Ignition Sensitivity of Composite Energetic Materials to Electrostatic Discharge
by
Eric Collins, BSME
A Dissertation
In
Mechanical Engineering
Submitted to the Graduate Faculty
of Texas Tech University in
Partial Fulfillment of
the Requirements for
the Degree of
DOCTOR OF PHILOSOPHY
Approved
Dr. Michelle Pantoya
Chair of Committee
Dr. Jharna Chaudhuri
Dr. Stephen Ekwaro-Osire
Dr. Tim Dallas
Dr. Ed Anderson
Dominick Casadonte
Interim Dean of the Graduate School
August, 2013
Copyright 2013, Eric Collins
Texas Tech University, Eric Collins, August 2013
ACKNOWLEDGMENTS
I owe much appreciation to many supporting individuals as I have reached my
goals as a graduate student. First of all I want to thank God who has provided me
opportunities to learn and grow and blessed my family and me greatly. I thank my
wife who has endured life with me as a graduate student and I look forward to our
future and what it entails. I also wanted to thank my family including my parents,
siblings, and in-laws who have provided support from afar.
There are several people at Texas Tech University I would like to thank for
their help with various different aspects of my research including my committee
members Dr. Jharna Chaudhuri, Dr. Tim Dallas, Dr. Dtephen Ekwaro-Osire, and Dr.
Ed Anderson. I owe much gratitude to my advisor Dr. Michelle Pantoya who has
provided me with the opportunity to develop skills to enhance my career as an
Engineer. I have learned so much these last four years and it is primarily because of
her as she taught me how to perform research, write technical papers, and present
technical presentations. She is a wonderful mentor who I hope to emulate throughout
my professional career. I also thank Mike Daniels at the Idaho National Laboratory for
his advice and valuable input for the manuscripts we have published.
I want to thank my colleagues in the Combustion Lab for being great friends
and offering great professional assistance; Charles Crane, Shawn Stacy, Oliver
Mulamba, Cory Farley, Billy Clark, Eric Nixon, Jeffery Gesner, Keerti Kappagantula,
Brandon Skelton, Evan Vargas, and Kelsey Meeks. I look forward to keeping in touch
and collaborating in the future.
ii
Texas Tech University, Eric Collins, August 2013
TABLE OF CONTENTS
ACKNOWLEDGMENTS .................................................................................... ii
ABSTRACT .......................................................................................................... vi
LIST OF TABLES ............................................................................................. viii
LIST OF FIGURES ............................................................................................. ix
CHAPTER 1: INTRODUCTION ........................................................................ 1
Thermite Combustion ....................................................................................... 2
Size distribution effects on combustion of Aluminum Particles ....................... 4
Oxidation of Aluminum particles ..................................................................... 6
Electrostatic Discharge...................................................................................... 6
CHAPTER 2: IGNITION SENSITIVITY AND ELECTRICAL
CONDUCTIVITY OF A COMPOSITE ENERGETIC MATERIAL WITH
CONDUCTIVE NANOFILLERS ....................................................................... 9
Abstract ............................................................................................................. 9
Introduction ....................................................................................................... 9
Experimental ................................................................................................... 11
Materials .................................................................................................... 11
Mixing Procedure ...................................................................................... 12
Adding CNTs and GNPs to CEM ............................................................. 12
Short Sonication Mixing Procedure .......................................................... 13
Long Sonication Mixing Procedure .......................................................... 13
Dry Mixing Procedure............................................................................... 14
Conductivity Measurement Setup ............................................................. 14
ESD Sensitivity Test ................................................................................. 16
Results and Discussion.................................................................................... 17
SEM images of the nanofiller dispersions ................................................ 17
Electrical Conductivity of Al + PTFE with GNP and CNT nanofillers .... 18
ESD Sensitivity Test ................................................................................. 22
Conclusion ...................................................................................................... 24
CHAPTER 3: PIEZOELECTRIC IGNITION OF NANOCOMPOSITE
ENERGETIC MATERIALS.............................................................................. 25
Abstract ........................................................................................................... 25
iii
Texas Tech University, Eric Collins, August 2013
Introduction ..................................................................................................... 25
Experimental ................................................................................................... 27
Results and Discussion.................................................................................... 31
Conclusion ...................................................................................................... 37
CHAPTER 4: SYNTHESIZING ALUMINUM PARTICLES TOWARDS
CONTROLLING ELECTROSTATIC DISCHARGE IGNITION SENSITIVITY
............................................................................................................................... 39
Abstract ........................................................................................................... 39
Introduction ..................................................................................................... 39
Experimental ................................................................................................... 42
Oxidation of Aluminum powders ............................................................. 42
Mixing powders ........................................................................................ 45
Test Setup .................................................................................................. 45
Results ............................................................................................................. 47
Ignition Delay............................................................................................ 47
Joule Heating of an Aluminum Particle .................................................... 50
Discussion ....................................................................................................... 53
Conclusion ...................................................................................................... 54
CHAPTER 5: CONCLUSION ........................................................................... 55
CHAPTER 6: COMPARISON OF ENGINEERED NANOCOATINGS ON THE
COMBUSTION OF ALUMINUM AND COPPER OXIDE NANOTHERMITES
............................................................................................................................... 58
Abstract ........................................................................................................... 58
Introduction ..................................................................................................... 59
Experimental ................................................................................................... 60
Materials .................................................................................................... 60
Pellets ........................................................................................................ 60
Nano-coating Tool .................................................................................... 62
Witness Samples ....................................................................................... 72
Combustion Test Setup ............................................................................. 72
Contact Angle Goniometer ....................................................................... 73
Spectrometry ............................................................................................. 74
Summary of Measurements ...................................................................... 74
Results ............................................................................................................. 75
Combustion Tests ...................................................................................... 75
iv
Texas Tech University, Eric Collins, August 2013
Witness Sample Surface Characterization ................................................ 79
Discussion ....................................................................................................... 82
Conclusion ...................................................................................................... 84
FUTURE WORK ................................................................................................ 86
REFERENCES .................................................................................................... 88
v
Texas Tech University, Eric Collins, August 2013
ABSTRACT
The Safe handling of powdered composite energetic materials requires an
understanding of their response to electrostatic ignition stimuli. It is important to focus
on and minimize characteristics of the reaction that promote sensitivity to ESD while
still maintaining high performance in combustion. This body of work investigates the
ignition response of energetic materials to ESD by examining various parameters
including packing density, electrical conductivity, and ignition delay.
Inter-particle connectivity plays a major role in the ignition sensitivity of
composite energetic materials when using electric ignition. Results show that the
ignition delay times are dependent on the powder bulk density with an optimum bulk
density of 50%. The packing fractions for particle geometries and electrical
conductivity were analyzed and assist in explaining the ignition behavior as a function
of bulk density.
The Al+PTFE composition has a low electrical conductivity and is not ignition
sensitive to EDS. Small concentrations of carbon nanotubes (CNTs) and graphene
nanoplatelets (GNPs) were added to Al+PTFE, which significantly increased the
electrical conductivity to approximately 100 S/cm with only 4 vol. % of GNPs and 1
vol. % of CNTs. ESD ignition was achieved only for a discrete range of conductivity
corresponding to approximately 2x10-3 S/cm.
The ESD ignition sensitivity of nano-scale Al particles, synthesized with
varying shell thicknesses ranging between 2.7 and 8.3 nm, and MoO3 was observed in
terms of ignition delay times. It was discovered that the ignition delay increased as the
alumina shell thickness increased. These results correlate with the resistivity of the
vi
Texas Tech University, Eric Collins, August 2013
sample which also increases as the alumina content increases. A model was developed
using COMSOL Multiphysics for a single Al particle and its initiation through joule
heating. The ignition delay in the model was consistent with the experimental results
suggesting that joule heating is a major contributor to the ESD ignition mechanism.
vii
Texas Tech University, Eric Collins, August 2013
LIST OF TABLES
1. Percent of nanofiller and mass for CNTs, GNPs, and 2 vol. % of
GNP/CNT combination................................................................. 13
2. Thickness of Al2O3 shell and weight percent active Al content for all
oxidized Al particles. .................................................................... 44
3. Coating types used to coat pellets ..................................................................... 63
4. Details of the Al2O3 ALD process. Chamber temperature was
maintained at 45°C. ....................................................................... 70
5. Details of the SiO2 ALD process. Chamber temperature was
maintained at 50°C. ....................................................................... 71
7. Average energy (kJ/kg) released from pellets with their respective
submersion time and coating type. ................................................ 77
viii
Texas Tech University, Eric Collins, August 2013
LIST OF FIGURES
1. Percent of active aluminum content for particles ranging from 10 nm
to 1 m. ........................................................................................... 5
2. Test setup to measure electrical conductivity of powder using the
HRLC meter. ................................................................................. 15
3. Ignition tests with electrostatic discharge apparatus. ........................................ 16
4. SEM images of CNTs in CEM for the a) short sonication mixing
method; b) long sonication mixing method; c) dry
mixing method.. ............................................................................ 17
5. Electrical conductivity of Al + PTFE with added CNT. ................................... 18
6. Electrical conductivity of Al + PTFE with added GNP. ................................... 19
7. Electrical conductivity of Al + PTFE with 1 vol. % CNT/GNP ratio. ............. 21
8. Electrical conductivity of Al + PTFE with 2 vol. % CNTs/GNPs
ratio ............................................................................................... 22
9. Electrical conductivity that promotes ignition as depicted in shaded
region. Ignition was achieved for data points marked
with an X. ...................................................................................... 23
10. a) Piezocrystal drop test setup for ignition of Al-MoO3 powder. b)
Electrical circuit. ........................................................................... 28
11. a) Voltage and current as spark from PZT bridged 0.5 cm air gap. b)
Voltage and current as spark from PZT bridged gap
filled with powder. c) Time delay between spark and
luminosity from photo diode. ........................................................ 30
12. a) Ignition time delay as a function of bulk density expressed in
terms of percent of theoretical maximum density (% of
TMD). b) Ignition voltage or dielectric breakdown
voltage resulting in ignition as a function of bulk
density, also expressed in terms of % of TMD ............................. 32
13. Electrical conductivity of Al+MoO3 pellets with bulk densities of
20%, 40%, 50%, and 60% of TMD. ............................................. 35
14. Neytech Qex oven used to oxidize Al particles. ............................................. 43
15. TEM images of the alumina shell after oxidation times of a) 150
min, b) 8 min, and c) 30 min ......................................................... 44
16. Electrodes in contact with the pellet in the acrylic channel ............................ 46
17. Electrical circuit used to monitor voltage and current and measure
time delay of the reaction .............................................................. 47
18. Voltage, current, and luminosity depicting time delay for Al+MoO3............. 48
ix
Texas Tech University, Eric Collins, August 2013
19. Ignition time delay for Al+MoO3 reactions with varying alumina
shell thicknesses ............................................................................ 49
20. Ignition voltage for Al+MoO3 reactions with varying alumina shell
thicknesses. ................................................................................... 49
21. Temperature of Al particle from joule heating COMSOL model ................... 51
22. Predicted ignition delay for a single Al particle with varying
alumina shell thicknesses .............................................................. 52
23. Electrical conductivity of Al+MoO3 powders with a varying Al2O3
shell thickness. .............................................................................. 53
24. Pellet die assembly used to press pellets from thermite powder. .................... 61
25. Schematic diagram of the nano-coating tool. P1, P2, etc are
Precursor chemicals. Ultra high purity nitrogen gas is
used to purge the chamber and gas stick lines .............................. 62
26. Illustration of the a) hydrolysis reaction and b) hydrogen bonding
steps in applying the F-SAM layer on the sample
substrate......................................................................................... 65
27. Illustration of an ideal Per-fluorooctyltrichlorosilane, F-SAM layer
on the sample substrate. ................................................................ 66
28. Recipe flow diagram for the deposition of a super-hydrophobic
coating. .......................................................................................... 67
29. Combustion of the thermite pellets test setup used to characterize
energy generation from submerged reactions. .............................. 73
30. Example images from contact angle measurements. A micro waterdrop on (a) hydrophobic sample surface and (b) superhydrophobic sample surface. ......................................................... 74
31. Still frame images of underwater combustion for Al-Cuo with
coating Type 1 after (a) 5ms, (b) 10ms, (c) 15ms, (d)
20ms, (e) 25ms, and (f) 30ms. ...................................................... 76
32. Energy released from coated thermite pellets for all coating types
with respect to submersion time.................................................... 76
33. FTIR spectra highlighting the peaks for bonds present for Al2O3
NP + FSAM coating. ..................................................................... 78
34. Absorbance trends from FTIR for major bonds in Type 6 coating
with respect to submersion time.................................................... 79
35. SEM images of nano-particle based coating (ALD–
Al2O3+NP+FSAM) on witness samples after; A) 0 day,
B) 5 day, and C) 10 day underwater submersion time. ................. 81
x
Texas Tech University, Eric Collins, August 2013
36. Graph showing the measured contact angle results against
submersion days. ........................................................................... 82
xi
Texas Tech University, Eric Collins, August 2013
CHAPTER 1:
INTRODUCTION
Composite energetic materials (CEMs) are mixtures of solid fuel and oxidizer
particles also known as metastable intermolecular composites or thermites. Some
examples of fuel particles include magnesium (Mg), aluminum (Al), boron (B), and
titanium (Ti). Oxidizer particles include metal oxides such as iron oxide (Fe2O3),
molybdenum trioxide (MoO3), copper oxide (CuO), and bismuth trioxide (Bi2O3) and
include fluorocarbons like polytetrafluoroethylene (C2F4). Upon ignition of these
materials an exothermic reaction ensues, an event that produces lots of heat and
energy very quickly.
The reaction of thermites was first patented in 1895 by Hans Goldschmidt, a
German chemist who discovered the application of thermite welding (Goldschmidt,
1895). The use of thermites and specifically Al+Fe2O3 as a weld material for welding
railroad rails has been applied ever since its discovery and has been heavily researched
and improved upon (Chen, Lawrence, Barkan, & Dantzig, 2006; Schroeder & Poirier,
1984; Yuan, Zhan, Jin, & Chen, 2010). Other applications for CEMs which require
rapid heating include alloying (L. L. Wang, Munir, & Maximov, 1993) and material
synthesis (Deevi, Sikka, & Swindeman, 1997; Shigeta & Watanabe, 2010; Yeh &
Huang, 2011). A positive aspect of CEMs is their ability to be tailored for specific
applications. For example, if either high gas generation or a specific flame temperature
is desired, the materials can be selected for the desired application.
1
Texas Tech University, Eric Collins, August 2013
Thermite Combustion
In 1908, soon after the discovery of thermites, Goldschmidt observed self
propagation of the thermite reaction and an oxidation reduction of the metal oxide
component into a pure liquid metal (Goldschmidt, 1908). This exothermic oxidation
reduction reaction can be written in a general form as
(1)
where A and B are metals, AO and BO are corresponding metal oxides, and
 is the heat of combustion defined as the chemical energy generated from the
reaction released as heat. This reaction is called oxidation reduction because the metal
oxide reactant (BO) loses its oxide and creates a pure metal product (B) and the metal
reactant (A) forms a metal oxide product (AO). Common examples of stoichiometric
thermite reactions are 2Al+Fe2O32Fe+Al2O3 and 2Al+3CuO3Cu+Al2O3 which
produce heats of combustion of 3955 and 4076 kJ/kg, respectively; and, exhibit
adiabatic flame temperatures of 3135 and 2843 K, respectively (Fischer & Grubelich,
1998). The heat of combustion is equal to the heat of reaction numerically, but is
positive whereas the heat of reaction is a negative value (Turns, 1996). The heat of
reaction at standard state, Hr at T0, can be calculated by subtracting the sum of the
standard heat of formation of the reactants from the sum of the standard heat of
formation of the products as Shown in Eq. 2 and Eq. 3.
(2)
2
Texas Tech University, Eric Collins, August 2013
(3)
In Eq. 2, ’’ is the moles of the product and ’ is the moles of the reactant for
each substance in the reaction (i). In Eq. 3,
standard state and
is the chemical enthalpy at the
is the sensible enthalpy at standard state (Kuo, 1986).
Thermal equilibrium codes such as REAL and CHEETAH have been developed to
approximate the heat of reaction as well as other characteristics of the reaction such as
adiabatic flame temperature and gas production to assist in material selection for
desired applications. Documents have been created which include these characteristics
for many thermite reactions based off of thermal equilibrium assumptions and are
tabulated in Fischer and Grubelich (Fischer & Grubelich, 1998). The tabulated heat of
combustion for thermite reactions provides an upper estimate and enables a
comparison between reactions; but, in practice, heat losses reduce the actual heat of
combustion and further reduce the energy available from a reaction impingent on a
substrate. For example, in a study by Crane et al. (Crane, Pantoya, & Dunn, 2010), a
thermite was partially embedded into a steel block and upon ignition the thermal
distribution of the steel block was measured using high speed infrared diagnostics.
They found that 11% of the theoretical heat of combustion for the Al + Fe2O3 reaction
was conducted into the steel and attributed significant loss to convection to the
environment (Crane et al., 2010).
Energetic materials have the potential to release a large amount of chemical
energy upon ignition and can be classified as composites or monomolecular
explosives. Propellants and pyrotechnics are composite materials that contain a fuel
and oxidizer discretely separated such that their reaction is diffusion limited and
3
Texas Tech University, Eric Collins, August 2013
therefore primarily deflagrate. Monomolecular explosives on the other hand, produce
more power because their bond-breaking reactions are kinetically controlled and
therefore they often detonate.
Size distribution effects on combustion of Aluminum Particles
While many different fuels have been researched such as B (Young, Sullivan,
Zachariah, & Yu, 2009) and Mg (Gorbunov & Shidlovskii, 1986), Al is predominantly
the fuel of choice because of its low cost, high energy density, and reactivity. Al
particle sizes are commonly found to be in the micrometer (m) range but recently
have decreased to nanometer (nm) (1 nm is equal to 1x10-9 meter). Manufacturing
techniques have been developed to produce particles in the nanometric regime. One
method that is becoming popular for nanometric particle production is the wire
explosion process. Sindhu and Sarathi superheated aluminum wires such that the
material was evaporated (Sarathi, Sindhu, & Chakravarthy, 2007). They also
discovered that smaller dimensions of particles can be produced by injecting energy
above the vaporization energy of the exploding conductor (Sindhu, Sarathi, &
Chakravarthy, 2008). As the Al particle size decreases, the percent of active Al
content also decreases with the assumption the Al2O3 shell thickness is constant (Fig.
1).
4
Texas Tech University, Eric Collins, August 2013
1
% Al Content
0.8
0.6
0.4
0.2
0
0
200
400
600
Diameter of Particle (nm)
800
1000
Figure 1. Percent of active aluminum content for particles ranging from 10 nm to 1
m.
The combustion of nanoscale Al and oxidizer particles has been researched and
shown to behave completely different than their micron scale counterparts primarily
due to increased surface contact between fuel and oxidizer particles. On the
nanometric scale, Al has been shown to melt at a lower temperature (J. Sun & Simon,
2007; Juan Sun, Pantoya, & Simon, 2006; Ward, Trunov, Schoenitz, & Dreizin, 2006),
the Al2O3 shell also melts at lower temperatures (Puri & Yang, 2010), higher flame
speeds are recorded (Bockmon, Pantoya, Son, Asay, & Mang, 2005; Spitzer, Comet,
& Baras, 2010), and the reaction is more sensitive to ignition requiring lower
activation energy (J. J. Granier & Pantoya, 2011; Hunt & Pantoya, 2005) than micron
size Al. The improved combustion performance and reaction rate for nanoscale Al and
oxidizer particles is primarily due to an increase in particle to particle contact of the
reactants (Shimizu & Saitou, 1990). In a study of oxidation behavior of Al
nanoparticles, Aumann et al. attributed the fast reaction rates of Al nanoparticles to be
due to reduced diffusion distance between individual reactant species (Aumann,
1995).
5
Texas Tech University, Eric Collins, August 2013
Oxidation of Aluminum particles
Aluminum naturally oxidizes in air, forming a passivated amorphous
aluminum oxide (Al2O3) layer around the spherical Al particle. The thickness of the
Al2O3 shell is on average about 2.7 nm (Gesner, Pantoya, & Levitas, 2012) and
continues to increase if left in an oxygen environment. When placed in an oxygen
environment at high temperatures, the oxidation of Al particles is accelerated and
thickness of the Al2O3 shell continues to increase (Gesner et al., 2012; Shih & Liu,
2006). The Al2O3 shell plays a major role in the ignition and combustion of Al
particles. Rai et al. observed from experimental approaches that the aluminum phase
change causes rupture of the Al2O3 shell, and is the initiator of aluminum particle
ignition. As the Al core reaches its melting temperature, the change in density between
its solid and liquid states causes rupture in the Al2O3 shell which results in ignition
due to the exposure of molten Al with the oxidizer (Rai, Lee, Park, & Zachariah,
2004).
Electrostatic Discharge
Sources for ignition of energetic materials include lasers, hot wire, impact, and
electric spark. The combustion characteristics of energetic materials changes
depending on the ignition source because of changes in heating rates and ignition
energy. An electric spark can be created from electrostatic discharge (ESD), the rapid
exchange of charge between two objects. A common cause of ESD is static electricity,
which can be generated by contact and separation of two different materials and
friction between two different materials. ESD occurs in the form of a spark during the
breakdown of the insulating properties of air as the electric field exceeds the dielectric
6
Texas Tech University, Eric Collins, August 2013
strength (Serway & Jewett, 2004). Accidental ignition of an energetic material can
occur from ESD and result in serious injury and damage. It is important to understand
the response of energetic materials to ESD to prevent serious injury.
In the early 1940's, Brown et al. from the Bureau of Mines began testing the
response of various energetic materials including explosives to ESD (Brown, Kusler,
& Gibson, 1953). They developed a capacitor discharge unit to represent the
electrostatic charge energy accumulation on a human body and discovered that small
particles were easier to ignite. They also discovered that the ignitability and ignition
energy varied depending on the level of confinement (i.e. metal powders were more
sensitive when unconfined and black powder was more sensitive when confined)
(Brown et al., 1953).
Recently, Beloni et al. conducted ignition experiments from ESD for Mg
powders and found that the energy from joule heating was about 50-90% of the total
measured spark energy for experiments without a binder. For experiments with a
binder however resulted in a higher joule heating energy than the total measured spark
energy indicating that the powder resistivity measurements were inaccurate (Beloni &
Dreizin, 2009). They also ignited Al and Ti powders but since the powders were
ignited from an open tray, both product and reactant particles were ejected due to the
forces of the spark resulting in incomplete combustion (Beloni & Dreizin, 2010,
2011).
In understanding the response of energetic materials to ESD, it is important to
focus on and minimize characteristics of the reaction that promote sensitivity to ESD
while still maintaining efficiencies in combustion. Some parameters that could affect
7
Texas Tech University, Eric Collins, August 2013
ESD ignition sensitivity are electrical conductivity, ignition energy and delay, and
particle morphology. This body of work investigates the ignition response of energetic
materials to ESD by examining various parameters including packing density,
electrical conductivity, and ignition delay.
8
Texas Tech University, Eric Collins, August 2013
CHAPTER 2:
IGNITION SENSITIVITY AND ELECTRICAL CONDUCTIVITY
OF A COMPOSITE ENERGETIC MATERIAL WITH
CONDUCTIVE NANOFILLERS
Abstract
The safe handling of powdered composite energetic materials requires an
understanding of their response to electrostatic ignition stimuli. A binary composite
comprised of Al and polytetrafluoroethylene (PTFE) was tailored with varied
concentrations of highly conductive nanofillers. The goal was to control the ESD
ignition response of the Al+PTFE with small concentrations of nanofillers that may
not significantly affect the overall combustion performance of the mixture. The
nanofillers examined include carbon nanotubes (CNTs) and graphene nanoplatelets
(GNPs). Adding CNTs created percolation at a lower volumetric percentage than
GNPs and were found to be the controlling nanofiller, creating percolation for the
mixture containing both CNTs and GNPs. ESD ignition was achieved only for adding
nanofillers at a volumetric percentage and mixing method that led to a bulk
conductivity of approximately 5x10-3 S/cm.
Introduction
A composite energetic material (CEM) is a mixture of solid fuel and oxidizer
particles at a reactive combination that is highly exothermic upon ignition. This class
of energetic materials enables tailoring reactants toward specific applications, unlike
explosives whose reactivity is kinetically limited by the monomolecular crystal
structure. Aluminum (Al) is a common fuel, and examples of oxidizers include metal
oxides, other metals, or fluoropolymers such as polytetrafluoroethylene (PTFE). In
9
Texas Tech University, Eric Collins, August 2013
fact, the use of PTFE as an oxidizer for reactions has been studied since the mid1950’s and found use in applications including flares, tracers, igniters, and propellants
(Koch, 2002). Kuwahara et al. compared the theoretical flame temperature of
aluminum, magnesium, boron, and titanium mixed with PTFE and discovered that the
composition with Al as the fuel produced a higher flame temperature (3764 K) than
any of the other fuels (Kuwahara, 2004). Densmore et al. found similar results in
measuring the temperature of the Al + PTFE reaction to reach as high as 3650 K
(Densmore, Biss, Homan, & McNesby, 2012).
When using CEMs in an application, their safe handling requires an
understanding of their response to ignition stimuli. Powders are particularly prone to
ignition from electrostatic energy. Weir et al. defined an ESD ignition sensitivity
threshold of 100 mJ such that mixtures ignitable under 100 mJ are deemed ESD
sensitive (C. M. Weir, Pantoya, & Daniels, 2013; C. Weir et al., 2012). They also
observed a correlation between electrical conductivity and ESD sensitivity;
specifically, compositions with a higher electrical conductivity such as Al + copper
oxide were ESD ignition sensitive and compositions with a lower electrical
conductivity such as Al + PTFE were not ESD ignition sensitive (C. Weir et al.,
2012).
The goal of this study was to understand how a CEM’s ESD ignition
sensitivity may be controlled by a nanofiller. The ideal nanofiller would be inert
relative to the CEM reaction and only a small percentage of the total mixture such that
the overall combustion performance is not significantly altered. The objective was to
increase the electrical conductivity of a CEM by using conductive nanofillers and
10
Texas Tech University, Eric Collins, August 2013
locate a percolation threshold where ESD ignition may be observed. The CEM used in
this experiment was a mixture of Al and PTFE powders, a composition which has a
low electrical conductivity and is not sensitive to electrostatic discharge (ESD) when
particle sizes are in the micrometer range (C. Weir et al., 2012). The electrically
conductive nanofillers were carbon nanotubes (CNTs) and graphene nanoplatelets
(GNPs). These nanofillers have high electrical conductivity and are anticipated to
affect the ESD ignition sensitivity of Al-PTFE. This objective was accomplished by
mixing the formulations using various mixing procedures to optimize dispersion of the
conductive nanofiller in the reactant matrix. Samples were imaged to observe
dispersion quality, and electrical conductivity was measured using established
techniques. Although Al + PTFE is not ESD ignition sensitive when the Al particles
have an average diameter in the micrometer regime (C. M. Weir et al., 2013), ignition
was achieved for select samples that achieved percolation with CNTs.
Experimental
Materials
Aluminum (Al) powder with particle sizes of 3-4.5 m was used as a fuel and
polytetraflouroethylene (PTFE) powder with an average particle diameter of 35 m
was used as the oxidizer; both purchased from Alpha Aesar. Multi-walled carbon
nanotubes (CNTs) and graphene nanoparticles (GNPs) were used as nanofillers to the
fuel and oxidizer and were purchased from Alpha Aesar and Graphene Supermarkets,
respectively. The size of the CNTs and GNPs were provided by the manufacturer
where the CNTs have an outer diameter of 3-20 nm, an inner diameter of 1-3 nm, and
11
Texas Tech University, Eric Collins, August 2013
a length of 0.1-10 m. The GNP flakes have a thickness of 8 nm with a length of 0.153.0 m.
Mixing Procedure
A stoichiometric equivalence ratio was prepared for each test based on fuel and
oxidizer particles. Once proportioned, hexane was added to the Al and PTFE powders
and sonicated. The hexane and solution was then poured into a Pyrex dish to evaporate
the hexane in a fume hood and leave behind the CEM.
Adding CNTs and GNPs to CEM
The CNTs, GNPs, and combinations of CNTs and GNPs (nanomaterials) were
mixed into the CEM as nanofillers. The masses of the nanofillers were determined by
first looking at the space or volume the CEM filled and calculating the bulk density or
fraction of the theoretical maximum density (TMD) of the CEM as shown in Eq. 4.
(4)
In Eq. 4, M is the mass fraction within the mixture and  is the density of the material.
The index of summation (i) represents each material in the mixture and the upper
bound of summation (N) is the total number of materials contained in the mixture. The
TMD for the Al+PTFE CEM in this study was calculated to be 1.73 g/cm3. The mass
of the nanofiller was calculated using a volumetric percentage of the CEM and the
density of the nanofiller. The percent of the nanofillers along with their respective
masses are listed in Table 1. Three different mixing methods were used to optimize
dispersion of the nanofillers in the CEM.
12
Texas Tech University, Eric Collins, August 2013
Table 1. Percent of nanofiller and mass for CNTs, GNPs, and 2 vol. % of GNP/CNT
combination.
CNT
Vol. %
Added
0.2%
0.5%
1.0%
2.0%
Mass
(mg)
1. 8
4.5
8.9
17.9
GNP
Vol. %
Added
0.5%
1.0%
2.0%
3.0%
4.0%
Mass
(mg)
5.6
11.2
22.4
33.5
44.7
Ratio of
GNP/CNT
0/100
20/80
40/60
60/40
80/20
100/0
GNP/CNT
Mass
Mass CNT
GNP (mg)
(mg)
(mg)
0
17.9
4.5
14.3
8.9
10.7
13.4
7.2
17.9
3.6
22.4
0
Short Sonication Mixing Procedure
An aqueous dispersant for multi-walled CNTs, provided by Alfa Aesar (no.
44276), was used to better disperse the nanofillers. The nanofillers were added to
solutions that consisted of 0.075 mL of the aqueous dispersant and 25 mL of water
which was then sonicated for 1 min forming nanofiller dispersions. The CEM was
mixed with isopropyl alcohol and added to the dispersions and again sonicated for 1
min. After sonication, the solvents were evaporated off, leaving a dry mixture of CEM
and nanofiller.
Long Sonication Mixing Procedure
The nanofillers were sonicated in distilled water for 30 min which allowed for
a complete dispersion in the solvent (i.e. no settling of nanofiller was visible in
solution). The dispersed solution was then sonicated for 1 minute with the CEM and
the solvent evaporated. During evaporation, the CEM settled on the bottom of the
solution and separated itself from the dispersed nanofiller, which settled on top of the
13
Texas Tech University, Eric Collins, August 2013
CEM. The dry powders were dry mixed as they were collected and placed in a storage
container.
Dry Mixing Procedure
A slurry was prepared by mixing the CNTs in water and immediately placed in
a freezer. The frozen solution was then freeze dried; a procedure where the water in a
solid or frozen state is directly evaporated into a gas. The freeze dried CNTs were dry
mixed with the CEM using a vortex mixer until the CNTs were no longer visible in the
powder.
Conductivity Measurement Setup
A two point probing method was used to measure the electrical conductivity of
the CEM powder (C. M. Weir et al., 2013; C. Weir et al., 2012). The test setup
consists of an acrylic plate with a channel, two copper electrodes, a conductive shield
container, and a high resistance low conductance (HRLC) HR2 meter from Alpha
Labs as seen in Figure 2. The powder was loaded in the cylindrical channel with a
bulk density of 35% of the TMD. The copper electrodes were used to plug the channel
and were in contact with the upper and lower surfaces of the powder.
14
Texas Tech University, Eric Collins, August 2013
Figure 2. Test setup to measure electrical conductivity of powder using the HRLC
meter.
The HRLC meter was then connected to the electrodes and the acrylic channel
loaded with powder was placed in a conductive shield box covered in aluminum foil to
avoid interference with charges on surrounding surfaces. The conductive shield box
was connected to a high impedance amplifier on the HRLC meter to hold it at ground
potential which increases the sensitivity of the measurement.
The resistance range of the HRLC meter measures from 1.0  to 2.0 T. The
HRLC meter passes current thru the sample at voltages less than 2.0 V. The current
that passes though the sample decreases by a factor of ten for each of the nine
resistance settings, ranging from 0.1 mA to1.0 pA on the 20.0 K and the 1999.9 G
settings respectively. When measuring conductance in pS, one volt is applied across
the sample.
15
Texas Tech University, Eric Collins, August 2013
ESD Sensitivity Test
The powders were tested for ignition sensitivity from an electrostatic discharge
(ESD) using an apparatus built by Franklin Applied Physics (C. Weir et al., 2012). It
is a human body model which transfers an electric charge to another object. The ESD
tester has a variable voltage output ranging from 1 to 10 kV and charges a 0.002 μF
capacitor. The stored electrical energy is discharged through a resistive network and
from an electrode pin into the sample. The samples had a bulk density of 35% of the
TMD which was the same density as was used in the conductivity measurements. The
sample was placed on the sample holder disk and the capacitor was lowered towards
the pellet to discharge its electric energy as seen in Figure 3. This test has a “go/no go”
result indicating ignition or no ignition of the sample.
Figure 3. Ignition tests with electrostatic discharge apparatus.
16
Texas Tech University, Eric Collins, August 2013
Results and Discussion
SEM images of the nanofiller dispersions
A scanning electron microscope (SEM) was used to image the samples and
determine nanofiller dispersion quality, as seen in Fig. 4. Mixing with short sonication
(Fig. 4a) created a dispersed mixture with Al particles in contact with the larger PTFE
particles and CNTs scattered throughout the sample, building a conductive network
between particles. Mixing with long sonication (Fig. 4b) also provides a good
dispersion of nanofiller but results in more agglomeration due to the separation of
nanofiller and CEM during mixing. The dry mixing procedure resulted in
agglomeration of nanotube structures throughout the sample such as the representative
CNT cluster seen in Fig. 4c. The SEM images show that the short sonication mixing
procedure provides the best dispersion of CNTs and the dry mixing procedure results
in clumps of aggregated CNTs represented in Fig. 4c.
Figure 4. SEM images of CNTs in CEM for the a) short sonication mixing method; b)
long sonication mixing method; c) dry mixing method. Dispersion of CNTs is
achieved through sonication and aggregation of CNTs occurs when dry mixed.
17
Texas Tech University, Eric Collins, August 2013
Electrical Conductivity of Al + PTFE with GNP and CNT nanofillers
CNTs were added to Al + PTFE ranging from 0.2 to 2.0 vol. %. The electrical
conductivity baseline for Al + PTFE with no nanofillers was measured to be 1x10-7
S/cm. Figure 5 shows the conductivity as a function of CNT concentration and was
consistent for the mixing procedures involving sonication (short and long) but
different for the dry mixing procedure. For the sonicated mixing procedures, the
conductivity between 0.5 and 1.0 vol. % of added CNTs significantly increased by 6.5
orders of magnitude; and, for the dry mix procedure, the jump in conductivity
occurred between 1.0 and 1.25 vol. %. The conductivity of the dry mixed powder
behaved differently than the sonicated powders in that an electrical conductivity
plateau was observed around 2.5x10-3 S/cm and before reaching the maximum
Conductivity in S/cm
conductivity (above 100 S/cm) (Fig. 5).
1.E+02
1.E+00
1.E-02
Difference of
~7 orders of
Magnitude
1.E-04
1.E-06
1.E-08
0.0%
Sonicated Mixtures
Dry Mix
2.0%
Vol. % CNT
4.0%
Figure 5. Electrical conductivity of Al + PTFE with added CNT.
Volumetric percentages of GNP nanofiller ranged from 0.5 to 4.0 vol. % of the
Al + PTFE. Figure 6 shows that the electrical conductivity begins to increase with 2
vol. % added GNPs and then increases exponentially by 7 orders of magnitude with
18
Texas Tech University, Eric Collins, August 2013
only 4 vol. % GNPs. Since the dispersion quality was the best for the short sonicated
Conductivity (S/cm)
mixing procedure with the CNTs, it was the only mixing procedure used for GNPs.
1.E+02
1.E+00
1.E-02
Difference of
~7 orders of
Magnitude
1.E-04
1.E-06
1.E-08
0.0%
2.0%
4.0%
Vol. % GNP
Figure 6. Electrical conductivity of Al + PTFE with added GNP.
The Al + PTFE with CNTs experienced a jump in electrical conductivity at
lower percentages (between 0.5 and 1.0 vol. % for sonication procedures and between
1.0 and 1.25 vol. % for dry mixing procedure) compared to that of the GNPs between
3.0 and 4.0 vol. %. Several models have been developed to approximate the
percolation threshold of polymers with conductive nanofillers, assuming homogeneous
and random distribution (Bao, Sun, Xiong, Guo, & Yu, 2013). Percolation theory
describes how sample properties, such as electrical conductivity, are affected by the
connectivity of objects within a network (Mutiso, Sherrott, Li, & Winey, 2012; J. K.
W. Sandler, Kirk, Kinloch, Shaffer, & Windle, 2003). From Figs. 5 and 6, the jump in
electrical conductivity is a sign of percolation, which is caused by the connectivity of
the nanofillers.
An analytical model was developed by Mutiso et al. to predict the percolation
threshold of an isotropic and monodispersed network containing CNTs and GNP with
19
Texas Tech University, Eric Collins, August 2013
modest aspect ratios (10-100) (Mutiso et al., 2012). The dimensions of the nanofillers
used in this model are the smallest particles as provided by the manufacturers (CNTs
L=100nm, D=3nm; GNPs L=8nm, D=150nm). This model assumes a fixed contact
resistance that is much greater than resistance of the CNTs.
The model predicts that the percolation threshold is equal to the nanofiller
volume divided by the excluded volume. Excluded volume is defined as a volume
surrounding a particle, which includes the center of mass of a second identical particle
with a different orientation that is in contact with the first (Balberg, Anderson,
Alexander, & Wagner, 1984). The percolation threshold (c) is dependent on the
length (L) and diameter (D) of the CNTs and GNPs as shown in Eq. 5 (Mutiso et al.,
2012).
(5)
Equation 5 predicts that the percolation threshold for GNPs occurs at 2.8 vol.
% and for CNTs occurs at 1.4 vol. %, assuming a perfect dispersion where all GNPs
and CNTs are in the form of individual disks and cylinders, respectively. Even with
the assumption of uniform particles with perfect dispersion, the model corresponds
well with the experimental results for CNTs that range from 0.5-1.0 vol. % for the
sonicated mixing procedures. The correspondence for GNP nanofillers is still
reasonable with the model’s prediction of 2.8 vol. % compared to 3.0-4.0 vol. %
experimentally measured. The experimental results shown in Figs. 5 and 6 are nicely
complemented by the estimates for percolation made using an analytical model
(Mutiso et al., 2012). Since the model is strictly dependent on the length and diameter
20
Texas Tech University, Eric Collins, August 2013
of the nanofillers, the difference in the model and experimental results is primarily due
to variance in the dimensions of the nanofillers. The quality of dispersion also plays a
role in the percolation threshold.
Note that the conductivity above the percolation threshold is different for the
two mixing procedures, which is not predicted by percolation theory. It is likely that
both of these conductive systems form percolating networks of CNTs or CNT
agglomerates. In the dry mixing case, there may be contact resistance between
conductive elements, which results in a lower conductivity.
A1 and 2 vol. % of a combination of CNTs/GNPs was added to the CEM using
the sonicated mixing method. The CNT/GNP ratio was varied from 0/100 to 100/0.
The electrical conductivity measurements for the 1 vol. % CNTs/GNPs combination
Conductivity (S/cm)
are shown in Figure 7.
1.E+02
Difference of
~5 orders of
Magnitude
1.E+00
1.E-02
1.E-04
1.E-06
1.E-08
0
20
40
60
80
% CNT in 1 Vol. % CNT/GNP
100
Figure 7. Electrical conductivity of Al + PTFE with 1 vol. % CNT/GNP ratio.
The electrical conductivity of the sample increased as the amount of CNTs in
the nanofiller of the mixture increased. Therefore the GNPs did not contribute to a rise
in electrical conductivity of the mixture at this concentration of 1vol. %. A 2 vol. %
21
Texas Tech University, Eric Collins, August 2013
CNT/GNP was also added to the CEM and the results of the electrical conductivity are
Conductivity (S/cm)
shown in Figure 8.
1.E+02
1.E+00
1.E-02
Difference of
~6.5 orders of
Magnitude
1.E-04
1.E-06
1.E-08
0
20
40
60
80
% CNT in 2 Vol. % CNT/GNP
100
Figure 8. Electrical conductivity of Al + PTFE with 2 vol. % CNTs/GNPs ratio
The trend in Fig. 8 is similar to that of Fig. 7 in that the electrical conductivity
of the sample increased as the amount of CNTs in the nanofiller of the mixture
increased. The difference is that the jump in electrical conductivity occurred at 60%
CNTs for 1 vol. % of nanofiller and at 20% CNTs for 2 vol. % nanofiller. Notice that
the percolation threshold corresponding with the volumetric percent of CNTs added in
Figs. 7 and 8 occurs between 0.4 and 0.6 vol. % and is consistent with Fig. 5 (i.e.,
mixtures with only CNTs added). The CNTs in the Al + PTFE mixture behave
different than the GNPs in that they wrap around particles and link together, creating a
conductive network throughout the mixture as seen in Fig. 4a.
ESD Sensitivity Test
The Al + PTFE and nanofillers were further examined for ESD ignition
sensitivity. The maximum voltage used to create a spark through the samples was
10kV and corresponds to 100mJ of energy. All the samples for the two mixing
22
Texas Tech University, Eric Collins, August 2013
procedures involving sonication resulted in no ignition but the samples dry mixed with
1.25 vol. % and 1.5 vol. % of added CNTs did achieve ignition below 100 mJ and
therefore deemed ESD ignition sensitive. The average electrical conductivities of the
ESD sensitive samples were 2.8x10-3 and 2.2x10-3 S/cm, respectively. These
measurements are located within an electrical conductivity region previously reported
(C. Weir et al., 2012) for aluminum and copper oxide that was shown to be ESD
ignition sensitive only within the conductivity limits between 8.8x10-4 and 1.2x10-2
S/cm as displayed in Fig. 9. The data points marked with an X indicate the reactions
that resulted in ignition from ESD.
Electrostatic
Dissipation
Threshold
Conductivity in S/cm
1.E+02
1.E+00
ESD Ignition
Sensitivity Region
1.E-02
1.E-04
Sonicated Mixtures
Dry Mix
1.E-06
1.E-08
0.0%
2.0%
Vol. % CNT
4.0%
Figure 9. Electrical conductivity that promotes ignition as depicted in shaded region.
Ignition was achieved for data points marked with an X.
The mixtures with a low electrical conductivity are not ignition sensitive to
ESD. Since conductance (G) and resistance (R) are inversely proportional, Power (P)
= V2 / R = V2 ∙ G, where V is the voltage. A composition with low conductivity results
in low power absorbed by the sample, which implies that the energy delivered to the
sample, does not reach the minimum energy required for ignition. The mixtures in the
medium conductivity range, approximately 2.5x10-3 S/cm, achieved ignition and are
23
Texas Tech University, Eric Collins, August 2013
just below the electrostatic dissipation threshold (~1x10-2 S/cm) as seen in Fig. 9.
When the electrical conductivity exceeds the electrostatic dissipation threshold, the
nanofillers create a conductive network such that the current travels through the
nanofillers (J. Sandler et al., 1999). Therefore, the mixtures with a high electrical
conductivity (around 100 S/cm) did not ignite because the current traveled through
the conductive nanofiller and bypassed the energetic material.
Conclusion
This study showed that nanofillers could be used to control the ESD ignition
sensitivity of an energetic composite. Carbon nanotubes (CNT) and graphene
nanoplatelets (GNP) were added to Al + PTFE to affect the overall mixtures electrical
conductivity and correspondingly ESD ignition sensitivity. Results showed that the
CNTs were the controlling nanofiller in affecting electrical conductivity and ESD
ignition sensitivity because their morphology wraps around fuel and oxidizer particles
achieving improved connectivity throughout the reactant matrix. In fact, the electrical
conductivity of Al + PTFE was 1x10-7 S/cm and significantly increased by almost 10
orders of magnitude to a conductivity of 100 S/cm with only 4 vol. % GNPs and 1
vol. % CNTs. When a combination of CNT/GNP nanofillers was examined, the low
volumetric percentages of CNTs created a jump in the electrical conductivity,
controlling the percolation threshold. The mixtures with a high electrical conductivity
did not ignite because the current traveled through the nanofillers bypassing the
energetic material and preventing it from heating and igniting. Mixtures within a
specific conductivity range (i.e., approximately 0.0025 S/cm) did ignite and showed
that a correlation exists between electrical conductivity and ESD ignition sensitivity.
24
Texas Tech University, Eric Collins, August 2013
CHAPTER 3:
PIEZOELECTRIC IGNITION OF NANOCOMPOSITE
ENERGETIC MATERIALS
Abstract
This study examines the ignition response of an energetic composite composed
of aluminum (Al) and molybdenum trioxide (MoO3) nano-powders to the voltage
generated from a lead zirconate and lead titanate (PZT) piezocrystal. The mechanical
stimuli used to activate the piezocrystal varied to assess ignition voltage, energy, and
delay time of Al-MoO3 for a range of bulk powder densities. Results show ignition
energy on the order of 5 mJ to ignite Al-MoO3 consistently and with high repeatability
under 0.4 ms, which is faster than observed with thermal or shock ignition. Electric
ignition of composite energetic materials is a strong function of inter-particle
connectivity yet the role of bulk density on electric ignition sensitivity is not well
understood. Results show that the ignition delay times are dependent on the powder
bulk density with an optimum bulk density of 50%. Packing fractions for particle
geometries studied here and electrical conductivity were analyzed and aid in
explaining the resulting ignition behavior as a function of bulk density.
Introduction
Piezocrystals are rarely used as igniters for energetic materials research and
there is limited reporting examining the response of energetic materials to piezocrystal
initiation (Valliappan, Swiatkiewicz, & Puszynski, 2005). Yet, piezoelectric crystals
offer a reliable and direct method for transforming mechanical energy into highvoltage electrical energy. Recent research shows that high voltage discharge into an
energetic composite produces ignition with as little as mJ ignition energy (C. Weir et
25
Texas Tech University, Eric Collins, August 2013
al., 2012). In particular, lead zirconate titanate (PZT)-based compounds develop a
voltage on the order of kV across two of its faces when compressed; this is an inherent
property of piezo-crystals and well documented for PZT (Goldfarb & Jones, 1999;
Lesieutre, Ottman, & Hofmann, 2004; Platt, Farritor, & Haider, 2005; Sodano &
Magliula, 2002; Xu, Akiyama, Nonaka, & Watanabe, 1998). Some advantages of PZT
ignition include: (1) extremely low cost (i.e., 10 cents per crystal); (2) highly
repeatable voltage response to a set mechanical stimuli; (3) tailorable to a specific
ignition sensitivity of an energetic material formulation; and (4) easy to implement.
Because the electrical output is directly proportional to the mechanical input, a PZT
ignition system could be designed to provide electric stimuli tunable towards a
specific energetic material formulation. For these reasons, understanding piezoelectric
ignition is paramount for the development of advanced igniters for miniaturized
propulsion systems, micro-thrusters and mobile robotics.
The physics governing electrical ignition of solid energetic materials is
uniquely different from other forms of ignition stimuli. The objective of this study is
to examine the response of a composite energetic material to piezoelectric ignition. In
particular, the response will be characterized for ignition delay time, ignition voltage
and ignition energy. The piezocrystal used here has also been used in other energetic
materials research (Bolyard, Neuber, Krile, & Kristiansen, 2010) and is a mixture of
lead zirconate and lead titanate, also known as PZT. This is a frequently used
piezocrystal for ignition of hydrocarbon fuels and is commonly found in propane
fueled gas igniters (Nwafor, 2004; Shudo, Nakajima, & Futakuchi, 1999). The
energetic composite is a mixture of aluminum (Al) spherical fuel particles combined
26
Texas Tech University, Eric Collins, August 2013
with molybdenum trioxide (MoO3) platelet particles. The mixtures are prepared for
varying bulk density to understand the influence of bulk density on ignition behavior.
Ignition of this particular mixture has been well documented, including ignition by
thermal and shock stimuli. For example, Al-MoO3 was shown to have an ignition
delay on the order of 20 ms when thermally ignited by a 50 W CO2 laser (J. Granier &
Pantoya, 2004) and 0.42 ms when ignited by a reflective shock (Bazyn et al., 2007). In
this way, comparison of Al-MoO3 PZT ignition sensitivity enables comparison to
other forms of stimuli and the use of PZT as an alternative ignition source.
Experimental
Figure 10 illustrates the experimental setup that is comprised of a piezoelectric
crystal with top and bottom electrically connected to two opposing copper electrodes.
27
Texas Tech University, Eric Collins, August 2013
a)
b)
Figure 10. a) Piezocrystal drop test setup for ignition of Al-MoO3 powder. b)
Electrical circuit.
A 0.8 kg drop weight with variable drop height serves to apply pressure to the PZT
and is held in place within a dielectric drop tube resting on a conducting support. The
electric charge released by the PZT is monitored with a high-speed current probe, and
a high voltage probe is utilized to monitor the potential of the developing discharge in
the electrode gap filled with Al-MoO3, shown in Fig. 10b. Note that the resistance of
the voltage probe is much greater than the resistance of the powder. Ignition of the
powder is witnessed through the sharp increase in luminosity detected with a
photodiode. The simultaneous current and voltage measurement along with the
28
Texas Tech University, Eric Collins, August 2013
luminosity enable determining ignition sensitivity quantified in terms of ignition delay
time.
The composition examined consists of 80 nm average diameter spherical Al
particles and 380 nm MoO3 flakes. The Al and MoO3 powders were mixed using a
sonification procedure described in detail elsewhere (Pantoya & Levitas, 2009). After
mixing, the powders were loaded into a channel with a volume of 31mm3. The bulk
density of the powder in the channel varied and was calculated as a percentage of the
theoretical maximum density (TMD) which is equivalent to 3.91 g/cc for Al+MoO3.
Bulk density and packing fraction are both ratios of the amount of material in the
controlled volume and are equivalent. For each bulk density examined, eight
experiments were performed to establish repeatability found to be the largest source of
uncertainty in the reported data.
Characterizing the PZT aided in understanding the influence of voltage output
on energetic material ignition response. The height of the drop weight above the PZT
was varied from 9.5 to 19.7 cm to provide a range of voltage outputs from the PZT. As
expected the peak voltage from the PZT into an open circuit increased as the drop
height increased. Initial tests were performed on the PZT using an air gap as the
reference condition. Figure 11 shows that for the 0.15 cm air gap, the voltage
amplitude increased on a millisecond timescale until the voltage potential between the
electrodes caused the electric field to exceed the dielectric strength of air and a spark
was produced.
29
Texas Tech University, Eric Collins, August 2013
4
20
0
0
-20
0.02
0.04
Time (s)
Voltage (kV)
6
Voltage
Current
Spark
60
4
40
2
20
0
0
-2
-0.05
-20
0.05
0.15
Time (s)
7
Voltage (kV)
0.06
Current (A)
0
c)
40
2
-2
b)
60
0.25
Voltage
Luminosity
5
100s
3
0.7
0.5
0.3
End
Delay Time
Start
1 Delay Time
-1
0.1
Luminocity (V)
Voltage (kV)
Voltage
Current
Current (A)
6
a)
-0.1
-100
0
100
Time (s)
200
Figure 11. a) Voltage and current as spark from PZT bridged 0.5 cm air gap. b)
Voltage and current as spark from PZT bridged gap filled with powder. c) Time delay
between spark and luminosity from photo diode.
30
Texas Tech University, Eric Collins, August 2013
When the spark bridged the gap, the voltage collapsed on a sub-microsecond
timescale and a distinct current pulse is observed, see Fig. 11a. For a spark gap
distance of 0.15 cm, the voltage output for all drop heights was repeatable with an
average value of 4.6 kV and a variation of 7% (shown in Figure 12b as a horizontal
line).
The air gap within the 0.03 cm3 volume channel was replaced with Al+MoO3
powder. The bulk densities of Al+MoO3 in the channel were 20%, 40%, 50%, and
60% TMD. The mixture was ignited using the PZT to create a spark between the two
electrodes as shown in Fig. 10b. As the spark ignited the powder, the high speed photo
diode captured the luminosity from the reaction (Fig. 11c). The length of time between
the spark and the instant luminosity producing a slope of ten (Fig. 11c) was recorded
as the ignition delay for the reaction.
The two point probing method as described in Chapter 2 was used to measure
the electrical conductivity of the Al-MoO3 for each bulk density. This robust design
was shown to provide repeatable measurements for a variety of composite energetic
material formulations and calibrated with alumina to within 1% of reported values (C.
Weir et al., 2012). All powders were tested up to five times to establish repeatability
of the measurement, the largest source of uncertainty.
Results and Discussion
Figure 12a and b show the average ignition voltage and average ignition delay
time as a function of sample bulk density, respectively.
31
Texas Tech University, Eric Collins, August 2013
a)
Time Delay (ms)
0.30
0.25
0.20
0.15
0.10
0.05
0.00
0%
20%
40%
60%
80%
% of TMD
b)
7
Voltage (kV)
6
Air gap
Voltage
5
4
3
2
1
0
0%
20%
40%
60%
80%
% of TMD
Figure 12. a) Ignition time delay as a function of bulk density expressed in terms of
percent of theoretical maximum density (% of TMD). b) Ignition voltage or dielectric
breakdown voltage resulting in ignition as a function of bulk density, also expressed in
terms of % of TMD
For both graphs, the bars indicate standard deviations. An interesting
observation from Fig. 12 is that the mid-density sample (i.e., 50% TMD) requires a
higher voltage for ignition and produces the shortest ignition delay time. This is also
the only bulk density sample studied that exceeds the 4.6 kV breakdown voltage for
the air gap case. Increasing the dielectric strength of the medium is shown to reduce
the ignition delay time by requiring a higher voltage for initiation.
Another observation from Fig. 12 is that the delay time from spark ignition is
significantly shorter than delay times reported for thermal ignition (e.g., 20 ms) (J.
32
Texas Tech University, Eric Collins, August 2013
Granier & Pantoya, 2004) and similar, i.e. about 50% faster, to the reflected shock
ignition (e.g., 0.42 ms) (Bazyn et al., 2007). The delay time varied for the tests that
produced voltages between 2-4 kV and was generally > 0.1 ms. However, as the
voltage increased to more than 5 kV, the time delay was < 0.1 ms. It is also noted that
the voltage with air as the medium had a steady value of 4.6 kV.
Studies on thermal ignition of Al-MoO3 show significant variation of ignition
delay time with variations in density. Lower density powders tend to form hot spots
that develop more prominently in highly porous (loose powder) media igniting the
mixture at reduced ignition delay times compared with consolidated media (Moore,
Pantoya, & Son, 2007). For impact ignition, increasing the bulk density corresponds to
a decrease in ignition sensitivity (Hunt, Malcolm, Pantoya, & Davis, 2009). In impact
ignition, breaking or damaging the Al passivation shell is a rate limiting step for
ignition, such that highly consolidated mixtures contain higher stresses on individual
particles due to the forced contact between particles, spurring ignition more readily for
highly consolidated mixtures (Hunt et al., 2009).
The results in Figs. 12a and b show a different trend. Ignition voltage peaks
and delay time drops for the 50% TMD bulk density samples and remains relatively
constant for all other bulk densities studied. A theory explaining this behavior involves
understanding that ignition from electric stimuli is a strong function of inter-particle
connectivity and thus the maximum packing fraction is observed to influence ignition
behavior. The Al+MoO3 particle composition consists of both spherical particles and
platelets, both with nano-scale dimensions. Literature for this combination of
morphologies is limited; however, for a random packing of spherical particles of equal
33
Texas Tech University, Eric Collins, August 2013
size, the maximum packing fraction is 58% (Al-Raoush & Alsaleh, 2007; Gotoh,
Jodrey, & Tory, 1978) and when the spherical particles vary in size, the maximum
packing fraction ranges from 58-60% (Roozbahani, Huat, & Asadi, 2013; Shi &
Zhang, 2008). Also, Coelho et al. found that the maximum packing fraction of equal
disks with an aspect ratio of 0.1 and randomly placed with the same orientation is less
than 60% (Coelho, Thovert, & Adler, 1997). Jia et al. confirmed those results with an
analytical model and also discovered that the maximum packing fraction of a
composite mixture of eight arbitrary shapes all had a maximum packing fraction less
than 60% (Jia, Gan, Williams, & Rhodes, 2007). Based on these studies, it is
reasonable to assume that the 60% TMD samples studied here at least correspond and
may exceed the maximum packing fraction for this composite, thereby inducing
deformations, fracturing and cracking within the particle matrix. It is noted that
Pantoya et al. (Pantoya & Levitas, 2009) observed cracking and deformations within
the particles of an Al+MoO3 pellet compressed to 75% TMD.
For the lowest bulk density samples (i.e., < 50% TMD), limited particle to
particle contact inhibits absorbance of electric energy. Yet for the 60% TMD case, the
highly compacted sample may exceed the maximum packing fraction such that
particle deformation and cracking inhibits absorbance of electric energy. For this case,
the highly compressed microstructure may reduce the conductive path of surface
currents channeling the electrical energy through the sample with less resistivity and
similar energy absorption observed with the lower density samples. In this way, this
highest bulk density samples behave similar to the low bulk density samples for the
pellets ability to absorb electric energy and exhibit similar ignition delay time. The
34
Texas Tech University, Eric Collins, August 2013
bulk density that provides the greatest inter-particle connectivity without deformation
may be achieved at 50% TMD, which may then also be considered as the density with
the lowest electrical field enhancement between particles and voids. That is, the local
electric fields at the particle boundaries are assumed to be lowest for the 50% TMD
for a given applied voltage. Since the local electric field drives the initiation of
electrical breakdown (Neuber, Butcher, Hatfield, & Krompholz, 1999), a higher
voltage needs to be applied to the bulk material to achieve a spark. This would explain
why the highest minimum ignition voltage is observed at 50% TMD. The shortest
ignition time delay at 50% TMD could also be considered a consequence of the higher
ignition voltage providing larger available ignition energy and the more favorable
energy deposition into the powder after the ignition is initiated. It is difficult to infer
which of the two mechanisms dominates, if at all, from the currently available
experimental data.
Figure 13 depicts the electrical conductivity for the various bulk densities of
Al+MoO3 measured using the two point probe method as described in Chapter 2.
Conductivity (S/m)
2.5E-07
2.0E-07
1.5E-07
1.0E-07
5.0E-08
0.0E+00
0%
20%
40%
60%
% TMD
Figure 13. Electrical conductivity of Al+MoO3 pellets with bulk densities of 20%,
40%, 50%, and 60% of TMD.
35
Texas Tech University, Eric Collins, August 2013
It is noted that the conductivity values in Fig. 13 are higher than those reported
in (C. Weir et al., 2012) due to a significantly larger MoO3 particle size. As the bulk
density of the powder increased, more particle to particle contact enhanced the
electrical conductivity of the powder. In other work, Robinson et al. submerged nonconducting acrylic spheres of equal size in a conductive solution and found the
electrical conductivity of the matrix decreased as the porosity decreased (Robinson &
Friedman, 2005). This finding is consistent with our result such that as the porosity of
the matrix is reduced with increasing bulk density, the electrical conductivity
increases. Beyond the maximum packing fraction, breaking and deformation of the
particles induces improved electrical conductivity, but exhibits similar ignition delay
times and ignition voltages as the low density pellets.
Research has shown a correlation between electrical conductivity and
electrostatic discharge (ESD) ignition sensitivity in binary energetic composites (C.
Weir et al., 2012). Figure 12 reveals that while the 60% TMD case shows an increase
in electrical conductivity, the 60% TMD case also results in a decrease in ignition
sensitivity (i.e. longer ignition delay time). This observation indicates that beyond the
maximum packing fraction, a linear and direct correlation no longer exists between
ESD ignition sensitivity and electrical conductivity. In other words, bulk density and
electrical conductivity are independent parameters in determining ESD ignition
sensitivity for samples that meet or exceed the maximum packing fraction.
The power (P), defined here as the rate at which energy is delivered to AlMoO3, is calculated from the voltage-current product, P = I * V. The duration of
significant power delivered to the mixture was on the time scale of about 200 ns,
36
Texas Tech University, Eric Collins, August 2013
which was also the duration of the current bridging the electrode gap through the spark
as shown in Fig. 11b. The power to the 20, 40 and 60% TMD samples averaged 130
W while the 50% TMD samples exhibited 200 W.
Conclusion
These results reveal that electric ignition of consolidated pellets composed of
nanometric Al and MoO3 exhibit reduced ignition delay times (i.e., on the order of 0.1
ms) compared to other forms of ignition stimuli (i.e., thermal, shock and impact).
Ignition energy is on the order of mJ and significantly less than other forms of ignition
stimuli (e.g. thermal requires three orders of magnitude more energy for ignition).
The trend in ignition sensitivity is a function of the bulk density of the sample.
For samples approaching the maximum packing fraction for spherical particle
powders, the ignition voltage exceeds the electric breakdown voltage in air. The
ignition of these samples (i.e., 50% bulk density) also experienced the shortest delay
times and largest minimum ignition voltage than the other bulk densities studied. For
the lower and higher density samples, the ignition voltage is less than the breakdown
voltage in air, the ignition energy is lower, and longer ignition delay times are
observed. The electrical conductivity of the samples increase as the particle to particle
contact is enhanced with an increase in bulk density. Beyond the maximum packing
fraction, no direct correlation between electrical conductivity and ESD sensitivity is
observed. In fact, electric ignition is different than other forms of ignition stimuli such
that it is a function of many parameters including electrical conductivity and packing
fraction of the medium. The PZT ignition system illustrated here offers a very low
37
Texas Tech University, Eric Collins, August 2013
energy source of electric stimuli that is controlled by a mechanical impact and highly
repeatable.
38
Texas Tech University, Eric Collins, August 2013
CHAPTER 4:
SYNTHESIZING ALUMINUM PARTICLES TOWARDS
CONTROLLING ELECTROSTATIC DISCHARGE IGNITION
SENSITIVITY
Abstract
The alumina shell around an aluminum core is believed to be a major
contributor in the initiation of a reaction containing aluminum particles. The ESD
ignition sensitivity of nano-scale Al particles, synthesized with varying shell
thicknesses ranging between 2.7 and 8.3 nm, and MoO3 was observed in terms of
ignition delay times. It was discovered that the ignition delay increased as the alumina
shell thickness increased. These results correlate with the resistivity of the sample
which also increases as the alumina content increases. A model was developed using
COMSOL Multiphysics for a single Al particle and its initiation through joule heating.
The ignition delay in the model was consistent with the experimental results
suggesting that the ESD ignition mechanism is joule heating.
Introduction
Composite energetic materials (CEM) are defined here as mixtures of
aluminum (Al) fuel and metal oxide particles, that ignite to produce exothermic
chemical energy. With the advent of nanotechnology, nano-Al fuel particles have
shown heightened reactivity compared to their micron scale counterparts (Spitzer et
al., 2010; Juan Sun et al., 2006; Ward et al., 2006). Safely handling these powder
mixtures requires a thorough understanding of their electrostatic ignition sensitivity
yet very few studies on electrostatic discharge (ESD) ignition have been reported in
39
Texas Tech University, Eric Collins, August 2013
the literature (Puszynski, Bulian, & Swiatkiewicz, 2007; C. Weir et al., 2012;
Williams, Dreizin, & Beloni, 2012).
Most ESD ignition research is performed for the discharge of electric energy
into a sample rather than by pouring induced inter-particle transport (e.g., electrostatic
sensitivity that can occur from pouring a powder sample). An interesting finding from
studying the literature on electrostatic ignition of powders is that a paradox exists
regarding the electrical properties of the powder and the corresponding electrostatic
ignition behavior. Glor (Glor, 1985) studied dust particles that had accumulated a
charge through inter-particle transport. He found that as the powder’s electrical
conductivity decreased, so does the minimum ignition energy. In other words,
materials with a decreased conductivity are more readily ignited by ESD (Glor, 1985).
In contrast, Foley et al.’s (Foley, Pacheco, Malchi, Yetter, & Higa, 2007) study on AlCuO showed that increasing electrical conductivity using additives actually decreased
the minimum ignition energy of the mixture. A striking difference in these two studies
is the way in which the electrical stimuli were introduced to the sample. For Glor
(Glor, 1985) electrostatic charge accumulated within the sample, while in Foley et al.
(Foley et al., 2007) the electrostatic charge was discharged into the sample. This
paradox poses new research questions that have potential for impactful development in
this field.
As a first step, we examined electrostatic discharge (ESD) ignition sensitivity
of nine different CEM formulations, limiting the study to only micron-Al inclusion (C.
Weir et al., 2012). The results showed that at the highest setting on the ESD apparatus
(i.e., which corresponded to 100 mJ), only Al-CuO ignited and its corresponding
40
Texas Tech University, Eric Collins, August 2013
electrical conductivity was measured to be two orders of magnitude above the next
mixture, Al-MoO3 which did not ignite (i.e., 1246 compared with 40 nS/m,
respectively). This was the first study to correlate electrical conductivity to ESD
ignition sensitivity in energetic materials (C. Weir et al., 2012).
The influence of alumina in ESD ignition sensitivity was further studied in (C.
M. Weir et al., 2013). Specifically, this study examined the electrical conductance and
ESD ignition sensitivity of aluminum and molybdenum trioxide (Al + MoO3) with
varying Al particle size ranging from nano to micron scales. The results showed that
as particle diameter decreased the electrical conductance increased by 7 orders of
magnitude and minimum ignition energy required for ESD ignition reduced
accordingly. As Al particle size is reduced, electrical conductivity increased and the
excessive presence of alumina overpowered this behavior. On the other hand,
discretely added alumina particles significantly reduce electrical conductivity,
desensitizing the mixture to ESD ignition. This study revealed that the alumina shell
may play a significant role in spurring ignition in Al+MoO3 by accumulating charge
and acting as a capacitive network, in contrast to discretely added alumina particles.
The objective of this work is to understand how the electrical conductivity and
ESD ignition sensitivity of Al+MoO3 varies as a function of the thickness of the
alumina passivation shell surrounding the Al particles. To accomplish this objective,
nano-scale Al particles were synthesized with varying shell thicknesses and combined
with nano-scale MoO3 particles. The mixture was further studied for electrical
conductivity measurements and ESD ignition sensitivity quantified in terms of ignition
delay time.
41
Texas Tech University, Eric Collins, August 2013
Experimental
The Al particles were supplied by Sigma Aldrich and had an average particle
diameter of 95 nm and the MoO3 particles were purchased from Nanostructured and
Amorphous Materials Inc and had an average platelet size of 380 nm. The Al particles
were oxidized in an isothermal oven to increase the thickness of the Al2O3 shell.
Oxidation of Aluminum powders
A thermogravimetric analyzer (TGA), model STA 409 PC by Netzsch, was
used to observe the weight gain of Al particles as the Al oxidized to Al2O3. During
oxidation in the TGA, 11.48 mg of Al powder was held in a platinum crucible while
the temperature increased at a rate of 40°C per minute in a controlled environment of
ultra-high purity oxygen. The oven continued to heat until 480°C, a temperature that
provided reasonable reaction rates where oxidation and shell growth was observed.
The Al sample remained in the isothermal environment for 180 min as the mass gain
was monitored. The precision of the change in mass of the sample in the TGA had a
variance of 0.001 mg. This data was used to control oxide shell growth on larger
quantities of aluminum powder.
A Neytech Qex oven was used to oxidize Al particles in an isothermal oxygen
environment and is shown in Figure 14 . Ultra-high purity oxygen was purged in the
oven chamber at a flow rate of 180 cm3/min for 10 minutes. This procedure ensured
that the volume was flushed five times and had a statistical purity of 99% of oxygen in
the oven chamber. The Qex oven was set to have a temperature ramp rate of 200
°C/min and a temperature of 480 °C. The settings were programmed in the oven for
automated control and repeatability between oxidation procedures. Six samples of 0.9
42
Texas Tech University, Eric Collins, August 2013
g of Al powder were prepared for each oxidation cycle. Once the chamber was purged
and the temperature of 480 °C was reached, the Al powder remained in the oven for
various durations ranging from 8 to 150 min. This variable oxidation time provided
different alumina shell thicknesses.
Figure 14. Neytech Qex oven used to oxidize Al particles.
A transmission electron microscope (TEM), model Jeol JEM-2100, was used
to view the Al particles and measure the thickness of the alumina shell. In the TEM, a
Gatan image analysis software was used to measure the thickness of the alumina shell.
Thickness measurements were taken from at least three different locations on several
Al particles. Figure 15 shows three representative images of the Al particles taken
from the TEM.
43
Texas Tech University, Eric Collins, August 2013
a)
b)
c)
Figure 15. TEM images of the alumina shell after oxidation times of a) 150 min, b) 8
min, and c) 30 min
The percent of active Al by weight (Y) was calculated using Eq. 6,
(6)
where R is the average particle radius,  is the Al2O3 shell thickness, and  is the
density of the metal (m) and metal oxide (mo). The average shell thickness and the
percent of active Al content are listed in Table 2 for all treated Al powders.
Table 2. Thickness of Al2O3 shell and weight percent active Al content for all oxidized
Al particles.
Oxidation Time (min)
0
8
15
30
60
90
150
Al2O3 Thickness (nm)
2.7
3.3
4.5
4.6
5.7
6.7
8.3
44
Wt % Active Al
78.1
73.9
66.3
65.7
59.4
54.2
46.7
Texas Tech University, Eric Collins, August 2013
Mixing powders
The Al + MoO3 mixtures were prepared to an equivalence ratio of 1.0,
corresponding to stoichiometric conditions. The percent of active aluminum in a
particle is multiplied by the actual mass of the fuel to account for the Al2O3. Once the
appropriate masses of the fuel and oxidizer were measured, they were added to a
hexane solution and sonicated. The hexane was then evaporated off, leaving a dry
composite energetic material (CEM). Details of this mixing procedure are reported in
(J. Granier & Pantoya, 2004).
Test Setup
An acrylic channel was loaded with 58 mg of the powder, which was pressed
into a pellet within the channel. The pellet occupied a volume of 31 mm3 and had a
bulk density of 1.89 g/cm3. The bulk density of the pellet was calculated to be 50% of
the theoretical maximum density (TMD). Two copper electrodes were positioned to
cover the openings in the channel such that the tips of the electrodes were in contact
with the surface of the pellet as shown in Figure 16. A voltage potential was applied to
the electrodes and when greater than the dielectric strength of the pellet material, a
spark was generated which ignited the pellet. The voltage source used in these
experiments was an electrostatic discharge (ESD) tester that is a human body model
developed by Franklin Applied Physics (C. Weir et al., 2012). A human body model
transfers charge from a capacitor (human) to another object. The range of voltage
output is 1-10 kV which is stored in a 0.002 μF capacitor producing up to 100 mJ of
energy.
45
Texas Tech University, Eric Collins, August 2013
Figure 16. Electrodes in contact with the pellet in the acrylic channel
A current monitor, model 2878 from Pearson Electronics, was used to measure
the electric charge released by the ESD tester. A high voltage probe, model P6015A
from Tektronix, reduced the voltage 1000 times, enabling the use of oscilloscopes to
record the voltage output. Ignition of the pellet in the channel was witnessed by a
steep increase in luminosity detected by a photo detector; model DET210 from
Thorlabs (see Figure 16). The luminosity along with simultaneous current and voltage
measurements were used to determine the ignition delay and ignition energy of the
pellet.
46
Texas Tech University, Eric Collins, August 2013
Figure 17. Electrical circuit used to monitor voltage and current and measure time
delay of the reaction
Results
Ignition Delay
As the voltage potential between the two electrodes increased beyond the
electric field of the medium, a spark bridged the electrode gap and ignited the CEM.
Figure 18 is a good representation of the voltage, current, and luminosity for all tests.
As the spark bridged the electrode gap, the voltage decreased and the current increased
as shown in Fig. 18a. The ignition delay was recorded as the length of time between
the spark and the luminosity as it produced a slope of ten, corresponding to ignition of
the sample (Fig. 18b).
47
Texas Tech University, Eric Collins, August 2013
3
Voltage (kV)
a)
2.5
Current (A)
2
1.5
1
0.5
0
-1
-0.5
0
0.5
Time (s)
3
2.5
2
1
Start Delay
Time
1.5
2
Voltage (kV)
Current (A)
Luminosity (V)
b)
1.5
1
0.7 ms
End Delay
Time
0.5
0
-0.25
0.25
Time (ms)
0.75
Figure 18. Voltage, current, and luminosity depicting time delay for Al+MoO3
The ignition time delay tests were conducted for CEM mixtures containing
each batch of oxidized Al powder identified in Table 2 combined with MoO3. Three
samples were tested for each Al powder with varied shell thickness and the ignition
time delay is shown in Figure 19.
48
Texas Tech University, Eric Collins, August 2013
Time Delay (ms)
5
4
3
2
1
0
0
2
4
6
Shell Thickness (nm)
8
10
Figure 19. Ignition time delay for Al+MoO3 reactions with varying alumina shell
thicknesses
All CEMs were deemed ESD ignition sensitive because all ignited below the
maximum energy threshold of 100 mJ supplied by the ESD apparatus. The thinner
shelled Al powders experienced the fastest ignition. Ignition delay increased as the
Al2O3 shell thickness increased. The voltage required to ignite the powders also
Voltage (kV)
increased as the Al2O3 shell thickness increased (Fig. 20).
3.5
3
2.5
2
1.5
1
0.5
0
0
2
4
6
Shell Thickness (nm)
8
10
Figure 20. Ignition voltage for Al+MoO3 reactions with varying alumina shell
thicknesses.
49
Texas Tech University, Eric Collins, August 2013
Joule Heating of an Aluminum Particle
The joule heating of a single Al particle was modeled using COMSOL
Multiphysics. The equation used for calculating joule heating is the time dependent
energy equation for incompressible flow as shown in Eq. 7.
(7)
In Eq. 7,  is the density, Cp is the specific heat, T is temperature, t is time, u is
a velocity vector, J is the current density, and E is the electric field strength. A single
Al particle was modeled and assumed to have a contact point with another particle on
its top, bottom, front, back, left, and right surfaces. A current of 1 Amp entered the Al
particle at one contact point while the rest of the contact points were held at ground
potential. In the model, the initial temperature of the Al particle was 20 °C and an
initial electric potential was 0 V. Because convective cooling was neglected in the
model due to the fast ignition of the particle, the boundary was insulated thermally.
The spherical Al particle was assumed to have a solid Al core with an axisymmetric
temperature distribution. The model also assumes homogeneous metals within the
particle when in reality, the metals are not homogeneous and contain impurities and
cracks.
The model results for transient heating of an Al particle with an Al2O3 shell
thickness of 2.7 nm are shown in Figure 21. As current travels through a single
particle, temperature quickly increased with time. Since the melting temperatures of
the fuel and oxide are similar, 933 and 1068 K respectively, ignition of the particle
was assumed to occur at the melting temperature of Al (Sullivan et al., 2012) and the
model predicts ignition times for these conditions to range from 3.0 - 4.5 ns. Note that
50
Texas Tech University, Eric Collins, August 2013
the predicted ignition time from the model is much less than the ignition time in the
experiments. One main reason for this difference is that the model predicts ignition
delay for a single particle whereas in the experiments, the delay is of a macroscopic
collection of particles. The heat transfer among neighboring particles within the
random media of the reactant sample is not accounted for in this model and beyond the
scope of this study. Instead, the trend in behavior as shell thickness is varied is of
interest and can be correlated to ignition delay times.
Temperature (K)
2000
1500
1000
500
0
0
1
2
3
Time (ns)
4
5
Figure 21. Temperature of Al particle from joule heating COMSOL model
Figure 22 shows the ignition delay of the Al particles for all the Al2O3 shell
thicknesses predicted by the COMSOL model.
51
Texas Tech University, Eric Collins, August 2013
Delay Time (ns)
5
4
3
2
1
0
0
2
4
6
Shell Thickness (nm)
8
10
Figure 22. Predicted ignition delay for a single Al particle with varying alumina shell
thicknesses
The model shows that the ignition delay increases as the shell thickness
increases. This result is consistent with the measured delay (Fig. 19) and implies that
the increase in shell thickness retards thermal energy buildup within a single particle
such that the ignition delay times are extended. This result implies that joule heating is
the main contributor for ignition when stimulated by electric stimuli.
The electrical conductivity was also measured using a high resistance low
conductance meter and a two point probe method (C. Weir et al., 2012). Figure 23
shows that electrical conductivity decreases as shell thickness increases.
52
Texas Tech University, Eric Collins, August 2013
Conductivity (S/m)
0.5
0.4
0.3
0.2
0.1
0
0
2
4
6
Shell Thickness (nm)
8
10
Figure 23. Electrical conductivity of Al+MoO3 powders with a varying Al2O3 shell
thickness.
Since the resistance is greater for Al2O3 than Al, as the percent of Al2O3
increases with shell thickness, the electrical conductivity decreases. Also as the
percent of Al2O3 content increases, time for current to travel through the Al2O3 shell
increases. In other words, the alumina acts as a heat sink and impedes electric energy,
decreasing the temperature of the Al and requiring more time and energy to reach the
melting temperature.
Discussion
These results show that all Al+MoO3 samples were ESD ignition sensitive
(Fig. 19) but their sensitivity is controlled by the shell thickness surrounding the Al
particle. Increasing shell thickness from 2.7 to 8.3 nm increases the ignition delay time
by 3 ms and requires up to 1500 V more electric input to achieve ignition (Figs. 19
and 20). Modeling the heat transfer through a single Al particle subjected to similar
electric input conditions reveals a comparable trend observed experimentally:
increasing shell thickness leads to increasing ignition delay time and decreasing
electrical conductivity (Figs. 22 and 23). Alumina retards energy propagation in the
53
Texas Tech University, Eric Collins, August 2013
core-shell structure of the aluminum-alumina particle and this is represented by the
COMSOL model (Figs. 21 and 22). This finding also suggests that joule heating is the
primary mechanism of ignition for ESD stimuli.
Conclusion
Aluminum particles were synthesized with varying oxide shell thicknesses that
ranged between 2.7 nm to 8.3 nm. Electrostatic discharge was used to ignite Al +
MoO3 and the ignition delay time was measured. The physics were also modeled using
COMSOL Multiphysics software to approximate the temperature increase of an Al
particle due to joule heating. The measured and modeled results both show similar
trends in that the ignition delay increased as the Al2O3 shell thickness increased. The
increase in ignition delay time suggests that the ESD ignition mechanism is joule
heating because the model calculates transient temperatures based on transforming
electrical stimuli into thermal energy. Alumina retards energy buildup within the
particle such that thicker alumina shells result in longer ignition delay times.
54
Texas Tech University, Eric Collins, August 2013
CHAPTER 5:
CONCLUSION
The ignition response of a composite energetic material was explored using
electrostatic discharge. Characterizations of the reactions such as packing density,
electrical conductivity, and ignition delay were investigated to analyze ignition
sensitivity.
Nanofillers can be used to control the ESD ignition sensitivity of an energetic
composite. Carbon nanotubes (CNT) and graphene nano particles (GNP) were added
to Al + PTFE to affect the overall mixtures electrical conductivity and
correspondingly ESD ignition sensitivity. Results showed that the CNTs were the
controlling nanofiller in affecting electrical conductivity and ESD ignition sensitivity
because their morphology wraps around fuel and oxidizer particles achieving
improved connectivity throughout the reactant matrix. In fact, the electrical
conductivity of Al + PTFE was 1x10-7 S/cm and significantly increased by almost 10
orders of magnitude to a conductivity of 100 S/cm with only 4 vol. % GNPs and 1
vol. % CNTs. When a combination of CNT/GNP nanofillers was examined, the low
volumetric percentages of CNTs created a jump in the electrical conductivity,
controlling the percolation threshold. The mixtures with a high electrical conductivity
did not ignite because the current traveled through the nanofillers bypassing heating
and ignition of the energetic composite. Mixtures within a specific conductivity range
(i.e., approximately 0.002 S/cm) did ignite and showed that a correlation exists
between electrical conductivity and ESD ignition sensitivity consistent with previous
results.
55
Texas Tech University, Eric Collins, August 2013
Electric ignition of consolidated pellets composed of nanometric Al and MoO3
exhibit reduced ignition delay times (i.e., on the order of 0.1 ms) compared to other
forms of ignition stimuli (i.e., thermal, shock and impact). Ignition energy is on the
order of mJ and significantly less than other forms of ignition stimuli (e.g. thermal
requires three orders of magnitude more energy for ignition). The trend in ignition
sensitivity is a function of the bulk density of the sample. For samples approaching the
maximum packing fraction for spherical particle powders, the ignition voltage exceeds
the electric breakdown voltage in air. The ignition of these samples (i.e., 50% bulk
density) also experienced the shortest delay times and largest minimum ignition
voltage than the other bulk densities studied. For the lower and higher density
samples, the ignition voltage is less than the breakdown voltage in air, the ignition
energy is lower, and longer ignition delay times are observed. The electrical
conductivity of the samples increase as the particle to particle contact is enhanced with
an increase in bulk density. Beyond the maximum packing fraction, no direct
correlation between electrical conductivity and ESD sensitivity is observed. In fact,
electric ignition is different than other forms of ignition stimuli such that it is a
function of many parameters including electrical conductivity and packing fraction of
the medium. The PZT ignition system illustrated here offers a very low energy source
of electric stimuli that is controlled by a mechanical impact and highly repeatable.
Aluminum particles were synthesized with varying oxide shell thicknesses that
ranged between 2.7 nm to 8.3 nm. Electrostatic discharge was used to ignite Al +
MoO3 and the ignition delay time was measured. The physics were also modeled using
COMSOL Multiphysics software to approximate the temperature increase of an Al
56
Texas Tech University, Eric Collins, August 2013
particle due to joule heating. The measured and modeled results both show similar
trends in that the ignition delay increased as the Al2O3 shell thickness increased. The
increase in ignition delay time suggests that the ESD ignition mechanism is joule
heating because the model calculates transient temperatures based on transforming
electrical stimuli into thermal energy. Alumina retards energy buildup within the
particle such that thicker alumina shells result in longer ignition delay times.
57
Texas Tech University, Eric Collins, August 2013
CHAPTER 6:
COMPARISON OF ENGINEERED NANOCOATINGS ON THE
COMBUSTION OF ALUMINUM AND COPPER OXIDE
NANOTHERMITES
Abstract
Water-repellent nano-coatings for submerged combustion of nano-energetic
composite materials were developed. These coatings may have applications for
oceanic power generation, underwater ordnance, propulsion, metal cutting, and torch
technologies. Nano-coatings were deposited on thermite pellets by a vapor-phase
technique. Two types of deposition techniques studied were chemical vapor deposition
(CVD) and atomic layer deposition (ALD). A total of six types of nano-coatings were
applied on the thermite pellets. Various process parameters to produce the coatings
were explored. Characterization of the nano-coatings was carried out using Fourier
Transform Infrared Spectroscopy (FTIR), Scanning Electron Microscopy (SEM),
Atomic Force Microscopy (AFM), and contact angle goniometry. Submerged
combustion tests of the nano-coated thermite pellets were performed as a function of
submerged time. The pellets were submerged in de-ionized water for 3, 5, and 10
days. The energy released by the thermite reaction was analyzed and compared to
other types of nano-coated pellets. Initial results of a fluorocarbon self-assembled
monolayer (FSAM) coating were compared with an ALD coating composed of Al2O3.
Results show that with increasing submerged time, there was a decrease in the ratio of
bubble energy to total energy of combustion (Kc=Kbubble/Kcombustion) for all coatings
tested. The initial bubble energy of the pellets coated with FSAM and ALD with
Al2O3 was 133.3 and 142.2 (KJ/Kg), respectively. After submersion for 10 days, the
58
Texas Tech University, Eric Collins, August 2013
bubble energy reduced to 10.4 and 15.6 (KJ/Kg), respectively. The value of Kc for the
FSAM coating decreased by a factor of 12.8 whereas the ALD with Al2O3 coating
decreased by a factor of 9.1. The hydrophobic coating is critical for energy generation
because without it, the pellets do not ignite, resulting in 100% loss of energy.
Introduction
An example of a thermite reaction is aluminum (Al) and copper oxide (CuO),
2Al + 3CuO  3Cu + Al2O3, which produces an adiabatic flame temperature of 2843
K and has a heat of reaction of 4075 kJ/kg (Fischer & Grubelich, 1998). A common
application for a thermite reaction is welding railroad ties together because of the high
heat and molten material produced in the products (Apperson et al., 2007).
Thermite reactions may have potential for underwater applications, such as
oceanic power generation, metal cutting, and welding. Aluminum has been used as an
additive for various types of underwater explosives and has been shown to improve
shock wave energy and bubble energy of the explosive (N. Wang, Wang, & Zhang,
2010). The Al and polytetrafluoroethylene (i.e., Teflon™) reaction has been studied
underwater because of its hydrophobicity (Stacy, Pantoya, Prentice, Steffler, &
Daniels, 2009). It was discovered that the energy released from the reaction can be
approximated by measuring the water displacement resulting from the reaction’s speed
and gas generation (Stacy et al., 2009).
Thermite reactions, such as Al+CuO, typically cannot ignite when submerged
underwater because the water quickly quenches the reaction. Nixon et al. discovered
that when a super-hydrophobic coating is applied to a thermite pellet, the pellet repels
59
Texas Tech University, Eric Collins, August 2013
the water and enables ignition while submerged underwater (Nixon, Pantoya,
Sivakumar, Vijayasai, & Dallas, 2011).
The objectives of this study are two-fold: (1) examine the energy generation
behaviors of the various coatings applied to thermites submerged underwater; and, (2)
examine the aging characteristics of six different types of hydrophobic coatings. The
coatings were all applied to thermite pellets consisting of Al+CuO and consolidated to
the same bulk density. To accomplish these objectives, the energetic samples were
ignited and examined in an aquarium apparatus. This enabled transient observation of
the submerged reaction.
Experimental
Materials
The Al and CuO formulation used in this study consists of nano-sized spherical
particles of Al and CuO mixed to a slightly fuel rich equivalence ratio of 1.1. The Al
particles were supplied by Nova Centrix (Austin, TX) and have an average diameter of
80nm and a specific surface area of 25m/g. The Al powder has 75% active aluminum
content by mass because of a thin, 2-3nm alumina passivation layer around the
periphery of the particle. The CuO particles were purchased from Sigma-Aldrich and
also have an average diameter of 80nm.
Pellets
Reactant powder mixtures were prepared using a sonification process, a
procedure which improves homogeneity. A detailed procedure for reactant mixing can
be found in (Kappagantula, Clark, & Pantoya, 2011). For each sample, 250mg of loose
60
Texas Tech University, Eric Collins, August 2013
thermite powder was pressed into cylindrical pellets based on the Theoretical
Maximum Density (TMD), defined by Eq. 8.
(8)
In Eq. (8), M is the percent mass of reactant i and  is the density for each reactant.
The volume of the pellet is determined by identifying a constant percent of
TMD, then calculating a height for the pellet which allows for that percent of TMD to
be achieved. The TMD calculated from (8) for Al-CuO was 4.9g/cc and throughout
this study, the percent of TMD for all the pellets was on average 52% TMD. The
pellet die shown in Figure 24 was used to press pellets and consists of a base, a die,
shims, and a plunger.
Figure 24. Pellet die assembly used to press pellets from thermite powder.
The die was placed on the base and the loose powder was poured into the die.
Shims, which are the same thickness as the height of the pellet, were placed between
61
Texas Tech University, Eric Collins, August 2013
the die and the plunger and the pellet die assembly was then compressed using a press
to form a pellet. The base and the shims were then removed and the pressed pellet was
plunged out of the die. Cold pressing of thermite powders to high compact densities
has been studied previously and Pantoya et al. elaborate further on the pressing
technique in (Pantoya & Levitas, 2009).
Nano-coating Tool
A commercially available vapor phase deposition tool from Integrated Surface
Technologies, model number RPX – 550, is used for depositing nano-coatings on
surfaces. The tool deposits by two methods; chemical vapor deposition (CVD) and
atomic layer deposition (ALD). Figure 25 shows a schematic of the nano-coating tool
which includes a vacuum chamber, vacuum pump, compressed nitrogen gas source,
precursor chemicals, and computer software to control the tool.
Figure 25. Schematic diagram of the nano-coating tool. P1, P2, etc are Precursor
chemicals. Ultra high purity nitrogen gas is used to purge the chamber and gas stick
lines
The vacuum chamber has a maximum volume of ~60,000 cm3 and the vacuum pump
can pump down the pressure inside the chamber to ~2.67 Pa. The chamber
62
Texas Tech University, Eric Collins, August 2013
temperature can be controlled on all sides of the chamber from 25 to 150°C. Five
precursor chemicals are stored inside a cartridge. Pneumatic valves precisely control
the opening of the carrier gas source (nitrogen) and the gas cartridges. The coating
applicator plate has an array of injection holes for injecting the precursor chemicals.
The chamber is also connected to a vent for vapor exhaust.
The nano-coating tool has the capability of depositing fluorocarbon monolayer coating (FSAM), alumina nano-particles (Al2O3 NP), and atomic layer
deposition (ALD) of Al2O3 and SiO2 on surfaces. The combination of these
depositions are arranged to form 6 Types of coatings as shown in Table 3.
Table 3. Coating types used to coat pellets
Type Coating
1
FSAM
2
Al2O3 NP + FSAM
3
ALD Al2O3
4
ALD SiO2 + FSAM
5
ALD Al2O3 + FSAM
6
ALD Al2O3 + Al2O3 NP + FSAM
The reaction chemistries for depositing the 6 types of coatings are described in detail
below. The ALD and nanoparticle coatings can be described as multilayers resulting
from multiple cycles of material deposition. The FSAM can be described as a
monolayer grown coating.
63
Texas Tech University, Eric Collins, August 2013
Type 1: FSAM
The chemical formulas for applying the FSAM coating are given in Eqs. (9)
and (10).
CF3(CF2)5(CH2)2 –Si–Cl3 + 3H2O  CF3(CF2)5(CH2)2 –Si–(OH)3 + 3HCl
(9)
CF3(CF2)5(CH2)2 –Si–(OH)3 + 3(-OH)  CF3(CF2)5(CH2)2 –Si–O3
(10)
The first step in applying FSAM is called hydrolyzation. The sample substrate
is hydrolyzed with vapor injection of a water and alcohol mixture as shown in Figure
26. Illustration of the a) hydrolysis reaction and b) hydrogen bonding steps in applying
the F-SAM layer on the sample substratea. This step ensures better hydroxyl surface
coverage. The tri-chlorosilane reaction with water, Eq. 9, is the next step in the
application process. This pre-cursor chemical (tridecafluoro-1,1,2,2tetrahydrooctyltrichlorosilane) reacts with water and forms silane with hydroxyl ion
termination as shown in Figure 26b.
64
Texas Tech University, Eric Collins, August 2013
Figure 26. Illustration of the a) hydrolysis reaction and b) hydrogen bonding steps in
applying the F-SAM layer on the sample substrate.
Then, the final step is cross-polymerization and hydrogen bonding, Eq. (10),
where the hydroxylated sample surface reacts with the hydroxyl terminated silane.
Adjacent silane molecules bond with each other, this process is called crosspolymerization. The byproduct in this final step is water and has a wait time of 15
minutes to ensure bonding transformation from hydrogen bonding to Si–O bonds, as
shown in Figure 27.
65
Texas Tech University, Eric Collins, August 2013
Figure 27. Illustration of an ideal Per-fluorooctyltrichlorosilane, F-SAM layer on the
sample substrate.
Figure 27 shows an ideal chemisorbed layer of the FSAM on the sample
surface. From experiments, defects in the surface are observed. Some sites on the
sample surface can be absent of the chemisorbed SAM layer, which is mainly caused
when the sample surface does not have hydroxyl ions (-OH) hence, there will be no
formation of a SAM layer. Other sites on the sample surface can have the SAM
adsorbed with imperfect or no cross-polymerization, this is mainly due to lattice
mismatching of silane-OH (X–Si–O) with the surface –OH. In this case, all three
hydroxyl bonds of the silane will bond with the surface hydroxyl groups. During the
process run, the excess chemicals and byproducts are pumped-out of the chamber. The
temperature of the chamber is set at 45°C. The temperature of the FOTS cartridge is
set at 85°C, whereas the temperature of water + alcohol cartridge is set at 45°C.
66
Texas Tech University, Eric Collins, August 2013
Type 2: Al2O3 NP + FSAM
The Type 2 coating is a super-hydrophobic coating and the application process
is explained in detail by Nixon et al. (Nixon et al., 2011). The ‘recipe flow’ for the
super-hydrophobic nano-coating deposition process is illustrated in Figure 28.
Figure 28. Recipe flow diagram for the deposition of a super-hydrophobic coating.
The cleaning steps are not shown in Figure 28, but consist of pre-process and
post-process steps that pump-down and purge the chamber for five iterations. The
deposition step includes injection of tri-methyl alumina (TMA) and water vapor +
67
Texas Tech University, Eric Collins, August 2013
alcohol into the chamber. This step is followed by an adhesive layer and water vapor +
alcohol injection. The final step is the injection of water vapor and FOTS into the
chamber. Excess chemicals and byproducts are pumped-out of the chamber. During
the process run, the temperature of the chamber is set at 45°C. The temperature of the
water + alcohol mixture is also set at 45°C. The temperature of the TMA is set at
45°C; the temperature of FOTS cartridge is set at 85°C.
Type 3: ALD Al2O3
Atomic layer deposition (ALD) is a technique wherein thin films are grown at
the atomic scale. This technique involves sequential chemical reactions occurring at
the substrate surface with a constant temperature of 45°C. The main characteristic of
these reactions is its self-limiting nature. Table 4 summarizes details of the ALD
process for Al2O3. The chemical reactions involved in the ALD of Al2O3 reaction
shown in Eqs. (11) and (12).
Al – OH* + Al(CH3)3  Al – O – Al(CH3)2* + CH4
(11)
Al – (CH3)2* + 2H2O  Al – (OH)2* + 2CH4
(12)
The first step in applying the ALD Al2O3 coating is trimethyl alumina (TMA)
injection into the chamber (i.e., Eq. (11)). One of the methyl groups reacts with the
hydroxyl ion on the substrate surface to form methane as the byproduct. Excessive
TMA inside the chamber and the methane are pumped out of the chamber. The second
step is water vapor injection into the chamber (i.e., Eq. (12)). The remaining two
methyl groups in Al–(CH3)2 oxidize with water to form Al–(OH)2. Among the two
hydroxyl groups in Al – (OH)2 molecule, one of them participates in hydrogen
bonding with adjacent molecule and the other acts as a surface for the next injection of
68
Texas Tech University, Eric Collins, August 2013
TMA. Hydrogen bonding is formed between adjacent Al–(OH) molecules to
eventually form Al–O–Al bridges. Excessive water vapor and methane are pumped
out of the chamber.
In both the steps, the pump-down cycle is very important, as it differentiates
the nano-particle generation from the atomic layer generation. If excessive TMA or
water vapor had been inside the chamber, the next sequential step would produce
nano-particles, which is a gas phase reaction. The thickness of the coating after one
ALD cycle is between 1.1 to 1.4 Å, and the variation in thickness is dependent on the
temperature of the ALD process run (Groner, Fabreguette, Elam, & George, 2004). All
ALD experiments were performed for 50 cycles, resulting in a coating thickness
between 55 to 70 Å.
69
Texas Tech University, Eric Collins, August 2013
Table 4. Details of the Al2O3 ALD process. Chamber temperature was maintained at
45°C.
TMA cartridge - 45°C, stick - 50°C;
Water+alcohol cartridge - 45°C,
stick - 50°C
Nitrogen
gas
N2-1
TMA
Gas
5
Nitrogen
gas
N2-5
Pump
Rough
Step name
Water +
alchohol
mixture
Gas 1
Pressure
(Torr) or
Time (sec)
TMA injection
-
-
ON
-
-
>0.05 Torr
Excess TMA Purge
-
ON
-
ON
ON
10 sec
Pump down
-
-
-
-
ON
<0.04 Torr
Intermediate Purge - 1
-
ON
-
ON
ON
10 sec
Pump down
-
-
-
-
ON
<0.04 Torr
Intermediate Purge - 2
-
ON
-
ON
ON
10 sec
Pump down
-
-
-
-
ON
<0.04 Torr
Water+alcohol mixture Injection
ON
-
-
-
-
>0.05 Torr
Excess water+alcohol mixture Purge
-
ON
-
ON
ON
10 sec
Pump down
-
-
-
-
ON
<0.04 Torr
Intermediate Purge - 1
-
ON
-
ON
ON
10 sec
Pump down
-
-
-
-
ON
<0.04 Torr
Intermediate Purge - 2
-
ON
-
ON
ON
10 sec
Pump down
-
-
-
-
ON
<0.02 Torr
Intermediate Purge - 3
-
ON
-
ON
ON
10 sec
Pump down
-
-
-
-
ON
<0.04 Torr
Intermediate Purge - 4
-
ON
-
ON
ON
10 sec
Pump down
-
-
-
-
ON
<0.02 Torr
Type 4: ALD SiO2 + FSAM
Table 5 summarizes details of the ALD process for SiO2. The chemical
reactions involved in the ALD SiO2 coating are described in Eqs. (13) and (14).
Si–OH* + SiCl4  Si–O–SiCl3* + HCl
(13)
Si–O–SiCl3* + 3H2O  Si–(OH)3* + 3HCl
(14)
The ALD SiO2 deposition process is similar to ALD Al2O3 except the injected
chemicals are tetrachlorosilane and water vapor. The main difference in the chemical
70
Texas Tech University, Eric Collins, August 2013
reactions between the two ALD coatings is the SiO2 growth process requires the
presence of a catalyst to execute the chemical reaction at temperatures <50°C. The
catalyst used here is pyridine. In ALD SiO2, Si–O–Si linkages are observed. Even
here, the pump down step is very important. During the process run for atomic layer
deposition, the temperature of the chamber is set at 50°C. The temperature of TMA,
water + alcohol mixture, tetrachlorosilane, and pyridine are all set at 50°C. FSAM is
then applied to the sample using the procedure described for Type 1.
The reaction chemistry for Type 5: ALD Al2O3 + FSAM and Type 6: ALD
Al2O3 + Al2O3 NP + FSAM are not expounded on here because they are derived from
the previously explained nano-coating recipes used for Type 1, Type 2 and Type 3.
Table 5. Details of the SiO2 ALD process. Chamber temperature was maintained at
50°C.
Water+alcohol,
Pyridine, siloxane gas
cartridges - 50°C, all
stick lines - 55°C
Water +
alchohol
mixture
Gas 1
Nitrogen
gas
N2-1
Pyridine
gas
Gas 2
Nitrogen
gas
N2-4
Siloxane
gas
Gas 4
Nitrogen
gas
N2-4
Pump
Rough
Pressure
(Torr)
or Time
(sec)
Linkerrix Carrer ON
-
-
-
-
-
ON
-
2 sec
Linkerrix - ON
-
-
ON
ON
ON
ON
-
>0.2
Torr
Linkerrix Purge
-
-
ON
ON
-
ON
-
5 sec
Diffusion Time
-
-
-
-
-
ON
-
5 sec
Reaction Step #1
-
-
-
-
-
-
-
8 sec
Process Purge 1
-
-
-
-
-
ON
ON
5 sec
Quick Pump
-
-
-
-
-
-
ON
<0.3
Torr
Zorrix Carrier - ON
-
ON
-
-
-
-
-
2 sec
Step name
Zorrix Injection
ON
ON
ON
ON
-
-
-
>0.34
Torr
Zorrix Purge
-
ON
ON
ON
-
-
-
5 sec
Reaction Step #2
-
-
-
-
-
-
-
8 sec
Process Purge 2
-
ON
-
-
-
-
ON
10 sec
Pump Out
-
-
-
-
-
-
ON
<0.125
Torr
71
Texas Tech University, Eric Collins, August 2013
Witness Samples
Characterization of each nano-coating with commercially available tools
requires the nano-coating deposited on clean, ‘witness sample’ substrates. Silicon
wafer pieces and glass slides were used as the sample substrates. Both substrates were
cleaned by wet and dry procedures. First, by wet procedure, the substrates were
alternatively washed in isopropyl alcohol and de-ionized water for 30 minutes. After
this, the samples were exposed to dry air at room temperature. Then the samples were
loaded in a UV-O cleaner; the tool used was the Novascan UV-4000; the cleaning
duration was 20 minutes. The majority of organic contaminants were removed and this
cleaning procedure ensures the surface termination with hydroxyl ions. Contact angle
goniometer measurements of the cleaned samples reveal complete wetting of the
surface. The measured value is always < 5° for a 2µl de-ionized water drop.
Combustion Test Setup
The experimental setup shown in Figure 29 consists of an aquarium, a variac
voltage regulator as the power source for pellet ignition, and continuous monitoring
using a Phantom v7 high speed video. An Al+CuO pellet was placed in a channel
approximately 3 mm deep in the pellet stabilizer within the aquarium and 1.0 liters of
de-ionized water were added to the aquarium. A Nichrome wire was placed on the
pellet and a variac transformer supplied current through the Nichrome wire to heat and
ignite the pellet. The high speed camera recorded the growth rate of the gaseous
bubble formed by the thermite reaction.
72
Texas Tech University, Eric Collins, August 2013
Figure 29. Combustion of the thermite pellets test setup used to characterize energy
generation from submerged reactions.
The energy generated by the reaction was calculated using Eq. 15 based an
energy displacement (Ed).
(15)
In Eq. 15, V is the volume of the bubble which is approximated as an ellipsoid, P is the
hydrostatic pressure, and M is the mass of the pellet.
Contact Angle Goniometer
Static contact angles were measured on the sample surface to quantify the
degree of hydrophobicity. A Rame Hart #100-00 goniometer measured static contact
angle and consists of a CCD camera mounted telescope, tilting sample base, microsyringe dispenser, and adjustable illuminator. By dispensing a few micro-liters of D.I.
water on the nano-coated samples, static contact angle measurements can be recorded.
The surface is hydrophobic if the contact angle is ≥ 90° and super-hydrophobic if the
contact angle is ≥ 150°. The degree of hydrophobicity increases with the measured
73
Texas Tech University, Eric Collins, August 2013
contact angle. Figure 30 shows sample images of 2µl water droplets dispensed on
hydrophobic and super-hydrophobic surfaces. Images are processed using ImageJ
software to calculate the contact angle. Fluorocta-trichlorosilane (FOTS) coated
surfaces produce a contact angle of 107±2°, while for super-hydrophobic coated
surfaces the angle is 167±2°.
a)
b)
Figure 30. Example images from contact angle measurements. A micro water-drop on
(a) hydrophobic sample surface and (b) super-hydrophobic sample surface.
Spectrometry
A Fourier Transform Infrared Spectrometer (FTIR) was used to examine the
absorbance properties of the coatings as a function of submersion time. A Bruker
Tensor 27 – Attenuated Total Reflectance (ATR) FTIR was used and can measure
absorption spectra from 650 to 4000 cm-1 with a ZnSe crystal ATR cell. The
measurement resolution is 1cm-1. The FTIR measures the interaction of infrared
radiation with the vibrating dipole moments of the sample molecules.
Summary of Measurements
FTIR measurements were conducted for type 2 and type 6 coatings. Contact
angle measurements and the underwater combustion tests were conducted for all 6
coating types. All the thermite pellets and witness samples were submerged in separate
74
Texas Tech University, Eric Collins, August 2013
Pyrex containers with de-ionized water. At 0, 3, 5, and 10 days, the witness samples
were removed from the containers and gently dried using a dry air gun and instantly
examined for contact angle measurements. After submersion time, the thermite pellets
were immediately transferred to the aquarium apparatus for combustion testing. Six
pellets for each of the six different coatings listed in Table 3 were examined for
repeatability throughout these experiments. Pellets were submerged for various times
and enabled aging characteristics of each coating to be examined.
Results
Combustion Tests
Figure 31. Still frame images of underwater combustion for Al-Cuo with
coating Type 1 after (a) 5ms, (b) 10ms, (c) 15ms, (d) 20ms, (e) 25ms, and (f)
30msshows still frame images from a representative underwater reaction. When the
pellet ignited, a bubble formed and continued to expand due to pressure and gas
generation from the reaction. When the reaction was complete, the bubble collapsed
and the products dispersed throughout the water. The collapsing of the bubble after an
explosion underwater is due to the decrease in pressure as the high temperature from
the reaction decreases. This collapsing phenomenon has been heavily studied (Kan,
Stuhmiller, & Chan, 2005; Menon & Lal, 1998; Zhang, Yao, & Yu, 2008).
75
Texas Tech University, Eric Collins, August 2013
Figure 31. Still frame images of underwater combustion for Al-Cuo with coating Type
1 after (a) 5ms, (b) 10ms, (c) 15ms, (d) 20ms, (e) 25ms, and (f) 30ms.
The displaced energy produced by the reaction is tabulated in Table 6 and
shown graphically in Figure 32.
Energy (KJ/Kg)
200
FSAM
Al2O3 NP+FSAM
ALD Al2O3
ALD SiO2 + FSAM
ALD Al2O3 + FSAM
ALD Al2O3 + Al2O3 NP + FSAM
150
100
50
0
0
2
4
6
8
Submersion Time (days)
10
Figure 32. Energy released from coated thermite pellets for all coating types with
respect to submersion time.
These results show that the pellets with Type 6 coating released the most
energy for the 0 day submersion time (immediate ignition when submerged). The
energy released after immediate ignition from Types 1, 5, and 6 are about 5% of the
76
Texas Tech University, Eric Collins, August 2013
heat of combustion for Al+CuO. This low percentage of the heat of combustion is due
to cooling and quenching from the water upon contact with the thermite reaction. The
results from Table 6. Average energy (kJ/kg) released from pellets with their
respective submersion time and coating typealso show that the pellet with Type 6
coating produced the most energy after being submerged for 10 days.
Table 6. Average energy (kJ/kg) released from pellets with their respective
submersion time and coating type.
Type 1
Type 2
Type 3
Type 4
Type 5
Type 6
Submersion
FSAM
Time
Al2O3
NP +
FSAM
ALD
Al2O3
ALD
SiO2 +
FSAM
ALD
Al2O3 +
FSAM
ALD Al2O3
+ Al2O3 NP
+ FSAM
136.1
6.3
8.4
8.9
142.2
32.6
30.4
15.6
0 day
3 day
5 day
10 day
133.3
----*
----*
10.4
123.4
59.3
41.1
10.5
48.7
89.4
6.3
60.9
5.6
----*
3.1
14.4
*Data not available
Figure 33 shows a representative FTIR spectrum of the ALD
Al2O3+NP+FSAM hydrophobic coating. The peak identified as H2O and OH on
Al2O3 NP is the water and OH molecules on the aluminum oxide nanoparticles (AlAbadleh & Grassian, 2003; Szymanski, Rowlette, & Wolden, 2008). Impurities have
been reported to be found in the aluminum oxide at 1300-1800 cm-1 (Szymanski et al.,
2008). One of the peaks identified is the Si-O-Si stretching mode at 1059 cm−1 (Nixon
et al., 2011). The peaks at 666 cm-1 and 860 cm-1 are assigned to O-Al-O bending and
Al-O stretching, respectively(Catherine & Talebian, 1988; Chowdhuri, Takoudis, Klie,
& Browning, 2012). The peak at 1241 cm-1 is from C-F stretching, while the peaks at
1212 and 1145 cm-1 are attributed to asymmetric and symmetric CF2 stretching
(Devaprakasam, Sampath, & Biswas, 2004; Kim, Kim, Kang, & Uhm, 2005). The
77
Texas Tech University, Eric Collins, August 2013
peaks at 2920 and 2850 cm-1 are due to CH2 asymmetric and symmetric virations
(Garside & Wyeth, 2012).
Figure 33. FTIR spectra highlighting the peaks for bonds present for Al2O3 NP +
FSAM coating.
Figure 34 shows absorbance for ALD Al2O3+NP+FSAM nano-coating with
respect to number of day submerged. The absorbance of the Si-O-Si bond, which
holds the FSAM together, and the bonds with aluminum decreases as submersion time
increases. The absorbance of the C-H bond slightly increases with increasing
submerged duration (until 5 days), and drops down at day 10. We attribute this
abnormal behavior as error in measurement.
78
Texas Tech University, Eric Collins, August 2013
Figure 34. Absorbance trends from FTIR for major bonds in Type 6 coating with
respect to submersion time.
Witness Sample Surface Characterization
Witness samples were examined using an optical microscope, a scanning
electron microscope (SEM), and an atomic force microscope (AFM). Surface
roughness of the substrates with ALD films were examined with an AFM at 3 sites on
each witness sample (AFM workshop model TT-AFM, probe – K-TE nano model#
CSG 01, scan speed – 0.3Hz, scan area 5 x 5 um, contact mode scanning method). The
surface roughness of the cleaned silicon substrate is 0.79 nm with a standard deviation
of 0.45 nm. The surface roughness of the ALD coated silicon substrate is 0.91 nm with
a standard deviation of 0.31 nm. Variation of the rms surface roughness is ~ ±3Å.
Even optical observations show no traces of nano-particle like agglomeration which is
common in a vapor phase reaction of precursor chemicals (Tri-methyl Alumina and
Water).
79
Texas Tech University, Eric Collins, August 2013
For a nano-particle based coating, the surface coverage of nano-particles was
dense at 0 days, but after 10 days, large voids (up to 100µm) were found on the
surface (indicating no nano-particles). Figure 35 shows images of “ALD Al2O3+
Al2O3 NP+FSAM” coating at 0, 5 and 10 days. The ALD and fluorocarbon based
coatings revealed very little particle contamination and the surfaces are as smooth as
observed on 0 day.
80
Texas Tech University, Eric Collins, August 2013
Figure 35. SEM images of nano-particle based coating (ALD–Al2O3+NP+FSAM) on
witness samples after; A) 0 day, B) 5 day, and C) 10 day underwater submersion time.
81
Texas Tech University, Eric Collins, August 2013
Figure 36 shows the contact angle results of witness samples for an increasing
number of days submerged in water. Every data point represents five tests averaged
and variation in the measurement for each data point is represented with error bars. All
the nano-coating types degrade with increasing submerged time. All the coatings
which were hydrophobic (90° to 150°) at day 0 lose their hydrophobicity (< 90°) after
being submerged for 3 days. Also, the super-hydrophobic coatings (> 150°) convert to
hydrophobic coating (90° to 150°) after being submerged for 3 days.
Figure 36. Graph showing the measured contact angle results against submersion days.
Discussion
From Figure 32, the energy decrease over time may be due to water
permeating through the nano-coatings and soaking into the pellet. After some of the
reactions, un-reacted solid chunks from the pellets were found in the water. It was
observed that the weight of the un-reacted chunks increased the longer the pellets were
82
Texas Tech University, Eric Collins, August 2013
submerged in the water. More water may have been absorbed into the pellet the longer
it was submerged underwater, quenching the reaction and resulting in incomplete
combustion.
One reason for water permeating into the pellet could be due to the breakdown
of bonds within the coatings. The absorbance of the bonds varied depending on the
quantity that was applied to the FTIR and magnitude of pressure applied by the screw
gauge in the ATR cell. The overall absorbance of the bonds decreased as submersion
time increased (Figure 34). As the coating was submerged in water for long periods of
time, the bonds became weak and allowed for water to permeate into the pellet.
Another cause for water permeation over time could be due to rough
surfaces on the pellet which would result in incomplete surface coverage for all nanocoating types. These gaps allow water to leak through the coating and soak into the
pellet. Images shown in Figure 35 support this theory by showing the surface is
affected over time. Also, the contact angle results (Figure 36) suggest that there are
three ranges of contact angle measurements. The first range corresponds to nanoparticle based coatings with contact angles ranging from 175º to 110º. The second is
FSAM based coating and final and lowest range is ALD coatings. The contact angle
measurements over increasing submerged time are consistent with the data in
Table 5 and Figure 32 which show nano-particle based coatings perform better
than both FSAM and ALD coatings.
The goal of future work is to modify the structure of the coating to enable
longer duration of submersion without reduction in combustion performance. From the
83
Texas Tech University, Eric Collins, August 2013
preliminary results provided here, this may be achieved through multiple approaches.
For example, strategic material selection of nanoparticle-based coatings (i.e., such as
carbon based materials including graphene) coupled with development of a denser
polymer chain monolayer and smoothing the surface roughness of the thermite pellet
may all lead to reduction or prevention of water permeation over time.
Conclusion
Thermite pellets composed of Al+CuO were examined for submerged
combustion with 6 different types of hydrophobic and super-hydrophobic coatings. All
coatings enabled submerged ignition and combustion but the multilayer coating that
included ALD of Al2O3, nanoparticles of Al2O3 capped with FSAM produced the
greatest energy generation of the thermite immediately upon submersion and for
extended submersion times. For all coatings, solid reactant chunks that increased in
mass with time of submersion, were found after ignition suggesting that water slowly
leaked through the coating resulting in degrading the combustion performance with
submersion time.
Water permeation into the pellet may be due to the breakdown of coating
bonds which is a strong function of submersion time. Results from the FTIR show that
the absorbance of the bonds decreased as submersion time increased. Another cause of
water permeating into the pellet may be associated with rough surfaces on the pellet
which may facilitate gaps in the coatings. Surface characterizations of the coated
pellets also reveal degradation of the coating with submersion time.
Results from this study have implications for developing next generation
coatings on energetic materials that will enable their use and application for
84
Texas Tech University, Eric Collins, August 2013
submerged energy generation technologies. Future work will provide greater insight
into the mechanisms that will maintain combustion performance with longer
submersion times.
85
Texas Tech University, Eric Collins, August 2013
FUTURE WORK
The results in this work opened the door to many avenues to improve ignition
sensitivity and achieve safer handling of energetic materials. This study discovered
that when carbon nanotubes reach a conductive network such that the electrical
conductivity exceeds a threshold of 1x10-2 S/cm, the current travels through the
carbon nanotubes and goes around the energetic material. Accidental ignition from
electrostatic discharge is a problem with some energetic materials such as Al+CuO
and can potentially be very dangerous. To prevent possible disasters, a composition
that is sensitive to ESD can be engineered to become insensitive to ESD but yet
maintain its combustion performance. An energetic material that is not sensitive to
ESD but still ignitable with other sources such as hot wire or laser could become very
attractive. The engineered energetic material would be prepared by adding carbon
nanotubes at a volumetric percentage where percolation is achieved to create a
conductive network. If the composition does not ignite, then it was desensitized to
ESD without significantly changing the combustion properties.
A conductivity range was observed where energetic materials were sensitive to
ignition from ESD. Studies could be performed that focus on the conductivity and
ignition sensitivity of materials would be beneficial to better understand ignition
behavior within the observed ESD sensitivity range. These studies could involve
adding carbon nanotubes or various mass percentages of a conductive oxidizer such as
copper oxide to a composition with a low conductivity. In this way, ignition sensitivity
can be observed for various compositions with varying electrical conductivities to
86
Texas Tech University, Eric Collins, August 2013
determine if electrical conductivity does in fact correlate to ignition sensitivity within
the range specified in this study.
Another study that would provide better understanding of Al combustion and
the significance of the alumina shell is keeping the percent of aluminum constant. In
this study, we learned that as the thickness of the alumina shell increased, time delay
of the reaction increased. But as the alumina shell increased, the percent of aluminum
content in the particles decreased, changing combustion of the reaction. The future
study would have varying alumina shell thicknesses but add additional oxide to some
compositions such that the percent of aluminum is constant. This would allow for
better understanding of the contribution of the alumina shell as the stoichiometry is
consistent.
87
Texas Tech University, Eric Collins, August 2013
REFERENCES
Al-Abadleh, H., & Grassian, V. (2003). FT-IR study of water adsorption on aluminum
oxide surfaces. Langmuir, 19(2), 341–347. Retrieved from
http://pubs.acs.org/doi/abs/10.1021/la026208a
Al-Raoush, R., & Alsaleh, M. (2007). Simulation of random packing of polydisperse
particles. Powder Technology, 176(1), 47–55. doi:10.1016/j.powtec.2007.02.007
Apperson, S., Shende, R. V., Subramanian, S., Tappmeyer, D., Gangopadhyay, S.,
Chen, Z., Gangopadhyay, K., et al. (2007). Generation of fast propagating
combustion and shock waves with copper oxide/aluminum nanothermite
composites. Applied Physics Letters, 91(24), 243109. doi:10.1063/1.2787972
Aumann, C. E. (1995). Oxidation behavior of aluminum nanopowders. Journal of
Vacuum Science & Technology B: Microelectronics and Nanometer Structures,
13(3), 1178. doi:10.1116/1.588232
Balberg, I., Anderson, C., Alexander, S., & Wagner, N. (1984). Excluded volume and
its relation to the onset of percolation. Physical review B, 30(7), 3933–3943.
Retrieved from http://prb.aps.org/abstract/PRB/v30/i7/p3933_1
Bao, H.-D., Sun, Y., Xiong, Z.-Y., Guo, Z.-X., & Yu, J. (2013). Effects of the
dispersion state and aspect ratio of carbon nanotubes on their electrical
percolation threshold in a polymer. Journal of Applied Polymer Science, 128(1),
735–740. doi:10.1002/app.37554
Bazyn, T., Glumac, N., Krier, H., Ward, T. S., Schoenitz, M., & Dreizin, E. L. (2007).
REFLECTED SHOCK IGNITION AND COMBUSTION OF ALUMINUM
AND NANOCOMPOSITE THERMITE POWDERS. Combustion Science and
Technology, 179(3), 457–476. doi:10.1080/00102200600637261
Beloni, E., & Dreizin, E. (2009). Experimental study of ignition of magnesium powder
by electrostatic discharge. Combustion and Flame, 156(7), 1386–1395.
doi:10.1016/j.combustflame.2009.02.009
Beloni, E., & Dreizin, E. (2010). Ignition of aluminum powders by electro-static
discharge. Combustion and Flame, 157(7), 1346–1355.
doi:10.1016/j.combustflame.2010.01.010
Beloni, E., & Dreizin, E. L. (2011). Ignition of Titanium Powder Layers by
Electrostatic Discharge. Combustion Science and Technology, 183(8), 823–845.
88
Texas Tech University, Eric Collins, August 2013
Bockmon, B. S., Pantoya, M. L., Son, S. F., Asay, B. W., & Mang, J. T. (2005).
Combustion velocities and propagation mechanisms of metastable interstitial
composites. Journal of Applied Physics, 98(6), 064903. doi:10.1063/1.2058175
Bolyard, D. W., Neuber, A. A., Krile, J. T., & Kristiansen, M. (2010). Simulation of
Compact Explosively Driven Ferroelectric Generators. IEEE Transactions on
Plasma Science, 38(4), 1008–1014. doi:10.1109/TPS.2010.2040819
Brown, F., Kusler, D., & Gibson, F. (1953). Sensitivity of explosives to initiation by
electrostatic discharges, (September). Retrieved from
http://scholar.google.com/scholar?hl=en&btnG=Search&q=intitle:Sensitivity+of
+explosives+to+initiation+by+electrostatic+discharges#0
Catherine, Y., & Talebian, A. (1988). Plasma Deposition of Aluminum Oxide Films,
17(2).
Chen, Y., Lawrence, F. V, Barkan, C. P. L., & Dantzig, J. a. (2006). Heat transfer
modelling of rail thermite welding. Proceedings of the Institution of Mechanical
Engineers, Part F: Journal of Rail and Rapid Transit, 220(3), 207–217.
doi:10.1243/09544097F01505
Chowdhuri, A. R., Takoudis, C. G., Klie, R. F., & Browning, N. D. (2012).
Metalorganic chemical vapor deposition of aluminum oxide on Si : Evidence of
interface SiO2 formation Metalorganic chemical vapor deposition of aluminum
oxide on Si : Evidence of interface SiO 2 formation. Applied Physics Letters,
4241(2002), 111–114. doi:10.1063/1.1483903
Coelho, D., Thovert, J.-F., & Adler, P. M. (1997). Geometrical and transport
properties of random packings of spheres and aspherical particles. Physical
Review E, 55(2), 1959–1978. doi:10.1103/PhysRevE.55.1959
Crane, C. a., Pantoya, M. L., & Dunn, J. (2010). Infrared measurements of energy
transfer from energetic materials to steel substrates. International Journal of
Thermal Sciences, 49(10), 1877–1885. doi:10.1016/j.ijthermalsci.2010.06.005
Deevi, S., Sikka, V., & Swindeman, C. (1997). Application of reaction synthesis
principles to thermal spray coatings. Journal of materials, 32(12), 3315–3325.
doi:10.1023/A:1018691826809
Densmore, J. M., Biss, M. M., Homan, B. E., & McNesby, K. L. (2012). Thermal
imaging of nickel-aluminum and aluminum-polytetrafluoroethylene impact
initiated combustion. Journal of Applied Physics, 112(8), 084911.
doi:10.1063/1.4762009
89
Texas Tech University, Eric Collins, August 2013
Devaprakasam, D., Sampath, S., & Biswas, S. K. (2004). Thermal Stability of
Perfluoroalkyl Silane Self-Assembled on a Polycrystalline Aluminum Surface.
Society, (13), 1329–1334.
Fischer, S. H., & Grubelich, M. . (1998). Theoretical Energy Release of Thermites,
Intermetallics, and Combustible Metals. 24th International Pyrotechnics Seminar
(Vol. 220, p. 56). Monterey. doi:10.2172/658208
Foley, T., Pacheco, A., Malchi, J., Yetter, R., & Higa, K. (2007). Development of
Nanothermite Composites with Variable Electrostatic Discharge Ignition
Thresholds. Propellants, Explosives, Pyrotechnics, 6(6).
doi:10.1002/prep.200700273
Garside, P., & Wyeth, P. (2012). Cellulosic Identification of FTIR Spectroscopy
Fibres by THREAD AND SINGLE FIBRE ANALYSIS BY ATTENUATED,
48(4), 269–275.
Gesner, J., Pantoya, M. L., & Levitas, V. I. (2012). Effect of oxide shell growth on
nano-aluminum thermite propagation rates. Combustion and Flame, 159(11),
3448–3453. doi:10.1016/j.combustflame.2012.06.002
Glor, M. (1985). Hazards due to electrostatic charging of powders. Journal of
Electrostatics, 16, 175–191.
Goldfarb, M., & Jones, L. (1999). On the Efficiency of Electric Power Generation
With Piezoelectric Ceramic. Journal of Dynamic Systems Measurement and
Control - Transactions of the ASME, 121(3), 566. doi:10.1115/1.2802517
Goldschmidt, H. (1895). Verfahren zur Herstellung von Metallen oder Metalloiden
oder Legierungen derselben (Process for the production of metals or metalloids or
alloys of the same). Process for the production of metals or metalloids or ….
Retrieved from
http://scholar.google.com/scholar?hl=en&btnG=Search&q=intitle:Verfahren+zur
+Herstellung+von+Metallen+oder+Metalloiden+oder+Legierungen+derselben#0
Goldschmidt, H. (1908). New Thermite Reactions. Iron Age, 82(232). Retrieved from
http://scholar.google.com/scholar?hl=en&btnG=Search&q=intitle:New+Thermite
+Reactions#0
Gorbunov, V., & Shidlovskii, A. (1986). Combustion of mixtures of magnesium and
oxides, sulfides, fluorides, and chlorides of lead (II) or copper (II). Combustion,
Explosion, and Shock Waves, 22(6), 726–727. Retrieved from
http://www.springerlink.com/index/W1X78222P748111X.pdf
90
Texas Tech University, Eric Collins, August 2013
Gotoh, K., Jodrey, W. S., & Tory, E. M. (1978). A random packing structure of equal
spheres — statistical geometrical analysis of tetrahedral configurations. Powder
Technology, 20(2), 233–242. doi:10.1016/0032-5910(78)80054-0
Granier, J. J., & Pantoya, M. L. (2011). Ignition and Combustion Behaviors of
Nanocomposite Al/MoO3. MRS Proceedings, 800, AA5.3. doi:10.1557/PROC800-AA5.3
Granier, J., & Pantoya, M. (2004). Laser ignition of nanocomposite thermites.
Combustion and Flame, 138(4), 373–383.
doi:10.1016/j.combustflame.2004.05.006
Groner, M. D., Fabreguette, F. H., Elam, J. W., & George, S. M. (2004). LowTemperature Al 2 O 3 Atomic Layer Deposition. Thin Solid Films, (16), 639–
645.
Hunt, E. M., Malcolm, S., Pantoya, M. L., & Davis, F. (2009). Impact Ignition of
Nano and Micron Composite Energetic Materials. International Journal of
Impact Engineering, 36(6), 842–846. doi:10.1016/j.ijimpeng.2008.11.011
Hunt, E. M., & Pantoya, M. L. (2005). Ignition dynamics and activation energies of
metallic thermites: From nano- to micron-scale particulate composites. Journal of
Applied Physics, 98(3), 034909. doi:10.1063/1.1990265
Jia, X., Gan, M., Williams, R. a., & Rhodes, D. (2007). Validation of a digital packing
algorithm in predicting powder packing densities. Powder Technology, 174(1-2),
10–13. doi:10.1016/j.powtec.2006.10.013
Kan, K., Stuhmiller, J. H., & Chan, P. C. (2005). Simulation of the collapse of an
underwater explosion bubble under a circular plate. Explosion, 12, 217–225.
Kappagantula, K. S., Clark, B., & Pantoya, M. L. (2011). Flame Propagation
Experiments of Non-gas-Generating Nanocomposite Reactive Materials. Energy
& Fuels, 25(2), 640–646. doi:10.1021/ef1012825
Kim, S. H., Kim, J., Kang, B., & Uhm, H. S. (2005). Superhydrophobic CF x Coating
via In-Line Atmospheric RF Plasma of He-CF 4 -H 2. Coating, (25), 12213–
12217.
Koch, E.-C. (2002). Metal-Fluorocarbon-Pyrolants: III. Development and Application
of Magnesium/Teflon/Viton (MTV). Propellants, Explosives, Pyrotechnics,
27(5), 262–266. doi:10.1002/1521-4087(200211)27:5<262::AIDPREP262>3.0.CO;2-8
91
Texas Tech University, Eric Collins, August 2013
Kuo, K. (1986). Principles of combustion. Retrieved from http://www.ulb.tudarmstadt.de/tocs/127150994.pdf
Kuwahara, T. (2004). Combustion and sensitivity characteristics of Mg/TF pyrolants.
Propellants, Explosives, Pyrotechnics, 202, 198–202. Retrieved from
http://onlinelibrary.wiley.com/doi/10.1002/prep.19970220403/abstract
Lesieutre, G. ., Ottman, G. ., & Hofmann, H. . (2004). Damping as a result of
piezoelectric energy harvesting. Journal of Sound and Vibration, 269(3-5), 991–
1001. doi:10.1016/S0022-460X(03)00210-4
Menon, S., & Lal, M. (1998). On the dynamics and instability of bubbles formed
during underwater explosions. Carbon, 16, 305–321.
Moore, K., Pantoya, M. L., & Son, S. F. (2007). Combustion Behaviors Resulting
from Bimodal Aluminum Size Distributions in Thermites. Journal of Propulsion
and Power, 23(1), 181–185. doi:10.2514/1.20754
Mutiso, R. M., Sherrott, M. C., Li, J., & Winey, K. I. (2012). Simulations and
generalized model of the effect of filler size dispersity on electrical percolation in
rod networks. Physical Review B, 86(21), 214306.
doi:10.1103/PhysRevB.86.214306
Neuber, A., Butcher, M., Hatfield, L. L., & Krompholz, H. (1999). Electric current in
dc surface flashover in vacuum. Journal of Applied Physics, 85(6), 3084–3091.
doi:10.1063/1.369647
Nixon, E., Pantoya, M. L., Sivakumar, G., Vijayasai, A., & Dallas, T. (2011). Surface
& Coatings Technology Effect of a superhydrophobic coating on the combustion
of aluminium and iron oxide nanothermites. Surface & Coatings Technology,
205(21-22), 5103–5108. doi:10.1016/j.surfcoat.2011.05.014
Nwafor, O. (2004). Effect of advanced injection timing on emission characteristics of
a diesel engine running on biofuel. International journal of ambient energy,
25(3), 115–122. Retrieved from
http://www.tandfonline.com/doi/abs/10.1080/01430750.2004.9674950
Pantoya, M., & Levitas, V. (2009). Effect of bulk density on reaction propagation in
nanothermites and micron thermites. Journal of Propulsion …, 25(2), 465–470.
Retrieved from http://cat.inist.fr/?aModele=afficheN&amp;cpsidt=21654681
Platt, S. R., Farritor, S., & Haider, H. (2005). On Low-Frequency Electric Power
Generation With PZT Ceramics. IEEE/ASME Transactions on Mechatronics,
10(2), 240–252. doi:10.1109/TMECH.2005.844704
92
Texas Tech University, Eric Collins, August 2013
Puri, P., & Yang, V. (2010). Thermo-mechanical behavior of nano aluminum particles
with oxide layers during melting. Journal of Nanoparticle Research, 12(8),
2989–3002. doi:10.1007/s11051-010-9889-2
Puszynski, J. a., Bulian, C. J., & Swiatkiewicz, J. J. (2007). Processing and Ignition
Characteristics of Aluminum-Bismuth Trioxide Nanothermite System. Journal of
Propulsion and Power, 23(4), 698–706. doi:10.2514/1.24915
Rai, A., Lee, D., Park, K., & Zachariah, M. R. (2004). Importance of Phase Change of
Aluminum in Oxidation of Aluminum Nanoparticles. The Journal of Physical
Chemistry B, 108(39), 14793–14795. doi:10.1021/jp0373402
Robinson, D. a., & Friedman, S. P. (2005). Electrical conductivity and dielectric
permittivity of sphere packings: Measurements and modelling of cubic lattices,
randomly packed monosize spheres and multi-size mixtures. Physica A:
Statistical Mechanics and its Applications, 358(2-4), 447–465.
doi:10.1016/j.physa.2005.03.054
Roozbahani, M. M., Huat, B. B. K., & Asadi, A. (2013). The effect of different
random number distributions on the porosity of spherical particles. Advanced
Powder Technology, 24(1), 26–35. doi:10.1016/j.apt.2012.01.006
Sandler, J. K. W., Kirk, J. E., Kinloch, I. a., Shaffer, M. S. P., & Windle, a. H. (2003).
Ultra-low electrical percolation threshold in carbon-nanotube-epoxy composites.
Polymer, 44(19), 5893–5899. doi:10.1016/S0032-3861(03)00539-1
Sandler, J., Shaffer, M. S. ., Prasse, T., Bauhofer, W., Schulte, K., & Windle, a. .
(1999). Development of a dispersion process for carbon nanotubes in an epoxy
matrix and the resulting electrical properties. Polymer, 40(21), 5967–5971.
doi:10.1016/S0032-3861(99)00166-4
Sarathi, R., Sindhu, T. K., & Chakravarthy, S. R. (2007). Generation of nano
aluminium powder through wire explosion process and its characterization.
Materials Characterization, 58(2), 148–155. doi:10.1016/j.matchar.2006.04.014
Schroeder, L. C., & Poirier, D. R. (1984). The mechanical properties of thermite welds
in premium alloy rails. Materials Science and Engineering, 63(1), 1–21.
doi:10.1016/0025-5416(84)90159-9
Serway, R., & Jewett, J. (2004). Physics for Scientists and Engineers (6th Editio.).
Belmont, CA. Retrieved from
http://books.google.com/books?hl=en&lr=&id=AttDBYgLeZkC&oi=fnd&pg=P
R5&dq=physics+for+scientists+and+engineers&ots=mDwUv9T_6&sig=K4Fujb9QuyhJLaJCD6Oa6ML8P00
93
Texas Tech University, Eric Collins, August 2013
Shi, Y., & Zhang, Y. (2008). Simulation of random packing of spherical particles with
different size distributions. Applied Physics A, 92(3), 621–626.
doi:10.1007/s00339-008-4547-6
Shigeta, M., & Watanabe, T. (2010). Growth model of binary alloy nanopowders for
thermal plasma synthesis. Journal of Applied Physics, 108(4), 043306.
doi:10.1063/1.3464228
Shih, T.-S., & Liu, Z.-B. (2006). Thermally-Formed Oxide on Aluminum and
Magnesium. Materials Transactions, 47(5), 1347–1353.
doi:10.2320/matertrans.47.1347
Shimizu, A., & Saitou, J. (1990). Effect of Contact Points Between Particles on the
reaction rate in the Fe2O3-V2O3 System, 38, 261–269. doi:10.1016/01672738(90)90431
Shudo, T., Nakajima, Y., & Futakuchi, T. (1999). Analysis of Thermal Efficiency in a
Hydrogen Premixed Spark Ignition Engine. ASME Advanced Energy Systems, 39,
350–354. Retrieved from
http://scholar.google.com/scholar?hl=en&btnG=Search&q=intitle:Analysis+of+T
hermal+Efficiency+in+a+Hydrogen+Premixed+Spark+Ignition+Engine#0
Sindhu, T. K., Sarathi, R., & Chakravarthy, S. R. (2008). Understanding nanoparticle
formation by a wire explosion process through experimental and modelling
studies. Nanotechnology, 19(2), 025703. doi:10.1088/0957-4484/19/02/025703
Sodano, H., & Magliula, E. (2002). Electric power generation using piezoelectric
devices. Proceedings of 13th International Conference on Adaptive Structures
and Technologies, 153–161. Retrieved from
http://books.google.com/books?hl=en&lr=&id=3k-ztDGERYC&oi=fnd&pg=PA153&dq=Electric+Power+Generation+using+Piezoelec
tric+Devices&ots=txy1lBiDKP&sig=TcuGHaKB7ulgaUFyiDsdm-hczg8
Spitzer, D., Comet, M., & Baras, C. (2010). Energetic nano-materials: Opportunities
for enhanced performances. Journal of Physics and …, 71(2), 100–108.
doi:10.1016/j.jpcs.2009.09.010
Stacy, S. C., Pantoya, M. L., Prentice, D. J., Steffler, E. D., & Daniels, M. A. (2009).
Nanocomposites for Underwater Deflagration. Advanced Materials, (October),
33–35.
Sullivan, K. T., Piekiel, N. W., Wu, C., Chowdhury, S., Kelly, S. T., Hufnagel, T. C.,
Fezzaa, K., et al. (2012). Reactive sintering: An important component in the
combustion of nanocomposite thermites. Combustion and Flame, 159(1), 2–15.
doi:10.1016/j.combustflame.2011.07.015
94
Texas Tech University, Eric Collins, August 2013
Sun, J., & Simon, S. L. (2007). The melting behavior of aluminum nanoparticles.
Thermochimica Acta, 463(1-2), 32–40. doi:10.1016/j.tca.2007.07.007
Sun, Juan, Pantoya, M. L., & Simon, S. L. (2006). Dependence of size and size
distribution on reactivity of aluminum nanoparticles in reactions with oxygen and
MoO3. Thermochimica Acta, 444(2), 117–127. doi:10.1016/j.tca.2006.03.001
Szymanski, S. F., Rowlette, P., & Wolden, C. a. (2008). Self-limiting deposition of
aluminum oxide thin films by pulsed plasma-enhanced chemical vapor
deposition. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and
Films, 26(4), 1079. doi:10.1116/1.2891258
Turns, S. (1996). An Introduction to Combustion, Concepts and Applications
McGraw-Hill. New York. Retrieved from
http://scholar.google.com/scholar?hl=en&btnG=Search&q=intitle:An+Introducti
on+to+Combustion:+Concepts+and+Applications#2
Valliappan, S., Swiatkiewicz, J., & Puszynski, J. a. (2005). Reactivity of aluminum
nanopowders with metal oxides. Powder Technology, 156(2-3), 164–169.
doi:10.1016/j.powtec.2005.04.027
Wang, L. L., Munir, Z. A., & Maximov, Y. M. (1993). Thermite reactions: their
utilization in the synthesis and processing of materials. Journal of Materials
Science, 28(14), 3693–3708. doi:10.1007/BF00353167
Wang, N., Wang, L., & Zhang, L. (2010). Effect of Aluminum Powder on Properties
of Emulsion Explosives in Underwater Explosion. In S. Li, W. Want, & Y. An
(Eds.), Progress in safety science and technology series (Vol. 8, pp. 2599–2605).
Science press Beijing.
Ward, T. S., Trunov, M. a., Schoenitz, M., & Dreizin, E. L. (2006). Experimental
methodology and heat transfer model for identification of ignition kinetics of
powdered fuels. International Journal of Heat and Mass Transfer, 49(25-26),
4943–4954. doi:10.1016/j.ijheatmasstransfer.2006.05.025
Weir, C. M., Pantoya, M. L., & Daniels, M. A. (2013). The role of aluminum particle
size in electrostatic ignition sensitivity of composite energetic materials.
Combustion and Flame. doi:10.1016/j.combustflame.2013.05.005
Weir, C., Pantoya, M. L., Ramachandran, G., Dallas, T., Prentice, D., & Daniels, M.
(2012). Electrostatic Discharge Sensitivity and Electrical Conductivity of
Composite Energetic Materials. Journal of Electrostatics, 71(1), 77–83.
doi:10.1016/j.elstat.2012.10.002
95
Texas Tech University, Eric Collins, August 2013
Williams, R., Dreizin, E., & Beloni, E. (2012). Ignition of Metal Powder Layers of
Different Thickness by Electrostatic Discharge. Journal of Propulsion and
Power, 28(1), 132–139. Retrieved from
http://arc.aiaa.org/doi/pdf/10.2514/1.56866
Xu, C. N., Akiyama, M., Nonaka, K., & Watanabe, T. (1998). Electrical power
generation characteristics of PZT piezoelectric ceramics. IEEE transactions on
ultrasonics, ferroelectrics, and frequency control, 45(4), 1065–70.
doi:10.1109/58.710589
Yeh, C. L., & Huang, Y. S. (2011). Thermite reduction of Ta2O5/SiO2 powder
mixtures for combustion synthesis of Ta-based silicides. Journal of Alloys and
Compounds, 509(21), 6302–6306. doi:10.1016/j.jallcom.2011.03.063
Young, G., Sullivan, K., Zachariah, M., & Yu, K. (2009). Combustion characteristics
of boron nanoparticles. Combustion and Flame, 156(2), 322–333.
doi:10.1016/j.combustflame.2008.10.007
Yuan, X. Y., Zhan, C. B., Jin, H. B., & Chen, K. X. (2010). Novel method of thermite
welding. Science and Technology of Welding and Joining, 15(1), 54–58.
doi:10.1179/136217109X12505932584691
Zhang, A., Yao, X., & Yu, X. (2008). The dynamics of three-dimensional underwater
explosion bubble. Journal of Sound and Vibration, 311, 1196–1212.
doi:10.1016/j.jsv.2007.10.007
96