Setting up a Tunable-Wavelength
Laser for Measuring Mid-Infrared
Emission from InSb Quantum Dots
A Thesis Presented in Partial Fulfillment of the Honors
Program
Ethan James Taylor
Physics and Astronomy
Under the supervision of Dr. William Rice
University of Wyoming
May 2019
Abstract
Small bandgap semiconductors, such as PbS, PbSe, and InSb, that are optically sensitive
to near- and mid-infrared (IR) light can be used to increase the wavelength solar spectrum
sensitivity of modern, multipartite photovoltaics. As with other semiconducting materials,
small bandgap semiconductors can be formed into quantum dots or colloidal nanocrystals,
thus enhancing both their absorptive cross-sections and emission efficiencies, which are cru-
cial optoelectronic parameters. Critically, the large exciton radii of many small bandgap
semiconductors allows researchers to control the semiconducting bandgap by altering the
quantum dot diameter around or below the exciton radius ( 10 to 20 nm). Here, we discuss
the design and construction of a setup to detect mid-IR (1.5 to 6 µm) photoluminescence
(PL) using a tunable, visible excitation source. Using this system, we measure PL from 15
nm-diameter InSb quantum dots produced using Sn-seeded growth and atomic layer depo-
sition. Using CdSe and PbS colloidal quantum dots, the system is tested in the visible and
near-IR regime. Current mid-IR measurements of InSb epitaxial quantum dots do not show
evidence of PL. In addition to these results, future measurements of InSb epitaxial quantum
dots using high-intensity visible and near-IR excitation will be discussed.
Contents
Abstract
Acknowledgments
1 Introduction 1
2 Background 3
2.1 Photovoltaics and Multiple Exciton Generation . . . . . . . . . . . . . . . . 3
2.2 Quantum Dots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2.1 Epitaxial InSb Quantum Dots . . . . . . . . . . . . . . . . . . . . . . 8
2.3 Lasers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3.1 Ring Cavity Dye Lasers . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.4 Spectrographs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3 Dye Laser Set Up 14
3.1 Beam block design and installation . . . . . . . . . . . . . . . . . . . . . . . 14
3.2 Creation of Water Chiller System . . . . . . . . . . . . . . . . . . . . . . . . 15
3.3 Dye Circulator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.4 Alignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
4 Photoluminescence System 17
4.1 Sample Excitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
4.2 Spectrograph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
4.3 PL Collection at Detector . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
4.4 LabView Program Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
5 Data 21
5.1 CdSe Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
5.2 PbS Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
5.3 InSb Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
6 Summary 26
7 Future Steps 27
References
List of Figures
1 Photovoltaic Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2 Solar Spectrum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3 Solar Photovoltaic Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
4 Multiple exciton generation . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
5 Quantum size effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
6 General ALD Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
7 Epitaxial QD formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
8 TEM Images and Initial Characterization. . . . . . . . . . . . . . . . . . . . 11
9 Czerny-Turner spectrograph . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
10 Beam Block Box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
11 Photoluminescence System . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
12 Sample holder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
13 LabView Control Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
14 CdSe PL and Absorbance . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
15 PbS PL and Absorbance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
16 InSb Absorbance with Reference . . . . . . . . . . . . . . . . . . . . . . . . . 23
17 InSb Absorbance Reference Subtracted . . . . . . . . . . . . . . . . . . . . . 24
18 InSb PL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Acknowledgments
I would like to express my deepest thanks to Dr. William Rice for his guidance, patience,
and his drive to help me become the best experimentalist I can be. I would also like to thank
Dr. Joseph Murphy for his daily assistance in the lab and his invaluable advice. I am also
thankful to Subash, Henry, Josh, and Shashank for their help on this project, and would
also like to thank Julian and Abby for their various forms of assistance and company.
1 Introduction
With societal focus shifting from a fossil fuel centered viewpoint to mentalities centered
among environmental concern and longevity, the necessity for efficient methods of renewable
energy only rises. The most prominent and promising of these renewable energy methods
being that of converting the consistent solar energy that impinges upon the surface of the
earth into electrical power. In order to convert the energy that is carried through light
into usable electrical energy, photovoltaics are used, which in the most basic sense take the
energy that light carries, and converts that energy into usable electrical through the use of
the photovoltaic effect.
The photovoltaic effect was discovered in the mid 1800’s, and since its discovery solar
cells have been created using silicon, one of the most common materials in earths crust, to
collect solar energy. While the efficiency of solar energy collection through photovoltaics has
increased over the last 100 years, the current efficiency of solar photovoltaics is still low,
peaking at around 30%. In order to make solar energy capture a viable, efficient, and cost-
effective way of generating electrical energy, improvements to current photovoltaics must be
made.
In the last 50 years, a novel material type called “quantum dots”, have been researched
and studied that exhibit unique properties. Previous investigation into these materials show
the possibility of a phenomenon called “multiple exciton generation”, which can help solve
the largest source of loss that occurs in current photovoltaics. This potential to dramatically
increase the efficiency of solar photovoltaics make these materials essential to investigate.
Indium Antimonide, a small bandgap semiconductor, can be formed into these quantum dots
and can be used to increase solar photovoltaic efficiency not only through this phenomenon
of multiple exciton generation, but also through its small bandgap. In order to study the
ability of these materials to collect light and convert it to the generation of electron-hole pairs
(excitons), specialized experimental setups must be created in order to properly characterize
these materials.
1
Through this thesis, the mechanisms of photovoltaics as well as the creation of novel
quantum dots for increasing the efficiency of solar photovoltaics will be discussed. An exper-
imental setup that was created for measurements of these materials will be shown, as well
as initial measurements of these novel materials.
2
2 Background
2.1 Photovoltaics and Multiple Exciton Generation
Photovoltaics, in the most basic sense, are materials which exhibit the photovoltaic effect,
with the photovoltaic effect being described as the creation of a voltage in a material due
to impinging photons [1]. The material required for this process is that of a semiconductor,
which is a material that does not fulfill the requirements to be a conductor or the requirements
to be an insulator, falling somewhere in the middle with regards to the electrical conductivity
of the material, with silicon being perhaps the most well known semiconductor.
To describe the electrical conductivity of materials, the idea of a valence band and a
conduction band are introduced. The valence band can be thought of as the location where
electrons are generally present in a material with little energy, and the conduction band can
be thought of as the location where electrons are not present in a material. This simpli-
fication allows us to describe the electrical conductivity of a material as being the ability
of a particular material to move electrons between the valence and conduction band. For
example with metals, the conduction and valence bands can be thought of as overlapping,
allowing easy movement between both bands, which is what allows metals to be considered
conductors. For insulators, the bands can be thought of as being separated by a large energy
distance, requiring a large portion of external energy to be applied to allow energy conduc-
tion within the material. For semiconductors, there still exists a separation between the
valence and conduction bands, but the energy separation is now small enough(∼ 0.5− 2eV )
that the ambient thermal energy, kBT , can populate the higher-energy conduction band with
electrons from the valence band. This energetic separation in semiconductors is called the
bandgap of the material.
This effect of semiconductors allows external energies to be imparted upon a semicon-
ductor material, producing a potential difference within the material. Since photons carry
energy, incident photons can provide the energy necessary to efficiently excite electrons to
3
Figure 1: (a) A photon with energy exceeding that of the bandgap excites an electron from
the valence band to the conduction band. (b) Energy in excess of the bandgap energy is
lost through thermalization. (c) Through coupling to a secondary material, the exciton can
be captured to generate a current. (d) Radiative recombination occurs where a photon with
energy corresponding to that of the bandgap is released as the electron returns to the valence
band.
the conduction band. This energetic promotion leaves an absence of a charge in the valence
band, which is physically identical to a positively charged particle; this is often called a
hole. When these two photon-created charges are seperated, a voltage is created; this entire
process is called the photovoltaic(PV) effect. Naturally, this process has been utilized for
the production of photovoltaics to collect solar light and convert it to usable electricity.
The general process of excitation in a photovoltaic is described in Figure 1. In Figure
1 (a), a photon of energy exceeding that of the band gap of the material excites an electron
from the valence band to the conduction band, creating a hydrogenically bound electron-hole
pair, which is often refered to as an exciton. In Figure 1 (b), the energy that was in excess of
the band gap is lost through phonon emission or ”thermalization” as the exciton relaxes back
to the bandgap energy. In Figure 1 (c), the electron and hole can be collected as a current
through coupling to a secondary material. If the exciton is not collected, recombination will
4
Figure 2: Solar radiation spectrum showing the wavelength density that impacts the surface
of the earth from the sun. The solar spectrum is broad, extending into the IR regime, but
peaking in the visible. The radiation at sea level does not correspond to that at the top of
the atmosphere due to molecular absorption in the atmosphere.[2]
occur where the electron will jump the bandgap, releasing a photon of energy corresponding
to that bandgap energy, as shown in Figure 1 (d). This process is called excitonic radiative
recombination.
Since semiconductors form the basis for the PV mechanism, the natural limitation in
these devices is that they can only absorb photons with energies greater than the semicon-
ductor bandgap. This restriction means that with a large solar spectrum, photons with
energy that is less than that of the bandgap of the semiconductor are not able to create
excitons in the material, so when viewing a solar spectrum like the one shown in Figure 2,
if a semiconductor like silicon is used which has a bandgap energy corresponding to approx-
imately 1100 nm, it is unable to collect any photons with wavelengths longer than 1100 nm.
This results in a large loss of available light which could be collected in the infrared (IR)
regime.
Small bandgap semiconductors such as lead sulfide (PbS)(∼ 0.37eV ) and indium anti-
5
Figure 3: From the available photons that make it to the surface of earth, the amount
collected peaks around 30%, with the two largest losses coming from not collecting in the
IR regime (20%), and thermalization due to photon energies exceeding the bandgap of the
material (30%).[4]
monide (InSb)(∼ 0.2eV ) have bandgaps that allow collection of some of this low energy IR
regime. While these small bandgap semiconductors allow collection of low-energy photons,
photons with energies exceeding that of the bandgap undergo the process shown in Figure
1 (b) where the energy that was in excess of the bandgap is lost through phonon emission or
thermalization within the material [1]. This energy loss can be very high in the case of visible
photons impacting a small bandgap semiconductor. The largest loss in solar photovoltaics
comes from this thermalization loss, as is shown in Figure 3. Thermalization loss can be
avoided through a process called multiple exciton generation (MEG) [3].
MEG begins with the same initial process of absorbing a photon with energy exceeding
that of the bandgap, but if the photon possesses energy that is at least twice that of the
band gap, a “splitting” can occur where that single exciton will transfer the excess energy
that would have been lost in thermalization to the generation of a secondary exciton. This
process is simplified in Figure 4, where the phonon emission from Figure 1 (b) is not
present and all that remains is a doubling of exciton generation efficiency.
6
Figure 4: A photon with energy exceeding at least twice that of the bandgap of the mate-
rial can cause a phenomenon called “multiple exciton generation” to occur where multiple
excitons are generated instead of only one.
2.2 Quantum Dots
Quantum dots can be thought of generally as particulates of a material which have been
reduced to a size where effects of quantum confinement begin to take place [3]. This quantum
confinement causes the bandgap energy to be altered as a function of the physical size of
the quantum dot, this effect is known as the “quantum size effect” and the low-dimensional
materials that exhibit it are referred to as quantum dots (QDs). The bandgap as the size of
the nanocrystal is decreased is blueshifted(energy increased), and can be estimated through
a potential well ‘box’ model. Also, as the regime of quantum confinement is entered, not only
is the bandgap blueshifted, but the continuous nature is split into that of discrete energy
levels similar to that of atomic electron shells. This is represented in Figure 5 along with
a first-order approximation of the size dependence of the bandgap energy of quantum dots,
with the R2 factor defining the physical QD size [Adapted from Fig 1.1 [3]].
For nanoparticles that fall within the regime of being able to be defined as quantum dots,
there are three primary types: laterally defined QDs, colloidal QDs, and epitaxial QDs. Lat-
erally defined QDs are created by a two-dimensional electron gas that has been sandwiched
between two substrates with the QDs defined by electrode placement on these substrates.
7
Figure 5: If the size of a material is reduced to the regime where quantum confinement
can take place, the continuous bands of a bulk will split into discrete energy levels. These
discrete energy levels are directly related to the lateral size of the quantum dot.
Construction of these types of QDs are complex, but provide a valuable application in quan-
tum computing applications [5].
Colloidal QDs are chemically synthesized in solution, and act as good absorbers and
emitters [6]. These QDs have tunable optical properties due to their small size, but are
difficult to couple to secondary materials due to the solution synthesis method. The final
common method for QD synthesis is epitaxial growth, which is a controlled layered growth
method. This growth technique produces films of QDs through induced strain between the
lattice mismatch of the layers. Since epitaxial QDs are grown on semiconductor substrates,
these QDs can be easily coupled to secondary materials, which makes these QDs ideal for
investigating solar photovoltaic applications.
2.2.1 Epitaxial InSb Quantum Dots
In creating epitaxial QDs, there exist a variety of methods for creation of these thin films
of QDs, ranging from site-controlled growth by substrate patterning and molecular-beam
epitaxy, to atomic layer deposition. In creating the indium antimonide (InSb) quantum dots
used in this investigation, atomic layer deposition using a tin (Sn) seed was utilized.
8
Figure 6: Precursor A is introduced to the system, and will react until all reactor sites
are filled, forming an atomically thin layer. Precursor A is pumped out and precursor B is
pumped in and will react until all reactor sites are filled. This process can be repeated as
necessary to grow atomically thin layers on a substrate.
Atomic layer deposition is a multi-step method allowing for deposition of atomically thin
films onto a desired substrate. The process typically involves two precursors, in this case A
and B, which are used to create the final sample. Precursor A is introduced to the reaction
chamber, where the reaction will occur until all reactor sites are filled, and then precursor A
will be pumped out of the system. Precursor B is then introduced, and undergoes reactions
again until all reactor sites are filled at which point it can be removed from the system.
This process of precursor A, pump, precursor B, pump, can be repeated as many times
as necessary until a sample of desired film thickness is created [7]. This generic process is
showcased in Figure 6.
For the particular sample that was investigated, an approximately 1 nm layer of tin oxide
was created using ALD before the deposition of the InSb layers, which was done by using
tetrakis(dimethylamido) tin and oxygen plasma, as the previously referenced precursors A
and B. This was done to help initiate growth of the InSb QDs. After the tin oxide layer is
9
Figure 7: Due to a lattice mismatch between the layers of the material, a strain is induced.
This energy is used to “compress” the material into mounds, which are defined as the epi-
taxial QDs.
deposited, trimethylindium and trisdimethyl amino antimony are deposited in the same ALD
pump process illustrated in Figure 6. Using this deposition method, as the trimethylindium
and trisdimethyl amino antimony react on the surface to produce InSb, due to a resulting
lattice mismatch between these thinly deposited layers, a strain is induced in the material.
Strain is essentially a lattice stress which raises the internal energy of the material. Due to
this increased energy, much like the increase in potential energy of a stretched spring, the
material seeks a way to release this energy, which it does through compressing into elevated
mounds of material, as shown in Figure 7. These mounds are what are then defined as the
QDs that are investigated.
Initial characterization of these dots were done using transmission electron microscopy
(TEM), which is a process that utilizes an electron beam to shoot it through a sample and
measure the received electron density on the opposite side, which allows imaging of objects
with sizes below that which is optically visible(<100 nm). Using this process, TEM images
of the QDs were taken and initial size characterizations were made, providing an initial
measurement of the QD size and the QD distribution throughout the film; These results are
shown in Figure 8. After initial characterization TEM was done, optical characterization of
the bandgap, quantum efficiency, and lateral size can be performed through characterization
of the photoluminescence (PL) released during excitonic radiative recombination, as shown in
10
Figure 8: (a) Large TEM image of the sample, where the black dots are the raised mounds,
the defined QDs. (b) Sweeping through sample and characterizing the size of each mound
provides a measurement of the size of these dots, where the red circle is a ∼15nm circle
around a dot. (c) The distribution of dot sizes for a particular sample.
Figure 1 (d) and the bandgap and lateral QD size relation shown in Figure 5. The ability
to optically characterize these QDs requires the design and construction of a system capable
of causing excitonic radiative recombination in these QDs, and measuring the resulting PL.
2.3 Lasers
In order to measure the PL from the QDs, a laser was used as an excitation source. Lasers
produce elecromagnetic radiation that differs from that of the solar spectrum or a generic
light bulb due to their ability to be spatially and temporally coherent. The spatial coherence
of lasers allows the emitted light to be collimated and focused in predictable patterns, as well
as travel large distances without extensive deviations. Temporal coherence allows the light
to be that of nearly a single wavelength, or have a very discrete emission spectra. This means
the laser light can be easily categorized and filtered in comparison to that of a dispersed light
source. All of these advantages of lasers provide themselves to be the ideal excitation source
in a setup that generates and collects PL.
11
2.3.1 Ring Cavity Dye Lasers
Lasers consist of a “gain medium”, which is a material which allows amplification of incident
light (or energy) through stimulated emission. In a dye laser, the standard gain medium is
an organic liquid dye and solvent solution, which is circulated through the system. As with
standard lasers, cavities are used to gain multiple passes through the gain medium in order
to achieve lasing. In a dye laser, a high-energy pump beam is often used as the excitation
source for the gain medium.
Therefore, a ring cavity dye laser is a laser which uses an organic dye as its gain medium,
an external laser source as its excitation source, and utilizes a ring cavity as the feedback
mechanism to efficiently circulate the light until lasing is achieved. Dye lasers are unique
in the broad emission ability of the organic dyes, which allows the length of the cavity to
be tuned so as to select specific output wavelengths from the cavity. This means that a
dye laser possesses the ability to be tuned to specific wavelengths, a function not possible in
many lasers. Wavelength tunability lends itself to versatility in measurements using the dye
laser as an excitation source.
2.4 Spectrographs
Spectrographs, which can also be called spectrometers or monochromators based on their
application, are instruments used to measure the wavelength-dependent intensity spectrum
of light. These instruments are often utilized in measuring the spectrum intensity of an
incoherent light source, which is their application in this project.
This project makes use of a Czerny-Turner monochromator with a triple grating turret to
laterally disperse the incoming incoherent light by diffracting the light off of a grating, where
individual wavelengths can be monitored at the output of the monochromator by rotating
the grating so as to select only specific wavelengths for exit. This type of monochromator
is highlighted in Figure 9. In this spectrograph configuration, incoherent light enters the
monochromator through the entrance slit, where it is directed to a collimating mirror, which
12
Figure 9: Light enters through an entrance slit, where it is then collimated and directed
to a diffraction grating. The diffraction grating laterally disperses the light, which is then
focused onto the exit slit. Through angular rotations of the grating, particular wavelengths
can be chosen for observation through the exit slit.
collimates the light and directs it to the grating. Once at the grating, the incoherent light is
laterally diffracted or ‘dispered’. This laterally dispersed light impacts the focusing mirror,
which focuses the light down at the exit slit. Through an angular rotation of the grating,
the laterally dispersed element of light which is exiting the monochromator is changed ac-
cordingly, allowing the grating to be rotated to specific angular locations in order to choose
specific wavelengths of light which exit the monochromator.
The dispersion of the light after impact with the grating is dependent upon the number
of grooves that the grating possesses, typically called the grooves/mm, and the angle of the
grooves themselves, called the blaze angle. The grooves/mm of a grating define the angular
spacing between adjacent wavelengths, or how dispersed the light becomes after impacting
the grating. The blaze angle is used to optimize gratings for a particular wavelength, and is
often not reported as an angle, but as a blaze wavelength.
13
3 Dye Laser Set Up
In order to utilize a dye laser excitation source, the dye laser that was present in the lab had
to be brought to working condition, as it was not working with regards to the gain medium,
being water cooled, or aligned. This section discusses the process of preparing the dye laser
for use.
3.1 Beam block design and installation
Before alignment could begin of the dye laser, safety measures needed to be prepared in order
to maintain proper laser safety in the lab. These safety measures revolved around preparing
laser beam blocks for the areas where the dye laser emission was possible, and preparing a
beam block system for the pump beam path, as high powers (>5 mW) would be present in
the pump beam.
In designing the pump beam path beam block, it was necessary to account for the location
of the 532 nm pump laser location and the end location of the dye laser. The high powers
from the pump beam needed to be brought to the dye laser with all reflection points properly
protected from the possibility of reflecting outside of the beam blocks. This was done by
creating a beam block system, which consisted of the walls of a light tight box being placed
around all reflecting mirrors and a light tight lid to be placed over the entire enclosure. The
wall and laser path can be seen in Figure 10, where M1 is a flipping mirror which allows
it to be flipped out of the way. Once this mirror is flipped, the pump beam can exit the
box through a small hole where it can be directed to another experimental system in the
lab. M2 is a stationary mirror which directs the pump beam into the dye laser and must
remain stationary due to the nature of the alignment. After the dye laser, beam blocks were
measured and cut to be placed along all the edges of the table, allowing laser safety for areas
parallel to the beam plane by preventing the beam from exiting the designated area.
14
Figure 10: A light tight box for laser safety. The high-power pump beam is directed to the
dye laser, and blocked from exiting the box.
3.2 Creation of Water Chiller System
In order for a liquid organic dye to be used as the gain medium, it must be circulated through
the pump beam path, which causes the dye solution to gradually heat up over time, thus
producing dye degradation and a decrease in the lasing quality. As such, it is necessary to
keep the dye solution from heating by cooling it in some way, which can be done by cooling
the dye circulator via a water circulation system. The water circulation system was created
through measurements of all the necessary connectors for the multiple water paths, tubing
sizes and lengths, and control mechanisms. The parts were then ordered, and the design
was implemented by splicing into an existing water chiller system and turning it into a T-
bracket intersection, allowing the operator to choose between which system to water cool.
This system was then tested and confirmed as operational in cooling the dye solution.
3.3 Dye Circulator
To obtain the circulation of the dye solution through the dye laser, an external dye circulator
is used which simultaneously cools the dye solution and pumps it through the necessary
tubing to allow it to be impacted by the excitation 532 nm beam within the laser itself. In
15
order to get this to working condition, the old dye solution that was present in the circulator
needed to be cleaned with ethylene glycol and methanol, which the dye solution that was
previously present, Rhodamine 6G, is soluble in. This process needed to be done while
wearing proper protection due to the toxic nature of Rhodamine 6G. The jet tube, which is
the tube that “shoots” the dye solution through the laser, needed to be cleaned and installed
into both the laser and the circulator. This also needed to be done with the “catcher” tube,
which is the tube that received the dye solution that was passed through the laser by the jet
tube, and returns it to the circulator for recirculation through the system.
After these initial steps were done, the system was aligned and brought to working
condition through a multi-step process of repeated circulation through the system with small
amount of new solution and alignment of the jet and catcher tube so as to allow proper flow,
and increasing the amount of solution and pressure until the system worked for dye solution
circulation at the appropriate pressure without leakage.
3.4 Alignment
Aligning and achieving lasing within the laser is a multi-step process that begins with hor-
izontally and vertically aligning the pump beam so that the entrance of the pump beam
is not angled with respect to the dye laser. After getting proper pump beam incidence,
alignment revolves around aligning the beam to each mirror in the setup consecutively. The
pump beam must be properly impacting the dye jet stream, which is necessary to be at the
proper angle to allow lasing to occur. The emmited fluourescence and reflected beam must
be aligned along each mirror in the cavity, until initial alignment occurs. Optimization of
each reflection point, along with the angle of the dye jet stream must then occur in order to
get optimized lasing through the cavity.
16
Figure 11: A photoluminescence system that collects radiative recombination from a sample
which has been excited with a 405 nm laser, and directs it through a Czerny-Turner spec-
trograph where the light is laterally dispersed and collected through a detector. The system
utilizes a mechanical chopper and a lock-in amplifier to suppress noise.
4 Photoluminescence System
In order to study excitonic radiative recombination in QDs, the ability to collect and char-
acterize the resulting PL after excitation is necessary. To do this, a setup was created that
was able to excite a sample at a variety of locations to allow for probing of various spots
on a sample, collect low intensity PL from the sample at a large range of wavelengths out
to the mid-IR, and characterize this low intensity temporally incoherent PL as a function of
intensity at specific wavelengths. This section highlights the creation of this experimental
setup and its operation.
4.1 Sample Excitation
In order to create an intensity illumination of the sample, a laser is used as the excitation
source. The laser can be tuned to have a desired power within safe regimes for the sample
17
and can be focused to small areas on the sample to allow for probing of various sample
locations. The laser light is mechanically chopped at a particular frequency through the use
of a mechanical chopper, which is a small wheel with ordered spacing to allow for transmission
only at particular frequencies. This chopper is used in conjunction with a lock-in amplifier
and the detector in order to suppress noise in the measurement.
The modulated excitation light is brought to the sample where the sample is mounted
on a glass slide and on a 3D translational mount. Translational mounts allow movement in
three dimensions in micrometer steps, allowing accurate sample translations. The mount for
the sample was designed and 3D printed; the design of this part is shown in Figure 12.
Once excitonic radiative recombination occurred at the sample, the PL was then col-
lected and brought to a spectrograph. The PL was collected through the use of two off-axis
parabolic, unprotected gold mirrors, which allowed the light to be collected without chro-
matic aberration and sent to the spectrograph. The gold coating allows large wavelength
reflectance(>600 nm), from the visible out to the far-IR regime, which is a necessary opera-
tion range due to the small bandgap nature of InSb (Eg ≈ 0.2 eV). The excitation path and
PL collection can be seen in Figure 11.
4.2 Spectrograph
As was mentioned in Section 2.4, spectrographs are frequently used in measuring the
wavelength spectrum, and this is the use of the spectrograph in Figure 11. The PL is
focused through the entrance slit of the spectrograph, where the light is reflected onto the
first collimating mirror before being brought to the grating that is chosen by the grating
turret. The spectrograph possesses three gratings that can be chosen from, one that is
optimized for UV, one for visible, and one for IR. The ones for visible and IR were the
two used in the experimental setup as it allowed alignment in the visible before moving to
IR. After the light is laterally dispersed at the grating, it impacts a focusing mirror, which
focuses the light down at the exit slit before the light exits the spectrograph.
18
Figure 12: 3D design of a sample holder that allows for mounting of various samples and
glass slides, and reproducible lateral translation.
4.3 PL Collection at Detector
After the light exits the spectrograph, it begins to diverge from the exit slit. This diverging
light is collimated through a calcium fluoride plano-convex lens, where the collimated light
then encounters an unprotected gold mirror. This mirror reflects the collimated light through
another calcium fluoride lens, where it is focused down to the detector. The mirror allows
easy alignment of the laterally dispersed PL to a detector of choice, which allows the detectors
within the setup to be switched in order to allow for large ranges of investigation. For
example, a Si detector can be used for the visible regime, an InGaAs detector for the near-
IR, and an InSb detector for the mid-IR. These detectors are then connected to a lock-in
amplifier, which acts to detect only at a reference frequency. The lock-in amplifier uses the
frequency of the mechanical chopper as the reference frequency, and since the excitation light
is chopped at a particular frequency, the resulting PL will also be at the frequency, meaning
19
Figure 13: A portion of the LabView code that was written to control the PL system. Here,
a loop steps through each wavelength of the spectrograph and records a lock-in measurement
at that wavelength and graphs it, before moving to the next wavelength.
the lock-in amplifier takes measurements from the detector only at the frequency that the
PL is occuring at, allowing high levels of outside noise suppression (>100 dB).
4.4 LabView Program Control
The lock-in amplifier and spectrograph are both connected to a control computer, which
allows commands to be passed to both of the devices from the computer itself. This design
allows the computer to pass a command to the spectrograph specifying an angle to move the
grating (and in turn, a wavelength to exit the spectrograph), and also to pass a command to
the lock-in amplifier to take a measurement at that particular position. Using this, a code was
written in LabView which takes user specified inputs such as initial and ending wavelengths,
step size, and measurement wait times, and can run to collect locked in measurements over
a range of laterally dispersed light. A sample of the code that was written for this is shown
in Figure 13.
20
Figure 14: The absorbance and PL curve of colloidal CdSe QDs, showing the blueshift of
the PL curve with regards to that of bulk CdSe.
5 Data
5.1 CdSe Data
The setup was initially tested with a known sample of colloidal cadmium selenide (Eg ≈
1.7 eV) QDs in the visible regime due to the ability to not only visibly align, but also
due extensive previous measurements in other experimental setups. Placing a thin film of
CdSe colloidal QDs onto a glass cover slip, and then onto a glass slide, the sample could
be mounted in the previously mentioned sample mount. A 405 nm laser was used as the
excitation source both for alignment ease and its high-energy photons (the photon energy
is well above the bandgap energy of CdSe). The off-axis parabolic mirrors could then be
aligned using this visible PL so as to allow the focal point to enter the spectrograph. Using
a 500 nm wavelength blaze grating, the PL could be laterally dispersed and brought out
of the spectrograph and focused onto a Si detector. Using a lock-in amplifier which both
21
Figure 15: The absorbance and PL curves for PbS, verifying setup operation in the near-IR
regime.
the mechanical chopper and detector were connected to, a measurement of the PL of CdSe
quantum dots could be taken, which is shown in Figure 14. At each point along the PL
curve, the grating within the spectrograph was rotated to the appropriate wavelength, and
then a measurement was taken. As expected, the PL shows a blueshift from the Eg ≈ 1.7 eV
that is the general bandgap energy of CdSe. From this bandgap energy increase, a lateral
size of these quantum dots can be estimated using the equation shown in Figure 5, giving
us a radius of approximately 2.7 nm.
5.2 PbS Data
Using known samples of colloidal lead sulfide (∼ 0.37eV ) QDs, measurements in the near-IR
were taken to verify IR operation of the system. In order to align the PL for IR measurements,
the detector was switched to InGaAs for IR measurements, and then aligned to the visible
PL of CdSe. This allows the sample to be switched to PbS, which does not effect alignment,
22
Figure 16: Absorbance measurement of InSb with a reference consisting of a blank TEM
grid.
and then a measurement of the PL of PbS was taken, and is shown in Figure 15. Using the
same process as that which was done for CdSe QDs, the radius of these PbS QDs is estimated
as approximately 2.2 nm. This confirms operation in IR, and allows measurements of the
untested InSb sample to be done using this experimental setup.
5.3 InSb Data
Moving on to untested epitaxially grown InSb QDs, absorption measurements needed to
first be taken. Absorption was taken in a large regime for the InSb QDs that had been
grown with a tin oxide seed on a silicon oxide substrate, and these measurements are shown
in Figure 16. After subtracting the references, the absorption of only the InSb QDs can
be seen in Figure 17. As can be seen from the scale and shape, there are no absorption
features, meaning that the sample is not absorbing at all over the regime, which would mean
that PL would also not be detected. PL measurements were attempted, and the results are
23
Figure 17: Absorbance of InSb after the subtraction of the reference, no evidence of an
absorbance peak.
Figure 18: PL measurement of InSb, no evidence of PL.
24
showcased in Figure 18, showing no PL peaks, the only detection being that of the noise
floor.
25
6 Summary
Modern solar photovoltaic efficiencies can be improved through the implementation of small
bandgap semiconductors that allow the capture of the large IR regime of solar light, and
through QD MEG behaviour which can capture the thermalization loss that occurs in pho-
tovoltaics. InSb is a small bandgap semiconductor, and can be formed into QDs epitaxially,
having the possibility to exhibit MEG, and the epitaxial growth allows ease of coupling to
secondary materials for collection of excitons. As such, InSb QDs are a prime candidate for
increasing the efficiency of solar photovoltaics, and should be investigated.
Throughout this thesis, the mechanics behind solar photovoltaics, QDs, and MEG were
introduced and discussed, as well as the creation of InSb QDs. A setup for investigating the
optical properties of these QDs through the use of a dye laser and photoluminescence system
was introduced, and measurements of these QDs in addition to colloidal CdSe QDs and PbS
QDs were presented. No PL was found for the epitaxial InSb QDs, meaning the intensity of
the PL was too low for detection in the current system, or the samples had degraded.
26
7 Future Steps
The next steps of this experiment revolve around improving experimental versatility, sensitiv-
ity, and sample variety. Obtaining optimized alignment of the dye laser for implementation
into the photoluminescence system will provide more experimental versatility in the choice
of excitation wavelengths, allowing a new degree of experimental freedom for each sample.
Sensitivity can be improved through further optimization of the PL system, particularly
with regards to the method of PL in-coupling to the spectrograph. Sample variety will pro-
vide this experiment with more parameters to test with regards to QD diameter, spacing,
coverage, and substrate coupling.
27
References
[1] R. A. G. Cook, L. Billman, Photovoltaic Fundamentals, ch. 2. National Renewable Energy
Laboratory, 1995.
[2] A. Skaaland, M. Ricke, K. Wallevik, R. Strandberg, and A. G. Imenes, “Potential and
challenges for building integrated photovoltaics in the agder region,” 01 2011.
[3] V. I. Klimov, Nanocrystal Quantum Dots Second Edition. CRC Press, 2010.
[4] O. E. Semonin, J. M. Luther, and M. C. Beard, “Quantum dots for next-generation
photovoltaics,” Materials Today, vol. 15, no. 11, pp. 508 – 515, 2012.
[5] P.-L. M. Bureau-Oxton C, Camirand Lemyre J, “Nanofabrication of gate-defined
gaas/algaas lateral quantum dots,” J. Vis. Exp., vol. 81, 2013.
[6] I. Moreels, K. Lambert, D. Smeets, D. De Muynck, T. Nollet, J. C. Martins, F. Van-
haecke, A. Vantomme, C. Delerue, G. Allan, and Z. Hens, “Size-dependent optical prop-
erties of colloidal pbs quantum dots,” ACS Nano, vol. 3, no. 10, pp. 3023–3030, 2009.
PMID: 19780530.
[7] S. M. George, “Atomic layer deposition: An overview,” Chemical Reviews, vol. 110, no. 1,
pp. 111–131, 2010. PMID: 19947596.