow optical features.
Monodispersity was achieved with a hot injection synthesis, where reactive cadmium
and selenium precursors were quickly injected to a high-boiling point solvent heated
to
-
300C. Rapid nucleation followed by slow crystalline growth produced monodis-
perse particle ensembles that were collected from the reaction mixture at regular time
intervals to halt particle growth and produce a size series of particles. Precursors and
reaction solvents have been optimized over time to tune precursor reactivity, reaction temperature, increase air stability, and reduce toxicity. Today, commonly used
precursors are cadmium oleate and tri-n-octylphosphine selenide (TOPSe), while octadecene, excess oleic acid, and tri-n-octylphosphine oxide (TOPO) are used as a high
boiling point solvent 2",
26 2 7
,
(Fig. 1-1). The resulting particles are capped with excess
ligands found in the growth solution, including TOPSe, TOPO, and oleic acid.
20
28,29
1.1.2
Passivation
Overcoating CdSe QDs with a ZnS3 0 ,3 1 or CdS 32 ,33 passivating shell improves the pho-
toluminescent quantum yield of QDs and reduces fluorescence intermittency known
as blinking. These improvements occur by reducing charge carrier access to defects
and dangling bonds on the QD surface with an insulating shell of high band gap material (Fig. 1-2). A ZnS shell insulates both electrons and holes from the QD surface
due to a large band gap, but suffers from a large lattice mismatch when compared to
CdSe. A CdS shell does not provide as large of an insulating barrier and allows the
electron wavefunction to delocalize to the surface of the core/shell QD, but a smaller
lattice mismatch provides fewer interfacial defects.
32,34
Novel synthetic procedures
have yielded CdSe/CdS core/shell QDs with improved photoluminescent properties,
and generally rely on successive ion layer adsorption and reaction (SILAR). 33,35 ,36
An alternate approach to overcoating, however, has yielded CdSe/CdS core/shell
QDs with
compact sizes, reduced defects, and improved passivation. Shell growth was
slowed by reducing the shell precursor reactivity and continuously injecting into the
reaction pot, which produced shells that maintained the original QD crystal structure. 37 , 38 Octanethiol was chosen as sulfur precursor due to its strong carbon-sulfur
bond. QDs synthesized using this method have a high quantum yield and very little fluorescence intermittency, and were used for much of the work presented in this
doctoral thesis.
1.2
1.2.1
Optical Properties
Initial Approximations
Colloidal QDs have interesting optical properties relating to the size of the particles.
Much like the quantum mechanical particle in a box model, the energy of each state
increases as the dimensions of the boundaries decrease. Following the absorption of
a photon, the excited state in a QD is an electron-hole pair that is confined within
the QD to dimensions smaller than the exciton Bohr radius, resulting quantization of
21
-2
-2.5
-3
E
_-3.5
>
-4
3
-4.5 --5
-5.5
-6
CdSe
CdS
ZnS
Figure 1-2: Band alignment for QD cores (CdSe) and common shell materials (CdS,
ZnS). Reproduced from Reiss et al., 2009.
the valence and conduction hands. This quantization is apparent in the QD absorption spectrum (Fig. 1-3) with discrete, low energy transitions observable in a neat,
monodisperse QD sample. The band edge energy for a QD is a function of particle
size (Fig. 1-4), and is given by
h 2 a2
E=Eq+ 2
2
2
-
1.8e 2
h2 a 2
V
-
,
(1.1)
where the first term, Eq, is simply the bulk semiconductor band gap. The second
and third terms contain information about the electron/hole kinetic energy, where
an(
an('c
7,1
are solutions (i.e. "zeros") to the Bessel function for quantum num-
bers n and 1. The Bessel function solutions are necessary to solve the Schr6dinger
equation with spherical boundary conditions and are responsible for quantizing the
kinetic energy of the electron /hole.
The Bessel function solutions also introduce
size dependence to the kinetic energy terms through r, the QD radius. The kinetic
energy terms also include the effective masses of the electron and hole, or m, and
m 1 . The effective mass approxination accounts for the periodic potentials of the
semiconductor lattice by adjusting the mass of the particle and then treating it as
a particle in a smooth background potential. This approximation relies on treating
22
E
w
C
0
Absorption
o
Emission
500
525
550
575
600
625
650
675
Wavelength (nm)
Figure 1-3: Absorption and emission spectra for CdSe/CdS core/shell QDs. The
absorption spectrum shows discrete low energy absorption transitions as evidence for
quantization of the semiconductor band gap.
electron/hole wavefunctions as Bloch functions, which can be broken down into a
periodic function (describing the periodic potentials of the QD lattice) and a plane
wave function, exp(-ik -r), where k is the wave vector. Separating out the periodic
function and treating the electron/hole wavefunction as a linear combination of plane
waves is called the envelope function approximation, and provides the kinetic energy
terms in equation 1.1 for an electron/hole in spherical boundary conditions. The
last term is a first-order correction for the coulombic attraction between the electron and hole where e is the charge of an electron and e is the dielectric constant of
the material. The Schr6dinger equation is not solvable when this term is included
in the Hamiltonian, so it is removed due to weak size dependence and replaced as
a first-order correction. 39 ,4' This approximation is justified because the system in
the strong confinement regime, where confinement by the QD boundary dominates
the electron/hole energies, making it possible to neglect coulombic attraction and
treat the electron and hole independently.
Figure 1-4 illustrates how quantization
and the applied approximations change the bulk semiconductor band diagram in this
simplified model.
23
E(k)
Conduction Band
460 nm
Band Edge
k
EBE
E
=2.7eV
2 nm
Valence Band
'
Bulk
Effective mass
approximation
E(k)
Conduction Band
550 nm
Band Edge
k
E
4 nm
Valence Band
(k)
640 nm
Band Edge
Conduction Band
,TkE
EBE
1.9 eV
Valence Band
8 nm
Figure 1-4: Illustration of a QD size series with accompanying band diagrams. The
band diagram for each QD compares the bulk semiconductor band (solid line) to
the approximated band (dashed line) with quantized energies that vary with QD
diameter. Adapted from Norris 2003.
24
1.2.2
Corrections
Corrections must be made to the above particle in a sphere model to accurately
describe QD electronic states and optical transitions. The conduction band, being
comprised of selenium 4p orbitals, is 6-fold degenerate. So spin-orbit coupling and
lattice structure must be taken into account to lift the degeneracy. Spin-orbit coupling breaks the valence band into two bands, a split-off band with total angular
momentum J
=
1/2, and a remaining band with J
=
3/2. The J = 3/2 band is
further split into heavy-hole and light-hole bands by the crystal field splitting of the
wurtzite CdSe lattice. This new set of bands, however, still does not accurately describe optical transitions observed in QDs. Valance band mixing is required, which
combines angular momentum for the atomic basis (the Bloch periodic function) and
the orbitals recovered from the spherically confined envelope function (the Bloch plane
wave function). This treatment, along with S-D mixing within the envelope function
orbitals themselves, accurately describes avoided crossings observed in QD optical
transitions indicative of state coupling in the valance band.41,39,40
1.2.3
Fine Structure
A final adjustment must be made to completely describe QD optical transitions, including band edge fine structure, long photoluminescent lifetimes, and the "dark exciton". 42 ,4 3 The band edge exciton is 8-fold degenerate when the QD is approximated
as spherical and with a cubic lattice. This degeneracy is lifted when the asymmetry
of the wurtzite crystal lattice, the prolate shape of the QD, and the electron-hole exchange interaction are included. These considerations are included as perturbations
to the spherical model. The resulting fine structure reveals that the lowest energy
transition, or the "dark exciton" is optically forbidden and that emission must occur
via less efficient phonon assisted emission, 4 2,43 ,3 9,4 4 resulting in a photoluminescent
lifetime that is orders of magnitude longer in QDs than in molecules.
25
(a)
x 10
2.5
(b)
100
2
10'
V ON
OFF
10-2
U'
1
c-
1.5
103
1Q-4
10-4
0.50
10-6
20
40 60 80 100 120
Time (s)
T
0.1
1
10
Time (s)
100
Figure 1-5: (a) An emission intensity trace from a single CdSe/ZnS core/shell QD
exhibits blinking. (b) Probability distributions for ON/OFF duration measured from
single QD intensity traces follow a power law distribution (adapted from Shimizu et
al., 2001).
1.3
Emission Dynamics
Shortly after the development of single molecule fluorescence microscopy, 4 5'4 6 colloidal
QDs
were individually investigated and single molecule emission dynamics were ob-
served that had previously been hidden by ensemble averaging. 42 Fluorescence intermittency, or blinking, and spectral diffusion in the single QD emission were revealed.
Spectral diffusion is the random movement of the band edge emission peak to different energies, often within 100 meV, and inhomogeneously broadens the QD emission
spectrum. 47' 48' 4 9 Blinking is characterized by the stochastic switching between an
emissive (ON) state to a dark, non-emissive (OFF) state during continuous excitation. 50 ' 5 1 An example of blinking is shown in figure 1-5a.
Blinking behavior was shortly thereafter quantified by histogramming the ON
and OFF durations.5 2,5
The resulting normalized distributions followed power law
behavior with the form
P(t) = Ct-exp(-t/T),
(1.2)
where a is the power law slope, generally equal to ~ 1.5, and T is the exponential
26
cutoff time observable in single QD blinking experiments for the ON time distribution but not for the OFF time distribution (Fig. 1-5b). 5 1 5'
This exponential cutoff is
effectively an upper limit to possible ON times, which is much shorter than the OFF
time upper limit, and has temperature and excitation flux dependence. Experiments
have shown that the ON time upper limit scales with biexciton formation, implicating
biexcitons (and thus Auger ionization) as a source for QD ON to OFF transitions.5 4
55
The power law behavior, however, introduces significant challenges to modeling QD
blinking mechanism. A true power law is scale invariant, so any mechanisms underlying QD blinking must have kinetics that span many orders of magnitude in time.
Additionally, power laws are less tractable without introducing artificial boundaries,
usually dictated by parameters of the experiment that can vary from experiment to
experiment.
1.3.1
Blinking Models
Despite the challenges of interpreting power law behavior, various blinking models
have emerged to describe the highly distributed blinking kinetics. The early charging model relied on an Auger ionization mechanisms, where an excited charge carrier (presumably an electron due to small mass) is ionized, leaving behind a charge
that quenches subsequent excitations via Auger recombination.
50 ,56
This explanation,
however, failed to explain the power law behavior of the ON time distribution with
a single ionization barrier.1 8 This was rectified with the inclusion of a distributed
tunneling barrier for ionization, 5 7 but the observation of power law blinking in QDs
tethered to a glass slide in aqueous solution suggested that complete ionization and
subsequent loss of the ejected electron is not likely the root cause of blinking. 58 A
complementary explanation in the deep trap model, describes ionization as localization of the charge carrier on the surface of the QD without full charge ejection. This
model is still widely discussed, but experiments have shown that a single trapped
charge is not sufficient to fully quench a
QD.
An alternate model escapes QD charging by introducing multiple recombination
centers (MRC). 60 '6 ' In the multiple recombination center model, a QD is rendered dark
27
when one or more recombination centers are open, providing a pathway through which
an excitation can recombine non-radiatively. Closing a number of these recombination
centers can be described with distributed kinetics. Lastly, the diffusion controlled
electron transfer model adequately describes power law blinking and spectral diffusion
behavior in QDs. 1 , 2 This model describes the ON and OFF states as anomalously
diffusing along a parabolic potentials in energy space, with a transition occurring
when the two states are in resonance.
1.3.2
Ensemble Behavior
Complex underlying blinking kinetics leads to interesting ensemble emission properties. Namely, emission dimming and brightening are observable in QD ensembles and
are purely statistical in nature (assuming photochemical processes are controlled).
The ensemble emission dynamics rely on the single molecule ON and OFF time distributions.5 3
Although an exponential cutoff is only observable for the ON time
distribution in single QD blinking experiments, an OFF time cutoff exits and is observable with ensemble dimming. In an ensemble measurement, the emission begins
to decrease as the experiment progresses past times longer than the ON time cutoff,
where it's almost certain that all QDs have turned off at least once. The OFF time
distribution, however, still has probability density beyond the ON time cutoff, so it
is possible to have QDs that have remained OFF for most of the experiment. An
alternate view is that at long experiment times, it becomes likely that some QDs will
sample very long OFF times while ON times are capped by the ON time cutoff. This
treatment illustrates the existence of an OFF time cutoff, because ensemble dimming
levels off to a steady state intensity after the maximum OFF time can be sampled. An
illustration of how ensemble emission relates to the ON/OFF time power law cutoff,
r in equation 1.2, is shown in figure 1-6.
28
-
10
ON
OFF
.9
--10-
-
-
10-
10
-
10'
102
10
Log10(t)
Figure 1-6: Experimental ensemble emission intensity on a log-log plot. The clitoff times, T, for the single QD ON OFF time probability distributions are labeled.
Adapted from Chung and Bawendi, 2004.
29
30
Chapter 2
Optical Trapping and Two-Photon
Excitation of Colloidal Quantum Dots
using Bowtie Apertures
2.1
Introduction
Optical tweezers have been a powerful tool to fix, control, and manipulate small objects since they were first demonstrated.
3
The introduction of plasmonic structures
has greatly advanced the field of optical trapping in the last decade. These structures
provide enhanced, localized, electric fields that require lower incident flux and can
trap smaller particles when compared to free space trapping.6 4,65 , 66' 67 68' 69 Trapping
is further enhanced in plasmonic apertures by self-induced back-action (SIBA), 70 a
positive feedback mechanism that increases trapping force due to dielectic loading of
the aperture when a particle is trapped. Recently, there have been many plasmonic
nanoapertures designed for trapping particles as small as tens of nanometers. Trapping with plasmonic apertures has been performed with circular 7 0 and rectangular
apertures. 71 Introducing a pinch point into the aperture, double nanoholes were used
to trap a 12 nm silica bead," and bowtie apertures were fabricated on films and on
fiber tips to implement 20 nm polystyrene bead trapping and 50 nm bead manipula31
tion. 73 The opposing prongs at the pinch point of the aperture act as dual sharp tips
to greatly enhance electric fields in the gap, 74 giving rise to a localized field gradient
suitable for optical trapping. This confined fundamental gap mode has also been used
to provide a narrower near field pattern for lithography, 75 brighter scanning near-field
optical microscopy, 76 and enhanced molecule fluorescence. 77
Various types of particles have been used in optical trapping studies, including
gold nanoparticles, 78 nanorods, 79' 80 globular proteins, 8 1 single-cell organisms, 8 2 ,6 6 and
polystyrene spheres with7 3 and without emissive dye.7 2 Colloidal quantum dots (QDs)
are attractive candidates for optical trapping and simultaneous electronic excitation
because their high index of refraction 83 increases trapping force, and their broad continuum of excited states makes them strong absorbers. 84 ' 8 Quantum dots have been
optically trapped 85 and nonlinearly excited in free space, 86' 8 7 but trapping with plasmonic structures renders the QDs non-emissive due to interactions with the nearby
metal.88 '89
In this experiment, QDs were overcoated with a silica (Si0 2 ) shell to mitigate
emission quenching and provide additional dielectric material to increase trapping
efficiency. The bowtie apertures were fabricated by collaborators in the Loncar group
at Harvard University using a lift-off procedure to provide larger aperture quantities
for higher throughput device testing. Bowtie apertures were used to trap silica coated
quantum dots (scQD) with a diameter of 30 nm with a trapping laser intensity of
1.56 MW/cm
2
at 1064 nm. Because of the strong field confinement inside the bowtie
aperture, 640 nm scQD emission was detected following two-photon excitation by the
1064 nm trapping laser. The enhanced two-photon excitation eliminates the need for
a separate excitation source and results in a system that self reports via emission
when trapping is achieved. Simulations show theoretical trapping performance and
experimental examples of single scQD trapping with simultaneously recorded laser
transmission and emission.
Scanning electron microscope (SEM) images of the apertures used in the experiments are shown in figure 2-la and 2-1b overlapped with field intensity enhancement
simulations, illustrating the dominant gap mode in the aperture when the trapping
32
45 m0.13
(c)
40
0.12
35
0.11
(Fig. 1b)
30
25
0.1
0(Fig.
1a)
1a)
Co(Fig.
E 0.09
C
20
S150.0815
10
0.07
5
0.06
1064 nm
____._._50
800
0
900 1000 1100
Wavelength (nm)
nm
1200
Figure 2-1: (a)(b) SEM images of the bowtie apertures used in the experiments, overlapped with field intensity enhancement profiles at 1064 ini. The confined gap mode
is dominant when the polarization is across the gap. (c) The simulated transmission
spectra of the two apertures used in trapping experiments, showing peak resonances
are blue-shifted from the 1064 nm trapping laser. (d) Transmission electron microscope image of the silica coated quantum dots used in trapping.
beam polarization is oriented across the gap. Enhancement is a unitless factor that
scales the intensity in the gap relative to the free space intensity. Both apertures, with
gaps of 38 nmn and 56 nm, were used to successfully trap scQDs. Given that the field
enhancement is lower in the 56 rim gap aperture, the required trapping laser intensity
is higher and the calculated trapping potential suggests it should only be able to trap
larger particles. The aperture is sandwiched between water and an underlying silicon
nitride (SiN) membrane, so it forms a low-Q Fabry-Perot cavity whose resonance can
be tuned by film thickness. 90 A 130 nm thick silver film was used to achieve resonances centered at 850 nm and 915 nmn (Fig. 2-1c), thus satisfying the requirement
imposed by the SIBA70 mechanism for a peak transmission resonance slightly blueshifted from the trapping laser. A transmission electron microscope (TEM) image of
the scQDs used in trapping shows particles with a CdSe/CdS core/shell,
38
center
and total sizes that are ~ 30 nmn in diameter (Fig. 2-1d), with a mean hydrodynamic
diameter of 39.2 nm as measured by dynamic light scattering (Fig. 2-3).
33
0.8
0.6
0.4
0.2
0
550
575
600
625
650
Wavelength (nm)
675
700
Figure 2-2: Normalized absorption (blue) and emission (red) spectra for scQDs. Continuous wave 532 nm excitation was used as an excitation source for the emission
spectrum.
2.2
2.2.1
Experimental Methods
Silica Coated Quantum Dot Synthesis and Characterization
Core/shell CdSe/CdS colloidal quantum dots (QDs) were synthesized as previously
described. 3 8 3 7 Silica overcoating was performed by loading 30 mL of anhydrous cyclohexane into a 100 mL round bottom flask. Under vigorous stirring, 4.75 mL of Igepal
CO-520 was added. After stirring for 10 minutes, 1 mL of QD-cyclohexane solution
(3 pM) was injected into the reaction followed by slowly adding 150 PL tetraethyl
orthosilicate (TEOS, 99%). After another 10 minutes of stirring, 0.5 mL of ammonium hydroxide solution (28% in water) was injected dropwise into the solution. The
final reaction solution was stirred for 18 hours at room temperature before purifying
the scQDs via precipitation using ethanol (~20 mL) and collecting by centrifugation.
The scQDs were washed with ethanol twice more and finally dissolved into 2 mL DIwater before storage at 4C until use. Normalized absorption and emission spectra
are shown in figure 2-2.
34
.
.
.
.
--
.
25
i--
Filtered
Unfiltered
20
15
10
5-
0
20
40
80
60
Particle Size (nm)
100
120
Figure 2-3: Size distribution of scQDs before (green) and after (blue) filtering measured with DLS. Sizes were calculated by volume.
Prior to some measurements, the scQD solution was passed through a 20 nmn pore
syringe filter (Whatman) to reduce the mean particle diameter. Dynamic light scattering (DLS) was performed on filtered and unfiltered particles and results calculated
by volume are plotted in figure 2-3. Mean hydrodynamic diameters are 21.1 nm and
39.2 nm for filtered and unfiltered particles, respectively. It should be noted that because DLS measures the hydrodynamic diameter of particles, these results are likely
an overestimation of the actual particle sizes.
2.2.2
Aperture Fabrication
Bowtie apertures were fabricated by collaborators in the Lonear group at Harvard
University using a lift-off procedure on a 100 nm thick silicon nitride (SiN) membrane with a silicon scaffold from Norcada Inc. The SiN substrate was spin-coated
with a negative tone electron-beam resist (FOX-16, Dow Corning) and bowties were
patterned with e-beam lithography (Elionix ELS-F125). The sample was developed
in tetramethylammonium hydroxide for 17 seconds, leaving behind 800 nm tall bowtie
posts. Electron bean evaporation (Denton) was used to evaporate a 2 nm layer of
titanium followed by a 130 nin layer of silver. The sample was briefly scrubbed with
35
a swab prior to performing a 130 second, 5:1 buffered oxide etch.
Scrubbing the
sample is crucial for high device yield because it breaks posts extending above the
silver surface that may have metal particles deposited on the sidewalls. Even though
e-beam deposition is directional, a small amount of sidewall deposition is unavoidable
and causes incomplete lift-off and poor device fabrication.
2.2.3
Simulations
In order to quantify and evaluate the trapping capability of the apertures, finitedifference time-domain (FDTD) simulations (Lumerical Solutions, Inc.) were performed by collaborators in the Lonear Group at Harvard University and the trapping
potential was calculated (Fig. 2-4). Simulations were performed on the 38 nm gap
aperture (Fig. 2-la) with the incident trapping beam set to have a 500 nm beam
waist focused on the entrance of the aperture. The scQD was simulated as a 6nm
CdSe core with silica coatings of varying thickness to produce final diameters of 20
nm, 25 nm, and 30 nm, and was placed close to the silver wall to get the strongest
trapping potential possible. The field intensity surrounding the scQD was recorded,
scaled to the experimental incident flux of 1.56 MW/cm 2 , and used to calculate trapping potential. The calculated trapping potential exhibits two local minima due to
field enhancement occurring on both faces of the aperture from operating near the
1st-order Fabry-Perot resonance, 90 with the deeper trapping potential at the waterSiN interface. Optical trapping is considered favorable when the trapping potential
overcomes the ambient thermal energy kBT, which was observed for particles of at
least 25 nm in this system at the water-SiN interface. The trapping potential at the
front surface of the aperture did not overcome kBT, regardless of particle size. However, factors not accounted for in the simulations could potentially enable trapping
particles smaller than 25 nm with this system. Van der Waals forces between the particle and the surrounding aperture surfaces could facilitate trapping when potentials
do not overcome kBT of ambient thermal energy, and reduced degrees of freedom for
particle motion inside the aperture should reduce the particle's kinetic energy, making
escape from the aperture more difficult.88
36
100
(a)
Ag
75
m
:3
v
CD
25
0
(b)
0
-~-0.5-_____
0
30 nm
-120
-90
-60
-30
0
30
z (nm)
Figure 2-4: (a) The simulated field intensity distribution inside the aperture showing
field enhancement on both faces of the aperture. The scQD is shown in its final
position at the bottom of the aperture touching the SiN membrane. (b) Potential
energy calculation results showing that scQDs of at least 25 nm will have a potential
lower than - 1 kBT at the bottom of the aperture.
37
0
-0.6
-
-0.4
-
-0.2
-0.8-
-1 -1
-1.2
-
0
30 nm
-1.4-
35 nm
-1.6-
44 nm
-1.811
-120
-90
-60
-30
0
30
z (nm)
Figure 2-5: The calculated potential for particle trapping with the 56 nm aperture in
figure 2-1b.
Figure 2-5 shows the calculated potential for the 56 nm aperture in figure 21b. It exhibits the same dual minima characteristic as the calculated potential for
the 38 nm aperture. Simulations show the minimum particle size this aperture can
trap is 35 nm by overcoming kBT of ambient thermal energy. However, non-optical
mechanisms such as van der Waals force and reduced particle degrees of freedom could
enable trapping of smaller particles.
2.2.4
Packaging and Instrumentation
Prior to trapping experiments, the aperture film was packaged with an aqueous scQD
solution.
2
A reservoir was made by cutting a 3 x 3 mm square from a 30 pm thick
polydimethylsiloxane (PDMS) spacer on top of a 80 pm thick cover slip. Then a small
drop of scQD solution (0.07% w/v) was placed in the reservoir and the aperture film
was placed face down on top of the reservoir.
Optical trapping was achieved by transmitting a continuous wave (CW) 1064 nm
trapping beam through an aperture packaged with scQDs as shown in figure 2-6.
The optical quality of a 1064 nm trapping beam (Laser Quantum Ventus 1064) was
cleaned with a polarizing filter and a 1064/10 nm laser line filter, expanded, and
38
slightly defocused to correct for chromatic aberration of the trapping objective. The
trapping objective was a 100x (1.25 NA) oil immersion objective that formed a spot
radius of 1 pum with 1.56MW/cm 2 of incident flux at 1064 nm. Emission from trapped
scQDs was collected with the same objective, separated from the 1064 nm trapping
beam with a 900 nm short pass dichroic mirror, and sent to either a silicon avalanche
photodiode (APD, Perkin Elmer SPCM -AQRH-13) or a spectrometer/ CCD camera
combination (Princeton Instruments Acton SP2750A/Pixis 1024) for detection of twophoton excitation upon trapping. Above the packaged film, 1064 nm transmission
intensity through the aperture was collected with a 60x (0.7 NA) air objective and sent
to a Ge photodiode (Thorlabs DET50B) to monitor jumps in transmission intensity
coinciding with trapping events. Sample positioning was achieved by a 3-axis stage
(Thorlabs Nanomax-TS) and broadband light from a quartz tungsten halogen lamp
was ported into the instrument for sample alignment and remained off during trapping
experiments.
2.3
2.3.1
Results
Single Particle Trapping
Single scQD trapping in the 38 nn aperture using scQDs shown in figure 2-1 is
demonstrated in figure 2-7 and is characterized by a stepwise increase in both emission and transmission intensities at 50 seconds. Intensity fluctuations are observed in
the emission channel at -110 seconds followed by a gradual decrease in intensity. Corresponding dynamics in the transmission channel are absent or undetectably small,
suggesting scQD emission dynamics (i.e. blinking, bleaching) are responsible for the
fluctuations observed in the emission channel.
Alternatively, the emission channel
may be far more sensitive to very small changes in particle position due to the nonlinear nature of two-photon excitation, resulting in large fluctuations in the emission
channel without corresponding transmission channel dynamics. The trapped particle
is likely around 30 nm in diameter given the mean particle size as determined by TEM
39
GePD
Liquid core
fiber
QAlignment lamp
'Spectrometer
640/30 nm band pass
-
720 nm short pass
1064/10 nm laser line
+ polarizing filter
1064 nm
laser
Dichroic Beam expander
Figure 2-6: Instrument schematic for simultaneous trapping with 1064 nm laser (gray
beam) and scQD emission detection at 640 nm (red beam).
40
x104
(a)
W2-
0
17
50
100
150
200
100
150
200
(b)
<0.3
0
0.29
C~c
CU
50
Time (s)
Figure 2-7: The (a) emission and (b) 1064 nm transmission channels show a stepwise
increase in signal at 50 seconds, suggesting individual scQD trapping.
and DLS, as well as the aperture trapping potentials and dimensions. It is unlikely
that the particle is larger than the gap width of 38 nm because larger particles are
prevented from peak trapping potentials at the bottom of the aperture.72
2.3.2
Spectrally Resolved Trapping
Figure 2-8 shows spectra collected from the same aperture before and after scQD
trapping and serves as evidence for two-photon excitation in the absence of a subbandgap excitation source. The spectral range between 520 nm to 700 nm is dark prior
to trapping, but a scQD emission peak appears at 640 nm after the particle is trapped.
Additionally, the absence of detected signal at 532 nm rules out second harmonic
generation in the aperture by the trapping beam. The emission and transmission
intensities for this trapping event are shown in figure 2-9, along with the scQD linear
41
1000
900
800
-
Before
-
After
700CL
600-
'C,
C
500
400300200
100
0
525
550
575 600
625 650
Wavelength (nm)
675
700
Figure 2-8: QD emission spectra before (blue) and after (red) optical trapping and
two-photon excitation.
emission spectrum excited with a 532 nm CW excitation source in figure 2-2. Given
the simulated peak intensity enhancement of ~ 60x for this aperture with a scQD
in the aperture (data not shown), the enhanced excitation flux at the trapped scQD
is calculated to be 93.60 MW/cm 2 . This enhanced excitation flux is appropriate for
two-photon excitation of QDs86 and was achieved with a very low incident flux of 1.56
MW/cm
2
at 1064 nm, relying on plasmonic enhancement from the aperture.91
The emission and the transmission channels for the trapping event in figure 2-8 are
shown in figure 2-9. The emission intensity was calculated by summing the spectral
intensity of each frame from spectrometer/CCD camera detection scheme. Spectra
were collected with 1 second integration times, resulting in poor time resolution in
the emission channel. Trapping occurred at 279 seconds and was stable until the trial
ended at 300 seconds.
2.3.3
Multiple Particle Trapping and Emission Dynamics
Emission dynamics for two trapping experiments are shown in figure 2-10, with emission and transmission channels for the 56 nm gap aperture in 2-10a and 2-10b, and
emission and transmission channels for the 38 nm gap aperture in figure 2-10c and
42
x 1
-
0n
C.
15 - (a)
c0 10-
0
E
280
285
290
295
300
280
285
290
295
300
.
275
0.25
0c
U)
(n
E
U)
0.2
~-0.151
275
Time (s)
Figure 2-9: (a) Emission and (b) 1064 nm transmission channels for spectrally resolved
emission detection presented in figure 2-8.
43
2500
10000 - (C)
(a)
.
CL 2000
r 1500
500
1000
500
200
0. 34
220
240
260
280
300
200
220
240
260
280
300
220
260
240
Time (s)
280
300
0.084
-(b)
(d)
.82
- 0.32
Dfl
E 0.3
0.08
0.078
0.076
-
.
0.074200
220
240
260
Time (s)
280
300
200
Figure 2-10: (a) Emission and (b) 1064 nm transmission for filtered scQDs in the 56
nm aperture shows evidence to QD blinking inside the optical trap. Multiple trapping
events are detected in the (c) emission and (d) 1064 nm transmission channels for
filtered scQDs in the 38 nm aperture that exhibit rapid quenching at 265 and 280
seconds in the emission channel only.
2-10d. Both experiments used filtered scQDs (Sup. Fig. 2-3) with a mean hydrodynamic diameter of 21.1 nm. Successful trapping of particles smaller than the minimum
size predicted by force calculations in both of these apertures can be rationalized by
two possible explanations. Firstly, non-optical mechanisms described in the main text
may contribute to trapping. Secondly, the trapped particle size could lie in the tail
end of the size distribution as determined by DLS measurements (Fig. 2-3), which
extends out to 45 nm in diameter for the filtered particles.
The first trapping experiment using the 56 nm aperture shows an increase in
1064 nm trapping laser transmission at 200 seconds (Fig. 2-10b), indicating particle
trapping. The corresponding emission trace (Fig. 2-10a), however, does not exhibit
emission until 260 seconds into the experiment, which is intermittent and resembles
QD blinking. The second trapping experiment using the 38 nm aperture shows initial
trapping in both the 1064 nm transmission (Fig 2-10d) and the emission (Fig 210c) at 210 seconds into the experiment. Subsequent trapping events, however, show
subtle increases in 1064 nm transmission at 265 and 280 seconds, with corresponding
emission events that rapidly decay in intensity after trapping.
44
2.4
Conclusion
The system presented here offers unique opportunities to study light-matter interactions inside a plasmonic cavity. The particle can be controllably placed and removed
from the aperture by toggling the trapping beam, allowing for convenient measurements of the emitter inside and outside of the cavity. The optical trap also provides
natural alignment of the nanoparticle to the peak field intensity in the aperture, alleviating concerns over particle placement in a resonant cavity.92 And the lift-off
nature of the aperture fabrication can provide large arrays of apertures, allowing for
high throughput experimentation or cavity design optimization. Lastly, experiments
need not be limited to scQDs. Emitter-cavity interactions can be investigated for
alternative quantum emitters including nitrogen-vacancy centers in diamond, 93 semiconductor nanorods, and hybrid structures.' 7
A particularly elusive experiment enabled by this system is the measurement of
a single particle broadband absorption spectrum. The aperture provides enhanced
transmission and a sub-diffraction limited window,94 combined with the ability to
toggle particle placement for recording signal and blank spectra. The effective mode
area9 5 was calculated for the aperture in figure 2-la at the scQD band edge (630 nm)
with a particle present in the aperture. The mode area of 3.31 x 104 nm 2 , is roughly 20
times smaller than a diffraction limited spot. With these advantages, measurement
of a single particle absorption spectrum can be achieved in integration times that
are shorter than our observed particle trapping duration (Fig.
2-7).
Additional
signal improvements are possible by reducing the temperature to consolidate oscillator
strength into narrower transition peaks, as well as tailoring the aperture design to
minimize effective mode area for this specific measurement. Changes in the aperture
transmission spectrum itself with the introduction of a transparent dielectric particle
are expected to be broad and relatively featureless in the visible region, but still must
be considered.
This platform also provides an alternative way to study QD emission dynamics in
enhanced electric fields. In particular, the Purcell effect" 6 increases the radiative rate
45
of an emitter in an enhanced electric filed, making emission competitive with much
faster non-radiative channels and thus allowing for investigation of pathways involved
in quenching and blinking.97 ,98 Evidence for scQD emission dynamics was observed,
including blinking and rapid quenching upon trapping.
In conclusion, bowtie apertures were designed and fabricated to optically trap 30
nm insulated QDs, yielding a system with stable single particle trapping and robust
two-photon excitation at modest flux. Lift-off aperture fabrication was introduced
and FDTD simulations revealed favorable trapping conditions that may be further
aided by non-optical mechanisms. This system may enable the high-throughput experimentation of light-matter interactions and multiphoton processes in various types
of emitters.
46
Chapter 3
Thermal Recovery of Colloidal
Quantum Dot Ensembles following
Photoinduced Dimming
3.1
Motivation
Colloidal quantum dots (QDs) have been considered for many applications due to
an array of attractive qualities, including stability, ease of synthesis, a broad excitation band, and a narrow, tunable emission linewidth. Fluorescence intermittency
(i.e. blinking) in individual QDs, however, reduces QD efficiency and can introduce
complications to potential applications. 99,100,'101 At the ensemble level, blinking is
manifested as photodimming to a steady state intensity53 and can limit the performance of applications that require high fluxes 5 1 '5 4 ' 55 and large collections of QDs.
Although synthetic procedures have been developed to mitigate blinking in CdSe
QDs,1 0 2,3 0 ,103,35,
104,38
researchers still seek a full understanding of mechanisms that
cause blinking as it affects other types of colloidal semiconductor nanoparticles including indium phosphide QDs, 18 indium arsenide QDs,1 9 cadmium selenide /cadmium
sulfide (CdSe/CdS) core/shell nanorods,
05
and lead sulfide QDs.2 0 ,2 1
Recovery from the photodimmed state is slow and must occur with thermal energy
47
in the absence of incident flux.106,53,107 The transition from a dark (OFF) QD to an
emissive (ON) QD can be examined by quantifying the time scale of recovery with
a range of ambient thermal energies. Allowing QDs to recover in the dark simplifies
the recovery process by removing potential light-assisted recovery channels available
with excess excitation energy. Using controlled amounts of thermal energy to drive
recovery simplifies complex power law behavior observed in single particle blinking
studies where QDs are continuously excited. 51
3.2
3.2.1
Experimental Methods
Synthesis, Packaging, and Instrumentation
Quantum dot CdSe cores were synthesized as previously described 38 and packaged
in an air-free environment. To package QDs, cores were precipitated out of growth
solution by heating and the addition of acetone followed by centrifugation. The pellet
was dissolved in hexanes and stored at 4 for at least 24 hours to allow precipitation of
excess salts. Hexanes were then evaporated prior to transfer to a glove box where they
were purified via precipitation twice more with anhydrous acetone and centrifugation.
The QDs were dissolved in anhydrous toluene with a 1 % w/v solution of 350 kD
poly(methyl methacrylate)(PMMA). Trace CdO was added to the QD solution to
reduce clustering and improve surface passivation before the solution was filtered
with a 20 nm pore syringe filter. The QD solution was spun-cast onto a glass cover
slip at 1000 rpm for 5 minutes to achieve a PMMA film thickness of 20 nm, previously
calibrated with a profilometer (Dektak 6M). The QD film was packaged by stacking
the cover slip face down on a rubber washer and a 1 mm thick glass slide, and
then sealed with UV-curing epoxy (Thorlabs NOA61) followed by 15 minutes of UV
exposure to cure the epoxy and photopassivate the QDs. 108
Measurements were performed on a home-built epifluorescence microscope with an
avalanche photodiode detector (Perkin Elmer SPCM-AQRH-13). A 4x air objective
(Olympus, 0.12 NA) was used for excitation and collection, and 532 nm continuous
48
wave excitation source (Laser Quantum Ventus) provided a flux of 90 W/cm 2 . Sample temperature was controlled with a ceramic heating element (Thorlabs HT19R)
in direct contact with the top surface of the sample. Spot position and sample exposure were controlled by galvo mirrors (Thorlabs GVS012) and an optical shutter
(Uniblitz D122/6932), respectively. Custom Matlab (MathWorks) drivers controlled
all instrumentation which allowed for scripted automation of the experiment.
3.2.2
Procedure
Photodimming was induced and recovery over time for CdSe cores was measured with
an epiflurosence microscope. Cores were chosen for their large degree of blinking and
were well dispersed in a poly(methyl methacrylate)(PMMA) thin film in an air-free
environment. The film was sealed prior to air exposure to eliminate oxidation during
experimentation.1 09 Five excitation spots were randomly chosen from the sample
plane and photodimmed for 30 seconds each with 90 W/cm 2 of continuous wave 532
nm light. Each of the five excitation spots was probed once for recovery following five
logarithmically spaced wait times ranging from 4.6 minutes to 3.4 hours as illustrated
in figure 3-1. It was crucial to probe each spot only once for recovery to eliminate the
effects of light-assisted recovery. Recovery over time was measured in triplicate at 295
K, 318 K, 338 K, and 358 K. Laser intensity was recorded during each experiment
and used to correct emission data for small fluctuations and drift in laser intensity.
3.3
3.3.1
Results
Scaling and Global Fit
In order to standardize recovery across different spots on the QD film, each recovery
intensity was scaled to the extent of dimming for its corresponding dimming trace. Explicitly, the recovery probability was calculated for each spot as P(t)
=
)Im
dim
where Irecov(t) is the emission intensity after wait time t, and
Im,"a
inal
Irecovt (t)-If
uni al
and
dim
im"a
are
the initial and final emission intensities of the dimming trace, respectively. Recovery
49
2.4
x 10
2.3
2.2
2.1
o. 2
.1.9
9
1.8
1.7
1.6
1.5
1.4/
0
10
20
30
280
Time (s)
1890
12400
Figure 3-1: Example of collected emission data for illustrative purposes. Photodimming and recovery for three separate excitation spots. Each spot was dimmed for 30
seconds and probed for recovery only once at logarithmically spaced wait times.
probability, P(t), is a cumulative distribution function describing the relative number
of QDs that have transitioned from OFF to ON at a time less than or equal to t, and
can also be thought of as S(t) = 1 - P(t), where S(t) is the survival probability of the
OFF state. This treatment relies on the assumption that dimming is caused solely
by individual QD blinking behavior and that non-reversible photochemistry does not
change the brightness of individual QDs over the course of dimming. All recovery
probabilities at each temperature are presented in figure 3-2.
The scaled recovery data were fit globally to a stretched exponential with the form
P(t) = 1 - exp [-(kot)0] ,
(3.1)
where ko is a characteristic rate constant that was allowed to vary with temperature and / is the stretching exponent that was shared across all temperatures. This
fitting procedure assumes that the mechanism for recovery is common across temperatures, with the temperature dependent effect being captured by the characteristic
rate constant ko. This global fitting procedure reduces the total number of fitting
parameters to five and resulted in a residual sum of squares (RSS) of 6.20 x 10-1.
50
10
0.9
0.80.7-
DO0.605
0.4
0.3
.
0.2
0.1
00
0
295K
0
0
318 K
338 K
0
358 K
0
0
2000
4000
6000 8000 10000 12000
Time (s)
Figure 3-2: Recovery probabilities collected at all of temperatures. Data from each
temperature is offset for clarity.
The values extracted for these data are 0.326 for the common stretching exponent /3,
and 0.72 x 10'4 s-, 0.97 x 10-1 s-1, 1.43 x 10-5 s-1, and 2.77 x 10-5 S-1 for k0 at
295 K, 318 K
,
338 K, and 358 K, respectively. A global power law fit with the form
p(t) = 1 - cta was also considered for these data but provided a larger RSS of 0.0186.
Averaged recovery probability at different temperatures and global fits are presented
in figure 3-3.
The stretched exponential function (Eq. 3.1).
ous physical processes" 0'
111, 112,113,114,115
has been used to describe vari-
and can provide physical insight by further
analysis of the functional form. Namely, the inverse Laplace transform of 1 - P(t)
transforms the distribution from the time domain to rate coefficient domain, giving
1
a probability density function of rate constants, H(k), present in the system.' 1"'
5
The closed form solution for H(k) is
H(k) =
r
! sin(7rrn) (01+1)
(3.2)
/3 is
the stretching exponent extracted from the global fit and F is the gamma
function."'
Equation 3.2 was numerically evaluated for a range of rate constants
where
51
(a)
-
-
-.
0.5
-
0.3 -
,
.-
9
0.3
.2
0295K
*
0.1-
2500
,@
5000
7500
Time (s)
-
0.2
-
0.1
318K
338 K
1
0
-
f
-
0.5
(b)
358K
10000
0
12500
0.1
0.2
0.3
ket
Figure 3-3: (a) Temperature-dependent recovery with global stretched exponential
fit. (b) Data plotted on a time axis corrected by k0 emphasizes the validity of the
fitting procedure. Data falls along a stretched exponential with #
0.326 (dashed
line).
normalized by k0 , or H(k/k0 ) (Fig. 3-4). Since all of the temperature-dependent
information in figure 3-3 is captured by k0 , increasing the temperature in the experiment only compresses the time axis while the underlying functional form for recovery
remains the same (Fig. 3-3b). Thus, the underlying temperature-independent functional form for the rate constant distribution can be recovered by using the relative
.
rate constant axis, k/k 0
3.3.2
Modeling
The recovery process was then modeled as a collection of first-order recovery processes
(i.e. a sum of single exponentials), resulting from a distribution of activation energy
barriers. 116 To transform the rate constant axis in figure 3-4 to an activation energy
axis, however, an accurate value for the pre-exponential factor in the Arrhenius equation is required. An Arrhenius plot was constructed (Fig. 3-5) with a linear fit to
the Arrhenius equation, ln(ko)
and T is temperature.
=
ln(Ao)
-
9,
where kB is the Boltzmann constant
A reference activation energy of Ea,o
52
188.03 meV and a
2
1.8
1.6
-
1.4
-
1.2
-
0.8
0.6
0.4-
0.2
0
0.5
0
1
k/kO
2
1.5
Figure 3-4: Temperature-independent rate constant probability distribution.
pre-exponential frequency factor of Ao = 0.0105 s- were recovered at a single point,
ko, in the underlying rate constant distribution.
The activation energy distribution was then calculated by a change of variables
from the rate constant probability density function. 11
The unscaled rate constant
axis was transformed to activation energy using the Arrhenius equation, while the
activation energy probability density function was calculated by relation with unscaled
rate constant distribution,
H(Ea) =
-
A0
kBT
exp(-Ea/kBT)H(k).
(3.3)
The recovered value of AO = 0.0105 s-1 was used under the assumption that the
exponential pre-factor is constant for all activation energies. The probability distribution H(Ea) plotted against the activation energy axis is shown in figure 3-6 and
represents the activation energy probability distribution measured for these QD cores.
The distribution has significant probability density at low activation energy values
and grows to a peak at 200 meV before decaying to zero.
53
-10.4
-10.6
0-
-10.8
4%-
-
-
-11
-11.2-
-
-11.4
''
2.8 .9
-
-11.6
3.1
3
.2
33
3
-11.8-12
1/Temperature (K~1)
x 10-3
Figure 3-5: Arrhenius plot of ln(ko) as a function of inverse temperature (K-') with
linear fit yields EaO = 188.03 meV and AO = 0.0105 s-
x10-3
54.5
43.53w
I
2.52-
1.5
1;
0.500
50
100
150
200
250
300
350
Ea (meV)
Figure 3-6: Probability distribution of activation energy barriers for air-free CdSe
cores.
54
3.4
Discussion
This underlying distribution of activation energy barriers intuitively explains the nonlinear thermal recovery observed in figure 3-3. The significant probability of lowenergy barriers on the order of room temperature thermal energy (kBT = 25.4 meV)
leads to barrierless recovery and explains the fast recovery observed (10% - 15% recovery after 280 s). Additionally, the large barriers are crossed exponentially faster as
temperature is increased, which is consistent with increased final recovery probability
P(t) with temperature. Finally, time to complete recovery is extremely long because
large barriers with energies ~ 10 times kBT are highly probable.
This experiment provides a clear picture of QD recovery in terms of a distribution of first-order processes, but care must be taken to avoid over-interpreting these
results. Nonlinear thermal recovery from distributed underlying kinetics is indicative
of progressively depleting random sinks and is consistent with commonly proposed
60 61
blinking mechanisms, including the multiple recombination center (MRC) model ,
and with the deep carrier trap model. 56' 59 Both of these mechanisms are possibly
activated processes with distributed kinetics, but the results presented here are not
sufficient to discriminate between the two mechanisms. It is also important to note
that the activation energy distribution does not provide any insight into the relative
energy levels of the OFF and ON states, but only the barrier for interconversion between the two. Thus the placement of the OFF state energy relative to the valence
or conduction band energies is not possible from these results. Lastly, the stretching
term (#) in equation 3.1 has been assigned physical significance in various relaxation
phenomena 1 6 ,1 2,1 4,1 1 5 and may provide physical insight into the thermal recovery
process upon further theoretical treatment. The value of # = 0.326 ~ 1/3 observed
in this study, for example, can be recovered for the resonant energy transfer rate
between a donor and a random spacial distribution of non-diffusing acceptors in twodimensions. 110,115 However, theoretical treatment of the stretching term /3 is beyond
the scope of this project.
In conclusion, air-free CdSe core QD ensembles were gently photodimmed and
55
allowed to recover using ambient thermal energy at a range of temperatures. Globally
fitting the nonlinear recovery to a stretched exponential function and transforming
to rate constant space provided an underlying probability distribution of activation
barriers that describes the data in terms of a collection of first-order processes and is
consistent with commonly proposed blinking mechanisms.
Chapter 4
Progress Towards Single Quantum
Dot Absorption Spectroscopy
4.1
Introduction
Fluorescence based techniques have been a powerful tool in uncovering photophysical properties of colloidal quantum dots (QDs). But these measurements are only
useful for QDs that are highly emissive and can only provide an indirect measure
of electronic transitions. To provide a more direct measurement of band edge exciton fine structure and higher order electronic transitions on the single QD level, it
is important to measure a single QD absorption spectrum. Additionally, remaining
questions regarding QD absorption may be answered, including if there are dynamics
in the absorption spectrum like spectral diffusion or blinking, what the linewidth of
transitions will be without inhomogeneous broadening, and if there are significant dot
to dot variations in the spectra.
Past work has made significant progress in describing band edge4 3 '1 1 7 and higher
lying41 absorption properties of QDs using photoluminescence excitation spectroscopy
(PLE). They made important contributions to the field by successfully identifying
excitonic transitions and validating approximations made in the derivation of QD
electron structure. 18 ' 44 These studies, however, provided results via an indirect measure of absorption transitions because PLE measurements are subject to a mixture of
57
absorptive and emissive properties. Specifically, these studies were unable to quantitatively analyze transition strengths and probe absorption dynamics due to emissive
character in PLE measurements.
Recently, progress has been made in the field of single molecule absorption by
direct transmission, 119 photothermal contrast,1 2 0 and by ground state depletion.1 2 1
These experiments were useful in describing changes in extinction for single organic
dye molecules during emissive and dark states and after bleaching, but suffer a common shortcoming of only probing absorbance at a single wavelength. In a study more
relevant to QDs, the emission and extinction of a single QD were simultaneously
measured.1 22 Extinction signals were found by measuring the interference between
a reference beam and a beam scattered by QDs. They observed that QDs maintain
the same extinction cross section during emissive and non-emissive periods, but become transparent after photobleaching. These results are interesting in terms of QD
fluorescence intermittency, but reveal little about optical transitions. All of these experiments could be extended to multiple wavelengths, but doing so would complicate
acquisition and require integration at each wavelength. The collection of broadband
spectra would allow for faster acquisitions and the possibility of uncovering dynamic
behaviors.
To speculate on the feasibility of single QD absorption spectroscopy, we first look
at the absorption probability by a single QD in a diffraction limited beam at room
temperature.
The probability of absorbing a photon under these conditions goes
as P = U/7rr 2 , where
- is the band edge absorption cross section of 0.32 nm 2 for
SiO 2 coated QDs12 3"
and r is the diffraction limited spot radius (for a 630 nm
24
light source). This calculation results in an absorption probability of approximately
5 ppm. The signal strength on a 16-bit camera with a maximum intensity value
of 216 = 65, 536 counts amounts to 0.33 absorbed counts per frame. With a pixel
shot noise of V216= 256, it will require (256/0.33)2 = 6.2 x 105 frames, or 4 hours
of integration time at an integration time of 24 ms per frame using an optimized
x
104 nm 2
,
CCD camera. Even a moderate decrease in spot size to a value of 3.31
as found with the bowtie apertures in section 2.4, leads to a significant reduction
58
in required averaging and thus overall integration time. Using the same parameters
described for the diffraction limited spot, the frames required to match the increased
absorption signal is only 1.6 x 10', requiring less than one minute of integration time in
bowtie apertures. It may also be possible to move our experiments to liquid helium
temperatures to improve the signal to noise ratio. At low temperatures, spectral
diffusion is mitigated, lifetimes are lengthened, and the linewidth is reduced by several
hundred times. This effectively condenses the oscillator strength into a more narrow
frequency range 125 and significantly increases the absorption cross section at peak
wavelengths.
It is therefore possible from a signal averaging perspective to perform this measurement with any system that provides a sub-diffraction limited spot with 1-2 orders
of magnitude reduction in spot size, including slot waveguides and apertures in silver
films. These strategies have a central theme in common, which is the employment of
a nanoscale structure to manipulate light to a scale what will provide better overlap
with a single QD.
4.2
Silica Coated Quantum Dots
Prior to all single QD absorption attempts, it was necessary to insulate the QDs from
surrounding nanoscale structures to prevent quenching, so that the QDs retain native
absorptive and emissive properties. Retaining these properties was important to facilitate QD placement via emission detection and to be confident that any absorption
results are artifact-free. The QDs were insulated by overcoating with a ~ 10 nm layer
of silica (SiO 2 ) as described in section 2.2.1.
The insulating properties of the silica coated QDs were examined by detecting
single QD emission for both silica coated QDs and uncoated CdSe/ZnS core/shell
QDs (QD655, Invitrogen) on both a glass and TiO 2 substrates. The QD655 QDs on
TiO 2 substrate showed signs of quenching (Fig. 4-1b), while the QD655 on glass and
the SiO 2 coated QDs on both substrates exhibited normal blinking behavior (Fig.
4-la,c-d). These results showed that the SiO 2 coating was successful in preventing
59
Glass substrate
Ti0 2 substrate
lxoo -a)
0)0b
10000
QD655
b)
8000
0
a
4000
Qms
C)
Si 02 coated
c)
126000 d
OWcaed
sn
.~6100
%
O
2oo 2ss 00
a
2
4
6
ibis (a)
a
10
00
2
TbW
2OW
6
4
1
)
Si~~~
m
2000
10000.
0WW
o
*E4000.
20000
4
ibis (S)
6
a
10
2
4
6
is(s)
10
a
Figure 4-1: a)QD655 on glass substrate, b)QD655 onl TiO 2 substrate, c)SiO 2 coated
QDs on glass substrate, and
d)SiO 2 coated QDs on TiO 2 substrate. Emission is
quenched only in b), where the QD shell is in direct contact with the TiO 2 substrate.
electron transfer from the QD conduction band to the TiO
2
conduction band, which
lies energetically below the CdSe core. The thin insulating ZnS shell on QD655 was
not able to prevent excited electron transfer in this case.
4.3
Slot Waveguides
Slot waveguides were explored for a single QD absorption measurement due to their
interesting waveguiding properties. Calculations show that the introduction of a nar-
row slot containing a low index material (i.e. air) along the length of a waveguiding
dielectric structure (i.e. amorphous TiO 2 , n = 2.4), causes the transmission mode to
focus into the low index slot providing a sub-diffraction limited a beam path.
126
The
mechanism for this phenomenon stems from the abrupt drop in permittivity when
going from the dielectric material to the air gap. Given that the electric displacement
60
field must remain constant, the electric field compensates for the drop in permittivity.
This leads to a strong field enhancement for a small, perturbative air gap in a waveguide. Placing a single QD inside the slot where the QD would occupy a significant
portion of the sub-diffraction limited beam path would provide a small enough mode
area to allow for single QD absorption spectroscopy.
4.3.1
Description
Slot waveguides were fabricated by collaborators in the Lon-ar group at Harvard
university using a lithographic procedure and is presented in figure 4-2. Rows separated by 20-100 nm gaps were lithographically written into a positive e-beam resist
poly(methyl methacrylate)(PMMA) on top of a TiO 2 chip. A chromium mask was
applied protect the tops of the slot waveguides during PMMA liftoff and TiO 2 removal with reactive ion etching. The slot waveguides were complete with the removal
of the chromium mask. An SEM image and schematic including dimensions of a slot
waveguide is presented in figure 4-3.
4.3.2
Transmission and Emission Measurements
Measurements on the slot waveguides were taken on an epifluorescence microscope
with a high numerical aperture (NA=0.9) air objective. Light was coupled into the
slot waveguide by focusing a beam at one end of the structure, and transmitted light
was detected after light is coupled out of the other end (Fig. 4-4 left). The entire slot
waveguide is visible when the image is projected onto a CCD camera. Alternatively,
slots potentially containing QDs were scanned with a 532 nm beam and emission from
the ends of the slot waveguides was detected after excitation light was coupled in and
emission light was coupled back out at the same spot (Fig. 4-4 right)
Characterizing the transmission of broadband light passing through the slot waveguide was important to establish an appropriate baseline for absorption measurements.
Light from a 100 W halogen lamp was transmitted through a slot waveguide with a
20 nm slot width and the output was spectrally resolved (Fig. 4-5). We observed
61
PMMA Resist
SiO 2
e-beam
lithography
-
I
I
PMMA
liftoff
Cr Mask
L
Reactive ion
etching and
mask removal
Figure 4-2: Slot waveguide fabrication procedure.
b
Figure 4-3: (Left) Scanning electron microscope image of a TiO 2 slot waveguide with
60 nm wide slot. (Right) Illustration of slot waveguide structure. Dimensions are 10
pm in length, a = 150 nm, and b = 200 nm.
62
Emission
Broadband
Excitation
Figure 4-4: Illustrations of the transmission (left) and emission (right) experimental
schemes.
fringes in the transmitted spectrum with a free spectral range (FSR) consistent with
that of two reflecting surfaces placed 10paM apart, indicating that the slot waveguide
structure is acting as a Fabry-Perot interferometer.
As a preliminary attempt to detect QDs inside a slot waveguide, a dilute solution
of silica coated QDs with band edge emission of 640 nm was spun cast onto a chip
containing a slot waveguide array. Physical features on the chip were visualized by
raster scanning a diffraction limited beam (532 nm) and collecting the scattered light
on an avalanche photodiode (APD). Scanning was then repeated with an emission
filter (600 nm longpass) placed in front of the APD to provide an image of emitting
QDs.
Scatter and emission images were overlaid to provide a map of QDs and their
placement on the slot waveguide chip (Fig. 4-6a and b). Emission at both ends of
a slot waveguide suggested that there were QDs located inside slot that were being
excited by coupled laser light. The excitation beam was then coupled into the slot
waveguide and the sample image was directed to a CCD camera where QD emission
from both ends of the slot waveguide was simultaneously imaged (Fig. 4-6c). Emission
fluctuations over time at both ends of the waveguide were identical (data not shown),
63
7.2nm
FSR
2nl cos 0
0.80.6
-
0.4 -
.2
Reference
Transmitted
-- Calculated
610
620
625
Wavelength (nm)
6i5
650
6'5
Figure 4-5: (Left) Reference (blue), transmitted broadband light (green), and calculated transmission (red) for a slot waveguide with a 20 nm slot width. The reference
spectrum was measured by reflecting the beam off a flat surface on the sample. The
calculated transmission has a FSR of 7.2 nm and was normalized by the reference
spectrum. (Right) Illustration of a slot waveguide structure and parameters used to
calculate FSR. A central transmission wavelength A 0
620 nm, length I = 10 prm,
index of refraction n = 2.7, and a coupling angle of 0 0 were used.
suggesting that the same QDs were contributing to the emission signal. Whether the
QDs
were deposited on one of the slot waveguide faces, on top of the slot waveguide,
or indeed inside the slot is unknown.
4.3.3
Placement of QDs
Initial attempts to place and detect QDs in a slot waveguide relied on evaporating
or spin casting a dilute solution of silica coated QDs onto slot waveguides. Optical
measurements were able to suggest the placement of QD inside a 20 nm gap slot
waveguide (Section 4.3.2), but imaging techniques were unable to image inside of
such a narrow gap. Atomic force microscopy (AFM) relies on probes that are larger
than the gap width, and SEM did not provide enough contrast inside the narrow
gap due to image artifacts from charging along edges of the slot waveguide structure
(e.g. Fig. 4-3 left). However, it was possible to image silica coated QDs inside slot
waveguides with a larger gap (100 nm) as shown in figure 4-7.
In order to controllably place a single QD inside the slot waveguide, an attempt
was made to physically manipulate a QD using and AFM tip in contact mode. The
64
Figure 4-6: (a) Overlay of emission channel (color) on laser scatter channel (grayscale)
for an array of slot waveguides. (b) Overlay image of slot waveguide with emission
at both ends as highlighted in (a). (c) The same 20 nm gap slot waveguide excited
at the top facet. Emission was collected on a CCD camera and showed emission at
both ends of the waveguide with blinking individual QDs along the waveguide.
65
Figure 4-7: Top-view SEM image of a 100 nm gap slow waveguide with 20 nm diameter
silica coated QDs. It was possible to image QDs inside the slot with larger gap slot
waveguides only.
66
(a)
97.4 nm
80.0
59.6
(b)lm
285.0 nm
I
200.0
152.3
Figure 4-8: AFM images of a slot waveguide spun cast with 20 nn silica coated QDs
(a) before and (b) after sweeping QDs into the slot with the AFM tip in contact
mode.
top image in figure 4-8 shows the initial placement of QDs deposited on the top
surface of a slot waveguide via spin casting, while the bottom image in figure 4-8 was
collected after sweeping multiple QDs into the slot with the AFM tip. The bottom
image contains many large depositions and artifacts, which are presumably from a
dirty AFM tip causing accumulation of unwanted material and loss of resolution from
changes in the tip geometry. It was therefore impossible to make any conclusions on
the effectiveness of the procedure to place a single QD inside the waveguide.
Optical trapping with a slot waveguide was considered for QD placement and
has been achieved with silicon waveguides.m
QD,
Controllable placement of a single
however, would remain difficult due to the linear geometry of the trap. It was
therefore decided that reducing dimensionality by moving to a zero dimension system
would be the best way to isolate a nanoparticle with matching geometry.
4.4
Circular Apertures in Silver Film
Circular apertures in a thin silver film were explored for single QD absorption spectroscopy to facilitate the placement of a single QD by reducing the dimensionality
of the nanoscale structure. Surface plasmon-assisted extraordinary transmission of
subwavelength apertures 94 provides a sub-diffraction limited beam with dimensions
67
comparable to those of a single QD absorption cross section, which increases the
signal strength in an absorption measurement by drastically reducing the amount
of background light transmitted. Other groups have used the extraordinary optical
transmission and plasmon resonances of arrays of subwavelength apertures to look at
absorption of monolayers on metal surfaces,
12 8
,12 9 , 13 0
but the absorption spectrum of
a single object through a single aperture has not been observed.
4.4.1
Fabrication and Characterization
The circular apertures were fabricated by collaborators in the Lonear Group at Harvard University in optically opaque silver films (- 250 nm thick) and had radii ranging
from 40 nm to 150 nm. Circular apertures were fabricated using a lift-off procedure, starting by spin casting a thin layer of negative electron-beam resist (Hydrogen
silsesquioxane, HSQ) on a silicon nitride substrate with a silicon scaffold (Fig. 4-9).
Posts of varying radius were written in the HSQ before evaporation of silver and then
hydrogen fluoride liftoff. An optical image of white light through circular apertures
shows increased light transmission with increasing radius, which agrees with measured
circular aperture transmission spectra as well as transmission spectra calculated using
finite-difference time-domain simulations (Fig. 4-10).
4.4.2
QD Absorption Spectrum through a Circular Aperture
A procedure for measuring broadband QD absorption and emission spectra inside
sub-wavelength circular apertures was developed. Broadband light was continuously
transmitted through apertures using a custom-built transmission absorption apparatus optimized for mechanical stability and minimal chromatic aberration. It was
crucial to reduce mechanical drift and chromatic aberration in the broadband light
passing through the aperture. Changes in aperture position in a focused spot with
aberration can cause artifacts due to changes in the transmission spectrum over the
course of the experiment. During the experiment, QDs were then evaporated on to the
aperture sample in situ as aperture transmission spectra were continuously collected,
68
2
Figure 4-9: (top) Circular aperture fabrication schematic. (1) Spin-cast HSQ on
silicon nitride membrane supported by silicon scaffold. (2) Write posts in HSQ. (3)
Evaporate silver film. (4) Liftoff with HF to remove posts. (bottom left) SEM image
of post structures. (Bottom Right) SEM image of circular apertures.
-3
(b)1
U
x 10
-- 150 nm
-- 140 nm
-- 120 nm
8
4%-
C
6
or
..
100 nm
*4
0
4
500
550
600
650
Wavelength (nm)
Figure 4-10: (a) Image of circular apertures back lit by a broadband source. (b)
Calculated (dashed) and measured (solid) transmission spectra for circular apertures
of varying radius.
69
I
Ensemble
Aperture
1
I
0.80
0.4
&
o 0.4
0.2
0
520
540
560
580
600
620
640
Wavelength (nm)
Figure 4-11: QD absorption spectra measured for an evaporated film ensemble and
for a collection of QDs inside a 140 nm radius circular aperture.
providing many frames of both blank and signal spectra. Emission spectra from QDs
incorporated with the apertures were also collected by providing a separate, spectrally
distinct excitation beam and measuring the transmitted emission. The emission was
then subtracted from the measured absorption spectrum. An absorption spectrum
for non-silica coated QDs deposited in a 140 nm radius aperture is shown in figure
4-11 (red) compared to a thin film ensemble measurement performed with the same
QDs
on the same instrument. Many QDs were observed inside a similar aperture
following evaporation (Fig 4-12).
Although an absorption spectrum was measured for a QD ensemble inside a circular aperture, limitations hindered development of this system into one capable of
achieving a single QD absorption measurement. Firstly, the experiment was purely
single shot. The transmission spectrum of the aperture was very sensitive to position, so measuring blank signal spectra had to be done without moving the sample.
Evaporating QDs onto the circular aperture film solved this problem, but led to an
experiment that was very inefficient because it could only be performed once for every
circular aperture array. Secondly, placement of individual QDs would still be challenging due to the size discrepancy between the QDs and the aperture. The use of
70
Figure 4-12: SEM of a 100 nm radius circular aperture coated by a thin film of QDs.
The bottom and the inside walls are coated with QDs.
silica coated QDs to increase the particle diameter was planned, but the silica coated
QDs
were suspended in water as opposed to hexane.
The evaporation procedure
was only possible with short evaporation times provided by non-silica coated QDs
in hexane.
The experiment took too long when evaporating a water/QD solution
and mechanical drift caused too much noise in the measured absorption spectrum.
Lastly, a large increase in transmission intensity was observed when particles entered
the aperture due to dielectric loading. The change in aperture transmission spectrum
induced by the dielectric nature of the particle itself could lead to artifacts in the
observed absorption spectrum if any spectral features are introduced in the visible
region of the aperture transmission spectrum. This is less of a concern for bowtie apertures because the peak transmission wavelength is in the infrared region, so spectral
changes occur outside of the region of interest. These limitations prompted a change
in the experimental procedure to optically trap with bowtie apertures as discussed
in chapter 2. Bowtie apertures alleviate these concerns by allowing for controllable
placement of the particle in the aperture for blank and signal spectra measurements.
71
72
Appendix A
Fast-Acquisition Absorption
Spectroscopy of Self-Assembled
Cyanine-Dye Nanotubes: A summary
of contributions made to Eisele et al.,
2014
A.1
Self-Assembled Nanotubes
Nature has produced various types of light harvesting complexes, which capture solar
energy and transport the resulting excitation to reaction centers to drive chemistry
that is essential for life. Photosynthetic systems in some green bacteria are comprised
of self-assembled chlorosomes with a high density of pigment molecules - a characteristic that maximizes absorption in low-light environments and allows for delocalized
Frenkel excitons that transport energy over many molecular units.131,132,133,134 Selfassembled cylindrical pigment structures, such as those found in green sulfur bacteria,
can further assemble to form large light harvesting antennae ultimately responsible
for energy harvesting. This hierarchal structure may play an important role in pho73
tosynthesis and understanding the effect of each structural contribution may reveal
how nature collects and transports solar energy so efficiently.
Self-assembled cyanine-dye nanotubes are a model system for studying the properties of natural photosynthetic systems, and are formed when an amphiphilic dye
(C8S3) is incubated water and methanol.
135
The dye monomers initially form double-
walled light harvesting nanotubes (LHNs) that can be microns in length, consisting of
concentric inner and outer cylinders with diameters of
~ 6 nm and
-
13 nm, respec-
tively (Fig. A-1a).136 This assembly results in a ~ 80 nm red shift in the monomer
absorption spectrum due to excitation transfer interactions between monomers, as
well as significant narrowing due to delocalization of the excitation'
31 " 37 (Fig.
A-1b,
band 1). Upon further incubation, the double-walled LHNs aggregate to form bundled LHNs, consisting of many inner cylinders bundled together within a single outer
wall. 136 There is a slight red shift in band 1, and band II is formed in place of bands
2 and 3 upon bundling. This self-assembled model system and provides an excellent
opportunity to investigate the role of hierarchical structure in the exciton transport
properties of light harvesting structures.
A.2
Fast-Acquisition Absorption Spectroscopy
In order to determine the impact that bundling inner cylinders has on the excitonic
properties of the model system, it was first necessary to understand the contribution
of the outer cylinder in both the double-walled LHNs and the bundled LHNs. The
inner cylinder spectroscopic contribution in double-walled LHNs was previously isorated using redox chemistry to reveal that the concentric cylinders are best described
as distinct but weakly coupled excitonic systems.137 A complementary strategy for
removing the outer cylinder contribution is flash dilution, where LHNs are rapidly
diluted in a native solvent that preferentially dissolves the outer surface.
Dissolution rate was found to be very rapid and to vary significantly with different addition volumes, stirring rates, and from LHN sample to sample. To accommodate for this, a fast-acquisition absorption spectrometer was constructed to monitor
74
Outer C finder
. . . . . . .
.
(b) ,
(
(a)
Double-walled LHNTs
Bundled LHNTs
(3
0
Monomers
-e
0
0
480 495 510 525 540 555 570 585 600 615
Wavelength / nm
Figure A-1: (a) Illustration of a double-walled LHN. (b) Absorption spectra reveal
changes in excitonic properties upon assembly from monomers to double-walled LHNs
and further assembly to bundled LHNs. The symbols 11and _L indicate the polarization of each band. Adapted with permission from Eisele et al., 2014. Copyright 2014
by the Proceedings of the National Academy of Sciences.
changes in absorption at high frame rates as flash dilution was performed (Fig. A-2).
A broadband quartz tungsten halogen source (66880, Newport/Oriel) was coupled to
a 3 mm liquid core fiber and collimated. It was then passed through a 4 mm quartz
cuvette mounted on a stir plate, and sent to a spectrometer/ camera combination (Acton SP2500/Pixisl024, Princeton Instruments). Spectra were collected for 20 minutes
with 10 ms frames starting immediately before dilution. Spectra were calculated relative to a reference spectrum collected on this instrument before dilution.
136
The flash dilution results presented in figure A-3 show clear differences between
the double-walled LHNs and the bundled LHNs. Band 2 of the double-walled LHNs,
13 7 rapidly decays as monomer absorption
originating mainly from the outer cylinder,
increases. The bundled LHNs, however, exhibit very few spectral changes upon flash
dilution other than an overall reduction in absorption strength along with an increase
in monomer absorption.
This indicates that the outer wall does not contribute to
spectral bands I and II, and likely only adds a broad underlying spectral contribution.
75
Spectrometer
Variable ND 4 mm
filter
Liquid core
cuvette fiber
Tungsten
Lamp
Mirror
Stir plate
Figure A-2: Fast-acquisition absorption spectrometer schematic.
Double-walled LHNs
Time evolution
start
end
500
500
E
C
Z 550
550 m
(3)
600
1)600
650
650
1b n
i0
Absorbance/normalized
5
10
15
Time / s
20
0
1
Absorbance/normalized
Bundled LHNs
start
Time evolution
end
500
500
E
C
550
550
-----------------*----------------------------U
r600
600
--
650
L
1
0 [
Absorbance/normalized
10
20
30
Time / s
40 0
650
_
1
Absorbance/normalized
Figure A-3: Flash dilution results. Reproduced with permission from Eisele et al.,
2014. Copyright 2014 by the Proceedings of the National Academy of Sciences.
76
A.3
Summary
The flash dilution results were used in concert with oxidation, cryo-TEM, polarizationcontrolled 2D excitation spectroscopy,1 38 and simulations 136 to conclude that bundling
does not adversely affect exciton transport properties even though there are spectroscopic changes associated with bundling. Band II, which arises upon bundling the
inner cylinders, is unlikely used for exciton transport because it lies energetically
above band I. Band II was reproduced with simulations by slightly changing the
monomer tilt, rotation, and wrapping angles, which is an unsurprising consequence
of close cylinder packing in bundled LHNs. Additionally, 2D spectroscopy revealed
that strong excitonic correlations between bands 1 and 3 in double-walled LHNs are
also preserved between bands I and II of the bundled LHNs. These results indicate
that bundling does not significantly alter the excitonic properties of LHNs, and likely
offers advantages unrelated to exciton transfer, including increased structural stability and increased absorption cross section. Bundling cylindrical building blocks can
have structural and absorptive advantages while the excitonic properties of the low
energy, highly delocalized, state is retained.
77
78
Bibliography
[1]
Istvdn Robel, Vaidyanathan Subramanian, Masaru Kuno, and Prashant V. Kamat. Quantum dot solar cells. harvesting light energy with CdSe nanocrystals
molecularly linked to mesoscopic TiO 2 films. Journal of the American Chemical
Society, 128(7):2385-2393, 2006.
[2] Prashant V. Kamat. Quantum dot solar cells. semiconductor nanocrystals as
light harversters. The Journal of Physical Chemistry C, 112(48):18737-18753,
December 2008.
[3] Igor Coropceanu and Moungi G. Bawendi. Core/shell quantum dot based luminescent solar concentrators with reduced reabsorption and enhanced efficiency.
Nano Letters, 14(7):4097-4101, 2014.
[4] Chia-Hao M. Chuang, Patrick R. Brown, Vladimir Bulovi6, and Moungi G.
Bawendi. Improved performance and stability in quantum dot solar cells
through band alignment engineering. Nat Mater, 13(8):796-801, August 2014.
[5] Nicholas J. Thompson, Mark W. B. Wilson, Daniel N. Congreve, Patrick R.
Brown, Jennifer M. Scherer, Thomas S. Bischof, Mengfei Wu, Nadav Geva,
Matthew Welborn, Troy Van Voorhis, Vladimir Bulovid, Moungi G. Bawendi,
and Marc A. Baldo. Energy harvesting of non-emissive triplet excitons in
tetracene by emissive PbS nanocrystals. Nat Mater, 13(11):1039-1043, November 2014.
[6]
Scott M. Geyer, Jennifer M. Scherer, Michael D. Jack, Moungi G. Bawendi, and
Frank B. Jaworski. Dual-band ultraviolet-short-wavelength infrared imaging via
luminescent downshifting with colloidal quantum dots. Journal of Nanophotonics, 7(1):073083-073083, 2013.
[71
Scott M. Geyer, Jennifer M. Scherer, Frank B. Jaworski, and Moungi G.
Bawendi. Multispectral imaging via luminescent down-shifting with colloidal
quantum dots. Opt. Mater. Express, 3(8):1167-1175, August 2013.
[8]
Daniel R. Larson, Warren R. Zipfel, Rebecca M. Williams, Stephen W. Clark,
Marcel P. Bruchez, Frank W. Wise, and Watt W. Webb. Water-soluble quantum
dots for multiphoton fluorescence imaging in vivo. Science, 300(5624):14341436, 2003.
79
[9] Igor L. Medintz, H. Tetsuo Uyeda, Ellen R. Goldman, and Hedi Mattoussi.
Quantum dot bioconjugates for imaging, labelling and sensing. Nat Mater,
4(6):435-446, June 2005.
[10] Sonal Mazumder, Rajib Dey, S. Mukherjee, and G. C. Das. Review: Biofunctionalized quantum dots in biology and medicine. Journal of Nanomaterials,
2009(ID 815734):17, 2009.
1111 Euan R. Kay, Jungmin Lee, Daniel G. Nocera, and Moungi G. Bawendi.
Conformational control of energy transfer: A mechanism for biocompatible nanocrystal-based sensors. Angewandte Chemie International Edition,
52(4):1165-1169, 2013.
1121 B. 0. Dabbousi, M. G. Bawendi, 0. Onitsuka, and M. F. Rubner. Electroluminescence from CdSe quantum - dot/polymer composites. Applied Physics
Letters, 66(11):1316-1318, 1995.
1131 LeeAnn Kim, Polina 0. Anikeeva, Seth A. Coe-Sullivan, Jonathan S. Steckel,
Moungi G. Bawendi, and Bulovid Vladimir. Contact printing of quantum dot
Light-Emitting devices. Nano Letters, 8(12):4513-4517, December 2008.
114] Eunjoo Jang, Shinae Jun, Hyosook Jang, Jungeun Lim, Byungki Kim, and
Younghwan Kim. White-light-emitting diodes with quantum dot color converters for display backlights. Advanced Materials, 22(28):3076-3080, 2010.
115] Tae-Ho Kim, Kyung-Sang Cho, Eun Kyung Lee, Sang Jin Lee, Jungseok
Chae, Jung Woo Kim, Do Hwan Kim, Jang-Yeon Kwon, Gehan Amaratunga,
Sang Yoon Lee, Byoung Lyong Choi, Young Kuk, Jong Min Kim, and Kinam
Kim. Full-colour quantum dot displays fabricated by transfer printing. Nat
Photon, 5(3):176-182, March 2011.
116] Brian J. Walker, Bulovi6 Vladimir, and Moungi G. Bawendi. Quantum dot/jaggregate blended films for light harvesting and energy transfer. Nano Letters,
10(10):3995-3999, 2010.
117] Ou Chen, Lars Riedemann, Fred Etoc, Hendrik Herrmann, Mathieu Coppey,
Mariya Barch, Christian T. Farrar, Jing Zhao, Oliver T. Bruns, He Wei, Peng
Guo, Jian Cui, Russell A. Jensen, Yue Chen, Daniel K. Harris, Jose M. Cordero,
Zhongwu Wang, Alan Jasanoff, Dai Fukumura, Rudolph Reimer, Maxime Dahan, Rakesh K. Jain, and Moungi G. Bawendi. Magneto-fluorescent core-shell
supernanoparticles. Nature Communications, 5, October 2014.
[18] Masaru Kuno, David P. Fromm, Alan Gallagher, David J. Nesbitt, Olga I.
Micic, and Arthur J. Nozik. Fluorescence intermittency in single InP quantum
dots. Nano Letters, 1(10):557-564, 2001.
[191 Thomas S. Bischof, Raoul E. Correa, Danna Rosenberg, Eric A. Dauler, and
Moungi G. Bawendi. Measurement of emission lifetime dynamics and biexciton
80
emission quantum yield of individual InAs colloidal nanocrystals. Nano Letters,
14(12):6787-6791, 2014.
120] Jeffrey J. Peterson and Todd D. Krauss. Fluorescence spectroscopy of single
lead sulfide quantum dots. Nano Letters, 6(3):510-514, 2006.
[21]
Raoul E. Correa, Eric A. Dauler, Gautham Nair, Si H. Pan, Danna Rosenberg,
Andrew J. Kerman, Richard J. Molnar, Xiaolong Hu, Francesco Marsili, Vikas
Anant, Karl K. Berggren, and Moungi G. Bawendi. Single photon counting
from individual nanocrystals in the infrared.
2012.
Nano Letters, 12(6):2953--2958,
[221 Tijana Rajh, Olga I. Micic, and Arthur J. Nozik. Synthesis and characterization
of surface-modified colloidal cadmium telluride quantum dots.
The Journal of
Physical Chemistry, 97(46):11999-12003, 1993.
[23] C. B. Murray, D. J. Norris, and M. G. Bawendi. Synthesis and characterization
of nearly monodisperse CdE (e = sulfur, selenium, tellurium) semiconductor
nanocrystallites. Journalof the American Chemical Society, 115(19):8706-8715,
1993.
[241 Daniel K. Harris, Peter M. Allen, Hee-Sun Han, Brian J. Walker, Jungmin Lee,
and Moungi G. Bawendi. Synthesis of cadmium arsenide quantum dots luminescent in the infrared. Journal of the American Chemical Society, 133(13):4676-
4679, 2011.
[25] Z. Adam Peng and Xiaogang Peng. Formation of high-quality CdTe, CdSe, and
CdS nanocrystals using CdO as precursor. Journal of the American Chemical
Society, 123(1):183-184, 2001.
126] Dmitri V. Talapin, Andrey L. Rogach, Andreas Kornowski, Markus Haase, and
Horst Weller. Highly luminescent monodisperse CdSe and CdSe/ZnS nanocrystals synthesized in a hexadecylamine - trioctylphosphine oxide - trioctylphospine
mixture. Nano Letters, 1(4):207-211, 2001.
[271 W. William Yu and Xiaogang Peng. Formation of high-quality CdS and other II
- VI semiconductor nanocrystals in noncoordinating solvents: Tunable reactiv-
ity of monomers. Angewandte Chemie InternationalEdition, 41(13):2368-2371,
2002.
[28] M. Kuno, J. K. Lee, B. 0. Dabbousi, F. V. Mikulec, and M. G. Bawendi. The
band edge luminescence of surface modified CdSe nanocrystallites: Probing the
luminescing state. The Journal of Chemical Physics, 106(23):9869-9882, 1997.
[29] Gregory Kalyuzhny and Royce W. Murray. Ligand effects on optical properties
of CdSe nanocrystals. The Journal of Physical Chemistry B, 109(15):7012-7021,
2005.
81
[301 Margaret A. Hines and Philippe Guyot-Sionnest. Synthesis and characterization
of strongly luminescing ZnS-capped CdSe nanocrystals. The Journal of Physical
Chemistry, 100(2):468-471, 1996.
1311 B. 0. Dabbousi, J. Rodriguez-Viejo, F. V. Mikulec, J. R. Heine, H. Mattoussi,
R. Ober, K. F. Jensen, and M. G. Bawendi. (CdSe)ZnS core - shell quantum dots: Synthesis and characterization of a size series of highly luminescent nanocrystallites. The Journal of Physical Chemistry B, 101(46):9463-9475,
1997.
1321 Xiaogang Peng, Michael C. Schlamp, Andreas V. Kadavanich, and A. P.
Alivisatos.
Epitaxial growth of highly luminescent CdSe/CdS core/shell
nanocrystals with photostability and electronic accessibility. Journal of the
American Chemical Society, 119(30):7019-7029, 1997.
[33] J. Jack Li, Y. Andrew Wang, Wenzhuo Guo, Joel C. Keay, Tetsuya D. Mishima,
Matthew B. Johnson, and Xiaogang Peng. Large-scale synthesis of nearly
monodisperse CdSe/CdS core/shell nanocrystals using air-stable reagents via
successive ion layer adsorption and reaction. Journal of the American Chemical
Society, 125(41):12567-12575, 2003.
1341 Peter Reiss, Myriam Protiere, and Liang Li.
Core/shell semiconductor
nanocrystals. Small, 5(2):154-168, 2009.
135] Yongfen Chen, Javier Vela, Han Htoon, Joanna L. Casson, Donald J. Werder,
David A. Bussian, Victor I. Klimov, and Jennifer A. Hollingsworth. "Giant"
multishell CdSe nanocrystal quantum dots with suppressed blinking. Journal
of the American Chemical Society, 130(15):5026-5027, 2008.
[361 Andrew B. Greytak, Peter M. Allen, Wenhao Liu, Jing Zhao, Elizabeth R.
Young, Zoran Popovic, Brian J. Walker, Daniel G. Nocera, and Moungi G.
Bawendi. Alternating layer addition approach to CdSe/CdS core/shell quantum
dots with near-unity quantum yield and high on-time fractions. Chem. Sci.,
3(6):2028-2034, 2012.
137] Ou Chen, Xian Chen, Yongan Yang, Jared Lynch, Huimeng Wu, Jiaqi Zhuang,
and Y. Charles Cao. Synthesis of metal - selenide nanocrystals using selenium
dioxide as the selenium precursor. Angewandte Chemie InternationalEdition,
47(45):8638-8641, 2008.
[381 Ou Chen, Jing Zhao, Vikash P. Chauhan, Jian Cui, Cliff Wong, Daniel K.
Harris, He Wei, Hee-Sun Han, Dai Fukumura, Rakesh K. Jain, and Moungi G.
Bawendi.
Compact high-quality CdSe - CdS core - shell nanocrystals with
narrow emission linewidths and suppressed blinking.
451, May 2013.
82
Nat Mater, 12(5):445-
[39] David Norris. Electronic structure in semiconductor nanocrystals. In Semiconductor and Metal Nanocrystals, Optical Science and Engineering. CRC Press,
2003. doi:10.1201/9780203913260.ch2.
140] Daniel E Gomez, Marco Califano, and Paul Mulvaney.
of single semiconductor nanocrystals.
8(43):4989-5011, 2006. 1463-9076.
Optical properties
Physical Chemistry Chemical Physics,
[411 D. J Norris and M. G Bawendi. Measurement and assignment of the sizedependent optical spectrum in CdSe quantum dots.
53(24):16338, 1996.
Physical Review B,
142] M. Nirmal, D. J Norris, M. Kuno, M. G Bawendi, Al L Efros, and M. Rosen.
Observation of the "Dark exciton" in CdSe quantum dots.
Letters, 75(20):3728, 1995.
Physical Review
[43] D. J Norris, Al L Efros, M. Rosen, and M. G Bawendi. Size dependence of
exciton fine structure in CdSe quantum dots. Physical Review B, 53(24):16347,
1996.
[441 Al L Efros and M. Rosen. The electronic structure of semiconductor nanocrystals. Annual Review of Materials Science, 30(1):475-521, 2000.
[45] M. Orrit and J. Bernard. Single pentacene molecules detected by fluorescence
excitation in a p -terphenyl crystal. Phys. Rev. Lett., 65(21):2716-2719, November 1990.
146] W. P. Ambrose, Th. Basch6, and W. E. Moerner. Detection and spectroscopy
of single pentacene molecules in a p - terphenyl crystal by means of fluorescence
excitation. The Journal of Chemical Physics, 95(10):7150-7163, 1991.
[47] S. A Empedocles, D. J Norris, and M. G Bawendi. Photoluminescence spectroscopy of single CdSe nanocrystallite quantum dots. Physical Review Letters,
77(18):3873, 1996.
[481 S. A Empedocles and M. G Bawendi. Quantum-Confined stark effect in single
CdSe nanocrystallite quantum dots. Science, 278(5346):2114-2117, 1997.
149] R. G Neuhauser, K. T Shimizu, W. K Woo, S. A Empedocles, and M. G
Bawendi. Correlation between fluorescence intermittency and spectral diffusion
in single semiconductor quantum dots. Physical Review Letters, 85(15):3301,
2000.
[50] M. Nirmal, B. 0 Dabbousi, M. G Bawendi, J. J Macklin, J. K Trautman, T. D
Harris, and L. E Brus. Fluorescence intermittency in single cadmium selenide
nanocrystals. Nature, 383(6603):802-804, 1996.
83
[51] K. T Shimizu, R. G Neuhauser, C. A Leatherdale, S. A Empedocles, W. K Woo,
and M. G Bawendi. Blinking statistics in single semiconductor nanocrystal
quantum dots. Physical Review B, 63(20):205316, 2001.
[52] M. Kuno, D. P. Fromm, H. F. Hamann, A. Gallagher, and D. J. Nesbitt. Nonexponential "blinking" kinetics of single CdSe quantum dots: A universal power
law behavior. The Journal of Chemical Physics, 112(7):3117-3120, 2000.
1531
Inhee Chung and Moungi G Bawendi. Relationship between single quantumdot intermittency and fluorescence intensity decays from collections of dots.
Physical Review B, 70(16):165304, 2004.
[54] Jeffrey J. Peterson and David J. Nesbitt. Modified power law behavior in
quantum dot blinking: A novel role for biexcitons and auger ionization. Nano
Letters, 9(1):338-345, January 2009.
[551 Amy A. Cordones, Teresa J. Bixby, and Stephen R. Leone. Evidence for multiple trapping mechanisms in single CdSe/ZnS quantum dots from fluorescence
intermittency measurements over a wide range of excitation intensities. The
Journal of Physical Chemistry C, 115(14):6341-6349, April 2011.
[56] Al L Efros and M. Rosen. Random telegraph signal in the photoluminescence
intensity of a single quantum dot. Physical Review Letters, 78(6):1110, 1997.
[57] M. Kuno, D. P Fromm, S. T Johnson, A. Gallagher, and D. J Nesbitt. Modeling
distributed kinetics in isolated semiconductor quantum dots. Physical Review
B, 67(12):125304, 2003.
[58] Vasiliy Fomenko and David J. Nesbitt. Solution control of radiative and nonradiative lifetimes: A novel contribution to quantum dot blinking suppression.
Nano Letters, 8(1):287-293, 2008.
159]
Jing Zhao, Gautham Nair, Brent R Fisher, and Moungi G Bawendi. Challenge
to the charging model of semiconductor-nanocrystal fluorescence intermittency
from off-state quantum yields and multiexciton blinking. Physical Review Letters, 104(15):157403, 2010.
1601 Pavel A Frantsuzov and R. A Marcus. Explanation of quantum dot blinking
without the long-lived trap hypothesis. Physical Review B, 72(15):155321, 2005.
[611 Pavel A Frantsuzov, Saindor Volkain-Kacs6, and Bolizsatr Jank6. Model of fluorescence intermittency of single colloidal semiconductor quantum dots using
multiple recombination centers. Physical Review Letters, 103(20):207402, 2009.
1621 Jau Tang and R. A Marcus. Diffusion-Controlled electron transfer processes and
Power-Law statistics of fluorescence intermittency of nanoparticles.
Review Letters, 95(10):107401, 2005.
84
Physical
[631 A. Ashkin, J. M. Dziedzic, J. E. Bjorkholm, and Steven Chu. Observation of a
single-beam gradient force optical trap for dielectric particles. Optics Letters,
11(5):288-290, 1986.
164] Maurizio Righini, Anna S. Zelenina, Christian Girard, and Romain Quidant.
Parallel and selective trapping in a patterned plasmonic landscape.
Physics, 3(7):477-480, 2007.
Nature
[651 A. N. Grigorenko, N. W. Roberts, M. R. Dickinson, and Y. Zhang. Nanometric optical tweezers based on nanostructured substrates. Nature Photonics,
2(6):365-370, 2008.
[661 M. Righini, P. Ghenuche, S. Cherukulappurath, V. Myroshnychenko, F. J. Garcia de Abajo, and R. Quidant. Nano-optical trapping of rayleigh particles
and escherichia coli bacteria with resonant optical antennas. Nano Letters,
9(10):3387-3391, 2009.
[67] Mathieu L Juan, Maurizio Righini, and Romain Quidant. Plasmon nano-optical
tweezers. Nat Photon, 5(6):349-356, June 2011.
[68] K. Wang, E. Schonbrun, P. Steinvurzel, and K. B. Crozier. Trapping and
rotating nanoparticles using a plasmonic nano-tweezer with an integrated heat
sink. Nat Commun, 2:469, 2011.
[69J Y. Tanaka, S. Kaneda, and K. Sasaki. Nanostructured potential of optical
trapping using a plasmonic nanoblock pair. Nano Lett, 13(5):2146-50, 2013.
[70] Mathieu L. Juan, Reuven Gordon, Yuanjie Pang, Fatima Eftekhari, and Romain
Quidant. Self-induced back-action optical trapping of dielectric nanoparticles.
Nat Phys, 5(12):915-919, December 2009.
[711 Chang Chen, Mathieu L. Juan, Yi Li, Guido Maes, Gustaaf Borghs, Pol
Van Dorpe, and Romain Quidant. Enhanced optical trapping and arrangement of nano-objects in a plasmonic nanocavity. Nano Letters, 12(1):125-132,
December 2011.
172] Yuanjie Pang and Reuven Gordon. Optical trapping of 12 nm dielectric spheres
using double-nanoholes in a gold film. Nano Letters, 11(9):3763-3767, August
2011.
[731
J. Berthelot, S. S. Acimovic, M. L. Juan, M. P. Kreuzer, J. Renger, and
R. Quidant. Three-dimensional manipulation with scanning near-field optical
nanotweezers. Nat Nano, 9(4):295-299, April 2014.
174] L. Novotny and B. Hecht. Principles of Nano-Optics. Cambridge University
Press, 2006.
85
[75] Xiaolei Wen, Luis M. Traverso, Pornsak Srisungsitthisunti, Xianfan Xu, and
Euclid E. Moon. Optical nanolithography with A/15 resolution using bowtie
aperture array. Applied Physics A, 2014.
[76] M. Mivelle, T. S. van Zanten, L. Neumann, N. F. van Hulst, and M. F. GarciaParajo. Ultrabright bowtie nanoaperture antenna probes studied by single
molecule fluorescence. Nano Lett, 12(11):5972-8, 2012.
[771 Guowei Lu, Wenqiang Li, Tianyue Zhang, Song Yue, Jie Liu, Lei Hou, Zhi Li,
and Qihuang Gong. Plasmonic-enhanced molecular fluorescence within isolated
bowtie nano-apertures. ACS Nano, 6(2):1438-1448, 2012.
1781 Poul Martin Hansen, Vikram Kjoller Bhatia, Niels Harrit, and Lene Odder-
shede. Expanding the optical trapping range of gold nanoparticles. Nano Letters, 5(10):1937-1942, 2005.
[79] Peter J. Pauzauskie, Aleksandra Radenovic, Eliane Trepagnier, Hari Shroff,
Peidong Yang, and Jan Liphardt. Optical trapping and integration of semiconductor nanowire assemblies in water. Nat Mater, 5(2):97-101, February 2006.
1801 Christine Selhuber-Unkel, Inga Zins, Olaf Schubert, Carsten Sdnnichsen, and
Lene B. Oddershede. Quantitative optical trapping of single gold nanorods.
Nano Letters, 8(9):2998-3003, 2008.
[81] Yuanjie Pang and Reuven Gordon. Optical trapping of a single protein. Nano
Letters, 12(1):402-406, 2012.
[82] A. Ashkin, J. M. Dziedzic, and T. Yamane. Optical trapping and manipulation
of single cells using infrared laser beams. Nature, 330(6150):769-771, December
1987.
[83] The idex of refraction is 2.5367 at 1064 nm for bulk CdSe.
[84] Sean A. Blanton, Ahmad Dehestani, Peter C. Lin, and Philippe Guyot-Sionnest.
Photoluminescence of single semiconductor nanocrystallites by two-photon excitation microscopy. Chemical Physics Letters, 229(3):317 - 322, 1994.
[851 Liselotte Jauffred, Andrew C. Richardson, and Lene B. Oddershede. Threedimensional optical control of individual quantum dots.
Nano Letters,
8(10):3376-3380, 2008.
[86] Liselotte Jauffred and Lene B. Oddershede. Two-photon quantum dot excitation
during optical trapping. Nano Letters, 10(5):1927-1930, 2010.
[871 Wei-Yi Chiang, Tomoki Okuhata, Anwar Usman, Naoto Tamai, and Hiroshi
Masuhara. Efficient optical trapping of CdTe quantum dots by femtosecond
laser pulses. The Journal of Physical Chemistry B, ASAP(0):0, June 2014.
86
[881 Yasuyuki Tsuboi, Tatsuya Shoji, Noboru Kitamura, Mai Takase, Kei Murakoshi, Yoshihiko Mizumoto, and Hajime Ishihara. Optical trapping of quantum dots based on gap-mode-excitation of localized surface plasmon. The Journal of Physical Chemistry Letters, 1(15):2327-2333, 2010.
[891 Ana Zehtabi-Oskuie, Hao Jiang, Bryce R. Cyr, Douglas W. Rennehan,
Ahmed A. Al-Balushi, and Reuven Gordon. Double nanohole optical trapping:
dynamics and protein-antibody co-trapping. Lab Chip, 13(13):2563-2568, 2013.
[90] I. A. Ibrahim, M. Mivelle, T. Grosjean, J. T. Allegre, G. W. Burr, and
F. I. Baida. Bowtie-shaped nanoaperture: a modal study. Optics Letters,
35(14):2448-2450, 2010.
1911
Wim Wenseleers, Francesco Stellacci, Timo Meyer-Friedrichsen, Timo Mangel,
Christina A. Bauer, Stephanie J. K. Pond, Seth R. Marder, and Joseph W.
Perry. Five orders-of-magnitude enhancement of two-photon absorption for
dyes on silver nanoparticle fractal clusters. The Journal of Physical Chemistry
B, 106(27):6853-6863, 2002.
192]
M. Geiselmann, R. Marty, J. Renger, F. J. Garcia de Abajo, and R. Quidant.
Deterministic optical-near-field-assisted positioning of nitrogen-vacancy centers.
Nano Lett, 14(3):1520-5, 2014.
[931 Jennifer T. Choy, Birgit J. M. Hausmann, Thomas M. Babinec, Irfan Bulu,
Mughees Khan, Patrick Maletinsky, Amir Yacoby, and Marko Loncar. Enhanced
single-photon emission from a diamond - silver aperture. Nature Photonics,
5(12):738-743, 2011.
194]
C. Genet and T. W. Ebbesen. Light in tiny holes. Nature, 445(7123):39-46,
January 2007.
195]
Irfan Bulu, Thomas Babinec, Birgit Hausmann, Jennifer T. Choy, and Marko
Loncar. Plasmonic resonators for enhanced diamond NV- center single photon
sources. Optics Express, 19(6):5268-5276, 2011.
196]
Edward M. Purcell. Proceedings of the american physical society. Phys. Rev.,
69(11-12):674-674, June 1946.
197]
K. T Shimizu, W. K Woo, B. R Fisher, H. J Eisler, and M. G Bawendi. Surfaceenhanced emission from single semiconductor nanocrystals. Physical Review
Letters, 89(11):117401, 2002.
1981 Eyal Shafran, Benjamin D. Mangum, and Jordan M. Gerton. Using the nearfield coupling of a sharp tip to tune fluorescence-emission fluctuations during
quantum-dot blinking. Phys. Rev. Lett., 107(3):037403, July 2011.
199]
Ute Resch-Genger, Markus Grabolle, Sara Cavaliere-Jaricot, Roland Nitschke,
and Thomas Nann. Quantum dots versus organic dyes as fluorescent labels.
Nat Meth, 5(9):763-775, September 2008.
87
11001
Fabien Pinaud, Samuel Clarke, Assa Sittner, and Maxime Dahan. Probing
cellular events, one quantum dot at a time. Nat Meth, 7(4):275-285, April
2010.
1101] Bhola N. Pal, Yagnaseni Ghosh, Sergio Brovelli, Rawiwan Laocharoensuk, Victor I. Klimov, Jennifer A. Hollingsworth, and Han Htoon. 'Giant' CdSe/CdS
core/shell nanocrystal quantum dots as efficient electroluminescent materials:
Strong influence of shell thickness on light-emitting diode performance. Nano
Letters, 12(1):331-336, 2012.
11021 Michal Danek, Klavs F. Jensen, Chris B. Murray, and Moungi G. Bawendi. Synthesis of luminescent thin-film CdSe/ZnSe quantum dot composites using CdSe
quantum dots passivated with an overlayer of ZnSe. Chemistry of Materials,
8(1):173-180, 1996.
[1031 Yongchi Tian, Theresa Newton, Nicholas A. Kotov, Dirk M. Guldi, and Janos H.
Fendler. Coupled composite CdS-CdSe and core - shell types of (CdS)CdSe and
(CdSe)CdS nanoparticles. The Journal of Physical Chemistry, 100(21):89278939, 1996.
11041 Xiaoyong Wang, Xiaofan Ren, Keith Kahen, Megan A. Hahn, Manju Rajeswaran, Sara Maccagnano-Zacher, John Silcox, George E. Cragg, Alexander L.
Efros, and Todd D. Krauss. Non-blinking semiconductor nanocrystals. Nature,
459(7247):686-689, June 2009.
11051 Siying Wang, Claudia Querner, Thomas Emmons, Marija Drndic, and Catherine H. Crouch. Fluorescence blinking statistics from CdSe core and core/shell
nanorods. The Journal of Physical Chemistry B, 110(46):23221-23227, 2006.
t1061 U. Banin, M. Bruchez, A. P. Alivisatos, T. Ha, S. Weiss, and D. S. Chemla. Evidence for a thermal contribution to emission intermittency in single CdSe/CdS
core/shell nanocrystals. The Journal of Chemical Physics, 110(2):1195-1201,
January 1999.
11071 T. A. Baker, J. L. Rouge, and D. J. and, Nesbitt. Single molecule studies
of quantum dot fluorescence intermittency: Evidence for both dark and lightassisted blinking dynamics. Molecular Phys., 107:1867-1878, 2009. JILA Pub.
8304.
11081 Harumi Asami, Yoshimi Abe, Takeshi Ohtsu, Itaru Kamiya, and Masahiko
Hara. Surface state analysis of photobrightening in CdSe nanocrystal thin films.
The Journal of Physical Chemistry B, 107(46):12566-12568, 2003.
11091 Wilfried G. J. H. M. van Sark, Patrick L. T. M. Frederix, Ageeth A. Bol,
Hans C. Gerritsen, and Andries Meijerink. Blueing, bleaching, and blinking of
single CdSe/ZnS quantum dots. ChemPhysChem, 3(10):871-879, 2002.
88
[110]
Theodor F6rster. Experimentelle und theoretische untersuchung des zwischenmolekularen iibergangs von elektronenanregungsenergie. Zeitschrift fur Naturforscher, 4a(5):321-327, 1949.
[1111] C. P. Lindsey and G. D. Patterson. Detailed comparison of the williams-watts
and cole-davidson functions. The Journal of Chemical Physics, 73(7):3348-3357,
1980.
1112] J. C. Phillips. Stretched exponential relaxation in molecular and electronic
glasses. Reports on Progress in Physics, 59(9):1133, 1996.
11131 R. Metzler, J. Klafter, Joshua Jortner, and M. Volk. Multiple time scales for
dispersive kinetics in early events of peptide folding. Chemical Physics Letters,
293(5 - 6):477 - 484, 1998.
1114] K.C. Benny Lee, J. Siegel, S.E.D. Webb, S. Leveque-Fort, M.J. Cole, R. Jones,
K. Dowling, M.J. Lever, and P.M.W. French. Application of the stretched
exponential function to fluorescence lifetime imaging. Biophysical Journal,
81(3):1265-1274, 2001.
1115] M. N. Berberan-Santos, E. N. Bodunov, and B. Valeur. Mathematical functions for the analysis of luminescence decays with underlying distributions 1.
kohlrausch decay function (stretched exponential). Chemical Physics, 315(12):171 - 182, 2005.
[1161 Larry S. Liebovitch and Tibor I. T6th. Distributions of activation energy barriers that produce stretched exponential probability distributions for the time
spent in each state of the two state reaction A ;- B. Bulletin of Mathematical
Biology, 53(3):443-455, 1991.
[117] Han Htoon, Paris J Cox, and Victor I Klimov. Structure of Excited-State transitions of individual semiconductor nanocrystals probed by photoluminescence
excitation spectroscopy. Physical Review Letters, 93(18):187402, 2004.
[118] Al L Efros, M. Rosen, M. Kuno, M. Nirmal, D. J Norris, and M. Bawendi. Bandedge exciton in quantum dots of semiconductors with a degenerate valence band:
Dark and bright exciton states. Physical Review B, 54(7):4843, 1996.
11191 Philipp Kukura, Michele Celebrano, Alois Renn, and Vahid Sandoghdar. SingleMolecule sensitivity in optical absorption at room temperature. The Journal of
Physical Chemistry Letters, 1(23):3323-3327, 2010.
[120] A. Gaiduk, M. Yorulmaz, P. V Ruijgrok, and M. Orrit. Room-Temperature
detection of a single molecule's absorption by photothermal contrast. Science,
330(6002):353-356, 2010.
11211 Shasha Chong, Wei Min, and X. Sunney Xie.
Ground-State depletion microscopy: Detection sensitivity of Single-Molecule optical absorption at room
89
temperature.
2010.
The Journal of Physical Chemistry Letters, 1(23):3316-3322,
[122] P. Kukura, M. Celebrano, A. Renn, and V. Sandoghdar. Imaging a single
quantum dot when it is dark. Nano Letters, 9(3):926-929, 2008.
[1231 C. A Leatherdale, W. K Woo, F. V Mikulec, and M. G Bawendi. On the
absorption cross section of CdSe nanocrystal quantum dots. The Journal of
Physical Chemistry B, 106(31):7619-7622, 2002.
1124] Y.-S. Park, A. V. Malko, J. Vela, Y. Chen, Y. Ghosh, F. Garcia-Santamaria,
J. A. Hollingsworth, V. I. Klimov, and H. Htoon. Near-unity quantum yields of
biexciton emission from CdSe/CdS nanocrystals measured using single-particle
spectroscopy. Phys. Rev. Lett., 106(18):187401, May 2011.
11251 W. William Yu, Lianhua Qu, Wenzhuo Guo, and Xiaogang Peng. Experimental
determination of the extinction coefficient of CdTe, CdSe, and CdS nanocrystals. Chemistry of Materials, 15(14):2854-2860, 2003.
[126] Qimin Quan, Irfan Bulu, and Marko Loncar. Broadband waveguide QED system on a chip. Physical Review A, 80(1):011810, 2009.
[127] Allen H. J. Yang, Sean D. Moore, Bradley S. Schmidt, Matthew Klug, Michal
Lipson, and David Erickson. Optical manipulation of nanoparticles and
biomolecules in sub-wavelength slot waveguides. Nature, 457(7225):71-75, January 2009.
[1281 Kenneth R. Rodriguez, Summit Shah, Shaun M. Williams, Shannon TeetersKennedy, and James V. Coe. Enhanced infrared absorption spectra of selfassembled alkanethiol monolayers using the extraordinary infrared transmission
of metallic arrays of subwavelength apertures. The Journal of Chemical Physics,
121(18):8671-8675, 2004.
[1291 J. Dintinger, S. Klein, F. Bustos, W. L. Barnes, and T. W. Ebbesen. Strong
coupling between surface plasmon-polaritons and organic molecules in subwavelength hole arrays. Phys. Rev. B, 71(3):035424, January 2005.
[1301 Shaun M. Williams, Kenneth R. Rodriguez, Shannon Teeters-Kennedy,
Amanda D. Stafford, Sarah R. Bishop, Ushani K. Lincoln, and James V. Coe.
Use of the extraordinary infrared transmission of metallic subwavelength arrays
to study the catalyzed reaction of methanol to formaldehyde on copper oxide.
The Journal of Physical Chemistry B, 108(31):11833-11837, 2004.
[1311 J. Frenkel. On the transformation of light into heat in solids. i. Phys. Rev.,
37(1):17-44, January 1931.
11321 Gregory D. Scholes and Garry Rumbles. Excitons in nanoscale systems. Nat
Mater, 5(9):683-696, September 2006.
90
[133] Gregory S. Orf and Robert E. Blankenship. Chlorosome antenna complexes
from green photosynthetic bacteria. Photosynthesis Research, 116(2-3):315--331,
2013.
[134] Shu Furumaki, Frantisek Vacha, Satoshi Habuchi, Yusuke Tsukatani, Donald A.
Bryant, and Martin Vacha. Absorption linear dichroism measured directly on a
single light-harvesting system: The role of disorder in chlorosomes of green photosynthetic bacteria. Journal of the American Chemical Society, 133(17):67036710, 2011.
[135] D6rthe M. Eisele, Jasper Knoester, Stefan Kirstein, Jurgen P. Rabe, and
David A. Vanden Bout. Uniform exciton fluorescence from individual molecular
nanotubes immobilized on solid substrates. Nat Nano, 4(10):658-663, October
2009.
[136] D6rthe M. Eisele, Dylan H. Arias, Xiaofeng Fu, Erik A. Bloemsma, Colby P.
Steiner, Russell A. Jensen, Patrick Rebentrost, Holger Eisele, Andrei Tokmakoff, Seth Lloyd, Keith A. Nelson, Daniela Nicastro, Jasper Knoester, and
Moungi G. Bawendi. Robust excitons inhabit soft supramolecular nanotubes.
Proceedings of the National Academy of Sciences, 111(33):E3367--E3375, 2014.
[1371 D. M. Eisele, C. W. Cone, E. A. Bloemsma, S. M. Vlaming, C. G. F. van der
Kwaak, R. J. Silbey, M. G. Bawendi, J. Knoester, J. P. Rabe, and D. A. Vanden Bout. Utilizing redox-chemistry to elucidate the nature of exciton transitions in supramolecular dye nanotubes. Nat Chem, 4(8):655-662, August 2012.
[1381 Daniel B. Turner and Keith A. Nelson. Coherent measurements of high-order
electronic correlations in quantum wells. Nature, 466(7310):1089-1092, August
2010.
91