A Thesis

entitled

Ignition and Combustion Characteristics of Nanoscale Metal and Metal Oxide Additives

in Biofuel (Ethanol) and Hydrocarbons

By

Matthew Jones

Submitted to the Graduate Faculty as partial fulfillment of the requirement for the

Master of Science Degree in Mechanical Engineering

____________________________________

Dr. Calvin Hong Li, Committee Chair

____________________________________

Dr. Matthew Franchetti, Committee Member

____________________________________

Dr. Yong Gan, Committee Member

____________________________________

Dr. Patricia R. Komuniecki, Dean

College of Graduate Studies

The University of Toledo

August 2011

Copyright 2011, Matthew Jones

This document is copyrighted material. Under copyright law, no parts of this document

may be reproduced without the expressed permission of the author.

iii

An Abstract of

Ignition and Combustion Characteristics of Nanoscale Metal and Metal Oxide Additives

in Biofuel (Ethanol) and Hydrocarbons

by

Matthew Jones

Submitted as partial fulfillment of the requirements for the

Master of Science Degree in Mechanical Engineering

The University of Toledo

August 2011

Metal energetic additives are added to propellants and explosives to improve

ignition and combustion performance. In particular, aluminum has been used as an

energetic material in solid-based propellant rockets and explosives for many years due to

its high combustion enthalpy and low cost. Recently, the introduction of nanotechnology

has led to significant developments in the field of energetic materials. Nanoscale

energetic materials, due to their surface area and unique thermal properties, are known to

exhibit many advantages over conventional micron sized particles. However, the current

mechanisms of nanoaluminum ignition and combustion are not fully understood.

Furthermore, studies involving suspensions of energetic nanomaterials in a liquid

medium (nanofluids) are very limited. A fundamental understanding of micron and

nanoscale aluminum combustion is critical to the design and implementation of practical

propulsion systems that use aluminum additives. Therefore, a comprehensive review on

the ignition and combustion of energetic nanoparticles was performed, with a primary

focus on aluminum, and two novel experimental studies were performed to investigate

the combustion characteristics of nanoscale aluminum (n-Al) and aluminum oxide (n-

Al2O3) in liquid fuels, namely, ethanol (C2H5OH).

iv

The first experimental study examined the heating values of several nanofluid

suspensions of n-Al (50 nm) and n-Al2O3 (36 nm) in ethanol. The primary objective of

this experimental study was to characterize the combustion reaction and gain a better

understanding of nanoaluminum oxidation in a multi-component heterogeneous system.

The heat of combustion was studied using a modified static bomb calorimeter system.

Combustion experiments were performed with volume fractions of 1%, 3%, 5%, 7%, and

10% for n-Al, and 0.5 %, 1%, 3%, and 5% for n-Al2O3. Combustion element composition

and surface morphology were evaluated using a scanning electron microscope and energy

dispersive spectroscopy system. The results indicate that the amount of heat released

volumetrically from ethanol combustion increases almost linearly with n-Al

concentration. Nanoaluminum volume fractions of 1% and 3% did not show

enhancement in the average volumetric heat of combustion, however, higher volume

fractions of 5%, 7%, and 10% increased the volumetric heat of combustion by 5.82%,

8.65%, and 15.31%, respectively. Aluminum oxide and heavily passivated n-Al additives

did not participate in combustion reactively, and there was no contribution from Al2O3 to

the combustion enthalpy in the tests. A combustion model that utilized Chemical

Equilibrium with Applications (CEA) was conducted as well and was shown to be in

good agreement with the experimental results.

Along with energy density enhancement, achieving precise control over the

reactivity of nanofluids is an opportunity for future nanoenergetic fuel applications. A

second experimental study involved the study in ignition probability of n-Al (50 nm) and

n-Al2O3 (36 nm) in ethanol and the commonly used No. 2 fuel oil (diesel). The primary

aims in this study were to study the effect of nanoenergetic additives on the ignition

v

probability of ethanol with a hot-plate setup, and to explore the underlying mechanisms

of ignition. Aluminum and Al2O3 were suspended in ethanol and diesel fuels from 0.1%

to 3% volume fractions, and dropped onto a hot plate at temperatures varying from

approximately 300 ºC to 600 ºC. The experimental probability was calculated and a

logistic regression approach was used to determine the 50% ignition threshold.

Nanoaluminum in ethanol was found to significantly increase the ignition probability:

ethanol suspensions with 1%, 2%, and 3% aluminum volume fractions ignited as much as

100 ºC lower than pure ethanol, and exhibited burning regimes similar to disruptive

combustion in slurry droplets, while ethanol suspensions with n-Al2O3 volume fractions

ignited at higher temperatures than pure ethanol. In addition, diesel mixtures with n-Al

and n-Al2O3 additives demonstrated relatively the same hot plate ignition probability as

pure diesel. Therefore, the likelihood of nanofluid ignition was strongly correlated with

volume fraction concentration, metal additive material, and surface tension (contact

angle).

vi

Contents

Abstract iii

Table of Contents vi

List of Tables ix

List of Figures x

List of Abbreviations xiv

List of Symbols xv

1 Introduction 1

1.1 Research Motivation……………………………………………….….............3

1.2 Research Objectives…………………………………………………………...5

1.3 Method of Approach……………………………………………………..…....6

2 Literature Review on Nanoenergetic Combustion 7

2.1 Combustion of Metal fuels ……………………………………………………7

2.1.1 Enthalpy of Combustion…………………………………………….8

2.1.2 Glassman‘s Criteria for Vapor-phase Metal Combustion……….....11

2.1.3 Kinetically-controlled vs. Diffusive-controlled Combustion…...…18

2.2 Micron and Nanoscale Aluminum Combustion……………………………..22

2.2.1 Characteristics of Nanoscale Materials……………………….……23

vii

2.2.2 Micron Aluminum Combustion……………………………………28

2.2.3 Nanoscale Aluminum Combustion …………………………..……34

2.3 Nanoenergetic Applications……………………………………………….…41

2.3.1 Nanocomposites…………………………………………………....42

2.3.2 Solid-based Composite Propellants……………………………..…43

2.3.3 Nanofluids……………………………………………………….....45

2.4 Summary……………………………………………………………………..49

3 Energetic Characteristics of Nanoscale Metal Additives in Biofuel 51

3.1 Experimental Methods……………………………………………………….54

3.2 Results and Discussion………………………………………………………65

3.2.1 Experimental Results…………………..…………………………..65

3.2.2 Thermodynamic Equilibrium Calculations………………………...71

3.3 Conclusions…………………………………………………………………..75

4 Ignition Characteristics of Nanoscale Metal Additives in Biofuel 78

4.1 Experimental Methods……………………………………………………….79

4.2 Results and Discussion…………………………………………………..…..89

4.2.1 Experimental Results………………………………………............89

4.2.2 Effect of Surface Tension……………………………………….…95

4.2.3 Heterogeneous Nucleation…………………….….…..…………....98

4.2.4 Theory of Thermal Droplet Ignition……………………………...101

4.3 Conclusions………………………………………………………………....104

5 Conclusions and Future Work 106

5.1 Concluding Remarks……………………………………………………......106

viii

5.2 Extensions of Research ………………………………………………….....107

References 110

Appendices

Appendix A: Sample CEA Output File for Ethanol in Air……….………….…124

Appendix B: Logistic Regression………………………………………..……..129

ix

List of Tables

Table 2.1 Thermodynamic properties of metals and their combustion products

in O2. Taken at 1 atm and initially at 298K….……………………....... 13

Table 3.1

Fuel properties of ethanol fuel and commercial diesel fuel……........

60

Table 3.2 Material properties of aluminum nanoparticle samples………………. 63

Table 4.1 Liquid fuel properties..………………………………………………... 84

Table 4.2 Additive material properties, as specified by the manufacturer…........ 84

Table 4.3 Experimental data recorded for the ignition experiments of pure

ethanol………………………………………………………………… 86

x

List of Figures

Fig. 2.1 Combustion enthalpy (heat of combustion) of energetic materials in

oxygen (O2)……………………………………………………………….. 8

Fig. 2.2 Heat of combustion of liquid fuels in O2..................................................... 9

Fig. 2.3 Calculated CEA adiabatic flame temperatures for Al in O2 and standard

air, stoichiometric reactants initially at 298 K………………..................... 15

Fig. 2.4 Calculated CEA adiabatic flame temperatures for Al in liquid water and

CO2, stoichiometric reactants initially at 298 K………………….............. 16

Fig. 2.5 Time-exposed free fall luminosity images of 215 µm Al particles in

mixtures of 21% O2 / 79% N2, 21%O2 / 79% Ar, and pure N2O, CO2,

CO. The side scale indicates distance (mm) from laser ignition…………. 17

Fig. 2.6 Schematic representation of the flame structures observed in aluminum

combustion…………………………………………………....................... 21

Fig. 2.7 Surface to bulk atom ratio for spherical iron crystals…………………….. 23

Fig. 2.8 Size dependence of melting points for Sn, and size dependent heat of

fusion for Sn……………………………………...………………………. 24

Fig. 2.9 Micron-sized aluminum vapor phase particle combustion……………….. 29

Fig. 2.10 Burning time of micron aluminum particles from 10 to 1000 μm……...... 31

xi

Fig. 2.11 Sequence of polymorphic phase transitions for micron aluminum

oxidation…………………………………………………………………... 33

Fig. 2.12 Aluminum ignition temperatures in literature………………………….…. 35

Fig. 2.13 Schematic of size-dependent self-heating rates against a constant external

heating rate………………………………………………………………... 36

Fig. 2.14 Nanoscale aluminum heterogeneous surface combustion………………... 37

Fig. 2.15 Levitas‘ diagram for nanoaluminum melting dispersion mechanism…….. 38

Fig. 2.16 Ignition temperatures for aluminum from various sources……………….. 39

Fig. 2.17 Proposed overall ignition stages of micro and nano-sized aluminum……. 40

Fig. 2.18 Combustion enthalpy for monomolecular materials and metal fuels…….. 43

Fig. 2.19 Competing effects for AP/HTPB/AL matrices, showing oxidizer/fuel

(O/F) diffusion flames and near-surface aluminum combustion…………. 45

Fig. 2.20 Microexplosion behavior of micron aluminum n-heptane slurries……….. 47

Fig. 3.1 Energy content of interest for select fuels and fuel additives in O2………. 55

Fig. 3.2 Schematic of the calorimeter system and combustion vessel…………….. 55

Fig. 3.3 Temperature vs. time data plot determined with thermocouple readings.... 56

Fig. 3.4 Calibration measurements for benzoic acid………………………….…… 59

Fig. 3.5 Experimental volumetric heat of combustion of pure ethanol fuel………. 61

Fig. 3.6 Experimental volumetric heat of combustion of sealed ethanol fuel…….. 61

Fig. 3.7 Experimental volumetric heat of combustion of diesel fuel……………… 62

Fig. 3.8 SEM images of n-Al and n-Al2O3 powder at 500 nm magnification……... 63

Fig. 3.9 Steric stabilization (left) and electrostatic stabilization (right) of additives

in a solution……………………………………………………………….. 64

xii

Fig. 3.10 a) Volumetric HoC of Eth + n-Al samples at 20 atm, and b) volumetric

and gravimetric HoC of Eth + n-Al samples……………………………... 66

Fig. 3.11 a) Volumetric HoC of ethanol + n-Al2O3 samples, and b volumetric and

gravimetric HoC of ethanol + n-Al2O3 samples………………………….. 67

Fig. 3.12 a) EDS spectra after combustion for Eth + 5% n-Al, and.b) EDS spectra

after combustion for Eth + 5% n-Al2O3…………………………………... 68

Fig. 3.13 SEM images of residual combustion products of Eth + 5% n-Al2O3 for a)

2.00 µm magnification and b) 500 nm magnification…………………….. 69

Fig. 3.14 SEM images of residual combustion products of Eth + 5% n-Al for 50

µm magnifications………………………………………………………... 70

Fig. 3.15 Calculated adiabatic flame temperatures for ethanol and aluminum

mixtures at stoichiometric conditions, for initial reactant temperatures of

298 K…………………………………………………………………….... 72

Fig. 3.16 Calculated adiabatic flame temperatures for ethanol and aluminum oxide

at stoichiometric conditions, for initial reactant temperatures of 298 K….. 72

Fig. 3.17 Calculated adiabatic flame temperatures for ethanol and aluminum

mixtures at stoichiometric conditions, for initial reactant temperatures of

298 K……………………………………………………………………… 73

Fig. 4.1 Schematic of experimental setup for the ASTM E 659 standard…………. 80

Fig. 4.2 Schematic of hot surface ignition operating parameters………………….. 80

Fig. 4.3 Schematic of the hot surface ignition experimental setup………………... 82

Fig. 4.4 Illustration of the hot surface ignition experimental setup……………….. 82

Fig. 4.5 Wire schematic of the hot surface ignition experimental setup…………... 83

xiii

Fig. 4.6 Low temperature distribution across the hotplate setup…………………... 83

Fig. 4.7 SEM image of particle size and distribution of pure Al nanoparticles

(left) and Al2O3 nanoparticles (right)……………………………………... 85

Fig. 4.8 Ethanol ignition data events, experimental probability, regression curve

and 95% confidence intervals…………………………………………….. 87

Fig. 4.9 Repeated ethanol tests over time, solid markers (■) designate the

regression fit, while empty markers the experimental probability (□)….... 88

Fig. 4.10 Ignition probability regression curves of aluminum/ethanol suspensions

compared to pure ethanol…………………………………………………. 89

Fig. 4.11 Regression curve with confidence band for pure Al in ethanol…………... 90

Fig. 4.12 Limits of ignition for pure Al in ethanol………………………………….. 90

Fig. 4.13 Regression curve with confidence band for Al2O3 in ethanol……………. 91

Fig. 4.14 Limits of ignition for Al2O3 in ethanol…………………...…..…………... 91

Fig. 4.15 Regression curve with confidence band for pure Al in diesel……………. 92

Fig. 4.16 Limits of ignition pure Al in diesel……………………………………….. 93

Fig. 4.17 Regression curve with confidence band for Al2O3 in diesel….…………... 93

Fig. 4.18 Limits of ignition for Al2O3 in diesel…………………………………….. 94

Fig. 4.19 Contact angles of ethanol/aluminum mixtures……………………………. 97

Fig. 4.20 Contact angles of ethanol and water for comparison….………………….. 97

Fig. 4.21 Schematic of a slurry fuel droplet, with regions A, B, and C….................. 98

Fig. 4.22 Characteristic ―S-curve‖ of ignition and extinction events……………….. 102

Fig. B.1 Experimental ignition data of pure ethanol (linear

model)…………………………………………………………………….. 131

xiv

List of Abbreviations

AFOSR… Air Force Office of Science Research

AIT… Auto Ignition Temperature

ALICE… Aluminum Ice Composite

AP… Ammonium Perchlorate

APS… Average Particle Size

ASTM… American Society for Testing and Materials

CEA… Chemical Equilibrium with Applications

CNT… Carbon Nanotubes

Da… Damkohler Number

DSC… Differential Scanning Calorimetry

DTA… Differential Thermal Analysis

EDS… Energy Dispersive X-Ray Spectroscopy

EPA… Environmental Protection Agency

FGS… Functionalized Graphene Sheet

HMX… Octogen

HoC… Heat of Combustion

HTPB… Hydroxyl-Terminated Poly-Butadiene

JANAF… Joint Army Navy Air Force

MDM… Melt Dispersion Mechanism

MIC… Metastable Intermolecular Composites

n-Al… Nanoaluminum

n-Al2O3… Nanoaluminum Oxide

NM… Nitromethane

SAM… Self-assembled Monolayer

SEM… Scanning Electron Microscopy

SHS… Self-Propagating High Temperature Synthesis

SRM… Solid Rocket Motor

SSA… Specific Surface Area

TEM… Transmission Electron Microscopy

TDI… Toluene Diisocyanate

TGA… Thermogravimetric Analysis

TNT… Trinitrotoluene

W… Watt

XPS… X-ray Photoelectron Spectroscopy

XRD… X-ray Diffraction

xv

List of Symbols

… Frequency factor

… Spalding transfer number

… Specific heat of surrounding gas

… System heat capacity

… Diffusivity coefficient

… Dimensionless Damkohler number

… Activation energy

… Mass consumption rate per unit area

… Gibbs energy of liquid

… Gibbs energy of vapor

… Effective latent heat of gasification ,

… Reference state of enthalpy at a temperature of 298K

... Available enthalpy

… Heat of combustion

... Heat of formation from a standard state of 298K

… Energy needed for droplet heating per unit mass of fuel gasified

… Reference state of enthalpy at a volatilization point

… Enthalpy of volatilization

… Stoichiometric fuel to oxygen ratio

K… Evaporation coefficient

… Knudsen number

… Heterogeneous specific reaction rate constant

… Specific latent heat of gasification

… Shell accumulated shell thickness

… Mass gasification rate

… Mass rate of fuel consumption

… Mass fraction (concentration) of oxidizer in the ambient stream

xvi

… Mass of vapor embryo

N… Gas concentration

… Ambient pressure, for liquid, for pressure of vapor

… Heat of combustion per unit mass

… Enthalpy of combustion

… Universal gas constant

… Activation temperature

… Boiling point

… Kinetic burning characteristic time

… Diffusive burning characteristic time

… Ambient temperature

… Stoichiometric fuel to oxygen ratio

… Volume of vapor embryo

… Average fuel molecular weight

… Pure fuel molecular weight

... Effective oxidizer concentration

… Ambient oxidizer mass fraction

… Standard normal probability density

… Statistical significance

… Beta heat transfer parameter

... Logistic regression coefficients for i = 1, 2, etc…

… Critical system Damkohler number

… Lower system Damkohler number

… Upper system Damkohler number

… Porosity

… Wetting contact angle

… Mean free path of a gas

… Thermal conductivity of air

… Logistic regression

… Ratio of the molecular mean free path length to the diameter

… Interfacial tension

... Solid volume fraction

… Free energy barrier to nucleate

1

Chapter One

Introduction

Metal additives have been utilized in solid propellants and fuels for some time

and have been shown to dramatically increase combustion enthalpies and quality. These

metalized additives in fuels combust to significantly heat up and expand the surrounding

gases, providing a larger specific impulse, the impulse per unit amount of propellant.

Common metallic materials of interest, including aluminum, boron, magnesium, and

zirconium [1], offer increases in the overall energy density of the fuel and effectively

reduce the tank storage volume. In the current state of the art implementation, energetic

additives offer a high volumetric enthalpy of combustion, facilitating transportation of

more payload per given fuel volume. However, given that the energetic additive sizes are

in the micron range and sometimes even the millimeter range, there are numerous side

effects to the combustion process, including ignition delays, slow burn rates, and

incomplete combustion of large (micron-sized) metal particles [2]. Furthermore, in

liquid-based fuels, a major challenge of dispersing metal additives is the settling of solid

particles in the fuel. Consequently, conventional liquid fuels may need to be remixed or

processed before use, due to the rapid settling of energetic additive particles.

2

This has led to new approaches to mitigate several of the disadvantages of larger metal

particle additions in fuels.

New advances in nanotechnology are being developed to understand and enhance

energy interactions in various fuel formulations. Fundamental research on nanomaterials,

such as nanoparticles suspended in fluids (nanofluids) and oxide reduction reactions

between nanoscale metals (superthermites), have led to entire new classes of energetic

materials. By taking advantage of the new functionalities offered by nanoparticles, the

reactivity and efficiency of propulsion vehicles has proven to be substantially improved.

For instance, nanoscale materials containing nanometer Al particles, can theoretically

release more than twice as much energy as the best molecular explosives [3]. Utilizing

aluminum nanoscale additives in solid-based composite propellants, it is possible to have

faster burning rates and more complete combustion [4, 5], thus improving the

performance of aluminized propellants. In liquid propellants, nanoparticles, due to their

high specific surface area, can be successfully suspended in fuels much longer than

micron-sized particles without precipitation out of the solution. In addition, recent studies

have reported reduced ignition delays in nanoaluminum suspended in ethyl alcohol and

JP-8 [6], increased catalytic activity in cerium oxide suspended in biodiesel [7], and

higher ignition probability of diesel with nanoaluminum additives [8] compared to the

base fuels. Based on these developments, the research on energetic nanomaterials has

become a topic of immense significance, and has led to a relatively new area of research

named ―nanoenergetics.‖ Current investigations are underway to further characterize the

energy contributions of nanoparticles, and to develop a unified understanding of the

relationships between nanoparticle additives and fuel performance.

3

1.1 Research Motivation

With recent advancements in nanoscale material production techniques, there has been

renewed interest in metal particle combustion and their energetic applications. The

energy consumed and cost to produce nanoparticles is reduced with the development of

new manufacturing technologies, offering nanoparticle additives that are increasingly

affordable [9]. As a result, it is now possible to routinely synthesize and characterize

nanoparticles, for instance, nanoaluminum is now commercially available as an

electroexplosively produced material named ALEX [10]. Furthermore, new low-cost

manufacturing techniques in functionalized graphene sheets have lead to their

consideration as alternatives to carbon nanotubes (CNT) [11]. Further development in

this area is expected to make these additives more available and affordable in the near

future. Presently, there are relatively few practical applications that are taking advantage

of nanoparticles ignition and combustion enhancements, and this opens up a large

window of opportunity for the research of new viable energetic fuels and their future

applications of nanoscale additives, including as advanced weapons, combustion

synthesis materials, and propellant enhancers [12].

There are many practical applications of nanoenergetic additives that have

attracted worldwide attention and given rise to new fuel formulations, however, there has

been considerable debate over the mechanism of nanoparticle ignition, and the conditions

at which nanoparticles burn. For example, considering aluminum, the most commonly

used metallic additive [13], current theoretical models cannot fully explain

nanoaluminum ignition in certain environmental conditions and particle size ranges. In

particular, bulk aluminum ignition is typically associated with the melting of the ambient

4

oxide layer at 1700-2400ºC, while the ignition temperatures of nanoscale aluminum has

been determined to be well below the bulk material value at 660 ºC [14]. A number of

experimental investigations on aluminum additive combustion have reported a wide

range of ignition temperatures even within the same particle distribution [14-17]. There is

also a critical diameter below which a vapor flame cannot be self-sustained; therefore

several investigators [18-20] have studied the conditions at which nanoparticles burn

diffusively or kinetically. Furthermore, the phenomena of the growth of the aluminum

oxide layer, effect of mechanical stresses or strains, and phase changes of the surrounding

oxide layer and core solid-liquid of core are not completely understood [21].

Consequently, more experimental studies are needed to fully characterize nanoaluminum

and along with other metals as nanoenergetic materials.

While there are a number of combustion enhancements resulting from the

addition of nanoparticles to gelled and solid-based propellants, there has been relatively

little investigative work on the combustion properties of nanoparticles suspended in

liquids (nanofluids). Nanoscale structures (< 100 nm) stably suspended in fluids give rise

to exciting new properties and phenomena. Previous studies [22, 23] have shown that the

addition of nanoparticles to liquids, such as water, may substantially improve the thermal

conductivity and mass transfer inside the liquid, even at low concentrations. For example,

it was found that carbon nanotubes suspended in oil enhance the thermal conductivity of

the mixture by as much as 250% [24]. This enhancement effect is dependent on

temperature [25], and thus provides enhanced thermal transport and heat sink capabilities

for fuels as well. Recent attention has focused on the potential large scale implementation

of nanoparticles as viable secondary energy carriers [9]. This concept proposes that pure

5

nanoenergetic materials or suspensions of nanoenergetic materials in a liquid medium can

be controllably ignited to provide a secondary release of thermal energy. The combustion

products can then be captured and reprocessed into their original nanoparticle form for

repeated use. Very few studies currently in press or within the last few years [6-8, 11, 26,

27] have investigated nanofluid combustion phenomena, and there is currently a need for

more insight into their behavior, particularly in alternative energy systems.

A projected decline in the world‘s oil production and concern of greenhouse gas

emissions have led to a recent push for increased propellant performance and

environmental sustainability. Current propellants in use, such as hydrazine, have many

negative health and environmental effects to human and animals. According to the U.S.

Environmental Protection Agency (EPA) [28], short term exposure to hydrazine can

result in permanent damage to the central nervous system, kidneys, and liver. In addition,

there is also concern about the occupational exposure of kerosene-based fuels, such as

military JP-8 and civilian equivalents Jet-A [29]. As a result, several alternative

propellants with nanoenergetic additives have been proposed such as graphene sheets or

nanoaluminum suspended in nitromethane [11, 30], and nanoaluminum suspended in

liquid or frozen water formulations [31-33]. Thus, nanoenergetics offer a range of

benefits, and it is of interest to explore other promising fuel candidates that minimize the

formation of greenhouse gases, ozone-forming pollutants, and harmful or toxic

emissions.

1.2 Research Objectives

The primary aims in this study were to study the effect of aluminum nanoenergetic

additives on the ignition and combustion characteristics of biofuel (ethanol). As a result,

6

this experimental study characterizes the combustion of commercially available

aluminum and aluminum nanoparticles to gain a better understanding of aluminum

oxidation in a multi-component heterogeneous system, and as a secondary energy carrier

suspended in ethanol. This study is one of the first studies investigating the combustion

nature of this fuel formulation. The basic combustion studies here may be extended to

more complex nanoenergetic systems, such as bimodal (combination of micron and

nanoscale) aluminum compositions, mechanically alloyed metals, or metastable

intermolecular composite (MIC) materials.

1.3 Method of Approach

The next chapter provides a brief introduction on nanoenergetic particle combustion,

particularly the work done on micron-sized and nanoscale aluminum, and the current

applications that utilize the unique features of the nanoscale additives. Chapter 3

describes a calorimeter study on the heat of combustion of the ethanol/aluminum fuel

formulations and their calculated flame temperatures, currently accepted by Nanoscale

Research Letters. Chapter 4 describes a hot plate ignition probability of

ethanol/aluminum and diesel/aluminum fuel formulations. Chapter 5 summarizes the

results and proposes future work. Finally, Appendices A and B are provided for

additional reference.

7

Chapter Two

Literature Review of Nanoenergetic Combustion

The combustion of nanoenergetics is a multidisciplinary subject, and due to their

propellant performance enhancing abilities, there has been a large body of research in

order to effectively model and understand the mechanism of nanoscale metal ignition and

combustion. A brief review of metal particle combustion is provided in section 2.1. As

shown in section 2.2, a better theoretical understanding of the thermodynamics and

reaction kinetics of micron-sized and nanoscale aluminum particles can lead to improved

combustion models and prediction capabilities. Ultimately, as reviewed in section 2.3,

this can lead to practical nanoenergetic applications that utilize the unique features of

micron and nanoscale energetic materials.

2.1 Combustion of Metal Fuels

Early works by von Grosse and Conway [34], and Glassman [35], serve as general

framework for current metal combustion models. The combustion of metal-oxygen

systems typically involves a significant release of energy, and the products of combustion

are usually condensed when at room temperature. Thus, the general equation for metal

8

( ) oxidation reactions is . Due to their high reactivity, they

are rarely found in their free (unbound) form in nature. Metals may react in the presence

of an oxidizer to form an oxide passivation layer - a thin 25Å (~2.5 nm) protective layer

that surrounds the metal particle, prevents spontaneous combustion, and impedes against

further oxidation of the metal. There are many of contributing factors that determine

which materials are best suited for a given application, including the enthalpy of

combustion (energy content), toxicity, chemical stability, cost and availability, and

properties of the metal and its oxide.

2.1.1 Enthalpy of Combustion

Metal additives are evaluated, in part, by the amount of energy they release for a given

mass or volume. As shown in Fig. 2.1, the heats of combustion for various energetic

materials are illustrated for comparison. Aluminum, boron, and beryllium offer the

Fig. 2.1: Combustion enthalpy (heat of combustion) of energetic materials in oxygen

(O2), compiled from Ref. [36].

9

highest energetic content in comparison to the other materials. Carbon has the advantage

of participating energetically, typically in the form of graphene sheets or carbon

nanotubes (CNT), without producing any residual solid oxide material [11]. In theory,

boron (a metalloid) has demonstrated to be an excellent energetic material, with the

highest volumetric heat of combustion of any known element. Furthermore, as shown in

Fig. 2.2, compared to hydrocarbon fuels boron has approximately 40% higher energy

density on a gravimetric basis and more than three times higher on a volumetric basis.

Fig. 2.2: Heat of combustion of liquid fuels in O2, compiled from Ref. [37].

Unfortunately, the practical implementation of boron-based propellants has thus far

resulted in poor combustion efficiency and low energetic performance [38]. The

combustion of boron is limited due to its inhibitive oxide coating, and ―energy trapping‖

due to the formation of HBO2 species within a hydrogen containing atmosphere [39]. As

a result, for the efficient combustion of boron particles, the oxide layer must first be

removed, and the formation of HBO2 reduced. Recent attention has focused on the

potential large scale implementation of nanoparticles as viable secondary energy carriers

[9]. This concept proposes that pure nanoenergetic materials or suspensions of

10

nanoenergetic materials in a liquid medium can be controllably ignited to provide a

secondary release of thermal energy. In addition to energetic content, a critical

consideration is the toxicity and chemical stability of the metal reactants and products.

The present concern for reduced toxicity significantly restricts the application of certain

metals. For instance, the high energy content of beryllium is particularly attractive;

however it is extremely hazardous to humans and animals even in small doses [9]. For

engineering applications, magnesium also has been used in fuel systems; though due to

its reaction with water and humid air, it is fairly unstable to handle. Other considerations

include the cost to produce, flame temperatures, material abundance, and burning

characteristics of the metal and its oxide, respectively.

Powdered spherical aluminum is the most widely used metal in energetic

materials [13], and the primary focus of the present study. Aluminum additives have been

utilized foremost in solid propellants and fuels for many years, and have been shown to

dramatically increase combustion enthalpies and quality. Aluminum is used due to its low

cost, high combustion enthalpy, high thermal conductivity, excellent surface absorption,

and low melting and ignition temperatures [13]. Furthermore, the main combustion

product of aluminum in oxygen, Al2O3, is relatively non-toxic, environmentally stable,

and may be recycled back to pure aluminum with an electrolytic reduction [9, 40]. As a

result, recently there has been a substantial amount of fundamental research in micron

[41, 42] and nanoscale aluminum [43], and also of applied research in aluminized gel

[30-32, 44], liquid [6-8, 11, 26, 27], and solid-based [4, 5, 33] propellant formulations.

There are many materials that have generated interest in regards to their

applications as energetic materials. Aluminum and magnesium have been used as

11

energetic additives to chemical rockets, mostly solid propellants, while boron has been

used for air breathing (jet engine) applications such as in ramjets [45]. When considering

energetic additives, there are many factors that determine which material to use. In

existing metalized propellants, aluminum is the leading additive used in combustion

systems due to its unique combination of low cost, commercial availability, and heat of

combustion. As a result, the primary focus of this work is on aluminum ignition and

combustion.

2.1.2 Glassman’s Criterion for Vapor-phase Metal Combustion

A significant portion of fundamental metal combustion research is whether a metal

particle will burn with a detached gaseous flame or with a surface reaction on the metal

surface. Consequently, the combustion process can be classified by the phase of the

reactants. If both the reactants are initially in the same phase, usually gases, then the

process is called homogeneous. Hence, a ―gas-phase chemical reaction‖ refers to a

homogenous reaction in which the fuel and oxidant are both gases. Heterogeneous

reactions are when the reactants initially exist in different phases, for instance a reaction

between an oxygen gas and solid coal. Furthermore, in some cases, it is possible to have

two combustion fronts, a heterogeneous one at the surface and a homogeneous one

detached. It is important to note that in some combustion literature, the terms

―heterogeneous‖ and ‗homogeneous‘ can also be defined as the interface (the location) at

which the chemical reaction occurs. To avoid misunderstanding, the terms heterogeneous

and homogeneous in this work will only denote the phase of the reactants and not the

location.

12

In liquid hydrocarbon combustion, the combustion process occurs in the gas

phase. The flame temperature exceeds the fuel saturation temperature (boiling point) at a

designated pressure, so that the fuel will vaporize and diffuse towards the oxidizing

atmosphere to react [35]. In contrast to liquid fuel droplet combustion, metal particles can

either burn homogenously as a vapor or heterogeneously as a liquid/gas surface

combustion process. In addition, metals can be classified by whether their combustion

products are volatile or non-volatile. According to Glassman‘s criterion for vapor phase

combustion of metals [35], the metal will burn in the vapor phase if the metal flame

temperature is greater than its boiling point (Tb). This flame temperature is a known

value, providing that the combustion products are non-volatile, and if the metal particle

meets this criterion, it will burn very much like a hydrocarbon droplet.

The volatility of a metal combustion product depends on the properties of the

metal and its oxide, and the available energy of the reaction. Since many metallic oxides

decompose (dissociate) into several smaller suboxides of the original metal oxide, a true

thermodynamic boiling point does not exist for most metal oxides [46]. For instance,

there is no gaseous aluminum oxide (Al2O3), only a variety of Al, O, and AlxOy species.

Rather, a ―vaporization-decomposition,‖ or ―volatilization‖ temperature exists, that is

more appropriately defined as a ―psuedo-boiling point.‖ If the enthalpy required to gasify

the metal oxide ( ), is greater than the available energy from the chemical reaction

( ), then at a given pressure the metal combustion product will be non-volatile, as

described in Eq. 2.1 below [35],

(2.1)

where is the enthalpy of combustion, and is the energy to bring the

13

metal to its volatilization temperature from an initial reference value of 298K. Analogous

to a boiling process, the metal oxide begins to decompose, and any further energy input is

absorbed by the metal oxide volatilization process without increasing the flame

temperature. Thus, for metal particle combustion with non-volatile product oxides, the

particle flame temperature can be assumed to be the metal oxide volatilization

temperature (Tvol).

On the other hand, if the energy available from the oxidation reaction is greater

than the heat of vaporization-dissociation of the metal oxide, then the combustion product

will be volatile. In this case, the flame temperature must be calculated, and compared to

the metal boiling point (Tb). Again, if the metal flame temperature is greater than its

boiling point (Tb), the metal will boil off and burn in the vapor phase; otherwise it will

occur as a heterogeneous surface reaction. As shown in Table 2.1, the thermodynamic

properties of metals and their oxides are listed. The boiling point of the metal Tb and the

Table 2.1:

Thermodynamic properties of metals and their combustion products in

oxygen. Taken at 1 atm and initially at 298K, adapted from Refs. [35, 47].

Material Tb (K) Combustion

Product Tvol (K)

(kj mol-1)

(kj

mol-1

) (kj mol-1

)

Aluminum

(Al) 2791 Al2O3 4000 1676 1860 2550

Boron (B) 4139 B2O3 2340 1272 360 640

Beryllium

(Be) 2741 BeO 4200 608 740 1060

Iron (Fe) 3133 FeO 3400 272 610 830

Lithium (Li) 1620 Li2O 2710 599 400 680

Magnesium

(Mg) 1366 MgO 3430 601 670 920

Silicon (Si) 3173 SiO2 2860 904 606 838

Titanium (Ti) 3631 Ti3O5 4000 2459 1890 2970

Zirconium

(Zr) 4703 ZrO2 4280 1097 920 1320

oxide volatilization temperature Tvol can be compared according to Glassman‘s

14

hypothesis, to determine the phase of combustion, whereas the energy required to heat up

to the volatilization temperature and then volatilize the oxide +

can be compared to the heat of formation to determine if the combustion

product is volatile. For example, for an aluminum-oxygen system initially at

stoichiometric conditions (1 atm and 298K), the heat to gasify Al2O3, 1860 kj mol-1

, is

more than the chemical energy released when the oxide is formed, 1676 kj mol-1

, hence

the oxide is regarded as a non-volatile product under those conditions. Since the flame

temperature is a known value limited to the metal oxide volatilization temperature, 4000

K, and this is larger than the metal boiling point, 2791 K, it is expected to burn in the

vapor phase. Boron, however, under the same conditions, is expected to have a volatile

combustion product B2O3 (g), and it does not necessarily have a limited flame

temperature, ultimately the calculated flame temperature would determine whether the

flame burns at the particle surface or as a vapor. In this context, for metal combustion in

pure oxygen at 1 atm, the metals Al, Be, Fe, Li, Mg, and Ti, would be expected to burn as

a vapor, while B, Si, and Zr would be expected to burn heterogeneously [47]. Regarding

Fe and Ti, the actual metal boiling points and oxide volatilization temperatures are close

enough to burn either way, due to thermal losses in the flame.

Before the availability of useful tools to calculate burning temperatures, such as

STANJAN [48] and NASA Chemical Equilibrium with Applications (CEA) [49] (also

known as the Gordon-McBride Computer Code), Glassman‘s criterion was used to

predict the combustion phase of metal reactions. With the development of

thermochemical databases such as JANAF [50] and computing machines, the adiabatic

flame temperature can be readily determined by equilibrium concentrations or the

15

minimization of free energy. In the case of high temperature metal-combustion reactions,

dissociation may occur of products back to reactants, and phase equilibrium needs to be

satisfied, hence equilibrium computations need to be performed to satisfy chemical

equilibrium. For CEA, its two combustion modules use a descent Newton-Raphson

numerical technique to minimize the Gibbs or Helmholtz free energy. As shown in Fig.

2.3, the adiabatic flame temperatures were calculated for aluminum in pure O2 and

standard dry air composition (N2 78.08%, O2 20.95%, Ar 0.94%, CO2 0.03%), with

reactants initially at 298 K.

Fig. 2.3: Calculated CEA adiabatic flame temperatures for Al in pure O2 and standard

air, stoichiometric reactants initially at 298 K, also seen in [35].

The aluminum boiling point, as shown as ―Al vapor,‖ will increase with pressure, which

can be validated by applying the Clausius–Clapeyron relation. Similarly, the adiabatic

flame temperatures will increase with increasing pressure due to less dissociation of

stable species according to Le Chatelier‘s principle. In addition, for most metal-oxygen

systems, with increased pressure there is a smaller enthalpy of volatilization ( ) for

16

the metal oxide, due to favorable equilibrium conditions of reduced dissociation [35].

Thus for isobaric combustion, aluminum metal is expected to burn in the vapor phase in

pure O2 and standard dry air composition, as predicted by Glassman‘s criterion. In

addition to total system pressure, the metal combustion phase can be influenced by the

form of the oxidizer. Depending on the oxygen reactant, the available enthalpy ( )

will be affected due to a change in the heat of reaction ( ), and thus, the phase of

combustion may change as well [47]. As shown in Fig. 2.4, stoichiometric aluminum

reactions with water and carbon dioxide initially at 298 K

are expected to start burning heterogeneously at approximately 1.5 atm and 5 atm,

respectively.

Fig. 2.4: Calculated CEA adiabatic flame temperatures for Al in liquid water and

carbon dioxide, stoichiometric reactants initially at 298 K, also seen in [35].

Experimentally, these temperatures and burning modes can be observed, Bucher reported

laser ignition of 215 µm aluminum particles burning in various atmospheres at room

17

temperature [51]. As shown in Fig. 2.5, combustion of Al in O2/N2, O2/Ar, and N2O

mixtures occur with a large envelope flame. While combustion of Al in CO2 and H2O

(steam) atmospheres have a reduced flame envelope reaction, whereas no flame envelope

exists in CO, indicating a transition to heterogeneous combustion. Furthermore, Legrand

et. al [52] observed pictures of laser ignition of Al in CO2, where the flame zone was very

near the surface, indicative of heterogeneous combustion.

Fig. 2.5: Time-exposed free fall luminosity images of 215 µm Al particles in mixtures

of 21% O2 / 79% N2, 21%O2 / 79% Ar, and pure N2O, CO2, CO. The side

scale indicates distance (mm) from laser ignition [51].

Therefore, the metal-combustion phase can change based upon the metal and metal oxide

properties and the available enthalpy from the reaction. A metal particle can be predicted

to burn with a detached gaseous flame or with a surface reaction on the metal surface, by

means of both experimental and numerical approaches, such as CEA. However, there are

some limitations to programs of this type. For instance, the CEA code specifies an output

based on a specific user input state and does not describe the reaction rate. Therefore, it

does not state how the reaction occurs, for example, if the chemical reaction occurs

18

instantly, slowly, or negligibly slow. Furthermore, it does not take into account the initial

metal oxide layer that is likely to be present. In addition, the particle size also plays a

critical role in determining the combustion phase. As a result, in the next section,

additional considerations of the chemical kinetics are required to determine the reaction

timescale and its effect on the combustion phase.

2.1.3 Kinetically-controlled vs. Diffusive-controlled Combustion

Combustion phenomena are understood to be controlled by two competing effects, the

reaction kinetics and the diffusion of chemical species and heat [53]. When the reaction

rate is slow compared to the rate of species and heat diffusion, there is adequate time to

reach spatial equilibrium. In this case, the reaction is said to be kinetically-controlled,

because the reaction rate controls the overall burning time. Examples of this combustion

regime include the initiation of combustion (ignition) and explosions. However, when the

reaction rate is fast compared to the rate of diffusion, spatial non-uniformities exist, and

the reaction can be characterized as diffusion-controlled. This regime is typical of candle

burning or fuel droplet burning. Furthermore, it is possible to have comparable kinetic

and diffusion rates, such as in premixed fuel burning. As a result, two reaction timescales

and an associated dimensionless ratio have been established as the current general theory

for metal particle combustion.

In diffusive-controlled burning, the burning regime is based on the classical ―D2‖

law for liquid droplet combustion in a quiescent atmosphere. First formulated by

Spalding [54] and Godsave [55], this approach states that the droplet lifetime is

proportional to the square of the diameter, and this holds experimentally for both droplet

19

evaporation and burning,

(2.2)

Where is the droplet diameter, the initial droplet diameter, the evaporation or

combustion coefficient, and the instant in time. By setting the final diameter to zero, the

total droplet lifetime can be described in terms of the initial diameter as,

(2.3)

Assuming infinitely fast kinetics with respect to heat and mass transport, the particle

mass consumption rate per unit area, , can be derived as [56],

(2.4)

where is mass rate of fuel consumption, the radius of the droplet, the oxidizer

density, the diffusivity coefficient, and is the Spalding transfer number. This is

exactly the expression used to describe liquid hydrocarbon droplet burning, except in the

case of metal combustion, takes the simple form of , due to no fuel volality.

Where, is the stoichiometric fuel to oxygen ratio, and is the mass concentration of

oxidizer in the ambient stream. After integration, the characteristic time scale for

diffusively controlled metal combustion is as follows,

(2.5)

Where and are the gas and particle density, respectively. Examining Eq. 2.5, the

mass diffusion coefficient is known to vary roughly as due to more molecules

being packed in a given volume, essentially hindering their movement. The gas density

( is proportional to pressure, therefore the product of is independent of pressure,

20

and thus, diffusive burning is said to be approximately independent of pressure. In

addition, because in the diffusively-controlled regime there is usually a flame limiting

temperature, it is said to be independent of the ambient combustion temperature. In other

words, it is normally unexpected that the flame temperature of a diffusively-controlled

particle is higher than its oxide volatilization point. While in kinetically-controlled

burning, i.e. - burning with fast diffusion, the particle mass consumption rate per unit area

can be written as,

(2.6)

where is the heterogeneous specific reaction rate constant for surface oxidation.

Similar to the diffusive case, after integration, the kinetic burning characteristic time is

derived as [56],

(2.7)

In this case, kinetic burning is said to be inversely dependent on pressure. To determine

which characteristic rate is dominant for a given particle, a dimensionless Damkohler

number ( ) is used, as follows,

(2.8)

indicating that for , the reaction is diffusively-controlled, the transition

point, and the reaction is kinetically-controlled. This relation predicts an inverse

relationship between particle size and pressure. Therefore, smaller particles at low

pressure favor kinetically-controlled conditions, and large particles at high pressures are

expected to burn diffusively. In general, the value of the number can vary across the

flame in accordance with the gas and particle temperature profiles and the change of the

21

particle diameter [57]. Thus, alternate regions of kinetic and diffusion particle

combustion modes may co-exist within the same flame sheet.

To illustrate this, as shown in Fig. 2.6, from left to right, there is a ―Diffusion-

Limited‖ combustion, in which there is a detached gaseous flame away from the particle

surface. In this case, the temperature peaks in the flame region, and is limited to the

Al2O3 volatilization or ―boiling‖ temperature. In the ―Transitional‖ schematic, the

diffusive rates become faster relative the kinetic rates and the oxidizer can diffuse closer

to the particle surface. In this case, a kinetic or a near surface diffusively controlled

regime may prevail. Finally, the ―Shrinking Core‖ schematic illustrates a kinetically-

controlled regime, where a fast heterogeneous reaction occurs, and the entire particle is

consumed in a heterogeneous fashion. This is expected to occur for very small

(nanoscale) particles.

Fig. 2.6: Schematic representation of the flame structures observed in aluminum

combustion [20].

Thus, in both vapor phase and heterogeneous combustion of metals, there are two known

22

competing characteristic timescales, kinetic and diffusive, and the number is used to

predict which timescale is present. A burn time relation of yields a classical

diffusive combustion, while yields a kinetically-limited flame [20]. The particle

enthalpy of combustion, combustion phase, and reaction timescales are all significant

features of aluminum combustion, particularly as the particle diameter is reduced to the

nanoscale. This leads to many practical applications of nanoscale materials, however, it is

also important to understand the underlying cause. In the next section, the unique features

of nanoparticles are explored and the current theoretical understanding of micron-sized

and nanoscale aluminum particles is briefly reviewed.

2.2 Micron scale and Nanoscale Aluminum Combustion

Nanoscale energetic materials, due to their surface area and unique thermal properties,

are known to exhibit many advantages over conventional micron sized particles. New

combustion regimes are being reported that have significant practical applications. There

has been a good amount of research on micron-sized aluminum combustion, as recently

reviewed by Beckstead [41]. However, for nanoscale aluminum combustion, there

currently has been considerable debate over the mechanism of nanoparticle ignition, and

the exact conditions under which the particle burns. Similar to very fine coal particles,

nanoscale aluminum has the ability to burn on its surface. Nanoaluminum synthesis and

combustion have been reviewed and describe the first nanoaluminum applications being

explored [1, 3, 13, 47]; however, there have been recent advances that deserve attention.

23

2.2.1 Characteristics of Nanoscale Materials

Nanoscale energetic materials are known to exhibit many advantages over conventional

micron sized particles. Several thermodynamic and heat transfer properties can be

changed, primarily attributed to an increase in surface-to-bulk atom ratio and specific

surface area. This high specific surface area leads to more surface available to react, and

thus leads to higher reactivity. As the size of a bulk material approaches the nanoscale, its

surface area will increase with respect to its volume. As shown in the Fig. 2.7, a reduction

in particle size has the ability to increase the surface atom to bulk atom ratio for spherical

iron nanocrystals [58].

Fig. 2.7: Surface to bulk atom ratio for spherical iron crystals [58].

Atoms at the surface of a particle are exposed to their environment and are in a state of

higher energy than interior bulk atoms [59]. Consequently, the cohesive energy, or the

energy required to loosen an atom from the surface, is decreased with decreasing particle

sizes. These surface atoms are more loosely bound than the bulk atoms, and surface

tension effects for nanoscale solid materials are no longer negligible. In addition, due to

their small diameters, the temperature uniformity within the particles is enhanced, and the

24

particles quickly reach equilibrium with their surroundings. Therefore, the non-

dimensional Biot numbers are usually below unity, and Fourier numbers increase [47].

As a result, nanoscale particles will often exhibit different properties and behaviors than

larger particles of the same substance.

In differential scanning calorimetry (DSC) and nanocalorimetric measurements, it

was discovered as the size of particles decrease, there is a corresponding decrease in

melting point temperature and heat required for melting (heat of fusion). As a result, there

has been considerable research interest in melting point depression and size-dependent

heat of fusion phenomenon. This phenomenon encompasses a wide range of materials, as

shown in Fig. 2.8.a, for tin (Sn) nanoparticle diameters ranging from 50 nm to 5 nm, the

melting point temperatures were reduced from 232°C to approximately 160°C. Also

illustrated in Fig. 2.8.b, the normalized heat of fusion for Sn can decrease as much as

70% [60].

Fig. 2.8: a) Size dependence of melting points for Sn, and b) size dependent heat of

fusion for Sn [60].

For experimental studies with nanoaluminum, first systematically studied by Eckbert et.

al [14], a wide range of ignition temperatures have been observed. Lai et al [15]

performed nanocalorimetric measurements and found a decrease in melting temperature

25

up to 140°C. Comparing to Lai and Eckbert‘s results, Sun and Simon [16] performed a

DSC study and found a reduced magnitude of melting point depression, proposing that

the compressive force of the oxide layer increased their melting point values. After using

an integrated form of the Clausius–Clapeyron equation, the effect of the compressive

pressure on the melting point temperature was calculated, and a corrected value was

obtained. Dreizin‘s group [17] performed differential scanning calorimetry on

nanoaluminum powders and found that nano-aluminum melting temperatures were much

lower than bulk value melting temperatures in all cases. They also found that data

between melting models that predict DSC signals and experimental results only

qualitatively matched. Furthermore, a melting point study based on molecular dynamics

suggested that Al clusters of 1 nm can ignite as low as 400 K (127° C) [61].

Presently, experimental DSC responses and predictive theoretical melting models

have met with limited success. Theoretically, size-dependent melting point depression

can be described using variants of the classical Gibbs-Thompson model [16], with the use

of different models based on the selected assumptions. With this model, it is assumed

that the heat of fusion decreases with particle size due to the increase in the solid-liquid

surface energy (surface tension). Although the melting phenomena are not fully

understood yet, it is suggested that the greater mobility of surface atoms and a decrease in

cohesive energy partly determines the fall or rise of the melting point of a nanosolid [62].

Authors have also suggested that the change in melting temperature is triggered by

increasing fraction of lattice defects or surface irregularities with decreasing particle size

[16], or metal/metalloid impurities [21]. Furthermore, a recent review reported that

discrepancies between the particle surface energies and those found in the literature

26

suggest that the conditions under which the Gibbs-Thompson equation is used needs to

be reconsidered [63]. As a result, there may be aspects of the melting process that remain

to be taken into account and there is a need for more detailed experimental analysis with

corresponding theory that is accurate on the nanoscale.

When developing models for nanoparticle combustion, it is also important to take

into consideration the mean free path of air with respect to the particle diameter. For

instance, the mean free path of air at ambient pressure is approximately m

(~66 nm) [64]. As will be shown, micron-sized aluminum particles tend to burn with a

flame stabilized around the particle. Because micron particles are much larger than the

mean free path of air, they can sustain a detached gaseous flame. However, when

developing combustion models on the nanoscale, it is possible that particles are smaller

than the mean free path of air. As a result, there may be a critical diameter below which a

vapor flame cannot be self sustained. The Knudsen number below is used to determine

whether the continuum assumption, that particles can be considered macroscopic objects

in a continuous gas, is valid [56],

(2.9)

where is the mean free path of the surrounding gas, and is representative length scale,

in this case the particle diameter. The mean free path, or the average distance of a particle

between collisions, can be defined as,

(2.10)

where is the ratio of the molecular mean free path length to the diameter, and N is the

gas concentration. In a free molecular regime , the nanoparticles can be

27

considered large burning particles, however, nanoparticles are typically smaller than the

mean free path, and a continuum burning model may not be valid. Therefore in

combustion modeling, there are three regimes that may be considered, that being a

continuum regime, free-molecule regime, and a transitional regime [65].

Other considerations for nanoscale additives are presence of an ambient metal

oxide layer, which is significant in determining the fuel additive energy potential.

Depending on the metal and layer thickness, the enthalpy of combustion and combustion

rate may be decreased, or oxidation may be inhibited altogether. For nanoparticle

diameters of 20 nm or less, the oxide layer becomes a more prominent part of the particle

volume, accounting for approximately 60% energy loss per volume [66]. Therefore, it is

expected that nanoparticles release more energy when taken as a whole, due to more

complete combustion; however, there may be a lower limit of the particle diameter on the

energetic enhancement of nanoparticles. As a result, to prevent any oxidation, the metal

must be first be synthesized and then stored in an inert atmosphere, such as argon.

Furthermore, post-processing techniques can selectively coat the outer layer with another

material altogether, such as with organic self-assembled monolayers (SAMs). For

example, a recent study investigated alkyl-substituted epoxide capping on the surface of

aluminum, and determined that the particles were effectively passivated, but were no

longer pyrophoric [66]. Nonetheless, the development of surface modifying and synthesis

techniques needs to be considered when utilizing nanoscale materials.

As will be described more in section 2.4, these unique configurations of increased

surface to volume ratio, temperature uniformity, melting point depression, and oxide

layers have many practical applications in nanoenergetic materials. An increase in surface

28

area can lead to increased chemical reactivity, allowing more fuel to be in contact with

the oxidizer. For instance, catalytic activity is essentially a surface area phenomenon, and

the active surface at which the reaction takes place is increased when the surface to

volume ratio is increased. Gold, being a noble metal, is normally regarded as chemically

inert, however, for particles 5 nm or less, gold was found to be catalytically active [67].

In particle thermal ignition, which may be defined as when the particle self-heating rate is

greater than the external heating rate, smaller particles self-heat more effectively,

therefore it is proposed the particles may ignite at lower temperatures under a constant

heating rate. Furthermore, physical properties of nanoparticles such as melting point are

important because they heavily influence the ignition temperature and oxidation of

nanoscale metals. Finally, the reaction time scales can change due to smaller diffusion

distances, and the close proximity of the fuel and oxidizer may reduce the diffusion

length and increase the reaction (burning) rate. Therefore, the experimental combustion

nature of a micron versus nanoscale metals, specifically aluminum, is described in the

following two sections.

2.2.2 Micron Aluminum Combustion

As predicted by Glassman‘s criterion of vapor phase combustion, aluminum is expected

to burn as a vapor because the boiling point, or vaporization-dissociation point, of

aluminum oxide is higher than the metal‘s boiling point. The particle is expected to

closely follow the diffusively-controlled D2 liquid hydrocarbon droplet combustion;

however, it cannot be directly extended to model aluminum combustion without several

modifications. The fact that the final diameter does not reduce to zero, the condensation

29

of an oxide cap and actual deposition of the oxide cap on the particle, and limited flame

temperature are three key modifications that need to be considered when modeling

aluminum. As mentioned, there are exceptions when micron aluminum will continuously

burn in a kinetically-controlled heterogeneous reaction, such as in CO atmospheres [52].

Nevertheless, in gaseous micron aluminum combustion, the ignition event is typically

associated with melting or cracking of the ambient oxide layer at approximately 2350 K

(2070 ºC), that protects the metal. In the case of cracking, the particle is first heated up,

and the core aluminum melts to form liquid aluminum. Liquid aluminum has a density of

2.4 (g/cm3) and solid aluminum has a density of 2.7 (g/cm

3), therefore there is an

approximate 12% volumetric expansion of the core, leading to cracking [68]. Ignition is

then initiated on the exposed aluminum with a heterogeneous surface reaction, the

interior metal begins to evaporate and diffuse towards the surrounding oxidizer, and the

product aluminum oxides (primarily ALO) condense as an oxide cap (Al2O3) on the

particle. The particles then usually burn with a detached diffusion flame as shown in Fig.

2.9.

Fig. 2.9: Micron-sized aluminum vapor phase particle combustion [69].

30

The maximum flame temperature is then maintained at the vaporization-dissociation

temperature of the oxide, until the entire oxide cap is dissociated. Consequently, the

flame temperature is predetermined by the vaporization–dissociation temperature of the

aluminum oxide, and the surrounding environment temperature exerts a small influence

on the particle burning time [20]. Overall, there are approximately 16 different

elementary reactions (surface, gas-phase, disassociation, and condensation) within the

presence of H2O, CO2, and O2 oxidizers [41].

Dreizin characterized the entire process of micron aluminum combustion, for

free-falling droplet diameters of 85, 120, 165, and 190 μm, in air [42]. There were three

proposed distinct phases of aluminum ignition in air. The first stage comprised of

spherically symmetric vapor phase combustion with a smoke cloud, modeled by the

conventional vapor phase droplet combustion theory. Subsequently, there is a

nonsymmetrical particle combustion characterized by particle spinning and an increase in

size and density of the smoke cloud. Finally, there is a formation of an oxide cap and

decrease in particle temperature. Experimentally, one common prediction of the metal

burning mode is to plot the burning time against the particle diameter, and fit a power law

relation with an exponent [20]. As shown in Fig. 2.10,

31

Fig. 2.10: Burning time of micron aluminum from 10 to 1000 μm, from Ref. [41].

Beckstead [41] compiled micron burning times from over 400 datum points and ten

sources to illustrate that micron-sized particles burn in the diffusive (D2) regime,

however, the effect of condensation of the oxide cap on the particle, and the fact that the

particle diameter does not reduce to zero, slightly reduce the exponent to around 1.5 or

1.8. Based on the combustion data from over ten different sources, Beckstead proposed

an empirical relation for burning times of aluminum particles,

where,

(2.11)

32

for

for

and an effective oxidizer concentration, , is expressed as a mole fraction ( ) of O2,

H2O, and CO2,

with O2 being approximately twice more effective as an oxidizer than H2O and about five

times more effective than CO2. Thus, Eq. 2.11 is the current expression commonly used

to describe micron aluminum combustion up to the current time.

Also in micron aluminum research, thermogravimetric analysis (TGA) has

identified an evolution of aluminum oxide material properties according to the

surrounding temperature rise. The natural amorphous aluminum oxide may exist in

several different crystalline forms (polymorphs) that change according to its temperature

rise, that being amorphous → γ → δ → θ → α-Al2O3 [70], leading to a step-wise particle

mass increase. Simultaneous differential thermal analysis (DTA) and thermogravimetric

analysis can determine the mass increase due to aluminum oxidation according to

temperature, and X-ray diffraction (XRD) patterns of samples oxidized determine the

polymorph phase. As shown in Fig. 2.11, Trunov et. al [71] identified four different

oxidation stages (I, II, III, and IV), in the first stage, as the metal is heated, the natural

amorphous alumina layer grows until it reaches a critical thickness (approx. 5 nm), and

then the oxide layer fractures and transforms into a crystalline γ-alumina phase.

33

Fig. 2.11: Sequence of polymorphic phase transitions for micron aluminum [1].

In the second stage, the γ-alumina oxide layer increases in density and molten aluminum

leaks through the γ-alumina faults, growing into the third stage as one of the similar

intermediary transitions such as δ or θ. In the final polymorph stage, the oxidation rate

increases and the crystalline structure becomes significantly dense as α-alumina.

Furthermore, it also was qualitatively proposed that these phase transformations are

responsible for the wide range ignition temperatures of nanoscale aluminum that are

currently being reported, as described in the next section.

In general, the simple hydrocarbon droplet combustion model can be extended to

model aluminum particle combustion; however, several modifications need to be

considered. Aluminum either ignites due to cracking under thermal stress or melting of

the oxide shell. As mentioned in previous experiments, an oxide cap develops and the

particle diameters do not reduce to zero, thus the exponent for the D2 law is reduced to

1.5 or 1.8. Very detailed computational models have been formulated, that are highly

dependent on the assumptions used. Beckstead‘s model [72] shows good agreement

34

between the model and available experimental data. However, most of these models

assume gaseous combustion with infinite kinetics, when aluminum approaches the

nanoscale, kinetics can have a much larger influence and both gas and surface phase

reactions need to be considered for an accurate representation. Nonetheless, even for

micron particles, some aspects, such as the physics of oxide condensation, are not

completely understood [35].

2.2.3 Nanoscale Aluminum Combustion

In nanoscale aluminum combustion, the conditions at which particles combust and the

cause of ignition are still unclear. Due to limited information on the burning of

nanoparticles, there is no current suitable model [18]. Due to their small scale, diffusion

becomes faster relative to the reaction rates, and it is expected to be a kinetically-

controlled regime. Similar to very fine coal particles, nanoscale aluminum has the ability

to burn on its surface. In this case, a fast heterogeneous reaction occurs, and the entire

particle is consumed in a heterogeneous fashion. In contrast with bulk aluminum ignition,

a number of experimental investigations on nanoaluminum combustion have reported a

wide range of ignition temperatures, even within the same particle distributions, as shown

in Fig. 2.12,

35

Fig. 2.12: Aluminum ignition temperatures in literature, sources are provided in Ref.

[71].

Thus, the ignition temperatures of nanoscale aluminum have been determined to be near

the aluminum melting point at 660 ºC, well below the bulk material value of 1700-

2400ºC. Similar to micron-sized ignition, the ignition event is typically associated with

melting or cracking of the ambient oxide layer. As mentioned at the end of last section,

Trunov et al. [71] suggested that the wide range of nanoparticle combustion is a result of

the sequence of four polymorphic phase transformations. A qualitative analysis suggested

that, within the four-stage oxidation, different particle self-heating rates were responsible

for the range of ignition temperatures, as shown in Fig. 2.13.

36

Fig. 2.13: Schematic of size-dependent self-heating rates against a constant external

heating rate [71].

Three different self-heating curves (fine, intermediate, and course particles), can intersect

with an external heating rate that is determined by the thermal analyzer. For a constant

external heating rate (―particle heating rate‖), smaller particle ranges are expected to

trigger the transition to the second oxidation stage (γ-alumina) at lower temperatures;

however, this transition to the second stage can be delayed under higher heating rates.

Again, thermal ignition may be defined as when the self-heating rate is greater than the

external heating rate [35]. Therefore, nanoparticles self-heat more efficiently, and this

may be a contributing factor to ignition.

Rai et al. [73] proposed that aluminum nanoparticle oxidation occurs in two

distinct regimes. At temperatures below the melting point of aluminum, a slow oxidation

occurs with oxygen-limited diffusion through the aluminum oxide shell. At temperatures

above the melting point of aluminum, a fast oxidation occurs with both aluminum and

oxygen diffusing through the oxide shell, followed by a hollowing of the aluminum core

37

at temperatures above 1000 °C. In the case of oxidation through the shell, it is a surface

phenomenon with transport of reactants through the oxide shell, as shown in Fig. 2.14.

Fig. 2.14: Nanoscale aluminum heterogeneous surface combustion, where the ―surface‖

is between ―Liquid Al‖ and ―Al oxide‖ [69].

In the case of shell rupturing, once the integrity of the surface coating is destroyed, the

exposed aluminum undergoes a direct kinetically-controlled surface reaction. Due to

higher curvature of the particle, the oxide shell is in tension and the aluminum core is in

compression, and this results in easier mechanical rupturing of the coating compared to

large particles [73]. However, under very high heating rates, a new fast oxidation

mechanism, named the melting dispersion mechanism (MDM), was discovered for

nanoaluminum particles in shock-tube experiments under heating rates in the order of 107

°C/ s [74, 75]. The change in volume due to fast melting of the nanoaluminum core

induces of pressures of 0.1–4 GPa, and causes spallation of the oxide shell, as shown in

Fig. 2.15.

38

Fig. 2.15: Levitas‘ diagram for nanoaluminum MDM [74], also observed in Ref. [75].

This and several other models cannot currently be modeled by diffusion oxidation

models. Furthermore, the phenomena of the growth of the aluminum oxide layer, effect

of mechanical stresses or strains, and phase changes of the surrounding oxide layer and

solid-liquid of the core are not completely understood [21].

Consequently, several investigators [18-20] have studied the conditions at which

aluminum particles burn diffusively or kinetically. As the size of the particle is reduced,

there is a transition at which diffusively-controlled burning particle will start to burn

kinetically-controlled. A formal study of nanoaluminum oxygen found that a transition is

expected to occur around 10 μm for a pressure of 8.5 atm [20]. In this case, Beckstead‘s

Eq. 2.11 in the last section is no longer valid for very small diameters. Similar to

Beckstead‘s expression, Huang [18] proposed the following kinetic expression, for the

39

burning of aluminum dusts in air,

(2.12)

Where is in cm, , , and is the universal gas constant.

Comparing Huang‘s model to the model for micron aluminum, experimental

data on the burning time can be plotted, as shown in Fig. 2.16.

Fig. 2.16: Ignition temperatures aluminum from various sources, sources are provided

in Ref. [18].

This illustrates that a transition between the and burning regimes is expected to

be anywhere for diameters of approximately 6 μm to 20 μm. However, there is still little

information on the conditions that this transition occurs. Several different techniques are

used to determine this transition, one being the empirical power law fit of burning time,

and others include measuring the effect of pressure and temperature on the burning time.

While diffusively-controlled particles burn at an approximately known temperature and

40

are independent of pressure, kinetically-controlled particles are known to have a high

dependence on the atmospheric temperature and pressure. As shown in Fig. 2.16, it can

be seen that nanoparticles burning at 3500 K have much shorter burning times than 1500

K or 2000 K, further suggesting that they burn kinetically-controlled. A more direct

approach is to observe if any gaseous phase combustion products exist. Therefore,

emission spectrography of flame fronts can determine which species are present, for

instance, significant amounts of the vapor product AlO in the surrounding gas-phase

indicates there is a diffusively-controlled mechanism [76]. Furthermore, when oxidizer

availability is changed, in other words, the diffusion coefficient is changed; this

should affect diffusive burning and not have an influence on kinetic burning [57].

Taking micron and nanoscale aluminum into consideration, the entire aluminum

combustion process can be divided into five major stages [69], as shown in Fig. 2.17.

Fig. 2.17: Proposed overall ignition stages of micro and nano-sized aluminum [69].

41

Starting with stage 1) there is particle heating with phase transformations, 2) core melting

due to cracks or melting of oxide layer, 3) heterogeneous reactions or healing of cracks,

4) particle consumption due to heterogeneous reactions for nanoscale particles, or melting

of oxide layer to form an oxide cap for micron-sized particles, and finally 5) a detached

vapor phase flame front for micron-sized particles. The critical stage is between stages 3)

and 4), in which the dominant characteristic rate determines whether the entire reaction

will be kinetically-controlled or switch to diffusively-controlled. Although many studies

provide useful information on the behavior of nanoaluminum combustion, further

experimental studies are needed to fully characterize nanoaluminum in certain

environmental conditions and particle size ranges. Furthermore, there are relatively few

practical applications are taking advantage of the unique characteristics of the nanoscale,

most notably the higher specific surface area, higher reactivity, shorter burning times,

and overall more complete combustion. In particular, the system design of new energetic

or fuel propellants can benefit from these characteristics.

2.3 Nanoenergetic Applications

There are many practical applications to nanoenergetic additives that have attracted

worldwide attention and given rise to new fuel formulations. Because aluminum has been

the most applied and experimented metal additive, the focus of this section will be on

aluminum. With recent advancements in nanoscale aluminum production techniques,

there has been renewed interest in metal particle combustion and their energetic

applications. Presently, there are relatively few practical applications are taking

advantage of nanoparticles ignition and combustion enhancements, and this opens up a

42

large window of opportunity for the research of new viable energetic fuels and their

future applications of nanoscale additives, including as advanced weapons, combustion

synthesis materials, and propellant enhancers [12]. Due to the relatively limited studies

involving nanoenergetics suspended in liquid fuels, other currently applied formulations

such as condensed metal reactions and solid-based composite fuel applications are also

reviewed.

2.3.1 Nanocomposites

The combination of decreased particle sizes, diffusion distances, and increased surface to

volume ratio have lead to a new class of condensed metal reactants called

nanocomposites. Also called thermites, or metastable intermolecular composite (MIC)

materials, they are a powdered mixture between a metal (fuel) and metal-oxide (oxidizer),

and can combust in an exothermic reduction-oxidation reaction between a metal and a

metallic or non-metallic oxide. The material is most commonly mixed with aluminum;

hence, a thermite reaction is sometimes called an aluminothermic reaction. However,

other common metals used include but are not limited to Mg, B, Ti, and Zr, along with

metallic oxides Fe2O3, CuO, MoO3, and C2F4 [47] For example, the classic thermite

reaction used is ferric oxide and aluminum, as shown below,

2 (2.13)

is commonly applied in welding. Monomolecular materials such as trinitrotoluene (TNT)

and octogen (HMX) are typically limited in their heat of combustion compared to metal

reactions, as shown in Fig. 2.18 [1].

43

Fig. 2.18: Combustion enthalpy for monomolecular materials and metal fuels [1].

Conventional micron-sized thermites are also limited due to their low rates of energy

release. However, for these nanocomposite materials, also known as ―super-thermites,‖

the combustion rate, velocity, and energy density can be controlled. In particular, the

unique combination of a high rate of energy release and heat of combustion make these

formulations attractive. For example, these nanoenergetic materials offer more than twice

the energetic release than traditional monomolecular materials [3]. Furthermore, it is

possible to add a relatively small amount of nanocomposite to micron-aluminum in order

to dramatically increase its burning rate, with little overall change to combustion

enthalpies [77]. This has many applications in materials processing, such as self-

propagating high temperature synthesis (SHS), and in explosives and military

applications as well.

2.3.2 Solid-based Composite Propellants

There also have been many new developments in solid based propellant formulations due

to nanoenergetic additives. For solid fuel chemical rockets, an entire block of premixed

44

fuel and oxidizer propellant, called a grain, is stored within the combustion chamber. The

current standard used solid based fuel-oxidizer system in solid rocket motor (SRM)

applications is an aluminum fuel mixture with an ammonium perchlorate (AP) oxidizer.

An elastomeric binding agent solidifies the fuel-oxidizer mixture, most commonly

hydroxyl-terminated poly-butadiene (HTPB), and curing agents are considered for

processing, such as epoxy thermosetting toluene diisocyanate (TDI). In addition, various

catalysts such as ferric oxide and titanium dioxide can be used to accelerate burning. In

this mixture (matrix), aluminum typically reacts with combustion products of the

energetic oxidizer and binder, such as H2O and CO2, at its volatilization temperature [35].

For solid aluminum metalized propellants, two large sources of combustion inefficiency

are the accumulation of slag, which are residue aluminum products, and incomplete

combustion. Conventional aluminum particles increase the chamber temperature, in order

to provide a larger specific impulse. However, because agglomerated molten particles at

the surface lift off, they do not significantly alter the burning rate. Dokhan reported that

enhanced rates were found for ultra-fine micron aluminum added to AP-based propellant,

with enhancement increases for reduced Al size [4]. It was discovered that when using

nanoaluminum additives, the particles burn near the binder matrix surface, and thus, it is

possible to have increase burning rates and reduce the size of slag particles, improving

the performance of aluminized propellants [20]. As shown in Fig. 2.19, the competing

effects of micron or nano aluminum additives can reduce (a, b) or increase (c, d) the

burning rate.

45

Fig. 2.19: Competing effects for AP/HTPB/AL binder matrices, showing oxidizer/fuel

(O/F) diffusion flames and near-surface aluminum combustion [5].

Nano-aluminum was found overcome the competing effects, as shown in (d), and

increases the propellant burning rates by 60–100% when compared to micro-aluminized

propellants [5]. As a result, bimodal burning can be used in solid propellant combustion

to tailor the combustion rate for particular applications.

2.3.3 Nanofluids

In contrast with solid propellants, gelled and liquid propellants can be throttled in the

engine, allowing longer range flight missions than SRM powered vehicles. Recent

attention has focused on nanoenergetic additives dispersed in liquid fuels, which offer

numerous capabilities and advantages to liquid fuel combustion. However, thus far there

have been relatively few advances in the ignition of colloidal suspensions. Previous

46

studies [22, 23] have shown that the addition of nanoparticles to liquids (nanofluids),

such as water, may substantially improve the thermal conductivity and mass transfer

inside the liquid even at low concentrations. For example, it was found that the measured

thermal conductivity of oil with suspended carbon nanotubes are anomalously (>250%)

greater than theoretical predictions and are nonlinear with nanotube loadings [24]. Thus,

this provides opportunities for enhanced thermal transport and heat sink capabilities, and

leads to several significant applications, such as in two-phase heat transfer or reducing

the pumping power in heat exchangers. In addition, there may be combustion instabilities

in both liquid-fueled and solid-fueled systems due to unsteady acoustical waves in the

combustion chamber. These traveling waves can result in high-frequency, or ―screech‖

oscillations, can result in pressure variations that can sometimes result in catastrophic

motor failure. Agglomerated metal particles have also been shown to dampen acoustic

vibrations that may lead to these combustion instabilities. The magnitude is dependent on

the mass fraction of particles in the chamber, and generally, larger particles have the

ability to damp lower frequencies, while smaller particles can damp out higher

frequencies [78]. It may be interesting to see what applications nanoparticle additives

have to offer in this area.

Recent attention has focused on the potential large scale implementation of

nanoparticles as viable secondary energy carriers [9]. This concept proposes that pure

nanoenergetic materials or suspensions of nanoenergetic materials in a liquid medium can

be controllably ignited to provide a secondary release of thermal energy. The combustion

products can then be captured and reprocessed into their original nanoparticle form for

repeated use. Previous studies have shown that suspended metallic colloids also have

47

the ability to be optically ignited by a simple light source, such as a camera, resulting in a

multipoint or ―distributed ignition‖ within a combustion engine [79]. Local disruptive

burning behavior similar to microexplosive slurry droplet combustion has recently been

observed for nanoaluminum suspended in ethanol [27]. As shown in Fig. 2.20, as the

droplet evaporates in a heated environment, there is a formation of a nonvolatile glass-

like shell from the particles left behind, leading to ignition.

Fig. 2.20: Microexplosion behavior of micron aluminum n-heptane slurries [80].

It has been suggested that this combustion regime enhances local turbulence and

promotes more complete combustion [80]. In addition, recent studies have reported

reduced ignition delays for rapid aerosol compression of nanoaluminum additives in

ethanol and JP-8 [6]. In monopropellants, the linear burning rates of nitromethane

(CH3NO2) were found to be enhanced by nanoaluminum (38 nm and 80 nm) additives in

deflagration studies [30]. It was demonstrated that nanoaluminum also shows promise to

replace any non-energetic gelling agents (i.e. - fumed silica), due to its large interfacial

area. In a later study, the linear burning rates of CH3NO2 were also found to be increased

by aluminum oxide (ALOOH), silica (SiO2), and functionalized graphene sheet additives

[11]. It was suggested that increased burning rates were due to enhancements in radiation

and thermal conductivity. Tyagi [8] demonstrated that hotplate ignition probability can be

48

enhanced in hydrocarbons (diesel fuel) with nanoaluminum (50 nm) and nanoaluminum

oxide (15 and 50 nm). With aluminum volume fractions of 0%, 0.1%, and 0.5%, hot plate

droplets were found to have much higher ignition probability regardless of the aluminum

size or form. Finally, experimental studies with cerium oxide in liquid fuels are known to

display increased catalytic activity, resulting in oxidation of hydrocarbons and

functioning as an oxygen buffer against NOx formation. Cerium oxide additives to

biodiesel resulted in reductions of NOx by approximately 30% and reductions of

hydrocarbon emissions by 25-40% [7]. Therefore, nanoparticles can function as a catalyst

and an energy carrier, as well. In addition, due to the small scale of nanoparticles, the

stability of the fuel suspensions should be markedly improved.

Several investigations with nanoaluminum in liquid and frozen H2O oxidizer

systems exemplify how nanoscale energetic materials can enhance the reactivity in

propellants while maintaining ―green‖ reaction products [31]. Risha et al. investigated the

combustion behavior of nanoaluminum and liquid water in an argon environment [32].

Nanoaluminum and liquid water with no gelling agent were heavily packed into quartz

microtubes and ignited, it was found that the mass burning rate was 157% faster when the

diameter was decreased from 130 nm to 50 nm, while micron-sized aluminum did not

ignite at all. Originally proposed for micron particles [81], unpassivated aluminum and

steam (H2O) can react to produce hydrogen, with three possible reactions [82],

(2.14)

(2.15)

(2.16)

By carrying and reacting in-situ hydrogen from an Al/H2O mixture, as opposed to

49

carrying cryogenic hydrogen along with a fuel delivery system [32], this offers new

capabilities to make H2 on demand without storage or transport of hydrogen. This has

great interest to offer more non-toxic alternatives to the currently deployed hydrazine

propellants and in underwater propulsion applications, where the oxidizer does not need

to be carried on board the vehicle. Building upon this research, latest fuel research with

nanoaluminum and frozen water (ice) composites, offer increased handling and safety to

the fuel [33]. The official name for this is formulation is aluminum-ice (ALICE), and is

being jointly researched and developed by Pennsylvania State University, NASA, Purdue

University, and the Air Force Office of Scientific research (AFOSR). These composites

show improved vacuum specific impulse over traditional AP/HTPB/AL composites. This

formulation is very difficult to ignite with larger sized aluminum, however, with

nanoparticles, their increased surface area provide more reactivity with water, and

scientists have successfully used nanoaluminum (80 nm) frozen in ice to launch a rocket.

Furthermore, there is an interest in retaining the aluminum combustion products and

reducing them back into the original aluminum fuel. As a result, current developments of

utilizing nanoaluminum in liquid and frozen water formulations show great promise, and

are being considered for lunar and Mars propulsion systems .

2.4 Summary

There has been a considerable amount of research on the combustion of micron and

nanoscale aluminum metal. As expected, the current costs of implementation may be an

obstacle; however, further development in this area is expected to make these additives

more available and affordable in the near future. Nonetheless, a fundamental

50

understanding of micron and nanoscale aluminum combustion is critical to the design and

implementation of practical propulsion systems that use aluminum additives.

Nanocomposites, solid-based, gelled, and liquid fuels all have demonstrated available

benefits to nanoparticle additives. However, relatively little investigative work has been

performed on nanofluids. More experiments on the feasibility of these fuels, especially

the combustion of nanofluid fuels, are needed to explore novel applications of

nanoparticles.

51

Chapter Three

Energetic Characteristics of Nanoscale Additives in

Biofuel

In this work, the combustion properties and performance of nanoaluminum (n-Al)

and nanoaluminum oxide (n-Al2O3) additions to liquid ethanol (C2H5OH) were

qualitatively and quantitatively investigated. As mentioned in chapter two,

nanoaluminum suspended in ethanol can lead to reduced ignition delays [6] over pure

ethanol. Nanoaluminum additives in ethanol also exhibit excellent stability

characteristics, in comparison to micron particles [27]. Furthermore, micron particles are

difficult to ignite and maintain the reaction. With limited studies on this fuel formulation,

the primary objective of this experimental study was to characterize the combustion and

gain a better understanding of n-Al oxidation in a multi-component system. As a result,

the heat of combustion (HoC) was studied using a modified static bomb calorimeter

system. Combustion experiments were performed with volume fractions of 1%, 3%, 5%,

7%, and 10% for n-Al, and 0.5 %, 1%, 3%, and 5% for n-Al2O3. Finally, thermodynamic

equilibrium calculations were carried out and compared to the experimental results. The

fundamental combustion studies in this study may be extended to more complex

nanoenergetic systems.

52

This experimental study characterizes the combustion of commercially available

aluminum (Skyspring, corp.) and aluminum oxide (Nanophase, corp.) nanoparticles in

ethanol. Thermally conductive nanoscale additives in ethanol are known to significantly

enhance thermal transport properties in the liquid mixture. Ali et al [83] performed a

systematic hot-wire beam displacement study of pure Al nanoparticles (18 nm) in water,

ethyl glycol, and ethanol at volume fractions from 0.085% to 0.42%. Their reported

enhancements of thermal conductivity and thermal diffusivity in ethanol were 24.27%

and 29.43%, respectively. Hu et al [84] utilized a transient hot-disk method to evaluate

the thermal conductivity of AlN nanoparticles (20 nm) in ethanol, resulting in a 20%

increase in thermal conductivity at a volume fraction of 4%. Furthermore, Allen et. al [6]

blended nanoaluminum (50 nm) within ethanol and kerosene-based JP-8 to reduce

ignition delays in a rapid compression machine by 32% and 50%, respectively, and the

authors proposed that a reduction in ignition delay may be a direct consequence of

increased thermal conductivity leading to accelerated internal heating and evaporation of

Al-ethanol nanofluid droplets. Furthermore, nanoaluminum additives in ethanol exhibit

excellent stability characteristics, and micro explosive behavior similar to disruptive

slurry droplet combustion [27].

Ethanol is widely used as a biofuel and oxygenated (oxygen containing) fuel

additive, in order to reduce greenhouse gases from fossil fuel use and promote energy

diversification. Alcohols contain oxygen and thus more fuel can be burned with the same

amount of air. The first chemical rockets, and the first manmade objects to achieve

spaceflight, such as the V2 rocket, consisted of liquid oxygen (LOx) and ethanol systems

[85]. Since then, other liquid fuels have been developed and used, such as kerosene-

53

based derivates JP-8 and Jet-A. Nonetheless, ethanol production and technology still

plays a major role in U.S. government incentives and public policies. Ethanol was

chosen based on its use as an oxygenated biofuel, and its complete combustion products

in pure O2 of CO2 and H2O [41]. Both of these products are possible oxidizers for

aluminum, under certain environmental conditions, yielding the following global

reaction mechanism, if O2 is assumed the reactant,

(3.1)

This global reaction does not express all the elementary features of combustion; however,

its detailed mechanism is currently known, with kinematic models available for the fuel

[86]. Furthermore, proponents of ethanol state that its low overall emissions compared to

gasoline, high octane number, and relatively low bio-toxicity make it an attractive

alternative fuel. Any spillage of pure ethanol may be simply diluted with water and

disposed of down the drain [87]. On the contrary, ethanol detractors believe that its low

energy content, poor ignition characteristics in general, and low flame temperatures are

drawbacks that make it less viable. It also easily contaminates soil and water supplies,

although it has been shown to biodegrade faster than hydrocarbon-based fuels [37].

Aluminum is used due to its numerous applications as an energetic material, high

volumetric heat of combustion, high thermal conductivity, excellent surface absorption,

and low melting/ignition temperatures [13]. If O2 is assumed the primary oxidizer for

aluminum combustion, the global reaction mechanism is as the following,

(3.2)

The main combustion product of aluminum, Al2O3, is environmentally stable and may be

recycled back to pure aluminum with an electrolytic reduction [9, 40]. Therefore, when

54

considering ethanol‘s lower energy density, aluminum combustion within ethanol could

be regarded as a viable formulation, if its energetic value is practical. Furthermore, if the

poor ignition characteristics of ethanol, especially in cold weather conditions, and low

flame temperatures are mitigated, that would greatly increase its attractiveness as an

alternative fuel. Obviously, these are just a small number of the considerations when

examining the use of alcohols, such as methanol and ethanol, within current propulsion

systems. Nevertheless, generally speaking, more research and development on the

feasibility of novel alternative fuels are needed to meet increased environmental

regulations and fuel efficiency in the near future.

3.1 Experimental Methods

As discussed in chapter two, fuels are evaluated, in part, by the amount of heat released

during combustion. Calorimetry is defined as the science of measuring the heat released

from chemical reactions or physical changes [88], and oxygen (O2) bomb calorimeters

are commonly used to determine the enthalpy of combustion for solid and liquid samples.

The enthalpy of combustion, defined as the amount of heat released per unit of sample, is

commonly expressed as calories per gram (cal/g), Btus per pound (Btu/lb.), or

Joules/gram (J/g). Oxygen bomb calorimeters are sufficiently precise that they are widely

used in laboratory instruction and commercial procedures with heat of combustion. It is

important to know that the enthalpy of combustion may be expressed volumetrically

(MJ/L) or gravimetrically (MJ/kg), as shown for liquid fuels and fuel additives in Fig.

3.1.

55

Fig. 3.1: Energy content of interest for fuels and fuel additives in O2 from [36, 37].

Regarding constant volume calorimeter experiments, the reactants are enclosed in a

stainless steel vessel and ignited in pure O2 to ensure complete combustion. The heat of

combustion of the fuel sample is determined indirectly from the heat transfer to the

surrounding distilled water. In the present study, the combustion experiments were

carried out with a modified static bomb calorimeter under a closed hood, as shown below

in Fig. 3.2.

Fig. 3.2: Schematic of the calorimeter system and combustion vessel [88, 89].

56

The calorimeter system is assumed to be adiabatic, resulting in no heat transfer (ideally)

between the system and the ambient environment. The only heat transfer that occurs is

assumed to be between the combustion vessel and the surrounding water within the

system itself. Approximate 1 gram samples were placed on the combustion capsule, and

combustion was initiated with an ignition unit via electrical discharge through a Ni-Cr

alloy fuse wire (length of 10 cm) in contact with the sample. Experiments were carried

out in the presence of 2 liters of distilled water in the oval bucket with pure O2 pressures

of 20 atm. In the case of liquid or solid samples, the total heat liberated in the combustion

process is primarily due to three sources: 1) the combustion of a fuel sample, 2) the

ignition of a fuse, and 3) the formation of extraneous acids such as nitric acid (HNO3)

and sulfuric acid (H2SO4). The water temperature rise was then determined from the

average of four T-type thermocouples embedded in the system, as illustrated in Fig. 3.3.

Fig. 3.3: Temperature vs. time data plot determined with thermocouple readings.

57

Note that the markers in Fig. 3.3 indicate a temperature reading every 50 seconds, while

the actual acquisition system recorded temperatures every 10 seconds. After combustion,

the interior of the vessel was washed with a jet of water and a titration procedure of the

residual products (bomb washings) using sodium carbonate (Na2CO3) and methyl red

indicator determined the amount of nitric acid formed. It was determined that no

significant amounts of extraneous acids were formed in this study.

In order to determine the heat of combustion of a given sample, the heat capacity

of the calorimeter system first needs to be determined by using a calorimetric standard.

The total system heat capacity is a summation of the heat capacities of the combustion

vessel and its contents, the surrounding water, any component submerged in the water,

and the calorimeter jacket. In this study, the heat capacity and accuracy of the system

were determined by calculating the heat transferred to the water by the combustion of

benzoic acid (C7H6O2), having a quoted energy of combustion of 6318 calories per gram.

For one gram ) samples of benzoic acid with predetermined heats of combustion ( ),

the system heat capacity ( ) was determined by the following equation,

(3.3)

where is the correction value for any sources of heat other than the

sample, such as the formation of nitric acid, or the combustion of fuse wire. Each

centimeter of consumed fuse wire represents 2.3 calories of extraneous heat released

during combustion of the wire, and each millimeter of alkali used to titrate the bomb

washings represents 1 calorie set free by the oxidation of nitrogen to form nitric acid, if a

concentration of 0.0709 N sodium carbonate was used for titration [89]. The term is

the recorded temperature rise accompanying the combustion event (°C). Ultimately, it

58

was discovered that several flushes of the vessel with O2 removed any residual nitrogen

in the vessel, and the samples used in this study did not contain the necessary sulfur to

form sulfuric acid.

In actual experiments, a small amount of radiation may have been introduced, in

this case, a radiation correction for is used, as derived in ASTM Designation D240

[90]. Before ignition, the system typically gains a small amount of heat from the

surrounding air, and after ignition, the system water is warmer than air and this results in

a loss of heat to the air. The D240 method attempts to correct the radiation effect, by

assuming that the calorimeter is equally heated by its surroundings before ignition, after

ignition, and during the first 60 % of the temperature rise. In all subsequent calculations;

the corrected net temperature rise is used,

(3.4)

where the following constants are selected from a plot, such as in Fig. 3.3,

a ≡ time of firing

b ≡ time when the temperature reaches 60 % of the total temperature rise

c ≡ time for steady state temperature, after the temperature rise

≡ temperature at time of firing

≡ temperature at time c

≡ rate of temperature rise during the 5 min. period before firing

≡ rate of temperature rise during the 5 min. period after time c

If the temperature was falling instead of rising after time a or c, the quantities

or become positive.

59

The standard procedure described in the literature recommends at least four

experiments for system calibration [88]. For eight consecutive calibrations, the

experimental heat capacity for the system was determined to be 2523 calories per °C, as

shown in Fig. 3.4.

Fig. 3.4: Calibration measurements for benzoic acid, represented stoichiometrically as

C7H6O2 (s) + 7.5 O2 (g) → 7 CO2 (g) + 3 H2O (l).

Not included in the Fig. 3.4 are periodic combustion of benzoic acid to check the system

for accuracy. Once the system has been calibrated with , the enthalpy of combustion

for any sample may be found with the following relation,

(3.5)

For liquid fuel samples, due to safety concerns, the fuse wire was connected to the liquid

by a cotton thread string. The cotton thread empirical formula CH1.686O0.843 was used with

an energetic value of 16250 J per g [91]. In all cases the fuel sample was to remain at

approximately one gram, and volumetric calorific values were determined from mass to

volume conversions and verified by sample experimental volume measurements.

60

In liquid fuel combustion, the higher heating value (HHV) & lower heating value

(LHV) are both different ways to express the energetic value heat of combustion of a

material. Higher heating value assumes that all of the water vapor combustion products

have condensed, and that the heat of vaporization given up to vaporize the water is

returned. The lower heating value assumes that all the moisture generated in combustion

is in the vapor phase; therefore the HHV is usually 10-20% higher than the lower heating

value. Constant volume bomb calorimetry measures the HHV; however, it may be

possible that moisture generated has not fully condensed to recover the heat of

vaporization given up, within the timeframe of data collection. Therefore, the

experimental calculations included in this report do not discriminate between phase

change and reaction enthalpies. It is assumed that all moisture generated in ethanol

combustion has condensed. To be conservative, an additional ±2.5% error could be

added. As shown in Table 3.1 and Fig. 3.5, for pure ethanol runs, the experimental

volumetric heat of combustion (HoC) was 21.67 ± 1.08 (MJ/L), this is in reasonable

agreement with published values. The approximated 2 MJ/L difference may be due to

using a different grade pure ethanol in this study.

Table 3.1: Fuel properties of ethanol fuel and commercial diesel fuel [36].

Fuel Density

(g/cc)

Literature

HHV (MJ/kg)

Literature

HHV (MJ/L)

Experimental

HoC (MJ/kg)

Experimental

HoC (MJ/L)

Ethanol

(99%) 0.789 28.86 23.4 27.44 ± 1.35 21.67 ± 1.08

Diesel 0.86 45.9 37.3 45.34 ± 2.2 38.54 ± 1.87

61

Fig. 3.5: Experimental volumetric heat of combustion of pure ethanol fuel.

Since ethanol is very volatile, mixtures were prepared and loaded into the vessel

immediately before experiments. In an effort to reduce the experimental error, adhesive

tape sealants were also tested, as shown in Fig. 3.6.

Fig. 3.6: Experimental volumetric heat of combustion of sealed ethanol fuel.

When sealing each combustion capsule, a sheet of 3M transparent tape was stretched

62

across the top of the capsule. Any excess tape over the side was trimmed off, and the

empty capsule was weighed. A small opening in the tape was made to prevent the tape

from collapsing onto the sample under high pressures. However, difficulties in applying

in the seal to the capsule, relatively high uncertainty values compared to pure ethanol,

and the required slower rate of oxygen supply when charging the vessel, resulted in not

using adhesive seals for all subsequent liquid samples. Commercial diesel fuel was also

investigated, as plotted in Fig. 3.7, which was determined to have a relatively large

margin of error. For additive suspensions in diesel, it was determined that additive

loadings could not exceed approximately 0.5% volume fractions without settling.

Fig. 3.7: Experimental volumetric heat of combustion of diesel fuel

In this study, n-Al particles were suspended in pure ethanol with volumetric

fractions of 1%, 3%, 5%, 7%, and 10%, and n-Al2O3 particles were suspended in pure

ethanol with volumetric fractions of 0.5%, 1%, 3%, and 5%. Additive n-Al and n-Al2O3

particles were of 50 nm and 46 nm size, respectively, with properties specified by the

manufacturer in Table 3.2. Since aluminum and aluminum oxide present extreme cases

63

with respect to energy content, aluminum oxide was used as an additive and was

hypothesized to not participate reactively in the experiments.

Table 3.2 Material properties of aluminum nanoparticle samples, including average

particle size (APS) and specific surface area (SSA).

Material Manufacturer Oxide phase

True density

(g/cc)

APS

(nm) SSA (m2/g)

Al (99.9%) Skyspring Amorphous 2.7 50 20-48

Al2O3

(99.5%) Nanophase 70:30, δ:γ 3.6 46 36

Scanning electron microscope (SEM) images in Figs. 3.8 and 3.9 below display the

similar size diameter and size distribution of the nanoaluminum materials. Note that n-

Al2O3 was less conductive than pure aluminum, and was less illuminated.

Fig. 3.8: SEM images of n-Al powder (left) and n-Al2O3 (right) and at 500 nm

magnification.

An Energy Dispersive X-Ray Spectroscopy (EDS) was performed before ignition and it

was determined that the atomic composition of n-Al samples were 78.53% Al, 19.48% O,

and of n-Al2O3 samples were 53.52 % Al, 46.48 % O. Both metals were suspended in

fuels until they exhibited a thick and claylike consistency (i.e. - to the observed threshold

of nanoparticle stability).

Five experiments were performed for each volume fraction and corresponding

64

additive. By applying alternating high and low pressure cycles of sound through the

solution, a simple mechanical dispersion method was used to separate agglomerated

particles. Fuel samples were sonicated in an ultrasonic cleaner for at least 30 minutes at

47 kHz with a power rating of 143 watts. Ultrasonic-induced cavitation applies

mechanical stresses between particles to break apart and reduce agglomeration. Two

additional approaches to stabilization include electrostatic stabilization and steric

stabilization, as shown in Fig. 3.9.

Fig. 3.9: Steric stabilization (left) and electrostatic stabilization (right) of additives in

a solution [92].

In steric or chemical stabilization, an absorbed compound layer alters the surface

properties to separate particles from each other. Previous studies of alumina powders

dispersed in ethanol have used surfactants, or have shown that absorbed acetic acid

generates a steric barrier between alumina particles [93, 94]. Therefore, by modifying the

acidity of the system, the suspending ability can be controlled. In electrostatic

stabilization, the particles are separated by Coulomb repulsion of like charges. In this

case, an electrical double-layer forms when a particle is placed in the dispersing medium.

The interior layer is called the surface charge, and the outer layer is the diffuse layer.

However, electrostatic stabilization is often not used due to the low dielectric constant of

ethanol. In the present author‘s experiments, no gelling agents, apart from the

65

nanoparticles themselves, or surfactants were used, in order to eliminate any contribution

from any other additives than nano-aluminum (n-Al) and nano-aluminum oxide (n-

Al2O3). Once ignited, the total corrected enthalpies of combustion were determined from

the net temperature increase and subtraction of extraneous heat of formations.

Despite being a well controlled instrument to measure thermodynamic properties,

it is important to note that there are errors inherent to using calorimeter-type systems. The

three sources of uncertainty can be attributed to the uncertainty in additive volume

fraction (sample mass and volume measurements), nonadiabaticity of the system, and the

performance variation of the ethanol suspensions themselves. Uncertainties in volume

fraction may be inclusive to the standard error in the samples graphed in the following

experimental figures.

3.2 Results and Discussion

3.2.1 Experimental Results

The nomenclature for the aluminum suspension samples will be as follows: for an

aluminum nanoparticle suspension volume fraction of 5% in ethanol, it will be indicated

by Eth + 5% Al, or Eth + 5% Al2O3 for alumina. As shown in Fig. 3.10a and 3.10b, the

energetic values are represented for volume fractions of Eth + n-Al samples at 1%, 3%,

5%, 7% and 10% with a standard deviation error.

66

(a) (b)

Fig. 3.10: a) Volumetric HoC of Eth + n-Al samples at 20 atm, and b) volumetric and

gravimetric HoC of Eth + n-Al samples.

Initially, at volume fractions of 1% and 3%, there was found to be a decrease in energetic

release for n-Al ethanol suspensions. With subsequently larger volume fractions, there

was an enhancement in volumetric energy release, indicating a transition to one of the

Al2O3 polymorphic phases. It was determined that the n-Al nanoparticles had an

oxidized n-Al2O3 layer on the surface, and that the volumetric HoC was lower than that

of pure ethanol at the volume fractions of 1% and 3% due to the existence of a surface

67

oxidization layer. Once the volume fraction was higher than 3%, more heat of

combustion was released from n-Al in the reaction process, and the volumetric heat of

combustion increased linearly. It is interesting to note that even though there was an

increasing trend in HoC versus volume fraction in Fig. 3.10a, there was a relatively

constant gravimetric HoC for all volume fractions in Fig. 3.10b.

The following Figs. 3.11a and 3.11b show the energetic values for n-Al2O3 samples.

(a)

(b)

Fig. 3.11: a) Volumetric HoC of ethanol + n-Al2O3 samples at 20 atm, and b)

volumetric and gravimetric HoC of ethanol + n-Al2O3 samples.

68

These nanoparticles have a dominating component of Al2O3 coating that was found to

increase the stability of the samples. However, as predicted, the n-Al2O3 nanoparticles did

not react with the ambient vessel O2. In Fig. 3.11.a, n-Al2O3 suspensions exhibited a

linear decreasing trend of energetic release due to the displacement of reactive ethanol. It

is readily apparent in Fig. 3.11.a that the volumetric HoC‘s were more than 2 MJ/L lower

than that of n-Al samples at equivalent volume fractions of 1% and 3%. This confirmed

that the volumetric HoC‘s of Eth + n-Al samples at 1 and 3% were lower than that of

pure ethanol due to oxidization layers. An EDS technique was performed on the residual

combustion products for Eth + 5% n-Al and n-Al2O3, and it was determined that in both

cases the Al:O atomic ratio was approximately 30:60, corresponding to a Al2O3 atomic

composition, as shown in Figs. 3.12a, and 3.12b.

(a)

(b)

Fig. 3.12: EDS spectra after combustion for a) Eth+5% n-Al, and.b) Eth+5% n-Al2O3.

69

The figures illustrate a near identical EDS response after combustion, which indicates a

thorough combustion of n-Al. Note that there is a background level of carbon due to

carbon mounts used in SEM analysis. Figure 3.13 shows the surface morphology of

residual nanoparticles after combustion. It was determined that, once ignited, the

nanoparticles will be quickly oxidized as n-Al2O3 and fused together in a nanoporous

structure. In particular, Eth + 5% n-Al samples will coagulate into droplets, while Eth +

5% n-Al2O3 will flake into a powdery substance.

Fig. 3.13: SEM images of residual combustion products of Eth + 5% n-Al2O3 for a)

2.00 µm magnification and b) 500 nm magnification.

This was characteristic of both n-Al and n-Al2O3 combustion products, however, some

parts of the residual combustion products of pure aluminum in ethanol exhibited surfaces

such as Fig. 3.14.

70

Fig. 3.14: SEM images of residual combustion products of Eth + 5% n-Al for 50 µm

magnifications.

Generally, it was experimentally shown that the amount of heat released from ethanol

combustion increases almost linearly with n-Al concentrations. Nanoaluminum volume

fractions of 1% and 3% deviated from the average volumetric heat of combustion from

that of pure ethanol by 3.78% and 0.66%, respectively. Higher volume fractions of 5%,

7%, and 10% increased the volumetric heat of combustion by 5.82%, 8.65%, and

15.31%, respectively. Nanoaluminum oxide or heavily passivated n-Al does not

participate reactively.

The experimental heat of combustion can be determined with an energy balance

within the system,

, (3.6)

where Q is the heat transfer between the interior of the vessel to the surrounding water,

W is the boundary work, and U is the internal energy of the system. Considering a

constant volume process,

, (3.7)

is the predetermined heat capacity of the vessel and water system, and ΔT is the

temperature change of the system after the combustion reaction. To determine the

71

enthalpy change within the vessel, the definition of enthalpy is used,

, (3.8)

where H is the enthalpy, P is the pressure within the vessel, and V is the vessel volume,

340 cm3. Assuming an ideal gas within the vessel, and combining Eqs. 3.7 and 3.8, the

experimental HoC can be rewritten as,

, (3.9)

where is the change in moles of gas reactants and products, and R is the ideal gas

constant (8.3145 J mol-1

k-1

). For the determined constant heat capacity ( of the

system, the first and final terms of Eq. 3.9 indicate that combustion products with higher

flame temperatures will have larger enthalpies of combustion (HoC).

3.2.2 Thermodynamic Equilibrium Calculations

The combustion thermodynamics was modeled using the NASA CEA computer program

[49]. This code calculates chemical equilibrium product concentrations and determines

thermodynamic properties for the product mixture. By the minimization the Gibbs or

Helmholtz free energy, or the maximization of entropy, adiabatic flame temperatures can

be estimated, as shown in Appendix A. It was assumed that all reactants were initially at

room temperature (298 K), undergoing a constant pressure combustion, as in most

practical applications. As shown in Fig. 3.15, ethanol with an added 10% Al by volume

resulted in a 6 to 9% increase in adiabatic flame temperature over the range of pressures,

and an 8.27% increase at the initial experimental pressure 20 atm.

72

Fig. 3.15: Calculated adiabatic flame temperatures for ethanol and aluminum mixtures

at stoichiometric conditions, for initial reactant temperatures of 298 K.

Fig. 3.16: Calculated adiabatic flame temperatures for ethanol and aluminum oxide at

stoichiometric conditions, for initial reactant temperatures of 298 K.

73

This 8.27% adiabatic flame temperature increase is comparable to the 8.65%

experimental HoC increase due to n-Al additives. A calculation for 20% aluminum in

ethanol predicts a much larger (17%) increase in flame temperature. As shown in Fig.

3.16, ethanol with 5% Al2O3 volumetric concentration resulted in a 1 to 2% lower flame

temperature than pure ethanol, agreeing with the experimental result that n-Al2O3 did not

participate in combustion. Similarly, a calculation for 10% Al2O in ethanol predicts a

further decrease of the flame temperature, displacing more ethanol fuel. Furthermore, as

shown in Fig. 3.17, the calculated adiabatic flame temperatures for solid and vaporized

aluminum in O2 were logarithmically compared to liquid ethanol with Al and Al2O3

volumetric concentrations, similar to Figs. 2.3 and 2.4 in chapter 2.

Fig. 3.17: Calculated adiabatic flame temperatures for ethanol and aluminum mixtures

at stoichiometric conditions, for initial reactant temperatures of 298 K.

74

The change in the combustion regime may also be predicted from Fig. 3.17. For

Al and fuel-oxidizer mixtures with flame temperatures below the Al vaporization

temperature, combustion is expected to occur as a heterogeneous surface reaction. While

mixtures with flame temperatures above the Al vaporization temperature typically occur

in a diffusive gas-phase. Again, this transition in combustion mode has been

experimentally measured; a transition for 10 µm Al in O2 was shown to occur at

approximately 10 atm [20]. In Fig. 3.17, pure ethanol in O2 has a higher adiabatic flame

temperature than the Al vaporization temperature up until approximately 4 atm. Over the

same range of pressures, ethanol with 10% Al additives exhibited flame temperatures

above the Al vaporization temperature up until approximately 14 atm. This indicates that

Al additives in biofuel could significantly influence the combustion regime of the

mixture.

The influence of the oxide layer was taken into account, by incorporating a

bimodal mixture of Al and Al2O3 into the fuel. For the reactant mixture of ―C2H5OH (L)

+ 5% Al : 5% Al2O3 in O2,‖ Fig. 3.17 illustrates a 3.8 to 5.5% decrease in adiabatic flame

temperature from the 10% Al ethanol mixture. This further illustrates the inert

characteristics of Al2O3, and that the presence of an oxide layer significantly reduces the

total combustion energy released from the Al-ethanol mixture. Furthermore, another

interesting result is a clear correlation with threshold of enhancement for the experiment

and thermochemical calculation. In the experiments, it was observed that the threshold

for the energetic enhancement of ethanol was with 3% volume fraction of pure Al. From

Fig. 3.15, the threshold for the flame temperature increase was with also with 3% volume

fraction of pure Al. Taking this into account for a bimodal mixture of 3% total volume

75

fraction, the calculated CEA threshold for the flame temperature enhancement was a

mixture of approximately 1.3% Al + 1.7% Al2O3. For this mixture, data processing of a 1

gram sample yields 36% and 43% active Al content in mass and volume. The thickness

of the oxide coating can then be estimated from the following equation [16]:

(3.10)

Where (2.7 g/cc) and (3.2 g/cc) are the aluminum and amorphous Al2O3

densities, is the outer mean particle radius, and is the pure Al content by mass. Based

on the threshold of experimental and simulation energetic enhancement (3%), the

estimated oxide layer thickness from this calculation was 6.6 nm. It is likely the oxide

layer thickness increased due to exposure to the atmosphere during storage, additional

uncertainty may be attributed to the exclusive nature of Al and Al2O3 in the software and

adiabatic flame assumptions. Authors commonly estimate this oxide thickness from the

particle size distribution and transmission electron microscopy (TEM) imagery. Other

ways to estimate this can be derived from thermogravimetric analysis [16] or binary

phase diagrams [69]. This experimental study proposes a basic new method of oxide

thickness approximation based on experimental HoC and thermochemistry data.

3.3 Conclusions

Combustion element composition and surface morphology were evaluated using a

scanning electron microscope and energy dispersive spectroscopy system. The results

indicate that the amount of heat released volumetrically from ethanol combustion

increases almost linearly with n-Al concentration. N-Al volume fractions of 1% and 3%

76

did not show enhancement in the average volumetric heat of combustion, however higher

volume fractions of 5%, 7%, and 10% increased the volumetric heat of combustion by

5.82%, 8.65%, and 15.31%, respectively. Aluminum oxide and heavily passivated n-Al

additives did not participate in combustion reactively, and there was no contribution from

Al2O3 to the combustion enthalpy in the tests. A combustion model that utilized Chemical

Equilibrium with Applications (CEA) was conducted as well and was shown to be in

good agreement with the experimental results.

Experiments have been conducted to investigate the combustion characteristics of

n-Al and n-Al2O3 in ethanol. To summarize, the conclusions of this study are as follows:

1) Aluminum nanoparticles may be stably suspended in ethanol fuel up to the

concentration of approximately 10% volume fraction for pure aluminum, and 5%

volume fraction for n-Al2O3. Although n-Al has demonstrated its ability as a

gelling agent, it is recommended in future work that a dispersant is incorporated

in the suspension for higher nanoparticle loadings.

2) It was experimentally shown that the amount of heat released from ethanol

combustion increases almost linearly with n-Al concentrations. Nanoaluminum

volume fractions of 1% and 3% deviated from the average volumetric heat of

combustion from that of pure ethanol by 3.78% and 0.66%, respectively. Higher

volume fractions of 5%, 7%, and 10% increased the volumetric heat of

combustion by 5.82%, 8.65%, and 15.31%, respectively. Nanoaluminum oxide or

heavily passivated n-Al does not participate reactively.

3) The oxide layer has a significant effect on reaction energetics. Scanning electron

microscope analyses and x-ray spectroscopy yielded almost identical final

77

element compositions, despite different initial compositions. Nanoaluminum

oxide displaces energetic ethanol fuel and active aluminum content, and may

function as a diffusion barrier, inhibiting phase transitions. Furthermore,

thermodynamic equilibrium modeling with CEA agreed with the reaction

energetics, predicting an 8.27% increase in adiabatic flame temperatures for Eth+

10% Al suspensions. As the size of aluminum fuel additive decreases, there is a

concern as to whether the energy content is significantly decreased. Because the

oxide thickness is relatively independent of the particle diameter, for nanoparticle

diameters of 20 nm or less, the oxide layer becomes a more prominent part of the

particle volume, accounting for approximately 60% energy loss per volume [66].

Therefore, it is expected that nanoenergetic fuel formulations release more energy

when taken as a whole, due to more complete combustion; however, there may be

a lower limit of the particle diameter on the energetic enhancement of

nanoparticles.

4) Furthermore, this may be extended to other burning parameters, such as

linear/mass burning rate and ignition delay that are influenced by the amount of

heat released. Although it is acknowledged that higher energetics may exceed the

limits set by the materials in the combustion system, increase the rate of thermal

NOx formation, and induce dissociation of products back to reactants, this study

provides valuable information on the role of the oxide layer in nanoaluminum

combustion and ability to tailor the reaction energetics.

78

Chapter Four

Ignition Characteristics of Nanoaluminum in Biofuel

Along with energy density enhancement, achieving precise control over the

reactivity of nanofluids is an opportunity for future nanoenergetic fuel applications. A

second experimental study involves the study of ignition probability of n-Al (50 nm) and

n-Al2O3 (36 nm) in ethanol and the commonly used No. 2 fuel oil (diesel). The primary

aims in this study were to study the effect of nanoenergetic additives on the ignition

probability of ethanol with a hot-plate setup, and to explore the underlying mechanisms

of ignition. Aluminum and aluminum oxide were suspended in ethanol and diesel fuels

from 0.1% to 3% volume fractions, and dropped onto a hot plate at temperatures varying

from approximately 300 ºC to 600 ºC. The experimental probability was calculated and a

logistic regression approach was used to determine the 50% ignition threshold.

Nanoaluminum in ethanol was found to significantly increase the ignition probability:

ethanol suspensions with 1%, 2%, and 3% aluminum volume fractions ignited as much as

100 ºC lower than pure ethanol, and exhibited burning regimes similar to disruptive

combustion in slurry droplets, while ethanol suspensions with n-Al2O3 volume fractions

ignited at higher temperatures than pure ethanol. In addition, diesel mixtures with n-Al

and n-Al2O3 additives demonstrated relatively the same hot plate ignition probability as

79

pure diesel. Therefore, the likelihood of nanofluid ignition was strongly correlated with

volume fraction concentration, metal additive material, and surface tension (contact

angle).

4.1 Experimental Methods

As mentioned previously, nanoenergetic additives dispersed in liquid fuels offer

numerous capabilities and advantages to liquid fuel combustion. The standard method to

determine the auto ignition temperature (AIT) of a fuel is the ASTM E659 ―Minimum

Auto Ignition Temperature of Chemical Liquid‖ standard [95]. In this ideal case, the fuel

vapor is uniformly heated in all directions, and insulated from any heat loss. In addition,

ASTM E659 is valid only for liquid and solid fuels that have no condensed phase when

ignition occurs. In accordance with the procedure described in ASTM E659, as shown in

Fig. 4.1, 100 μl of liquid fuel is dropped into a uniformly heated 500 ml glass flask

containing air that is insulated with aluminum foil. The liquid fuel evaporates, and the

temperature is then adjusted until ignition occurs. This is the ideal case in which the

lowest temperature of a fuel will ignite without an external source of ignition. However,

as shown in Fig. 4.2, hot surface or hot plate ignition occurs under less ideal conditions

due to differences in environmental heat losses and test conditions. In this case, the fuel is

only heated from primarily one direction, and hot plate ignition temperatures are

expected to be higher than autoignition temperatures.

80

Fig. 4.1: Schematic of experimental setup for the ASTM E 659 standard [95].

Fig. 4.2:

Schematic of hot surface ignition operating parameters [96].

81

Hotplate ignition, although simple in its geometry, is a complex phenomenon governed

by various factors such as surface geometry, composition, and external environmental

conditions [97]. These conditions are important in determining ignition probabilities for

hot surface ignition events. Ignition then occurs over a range of temperatures, rather than

at a set point. Depending on the application, hot surface ignition of flammable liquids can

be regarded as a risk, or desirable in the case of the combustion of a working fluid. There

is no accepted approach for ignition probability, rather, for each setup; there are several

possible methods that can be used. There are generally two methods for reporting the

results, either the experimental ignition probability for each temperature is reported, or a

curve fitted through the entire range of ignition events. For solid fuels, the Bruceton

method, or ―up and down‖ method is widely used, along with logit, probit, and Wiebull

distribution techniques [98]. In this study, a logistic regression model was used to

calculate the ignition probability, due to its convenience and its successful use with

working fluids by previous authors [99]. There are currently two previous studies of

nanoparticle additives in hot surface ignition experiments: 1) nanoaluminum additives in

diesel [8], and 2) carbon nanotube additives in ethanol [26]. In both cases, there was

determined to be a maximum 20-30% increase in ignition probability at each surface

temperature. In this respect, a single droplet burning may serve as a stepping stone for

understanding spray combustion; however, its practical use is primarily in research.

In this study, a horizontal hot plate setup was used to estimate the hot plate

ignition temperature. The experimental setup consisted of two main elements: an 18 cm

diameter, 2.5 cm thick cast iron heating plate, and a fluid delivery system as shown in

Figs. 4.3, 4.4, and 4.5. A 1500 W spiral copper heating coil was set inside the plate

82

assembly, powered by a heater control unit. Aluminum silicate supported the base of the

heater element, and fiberglass surrounded the entire assembly to minimize heat losses.

The center of the plate was set beneath the top surface to prevent droplets from escaping.

Around the outer radius of the plate, a channel caught any fuel material that was ejected

from the inner region during the experiments. Several high temperature k-type

thermocouples were used to monitor the temperature, with an internal cross-sectional

measurement, an array of surface measurements, and a surface probe closest to the

droplet surface, for use up to 1335°C.

Fig. 4.3: Schematic of the hot surface ignition experimental setup.

Fig. 4.4: Illustration of the hot surface ignition experimental setup.

83

Fig. 4.5: Wire schematic of the hot surface ignition experimental setup.

As shown in Fig. 4.6, low temperature calibration experiments calculated the average

temperature for 600 seconds between the center and four outer circular locations, in

which the total average was within 10% error.

Fig. 4.6: Low temperature distribution across the hotplate setup.

A separate assembly was used to control the fluid delivery. Ultrasonic agitation was used

84

to disperse the mixtures for at least 30 minutes. The volume and flow rate of the liquid

suspensions were controlled with a programmable syringe pump (Cole-Parmer, model

No. 780100C). To ensure that each drop was impinged on the center of the plate, a

swiveling support beam was used to drop in the same region approximately 20 mm above

the plate. The supporting beam also mounted a k-type thermocouple probe to

approximate the initial droplet temperature. A ventilation hood enclosed the entire

experimental setup to continuously remove any gaseous combustion products. Overall,

the heater size and orientation was configured for a reproducible setup, and not to be

representative of any component or device.

In the present study, commercially available aluminum nanoparticles (Skyspring,

corp.) and aluminum oxide (Nanophase, corp.) were used. Tables 4.1 and 4.2 tabulate the

fuel and material properties, respectively.

Table 4.1: Liquid fuel properties [37], all properties at atmospheric pressure.

Fuel Density,

(g/cc)

AIT,

(°C) Tb, (°C)

Flash Point,

(°C)

Volume,

(µl)

k,

(W/mK)

Ethyl Alcohol 0.789 363 78 13 24.7 0.171

Diesel 0.832 256 260-371 66 20.1 0.15

Table 4.2: Additive material properties, as specified by the manufacturer.

Particle Type True Density,

(g/cc ) Diameter, (nm) Crystal Phase

SSA,

(m2/g)

k, (W/mK)

Skyspring Al (99.9%) 2.7

40-60 Amorphous 20-48 250

Nanophase Al2O3

(99.5%) 3.1 36 70:30, δ: γ 36 30

The average liquid droplet diameter was estimated from the total volume of 100 droplets.

From a previous study [100], EDS data reported that the Al/O atomic composition for the

pure Al samples were 78.5/19.5 and 53.5/46.5 for the Al2O3 samples. Scanning electron

85

microscope (SEM) (Hitachi S-480) images of the particle size and distribution can be

seen in Fig.4.7. Stabilizing agents were not used, because this may affect the ignition

process. For instance, a decrease in droplet surface tension has been known to

significantly affect the evaporation rate [101].

Fig. 4.7: SEM image of particle size and distribution of pure Al nanoparticles (left) and

Al2O3 nanoparticles (right).

Ignition experiments were conducted by heating the surface to a starting

temperature of 300 to 400°C. The temperature of the surface was increased by 10°C

increments for approximately 10 to 15 increments, with 15 droplets of fuel dropped onto

the hot plate on each increment. The leidenfrost effect was observed in all cases, in which

a film of fuel vapor is immediately formed under the droplet, physically lifting the

droplet by a cushion of fuel vapor. For each trial, either an ignition event or a pure

evaporation event was recorded. Once ignited or evaporated, any gaseous products were

exhausted by the vent. After each experiment, a silicon carbon abrasive (CAMI Grit

designation 190) and a jet of water was used to remove any excess material. Overall, the

results of over 3500 ignition events on a reproducible setup were recorded during the

86

study. Under this method, a good quantitative indication of the ignitibility of nanofluids

under different conditions can be gathered, as shown in Table 4.3, for pure ethanol.

Table 4.3:

Experimental data recorded for the ignition experiments of pure ethanol.

Temperature, °C Events Trials Experimental Ignition Probability

400 0 15 0

410 0 15 0

420 0 15 0

430 0 15 0

440 0 15 0

450 0 15 0

460 0 15 0

470 0 15 0

480 0 15 0

490 4 15 0.267

500 9 15 0.6

510 11 14 0.786

520 15 15 1

530 15 15 1

540 15 15 1

550 15 15 1

560 15 15 1

In the present study, a logistic regression model is utilized to statistically quantify the

ignition response from surface temperature increase, as described by Hosmer and

Lemeshow [102]. For each temperature, there was a corresponding dichotomous

response: ignition or no ignition. Consequently, the ignition probability and

associated 95% confidence interval can be represented in Eqs. 4.1 and 4.2,

(4.1)

87

(4.2)

where the regression coefficients are and , Z represents the standard normal density

curve at significance, and the standard deviation of the fit. To

determine the coefficients and , an iterative maximum likelihood technique is

performed. Most statistical packages will generate the coefficient estimates along with

the associated deviations. The experimental ignition probabilities at each temperature was

in good agreement with the regression fit, for all fuel suspensions. This may be simply

regarded as a regression tool, however, a more detailed description is provided in

Appendix B. As shown in Fig. 4.8, the experimental ignition probability, regression fit,

along with regression confidence interval can be plotted.

Fig. 4.8: Ethanol ignition data events, experimental probability, regression curve and

95% confidence intervals.

88

In further calibration studies, it may be possible that repeated droplets interact and change

the hot plate surface used in the experiments, despite repeated abrasion and cleaning. As

shown in Fig. 4.9, the ignition of pure ethanol was repeatedly performed throughout the

experiments, due to its ease of handling. It was determined that from the first ethanol

experiment to the last experimental test, ethanol did not significantly deviate from its

initial value. The fifth test did show a decreased value, possibly due to Al2O3

accumulation on the surface. In general, over months of surface testing, the surface

ignition behavior did not deviate significantly.

Fig. 4.9: Repeated ethanol tests over time, solid markers (■) designate the

regression fit, while empty markers the experimental probability (□).

89

4.2 Results and Discussion

4.2.1 Experimental Results

The fuel droplets introduced onto the experimental setup displayed a wide range of

ignition behavior. As previously mentioned, the leidenfrost phenomena was observed for

all fuel mixtures, in which the droplet hovers over an insulating layer of fuel vapor. As

shown in Fig. 4.10, nanoaluminum suspended in ethanol resulted in enhanced ignition.

Fig. 4.10: Ignition probability regression curves of aluminum/ethanol suspensions

compared to pure ethanol.

As shown in Figures 4.11 and 4.12, neat ethanol was observed to start igniting at no less

than 480°C, with a stable flame enveloping the droplet until all of the fuel was consumed.

At low volume fractions of pure Al in ethanol (0.1%, 0.5%), the fuel droplet shrinks

90

axisymmetrically, and either ignites or purely evaporates, leaving agglomeration of pure

Al that ignited at higher hot plate temperatures. Ethanol suspensions with 1%, 2%, and

3% Al volume fractions ignited much lower (100 ºC) than pure ethanol, and exhibited

burning regimes similar to disruptive combustion in slurry droplets.

Fig. 4.11: Regression curve with confidence band for pure Al in ethanol.

Fig. 4.12: Limits of ignition for pure Al in ethanol.

As the droplet burns and evaporates, there is a formation of a glass-like shell from the

91

non-volatile solid left behind; once the droplet ignites there was pronounced spinning and

upward jetting of particle streaks. It has been suggested that this regime enhances local

turbulence and promotes more complete combustion [80]. Furthermore, there was

observed droplet rippling similar to buckling of colloidal suspensions in leidenfrost

boiling [103]. Figures 4.13 and 4.14 display the ignition of Al2O3 mixtures in ethanol.

Fig. 4.13: Regression curve with confidence band for Al2O3 in ethanol.

Fig. 4.14: Limits of ignition for Al2O3 in ethanol.

In all cases, Al2O3 volume fractions ignited at higher temperatures than pure ethanol,

92

igniting at no less than 510°C, thus reducing the ignition probability at the same

temperatures.

The ignition behavior of diesel was also investigated in order to determine its

relative ignition performance on the experimental setup. Diesel fuel mixtures were

significantly more explosive than ethanol, with flames exploding across the surface of the

plate. In addition, diesel overall had the greatest uncertainty in ignition events. Diesel

may exhibit a very different ignition profile due to being mix of complex hydrocarbons,

whereas the ethanol used was primarily one molecule (C2H5OH). There also may be

variations in the fuel depending on the vendor. It is suggested that the lighter, more

volatile components of diesel immediately evaporate, followed by the heavier

components. Starting at temperatures as low as 300°C, there is rapid evaporation of more

volatile components up until ignition events at ~560°C. As mentioned in chapter three,

without any dispersing agent, diesel mixtures could not successfully suspend higher

volume fractions of metallic additives. Experimental results are displayed in Figs. 4.15,

4.16, 4.17 and 4.18 for pure Al and Al2O3 at volume fractions of 0.1 and 0.5%.

Fig. 4.15: Regression curve with confidence band for pure Al in diesel.

93

Fig. 4.16: Limits of ignition for pure Al in diesel.

Fig. 4.17: Regression curve with confidence band for Al2O3 in diesel.

94

Fig. 4.18: Limits of ignition for Al2O3 in diesel.

Overall, diesel mixtures with Al and Al2O3 additives demonstrated somewhat similar

surface ignition temperatures to pure diesel. In comparison to Tyagi‘s [8] experiments

with Al in diesel, the results here do not show comparable enhancement. This may

indicate that enhancement is dependent on the materials used and highly dependent on

the experimental temperature range observed. Several differences in experimental

procedure, such as the use of a syringe pump, a heated cast iron surface, and materials

used may also explain the differences in data. In comparison with ethanol suspensions,

wetting behavior may also play a role, as shown in the next section.

The results suggest that a decrease of ignition temperature may be due to several

underlying mechanisms. As previously mentioned, thermally conductive nanoscale

additives in ethanol are known to significantly enhance thermal transport properties in the

liquid mixture, leading to reduced ignition delays. Similarly, as shown in Fig. 4.12 for

pure Al nanoparticles in ethanol, an increase in ignition probability or decreased ignition

95

temperature may be due to enhanced thermodynamic properties of nanofluids.

Comparing the ignition behavior from Al and Al2O3 in ethanol (Figs. 4.12 and 4.14),

suggests that ignition characteristics may be enhanced due to the reactive nature of the

additive material. In the present study, experimental temperatures are below the melting

point commonly reported for aluminum particles; however, there may be forced (spark)

ignition due to the increased specific surface area and size-dependent melting point

depression found in nanoparticles. Previous studies [15, 16] found depressed melting

temperatures and heat of fusion for nanoscale Al, with ignition temperatures as low as

520 °C. Despite the lowest ignition event of the ethanol/Al mixture being at the

experimental temperature of 390°C, it cannot be ruled out that there was a spark ignition

event. Nonetheless, in this case, it may be possible to enhance or inhibit the ignition of

nanoenergetic liquid fuels by using different additives or by selective passivation of the

additive.

4.2.2 Effect of Surface Tension

Surface tension is also an important parameter for evaporation rate and lifetime [98], due

to determining the contact area between the hotplate and liquid, and altering the

evaporation properties of the liquid-vapor interface. Taking this into consideration, a

reduction in surface tension may facilitate the occurrence of ignition. However, there

have been limited works and lack of consistent studies on the effect of size and

concentration of nanoparticles on the surface tension of the nanofluid. In previous sessile

droplet contact angle studies, Vafaei et al [104] found that with increasing bismuth

telluride (Bi2Te3) nanoparticle concentrations, contact angles of iso-octane mixtures on

96

glass and silicon substrates were found to initially increase and then decrease after critical

concentrations. In a later investigation [105], it was numerically calculated that Bi2Te3

nanofluid suspensions containing 2.5 nm and 10.4 nm diameter in iso-octane significantly

reduced the liquid-gas interface surface energy. The authors propose that the

nanoparticles accumulate and assemble at the liquid–gas interface, where the effective

surface tension can be reduced up to 50% for 2.5 nm particles, compared to the bulk

liquid surface. Tensiometer measurements by Kim et al [106] reported no significant

change in surface tension (±3%) of Al2O3 blended in water, compared to that of pure

water. Furthermore, Sefiane and Bennacer [107] found that passivated Al nanoparticles

suspended in ethanol at 0.005 to 0.02 % volume fractions exhibited no change in surface

tension, however, there was found an enhancement in overall evaporation of ethanol by

extending the fluid film on the substrate. Therefore currently, for nanoparticles suspended

in fluids, contact angles have been confirmed to be affected while further studied on

surface tension are needed.

Therefore, contact angles were determined experimentally for ethanol and

ethanol-aluminum suspensions using a Tantec Inc. contact angle meter (Model CAM-

PLUS MICRO) and with the half-angle method at 22 °C in air. As shown in Fig. 4.19,

and compared to water in Fig. 4.20, nanoaluminum additives in ethanol have a tendency

to reduce the contact angle of the droplets. A value of 5° was assigned for complete

wetting of the substrate surface. A decreasing trend of contact angles was observed for

stainless steel, glass, copper, and acrylic resins for higher pure aluminum loadings.

97

Fig. 4.19: Contact angles of ethanol/aluminum mixtures.

Fig. 4.20: Contact angles of ethanol and water for comparison.

However, similar results were confirmed for nanoaluminum oxide, which implies that the

decrease in contact angle may be due to nanoparticle accumulation around the substrate

film edge; in this case, it may be expected to increase the contact area on a heated

surface. Furthermore, diesel fuel mixtures were completely wetted on all surfaces, for all

additive loadings. Similar wetting results for diesel droplets were observed when the

98

droplets impinged on the hot plate. As a result, based on previous studies and the

experimental contact angles of Fig. 4.19, the presence of nanoparticles suspended in these

fuels affects the hot plate contact area, and could possibly have the ability to alter the

droplet interfacial contact area or liquid–gas interface behavior.

4.2.3 Heterogeneous Nucleation

The importance of heterogeneous nucleation may play a vital role in the enhanced

ignition behavior for higher volume fractions (1%, 2%, and 3%) of ethanol and n-Al. In

this study, compared to the well documented ―microexplosion‖ phenomena [27, 80], a

mild disruptive burning phenomena was observed for Eth + Al, in which there is

shattering or fragmentation of parent droplet into smaller ones. With the vaporization of a

volatile liquid, there is the formation of a porous non-volatile shell on the exterior of the

droplet, with subsequent pressure build up and ignition of the interior liquid. A slurry fuel

droplet schematic is shown in Fig. 4.21,

Fig. 4.21: Schematic of a slurry fuel droplet, with regions A, B, and C [80].

99

Where is the accumulated shell thickness, initial droplet radius, and transient

droplet radius. If the internal droplet circulation is not strong enough, or if more volatile

fuel components are entrapped within the accumulated porous shell, this has been known

to lead to homogeneous nucleation by attaining the limit of superheat of the liquid carrier

[108, 109]. The interior region is superheated above its boiling point, then internal

bubbling, and subsequent rapid gasification of these droplets can be reached,

characterized as ―microexplosive,‖ or disruptive burning.

However, if potential nucleation sites are present, this can lead to heterogeneous

nucleation and internal boiling within the droplet. It is likely that this type of boiling

plays a role in the hot surface ignition of the present experimental studies in this chapter.

In contrast to homogeneous nucleation, as in a uniform liquid boiling, heterogeneous

nucleation sites reduce the free energy barrier to form a gas phase and require less energy

to facilitate nucleation. Thus, the energy barrier to attain heterogeneous nucleation is

smaller than homogeneous nucleation. The total energy barrier needed for homogeneous

nucleation in a uniform spherical droplet is,

(4.3)

Where is the energy barrier to nucleate, is the mass of vapor embryo, is the

Gibbs energy of vapor, is the Gibbs energy of liquid, is the volume of vapor

embryo, is the radius of the vapor embryo, is the interfacial tension. Once the

thermal equilibrium and mechanical equilibrium have been achieved, the energy barrier

will be,

100

(4.4)

For colloids, the distinguishing feature of heterogeneous nucleation is the presence of

foreign solid particles in the liquid medium. In contrast to homogeneous nucleation, the

total energy barrier needed for heterogeneous nucleation is the product of homogeneous

nucleation and a function of the wetting contact angle,

(4.5)

Where is the pressure of liquid and is the pressure of vapor, and is the contact

angle of liquid on the solid surface. Once the thermal and mechanical equilibriums have

been reached, the expression may be reduced to,

(4.6)

Depending on the wetting contact angle, the final term

can range from zero to one. If the contact angle is a

non-zero value, as in most cases, the final f is less than 1. Therefore, the energy

barrier to nucleate for a heterogeneous nucleation demands less energy, and internal

bubbling will occur around potential nucleation sites inside the droplet mixture.

It is most likely that heterogeneously induced internal boiling occurred due to the

presence of foreign solid particles, rather than homogeneous nucleation of the interior

mixture due to internal superheating. With higher loadings of particle additives, it is

likely that there are enough suitable heterogeneous nucleation sites that internal boiling

can be initiated. Consequently, internal heterogeneous nucleation is relevant to the

formation of saturated fuel vapor to drive the combustion event. Furthermore, as the

101

outer region of the droplet becomes concentrated with a less volatile, higher boiling point

porous shell, rapid internal bubbling induces the observed disruptive burning event.

Therefore, there is a critical concentration of pure aluminum additives in the fuel

facilitates the onset of heterogeneous nucleation of the liquid carrier, which increases the

ignition probability at a given temperature. However, the mechanisms have not been fully

studied and more work is needed to understand and characterize heterogeneous

nucleation in fuel suspensions.

4.2.4 Theory of Thermal Droplet Ignition

Similar to spherical metal particle combustion, finite rate chemical kinetics also need to

be considered for liquid droplet ignition. The criteria for thermal droplet ignition for fuel

droplets may be modeled by a dimensionless Damkohler or number, which is the ratio

of the characteristic flow time to the characteristic chemical reaction time, as described in

Eq. 4.7 [109, 110],

(4.7)

For a frequency factor of , specific heat of , thermal conductivity of air , ambient

pressure , droplet radius of , universal gas constant , heat of combustion per unit

mass , stoichiometric fuel to oxygen ratio , and pure fuel molecular weight and

average weight . Once a preliminary number has been defined, the system

Damkohler number ) can take two different forms, one for an upper reactive branch

( , and the other for a weakly-reacting branch ( ). As shown in Fig. 4.22, for

, a frozen state corresponds to pure droplet evaporation, while represents

102

an infinitely fast chemical reaction (equilibrium state), corresponding to the D2-law of

droplet combustion.

Fig. 4.22: Characteristic ―S-curve‖ of ignition and extinction events as a function of a

burning rate parameter [109].

Of particular interest in ignition, is the lower branch, where the burning rate slowly

increases until a critical ignition state is reached ( ). The droplet ignition criteria of Law

[110] states that a droplet can achieve ignition, if the number of the system

exceeds a critical number ( , expressed as,

(4.8)

where,

(4.9)

For a given ambient temperature , an activation temperature with

activation energy , ambient oxidizer mass fraction , effective latent heat of

103

gasification , energy needed for droplet heating per unit mass of fuel gasified

, and specific latent heat of gasification . Where,

(4.10)

is the heat transfer parameter. For small numbers of , can be accurately represented

as [110] as,

for (4.11)

Theoretically, there are two different processes that are investigated, first the there

is transient droplet vaporization and heating from the ambient environment. As the

internal heating is increased, the rate of fuel vapor generation is also increased. At an

instantaneous critical point , there is a transition from pure vaporization of the

droplet to active chemical reactions in the surrounding gas-phase, favoring increasing

pressure, temperature, and oxidizer concentration for ignition. Finally, there is an

established gas-phase reaction after which the droplet is said to be ignited. An energy

conservation on the droplet surface yields [109], for a mass gasification rate of ,

(4.12)

or equivalently as,

(4.13)

In the first combustion stage, internal heating with little gasification, the Eq. 4.12 may be

reduced to,

(4.14)

In the second stage, there is little droplet heating with significant vaporization,

(4.15)

104

The mass gasification rate in the first stage of combustion (Eq. 4.14), can be expected to

be larger due to an experimentally confirmed enhanced thermal conductivity , or

effective thermal conductivity in the nanofluid. When considering the thermal

conductivity of ethanol 0.171 (W/mK), and its corresponding fuel additives Al (250

W/mK) and Al2O3 (30 W/mK), enhanced thermal transport may play an important role in

triggering ignition event by rapidly heating the mixture. As a result, there may be

enhanced internal heating and mass transfer rate away from the droplet surface, to initiate

the second combustion stage.

4.3 Conclusions

In the present study, the ignition behavior of ethanol and diesel with nanoaluminum and

nanoaluminum oxide additives were experimentally investigated with hot plate ignition,

and several possible underlying mechanisms are explored. Depending on the

experimental conditions, liquid fuel, and nanoparticle additive, colloidal fuel suspensions

of metallic additives may undergo hot surface ignition at lower temperatures than the

bulk fuel. Therefore, this implies that hot plate ignition probability is highly coupled with

the experimental procedure. As shown in Fig. 4.9, the ignition of ethanol was repeatedly

performed throughout the experiments, and was determined to not significantly deviate

from its initial value. However, it may be possible that within testing one fuel sample,

previous droplets may affect future droplets at higher temperature ranges.

There was found a strong correlation of minimum surface ignition temperatures

with volume fraction concentration and fuel type. Theories of heterogeneous nucleation

and thermal ignition for fuel droplets were also considered. Due to several possible

105

mechanisms, it was found that the greatest ignition enhancement was with aluminum

energetic additives dispersed in ethanol; however, the data suggest that the primary

mechanisms responsible for combustion are within the heat up phase via thermal

transport enhancements within the fuel, or due to heterogeneous nucleation of the

particles present. Therefore, the ignition enhancement in ethanol may be due to a

combination of several other contributing factors: 1) the reactive nature of aluminum, 2)

the ability of ethanol to suspend additives at higher (>2%) volume fractions, 3) enhanced

thermal transport properties, and 4) changing of properties at the liquid-vapor interface.

As a result, along with energy density enhancement, achieving precise control over the

reactivity of nanofluids is an opportunity for future nanoenergetic fuel applications.

106

Chapter Five

Conclusions and Future Work

5.1 Concluding Remarks

In this study, it was experimentally determined that the combustion energetics and

ignition probability of liquid fuel ethanol can be modified with nanoparticle additives. On

the basis of these two experimental observations, along with those available in literature

[6, 26, 27], it was concluded that nanoaluminum suspended in ethanol has the ability to

tailor the reactivity and energetic content of pure ethanol and is relatively easy to

synthesize and handle. Furthermore, the oxide layer has a significant effect on reaction

energetics and aluminum nanoparticles may be stably suspended in ethanol very well.

Investigating new and currently used fuels with nanoenergetics has demonstrated to be

promising in terms of maximizing the efficiency and minimizing the environmental

impact. This study also provides basic research on the combustion of nanofluids, which is

currently an area with relatively little investigative work. Further investigations into the

ignition mechanisms of nanofluids may provide insight into their impact as a viable fuel

in the future. At the present time, there is a need to a unified fundamental understanding

of nanoenergetic additives within colloidal dispersion compositions.

107

5.2 Extensions of Research

Extensions of the current work include the following:

1) Future work could investigate heavier weight loadings of n-Al with the use of

dispersant, and identifying the most effective surfactant for long-term fuel

suspension stability. Furthermore, more advanced materials analysis of the

combustion products, such as X-ray Photoelectron Spectroscopy (XPS), can

determine the actual molecular formula and phase (i.e. - α-Al2O3 or γ-Al2O3) in

addition to the element compositions.

2) Combustion characteristics of different sizes, size distributions, and shapes of

nanomaterials can be investigated. It has been shown that for a given shape, the

surface to volume ratio increases almost linearly with decreasing size, however,

most nanoparticle materials are spherical, which is the smallest surface area

among any surface enclosing a given volume. Thus, different shapes, such as

flakes, that are available may exhibit vastly different behavior. As a result, further

experimental studies on more complex nanostructured fuels, such as nanotubes or

nanorods that take advantage of self-assembly properties of nanoparticles can be

studied.

3) The ignition behavior of powders and fuel materials can also be investigated by

heated wire filaments. Wires can be coated with metallic particles and optically

monitored to determine the ignition temperature of the metal. Previous studies by

Shoshin et. al. [2, 111] determined the ignition temperature of mechanically

alloyed Al-Ti and Al-Mg powders by optically monitoring a heated nichrome

wire with powder coatings. A simple car battery, or multiple connected in series,

108

can provide the DC current through the filament, and a resistor network or

autotransformer can vary the heating rate. Uncoated portions of the wire are then

monitored by an infrared pyrometer to determine the temperature history.

However, the finite thickness of the powder on the wire may introduce error.

Preliminary experiments of the technique in the laboratory have shown that it is

possible to use nickel- chromium fuse wire, a 2901 ignition unit, and a variac

control power supply through the wire. There is interest in experimentally

determining the ignition temperature of nanoparticles, or the ignition temperature

of a fuel formulation, such as a nanofluid.

4) In constant volume experiments, it is possible to determine the pressure history of

the combustion event. The maximum rate of pressure rise is generally

proportional to the flame speed and serves as an indicator of the burn rate [77]. By

fitting a pressure transducer on to a combustion (calorimeter) vessel, it is possible

to estimate the burning rate of fuel samples.

5) Regarding thermochemistry calculations, other CEA combustion modules can be

used, such as Chapman-Jouguet detonations or ―rocket‖ problems in order to

determine theoretical rocket parameters, such as specific impulse. Typical

conditions for a combustion chamber could be 1000 psia and a nozzle expansion

ratio of 40.

6) Other proposed future work may entail flow visualization, a high-speed camera or

video system can visualize droplet lifetime and ignition events, in comparison

with micron-sized aluminum samples, or other metals. In addition, a combustion

109

finite difference model or the incorporation of a micron/nano particle combustion

model into commercial simulation software could be attempted.

110

References

1) E. L. Dreizin, ―Metal-based reactive nanomaterials,‖ Progress in Energy and

Combustion Science 35 (2009) 141–167.

2) Y. L. Shoshin, and E. L. Dreizin, ―Particle combustion rates for mechanically

alloyed Al–Ti and aluminum powders burning in air,‖ Combustion and Flame 145

(2006) 714–722.

3) D. D. Dlott, ―Thinking big (and small) about energetic materials: a review,‖

Materials Science and Technology 2006, Vol. 22, No. 4.

4) A. Dokhan, E.W. Price, J.M. Seitzman and R.K. Sigman, ―The effects of bimodal

aluminum with ultrafine aluminum on the burning rates of solid propellants,‖

Proceedings of the Combustion Institute Volume 29, Issue 2, 2002, p. 2939-2946.

5) K. Jayaraman, K. V. Anand, S. R. Chakravarthy, and R. Sarathi, ―Effect of nano-

aluminium in plateau-burning and catalyzed composite solid propellant

combustion,‖ Combustion and Flame 156 (2009) p. 1662–1673.

6) C. Allen, G. Mittal, C. Sung, E. Toulson, and T Lee, ―An aerosol rapid

compression machine for studying energetic-nanoparticle-enhanced combustion

of liquid fuels,‖ Proceedings of the Combustion Institute, Volume 33, Issue 2,

2011, Pages 3367-3374.

111

7) V. Sajith, C. B. Sobhan, and G. P. Peterson, ―Experimental Investigations on the

Effects of Cerium Oxide Nanoparticle Fuel Additives on Biodiesel,‖ Advances in

Mechanical Engineering, 6 pages Volume 2010 (2010), Article ID 581407, p. 6.

8) H. Tyagi, P. E. Phelan, R. Prasher, R. Peck, T. Lee, J. R. Pacheco, and P.

Arentzen, ―Increased Hot-Plate Ignition Probability for Nanoparticle-Laden

Diesel Fuel,‖ Nano Lett., 2008, 8 (5), p. 1410–1416.

9) D. Wen, ―Nanofuel as a potential secondary energy carrier,‖ Energy &

Environmental Science (May 2010), 3 (5), p 591-600.

10) ―Metal Nanopowders High-Energy Materials and Processes Advanced Powder

Technologies LLC,‖ http://www.nanosized-powders.com/en/35/. Accessed March

21, 2011.

11) J. L. Sabourin, D. M. Dabbs, R. A. Yetter, F. L. Dryer and I. A. Aksay,

―Functionalized Graphene Sheet Colloids for Enhanced Fuel/Propellant

Combustion,‖ ACS Nano, 3 (12), 2009, p. 3945–3954.

12) W. Miziolek, ―Nanoenergetics: An emerging technology area of national

importance,‖ The AMPTIAC Newsletter, Vol. 6(1) (2002) 43-48.

13) C. Rossi, K. Zhang, D. Estève, P. Alphonse, P Tailhades, and C. Vahlas,

―Nanoenergetic Materials for MEMS: A Review,‖ J. Micromech. Sys. 16 (2007)

p. 919-931.

14) J. Eckert, J.C. Holzer, C.C. Ahn, Z. Fu and W.L. Johnson, ―Melting Behavior of

Nanocrystalline Aluminum Powders,‖ Nanostructured Materials Vol. 2, pp. 407-

413, 1993.

112

15) S. L. Lai, J. R. A. Carlsson, and L. H. Allen, ―Melting point depression of Al

clusters generated during the early stages of film growth: nanocalorimetry

measurements,‖ Appl. Phys. Lett. 72, p. 1098 (1998).

16) J. Sun, S. L. Simon, ―The Melting Behavior of Aluminum Nanoparticles,‖

Thermochimica Acta 463 (2007) p. 32–40.

17) M. A. Trunov, S. M. Umbrajkar, M. Schoenitz, J. T. Mang, and E. L. Dreizin,

―Oxidation and Melting of Aluminum Nanopowders,‖ J. Phys. Chem. B, 2006,

110 (26), pp 13094–13099.

18) Y. Huang, G. A. Risha, V. Yang, R. Yetter, ―Combustion of bimodal

nano/micron-sized aluminum particle dust in air,‖ Proceedings of the Combustion

Institute 31 (2007) 2001-2009.

19) K. Park, D. Lee, A. Rai, D. Mukherjee, M.R. Zachariah, ―Size-Resolved Kinetic

Measurements of Aluminum Nanoparticle Oxidation with Single Particle Mass

Spectrometry,‖ Journal of Physical Chemistry B 109 (2005) 7290–7299.

20) T. Bazyn, H. Krier, and N. Glumac, ―Evidence for the transition from the

diffusion-limit in aluminum particle combustion,‖ Proceedings of the Combustion

Institute, Volume 31, Issue 2, January 2007, Pages 2021-2028.

21) B. Rufino, F. Boulc‘h, M.-V. Coulet, G. Lacroix, and R. Denoyel, ―Influence of

particles size on thermal properties of aluminum powder,‖ Acta Materialia 55

(2007) p. 2815–2827.

22) C. H. Li, and G. P. Peterson, ―Experimental Investigation of Temperature and

Volume Fraction Variations on the Effective Thermal Conductivity of

Nanoparticle Suspensions (Nanofluids),‖ J. Appl. Phys., 99, (2006) p. 084314.

113

23) S. Krishnamurthy, P. Bhattacharya, and P. E. Phelan, ―Enhanced Mass Transport

in Nanofluids,‖ Nano Lett., 2006, 6 (3), p. 419–423.

24) Choi, S. U. S.; Zhang, Z. G.; Yu, W.; Lockwood, and F. E.; Grulke, ―Anomalous

thermal conductivity enhancement in nanotube suspensions,‖ E. A. Applied

Physics Letters, Volume 79, Issue 14, id. 2252 (2001).

25) W. Yu, D. M. France, J. L. Routbort, and S. U. S. Choi, ―Review and Comparison

of Nanofluid Thermal Conductivity and Heat Transfer Enhancements,‖ Heat

Transfer Engineering, 29(5):432–460, 2008.

26) S. Fu, L. Yu, Y. Hu, C. Wang, X.Hu, ―Impact of Carbon Nanotube Suspensions

on the Hotplate Ignition of Liquid Fuels,‖ IEEE 978-1-4244-7739-5/10 (2010).

27) Y. Gan, and L. Qiao, ―Combustion characteristics of fuel droplets with addition of

nano and micron-sized aluminum particles,‖ Combustion and Flame Volume 158,

Issue 2, February 2011, Pages 354-368.

28) United States Environmental Protection Agency, ―Hydrazine Hazard Summary,‖

Created in April 1992; Revised in January 2000. Retrieved on March 7th, 2011.

29) G. Ritchie, K. Still, J. Rossi, M. Bekkedal, A. Bobb, and D. Arfsten, ―Biological

And Health Effects Of Exposure To Kerosene-Based Jet Fuels And Performance

Additives,‖ Journal of Toxicology and Environmental Health, Part B, Volume 6,

Issue 4 January 2003 , pages 357 – 45.

30) J. L. Sabourin, R. A. Yetter, B. W. Asay, J. M. Lloyd, V. E. Sanders, G. A. Risha,

and S. F. Son, ―Effect of Nano-Aluminum and Fumed Silica Particles on

Deflagration and Detonation of Nitromethane,‖ Propellants Explos. Pyrotech. 34,

2009, p. 385 – 393.

114

31) J.L. Sabourin, G.A. Risha, R.A. Yetter, S.F. Son, and B.C. Tappan, ―Combustion

characteristics of nanoaluminum, liquid water, and hydrogen peroxide mixtures,‖

Combustion and Flame 154 (2008) 587–600.

32) G.A. Risha, S.F. Son, R.A. Yetter, V. Yang, and B.C. Tappan, ―Combustion of

nano-aluminum and liquid water,‖ Proceedings of the Combustion Institute 31

(2007) 2029–2036.

33) G. A. Risha, T. L. Connell Jr., R. A. Yetter, V. Yang, T. D. Wood, M. A. Pfeil, T.

L. Pourpoint, and S. F. Son, ―Aluminum-Ice (ALICE) Propellants for Hydrogen

Generation and Propulsion,‖ 45th AIAA/ASME/SAE/ASEE Joint Propulsion

Conference & Exhibit.

34) von Grosse, A., and Conway, ―Combustion of Metals in Oxygen,‖ J. B., Indust.

Eng. Chem. 50:663 (1958).

35) I. Glassman, ―Combustion,‖ 3rd ed., Academic Press, Orlando, 1996 (Chapter 9).

36) S.H. Fischer and M.C. Grubelich, ―Theoretical energy release of thermite,

intermetallic, and combustible metals,‖ International Pyrotechnics Seminar,

(1998), SAND98-1176C.

37) G. R. Astbury, ―A review of the properties and hazards of some alternative fuels,‖

Process Safety and Environmental Protection Volume 86, Issue 6, November

2008, p. 397-414.

38) G.M. Faeth, ―Status of Boron Combustion Research, Specialist Meeting on Boron

Combustion,‖ Air Force Office of Scientific Research, Bolling Air Force Base,

Washington, D.C. (June 1984).

115

39) A. Ulas, K. K. Kuo, and C. Gotzmer, ―Ignition and Combustion of Boron

Particles in Fluorine-Containing Environments,‖ Combustion and Flame

127:1935–1957 (2001).

40) F. Franzoni, M. Milani, L. Montorsi, and V. Golovitchev, ―Combined hydrogen

production and power generation from aluminum combustion with water: analysis

of the concept,‖ International Journal of Hydrogen Energy Volume 35, Issue 4,

February 2010, p. 1548-1559.

41) M.W. Beckstead, ―A summary of aluminum combustion,‖ RTO/VKI special

course on ―internal aerodynamics in solid rocket propulsion‖. Report number

RTO-EN-023 (2002).

42) E. L. Dreizin, ―Experimental Study of Stages of Aluminum Particle Combustion

in Air,‖ Combustion and Flame 105:541-556 (1996).

43) A. Rai, K. Park, and L. Zho, ―Understanding the mechanism of aluminum

nanoparticle oxidation,‖ Combustion Theory and Modeling Vol. 10, No. 5,

October 2006, p. 843–859.

44) B. Palaszewski, J. Jurns, K. Breisacher and K. Kearns, ―Metalized Gelled

Propellants Combustion Experiments in a Pulse Detonation Engine,‖

NASA/TM—2006-214119.

45) F. Dryer and R Sawyer, ―Physical and Chemical Aspects of Combustion: A

Tribute to Irvin Glassman,‖ Taylor & Francis; 1 edition (1997).

46) I. Glassman, ―Comment on ‗The Combustion Phase of Burning Metals‘ By

T.A.Steinberg, D.B.Wilson, and F.Benz,‖ Combustion and Flame 93:338-342

(1993).

116

47) R. A. Yetter, G. A. Risha, and S. F. Son, ―Metal particle combustion and

nanotechnology,‖ Proceedings of the Combustion Institute Volume 32, Issue 2,

2009, p. 1819-1838.

48) J. W. Reynolds, ―The Element Potential Method for Chemical Equilibrium

Analysis: Implementation in the Interactive Program,‖ STANJAN; Department of

Mechanical Engineering: Stanford University: Stanford, CA, 1986.

49) B.J. McBride and S. Gordon, ―Computer Program for Calculation of Complex

Chemical Equilibrium Compositions and Applications,‖ NASA-RP-1311, 1996.

50) M.W. Chas, C.A. Davies, J.R. Davies, D.J. Fulrip, R.A. McDonald, and A.N.

Syverud, A.N., ―JANAF Thermochemical Tables,‖ J. Phys. Chem. Ref. Data 14

(1985).

51) P. Bucher, R. A. Yetter, F. L. Dryer, E. P. Vicenzi, T. P. Parr, and D. M. Hanson-

Parr, ―Condensed-phase Species Distributions about Al Particles Reacting in

Various Oxidizers,‖ Combustion and Flame 117:351–361 (1999).

52) B. Legrand, M. Marion, C. Chauveau, I. Gokalp, and E. Shafirovich, ―Ignition

and Combustion of Levitated Magnesium and Aluminum Particles in Carbon

Dioxide,‖ Combust. Sci. Technol. 165 (1-6) (2001) 151–174.

53) A. Kanury, ―Introduction to Combustion Phenomena,‖ Gordon and Breach

Science Publishers (1995).

54) Spalding, D. B, ―The burning of liquid fuels,‖ Proc. Combust. Inst. 4, 847–864.

55) G. A. E. Godsave, ―Studies of the combustion of drops in a fuel spray: The

burning of single drops of fuel,‖ .Proc. Combust. Inst.4, 818–830 (1953).

117

56) I. Glassman and R. A. Yetter, ―Combustion,‖ 4th ed., Academic Press, Orlando,

2008 (Chapter 9).

57) F. Tang, S. Goroshin, and A.J. Higgins, ―Modes of particle combustion in iron

dust flames,‖ Proceedings of the Combustion Institute 33 (2011) 1975–1982.

58) K. J. Klabunde, J. Stark, O. Koper, C. Mohs, D. G. Park, S. Decker,Y. Jiang, I.

Lagadic, and D. Zhang, ―Nanocrystals as Stoichiometric Reagents with Unique

Surface Chemistry,‖ J. Phys. Chem. 1996, 100, 12142-12153.

59) W. H. Qi, and M. P. Wang, ―Size effect on the cohesive energy of nanoparticle,‖

Journal of Materials Science Letters 21, 2002, 1743-1745.

60) S. L. Lai, J. Y. Guo, V. Petrova, G. Ramanath, and L. H. Allen, ―Size-Dependent

Melting Properties of Small Tin Particles: Nanocalorimetric Measurements,‖ Phys

Rev. Lett. 77, p. 99–102 (1996).

61) S. Alavi, ―Molecular Dynamics Simulations of the Melting of Aluminum

Nanoparticles,‖ J. Phys. Chem. A 2006, 110, 1518-1523.

62) S. Li, H. Huang, Y. B. Zhang, C. Q. Sun, Y. Wang, and B. K. Tay, ―Correlation

between the Melting Point of a Nanosolid and the Cohesive Energy of a Surface

Atom,‖ J. Phys. Chem. B 2002, 106, 10701-10705.

63) M. Alcoutlabi and G. B McKenna, ―Effects of confinement on material behaviour

at the nanometre size scale,‖ J. Phys.: Condens. Matter 17 (2005) R461–R524.

64) S. G. Jennings, ―The mean free path in air,‖ J. Aerosol Sci., Vol 19, No. 2, pp.

159-166, 1988.

118

65) S. Mohan, M. A. Trunov, and E. Dreizin, ―On the possibility of vapor-phase

combustion for fine aluminum particles,‖ Combustion and Flame 156 (2009)

2213-2216.

66) S. W. Chung, E. A. Guliants, C. E. Bunker, D. W. Hammerstroem, Y. Deng, M.

A. Burgers, P. A. Jelliss, and S. W. Buckner, ―Capping and Passivation of

Aluminum Nanoparticles Using Alkyl-Substituted Epoxides,‖ Langmuir 2009,

25(16), 8883–8887.

67) B. Hvolbæk1, T. V. W. Janssens, B. S. Clausen, H. Falsig, C. H. Christensen, and

J. K. Nørskov, ―Catalytic activity of Au nanoparticles,‖ Nanotoday, August 2007,

Volume 2, Number 4.

68) A. Rai, D. Lee, K. Park, and M. R. Zachariah, ―Importance of Phase Change of

Aluminum in Oxidation of Aluminum Nanoparticles,‖ J. Phys. Chem. B 2004,

108, 14793-14795.

69) P. Puri and V. Yang, ―Multi-Scale Modeling of Nano Aluminum Particle Ignition

and Combustion,‖ The Pennsylvania State University, NEEM MURI Presentation

November 16, 2005.

70) I. Levin and D. Brandon, ―Metastable Alumina Polymorphs: Crystal Structures

and Transition Sequences,‖ J. Am. Ceram. Soc., 81 [8] 1995–2012 (1998).

71) M. A. Trunov, M. Schoenitz, X. Zhu and E. L. Dreizin, ―Effect of polymorphic

phase transformations in Al2O3 film on oxidation kinetics of aluminum powder,‖

Combustion and Flame. Volume 140, Issue 4, March 2005, p. 310-318.

119

72) M. W. Beckstead, Y. Liang, and K. V. Pudduppakkam. ―Numerical Simulation of

Single Aluminum Particle Combustion (Review),‖ Combustion, Explosion, and

Shock Waves, Vol. 41, No. 6, pp. 662-638, 2005.

73) A. Rai, K. Park, and L. Zho, ―Understanding the mechanism of aluminum

nanoparticle oxidation,‖ Combustion Theory and Modelling Vol. 10, No. 5,

October 2006, p. 843–859.

74) Valery I. Levitas, ―Melt dispersion mechanism for fast reaction of

nanothermites,‖ Appl. Phys. Lett. 89, 071909 (2006).

75) T. Bazyn, H. Krier, and N. Glumac, ―Combustion of nanoaluminum at elevated

pressure and temperature behind reflected shock waves,‖ Combust. Flame 145

(2006) 703-713.

76) S. Goroshin, J. Mamen, A. Higgins, T. Bazyn, N. Glumac, and H. Krier,

―Emission spectroscopy of flame fronts in aluminum suspensions,‖ Proceedings

of the Combustion Institute 31 (2007) 2011-2019.

77) C. Stamatis, X. Jiang, E. Beloni, and E. L. Dreizin, ―Aluminum Burn Rate

Modifiers Based on Reactive Nanocomposite Powders,‖ Propellants Explos.

Pyrotech. 2010, 35, 260 – 267.

78) F. S. Blomshield, ―Lessons Learned In Solid Rocket Combustion Instability,‖

American Institute of Aeronautics and Astronautics, Missile Sciences Conference,

Monterey, CA November 2006.

79) P. M. Ajayan, M. Terrones, A. de la Guardia, V. Huc, N. Grobert, B. Q. Wei, H.

Lezec, G. Ramanath, and T. W. Ebbesen, ―Nanotubes in a Flash--Ignition and

Reconstruction,‖ Science 26 April 2002: Vol. 296. no. 5568, p. 705.

120

80) D. Byun, S. Baek, and J. H. Cho, ―Microexplosion of aluminum slurry droplets,‖

International Journal of Heat and Mass Transfer 42 (1999) 4475-4486.

81) A. Ingenito and C. Bruno, ―Using Aluminum for Space Propulsion,‖ AIAA paper,

Vol. 20, No. 6, Journal of Propulsion and Power, Copyright Clearance Center,

Danvers, November-December 2004.

82) J. Petrovic and G. Thomas, ―Reaction of Aluminum with Water to Produce

Hydrogen,‖ A Study of Issues Related to the Use of Aluminum for On-Board

Vehicular Hydrogen Storage, U.S. Department of Energy, 2008.

83) F. M. Ali, W. Mahmood Mat Yunus, M. M. Moksin, and Z. A. Talib, ―The effect

of volume fraction concentration on the thermal conductivity and thermal

diffusivity of nanofluids: Numerical and experimental,‖ Review of Scientific

Instruments 81, 074901, 2010.

84) Peng Hu, Wan-Liang Shan, Fei Yu, and Ze-Shao Chen, ―Thermal Conductivity of

AlN–Ethanol Nanofluids,‖ Int. J. Thermophys (2008) 29:1968–1973.

85) C. Reuter, ―The V2 and the German, Russian and American Rocket Program,‖

S.R. Research & Publishing; 2 edition (May 2000).

86) N. M. Marinov, ―A Detailed Chemical Kinetic Model for High Temperature

Ethanol Oxidation,‖ International Journal of Chemical Kinetics, Volume 31, Issue

3, pages 183–220, 1999.

87) Renewable Fuels Association, ―2011 Ethanol Industry Outlook,‖ (2011).

88) Operating Instruction Manual, No. 204M. Parr. Oxygen Bomb Calorimeter, 2008.

89) Analytical Methods for Oxygen Bombs, No. 207M. Parr. Oxygen Bomb

Calorimeter, 2004.

121

90) ASTM Standard D240, 2009, ―Standard Test Method for Heat of Combustion of

Liquid Hydrocarbon Fuels by Bomb Calorimeter," ASTM International, West

Conshohocken, PA, 2003, DOI: 10.1520/D0240-09, www.astm.org.

91) J. Coops, R.S. Jessup, and K. Van Nes, ―Calibration of Calorimeters for Reactions

in a Bomb at Constant Volume,‖ in: F.D. Rossini (Ed), Experimental

Thermochemistry Interscience, vol. 1, New York, 1956, p. 27–58 (Chapter 3).

92) Ink Jet, ―How to disperse and stabilize pigments,‖

http://www.inkline.gr/inkjet/newtech/tech/dispersion/. Accessed March 21, 2011.

93) P. C. Hidber, T. J. Graule, and L. J. Gauckler, ‗‗Citric Acid — A Dispersant for

Aqueous Alumina Suspensions,‘‘ J. Am. Ceram. Soc., 79 [7] p. 1857–67 (1996).

94) M. Menon, S. Decourcelle, S. Ramousse, and P. H. Larsen, ―Stabilization of

Ethanol-Based Alumina Suspensions,‖ J. Am. Ceram. Soc., 89 [2] p. 457–464

(2006).

95) ASTM E 659-78, ―Standard Test Method for Autoignition Temperature of Liquid

Chemicals,‖ ASTM, West Conshohocken, PA, 1994.

96) ―Wendell Hull and Associates – Auto Ignition Temperature,‖ Autoignition

Temperature, http://www.wendellhull.com/autoignition_temperature.php,

Accessed March 21, 2011.

97) J. M. Bennett, ―Ignition of Combustible Fluids by Heated Surfaces,‖ Process

Safety Progress (Vol. 20, No.1).

98) J.A. Sanchidrian, L.M. Lopez, and P. Segarra, ―The influence of some blasting

techniques on the probability of ignition of firedamp by permissible explosives,‖

Journal of Hazardous Materials 155 (2008) 580–589.

122

99) J. D. Colwell and A. Reza, ―Hot Surface Ignition of Automotive and Aviation

Fluids,‖ Fire Technology, 41, 105–123, 2005.

100) M. Jones, C. H. Li, A. A. Afjeh, and G. P. Peterson, ―Experimental Study of

Combustion Characteristics of Nanoscale Metal and Metal Oxide Additives in

Biofuel (Ethanol),‖ Nanoscale Research Letters 2011, 6:246.

101) R. Chen, T. X. Phuoc, and D. Martello, ―Effects of nanoparticles on nanofluid

droplet evaporation,‖ International Journal of Heat and Mass Transfer 53 (2010)

3677–3682.

102) D.W. Hosmer and S. Lemeshow, ―Applied Logistic Regression,‖ 1st ed., New

York: John Wiley & Sons, 1989.

103) G. Marty and N. Tsapisa, ―Monitoring the buckling threshold of drying colloidal

droplets using water-ethanol mixtures,‖ Eur. Phys. J. E 27, 213–219 (2008) DOI

10.1140/epje/i2008-10375-6.

104) S Vafaei, T Borca-Tasciuc, M Z Podowski, A Purkayastha, G Ramanath and P M

Ajayan, ―Effect of nanoparticles on sessile droplet contact angle,‖

Nanotechnology 17 (2006) 2523–2527.

105) S. Vafaei, A. Purkayastha, A. Jain, G. Ramanath and T. Borca-Tasciuc, ―The

effect of nanoparticles on the liquid–gas surface tension of Bi2Te3 nanofluids,‖

Nanotechnology 20 (2009) 185702 (6pp) doi:10.1088/0957-4484/20/18/185702

106) S. J. Kim, I. C. Bang, J. Buongiorno, and L. W. Hu, ―Effects of nanoparticle

deposition on surface wettability influencing boiling heat transfer in nanofluids,‖

Applied Physics Letters 89, 153107, 2006.

107) K. Sefiane, R. Bennacer, ―Nanofluids droplets evaporation kinetics and wetting

123

dynamics on rough heated substrates,‖ Advances in Colloid and Interface Science

147–148 (2009) 263–271.

108) C. K. Law, ―Internal Boiling and Superheating in Vaporizing Multicomponent

Droplets,‖ AlChE Journal (Vol. 24, No. 4) 1978.

109) C. K. Law, ―Recent Advances in Droplet Vaporization and Combustion,‖ Prog,

Energy Combust. Sci., Vol. 8. pp. 171-201.

110) C. K. Law, ―Theory of Thermal Ignition in Fuel Droplet Burning,‖ Combustion

and Flame 31, 285-296 (1978).

111) Y. L. Shoshin, R. S. Mudryy, and E. L. Dreizin, ―Preparation and

Characterization of Energetic Al-Mg Mechanical Alloy Powders,‖ Combustion

and Flame 128:259-269 (2002).

112) Jay L. Devore, ―Probability and Statistics for Engineering and the Sciences,‖ 7th

ed., Duxbury Press, 2008.

124

Appendix A

Sample CEA Output File for Ethanol in Air

The following sample output file was generated by NASA‘s CEA2 chemical equilibrium

package. A separate program, CEAgui, provides a graphical user interface and prepares

the output file for viewing.

125

126

127

128

129

Appendix B

Logistic Regression

A logistic regression model was utilized to statistically quantify the ignition response

from a surface temperature increase. The dichotomous (binary) variable y, with values of

1 and 0 corresponding to a success response or a fail response, is measured at each

temperature. In simple linear regression, the method of least squares is the standard

approach to approximate the best-fit curve, where the best-fit line y is

that which minimizes the sum of squared deviations, with regression coefficients , ,

and the random error. In this case, the best-fit line may take the values of -∞ to ∞.

However, for binary data, there are some difficulties in using linear and multiple

(polynomial) regression techniques [112]:

For binary data, there is a constraint on the response range; hence, the probability

may only take the values between 0 to 1.

There is a nonlinear probabilistic relationship between y and x.

There is a non-constant variance ( ) in the response (heteroscedasticity); one of

the underlying assumptions of least square techniques are constant variance across

all observations.

130

Finally, there are non-normal error terms; normal error terms are also an

assumption of least squares fits.

Therefore, with binary data sets, there is a need for more suitable regression techniques

such as weighted least squares or nonlinear least squares. Logistic regression is just one

of the techniques to fit binary data. It is useful to use link functions to transform

variables, so that the relationship between the new variables is linear. In this case, a

logistic regression approach was used where the link function is the log of the odds ratio.

The odds ratio is defined as the probability of an event over the probability of

failure, and in logistic regression literature, the probability of a dichotomous outcome is

generally referred to as ,

(B.1)

The probability in nonlinear with respect to y, however, taking the log of the odds

ratio, it is possible to define a new function to achieve linearity.

g (B.2)

Once the regression coefficients are determined, and once the odds ratio is ―log

transformed,‖ the functional relationship is linear, as shown in Fig. B.1,

131

Fig. B.1: Experimental ignition data of pure ethanol (linear model).

With algebraic manipulation it is now possible to calculate the probability , defined

as,

(B.3)

Where represents the conditional probability of , given x, or the

probability that there is an ignition event, at a designated temperature. In this case, the

ignition response is constrained between the values of 0 or 1; therefore the regression

curve is restricted between 0 and 1. To determine the coefficients and , an iterative

maximum likelihood technique is performed. Most statistical packages such as

MINITAB or JMP generate the coefficient estimates along with other useful inferences;

the coefficients and were determined from MINITAB, and the

maximum likelihood procedure was performed in Excel to verify the constants. Finally,

the experimental data, regression curve, and confidence interval can then be plotted as

shown in Fig 4.8.

132

To determine the confidence intervals, a variance summation law must be used

for the variance of a sum [112]. In this case, the two variables are interdependent and the

following variance operator for the function g is used,

(B.4)

Where denotes the covariance of the function g calculated in the maximum

likelihood technique, or reported from a statistical package. To determine the standard

deviation, the square root of the variance is calculated,

(B.5)

Therefore, the upper and lower confidence intervals of the function are,

where Z represents the standard normal density curve. Finally, the confidence intervals

can then be incorporated into the logit model , to calculate the confidence intervals

of the entire logistic curve, as,

(B.6)

or as,

(B.7)