DEVELOPMENT AND EVALUATION OF TRANSPARENT, ALIGNED

POLYCRYSTALLINE ALUMINA AS AN INFRARED WINDOW

CANDIDATE FOR HYPERSONIC FLIGHT APPLICATION
by
Ashwin Sivakumar Dharmakash

A Dissertation
Submitted to the Faculty of Purdue University
In Partial Fulfillment of the Requirements for the degree of

Doctor of Philosophy

School of Materials Engineering

West Lafayette, Indiana
May 2024
THE PURDUE UNIVERSITY GRADUATE SCHOOL
STATEMENT OF COMMITTEE APPROVAL

Dr. Rodney Trice, Co-Chair

School of Materials Engineering

Dr. Jeffrey Youngblood, Co-Chair

School of Materials Engineering

Dr. Eric Kvam

School of Materials Engineering

Dr. Kevin Trumble

School of Materials Engineering

Approved by:
Dr. David Bahr

2
Dedicated to my wife

3
 ACKNOWLEDGMENTS

I would like to thank Lockheed Martin Corporation for their financial support of this
research including the support of Mrs. Clara Rivero-Baleine, without whom, aspects of this
research would not be possible, as well as Mr. Barry Bauer and program manager Mr. Christopher
Iannello for their constant support throughout the years. I want to thank the U.S. Air Force and
Purdue Military Research Institute for their financial support and Col. Dave Hankins and Dr. Eric
Dietz for their guidance and mentorship throughout the years. I want to thank Dr. David Bahr and
the Purdue Materials Science and Engineering Department, whose facilities were used for most of
this research. I would like to thank my advisors Dr. Rodney Trice and Dr. Jeffrey Youngblood for
their ever-present advice and sincere help in my work throughout the years. I want to thank my
committee members for their invaluable feedback and thank faculty members such as Dr. John
Blendell, Dr. Eric Kvam, Dr. Kevin Trumble, Dr. Ernesto Marinero, and Dr. El-Azab for
discussing aspects of my research and beyond. I would like to thank the priceless mentorship and
friendship of Dr. Andrew Schlup over the years, who inspired me to explore various avenues of
research. I would like to thank the mentorship and assistance of Mr. Lester Gomez of Centorr
Vacuum Industries. I would like to thank Dr. Alexei Lagoutchev of Burke Nanotechnology Center
at Purdue University for his invaluable assistance in critical aspects of research and mentorship. I
would like to thank my peers who helped me throughout my time here, including Abdullah Saad,
Rodrigo Guerra, Olivia Brandt, Averyonna Kimery, Matthew Thompson, and Edwin Romero. I
would like to thank our kind building crew, including Timothy VanMeter, Joshua Contrares,
Phillip Qualio, Talukdar Alam, Jenni Fifer, and Steve Bowen. I would like to thank my neighbor
Harris and friends at the International Home Group for their support and kinship. I am indebted to
my wife for her love and constant support. I am thankful for my family for supporting my
endeavors. Finally, I would like to thank Klaus Himmel at Merck KGaA (EMD Performance
Materials) for providing the RonaFlair® White Sapphire powder, and Rick Lamar and AdValue
Technologies, without whom this work would not be possible.

4
 TABLE OF CONTENTS

LIST OF TABLES .......................................................................................................................... 8

LIST OF FIGURES ........................................................................................................................ 9
ABSTRACT.................................................................................................................................. 12
1. INTRODUCTION .................................................................................................................... 13
1.1 Background and Motivation – IR Windows and Hypersonic Flight ........................... 13
1.2 Challenges in Designing an IR Window for Hypersonic Flight .................................. 18
1.2.1 IR Windows – Sources of Scattering and Loss ....................................................... 18
1.3 Literature Review: Alumina as a Hypersonic IR Window Candidate ......................... 21
1.3.1 Single-Crystal Sapphire (Alumina) ......................................................................... 21
1.3.2 Polycrystalline Sapphire (PCA) .............................................................................. 22
1.3.3 Efficacy in a Hypersonic Environment ................................................................... 31
1.4 Research Motivation and Hypothesis........................................................................... 32
1.5 References .................................................................................................................... 32
2. HOT-PRESSING PLATELET ALUMINA TO TRANSPARENCY ...................................... 35
2.1 Contributions and Introduction .................................................................................... 35
2.2 Experimental Procedures ............................................................................................. 35
2.2.1 Powder Preparation ................................................................................................. 35
Chemical Composition of RonaFlair® Batches ............................................................. 38
Effects of Washing ......................................................................................................... 39
2.2.2 Die Prep and Hot-Pressing Parameters ................................................................... 40
2.2.3 Post-Processing ....................................................................................................... 44
Rough Polishing and Surface Grinding ......................................................................... 44
Optical Polishing............................................................................................................ 44
2.3 Physical Measurements ................................................................................................ 46
2.4 Summary and Conclusions........................................................................................... 46
2.5 References .................................................................................................................... 46
3. EVALUATING INFRARED TRANSPARENCY OF PLATELET POLYCRYSTALLINE
ALUMINA .................................................................................................................................... 48
3.1 Contributions and Introduction .................................................................................... 48

5
 3.2 Experimental Methods ................................................................................................. 48
3.2.1 Measuring In-Line Transmission (TILT)................................................................... 48
3.2.2 Normalizing Optical Properties to Sample Thickness ............................................ 49
3.3 Results and Discussion................................................................................................. 49
3.3.1 Comparison of Spectroscopic Methods................................................................... 50
3.3.2 Results on Effect of Sintering Profiles .................................................................... 51
3.3.3 Effects of Calcination on Aligned PCA .................................................................. 54
3.3.4 Effects of Washing on Aligned PCA Samples ........................................................ 56
3.4 Summary and Conclusions........................................................................................... 60
3.5 References .................................................................................................................... 60
4. MECHANICAL PROPERTIES OF ALIGNED HOT-PRESSED ALUMINA ....................... 62
4.1 Contributions and Introduction .................................................................................... 62
4.2 Experimental Procedures ............................................................................................. 62
4.2.1 Modified Die Preparation ........................................................................................ 62
4.2.2 Post-Processing ....................................................................................................... 63
Surface Grinding and Polishing ..................................................................................... 63
Machining ...................................................................................................................... 67
4.3 Mechanical Testing ...................................................................................................... 68
4.3.1 Experimental Apparatus .......................................................................................... 68
4.4 Results and Discussion................................................................................................. 70
4.4.1 Weibull Modulus Calculations ................................................................................ 70
4.4.2 Effects of Sintering Parameters on Weibull Modulus ............................................. 71
4.4.3 Effects of Grain Size on Mechanical Properties of Aligned PCA .......................... 72
4.5 Summary and Conclusions........................................................................................... 78
4.6 References .................................................................................................................... 78
5. EFFECT OF CRYSTALLOGRAPHIC ALIGNMENT OF Polycrystalline ALUMINA ON
SURVIVAL IN A SIMULATED HYPERSONIC ENVIRONMENT ......................................... 80
5.1 Contributions and Introduction .................................................................................... 80
5.2 Rapid Oxyacetylene Ablation Testing ......................................................................... 80
5.2.1 Ablation Testing Apparatus .................................................................................... 81
5.2.2 Ablation Testing Parameters ................................................................................... 83

6
 5.2.3 Sample Testing Parameters ..................................................................................... 83
5.3 Results and Discussion................................................................................................. 85
5.3.1 Preliminary Findings and Hypothesis ..................................................................... 85
5.3.2 Characterization Before and After Ablation Tests .................................................. 87
5.3.3 Effect of Crystallographic Orientation .................................................................... 91
5.3.4 Effect of Sintering Parameters ................................................................................ 92
5.4 Summary and Conclusions........................................................................................... 93
5.5 Overall Conclusions and Future Work ......................................................................... 93
5.6 References .................................................................................................................... 96
APPENDIX A. EVALUATING HOT-PRESSED MgO:Y2O3 NANOCOMPOSITE POWDER
FOR INFRARED TRANSPARENCY ......................................................................................... 98
A.1 Contributions and Introduction as an Infrared Window Candidate ............................... 98
A.2 Experimental Procedure ............................................................................................... 101
A.2.1 Powder Morphology ............................................................................................ 102
A.2.2 Powder Processing and Heat-Treatment .............................................................. 104
A.2.3 Hot-Pressing ......................................................................................................... 104
A.2.4 Post Processing ......................................................................................................... 106
A.3 Results and Discussion ................................................................................................ 106
A.4 Summary and Future Work.......................................................................................... 110
A.5 References .................................................................................................................... 110
APPENDIX B. INFRARED TRANSMISSION OF REFRACTORY OXIDES AT HYPERSONIC
TEMPERATURE SCALES........................................................................................................ 112
VITA ........................................................................................................................................... 114

7
 LIST OF TABLES

Table 2.1 Chemical Compositions of As-Received (Batch A) and Washed (Batch C) RonaFlair®
....................................................................................................................................................... 39

Table 2.2 Auto-Polishing Procedure for 25 mm samples ............................................................. 45

Table 4.1 Auto-Polishing Procedure for 55 mm samples ............................................................. 65

Table 4.2 Weibull Modulus, Load, and Stress Values for Standard and Long Aligned PCA ...... 71

Table 4.3 Relative Densities and Average Grain Size of Standard and Long Aligned PCA ....... 73

Table 5.1 Comparison of Relevant Sample Properties from Ablation Testing Commercial and In-
House Alumina ............................................................................................................................. 91

8
 LIST OF FIGURES

Figure 1.1 Illustration of a Hypersonic Environment on a Leading Edge1 ................................... 13

Figure 1.2 Atmospheric Windows over a Broad Wavelength2 ..................................................... 14

Figure 1.3 Wien’s Displacement Law for Blackbody Radiation4................................................. 16

Figure 1.4 Hypersonic IR Window Location on a Leading Edge5 ............................................... 17

Figure 1.5 Comparison of Single and Poly-Crystal Material Order9 ............................................ 19

Figure 1.6 Sources of Optical Loss when Light Passes Through a Polycrystalline Material10 .... 20

Figure 1.7 Transparent Pore-Free Alumina as Lucalox™ Sodium-Vapor Lamp12 ...................... 23

Figure 1.8 Crystallographic Orientation and Refractive Mismatch in PCA18 .............................. 24

Figure 1.9 Sources of Optical Loss in PCA18 ............................................................................... 26

Figure 1.10 Spherical Morphology Equiaxed Polycrystalline Alumina (Left): Random Grain
Orientation (Right) ........................................................................................................................ 27

Figure 1.11 Platelet Morphology Aligned Polycrystalline Alumina (Left): Basally Aligned Grain
Orientation (Right) ........................................................................................................................ 28

Figure 1.12 In-Line Transmission of CeraLumina™, Platelet, and Equiaxed Alumina10 ............ 29

Figure 1.13 Grain Growth During Sintering of PCA10 ................................................................. 30

Figure 2.1 (a-e) SEM Images of RonaFlair® Under Various Powder Processing Methods ......... 37

Figure 2.2 (a,b) Cross-Sectional (Front) Illustration of Graphite Die Anatomy During Preparation
(left) and Centorr Furnace Loading Conditions1 (right) ............................................................... 41

Figure 2.3 Sintering Profiles of Polycrystalline Alumina (Red = Standard, Blue = Long, Yellow =
Two-Step) ..................................................................................................................................... 42

Figure 2.4 Two-Step Sintering (TSS) profiles: described by Chu.et al [TSS-C (Left)3] and Chen
and Wang [TSS-CW (Right)4]2..................................................................................................... 43

Figure 3.1 In-Line Transmission using FTIR and UV-VIS Spectroscopy Methods of Sapphire and
PCA Samples (Batch A) ............................................................................................................... 50

Figure 3.2 Results (Batch A) on Effect of Sintering Profiles on In-Line Transmission in the Optical
Regime .......................................................................................................................................... 52

9
Figure 3.3 In-Line Transmission for Standard and Long Batch E Test Cases ............................. 55

Figure 3.4 Polycrystalline Alumina Samples Based on Batch and Sintering Profile with Relative
Densities........................................................................................................................................ 57

Figure 3.5 In-Line Transmission of Polycrystalline Alumina vs. Wavelength as a Function of

Powder Condition ......................................................................................................................... 58

Figure 4.1 Cross-Sectional (Front) Illustration of Graphite Die Anatomy During Preparation (55
mm) ............................................................................................................................................... 63

Figure 4.2 (a,b) Size Comparison of Pre-Polished Standard 55 mm (thickness 3 mm) Aligned PCA
to the Standard 25 mm Alumina (thickness of 1.5 mm) ............................................................... 66

Figure 4.3 ASTM-1161 “A” Flexure Bar2 .................................................................................... 67

Figure 4.4 55 mm Hot-Pressed Sample Orientation During Machining ...................................... 68

Figure 4.5 (a,b) 4-Point Flexure Test Apparatus (a) and Prepared Sample Ready for Testing in
MTS® Electromechanical Test System (b)2 ................................................................................. 69

Figure 4.6 Weibull Modulus Distribution of Standard and Long Aligned PCA .......................... 71

Figure 4.7 SEM Images of Prior Work (scale bars = 200 µm) and Current Approach Standard and
Long Aligned PCA (scale bars = 100 µm) ................................................................................... 74

Figure 4.8 Grain Size (left) and Normalized Average Grain Distribution (right) for Standard and
Long PCA ..................................................................................................................................... 75

Figure 5.1 Schematic of Rapid Oxyacetylene Ablation Testing Apparatus ................................. 82

Figure 5.2 In-Line Transmission Comparison of Batch E Standard and Long PCA and Commercial
Sapphire ........................................................................................................................................ 84

Figure 5.3 Temperature vs. Time Distribution During 60 second Ablation Test of PCA ............ 86

Figure 5.4 “Throw-Away” Platelet PCA Before (left) and After (right) Ablation Test at 20 mm 86

Figure 5.6 (left) Commercial Equiaxed Before and After Ablation Test ..................................... 88

Figure 5.5 (right) Sapphire Before and After Ablation Test ......................................................... 88

Figure 5.7 (left) Basally Aligned Platelet Polycrystalline Alumina – Standard Test Cases Before
and After Ablation test. ................................................................................................................. 89

Figure 5.8 (right) Basally Aligned Platelet Polycrystalline Alumina – Long Test Cases Before and
After Ablation test......................................................................................................................... 89

10
Figure 5.9 In-Line Transmission Results Before and After Ablation Test for Commercial and In-
House Alumina (1.5 mm Thickness) ............................................................................................ 90

11
 ABSTRACT

Hypersonic flight is the key to unlocking a nation’s strategic advantage in this century’s military
theater. Military powerhouses such as the United States, Russia, India, China, Australia, and the
EU publicly possess hypersonic weapons capabilities. Such technology enables intercontinental
travel orders of magnitude faster than conventional flights. A trip halfway across the world would
take not twenty hours, but two. However, the level of thermal and chemical load the aircraft and
these electronic equipment experience while at such high speeds cause them to fail. Thus, ceramic
window materials are used to act as a barrier between the hypersonic flight environment and this
sensitive electronic equipment. Such materials need to be both mechanically robust, but
transparent within the relevant infrared ranges used for target detection. Single-crystal sapphire
(alumina) is an infrared window material readily available, plentiful, and easy to microstructurally
control and manufacture, but not optimal. Its transparency range is limited to the optical and near-
infrared, while it exhibits poor mechanical and dielectric strength. Polycrystalline alumina (PCA)
has recently been shown to possess more favorable infrared window characteristics as opposed to
its single-crystal counterpart. This is achieved by processing using a platelet powder morphology
in a single processing step – hot-pressing. Full densification (> 99.5%) of PCA samples was
achieved, demonstrating maximum of 84% optical transparency, but accompanied by grain growth
(60+µm), resulting in lower mechanical strength. This research thus works on a two-fold approach
to minimizing the grain growth of PCA. Optical tests demonstrated favorable results for lowering
isothermal temperatures to reduce grain growth. Weibull values of m = 28.8 and m = 9.7 from 4
point-flexure tests were obtained (ASTM 1161a). Thermal loading via ablation testing compared
PCA samples to industry alternatives (single-crystal sapphire) and (equiaxed alumina). Ablation
tests revealed the benefit of polycrystalline alumina over sapphire. The benefit of lower isothermal
sintering temperatures for reduced grain growth resulted in higher peak load before failure,
resulting in greater characteristic strength and minimal transmission lost during a minute of
oxyacetylene heat flux exposure. Finally, additional work was done on nanoceramic MgO-Y2O3,
in a ceramic-processing method like that of PCA. These findings will also be discussed.

12
 1. INTRODUCTION

1.1 Background and Motivation – IR Windows and Hypersonic Flight

Hypersonic flight is the key to unlocking a nation’s strategic advantage in this century’s
military and defense theater. Military powerhouses such as the United States, India, Russia, and
China publicly possess hypersonic aircraft and weapon capabilities. Such technology enables
intercontinental travel that is orders of magnitude faster than most conventional flights1. A trip
halfway across the world from Australia to the United Kingdom that would normally take over
twenty hours, now takes less than two. A nation that possesses such technology would not only
experience a massive growth in its commercial sector and productivity, but also pose a precise and
lethal military threat.
A hypersonic flight travels at speeds greater than Mach V in the atmosphere, where it
experiences compressive failure stresses, thermal loads, and chemical and oxidation reactions.
Temperatures reached on the surfaces of these aircraft can be three times as hot as molten lava.

Figure 1.1 Illustration of a Hypersonic Environment on a Leading Edge1

Figure 1.1 illustrates the extreme conditions experienced by a leading edge of a hypersonic
vehicle. When airspeed greater than Mach V reaches a stagnation point (V∞ = 0) on a leading edge
of an aircraft, a bow shock is created, followed by a resulting boundary layer, shear layer, and
ablation layer. The laminar airflow rapidly increases the thermal load experienced, and the
temperature rises drastically, up to and above 1600°C. A shear layer is created on the superior side

13
that results in a combination of rapid heat flux and shear forces. Ablation and pyrolysis can occur,
demanding a material with a high thermal and corrosive resistance and mechanical strength.
Dissociation of air molecules becomes significant, demanding chemical and oxidation resistance.
The level of thermal load, differential heat flux, and ablative reactions demand precise and durable
materials.
The technology that goes into successfully launching a hypersonic weapon or aircraft
therefore relies heavily on having materials to withstand this environment. Ground-based
controllers or pilots need an accurate picture of important parameters to navigate, maneuver, or
adjust trajectory of flights travelling at these speeds. These include factors such as the aircraft’s
surrounding environment, incoming heat flux, distance to target, and temperature, among others.
One critical parameter is tracking and communication, whereby remote-sensing
instruments onboard aircraft or space-based platforms operate in a particular wavelength range to
collect, measure, and determine the energy associated for an object within line-of-sight. This is
done in both a passive and active remote-sensing method.
Passively, remote sensing takes advantage of Earth’s atmospheric windows (not to be
confused with hypersonic windows), which arises from incoming electromagnetic radiation from
outer space and the ways in which it interacts with various molecules and elements in Earth’s
atmosphere.

Figure 1.2 Atmospheric Windows over a Broad Wavelength2

14
 Figure 1.2 is a plot of transmittance over wavelength. The broad wavelength range in the
figure is representative of Sunlight that enters the Earth. When this light from the Sun interacts
with the elements in Earth’s atmosphere, it may be absorbed. For example, absorption of sunlight
in the 5.5-7.5 µm range is due to the presence of water (H2O) in the atmosphere. It can also be
observed that there are some wavelength regions in which atmospheric molecules absorb less light,
allowing for higher transmission in those regions. These specific regions are known as atmospheric
windows. For example, around the 1.4-1.7 µm range lies a window of transmittance, and around
the 2.0-2.5 µm range lies another. For a hypersonic aircraft, a remote sensor can passively detect
the subsequent warm body radiation of an object that corresponds to a particular wavelength, with
the near-IR and mid-IR bands are particularly useful for military application of hypersonic flight.
The IR band according to the Commission Internacional de’Leclairage (CIE) sets the relevant
standards for the near-IR range as 0.78-1.4 µm, the mid-IR range as 1.4-3.0 µm, and the far-IR
range as larger than 3.0 µm but less than 10.0 µm 3.
Actively, remote sensors detect the wavelength of the signal emitted from an object, which
varies based on the temperature of the object. Such a relationship is described theoretically with a
Blackbody object in Wien’s Displacement Law.

(1.1)
Equation 1.1 describes the spectral radiation of this theoretical blackbody, whereby the
peak wavelength, λmax, is inversely proportional to the absolute temperature, T, with b being Wien’s
displacement constant (2898 µm K)4. This relationship is illustrated over a broad wavelength range
in a distribution.

15
 Figure 1.3 Wien’s Displacement Law for Blackbody Radiation4

Figure 1.3 describes the relationship between the wavelength of a blackbody, temperature,
and peak intensity, typically measured in spectral energy density (watts/m3). From the graph, it
can be shown that a blackbody object with high absolute temperature leads to large intensity peaks
followed by flattening-out of the intensity distribution. As the absolute temperature decreases, the
intensity peak lowers and the curve flattens substantially, leading to an increase in the peak
wavelength λmax. For example, a blackbody with a 6000K absolute temperature results in a peak
wavelength at 0.483 µm, which lies in the visible range. Whereas a blackbody with an absolute
temperature of 4000K will have a peak wavelength of 0.235 µm, which lies in the mid-IR. In
practice, the temperatures achieved during hypersonic flight (800-1600°C) result in peak detection
for active remote sensing occurring in the mid-IR range (2.7-1.5 µm, respectively). Utilizing
atmospheric windows that occur in the “infrared windows” shown in Figure 1.2 is thus the primary
function of electronic remote sensing equipment onboard hypersonic aircraft.
However, having sensitive electronic equipment onboard responsible for tracking on a
Mach V aircraft exposed to the aerothermal environment would be an engineering failure.
Therefore, infrared windows are used along with injection coolant to serve as a barrier between
remote-sensing guidance electronics and the hypersonic aerothermal environment.

16
 Figure 1.4 Hypersonic IR Window Location on a Leading Edge5

Figure 1.4 shows an example of an IR window on a leading edge. Remote-sensing

technology (“Seeker” shown in Figure 1.4) needs to receive a radiation signal of sufficient strength
through a boundary layer and an IR window to operate successfully for precise guidance. The
presence of oxygen in the air makes the window material vulnerable to oxidation reactions (making
oxide materials particularly attractive). The aerothermodynamic effects make the window
susceptible to cracking, risking the protection of the guidance electronics. As such, the location of
the optical window is usually coupled with a form of injection coolant.
These IR windows must thus be made of materials that accomplish several mission critical
objectives, while being able to withstand the hypersonic environment. The first is to prevent failure
of sensitive optoelectronic equipment which lies onboard high visibility areas of a hypersonic
vessel, such as a leading edge6. The second is to provide transparency in the IR range used for
sensing, surveillance, detection, and communication applications. Finally, it serves as a physical
barrier and means of thermal protection to sensors from a vehicle’s extreme aerothermal condition
– preventing oxidation, chemical, and ablative reactions. These reactions lead to mass loss,
pyrolysis, phase reactions or melting of the surface, all of which reduce a hypersonic projectile’s
ability to sense or thermally track its environment via onboard infrared remote-sensing equipment6.

17
 1.2 Challenges in Designing an IR Window for Hypersonic Flight

During the mid-20th century, materials scientists began using refractory oxide ceramics to
advance high-speed aircraft and communication technology for military applications in the Cold
War. Their high compressive failure stresses, thermal load resistance, oxidation and chemical
resistance made ceramics a promising material class for infrared and optical windows. As such,
they saw extensive use as missile-protection windows in the infrared (IR) range for both the U.S.
Navy and Air Force throughout the 1960s and 1970s for shielding, radomes, and nose cones. Some
modern refractory oxide ceramics, such as alumina (Al2O3), yttria (Y2O3), and magnesia (MgO),
are of primary interest for these transparent armor solutions, given their high melting points, high
relative densities, and favorable thermal and corrosive-resistant properties7,8. Most importantly,
their ability to be finely and microstructurally tuned gives material engineers the greatest agency
over tailoring these ceramics for the engineering objectives in a hypersonic environment.
This does not mean however, that these refractory oxides are most optimal for the
application, but rather they are currently the best suited and most available based on decades of
research. They have some clear limitations in a flight environment, and these limitations are greatly
exacerbated at hypersonic velocities.

1.2.1 IR Windows – Sources of Scattering and Loss

One of the critical functions of IR windows is to be transparent in relevant IR wavelengths

in the hypersonic environment. Therefore, it is imperative to look at ways in which an incoming
signal strength can be reduced as it passes through the window. We will first look at sources of
loss in terms of physical properties; but to do so, it is important to understand how these materials
are built atomically, and the difference between a single crystal and polycrystalline material. The
preferential orientation of the crystals (grains) in a material will influence how an incoming light
signal interacts and finally transmits through it.

18
 Figure 1.5 Comparison of Single and Poly-Crystal Material Order9

Both single and polycrystalline orders are shown in Figure 1.5. Single-crystal materials
consist of a continuous, ordered lattice structure in which the atoms or molecules are arranged in
a repetitive pattern. The entire material is composed of a single crystal (no grain boundaries), and
all crystallographic directions have the same atomic arrangement, resulting in uniform properties
in all directions. As a result, single-crystal materials exhibit specific anisotropic properties,
meaning their optical, mechanical, thermal, and electrical properties can vary significantly
depending on the direction of measurement relative to the crystal lattice. This ordered structure
results in good optical and electronic properties compared to polycrystalline materials. However,
producing large single crystals is challenging and expensive, which limits their practical
engineering potential.
Polycrystalline materials, which in contrast, consist of numerous small, randomly oriented
grains, separated by grain boundaries. Each grain has a specific crystallographic orientation, and
the overall material exhibits more isotropy because the individual grains are randomly oriented.
However, since polycrystalline materials' properties are the average of individual grains' properties,
microstructural control of grain size and distribution changes the properties of the entire material.
This therefore makes engineering polycrystalline materials more practical than single crystals.
The caveat is that because of the anisotropy in polycrystalline materials, the presence of
grain boundaries can lead to mechanical weaknesses and reduced electrical and thermal
conductivities (along certain planes) compared to single crystals. Grain boundaries significantly
affect the way light reacts throughout a polycrystalline material and typically reduces overall
transmittance. In general, the below equation describes three ways in which light may interact with
an object.

19
 I = R + T + A = 1 (100%) (1.2)
In Equation 1.2, I is the total signal intensity, R is the total reflectivity, T is the total
transmission, and A is the total absorption. For a polycrystalline material, as opposed to single-
crystal, porosity may be more present – these pores are a volume fraction of void in the material
that can be either along the grain boundaries (intergranular) or within the grains (intragranular).
The reflection and absorption in the material can be broken down into components and used to
mathematically understand the transmission desired for IR windows.

Figure 1.6 Sources of Optical Loss when Light Passes Through a Polycrystalline Material10

Figure 1.6 shows how light reacts with a polycrystalline material. Analysis of the figure
results in the following equation, applicable for a polycrystalline material.

TILT = 1 - (RBS + RDif + RSpec) - A - (TFS(Pore) +TFS(GB)) (1.3)

To understand each component of Equation 1.3 requires examining the figure left to right.
From the left, the first way light can interact with the material is with the surface of the object

20
through diffuse or specular reflection. RDif is the diffuse reflection off the irregularities in the
material surface (as well as reflection off the material back surface). RSpec is the specular reflection
off the smooth portions of the material surface (as well as that on the material back surface). As
light continues to pass through the material, some will be absorbed naturally as light converts to
an increase in heat or thermal energy of the material. It will encounter any number of pores, voids,
or defects in the material that can cause any number of reactions. It can be reflected by the surface
of these pores as RBS, or backward scattering. The light that continues to pass through these pores
is refracted based on the refractive index of the pores and continues as forward scattering. This is
because for polycrystalline grains, the refractive index value varies based on crystallographic
orientation. When grains are oriented in different crystallographic directions. the refraction of light
countless times throughout all the grains of a polycrystalline material results in some of the light
being transmitted, but some being reflected and absorbed each time. The more refraction that
occurs, the less transparent the material. The accumulation of these scattering losses at the grain
boundaries is known as birefringence scattering and is a problem every polycrystalline material
faces when attempting to maximize transparency. The final remaining portion of light transmitted
is known as the in-line transmission, TILT. Maximizing this term is desired.
For a polycrystalline material, controllable microstructural aspects include minimizing the
transmission lost in the grain boundaries and pores. Thus, material with high relative densities (low
porosity) and/or materials whose orientation minimize their refractive index mismatch is preferred.

1.3 Literature Review: Alumina as a Hypersonic IR Window Candidate

1.3.1 Single-Crystal Sapphire (Alumina)

During the mid-20th century, Kingery, Pappis, and Franel (along with other collaborators)
conducted hallmark research on the thermal and electrical properties of ceramic materials11. It was
during this time when refractory oxides were first researched and applied as IR window candidates.
Few materials are more researched or documented than single-crystal sapphire (a form of
corundum – alpha-Al2O3).
Single-crystal sapphire has been known and used as a material for thousands of years, as
aluminum oxide is a plentiful and available material. However, it was only during the early 20th
century when its material properties were thoroughly explored. It was first explored during 1916

21
when the Polish scientist, Jan Czochralski, developed a method to fabricating large single-crystal
sapphire crystals23. This process was essential in creating larger and high-quality crystals. Later
during the Cold-War era, sapphire was used in flight window applications6,7.
Sapphire as a window material is one of the hardest known materials, ranking 9 on the
Mohs scale23. It is scratch-resistant, has high mechanical strength, and possesses excellent optical
properties from ultraviolet to near-infrared wavelengths. As a result, sapphire was used in
numerous commercial aviation and aerospace applications.
However, sapphire also has inherent drawbacks as a material. Its single-crystalline
structure demands the production of large, high-quality single-crystals, that increase
manufacturing cost and precision. This also limits the size of the window material. Crystalline
defects, if present, also affect the entirety of the window material. Despite sapphires hardness, it
is a brittle material and prone to fracture. While authors like Harris et al. showed sapphire holds
mechanical strength advantages to other IR window candidates and considerable optical
transparency (visible range), sapphire also suffers in its IR transmission comparatively, as its peak
transmission drops off in the mid and far-IR range (~2.5-5 µm)13. While this makes it a suitable
window candidate in many flight conditions, it limits its capabilities in the hypersonic environment,
specifically in military applications that require a broader range of transmission that extends into
the near, mid, even far-IR. Working with higher-purity sapphire has been shown to result in more
favorable electrical resistivity properties, but sapphire has several inherent drawbacks as an IR
window candidate for hypersonic flight13.

1.3.2 Polycrystalline Sapphire (PCA)

Enter polycrystalline sapphire (PCA), which has also been researched since the 1960s.
General Electric in 1961 first developed the Lucalox™ sodium-vapor lamps, first demonstrating
it was possible to industrially densify and manufacture transparent alumina12.

22
 Figure 1.7 Transparent Pore-Free Alumina as Lucalox™ Sodium-Vapor Lamp12

In comparison to its single-crystal counterpart, PCA has several advantages. As a

polycrystalline material, it is more cost effective, as densification methods through sintering and
hot-pressing are more friendly to the manufacturing process than the Czochralski method. Its
polycrystalline microstructure results in increased mechanical toughness and corrosive
resistance8,10. Most importantly, its ability to be more finely microstructurally controlled through
improved processing flexibility makes it more suitable for a broader range of applications. These
benefits come at the cost of optical transparency via grain boundary effects and material
homogeneity.
It was over half a century ago when notable authors Heuer, Rhodes, Sellers, and Vasilos
conducted research on various material properties of polycrystalline alumina for the U.S.
Department of the Navy7,8. Recrystallization texture during hot-working was observed in some
cases, and most notably, no structural integrity was compromised during hot-working that resulted
in diminished mechanical properties7. Cannon, Rhodes, and Heuer showed deformation in the
presence of neighboring grain boundaries provides intrinsically more resistance to fracture,
making alumina a more suitable candidate for IR windows8. Peelen showed it is possible to densify
PCA to achieve transparency in the visible and near-IR via continuous hot-pressing, which is the
method this thesis takes14. PCA also has a lower dielectric constant, favorable for IR window
materials. These notable works on PCA make it not only comparable to single-crystal sapphire but
make it favorable for flight applications and potentially suitable in a hypersonic environment.

23
 Table 1.1 Material Properties of Single-Crystal Sapphire and Polycrystalline Alumina15,16

Material Property (⊥basal plane) Sapphire Polycrystalline Alumina

Density (g/cm3) 3.98 3.98
ϵr at 10 GHz 10.0 9.6
Loss at 10 GHz 0.001 0.001
Modulus of Elasticity (GPa) 430 390
Flexure Strength (MPa) 430 390
Bending Strength (MPa) 630 280
Compressive Strength (MPa) 2100-4100 2400
Tensile Strength (MPa) 1800-2600 210-310
Fracture Toughness K1c (MPa m0.5) 2.3 5-6

However, the critical drawback with PCA is due to birefringent scattering, which occurs
when light is refracted away from its optimal path via refractive index mismatches. These
mismatches are from crystallographic misalignment along the thickness of the window material.
Due to PCA's anisotropic rhombohedral crystal structure, the resulting refractive index mismatch
between grains yields a lower in-line transmission, which lowers PCA's efficacy as an IR window.

Figure 1.8 Crystallographic Orientation and Refractive Mismatch in PCA18

In Figure 1.8, the differing refractive indices along various orientations can be seen. The
difference in these values are compounded throughout the thickness of the material when light

24
passes through, resulting in scattering and loss in total transmission. Decreasing grain size, which
results in greater grain boundary density, should also increase birefringent losses. Pores possess
their own refractive index value and contribute to total losses in transmission.
Still, as a polycrystalline material, alumina does have the stark benefit over single-crystal
sapphire of being more easily microstructurally tuned in manufacturing. There are, in fact, several
various engineering methods to retain the mechanical benefits of PCA, whilst minimizing the
optically negative effects of birefringence. The approaches to these methods are supported by the
Rayleigh-Gans-Debye model10.

(1.4)
In Equation 1.4, Rs is the surface reflection contributed by both RDif and RSpec from
Equation 1.3, rg is the average grain size radius, Δng is the difference in refractive indices between
adjacent grains, d is the material thickness, and λ0 is the wavelength of incoming light. By reducing
the values of grain size, refractive index mismatch, or sample thickness it is possible to increase
in-line transmission. The incoming wavelength of light is either a fixed, desired value or constant.
It is also possible to increase the in-line transmission by reducing the surface reflection of the
material. However, theoretically, there will always be some in-line transmission lost due to the
refractive index difference between sample medium and material medium as light passes through.

(1.5)
In Equation 1.5, Rs is the surface reflection and n is the refractive index of alumina. The
value for Rs is ~0.14. Therefore, the maximum theoretical TILT for alumina is ~0.86. PCA would
have a lower theoretical limit, due to birefringence or porosity losses. Ideally, eliminating
birefringence in nonporous PCA would make it optically equivalent to sapphire.
Reducing surface reflection to bring TILT close to its theoretical value is accomplished by
surface grinding and finely polishing the material for an extended period. Polishing smooths the

25
outer surface of the material, effectively reducing the value of RDif . It is a macroscopic method
used universally not only for all PCA samples but ceramics window materials in general.
The first microstructural method to maximize TILT is reducing the size of the grains such
that they are drastically smaller than the wavelength of entering light. This method was explored
using hot isostatic pressing (HIP) first by Hayashi et al. and then by Krell and Apetz. Krell and
Apetz also attempted to reconcile a theoretical discrepancy in the literature regarding the role of
grain boundaries in light scattering in PCA17,18. Some authors such as Peelen showed that light
scattering caused by the grain boundaries is negligible, counter to the fact that the grain size of the
sample was significantly larger than the incoming light signal (1-2 µm)14. Other authors like
Hayashi et al. did not find evidence of light scattering due to pores17. There may be an explanation
found in this dissertation, though this paper does not deliberately seek to directly explain the roles
of grain boundaries or pores on birefringence.

Figure 1.9 Sources of Optical Loss in PCA18

The second microstructural method to maximize TILT is reducing the porosity of the
material, which decreases the frequency to which light scatters and interacts with each pore along
the thickness of the material (a reduction in TFS,Pore in Equation 1.3). The effect of pore size on in-
line transmission was thoroughly explored by Pabst, Hostaša, and Esposito19. These first two
methods are the ones most found and discussed in literature.

26
 The third microstructural method – the one least found in literature, but the one this
research takes greatest advantage of – is alignment of the grains. This makes PCA such that the
refractive index mismatch along different crystallographic faces of PCA is minimized, reducing
the degree to which birefringent scattering occurs along grain boundaries (a reduction in TFS,GB in
Equation 1.3). This can potentially be accomplished using several methods. Early research by
Heuer et al. discussed orientation of grains via press-forging, whereby primary recrystallization
(or the formation of a new, strain-free grain matrix in the presence of plastic deformation) occurred
simultaneously. Later work by Ashikaga et al. as well as from Sokolov and Harris produced
orientation via magnetic field alignment along the c-axis (basal plane) of PCA20,21. However, the
strength of the B fields required (>12T), are a barrier to large-scale reproducibility.
However, recently in 2020, Schlup et al. confirms that it is possible to align the platelet
alumina powders crystallographically through a single processing step – hot-pressing10. This
alignment is a consequence of a platelet morphology alumina powder (as opposed to spherical
morphology). Densification > 99.8% was simultaneously achieved, which has profound impact on
scalability and reproduction.

Figure 1.10 Spherical Morphology Equiaxed Polycrystalline Alumina (Left): Random Grain
Orientation (Right)

27
Figure 1.11 Platelet Morphology Aligned Polycrystalline Alumina (Left): Basally Aligned Grain
Orientation (Right)

The samples in Figures 1.10 and 1.11 show a clear illustration of the potential of platelet
morphology for PCA. The sample in Figure 1.10 is a non-porous (99.9% high relative density)
equiaxed PCA sample, commercially available in McMaster-Carr. From observation, the sample
is completely opaque, a result of the randomly aligned crystal structure. The amount of reflection
due to grain boundary and surface makes it a poor window candidate for IR application, let alone
hypersonic flight. However, in Figure 1.11 is an in-house platelet, aligned PCA sample. Utilizing
platelet morphology, where the basal planes correspond to the c-plane of PCA's rhombohedral
crystal structure, allows the settling and alignment of the crystals during densification, resulting in
a transparent sample. This is a promising result if the transparency is comparable to that of sapphire.
Interestingly, in the past decade, companies such as CeraNova specialize in advanced
ceramic processing and application manufacture PCA for ballistic and flight applications. This
CeraLumina™ is a form of polycrystalline alumina using their own unique powder processing
methods to develop high strength, high hardness, and transparent nose cones and optical windows
for flight applications. This industry sample is replicated by Schlup et al. in 2020 and compared in
Figure 1.10 below10.

28
 Figure 1.12 In-Line Transmission of CeraLumina™, Platelet, and Equiaxed Alumina10

The in-line transmission values for the various PCA samples in Figure 1.12 compare
commercial single-crystal sapphire with CeraLumina™, platelet PCA, and equiaxed PCA values.
Alumina had the highest transmission throughout a broad range, spanning the optical, near, and
mid-IR, close to the theoretical value of ~84%. CeraLumina™ had a more interesting transmission
distribution across the wavelength range. While its transmission values in the optical are poor,
with TILT = 45% at 500 nm, its transmission approached close to that of alumina in the near and
mid-IR, with TILT = 77% at 2000 nm. Platelet PCA, while having a lower peak transmission value
overall, with TILT = 70% at 2000 nm, had a more consistent transmission distribution in the visible
range, comparable to sapphire. This is likely due to minimization of birefringence scattering.
Equiaxed PCA behaved similarly, albeit having the lowest transmission out of all the samples at
TILT = 54% at 2000 nm.
The transmission values shown in Figure 1.12 show the promise of PCA, and more
importantly, the alignment of platelet morphology alumina and its method achieving densification.
However, the primary issue Schlup et al. encountered during the hot-pressing process was grain
growth at operating temperatures10. While higher sintering temperatures lead to higher relative
densities and a reduction in grain boundary surface area, resulting in fewer losses due to

29
birefringence scattering on average throughout the thickness of the material, Schlup et al. also
noted higher operating temperatures result in low fracture stresses (~175 MPa) and Vickers
hardness (~17 GPa) and large grain sizes exceeding 75 µm10.

Figure 1.13 Grain Growth During Sintering of PCA10

In Figure 1.13, it is shown that as the dwell temperature increases for platelet PCA, the
grain size is positively correlated along with relative densification. This is qualitatively seen by
observing the amount of intergranular or intragranular porosity present in all 4 images. For
example, the image with a dwell temperature of 1750°C has a higher level of porous defects present
ubiquitously throughout the material, whereas the image with a dwell temperature of 1825°C has
fewer noticeable porous defects present. Theoretically, dwelling at higher temperatures for an
extended length of time would continue grain growth until most are consumed within a new grain
matrix, approaching crystallography close to single-crystal. However, this comes at the cost of the
mechanical and electrical property benefits of PCA. Therefore, the engineering problem at hand is

30
minimizing grain growth of PCA whilst maximizing in-line transparency and retaining its unique
benefits over single-crystal sapphire.

1.3.3 Efficacy in a Hypersonic Environment

While it is well known that while the strength of sapphire is still better than other refractory
oxide IR windows, its strength – along with most refractory oxides – drastically decreases at
temperatures above 600°C24. Thus, while optically transparent, it fails mechanically in a
hypersonic environment – almost instantly to exposure. Equiaxed alumina also possesses weak
thermal and mechanical characteristics, attributed to the randomly oriented grains and isotropic
CTE values. Additionally, its low in-line transmission makes it unsuitable as a window candidate.
With the introduction of aligned platelet alumina, there now exists a contender to single-
crystal sapphire that has the potentially to be more mechanically robust, whilst minimizing
transparency lost by crystallographically aligning during sintering. A low dielectric constant,
corrosion and oxidation resistance, in addition to mechanical strength and IR transparency could
make aligned polycrystalline alumina a permanent replacement to single crystal sapphire as a
hypersonic IR window.
Unfortunately, when it comes to work specifically related to PCA as a hypersonic IR
window candidate, the literature is quite sparse in this area. Most thermal testing of PCA consists
of thermal shock tests, and while such data is useful, IR windows during hypersonic flight more
often experience rapid heat flux, as opposed to thermal shock. Only a few papers involve rapid
heat flux or thermal loading of alumina or PCA at higher temperatures comparable with those
experienced during hypersonic flight. For example, Ali et al. conducted an oxyacetylene fueled
ablation test for alumina modified with carbon/phenolic composites to provide higher strength,
stiffness, and heat dissipation at maximum temperatures of 800°C22. However, samples in the
hypersonic regime experience temperatures near 1600°C, if not greater. This provides this paper
an opportunity for data on an aggressive test in a rapid heat flux environment to determine the
practical capabilities of single-crystal alumina, equiaxed alumina, and aligned polycrystalline
alumina.

31
 1.4 Research Motivation and Hypothesis

It is thus the goal of this research to investigate and develop methods to achieve IR
transparency for basally aligned polycrystalline alumina and evaluate its effectiveness as a
hypersonic IR window candidate in terms of mechanical strength, transparency, and ability to resist
rapid thermal heating and ablation. It is to manage the engineering tradeoff of the unique benefits
of PCA as a hypersonic IR window candidate over single-crystal sapphire. These benefits include
increased manufacturing effectiveness and improved toughness at the cost of optical strength and
microstructure homogeneity. This paper’s approach to this is to experiment with longer sintering
dwell times and lower sintering isothermal temperatures, to maximize the positive effects of full
densification, and minimize grain growth. In addition, the alignment of PCA grains (which would
increase its anisotropic CTE values along the basal plane), as well as the presence of grain
boundaries (that may act as a source of heat dissipation in polycrystalline alumina), might have a
net positive effect on its candidacy as a hypersonic window. Being able to simultaneously densify
and sinter in a single-processing step is a profound benefit that increases its large-scale ability to
be mass-produced efficiently. In section 2, this thesis discusses the specific procedures taken to
take platelet PCA from powder to IR window. Section 3 discusses IR transparency, comparing
platelet PCA under various processing conditions with commercial sapphire. In section 4, effects
on mechanical properties are discussed. Section 5 tests various PCA and sapphire IR window
candidates in an aggressive ablation test to test thermal load and heat flux capabilities of alumina.
Each section will explain procedures and testing parameters, results, a comprehensive comparison
to existing literature, and draw conclusions to determine the benefits of basally aligned
polycrystalline alumina as an IR window candidate over single-crystal sapphire. Finally, Section
6 (Appendix A.) discusses and evaluates the future potential of another promising refractory oxide
candidate – nanocomposite magnesia and yttria.

1.5 References
1
[1] Chou, A. (2022, January 10). Hypersonic technology. Materials Research Furnaces, LLC.
https://mrf-furnaces.com/hypersonic-technology/

[2] 1: The atmospheric transmission windows from radio to X-ray... ... (n.d.-a).
https://www.researchgate.net/figure/The-atmospheric-transmission-windows-from-radio-
to-X-ray-wavelengths-Figure-created-by_fig1_327860369

32
[3] Karine. (n.d.). Infrared radiation. ICNIRP.
https://www.icnirp.org/en/frequencies/infrared/index.html

[4] Star temperatures. Blackbody Radiation. (n.d.). http://hyperphysics.phy-

astr.gsu.edu/hbase/wien.html

[5] (PDF) response of hypersonic compression corner flow to upstream ... (n.d.-b).
https://www.researchgate.net/publication/371158399_Response_of_hypersonic_compressi
on_corner_flow_to_upstream_disturbances

[6] Hypersonic seeker window attachment for Hypersonic Flight Systems. Hypersonic Seeker
Window Attachment for Hypersonic Flight Systems | SBIR.gov. (n.d.).
https://www.sbir.gov/node/2214993

[7] Heuer, A. H., Sellers, D. J., & Rhodes, W. H. (1969). Hot working of aluminum oxide. 1:
Primary recrystallization and texture. J. Amer. Ceram. Soc., 52, 468–474.

[8] Heuer, A. H., Rhodes, W. H., Sellers, D. J., & Vasilos, T. (1967). Microstructure Studies of
Polycrystalline Refractory Oxides. March 1966.

[9] The difference between Single Crystal and Poly Crystal? | ResearchGate. (n.d.-c).
https://www.researchgate.net/post/The-difference-between-single-crystal-and-poly-crystal

[10] Schlup, A. P. (2020). PROCESSING ALUMINUM OXIDE FOR THE CONTROL OF

MICROSTRUCTURAL TEXTURE AND OPTICAL PROPERTIES.

[11] PAPPIS, J., & KINGERY, W. D. (1961). Electrical Properties of Single‐Crystal and
Polycrystalline Alumina at High Temperatures. Journal of the American Ceramic Society,
44(9), 459–464. https://doi.org/10.1111/j.1151-2916.1961.tb13756.x

[12] Burke, J. E. (1996). Lucalox alumina: The ceramic that revolutionized outdoor lighting. MRS
Bulletin, 21(6). https://doi.org/10.1557/S0883769400046133

[13] Harris, D. C., Johnson, L. F., Zelmon, D. E., Poston, W. B., Kunkel, J. D., Pascucci, M. R.,
Gannon, J. J., Wen, T., Harris, D. C., Johnson, L. F., Cambrea, L., Baldwin, L., Baronowski,
M., Zelmon, E., Poston, W. B., Kunkel, J. D., Parish, M., Pascucci, M. R., Gannon, J. J., &
Tzu-, J. (2017). Refractive index of infrared- transparent polycrystalline alumina. Optical
Engineering, 56(7). https://doi.org/10.1117/1.OE.56.7.077103.Downloaded

[14] Peelen, J. G. J. (1979). Transparent Hot-pressed Alumina. I: Hot pressing of alumina.

Ceramurgia International, 5(2), 70–75. https://doi.org/10.1016/0390-5519(79)90032-2

[15] Hemofilia Comunidad Valenciana Ashecova. Facebook. (n.d.).

https://www.facebook.com/ashecova/posts/pdfsciencedirectassetscom2724421-s20-
s1079979619x000661-s20-s1079979619303146mai/1683150378488335/

33
[16] Technical information single crystal sapphire. (n.d.).
https://www.advaluetech.com/technical-single-crystal-
sapphire#:~:text=Single%20crystal%20sapphire%20possesses%20excellent,chemical%20r
esistance%20at%20high%20temperatures.

[17] Hayashi, K., Kobayashi, O., Toyoda, S., & Morinaga, K. (1991). Transmission Optical
Properties of Polycrystalline Alumina with Submicron Grains. Materials Transactions, JIM,
32(11), 1024–1029. https://doi.org/10.2320/matertrans1989.32.1024

[18] Apetz, R., & Bruggen, M. P. B. van. (2003). Transparent Alumina : A Light-Scattering
Model. Journal American Ceramic Society, 86(3), 480–486.

[19] Pabst, W., Hostaša, J., & Esposito, L. (2014). Porosity and pore size dependence of the real
in-line transmission of YAG and alumina ceramics. Journal of the European Ceramic
Society, 34(11), 2745–2756. https://doi.org/10.1016/j.jeurceramsoc.2013.12.053

[20] Ashikaga, T., Kim, B. N., Kiyono, H., & Suzuki, T. S. (2018). Effect of crystallographic
orientation on transparency of alumina prepared using magnetic alignment and SPS. Journal
of the European Ceramic Society, 38(7), 2735–2741.
https://doi.org/10.1016/j.jeurceramsoc.2018.02.006

[21] Sokolov, A. S., & Harris, V. G. (2018). 3D crystallographic alignment of alumina ceramics
by application of low magnetic fields. Journal of the European Ceramic Society, 38(15),
5257–5263. https://doi.org/10.1016/j.jeurceramsoc.2018.06.035

[22] Ali, S., Ahmad, S., Baluch, A. H., Ramzan, M., Saleem, M., Bilal, M., & Naseem, K. (2022).
Microstructural and thermal analysis of alumina modified carbon/phenolic composites after
ablation testing. Journal of Composite Materials.
https://doi.org/10.1177/00219983221147389

[23] Sarukura, N., Nawata, T., Ishibashi, H., Ishii, M., & Fukuda, T. (2015). Czochralski Growth
of Oxides and Fluorides. In Handbook of Crystal Growth: Bulk Crystal Growth: Second
Edition (Vol. 2, pp. 131–168). Elsevier Inc. https://doi.org/10.1016/B978-0-444-63303-
3.00004-3

[24] Stefanik, T., Gentilman, R., & Hogan, P. (2007). Nanocomposite optical ceramics for
infrared windows and domes. Window and Dome Technologies and Materials X, 6545,
65450A. https://doi.org/10.1117/12.719312

34
 2. HOT-PRESSING PLATELET ALUMINA TO TRANSPARENCY

2.1 Contributions and Introduction

The persons listed contributed to the following tasks relevant to this section:
• Dr. Klaus Himmel of Merck KGaA (EMD Performance Materials) for supplying
RonaFlair® White Sapphire powder used for aligned alumina.
• Dr. Andrew Schlup in his mentorship of graphite die set-up, modifying hot-press
parameters on the Centorr™, and post-processing care.
• Mr. Dawson Smith on assistance during Au-Pd sputter coating.
• Dr. Talukder Alam for training on FEI Quanta 650 SEM.
• NSL Analytical for ICP-MS results on powder composition.
• Mr. Tim VanMeter for technical assistance during centrifuging.

As a preface to this section, do note some of the many approaches taken to prepare the
platelet alumina powder were inspired by the work of Schlup et al., whose promising results of the
capabilities of platelet morphology polycrystalline alumina is a motivation for this research. Dr.
Schlup was not only part of the same laboratory in the past, but a personal mentor in helping me
with the process of taking alumina powder and successfully turning it into window material.
Section 2 goes into detail of the specific procedures taken to undertake such a task, in the hopes of
helping future engineers more easily understand and replicate the ceramic processing methods
used.

2.2 Experimental Procedures

2.2.1 Powder Preparation

Polycrystalline alumina powder with platelet morphology was exclusively used. The
RonaFlair® White Sapphire had a diameter of approximately 11 µm with thickness of 0.5 µm. Four
variants of this initial powder batch were made and are shown in Figure 2.1, resulting in five
overall powder batches. Two primary methods of purification were used. These include washing
and calcining (heat-treating).

35
 The first was a washed batch. The powder was washed in high-purity de-ionized water with
purity of 18.1 Ω-cm and roller-milled overnight (approximately 8 hours). The milled powder was
subsequently centrifuged and decanted. This washing process was repeated for 5 cycles.
Subsequently, the powder was oven-heated in an alumina crucible to 120°C for 4 hours to
evaporate the remaining water and moisture. The powder was finally sieved through a 250 µm
nylon mesh to de-agglomerate the powder more evenly.
The next is a calcined batch. Powders were calcined using a Muffle™ Laboratory Furnace.
Alumina crucibles were cleaned and washed with ethanol, hand-dried, then left to fully dry for 5
minutes. The sieved powder was placed in the crucibles at 50 g at a time, and set to the following
ramping cycle: room temperature to 1100°C at 10°C/min, one-hour isothermal hold at 1100°C,
then cooled down to room temperature at 10°C/min. The newly calcined powders were left to cool
overnight (~minimum of 6 hours).
SEM Micrographs of the five powder variants were obtained using the FEI Quanta 650
Scanning Electron Microscope. Grain size and diameter measurement and statistical analysis was
done through ImageJ program processing. Annealing for SEM was done with a Muffle laboratory
furnace with a max temperature of 1500°C in an alumina crucible. Two coats of Au-Pd each for
60 seconds were applied on every SEM sample using an MSE PRO™ Magnetron Sputter Coater.
Figure 2.1 captures each variant in the following SEM images.

36
37

Figure 2.1 (a-e) SEM Images of RonaFlair® Under Various Powder Processing Methods
 From Figure 2.1, each of the five variants of powder used are shown. Henceforth, each
batch of powder discussed will be referred to as powder batch “A”, “B”, etc., according to each
image in Figure 2.1, respectively. Some common trends can be seen amongst all five of the images.
First, the platelet nature of these powders is shown. Notice, for example, in image 2.1(a), the SEM
observed multiple discs (or plate) like grains stacked atop one another. It’s important to understand
that while these images show such morphology in a two-dimensional manner, the actual
orientation of the crystals in the image extends into the third dimension. These morphological
patterns are shown in images 2.1(b), 2.1(c), 2.1(d), and 2.1(e).
Between images 2.1(a) and 2.1(b) are the effects of sieving the platelet powder. The
differences are subtle but can be shown by observing small areas of the SEM images. Some areas
seem less agglomerated in 2.1(b), but this observation is purely qualitative. Such an observation,
if quantitively shown as well, would be supported, as the practice of sieving is to de-agglomerate
the starting powder by separating the basal planes, allowing for a more separated and finer powder
bed for sintering. However, considering the sieve mesh was maximized at 250 µm, it would explain
the reason why little difference is shown between the images, as the grain scale is below the 10
µm level. Sieving differences would not qualitatively be manifested as topographical differences
under the SEM.
Between images 2.1(b) and 2.1(c), we can observe the effects of water-washing the
RonaFlair® as it relates to topography of the platelet powder. Qualitatively, it’s difficult to compare.
As such, it appears both variants have similar surface topography at this scale.
The effects of calcination would be shown between images 2.1(d) and 2.1(e), as it relates
to topography of the powders. While it is difficult to observe differences in the images in Figure
2.1, Schlup et al. showed calcination of powders resulted in a decrease of up to 4% in absorption
losses1. The process of calcination effectively heated the sample to remove residual impurities or
oxides. When compared to the as-received variant of RonaFlair® in 2.1(a), subtle qualitative
differences may exist, however, such support isn’t shown in the SEM image but may be shown in
the resultant PCA samples produced based on the calcined powder variant.

Chemical Composition of RonaFlair® Batches

100-gram batches of raw RonaFlair® powder and washed+sieved powder (batches A and
C) was sent to NSL Analytical Services. The composition tests used inductively coupled plasma

38
mass spectrometry (ICP-MS) to identify trace elements present in each powder batch. Table 2.1
below contains results for the two batches.

Table 2.1 Chemical Compositions of As-Received (Batch A) and Washed (Batch C) RonaFlair®
% Composition %Composition
Element
(As Received/Batch A) (Washed+Sieved/Batch C)
Al 52.8 52.7
C 0.0016 0.018
Ca <0.005 <0.005%
Cr <0.001 <0.001%
Fe <0.001 0.005%
K 0.13 0.025%
Na <0.005 0.24%
Si <0.005 0.009%
Ti <0.001 <0.001%

From Table 2.1 we observe the chemical compositions of batches A and C. While the
washing process did not influence the primary element present in the RonaFlair® batch, there are
some key differences that can be determined by analyzing the changes in the trace element
compositions. First, we observe that in batch A, the RonaFlair® powder contains both carbon and
potassium in not an insignificant quantity, with the primary contaminant being potassium at 0.13%
composition. Washing and sieving the RonaFlair® powder reduced the potassium composition to
0.025% (by more than half). However, this is at the cost of an increased sodium content (0.24%),
which was originally a trace element at a composition of <0.005% in batch A. The optical effects
of washing techniques will be discussed in section 3.

Effects of Washing

In terms of powder composition, the effect of washing was the increase in the content of
sodium, at the expense of potassium. It was shown in Schlup et al. that the primary alkali impurity
in the RonaFlair® was sodium, likely in the form of either sodium carbonate or sodium oxide1.
However, in this study’s RonaFlair® batch, the primary alkali impurity was potassium. There could
be several reasons as to the increase in sodium content during the washing process. The most likely
contender, being natural impurities in the water used for washing. High purity de-ionized water
was used for all washing steps, but it is possible for trace alkalis such as sodium to be present.

39
 It was shown by Rasmussen et al. that the effect of alkali impurities on polycrystalline
alumina results in a loss in some mechanical properties, such as hardness6. Schlup et al. found that
reducing the content of alkali impurities such as sodium subsequently reduces absorption losses1.

2.2.2 Die Prep and Hot-Pressing Parameters

Polycrystalline alumina discs 25 mm in diameter and 1.5 mm in thickness were sintered in

a graphite die using Centorr furnace, TestorrTM series, mounted in an MTS Hydraulic Frame,
Model 312.21. The anatomy of the die-prep procedure from top to bottom is as follows: a graphite
ram (pushrod), followed by 2 layers of 0.26 mm thick graphoil, followed by a graphite spacer,
followed by 2 layers of graphoil, followed by a graphite spacer, followed by 2 layers of graphoil
followed by boron nitride spray, followed by 0.127 mm high-purity molybdenum foil, followed
by the platelet alumina powder bed (6.08g). From the powder bed upward is another layer of
molybdenum foil, followed by boron nitride spray, followed by 2 layers of graphoil, followed by
a graphite spacer, followed by 2 layers of graphoil, followed by a graphite spacer followed by 2
layers of graphoil, followed by a graphite ram (pushrod). A detailed cross-sectional illustration is
shown on the next page (the diameter of which is 25 mm)
The graphite die was then uniaxially cold-pressed to a pressure of 7 MPa. The prepped
graphite die was then placed in the furnace chamber and a vacuum was pulled for at least 6 hours
(or until a vacuum of better than 65 millitorr). This vacuum was maintained from room temperature
to 1550°C, heated at a rate of 25°C/min, then backfilled with gettered nitrogen with a flow rate of
4L/min. This backfill occurs at this temperature because alumina and graphite (carbon) exhibit
phase reactions at temperatures commonly above 1650°C. This nitrogen atmosphere is held until
the end of the hot-press run. The top platen of the MTS load frame is placed 6.5 mm away from
die contact to allow room for thermal expansion during heating. Upon reaching maximum
isothermal temperatures (1700°C and 1800°C) for the sintering run, an initial pressure of 1 MPa
was applied to first raise the platen to contact the die, at a rate of 1.3 MPa/min. Upon contact, a
final 20 MPa of pressure was applied at the same rate, taking approximately 15 minutes for total
pressure application. These conditions were maintained isothermally for various holding times.
Afterwards, cooling began at 25°C/min to a 1000°C and held 10 min, during which time the load
was removed. Finally, the die was cooled to room temperature at 25°C/min. A graph illustrating
the hot-press parameters with a total of three independent testing levels (time and temperature

40
variant) is shown in Figures 2.2 and 2.3 in the following pages. These testing levels are henceforth
known as the “Standard” (control group), “Long” (longer dwell), and “Two-Step” (two-step longer
dwell).

Figure 2.2 (a,b) Cross-Sectional (Front) Illustration of Graphite Die Anatomy During Preparation
(left) and Centorr Furnace Loading Conditions1 (right)

41
42

Figure 2.3 Sintering Profiles of Polycrystalline Alumina (Red = Standard, Blue = Long, Yellow = Two-Step)
 From the three hot-pressing profiles shown in Figure 2.3, the Standard profile of a 5-hour
isothermal hold at 1800°C is the control group. This is the profile previously shown in Schlup et
al. to have experienced high in-line transmission, but grain growth at higher hot-pressing
temperatures. The primary intentions of the two other sintering profiles (Long and Two-Step) are
to minimize grain growth, while maximizing densification. This is achieved with lower isothermal
temperatures (1700°C) and longer dwell times (10 hours). The Long test case was the first profile
variant tested along the Standard test case.
The Two-Step test is to assess the impact of a short period of high isothermal temperature
densification, followed by longer dwell times. This test case is identical to the Long test case, but
includes a 15-minute period during pressure application where a higher isothermal temperature is
held at 1800°C, followed by a 10-hour dwell at 1700°C.
The two-step sintering (TSS) process is a technique used in recent decades to obtain higher
density ceramic bodies while minimizing grain growth. Lóh et al. reviewed two different
approaches to TSS by both Chu et al. and Chen and Wang2,3,4. Chu et al. describes a TSS profile
used to increase material densification (Figure 2.4 (left))3. This involves pre-coarsening the
material first at a lower temperature to eliminate residual particles with the goal of a more
homogeneous microstructure. However, in the case of alumina, it was found that the second, higher
temperature stage resulted in significant grain growth, resulting in “great structural disarray”. The
approach described by Chen and Wang involves first a stage of higher temperature, followed by a
longer dwell at lower temperature4. Plots of these profiles are shown in Figure 2.4 below.

Figure 2.4 Two-Step Sintering (TSS) profiles: described by Chu.et al [TSS-C (Left)3] and Chen
and Wang [TSS-CW (Right)4]2

43
 In the plots shown in Figure 2.4, T1 refers to the temperature during the first isothermal
step and T2 refer to the temperature during the second isothermal step, whereby t1 and t2 describe
the duration of T1 and T2, respectively. For alumina ceramics, the technique by Chen and Wang
has been widely used4. Li and Yi applied the TSS-CW profile to alumina with particle size 10 nm
with polyacrylamide gel to an initial temperature for t1 = 1 hour, followed by a lower isothermal
temperature (by 100°C) for t2 = 34 hours2. Relative density of resultant samples was greater than
96%. Hesabi et. Al was able to densify alumina with similar starting grain size to the RonaFlair®
to greater than 99%, while minimizing grain growth2,5. Therefore, this paper uses a similar
experimental approach following the TSS-CW profile. Variations of the TSS-CW profile, as well
as this thesis’ Two-Step profile, would be worthwhile investigating for future work related to PCA.

2.2.3 Post-Processing

Rough Polishing and Surface Grinding

Alumina samples were carefully taken out of the hot-press die and carefully hand-polished
for approximately 15 minutes on either side using an Apex Terra Magnetic Diamond Grinding
Disc, 165 µm and 10 in. diameter. This was to eliminate any residual molybdenum foil or graphoil
that might have attached itself to the sample in thermal contraction during cooling. The specimen
was then mounted on a metal plate, liberally binding with a thermoplastic polymer (CrystalbondTM,
509-3). A diamond wheel with 600 grit was used for grinding down to 2.0 mm in thickness, equally
from both sides. The final 0.5 mm on either side of the alumina specimens is grinded using a 420
grit wheel to bring the thickness to 1.5 mm.

Optical Polishing

The final, and most critical, polishing step was done using a LECO® Autopolisher with
diamond suspension and lubrication. The first three, which were Diamat pads with water
lubrication, were 74 µm (6 min), 44 µm (10 min), and 30 µm (10 min). This was followed by the
diamond suspensions of 15µm (30 min.), 9 µm (1 hr.), 6 µm (6 hrs.), 3 µm (2 hrs.), and 1 µm (3
hrs.). The procedure is outlined in the following Table 2.2.

44
 Table 2.2 Auto-Polishing Procedure for 25 mm samples
45
 2.3 Physical Measurements

Archimedes’ density test method was conducted to obtain relative densities of fully
polished 1.5 mm thick alumina specimens. Dry weight was calculated after all wet calculations.
Samples were dried in an alumina crucible for a minimum of 2 hours in an oven.

2.4 Summary and Conclusions

The effects of various powder processing parameters were shown through SEM
micrographs, illustrated in Figure 2.1. Topographical differences between variants in the SEM
cannot be qualitatively determined. Compositional differences between washed and unwashed
powders show the dissolving of some residual alkalis in place of new ones. The differences in the
powder batches can be determined by testing hot-pressing parameters. The various hot-pressing
parameters described are done so with the goal of improving upon the results published by Schlup
et al., by minimizing grain growth. This is approached primarily with lower dwell times and overall
lower isothermal operating temperatures, also utilizing a two-step sintering approach. The next
section discusses the optical results for the powder variants outlined in this section.

2.5 References
1
[1] Schlup, A. P. (2020). PROCESSING ALUMINUM OXIDE FOR THE CONTROL OF
MICROSTRUCTURAL TEXTURE AND OPTICAL PROPERTIES.
2
[2] Lóh, N. J., Simão, L., Faller, C. A., de Noni, A., & Montedo, O. R. K. (2016). A review of
two-step sintering for ceramics. In Ceramics International (Vol. 42, Issue 11, pp. 12556–
12572). https://doi.org/10.1016/j.ceramint.2016.05.065
3
[3] M.-Y. Chu, L.C. De Jonghe, M.K.F. Lin, F.J.T. Lin, Precoarsening to improve

[4] microstructure and sintering and sintering of powder compacts, J. Am. Cer-

[5] am. Soc. 74 (11) (1991) 2902–2911

4
[6] X.-H. Wang, P.-L. Chen, I.-W. Chen, Two step sintering of ceramics with

[7] constant grain-size, I. Y2O3, J. Am. Ceram. Soc. 89 (2) (2006) 431–437.
5
[8] Z.R. Hesabi, M. Haghighatzadeh, M. Mazaheri, D. Galusek, S.K. Sadrnezhaad,

46
[9] Suppression of grain growth in sub-micrometer alumina via two step sin-

[10] tering method, J. Eur. Ceram. Soc. 29 (8) (2009) 1371–1377.

6
[11] Rasmussen, J. J., Stringfellow, G. B., Cutler, I. B., & Brown, S. D. (1965). Effect of
Impurities on the Strength of Polycrystalline Magnesia and Alumina. Journal of the
American Ceramic Society, 48(3), 146–150. https://doi.org/10.1111/j.1151-
2916.1965.tb16051.x

47
 3. EVALUATING INFRARED TRANSPARENCY OF PLATELET
POLYCRYSTALLINE ALUMINA

3.1 Contributions and Introduction

The persons listed contributed to the following tasks relevant to this section:
• All in-line transmission values obtained using the UV-VIS spectrophotometer were done
at Purdue’s Birck Nanotechnology Center, courtesy of Dr. Alexei Lagoutchev.
• Calculations for sample normalization to thickness explained by Dr. Andrew Schlup.
• Mr. Lester Gomez of Centorr Vacuum Ind. for assistance in vacuum operational capacity.
• Mr. Timothy VanMeter for assistance with mechanical pump installation.
• Mr. Tony Johnson of MTS™ in replacement of servo valve.

The effects of various powder processing methods and different sintering parameters on in-
line transmission are discussed in detail. The goal of this section is to evaluate the effectiveness of
these processing changes on PCA as an IR window candidate.

3.2 Experimental Methods

Two methods of optical spectroscopy were used. The first is using an ultraviolet-visible-
near-infrared spectrophotometer (UV-VIS Spectrophotometer). The second is using a Fourier
transform infrared spectrophotometer (FTIR Spectrophotometer). For much of the research
duration, only a UV-VIS spectrophotometer was available. Through the support of Lockheed
Martin, an FTIR was acquired towards the latter half of the research period. This greatly extends
the breadth of transmission data into more relevant IR wavelengths for hypersonic flight.
Therefore, while most of the results of this section come from a UV-VIS Spectrophotometer, the
section end discusses the specific capabilities of the FTIR as it pertains to PCA samples.
Transmission data in subsequent sections exclusively contain FTIR transmission results.

3.2.1 Measuring In-Line Transmission (TILT)

The UV-VIS spectrophotometer is a PerkinElmer® Lambda 950 UV-VIS-NIR with

integrating sphere. In-line transmission, total-transmission, reflection, and absorption can be

48
measured from the instrument, and forward and backward scattering can be derived from these
values.
The FTIR is a PerkinElmer® Spectrum™ 3 Tri-Range MIR/NIR/FIR Spectrophotometer.
The FTIR measures the absorption, transmission, or reflection of IR light in the sample by analysis
of molecular vibrations. In-line transmission was obtained for 25 mm diameter samples on a
sample mount provided by PerkinElmer® and placed in line with laser direction, with no
obstructions and lid closed, to minimize effects of ambient light. The FTIR possesses a
significantly broader range than the UV-VIS spectrophotometer.

3.2.2 Normalizing Optical Properties to Sample Thickness

Samples were normalized to 0.8 mm thickness, as is common in the transparent alumina

literature*. The measured transmission, which was percent in-line transmission, was taken using a
PerkinElmer® Ultraviolet-Visible Spectrophotometer. Some of these transmission measurements
are cross-referenced with a PerkinElmer® Fourier Transform Infrared Spectrophotometer (FTIR).
The measured values were used in normalization equation for calculations, shown below in
Equation 3.1

𝑑2
𝑇1 ( )
𝑑1
𝑇2 = (1 − 𝑅) ((1−𝑅)) (3.1)

In Equation 3.1, T2 is the normalized %in-line transmission, R is the theoretical reflection (14%
or 0.14 for alumina) based on index of refraction losses, d2 is the normalized thickness (0.8 mm)
and d1 is the measured thickness. Note when evaluating T2 during calculations, a whole number
value of R = 14% was used, to remain consistent with the percent transmission.

3.3 Results and Discussion

We will first be comparing methods of obtaining in-line transmission via two different
spectroscopic methods. Then, we will discuss effects of calcination, as outlined in section 2.2.1,
on in-line transmission in the near-IR. We will then discuss preliminary results on the effect of
sintering profiles on in-line transmission in the optical regime. Finally, we will then discuss the
effects of washing techniques and potential effects of changes in powder composition that were
discussed in section 2.2.1 on in-line transmission in the optical and near-IR. Note in testing,

49
vacuum capabilities were fully functional and the presence of vacuum during the ramping process
assisted in higher transmission values for Figures 3.2 and 3.3. However, during testing of the
washed samples (Figures 3.4 and 3.5), issues in maintaining the usual vacuum during sintering up
to a temperature of 1550°C were compromised, and therefore a standard gettered nitrogen
atmosphere was used. This has been shown to influence optical results by Schlup et al. and Coble5,6.
It was shown by Coble that in the presence of a vacuum, hydrogen, or oxygen atmosphere, it is
possible to full elimination of porosity in alumina. However, this is not possible in a nitrogen,
argon, or helium atmosphere. This therefore limits the peak in-line transmission we will observe
in the effects of washing on in-line transmission. While it is still possible to compare and analyze
results of washing, future testing could consider comparisons in a full vacuum atmosphere up to
1550°C.

3.3.1 Comparison of Spectroscopic Methods

The UV-VIS and FTIR Spectrophotometer use different methods to calculate in-line
transmission for samples. Both methods were used with commercial sapphire, Standard, and Long
test cases using Batch A powder, and were superimposed on the same transmission plot. Figure
3.1 below describes a wavelength over a wide breadth, from the optical up to far-IR.

Figure 3.1 In-Line Transmission using FTIR and UV-VIS Spectroscopy Methods of Sapphire and
PCA Samples (Batch A)

50
 From Figure 3.1, we can observe that sapphire remains the most transparent, while the
Standard and Long test cases follow, respectively in terms of transparency. The Standard and Long
test cases used were sintered from Batch A powder. From the plot, a combination of both UV-VIS
and FTIR data allows for a better understanding of in-line transmission over a broader wavelength
range. Many authors have published UV-VIS or FTIR results for sapphire*, but no literature
currently exists with a combined transmission illustration of sapphire and polycrystalline alumina
from the optical to far-IR range. The data from observing all three samples illustrates that while
different spectroscopic methods were used, similar transparency data is achieved in both instances.
This gives confidence in the in-line transmission values determined, as well as the consistency in
the transition range, between the near and mid-IR.

3.3.2 Results on Effect of Sintering Profiles

The Standard, Long, and Two-Step test cases were compared using a UV-VIS
Spectrophotometer. Transmission values were obtained and normalized to 0.8 mm thickness.
Archimedes’ density was calculated. The platelet powder used was Batch A. These results are
shown on the following page in Figure 3.2.

51
52

Figure 3.2 Results (Batch A) on Effect of Sintering Profiles on In-Line Transmission in the Optical Regime
 Figure 3.2 describes the in-line transmission of the Standard, Long, and Two-Step test
cases in the optical regime, also including their respective sample images. By solely examining
the plot, several observations are apparent. First, the standard test case displayed the highest levels
of optical transparency among the three, with peak transmission at 800 nm occurring at 73%.
Optical transmissions for the Long and Two-Step test case were lower in comparison, peaking at
60% and 64%, respectively, at 800 nm. Optical transmission for all samples was low until about
320 nm. This is consistent with the literature, as the wavelength range for alumina drops off in the
ultraviolet wavelength range1.
These trends can be qualitatively matched by observing the sample images for each test
case. Since the wavelength range is in the optical regime, observing the samples through the human
eye provides valuable feedback on the quality of transmission. The Standard test case was visually
the clearest. Both the Standard and Two-Step test cases possess an outer ring of significant
diameter to the total sample diameter. This could be due to the higher temperature at pressure
application, as this feature is not seen in the Long test case. Recrystallization of the samples is
known to occur sometime during the sintering stage after pressure application, as the pressure from
outward expansion of the green body causes the creation of a new strain-free grain matrix under
plastic deformation. Such phenomena were explained by Heuer et al. and Rhodes in their work on
polycrystalline alumina and recrystallization1,2. Heuer et al. observed temperature uniformity was
a key factor in the production of a homogenous sintered body1. The results from Figure 3.2 indicate
that temperature may also be a key factor in homogeneity of samples.
The Long test sample possessed an even optical surface, which might be explained by the
longer dwell times and lower temperature during pressure application. Visually the Long test case
contained a level of carbon contamination originating in the center, resembling soot, that is not as
present in the Standard and Two-Step test cases. Interestingly, this was not contamination that
could be removed by surface grinding or polishing and most heavily present in the center of the
sample. To verify, several subsequent Long test cases were all run, producing near identical results.
The resultant contaminant in the middle of the Long test case is likely a remnant of radially
moving plastic deformation over the longer dwell time, recrystallizing grains more favorably
towards the outer diameter due to the lower sintering temperature and a comparably weaker kinetic
driving force. Heuer et al. explains increasing the interfaces along the vertical path by adjusting
the thermal load results in less heat loss in the sintered body and a more homogeneous material 1.

53
Future work regarding longer dwell times of PCA could test modifications to the graphite die set-
up to involve more graphoil, higher quantity of spacers, or different layering materials. This might
reduce thermal load and vertical heat loss that might have contributed to the visible carbon
contamination around the center of the Long test cases. Various tests were run with Tungsten and
Boron Nitride (in place of Molybdenum) as a layering material but resulted in the same tinted
effect. Some papers combine these longer dwell periods with hot-isostatic pressing (HIP) and
observe no such effect1,2. This comes at the cost, however, of the main production benefit of
densifying platelet alumina in a single hot-press step.
The Two-Step test case is arguably more interesting, as it seems to possess a higher optical
transparency, exhibiting the same visual appearance, but contains significant carbon contamination
along the outer diameter. The higher initial temperature likely results in a transparent center and
more rapid growth of a strain-free grain matrix. However, reduction of the dwell temperature likely
doesn’t allow carbon contamination to leave the outer diameter and resulted in inconsistent Two-
Step test cases. Some transparency levels were higher than the Long test case, while some were
significantly less (~39%). Many samples were prone to cracking along the outer edges. Lóh et al.
examined several two-step sintering cases with higher t1 durations3. This thesis did not investigate
changing t1 durations, but future work might find success with a longer t1 duration for the Two-
Step case.

3.3.3 Effects of Calcination on Aligned PCA

The calcined and sieved powder batch (E) was tested for in-line transmission using
a UV-VIS Spectrophotometer. A Standard and Long test case were measured in near-IR
wavelengths, and the results are displayed in Figure 3.3 below.

54
 Figure 3.3 In-Line Transmission for Standard and Long Batch E Test Cases

From Figure 3.3, a few observations are apparent. The first is that the Long test case, when
calcined had a higher transmission value in the near-IR when compared to the Standard test case.
A peak transmission difference of 8% is consistent with the results from Figure 3.2, in both the
calcined and washed instances. Compared to the effects of calcination in Schlup et al., calcination
for the Long test case is twice as effective as calcination for the Standard test case (Schlup observed
a 4% increase)5. The transparency results, while promising, were less consistent than during a
Standard test case. One such explanation is that the doubling of the isothermal time makes the
sintered body more susceptible to mechanical instabilities in the furnace, load frame, atmospheric
level, etc. This could explain the common theme of the Standard test case being more consistent
than the Long test case, as well as the Two-Step test case.
With the primary impurity of the powder used being sodium based, it is likely through
calcination any sodium oxides or sodium carbonate dissolved, resulting in higher in-line
transmission values. This result is supported by the fact that the sodium oxide vapor pressure
results in sodium dissolving during calcination between 1000-1050K (~726-776°C)7. Hillenbrand
and Lau measured the sublimation of sodium oxides and found these lower and upper limits are in
accord with theoretical calculations7. Since our calcination temperatures were above this value, it

55
is likely the residual alkali impurity, sodium, is sublimated during the calcination process,
increasing the in-line transmission for the Standard and Long test cases.
The Long test case experienced a greater increase in its in-line transmission than the
Standard. This contrasts with the in-line transmission data observed from the Batch A powders,
shown in Figure 3.2. There are several likely explanations for this observation. The first is that the
Long test case (as is supported in section 4), has a smaller grain size because of the lower
isothermal sintering temperatures. The greater frequency of smaller grains means there is a larger
surface area per unit mass compared to a sample with larger grains. This increased surface area
would allow for more efficient heat transfer during sintering, and greater densification. This is
supported by the density values obtained for both samples. Smaller grain sizes also contribute to
denser packing and less transmission losses due to birefringence for aligned PCA samples,
resulting in the higher transmission observed. For these reasons, the calcination of the powder
(Batch E) coupled with the Long test case was by far the most successful in achieving full density
and minimizing transmission losses, while successfully eliminating the presence of the primary
alkali impurities present.

3.3.4 Effects of Washing on Aligned PCA Samples

The three PCA test cases based on sintering profile are further stratified by powder batch
A and C (as discussed in 2.2.1.1). Do note the subsequent tests described in Figures 3.4 and 3.5
involve the sintering of alumina in a gettered nitrogen atmosphere, and not a full vacuum (up until
1550°C due to mechanical limitations at the time). Figure 3.4 below illustrates six PCA samples
made based on these levels. The UV-VIS spectrophotometer was used to determine in-line
transmission, normalized to 0.8 mm. The transmission results are shown in Figure 3.5 below.

56
 Batch A: As-Received Standard Long Two-Step

ρ=99.89% ρ=99.84% ρ=99.81%

57

Batch C:Washed+Sieved

Figure 3.4 Polycrystalline Alumina Samples Based on Batch and Sintering Profile with Relative Densities
ρ=99.70% ρ=99.76% ρ=99.74%
58

Figure 3.5 In-Line Transmission of Polycrystalline Alumina vs. Wavelength as a Function of Powder Condition
 Figure 3.5 describes the relationship between TILT and wavelength for the six
polycrystalline samples. The two variables tested are testing batch (A and C) and sintering profile
(Standard, Long, and Two-Step). It should be noted that all samples tested, while normally done
in the presence of a vacuum until 1550°C during the sintering process, were backfilled at room
temperature. This is due to mechanical issues with the vacuum pump and a vacuum leak in the
sintering furnace. The presence of vacuum has been shown to help remove residual impurities,
such as carbon and alkali impurities in ceramic powders such as PCA, improving transmission.
This is supported by Rossi and Fulrath, as well as other sections of this paper4.
To understand the plot in Figure 3.5, it is important to analyze the qualitative properties of
the samples in Figure 3.4. Immediately, clear differences can be seen between Batch A and Batch
C. Starting with the Standard test case, Batch A showed no visible cracking or degradation. Carbon
contamination was minimal and visually displayed a high level of optical transparency. The
Standard case for Batch C displayed visible cracking, particularly around the outer edges. Parts of
the sample were irrecoverable during processing, and the sample was more translucent. For the
Long test case, the Batch A sample also appeared to have a high level of transparency but had a
noticeable gray tint homogenous along the entire material. While small cracks beginning from the
edges and travelling inward are apparent in some areas of the sample, the sample itself maintained
integrity during post processing. However, for Batch C, the sample possessed a long,
hemispherical crack, along with visible crystalline defects, causing a “shine” or “glow” to the
sample. This is reflected in poor visible transparency. Finally, for the Two-Step case, Batch A
possessed few cracks but appeared to have a gray haze when viewing optically. Reflected in Figure
3.4, the TILT was very low at (~27% at 700 nm). This was an interesting result. The Batch C case
possessed large carbon contamination, which was common amongst all Batch C cases, but
surprisingly had a higher transmission than its Batch A counterpart. While difficult to conclude
from these two test cases, it was commonly observed that the two-step sintering cases were more
prone to inconsistencies in their relative densification, supported by Lóh. et al.
While the human eye is a good indicator of optical properties, it is unable to determine
effectiveness of samples in the near-IR. Observing Figure 3.5, the Long Batch A test case
surpassed the Standard Batch A test case by a margin of 6% at 1200 nm. This was a promising
result for the Long test case, as one of the primary weaknesses of PCA was its poor IR transparency
past the near IR. Should both mechanical and thermal load capabilities of the Long test case prove

59
comparable or better than the Standard test case, it would make the Long sintering profile a more
suitable contender as an IR window.
The inconsistencies with the Two-Step test case, as well as the overall loss of transparency
due to the washing process, means for future tests and results discussed henceforth in this thesis
do not include either washing as a powder purification process, nor a Two-Step sintering profile.
Future testing on Two-Step sintering cases may prove more successful with longer t1 times or other
sintering or powder modifications.

3.4 Summary and Conclusions

The results from UV-VIS and FTIR showcase the consistency of sample transmission over
a broad wavelength spanning the optical to far-IR. Preliminary results compared effects of the
three sintering profiles discussed in section 2.2.2. These results used Batch A platelet powder.
Standard test cases produced the most consistent in-line transmission data. While Two-Step test
cases showed initial promise, Two-Step test cases resulted in inconsistent in-line transmission in
the optical and near-IR upon several subsequent tests in at least six continuous samples. The
process of washing outlined in 2.1.1 was overall ineffective in increasing in-line transmission.
Alkali impurities in the form of potassium were replaced by sodium during the washing process,
likely related to either contamination with milling media or purity level of water. Some carbon
impurities were present in samples along the outer edges, but likely had no effect on transmission
values, which were taken about the center of the sample. Calcining the platelet powder drastically
improved transparency values for particularly the Long test case, comparable and even exceeding
the Standard test case in the near-IR range. Calcination successfully reduced the sodium alkali
impurities present in the powder, and this is supported by Hildenbrand and Lau as well as the
densities obtained for both samples7.

3.5 References
1
[1] Heuer, A. H., Sellers, D. J., & Rhodes, W. H. (1969). Hot working of aluminum oxide. 1:
Primary recrystallization and texture. J. Amer. Ceram. Soc., 52, 468–474.
2
[2] Rhodes, W. H., Sellers, D. J., Heuer, A. H., & Vasilof, T. (1967). Development And
Evaluation Of Transparent Aluminum Oxide.

60
 3
[3] Lóh, N. J., Simão, L., Faller, C. A., de Noni, A., & Montedo, O. R. K. (2016). A review of
two-step sintering for ceramics. In Ceramics International (Vol. 42, Issue 11, pp. 12556–
12572). https://doi.org/10.1016/j.ceramint.2016.05.065
4
[4] ROSSI, R. C., & FULRATH, R. M. (1965). Final Stage Densification in Vacuum Hot-
Pressing of Alumina. Journal of the American Ceramic Society, 48(11), 558–564.
https://doi.org/10.1111/j.1151-2916.1965.tb14672.x
5
[5] Schlup, A. P. (2020). PROCESSING ALUMINUM OXIDE FOR THE CONTROL OF
MICROSTRUCTURAL TEXTURE AND OPTICAL PROPERTIES.
6
[6] COBLE, R. L. (1962). Sintering Alumina: Effect of Atmospheres. Journal of the American
Ceramic Society, 45(3). https://doi.org/10.1111/j.1151-2916.1962.tb11099.x
7
[7] Hildenbrand, D. L., & Lau, K. H. (1993). Mass spectrum and sublimation pressure of sodium
oxide vapor: Stability of the superoxide molecule NAO2. The Journal of Chemical Physics,
98(5), 4076–4081. https://doi.org/10.1063/1.465043

61
 4. MECHANICAL PROPERTIES OF ALIGNED HOT-PRESSED
ALUMINA

4.1 Contributions and Introduction

The persons listed contributed to the following tasks relevant to this section:
• Dr. Andrew Schlup, Dr. Olivia Brandt, Dr. Tess Martinez, Dr. Rodrigo Orta, and Mr. Kyle
Cox on their mentorship in surface grinding and machining ASTM flexure bars.
• Dr. Abdullah Saad in his technical assistance with graphing programs.

One of the most mission-critical characteristics of an infrared window is its ability to

withstand high levels of mechanical stress, where flexure strength typically is the weakest loading
condition. This section tests the mechanical strength of polycrystalline alumina for a desired room
temperature flexure strength of 150 MPa and calculates experimental Weibull modulus1.

4.2 Experimental Procedures

Mechanical testing requires flexure bars according to the American Society for Testing and
Materials (ASTM) 1161a standard. Thus far in this thesis, all PCA samples are 25 mm in diameter.
To machine ASTM standard flexure bars, larger diameter samples are required. Therefore, a
modified die preparation and processing methodology was created to compare the differences
between the Standard and Long test cases for PCA.

4.2.1 Modified Die Preparation

The die prep procedure has been modified by using a larger die opening diameter of 55
mm as opposed to 25 mm. This was done so alumina bend bars can be machined and polished for
flexure tests. The isometric pressure during sintering was also increased to accommodate the larger
area of the graphite push ram, to keep the parameters equivalent to the 25 mm test cases. The
detailed die illustration is shown in Figure 4.1 below and is like Figure 2.3 minus the diameter of
the graphite die.

62
Figure 4.1 Cross-Sectional (Front) Illustration of Graphite Die Anatomy During Preparation (55
mm)

The procedure for hot-pressing follows the conditions set out in section 2.2.2 using Batch
A powder, with one key difference. Due to the larger diameter of the die, significantly more
powder was used (29.31g vs. 6.08g).

4.2.2 Post-Processing

Surface Grinding and Polishing

The Standard and Long test cases for the 55 mm samples were subject to the same sintering
profiles outlined in section 2.2.2. After sintering, the larger 55 mm PCA samples were gently taken
out of the hot-press die and carefully hand-polished for approximately 15 minutes on either side,
again using an Apex Terra Magnetic Diamond Grinding Disc, 165 µm and 10 in. diameter to
eliminate residual graphoil and molybdenum. The specimen was then mounted on a metal plate,
liberally binding with a thermoplastic polymer (CrystalbondTM, 509-3).

63
 A diamond wheel with 600-grit was used for grinding down samples from about 5.0 mm
in thickness to 3.0 mm in thickness, equally from both sides. The final 0.5 mm on either side of
the alumina specimens is grinded using a 420-grit wheel to bring the thickness to 2.5 mm. Once
the samples were at this thickness level, optical polishing was done in a process like the 25 mm
samples. The polishing procedure is outlined in Table 4.1 below. This polishing procedure is like
the procedure for the 25 mm sample but does not include the finer grit polishing scheme (9 µm
and below).

64
 Table 4.1 Auto-Polishing Procedure for 55 mm samples
65
 The resultant 55 mm samples after polishing were approximately 3 mm in thickness and
ready for machining into bend bars. A physical comparison between these larger samples and the
25 mm optical samples is shown below in Figure 4.2.

(a) (b)

Figure 4.2 (a,b) Size Comparison of Pre-Polished Standard 55 mm (thickness 3 mm) Aligned PCA
to the Standard 25 mm Alumina (thickness of 1.5 mm)

From the figure, we can immediately observe several consequences of the larger sample
diameter and thickness. Referring to the Rayleigh-Gans-Debye Equation 1.4, the additional
thickness contributes to visible losses in transparency to the naked eye. Therefore, we can speculate
the transmission for the larger sample is significantly less than that of the smaller. The high level
of polish present in the 25 mm sample is not fully present in the 55 mm sample, which would also
contribute to transmission losses through specular reflection. Portions of the sample contain carbon
contamination from sintering that were unable to be polished out. However, through the machining
process, which involves drastically reducing the thickness of the material from 3 mm to 1.5 mm,
most of these contaminants were eliminated.
The load conditions during sintering for both samples were identical. Given that the 55 mm
sample had a larger diameter and greater surface area, hot-pressing should have resulted in a
sample with less densification. However, this was not the case. Density values post-processing
were within 0.5%. This hints that a higher sintering pressure has little effect on densification after
a certain point with aligned PCA samples. This result is supported by Schlup et al., who discussed

66
the effects of sintering pressure on crystallographic orientation of aligned PCA5. It was found that
the relative density as a function of hot-pressing pressure varied little between 10 MPa and 25 MPa
for the aligned PCA samples. Therefore, the similar density values obtained by using the same
pressure for both the 55 mm and 25 mm samples can be explained by particle rearrangement that
occurs during hot-pressing to crystallographically align PCA than the hot-pressing pressure
parameter. It is shown that temperature is a far more important parameter for achieving
densification than loading pressure, with the majority of densification occurring within the first 15
minutes5.
For the following mechanical tests, we therefore machine the larger 55 mm samples for the
purposes of flexure testing and use the normal 25 mm samples for characterization and grain size
measurements, considering the relative densities of both samples is comparable.

Machining

The bars were machined from 55 mm samples according to ASTM-1161a. The

dimensions of the bend bar are shown below:

Figure 4.3 ASTM-1161 “A” Flexure Bar2

The orientation of the 55 mm alumina samples is shown below in relation to the hot-press
direction:
10 machined samples were created for the Standard and Long test case. All samples were
polished on the tensile side following the polishing procedure outlined in Table 4.1. All samples
were measured within the tolerances outlined in ASTM-1161a and at minimum 25 mm in length,
although most samples exceeded this minimum length.

67
 Figure 4.4 55 mm Hot-Pressed Sample Orientation During Machining

4.3 Mechanical Testing

4.3.1 Experimental Apparatus

Mechanical testing was done using an MTS® Electromechanical Test System, a load frame
that would apply specified compressive force until fracture. Flexure testing was done using a 4-
point flexure test apparatus from McMaster Carr. Machined samples would be placed, polished
side facing up, in the apparatus shown below in Figure 4.5a. After samples were individually
placed, they were placed in the electromechanical load frame for flexure testing. An example of
the set-up is illustrated in Figure 4.5b below.

68
 (a) (b)
Figure 4.5 (a,b) 4-Point Flexure Test Apparatus (a) and Prepared Sample Ready for Testing in
MTS® Electromechanical Test System (b)2

Samples were applied stress until failure, upon which the load frame halted. Average load
and peak stress were recorded for all samples. These recorded values, along with values for
specimen dimensions, were used in Equation 4.1 below to determine the strength of the beam for
4-point flexure testing.

(4.1)
In Equation 4.1, S represents the characteristic strength of the beam, P represents the peak
force at beam failure, L represents the length of the outer support, b represents the specimen width,
and d represents the specimen thickness. These values varied from sample to sample within the
acceptable tolerances and were used in Equation 4.1 to determine characteristic strength. These
calculated strength values were used for Weibull modulus calculations, discussed below.
Griffith failure criteria was not calculated, as the measurement of crack length in samples
was difficult and inconsistent, but is an important theory based on the critical flaw size in materials.
These flaws are stress concentrators that when a critical stress value is exceeded, propagate cracks,

69
leading to brittle fracture. This is experienced in our samples when the material goes under 4-point
flexure testing. The equation for the Griffith criteria is shown in Equation 4.2 below.

(4.2)
In Equation 4.2, σG is the applied tensile stress, E is the Young’s Modulus, γs is the surface
energy per unit area, and a is the length of the flaw or crack. Future work could evaluate the Griffith
criteria to provide more insight into the distribution of flaws and mechanical properties.

4.4 Results and Discussion

This results section will compare the Standard and Long test cases in calculations of
Weibull modulus, as well as the effect of sintering profile on average grain size and microstructure.

4.4.1 Weibull Modulus Calculations

The Weibull modulus is a dimensionless parameter used for ceramics and brittle materials
to describe the variability in the distribution of the maximum stress a material can withstand before
failure3. A high Weibull modulus is representative of little variability in the distribution of physical
flaws in the material, instilling more confidence in the mechanical strength of the material.
Inversely, a low Weibull modulus is representative of a high variability in the distribution of
material flaws, which is less desired. For both the Standard and Long test cases for flexure testing,
Weibull distributions are outlined below in Figure 4.6.

ln(Peak Stress [MPa])

70
 ln(ln(1/(1-F)))

Figure 4.6 Weibull Modulus Distribution of Standard and Long Aligned PCA

Overall, 10 samples were tested for each test case, resulting in 20 overall samples. While
it is always preferrable to obtain large amounts of data for analysis, a sample size of at least ten
for each failure case has been used previously in literature3. Details of Figure 4.6 will be discussed
in 4.4.2. A table outlining the average peak stress, maximum load, and Weibull modulus is in Table
4.2 below. It should be noted that eliminating the two lowest outliers for each case results in a
Weibull modulus for the Standard and Long Case m = 24.2 and m = 8.8, respectively.

Table 4.2 Weibull Modulus, Load, and Stress Values for Standard and Long Aligned PCA
Standard Long
Average Peak Stress (MPa) 158.5 ± 6.3 195.1 ± 8.3
Maximum Peak Stress (MPa) 167.3 ± 7.2 234.7 ± 24.0
Weibull Modulus (m) 28.8 ± 3.5 9.5 ± 0.8

4.4.2 Effects of Sintering Parameters on Weibull Modulus

From Figure 4.6 several observations can be made. First are the Weibull Modulus values
for both the Standard and Long test cases. The Weibull modulus value for the Standard is

71
significantly higher than that of the Long case, by three-fold. It is indicated graphically by the
steeper slope, indicating the distribution of physical flaws along the material are less varied than
those in the Long test case. Some reasons for these differences would include some difference in
the processing parameters. Average grain size would be an indicator for these differences, and this
is discussed in the following section 4.4.3. We can see for the Standard test case, the distribution
of values is much less varied than for the Long test case, while the Standard test case also had a
lower maximum and average peak stress. There are several reasons for these observations. The
first is understanding the Weibull modulus is character of the scatter in the distribution of flaws
along the material. Differences in grain size distribution between the Standard and Long test case
could lead to differences in the Weibull modulus values. This will also be discussed in section
4.4.3.
In terms of the values for the Weibull modulus, Carisey et al. found Weibull modulus
values for platelet alumina with 3-7 µm grain size at approximately m = 8 4. Schlup et al. found in
1,5
pre-aligned polycrystalline alumina samples, the Weibull Modulus had a value of m = 7 . The
processing conditions were like the Standard test case used in this research, and Schlup et al.
attributed the finding to a tightening of the grain distribution. Approximate values of m = 29 for
the Standard test case and m = 10 for the Long test case show significant improvements in terms
of Weibull value. By examining grain size distributions, a more detailed understanding of these
high Weibull values can be found.
A more recent paper by Schlup et al. found a Weibull modulus of m = 24 for aligned
polycrystalline alumina samples, which is consistent with the findings for the Standard test case in
this study5. This highlights the importance of crystallographic alignment, as the difference in
Weibull values between pre-aligned and aligned samples are consistent (at least three-fold the
value for aligned PCA).

4.4.3 Effects of Grain Size on Mechanical Properties of Aligned PCA

Grain size was determined using line-intercept method and processed through ImageJ. An
average of 1200 grains was taken for each test case. These test cases were made from Batch A
powder. Table 4.3 reports these findings, along with the relative densities via Archimedes’
Method, and Figure 4.7 on the following page displays samples of SEM images for each test case.
The line-intercept method was used to obtain grain size values and error bars. The SEM represents

72
a cross-sectional view in the direction normal to the hot-press direction. Statistical analysis was
also performed, and this is shown in Figure 4.8.

Table 4.3 Relative Densities and Average Grain Size of Standard and Long Aligned PCA
Standard Long
Relative Density (%) 99.54 99.86
Average Grain Size (µm) 22.2 ± 2.9 11.1 ± 1.4
Maximum Peak Stress (MPa) 167.3 ± 7.2 234.7 ± 24.0
Weibull Modulus (m) 28.8 ± 3.5 9.5 ± 0.8

73
 Standard Long

Hot-Press Direction
Hot-Press Direction

Prior Work (Schlup et al. Larger Grains) Current Approach (Smaller Grains)
74

Figure 4.7 SEM Images of Prior Work (scale bars = 200 µm) and Current Approach Standard and Long Aligned PCA (scale bars = 100 µm)
 Standard Long
75

Figure 4.8 Grain Size (left) and Normalized Average Grain Distribution (right) for Standard and Long PCA
 From the SEM images, several observations can be made. First, looking at the Standard
test case, we can see that intragranular porosity is present in much greater concentration than
compared to the Long test case. This result is consistent with Schlup et al., among other papers1,5,8.
In the Long test case, some areas (likely a colloidal residue from post-processing) are present.
Ignoring them, the porosity, both intragranular and intergranular, in the Long test case is
significantly less than that of the Standard test case. This is supported by the relative density
calculation. These results also explain the Weibull modulus values and characteristic strength of
each test case. The Long test case was able to withstand a much greater peak load on average
before fracturing when compared to the Standard test case. This is supported by the visual analysis
of the grain size distribution in Figure 4.7, and it is also supported in the plots of average grain
size distribution in Figure 4.8.
One observation, alluded to previously, is the Standard test case has a significantly higher
Weibull modulus, while having a lower maximum and average peak failure stress. To understand
this phenomenon, it might be helpful to consider the cause of fracture for these materials. For
aligned PCA, a smaller grain size would result in a greater grain boundary distribution. Materials
with stronger or more coherent grain boundary distributions may exhibit higher a peak failure
stress, despite having lower a Weibull modulus. This is supported by Klein et al., who observed
similar behavior in fused silica9. For both Standard and Long test cases, the spread of the flaw size
distribution is homogenous, but much more so for the Standard test case. The difference here could
be a result of several factors. This could be the surface finish, texture, and grain boundary
distribution. From Figure 4.7, we can observe that there is a textural difference between the
Standard and Long SEM images, in that the Long test case shown elongation perpendicular to the
hot-press direction. This could indicate the presence of grain boundary sliding or some other
deformation mechanism that occurs during hot-pressing. The most likely contender is primary
recrystallization, which was observed by Heuer et al. in several instances8. Heuer et al. observed
at temperatures as low as 1750°C elongation, like the ones observed in this study, for
polycrystalline alumina in press-forging conditions. A similar deformation texture perpendicular
to the press-forging direction was observed for samples, whose relative densities exceeding 99.5%,
the reduction in matrix strain energy being the driving force for this interaction. It was concluded
that basal slip was the primary method of deformation that contributed to recrystallization,
whereby dislocation move along planes parallel to the basal plane in alumina’s hexagonal close-

76
packed crystal structure. The presence of basal slip deformation can contribute to the increased
flexure strength and maximum peak strength observed in the Long test case, that is not seen in the
Standard test case. It is possible the more homogeneous texture in the Standard SEM image in
Figure 4.7, when compared to the Long case’s texture is a result of some other mechanism. Despite
the Standard having a higher operating temperature than the Long (1800°C vs. 1700°C,
respectively), and therefore a higher activation energy to allow for basal slip deformation, this
texture is not observed in the Standard test case and is almost always observed in several samples
for the Long test case. It is possible the process of grain boundary diffusion or rate dependent
diffusion is ideal at a particular temperature range. This is supported by the SEM images provided
by Schlup et al. on the left side of Figure 4.7. We can observe that at temperatures below 1800°C,
we observe a similar recrystallization texture consistent with this study and work by Heuer et al1,8.
This leads to the conclusion that at a temperature below 1800°C lies a critical temperature to allow
for basal slip deformation, resulting in recrystallized texture. This pattern is not seen above this
temperature. The effect this has on Weibull modulus is unknown but might be a possible
explanation as to the lower Weibull values for the Long test case, despite the higher characteristic
strength of the Standard test case.
From Figure 4.8, we can see for a random sample size of 120 grains for each test case, the
average grain size for the Long test case is lower than the Standard case. This is also true for the
average grain area. It is most likely the lower processing temperatures for the Long test case
resulted in not only less kinetic force for grain growth to occur, but fewer exaggerated grains and
a more even grain distribution. This conclusion is supported by Rice, who worked on hot-working
oxides, such as alumina7. His results based on hot-extrusion temperatures were like that of Heuer
and Rhodes, in that lower sintering temperatures lessen grain growth observed7,8.
Referencing optical results from section 3.3.3, the Long test case, when calcined also
exhibited higher in-line transparency values when compared to the Standard test case. The Long
test case also exhibited a higher average peak load before 4-point failure of 234.7 N as compared
to 167.3 N for the Standard test case. Despite the Standard test case having a Weibull modulus of
m = 28.8, the Long test case also displays a Weibull modulus of m = 9.7 that is still high for platelet
alumina when compared to current literature5,6.

77
 4.5 Summary and Conclusions

The goal of this section was to determine flexure strength of polycrystalline alumina based
on the Standard and Long testing conditions. Larger 55 mm PCA samples were machined to make
ASTM “a” bars. Mechanical tests were conducted using 4-point flexure tests according to
ASTM1161a. The characteristic strength was determined, and Weibull modulus was calculated
and plotted based on test case. The Standard test case had a Weibull value of m = 28.8 and the
Long test case had a Weibull value of 9.7. These values are high for platelet alumina, when
compared to literature4,5,6. This is reflected in the very narrow spread of grain distribution in both
samples, particularly the Long test case. The Long test case hints at undergoing a form of primary
recrystallization under plastic deformation based on SEM micrographs and texture. This is
consistent with other literature1,5,7,8. Overall, grain size was successfully decreased and
mechanically, both mark improvements to aligned alumina as a potential hypersonic IR window.

4.6 References
1
[1] Schlup, A. P. (2020). PROCESSING ALUMINUM OXIDE FOR THE CONTROL OF
MICROSTRUCTURAL TEXTURE AND OPTICAL PROPERTIES.
2
[2] Flexural strength of advanced ceramics at ambient ... (n.d.-b).
https://sites.utexas.edu/taleff/files/2019/10/astm_c1161_13.pdf
3
[3] (pdf) 3.3.2 Model Weibull.pdf | AAN Junior - academia.edu. (n.d.-c).
https://www.academia.edu/37584350/3.3.2_Model_Weibull.pdf
4
[4] Carisey T, Leviri I, Brandon DG. Microstructure and mechanical properties of textured
Al2O3. J Eur Ceram Soc. 1995; 15(4): 283–9. Carisey T, Leviri I, Brandon
DG. Microstructure and mechanical properties of textured Al2O3. J Eur Ceram
Soc. 1995; 15(4): 283–9.
5
[5] Schlup, A. P., Costakis, W. J., Youngblood, J. P., & Trice, R. W. (2023). Mechanical
properties of hot-pressed transparent platelet alumina. International Journal of Applied
Ceramic Technology, 20(6), 3656–3665. https://doi.org/10.1111/ijac.14462

[6] 6
Munro, R. G. (1997). Evaluated material properties for a sintered α-alumina. In Journal of
the American Ceramic Society (Vol. 80, Issue 8, pp. 1919–1928). Blackwell Publishing Inc.
https://doi.org/10.1111/j.1151-2916.1997.tb03074.x
7
[7] Rice, R. W. (1970). Hot-Working of Oxides. In Magnesia, Alumina, Beryllia Ceramics:
Fabrication, Characterization and Properties: High Temperature Oxides, Part III (Vol. 5).
ACADEMIC PRESS, INC. https://doi.org/10.1016/B978-0-12-053303-9.50013-7

78
 8
[8] Heuer, A. H., Sellers, D. J., & Rhodes, W. H. (1969). Hot working of aluminum oxide. 1:
Primary recrystallization and texture. J. Amer. Ceram. Soc., 52, 468–474.
9
[9] Klein, C. A. (2009). Characteristic strength, weibull modulus, and failure probability of fused
silica glass. Optical Engineering, 48(11), 113401. https://doi.org/10.1117/1.3265716

79
 5. EFFECT OF CRYSTALLOGRAPHIC ALIGNMENT OF
POLYCRYSTALLINE ALUMINA ON SURVIVAL IN A SIMULATED
HYPERSONIC ENVIRONMENT

5.1 Contributions and Introduction

The persons listed contributed to the following tasks relevant to this section:
• Dr. Abdullah Saad in his supervision during oxyacetylene ablation tests and for his CAD
template used to modify a graphite holder design.
• Rick Lamar of AdValue Technology for supplying aligned single-crystal sapphire.
• McMaster-Carr for supplying equiaxed polycrystalline alumina.

This chapter dives into one of the most interesting testing avenues of platelet polycrystalline
alumina as an IR window candidate. Despite alumina being one of, it not the most researched
refractory oxide ceramic, few works in literature test the efficacy of polycrystalline alumina as an
IR window in a high thermal loading environment. Alumina is an anisotropic material, meaning it
possesses different material properties based on orientation. Single-crystal sapphire (basally
aligned) and equiaxed (randomly oriented) are commercial forms of alumina used as IR window
candidates. Purdue’s platelet polycrystalline aligned alumina is to be tested as a potential
replacement to these two commercial alternatives. This chapter will discuss findings in preliminary
ablation tests, a detailed experimental methodology, results of various physical and optical
properties, and potential theories as to explain the effects of crystallographic alignment on
polycrystalline alumina as an IR window candidate.

5.2 Rapid Oxyacetylene Ablation Testing

Purdue’s ablation rig can use fuel and oxidizer to apply thermal load to a sample over its
entire material surface, raising it to temperatures of up to 1800°C and subjecting it to immense
thermal gradient for an extended period. While not a direct simulation of a hypersonic environment,
it does help illustrate a thermal image to determine the efficacy of PCA as an IR window candidate.

80
5.2.1 Ablation Testing Apparatus

The ablation testing rig is shown in Figure 5.1 below. It is important to note the graphite
holder for the sample was specially designed for a 25 mm sample, allowing for room for thermal
expansion.

81
82

Figure 5.1 Schematic of Rapid Oxyacetylene Ablation Testing Apparatus

5.2.2 Ablation Testing Parameters

From Figure 5.1 above, 25 mm samples would be placed in the graphite holder. By varying
the distance from the nozzle, the sample would experience different heat flux conditions. Estimates
of the front-facing temperatures based on several trials is shown in Figure 5.1 above. Thermal
resistance testing was conducted on this apparatus. The fuel consumption rate was 10 L/min while
the oxidizer consumption rate was 12 L/min. The ablation test was done for the duration of 60
seconds upon moving a graphite holder 40 mm from the tip of the torch. This temperature was
obtained using a K-Type Thermocouple, with a maximum temperature of 1600°C. After the
heating period, the material is left to quench in air for a period of 5 min. Subsequent density
measurements were taken using the Archimedes’ method and compared to pre-ablation values.
Optical transmission was obtained using a PerkinElmer® Fourier Transform Infrared
Spectrophotometer, at a wavelength range of 1500 nm to 7500 nm (near to far-IR).

5.2.3 Sample Testing Parameters

Samples tested for this experiment included single-crystal sapphire (c-plane), acquired by
AdValue Technology, equiaxed polycrystalline alumina, acquired by McMaster-Carr, and the in-
house processed Standard and Long test cases. The test cases produced followed the
methodologies outlined in section 2, using powder batch A. Initial optical results for the samples
mentioned are shown in Figure 5.2 below (note, equiaxed alumina’s in-line transmission was < 1%
and not included). All samples were machined to 1.0 mm thickness and optically polished using
the polishing procedure in Table 2.2.

83
84

Figure 5.2 In-Line Transmission Comparison of Batch E Standard and Long PCA and Commercial Sapphire
 It should be noted that after multiple experimental improvements and trials, maximum
transparency for the in-house Standard sample was improved to 80% maximum in-line
transmission at 3.7 µm. It is evident that commercial sapphire possesses the highest and most
consistent transmission curve along the entire wavelength range from near-IR to far-IR. It is worth
restating the benefits of polycrystalline alumina are increased mechanical toughness, lower
dielectric constant, and increased corrosive resistance, as a trade-off for transparency. Achieving
high optical quality and polish in samples is important, as it is found to directly correlate with
thermal resistance, as supported by Ali et al.1. It was also noted samples of 1.0 mm thickness failed
often during ablation testing, and therefore samples were ubiquitously polished to 1.5 mm
thickness, which will influence the starting optical conditions in the ablation results presented.

5.3 Results and Discussion

5.3.1 Preliminary Findings and Hypothesis

Before testing the main batch of higher quality optical samples, the first samples used in
the ablation test were “throw-away” samples. These were not polished optically, nor were they of
the highest quality. These samples typically contained some graphite contamination that was
difficult to post-process or some other defect that prevented them from being used as either a high-
quality optical sample or mechanical testing sample. It was during a preliminary experiment with
one of these samples where a surprising result was made.
A temperature-time plot of a 60 second ablation test at 20 mm distance was recorded and
shown below in Figure 5.3.

85
 Figure 5.3 Temperature vs. Time Distribution During 60 second Ablation Test of PCA

With alumina’s melting point being approximately 2050°C, the results found in Figure 5.3
can be interpreted as follows. First, it is likely the front pyrometer reading became inaccurate after
a certain temperature threshold (likely around 1800° based on the pyrometer). However, the rear
thermocouple reading is accurate up to 1200°C. This temperature time distribution indicates that
a throw away PCA sample was able to withstand a temperature gradient of at least 1200°C.
Alumina, compared to other ceramics, is historically susceptible to cracking upon experiencing
thermal shock or resistance2. The sample before and after is shown in Figure 5.4 below.

Figure 5.4 “Throw-Away” Platelet PCA Before (left) and After (right) Ablation Test at 20 mm

86
 From Figure 5.4, the sample experienced immense thermal load, with visible ablation on
the surface of the sample and various cracks. Surprisingly, the sample neither shattered, nor did it
break during post-processing after the ablation test (which was common for many samples due to
PCA’s weak thermal shock). The preceding result leads to the following explanation for the in-
house platelet polycrystalline alumina, and why it performed well during the ablation test.
Understanding that there is anisotropy in the CTE values along certain crystallographic planes in
the in-house PCA samples, the reason the PCA didn’t break may be that the CTE values along one
of the planes is higher than a plane normal to it. This implies that during thermal expansion, there
is less localized stress between crystallographic faces because two adjacent faces will be
expanding at the same rate in a highly aligned sample, versus a randomly oriented one. Based on
this theory, single-crystal sapphire should neither experience cracking during excessive thermal
load, despite this not being the case in practice; however, a possible reason is that the presence of
the grain boundaries in the in-house polycrystalline sample regions of (disorder) act as thermal
sinks to increase the ability of our highly aligned samples to convect heat on its surface that single-
crystal sapphire will not be able. This is only possible because the in-house polycrystalline alumina
is also basally aligned. To test the hypothesis that the in-house highly aligned sample will have the
highest thermal shock resistance when compared with randomly aligned alumina (equiaxed), along
with single-crystal sapphire in the same crystallographic (basal) plane.

5.3.2 Characterization Before and After Ablation Tests

The following showcases the before and after images following a 60 second ablation test
at 40 mm distance, resulting in approximately a 1500°C front surface and at minimum a 700°C
thermal gradient across all samples. In the literature, cases of %mass loss are typically measured
for ablation test materials1,2. However, initial calculations for mass loss for many tests were either
negligible, or due to failure, were not as useful in understanding material evolution under thermal
load. Therefore, before and after in-line transmission data was taken using an FTIR, which
provides useful information as to a material’s capability to withhold its transmission value after a
period of high heat flux. Figures 5.5-5.8 showcase the visual evolution of each sample whereas
Figure 5.9 describes their transmission evolution.

87
88

Figure 5.6 (left) Commercial Equiaxed Before and After Ablation Test Figure 5.5 (right) Sapphire Before and After Ablation Test
89

Figure 5.7 (left) Basally Aligned Platelet Polycrystalline Alumina – Figure 5.8 (right) Basally Aligned Platelet Polycrystalline Alumina –
Standard Test Cases Before and After Ablation test. Long Test Cases Before and After Ablation test.
Figure 5.9 In-Line Transmission Results Before and After Ablation Test for Commercial and In-
House Alumina (1.5 mm Thickness)

Figure 5.9 describes the in-line transmission evolution of the alumina test cases before and
after 60 seconds of ablation at approximately 1500°C on the front surface. A table detailing the
%transmission lost at a relevant 3 µm is shown in the following Table 5.1.

90
Table 5.1 Comparison of Relevant Sample Properties from Ablation Testing Commercial and In-
House Alumina
Sample Ablation Time (s) Failure (Y/N) %Transmission Lost (3 µm)
Commercial Equiaxed
60 N 7.4%
PCA
Commercial Sapphire
7 Y 54.9%
(C-plane)
Standard Aligned PCA 60 N 80.9%
Long Aligned PCA 60 N 7.5%

5.3.3 Effect of Crystallographic Orientation

From Figure 5.9, several observations are apparent. Equiaxed PCA, as previously
mentioned had <1% transmission to begin with, and therefore the transmission lost value can be
considered extraneous. Commercially aligned sapphire had peak transmission at 85% at 3 µm but
failed within 7 seconds of ablation. It was the only sample to do so of the 4 test cases, most likely
due to its single crystal nature. Despite being aligned for favorable transmission properties, the
absence of grain boundaries provided no pathway for kinetic travel for the rapid increase in heat
flux, increasing the thermal gradient experienced by the sample and causing it to quickly shatter.
The equiaxed PCA, despite not being transparent at any range, did not fail during 60
seconds of testing, supporting this hypothesis, and highlighting the stark thermal performance
capabilities of polycrystalline alumina vs single crystal. It should also be noted that the equiaxed
alumina sample performed best in terms of visual performance of all samples, showing neither
signs of failure, nor residual signs of ablation on the surface, as supported in Figure 5.5. Visually
observing the Standard and Long test cases show signs of ablation on the bottom surface, and this
is because the samples were in direct contact with the graphite holder at the bottom. The direct
contact during nozzle flow would transfer heat from the solid graphite to the bottom of the sample
surface, in addition to the heat flux from the nozzle. This was present in all cases except the
commercial equiaxed sample.
The temperature dependence of the lattice parameters of alumina in the polycrystalline
samples play a critical role, due to their anisotropic thermal expansion values. Munro et al. shows
that for small grain size α-alumina (1-6µm), characteristic strength of materials is retained at high
temperatures3. It is possible the randomly oriented nature of the alumina crystals in the equiaxed
sample reduces the level of anisotropy in CTE such that it allows for more equivariant heat flow

91
across the material. The grain size being the smallest (10 µm) supports the fact that the greater
concentration of grain boundaries present lessen the thermal load experienced. Therefore, despite
being the least transmissive test case, the equiaxed alumina performed best in terms of survival in
the ablation test. This would also explain why the single-crystal sapphire performed poorly in the
environment. The primary properties of interest in terms of thermal performance would be thermal
conductivity, diffusivity, and specific heat. Munro shows that for α-alumina, in contrast to
sapphire, the specific heat increases with temperature, which makes it more favorable for high heat
flux and flight applications. Similarly

5.3.4 Effect of Sintering Parameters

The in-house platelet PCA samples are, however, vastly more interesting test cases
compared to the commercial equiaxed sample. Looking at the in-house platelet aligned alumina
cases, we observe the Standard test case had the most comparable transmission to sapphire, but
also experienced the greatest transmission loss of any sample. Figure 5.7 after ablation shows
much of the Standard surface was ablated with a degree of surface melting taking place. However,
in Figure 5.8, it can be observed that the Long test case, while displaying some degree of melting
on the bottom surface, is of lesser degree when compared to the Standard test case. The difference
in these two cases, not only being their sintering parameters, was also their grain size (as recalled
from Table 4.3). The Long test case had less than half the grain size of the Standard test case, and
more closely resembled the performance of the commercial equiaxed test case. Microstructurally
between the sintering parameters (referring to Figure 4.7), more intragranular porosity is present
in the Standard test case than the Long test case. From Hasselman, the strength behavior of
polycrystalline alumina during thermal shock tests is in accordance with expected behavior2. Based
on the ablation test results, a similar conclusion can be made regarding the behavior of aligned
PCA during thermal gradient testing. Samples with greater mechanical strength (such as the Long
test case) and smaller grain size, display favorable behavior than samples with less mechanical
strength or exaggerated grain size. Future work could confirm this explanation by testing various
samples with differing characteristic strength values.
Another possible explanation is the evidence of primary recrystallization seen in section
4.4.3 for the Long test case. The plastic deformation that occurs, likely as a result of basal slip may
contribute to maintenance of transmissive strength from ablation.

92
 The primary properties of interest in terms of thermal performance would be thermal
conductivity, diffusivity, and specific heat. Munro shows that for α-alumina, in contrast to
sapphire, the specific heat increases with temperature, which makes it more favorable for high heat
flux and flight applications3. Similarly, for the PCA test cases, it is likely that as temperature during
ablation increased, the specific heat for alumina resulted in the transient response of heat being
added to the face of the material. As for the difference between the Standard and Long test cases,
Munro also showed a temperature dependence in the strength characteristics of alumina with the
most used statistic being flexure strength, which was calculated for these test cases in section 4.
The Long test case performed the most consistently because of its smaller grain size and ubiquitous
microstructure as opposed to the Standard test case, leading to a higher characteristic strength and
less dependence on anisotropic thermal expansion as opposed to single-crystal sapphire.

5.4 Summary and Conclusions

Four independent levels of alumina were tested, allowing comparisons of the effects of
crystallographic orientation and sintering parameters. The single-crystal sapphire performed the
worst in all test cases, failing after 7 seconds, despite being most transmissive. All polycrystalline
samples survived a full minute of high heat flux, with the commercial equiaxed case performing
visually the best, albeit having almost no transmissive capabilities. The Long test case offers a
surprising compromise, little transmissive loss when compared to both the Standard test case and
single-crystal sapphire. This is a qualitatively remarkable result, as by not only aligning the platelet
grains, was the Long test case able to achieve a high level of near-IR transmission (as shown in
Figure 3.4) it was also one of the most resilient test cases to the application of immense thermal
load, comparable to that of equiaxed PCA. This leads to the belief that microstructurally preferring
smaller grain size during the development of platelet polycrystalline alumina offers significantly
improved thermal loading capabilities as a hypersonic IR window, with a potential tradeoff in
transmission. This tradeoff can be mitigated by calcination experimenting with longer dwell times
or lower isothermal temperatures.

5.5 Overall Conclusions and Future Work

The key findings unique to this study on aligned PCA are as follows:

93
 • Lower isothermal sintering temperatures successfully resulted in smaller grains.

• Calcination of samples at lower sintering temperatures resulted in fewer absorption losses.

• Evidence of primary recrystallization is present for aligned alumina sintered below 1800°C,
consistent with prior literature4,8.

• Weibull modulus value of m = 28.8 was calculated for the Standard test case, consistent
with literature, and among the highest.

• Crystallographic alignment and primary recrystallization of aligned alumina both contribute

to favorable thermal performance in a high heat flux environment as it relates to
transmissive losses.

• Processing in a single-sintering step at temperatures lower than 1800°C for more than at
least 5 hours in combination with calcination results in a sample with peak in-line
transmission of 76% in the near-IR, maximum peak stress of 234.7 MPa, small grain sizes
of 11 µm, Weibull modulus of m = 9.5, and minimal transmissive losses after 60s of ablation
at surface temperatures of 1500°C. This particular sample offers superior thermal and
mechanical performance of single-crystal sapphire with comparable transmission that also
extends into the far-IR.

Polycrystalline alumina is a material that has been researched for decades as a potential
replacement for single-crystal alumina (sapphire) as an infrared window candidate for hypersonic
flight. While the main drawback of polycrystalline alumina is birefringence scattering, resulting
in weaker in-line transmission, the process of crystallographic alignment results in favorable
optical properties, comparable to that of single-crystal sapphire. The process of grain alignment,
however, results in grain growth4. This thesis’ focus was to experiment the effects of various
sintering and powder processing approaches on grain size, in-line transmission, flexure strength,
and ability to survive in a rapid heat flux environment.
Alumina has a rhombohedral crystal structure and anisotropic material properties based on
crystallographic face. Differing refractive index values limit transmission due to forward
scattering. Porosity contributes to any refractive index mismatch that results in transmission losses.
Purifying the powders before processing, as well as achieving full densification during sintering,
are two methods of reducing these transmissive losses. The process of washing was not successful

94
in reducing overall alkali impurities in the alumina samples and resulted in poor in-line
transmission from the visible to near-IR. The process of calcination was successful in improving
the in-line transmission of all test cases used. It is recommended that future work involves
calcination of the platelet alumina powder to achieve the maximum possible in-line transmission.
Water or ethanol washing may be worth investigating but did not yield favorable results in this
case. Sintering parameters included a Standard, Long, and Two-Step method. The Two-Step
method, while yielding inconsistent results in this work, may be worth investigating uses the
different approaches outlined by Lóh et al.5. Overall, the Long test case, while having weaker in-
line transmission throughout the optical range, had superior transmission in the near and mid-IR,
which was only further improved through calcination. This shows promise in the approach in
reducing isothermal sintering temperatures and densifying over a longer dwell period.
To test mechanical strength of the Standard and Long test cases, ASTM1161a was
followed. Flexure bars required the fabrication of larger 55 mm samples, and modified die and
processing procedure. These larger samples were thicker and contained less over in-line
transmission than the conventional 25 mm samples. Flexure strength for both test cases were
calculated using 4-point flexure tests. Microstructural analysis and grain size statistics were
determined for both cases. Weibull modulus was calculated using characteristic strength values.
The results favored the Long test case, which resulted in significantly less grain growth than the
Standard test case. The most interesting result was evidence of primary recrystallization in the
Long test case, which is supported in prior literature7. This resulted in a peak stress greater than
195 MPa and a high Weibull modulus (m = 28.8), which is consistent for aligned polycrystalline
3,6
alumina according to prior work (m = 24), and high for other literature (typically m = 8-11) .
Future work could test various lower isothermal conditions below 1800°C for a variety of dwell
times.
Alumina’s crystallographic anisotropy also results in different thermal properties based on
orientation. These effects were tested in a rapid heat flux ablative environment. From these results
several conclusions can be drawn. First, alignment of alumina had a negligible effect on thermal
properties. Alignment most benefits optical quality of the sample during testing, and a highly
aligned sample is desired for the in-line transmission benefits in the IR range. However, grain
boundary had a significant impact on test results. Single-crystal sapphire performed poorly. It is
possible grain boundaries act as a thermal sink for transient increases in temperature, and this is

95
supported by Hasselman and Munro2,3. Such an explanation would account for the superior
performance of the Long test case as compared to the Standard test case during ablation testing.
Given the smaller grain size, and thus the higher density of grain boundaries, the Long test case
was least impacted by thermal CTE anisotropy and therefore experienced little change before and
after rapid oxyacetylene ablation testing. Future work could confirm these results by testing
independent levels with similar sintering conditions to the Standard and Long test case. Additional
equiaxed samples could be tested, as only one such equiaxed sample was provided for testing in
this case. Furthermore, single-crystal sapphire aligned along different crystallographic orientations
could be tested but would likely not yield fruitful results for comparison. Future work might also
test effects of recrystallization present in samples as it pertains to high heat flux test.
Overall, through calcination methods and sintering for lower isothermal temperatures and
longer dwell times, we have been able to maximize the in-line transmission of alumina, while
minimizing grain growth, improving upon existing mechanical strength, and producing promising
results in a simulated hypersonic heat flux environment.

5.6 References
1
[1] Ali, S., Ahmad, S., Baluch, A. H., Ramzan, M., Saleem, M., Bilal, M., & Naseem, K. (2022).
Microstructural and thermal analysis of alumina modified carbon/phenolic composites after
ablation testing. Journal of Composite Materials.
https://doi.org/10.1177/00219983221147389
2
[2] HASSELMAN, D. P. H. (1970). Strength Behavior of Polycrystalline Alumina Subjected to
Thermal Shock. Journal of the American Ceramic Society, 53(9), 490–495.
https://doi.org/10.1111/j.1151-2916.1970.tb15997.x

[3] 3
Munro, R. G. (1997). Evaluated material properties for a sintered α-alumina. In Journal of
the American Ceramic Society (Vol. 80, Issue 8, pp. 1919–1928). Blackwell Publishing Inc.
https://doi.org/10.1111/j.1151-2916.1997.tb03074.x
4
[4] Schlup, A. P. (2020). PROCESSING ALUMINUM OXIDE FOR THE CONTROL OF
MICROSTRUCTURAL TEXTURE AND OPTICAL PROPERTIES.
5
[5] Lóh, N. J., Simão, L., Faller, C. A., de Noni, A., & Montedo, O. R. K. (2016). A review of
two-step sintering for ceramics. In Ceramics International (Vol. 42, Issue 11, pp. 12556–
12572). https://doi.org/10.1016/j.ceramint.2016.05.065
6
[6] Schlup, A. P., Costakis, W. J., Youngblood, J. P., & Trice, R. W. (2023). Mechanical
properties of hot-pressed transparent platelet alumina. International Journal of Applied
Ceramic Technology, 20(6), 3656–3665. https://doi.org/10.1111/ijac.14462

96
 7
[7] Heuer, A. H., Sellers, D. J., & Rhodes, W. H. (1969). Hot working of aluminum oxide. 1:
Primary recrystallization and texture. J. Amer. Ceram. Soc., 52, 468–474.

97
 APPENDIX A. EVALUATING HOT-PRESSED MGO:Y2O3
NANOCOMPOSITE POWDER FOR INFRARED TRANSPARENCY

A.1 Contributions and Introduction as an Infrared Window Candidate

The persons listed contributed to the following tasks relevant to this section:
• U.S. Research Nanomaterials Inc. for supplying the high purity nano powders.
• Dr. Clara Rivero-Baleine of Lockheed Martin Corporation for her financial support of
the project.
• Dr. Venkat Kasi and Dr. Mahabubur Rahman for assistance in FTIR measurements.

Infrared (IR) window materials demand elevated levels of thermal shock resistance,
transparency in the IR, mechanical strength, and durability. In a hypersonic environment, ceramics
conventionally used for IR windows are single-phase. Ceramics like single-crystal sapphire
(alumina – Al2O3) are most widely used in IR window applications, not because of their optimal
performance, but rather because of their availability. While optical transmission of sapphire is high,
it does not extend well into the IR range and drops off after 4.5 µm at room temperature1. In this
thesis investigating alumina as an IR window, it was shown that one of the primary hindrances to
optical and IR transparency is the refractive index mismatch between adjacent grains. In recent
years, the possibilities of using nanocomposite, multi-phase ceramic materials as IR windows have
gained attraction due to their ability to reduce optical and IR scattering. This increased
transmission is accomplished using a smaller nanoscale grain size. In simple terms, it is the wave
nature of light that effectively passes through grains of size significantly smaller than its own
wavelength with little inhibition. However, while reducing grain size typically comes at the cost
of mechanical toughness, the refractive indices of some nanocomposites have been found to vary
linearly with the volume fraction of particles. Thus, at smaller grain sizes, it is possible to use a
refractory oxide compound as a secondary phase to minimize losses due to refractive index
mismatch, reducing light scattering dramatically compared to single-phase ceramics, with the
benefit of maintaining mechanical toughness. Such two-phase ceramics have potential as
candidates for IR windows for hypersonic flight.

98
 Research on infrared composite nanomaterial for window and dome applications in flight
began with nanoscale tetragonal zirconia (ZrO2) and thoria (ThO2) as early as 19812. Yttria (Y2O3)
is a material widely sought for its high transmission over a wide wavelength range spanning the
ultraviolet to far infrared, offering it flexibility in many applications. However, its weak
mechanical strength makes it a poor candidate as a window material during flight. Companies such
as Raytheon in the past several decades have experimented with adding 10% vol. of thoria (ThO2)
or lanthia (LaO2) to yttria, the goal being to use a material with similar refractive index that would
both mix homogenously, inhibit grain growth, and maximize transparency3. Seminal work since
the mid-2000s has investigated combining Y2O3 with MgO as a MgO-Y2O3 composite to
overcome the weak mechanical strength of Y2O3, as MgO has been found to effectively inhibit
grain growth by a pinning effect at grain boundaries. Yttria having a refractive index of n = 1.874
at 3 µm and magnesia having a refractive index of n = 1.691 at 3 µm make magnesia a suitable
candidate as a secondary phase for a nanoscale composite material.
Initial results in literature were promising. Ma et. al achieved 150 nm grain size and
investigated sintering parameters to achieve densification4. Harris et. al stated, “... the strength of
the nanocomposite [MgO-Y2O3] at 21°C is twice as great as strengths achieved for its
constituents...”5. Stefanik et al. also posited a maximum theoretical transmittance number of ~85%
possible1. This theoretical transmission aligns with several other papers4,5,6,7.

99
 (a) (b)
100

(c)

Figure A.1 (a,b,c) Transmission vs. Wavelength MgO-Y2O3 [(a) Stefanik et al.1 (top-left), (b) Ma et al.4 (top-right), (c) Harris et al.5 (below)]
 While a 50:50 vol.% is not ideal, given the slight differences in refractive indices of the
materials, Stefanik et al., amongst others, found the greatest potential for transmission lay in a
50:50 vol% composite material1,4,5,6. It is important to note that while at 50:50 vol% shown in
Figure A.1(a), it is possible grain growth was most inhibited, it is also at the cost of increased grain
boundary surface area. In Figure A.1(c) from Harris et al., it can also be shown that as the grain
scale decreases there is little difference in maximum transmission at longer wavelengths. The
benefits attributed to a 50:50 vol% MgO-Y2O3 at small grain scales are reflected at shorter
wavelengths in the visible and near-IR. This is because light travelling at a wavelength smaller
close in size to the grain scale of the material results in both more surface reflectance and
birefringent losses. However, many papers, apart from the abovementioned, investigated powder
morphology, particle size, mechanical strength, and optical transmittance6,7. Nanocomposite
ceramic material MgO-Y2O3 is one such material that has garnered attention as it shows promise
as an infrared window given its broad optical range, small grain size, and ability to be
microstructurally improved to better its mechanical properties. This section of the paper
investigates MgO-Y2O3 as an infrared window material in terms of optical transmittance,
discussing effects of sintering parameters and powder processing methods. The research goal being
to achieve a high level of transparency utilizing a similar, single-step hot-pressing procedure as
was done with polycrystalline alumina in the previous sections.

A.2 Experimental Procedure

It should be noted that the experimental approach for processing and densifying the
nanoscale MgO-Y2O3 is like the approach taken in this thesis for polycrystalline alumina, in section
2. With the goal being to sinter and densify the two-phase nanoscale ceramic powders, while
minimizing refractive losses due to birefringence, it is appropriate to utilize a similar approach that
led to the successful creation of transparent polycrystalline alumina. Experience and precedent in
the ceramic processing of such samples as IR windows resulted in a similar methodology, reflected
in the procedures that follow.

101
A.2.1 Powder Morphology

The infrared window ceramic tested is comprised of a 50:50 vol% MgO and Y2O3. The
powders were obtained from U.S. Nano Inc. and contain the starting diameters and purity levels
outlined in Table A.1 below.

Table A.1 Starting Properties of Nanocomposite Powders

Powder Type Powder Diameter (nm) Purity Level

1. Yttrium Oxide (Y2O3) Nanopowder 20-40 99.999%
2. Magnesium Oxide (MgO)
20 99+%
Nanopowder
3. Magnesium Oxide (MgO)
10 99.9+%
Nanopowder

High purity level powders were purchased for this study, supported in Ma et. al, as they
would contribute to higher levels of optical transmittance. The figures below show the SEM for
the starting powders, obtained with an FEI Quanta 650™:

Figure A.2 SEM Image of 50:50 vol% MgO-Y2O3 powder (wet milled and calcined)

102
 From the mixed powder morphology in Figure A.2, a few observations are apparent. The
first and foremost is the presence of heavy agglomeration, that likely resulted in uneven mixing
during the processing step. While some areas can be shown to have a finely blended mixture, many
areas seen under SEM revealed that large micron scale grains. This certainly contributed to a loss
in transparency. Referencing the figure from Harris et al., transparency can be expected to increase
as grain size decreases. The presence of this uneven grain distribution is likely the key culprit to
the significant transmission loss and general opaqueness of the samples, despite post-processing
and calcination.
Beyond powder mixing effects on transparency, the effects of calcination temperature and
resulting hardness and transmittance were also investigated by Yong et. al. Harris et. al, and Xu et.
al5,7,8. Similar calcination temperatures and ramp rates were used in this research that would
maximize hardness and transmittance.

Figure A.3 Vicker’s Hardness vs. Calcination Temperature8

From Figure A.3 above, Yong et al. illustrates the effect of calcination temperature on
measured Vicker’s hardness8. A temperature of approximately 1050°C was used, and this

103
coincides with methodologies from Harris et al. and Ma et al. (both of whom calcined at 1100°C)1,5.
Calcination temperature remained consistent in this paper.

A.2.2 Powder Processing and Heat-Treatment

The process of combining such small-scale powders to achieve homogeneity proved

challenging. Various factors influence the degree of mixing, such as the batch size, purity, and
volume of mixing media, and mixing time. Given the densities of MgO and Y2O3, a 50:50% vol
was measured for a batch of 100g, which translated to 41.68g of MgO and 58.32g of Y2O3. The
mass solution was wet mixed in a 1L Nalgene milling container along with 5% vol Yttrium
Stabilized Zirconia (YSZ) milling media and 50% vol of high purity de-ionized water (18 Ω*cm).
Note, a preliminary batch was made by dry mixing; however, resulted in severe agglomeration,
and thus wet milling was used exclusively. Therefore, the testing batch container was wet milled
for at least 24 hours. The dried batch was then decanted and sieved through a 250 µm Nylon mesh
to separate the YSZ milling media from the washed powder and baked in an oven at 50°C overnight
to evaporate any residual moisture. The batch was sieved once more to break apart any
agglomerates, and the final batch was then weighed in alumina crucibles for heat treatment.
The powder batch was calcined in air at 1100°C for 1 hour at a 10°C/min heating and
cooling ramp rate, the goal being to evaporate residual impurities and oxides. Harris et al. went
through a similar procedure of heat treatment for their MgO-Y2O3 solution, and further discussion
on effects of calcination temperature is given in the results discussion5. The calcined powder was
then sieved again to break apart any soft aggregates and placed in a graphite hot-press die with a
sample diameter of 25 mm.

A.2.3 Hot-Pressing

The die anatomy followed a similar set-up to the polycrystalline alumina discussed in
Chapter 2. An illustration of the vertical set-up is shown below:

104
 Figure A.4 Die Anatomy for MgO-Y2O3 (25 mm diameter)

Molybdenum was used as a non-reactant and separation agent between the powder bed and
graphite. Enough powder was used to result in a 3mm hot-pressed sample thickness. The die was
isostatically cold-pressed to 1.5 psi and placed in a sintering furnace at room temperature in
vacuum overnight.
Three variants of the sample were hot-pressed. The sintering parameters for each are shown
in the table below:

Table A.2 Sintering Parameters for MgO-Y2O3 (25 mm) samples

Heating/Cooling Rate
Sample Pressure (MPa) Dwell (min)
(°C/min
1 25/25 20 60
2 25/25 40 60
3 15/5 40 60

An isothermal temperature of 1300°C was shown to be used in Harris et al., Yong et al.
and Ma et al. to produce transparent samples4,5,8. Thus, the same temperature was used as a
constant for these various samples along with an inert Argon atmosphere.

105
A.2.4 Post Processing

Samples were gently taken out of the hot-press die and residual Molybdenum and graphite
was hand-polished for approximately 10-15 minutes each side using an Apex Terra Diamond Pad.
It was noted that many samples tended to fracture along the centerline while grinding or polishing,
and felt noticeably more difficult to handle, when compared to a sample like polycrystalline
alumina. However, most fractures if they occurred happened hemispherically, and so optical data
was still able to be obtained.
Following a hand polish, samples were mounted on a metal plate using thermoplastic
polymer and surface ground on each side to a thickness of about 1.5 mm. The grinding and
polishing process follows identically to the process used in section 2.1.3 for polycrystalline
alumina.
Completely polished samples were then calcined in air for 21 hours at 1100°C with a
15°C/min ramp rate, to reduce any residual oxygen vacancies. This was a practice reported in
Harris et al. that was followed in this procedure to minimize potential discrepancies5. The annealed
and fully polished samples were then tested for relative densities and optical transmission.
In-line transmission was measured using a Nicolet™ iS™ 5 FTIR Spectrophotometer
along with ThermoScientific™ OMNIC software. The results of the FTIR for all three samples are
shown in Figure A.6, whereby the thickness of the samples for FTIR testing were normalized to
0.8 mm.

A.3 Results and Discussion

Results for the three sample conditions are shown in the figures below. The relative densities of
the samples were determined via the Archimedes density technique.

106
 ρ = 98.7% ρ = 97.6% ρ = 98.4%

Figure A.5 Optical Images and Relative Densities of 50:50 vol% MgO-Y2O3 Samples (1.5 mm
thickness, 25 mm diameter)

From the images, a few observations can be made. In samples 1 and 3, fractures occurred
along centerline during post-processing. Sample 2 remained intact, but also felt noticeably weak
along centerline. It has been shown in Ma et al. that mechanical strength is strongly dependent on
residual pores and the square root of the grain size of the samples, albeit mechanical properties
were not a subject of testing for these samples4.
As for the relative densities, Xu et al. reported a density level of less than 95% when
pressurelessly sintering to 1600°C but reheating to 1700°C eliminated nearly all pores7. Harris et
al., whose procedure this closely followed achieved a relative density of >99% 5. The reasons for
discrepancies in the relative densities are likely two-fold. The first could be a result of the wet
milling process. Trace impurities from the milling media or high purity de-ionized water could
cause residual pores to remain that aren’t related to oxygen vacancies. This is supported by
observing the agglomerated powder mixture in Figure A.2. The second, and more impactful reason,
is that this methodology did not contain a second-processing step. Prior literature almost always
involves a step of hot-isostatic pressing (HIP), heating and pressurizing the sintered body under
extreme loads to achieve very high relative densities1,5. While the relative densities of the samples
shown in the table are high (with Sample 1 at ρ = 98.7%), Xu et al. and Ma et al. both show MgO-
Y2O3 samples of only greater than ρ = 99.8%, were successful4,7. To achieve a higher densification,
a two-step sintering procedure or coarsening may be possible.

107
 Another observation ubiquitous amongst all three samples is lack of optical transparency
and opaqueness in line with the sintering direction. Despite the polishing routine, and calcination
processes, samples had very low optical transparency, which is reflected by lower relative densities.

Figure A.6 FTIR Plot Transmittance of 50:50 vol% MgO-Y2O3 from 2.0 - 8.5 µm

The effects of sintering temperature on transmittance can be determined from the FTIR
results, indicating that a shallower ramping and cooling rate is preferable to a steeper ramp rate.
From these results, a few observations can also be made. The first is that the peaks that occur in
all samples appear ubiquitous, other than the peak in Sample 2 that occurs at about 6.3 µm. This
peak is shown in Ma et al., whose specimens reached full density and a peak transmission close to
the Stefanik theoretical1,4. However, this peak is distinctly muted in samples 1 and 3.
A clear distinction between our sample results, and results from others such as Harris et al.
and Ma et al. is the transmission4,5. The samples were nearly opaque, with a maximum
transparency of <1%.
There are some possible explanations for such vast discrepancies in transmission levels.
The first could be the sintering time and pressure. The sintering furnace has a limited maximum
isothermal pressure based on the tolerances of the graphite rams and graphite die. Pressures

108
ranging from 40MPa-60MPa were used by Harris et al., Ma et al., and Xu et al.4,5,7. Future testing
would benefit from constructing a higher tolerance die and testing pressure up to and exceeding
60 MPa. Applying a higher isothermal sintering pressure may help contribute to a fuller
densification of the sample and result in higher transparency.
Harris et al. shows similar transmission data to Ma et al. reaching peak transmission of
84%4,5. They show the effect grain size has on transmission, where a smaller grain size reduces
the optical scattering of light. A grain size of 25nm showed significant transmission differences in
the optical and near infrared wavelength range, and minimal improvements in the mid infrared and
far infrared range.

Figure A.7 SEM Imaging of Samples 1 (left), 2 (center), and 3 (right) (WD = 10 mm, Scale Bars
= 100 µm)

Referencing their transmission results, another explanation for such low transmission
values in our samples could be from the effects of powder processing. The SEM images for the
50:50 vol% powder batch is shown in Figure A.7 below
From Figure A.7, observations can be made about each sample. Sample 1 contained
aggregates of the powder mixture, visible by the dark and light regions in the image. Sample 2
contained a significantly more homogeneous topography. With the only processing differences
between Samples 1 and 2 being the maximum isothermal pressure, and with the same homogeneity
present in Sample 3, it can be concluded that a higher isothermal pressure is needed (more than 40
MPa) to eliminate larger aggregates. This difference in pressure is a source of much more
significant change in sample microstructure than is ramp rate, as evident by comparing images of
Samples 2 and 3. With the only difference between Samples 2 and 3 being ramp/cooling rates, not
as stark of a difference is present in the SEM. However, this difference was reflected optically
with higher in-line transmission, as recalled from Figure A.6.

109
 A.4 Summary and Future Work

50:50 vol% MgO-Y2O3 as an IR window was developed in a single-processing step,

achieving up to 98.7% densification. In-line transmission peaked in the 4.25 µm range, albeit the
samples displayed overall low transmission. The approach shows promise in achieving this level
of density, and greatest potential for improvements may lie in increasing maximum isothermal
pressure, adjusting ramp/cooling rates, and operating in a vacuum atmosphere.

A.5 References
1
[1] Stefanik, T., Gentilman, R., & Hogan, P. (2007). Nanocomposite optical ceramics for
infrared windows and domes. Window and Dome Technologies and Materials X, 6545,
65450A. https://doi.org/10.1117/12.719312

[2] 2
R. M. Cannon, T. D. Ketcham, and T. W. Coyle, “Toughened Optical Ceramics,” Proc. SPIE,
297, 173-87 (1981)
3
[3] Keyvani, A., Mostafavi, N., Bahamirian, M., Sina, H., & Rabiezadeh, A. (2020). Synthesis
and phase stability of zirconia-lanthania-ytterbia-yttria nanoparticles; a promising advanced
TBC material. Journal of Asian Ceramic Societies, 8(2), 336–344.
https://doi.org/10.1080/21870764.2020.1743419
4
[4] Ma, H. J., Jung, W. K., Baek, C., & Kim, D. K. (2017). Influence of microstructure control
on optical and mechanical properties of infrared transparent Y2O3-MgO nanocomposite.
Journal of the European Ceramic Society, 37(15), 4902–4911.
https://doi.org/10.1016/j.jeurceramsoc.2017.05.049
5
[5] Harris, D. C., Cambrea, L. R., Johnson, L. F., Seaver, R. T., Baronowski, M., Gentilman, R.,
Scott Nordahl, C., Gattuso, T., Silberstein, S., Rogan, P., Hartnett, T., Zelinski, B., Sunne, W.,
Fest, E., Howard Poisl, W., Willingham, C. B., Turri, G., Warren, C., Bass, M., … Goodrich,
S. M. (2013). Properties of an infrared-transparent MgO: Y2O3 nanocomposite. Journal of
the American Ceramic Society, 96(12), 3828–3835. https://doi.org/10.1111/jace.12589
6
[6] Safronova, N. A., Kryzhanovska, O. S., Dobrotvorska, M. v., Balabanov, A. E., Tolmachev,
V. v., Yavetskiy, R. P., Parkhomenko, S. v., Brodskii, R. Y., Baumer, V. N., Kosyanov, D.
Y., Shichalin, O. O., Papynov, E. K., & Li, J. (2020). Influence of sintering temperature on
structural and optical properties of Y2O3–MgO composite SPS ceramics. Ceramics
International, 46(5), 6537–6543. https://doi.org/10.1016/j.ceramint.2019.11.137
7
[7] Xu, S., Li, J., Li, C., Pan, Y., & Guo, J. (2015). Hot pressing of infrared-transparent Y2O3-
MgO nanocomposites using sol-gel combustion synthesized powders. Journal of the
American Ceramic Society, 98(3), 1019–1026. https://doi.org/10.1111/jace.13375

110
 8
[8] Yong, S. M., Choi, D. H., Lee, K., Ko, S. Y., & Cheong, D. I. (2018). Influence of the
calcination temperature on the optical and mechanical properties of Y2O3-MgO
nanocomposites. Archives of Metallurgy and Materials, 63(3), 1481–1484.
https://doi.org/10.24425/123834

111
APPENDIX B. INFRARED TRANSMISSION OF REFRACTORY OXIDES
AT HYPERSONIC TEMPERATURE SCALES

The purpose of this section was to understand the various methods by which in-line
transmission may be improved in polycrystalline alumina for its use as a potential hypersonic IR
window candidate. However, in a true hypersonic environment, temperatures experienced by the
IR window can be anywhere between 1200-1600°C. Preliminary work was done along with a
fellow Purdue lab partner, Abdullah Al Saad, in modifying the PerkinElmer® FTIR with a thermal
hot stage, allowing NIR/MIR/FIR transmission measurements at temperatures up to 1500°C. This
is a novel method developed, that is not currently found in the literature, to test the efficacy of
hypersonic IR ceramic windows in a more realistic temperature environment.

Figure B.1 PerkinElmer® FTIR and Hot-Stage Apparatus with Vacuum Pump and ECU

112
 Figure B.2 PerkinElmer® FTIR and Hot-Stage During Successful Preliminary Test to 1000°C

Preliminary test runs were successfully conducted up to 1000°C, as shown in Figure A.2.
The thermal stage and vacuum pump were operationally ready for testing. Some engineering
obstacles were present, however, included modifying the thermal stage holder (developed using
3D printed parts) and modifying the thermal stage set-up for different types of ceramic samples
(i.e. zirconia vs. alumina). These obstacles have been sorted out with the technical assistance of
Ryan Smith and Samantha Nania at PerkinElmer®. Future tests using the modified PerkinElmer®
FTIR and thermal-stage apparatus would produce novel results in the field of IR transparency of
hypersonic window candidates.

113
 VITA

Ashwin Sivakumar Dharmakash was born in Madras, India, and moved with his mother and father
to the United States at the age of one. He lived all along the east coast as a child, but he and his
family settled in Virginia, where his younger sister, Harsha, was born. He completed his primary
education there and attended Thomas Jefferson High School for Science in Technology from 2013-
2016. While often studying in the school’s planetarium, he grew a passion for space and
engineering. He cemented this passion at several internships at NASA Langley Research Center.
He then attended Purdue University from 2016-2020 and completed his degree in Aerospace
Engineering, also commissioning as a second lieutenant in the United States Space Force through
the ROTC program. Following his undergraduate education, Ashwin was awarded the DAWN-
ED scholarship program through the U.S. Air Force, which with the help of the Purdue Military
Research Institute, led to his Ph.D at Purdue from 2020-2024. Upon graduation, Ashwin will
complete a military assignment in Peterson Space Force Base, Colorado, where he will attend
Undergraduate Space Training (UST). Ashwin was married to his wife, Saumya, of 6 years in
December 2023 and is excited for a loving and beautiful life with her and her family. His wife
lives in Virginia, where he will be stationed upon completion of UST. When Ashwin is not working
on his Ph.D., he thoroughly enjoys learning about engines of all kinds. He is often playing around
and tinkering with his motorcycle when he’s not riding it. He enjoys weightlifting, running,
cooking, reading books, and enjoying both nature and life with his loved ones.

114