ABSTRACT 
 
Title:   METALLIC NANO-PARTICLES AS FUEL ADDITIVES 
IN AIR-BREATHING COMBUSTION 
 
   Gregory Young, Doctor of Philosophy, 2007 
 
Directed By:  Associate Professor, Dr. Kenneth H. Yu, Department of 
Aerospace Engineering   
 
 
 
For volume limited propulsion systems, the use of metals as fuel additives offers 
to increase performance by increasing the volumetric energy stored in the fuel.  High-
speed airbreathing combustors generally have very short residence times which do not 
allow for complete combustion of traditional metal particles such as aluminum and 
boron, which are typically larger than 1 ?m.  The development of nano-scale metals has 
opened up the possibility of utilizing metals in high-speed propulsion systems. 
This research effort was conducted in order to determine the viability of utilizing 
metallic nanoparticles as fuels or fuel supplements for high-speed airbreathing 
applications.  The study was broken into two main parts; the first was a fundamental 
investigation into the combustion behavior of boron nanoparticles in a controlled setting, 
while the second was a demonstration of the performance potential of the nanoparticles in 
a realistic airbreathing combustor configuration.     
Initially, the combustion behavior of boron nanoparticles was studied in a 
controlled flame environment to extract basic combustion parameters, such as ignition 
criterion and burning time as functions of surrounding temperature and oxygen 
concentration.  Ensemble average burning times of boron nanoparticles were obtained for 
the first time, while the range of ignition time data was extended into a lower temperature 
range.  The burning time results were compared to both diffusion and kinetic limited 
theories of particle combustion.  It was found that the size dependence on particle 
burning times did not follow either theory.  A kinetic limited burning time correlation has 
been proposed based upon Arrhenius parameters extracted in this study.   
Finally, the fundamental combustion behavior of both boron and aluminum 
nanoparticles was studied extensively in an airbreathing combustor simulating ramjet 
conditions in an effort to determine the viability of using metallic nanoparticles as fuels 
or fuel supplements in high-speed airbreathing propulsion system.  Experiments were 
conducted with hydrocarbon only (ethylene), mixtures of ethylene and boron, and 
mixtures of ethylene and aluminum.  The oxidation of the metals was studied through the 
emission of BO2 and AlO at wavelengths of 546 nm and 488 nm respectively.  
Temperature measurements inside the combustor using thermocouples were made in 
order to determine boundaries in which the addition of boron or aluminum provided a 
positive thermal output for a variety of equivalence ratios ranging from 0.52 to 0.7, metal 
loadings ranging from 9.7 to 15.2% by weight, and combustor residence times ranging 
from 6 to 10.5 milliseconds (combustor inlet velocities of 40 ? 70 m/s).  Both BO2 
emission data and temperature measurements indicated that a critical temperature exists 
for sustained combustion of the boron particles.  Tests with measured peak temperatures 
below 1700 K indicated no benefit of boron addition, while experiments with measured 
peak temperatures above 1770 K showed a positive thermal contribution from boron 
addition.  All tests using nanoaluminum displayed a noteable positive addition to the 
thermal output of the system.  These results suggest that even when employing nano-
sized boron, only a small envelope for complete energetic extraction exists in combustors 
with short residence times.  On the other hand, great hope exists for the use of 
nanoaluminum in high speed airbreathing combustors.    
 
 
 
METALLIC NANOPARTICLES AS FUEL ADDITIVES IN AIRBREATHING 
COMBUSTION 
 
 
 
 
 
 
 
By 
 
 Gregory Young 
 
 
Dissertation Submitted to the Faculty of the Graduate School of the  
University of Maryland, College Park, in partial fulfillment  
of the requirements for the degree of  
Doctor of Philosophy 
2007 
 
 
 
 
 
 
 
Advisory Committee: 
Dr. Kenneth Yu  
Dr. James Baeder 
Dr. Michael Zachariah 
Dr. Ashwani Gupta 
Dr. Bao Yang 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
? Copyright by 
Gregory Young 
2007 
 
 
 
ii 
DEDICATION 
 
 This work is dedicated to my family without whom I could not have hoped to 
accomplish anything.  By family I include more than just my mother and siblings.  I 
include everyone who has helped me along the way, no need to name names; I think you 
know who you are.  I don?t have the words to show my appreciation; all I can say is 
thanks.  I?d specifically like to dedicate this work to my wife Chrissy, who has supported 
me every step of the way.  Thanks. 
 
 
 
iii 
ACKNOWLEDGMENTS 
 
 I would like to express my sincere thanks to all of those who lent their support to 
me through the course of this research.  In particular, I would like to thank Dr. Kenneth 
Yu, my advisor, for providing guidance and encouragement through the entire effort.  I 
would also like to specifically thank Dr. Michael Zachariah for his input and support 
throughout.  Additionally I would like to express my gratitude towards my remaining 
committee members, Dr. James Baeder, Dr. Ashwani Gupta, and Dr. Bao Yang. 
 I am especially grateful to the Naval Surface Warfare Center-Indian Head 
Division for providing me with the opportunity to conduct this research.  Specifically I 
would like to thank Mr. Gary Prybyla, my supervisor, for his support and patience.  In 
addition I would like to express my thanks to Mr. Kevin Gessner, Mr. Bob Kaczmerak, 
and Mr. Chris Fawls. 
 Finally I would like to thank all of the members of the Advance Propulsion 
Research Lab, who have made the last few years extremely enjoyable.  I would 
specifically like to thank Mr. Rama Balar and Mr. Mike Krasel for all of their time and 
effort involved in this study.  Lastly, I would like to thank Mr. Kyle Sullivan of the Co-
Laboratory for Nanoparticle Based Manufacturing and Metrology for his time and effort 
in support of this research. 
 
 
 
iv 
TABLE OF CONTENTS 
 
ABSTRACT??????????????????????????????i 
DEDICATION???????????????????????????.?.ii 
ACKNOWLEDGEMENTS???????????????????????.iii 
TABLE OF CONTENTS????????????????????????.iv 
LIST OF FIGURES..??????????????????????????ix 
LIST OF TABLES??????????????????????????..xvi 
NOMENCLATURE?????????????????????????..xvii 
CHAPTER 1: INTRODUCTION?????????????????????.1 
1. Introduction?????????????????????????????..1 
1.1 Jet Propulsion Classification ?????????????..???????.2 
1.2 Air Breathing Propulsion Systems?????..?????????????.2 
1.2.1 Ramjets/Ducted Rockets???????????????????..4 
1.3 Brief History of Ramjets???..??????????????...???...10 
1.4 Motivation?????????..??????????????...???.11 
1.5 Objectives/Goals??????..???????????????...??...18 
1.6 Scope???..???????????????????????...??.18 
1.7 Benefits for Commercial Applications Based Upon the Results of this 
Research????????????????????????????.20 
 
CHAPTER 2: REVIEW OF RELEVANT PREVIOUS STUDIES???????21 
2. Review of Relevant Previous Studies???????????????????21 
 
 
 
v 
2.1 Brief Review of Droplet Combustion and the D2 Law??????????..21 
2.2 Metal Particle Combustion?????????????????????22 
2.2.1 Aluminum Particle Combustion Fundamentals????..??????..25 
2.2.1.1 Aluminum ? Single Particle Measurements???.??????28 
2.2.2 Boron Particle Combustion Fundamentals??..??????????.33 
2.2.2.1 Boron ? Single Particle Measurements??????????.?37 
2.3 Metal Combustion in Air Breathing Engines??????????????.40 
2.4 Nanoparticle Work????????????????????????.42 
2.4.1 Nanoparticles in Propulsion????????????????...?...44 
2.4.2 Advantages and Disadvantages of Nanoparticles?????????.?46 
2.5 Laminar Jets Exhausting into a Quiscent Atmosphere??????????..47 
 
CHAPTER 3: EXPERIMENTAL APPROACH??????..????...???.51 
3. Experimental Approach????????????????????????.51 
3.1 Flat Flame Burning Experiment....?..????????????????...51 
3.1.1 Flow Controls for Flat Flame Burning Experiment.?????????.56 
3.1.2 Particle Image Velocimetry (PIV)??..????????????..?56 
3.1.3 Temperature Measurements??????????????????..60 
3.1.4 Chemiluminescence for Flat Flame Burning Experiments??????...63 
3.2 Airbreathing Combustor Experiments?????..???????????.64 
3.2.1 Overall Design Strategy????????????????????65 
3.2.2. Flow Facilities???????????????????????..66 
3.2.3 Particle Delivery System Design???????????????......66 
 
 
 
vi 
3.2.4 Pilot Flame Design??????????????????????68 
3.2.5 Main Combustor Design????????????????...??....71 
3.3 Combustor Diagnostics??????..?????????????...??.75 
3.3.1 Flow Control for Airbreathing Combustor??????????...??75 
3.3.2 Mie Scattering??????????????????????...?76 
3.3.2 Chemiluminescence???????..???????????...??.77 
3.3.4 Temperature Measurements?...???????????????...?78 
 
CHAPTER 4: BASIC CHARACTERIZATION STUDIES???..??????..80 
4. Basic Characterization Studies?????????????????????..80 
4.1 Particle Characterization?..???????????????????.?.80 
4.1.1 Thermal Gravometric Analysis (TGA)?????????????.?81 
4.1.1.1 Determination of Active Content by TGA??...??????.?81 
4.1.2 Scanning Electron Microscope Imaging (SEM)???.???????..84 
4.2 Flat Flame Burner and Aerosol Injection Characterization ?...?..?????.87 
4.2.1 Velocity Measurements...???????????????????.87 
4.2.2 Temperature Measurements??????????????????..96 
4.2.3 Particle Size Measurements?????????????????..?97 
4.2.3.1 TEM??????...?????????????????...97 
4.2.3.2 DMA????????????????????????.99 
4.3 Airbreathing Combustor Characterization??????????????.103 
4.3.1 Gas Flow Calculation????????????????????..104 
4.3.2 Particle Seeder Characterization????????????????.104 
 
 
 
vii 
4.3.3 Pilot Flame Plume Measurements???????????????..107 
4.3.4 Blowoff Limits???????????????????????114 
 
CHAPTER 5: FLAT FLAME BURNER RESULTS AND ANALYSIS 
????????????????..????????..???..117 
5. Flat Flame Burner Results and Analysis??????.. ??????????...117 
5.1 Brief Discussion of Boron Combustion Physics????.....???????.117 
5.1.1. Ignition and Combustion Data for SB99 Nano-Boron??????.?.119 
5.1.2 Ignition Stage, t1, Results???????????????????131 
5.1.3 Second Stage/Full-Fledged Combustion, t2, Burning Time Results??...138 
5.2 Studies on Nanoaluminum Combustion????..???????????146 
5.3 Implications on High-Speed Airbreathing Combustors??...???????148 
 
CHAPTER 6: AIRBREATHING COMBUSTOR RESULTS????????..151 
6. Airbreathing Combustor Experiments??????????????????.151 
6.1 Test Conditions????????..????????????????..151 
6.1.1 Mie Scattering Results????????????????????154 
6.1.2 Chemiluminescence Results?????????????????...158 
6.1.2.1 AlO Emission??...?????????????????...158 
6.1.2.2 BO2 Emission???...????????????????...163 
6.1.2.3 Chemiluminescence Summary???...??????????.169 
6.1.3 Temperature Measurements??????????????????172 
6.2 Data Interpretation??????????????..?????????.184 
 
 
 
viii 
6.3 Additional Considerations????????????????????.189 
 
CHAPTER 7: CONCLUSIONS AND FUTURE WORK??????????..195 
7. Conclusions and Future Work?????????????????????.195 
7.1 Particle Characterization?..????????????????????196 
7.2 Flat Flame Burning Experiment?????..?????????????.197 
7.2.1 Ignition and Burning Time Conclusions?????????????.198  
7.2.1.1 Stage 1 Combustion of SB99?????...????????...199 
7.2.1.2 Stage 2 Combustion of SB99??????...???????...200 
7.3 Airbreathing Combustor Conclusions????..????????????201 
 7.3.1 Boron Combustor Conclusions????????????????...202 
 7.3.2 ALEX Combustor Conclusions????????????????..204 
7.4 Overall Study Conclusions??..??????????????????205 
7.5 Benefits of this Study to the Scientific Community?..?????????..207 
7.6 Recommendations for Future Work??..??????????????...207 
 
APPENDIX A ? PHOTOGRAPHS OF HARDWARE AND DESIGN DRAWINGS 
???????????.?????????????????.209 
APPENDIX B ? UNCERTAINTY ANALYSIS??????????????..229 
APPENDIX C ? ROTAMETER CALIBRATIONS???????????..?.231 
REFERENCES???????????????????????????..234 
 
 
 
ix 
LIST OF FIGURES 
 
FIGURE 1.1 Performance Envelope of Jet Propulsion Systems???????..??...3 
FIGURE 1.2 Brayton Cycle Temperature-Entropy Diagram for Ramjet??..???..?5 
FIGURE 1.3 Schematic Diagram of a Liquid Fueled Ramjet??????..????...7 
FIGURE 1.4 Schematic Diagram of a SFRJ???...??????????????.8 
FIGURE 1.5 Schematic Diagram of a Ducted Rocket??????????????9 
FIGURE 1.6 ISP of Aluminum Loaded RP-1 Fuels????????????.??..16 
FIGURE 1.7 ?ISP of Aluminum Loaded RP-1 Fuels???...????..????...?.16 
FIGURE 1.8 ISP of Boron Loaded RP-1 Fuels??...?...?.????..??????..17 
FIGURE 1.9 ?ISP of Boron Loaded RP-1 Fuels???...??.????..?????..17 
FIGURE 2.1 Comparison of Heats of Combustion of Selected Metals...??????.23 
FIGURE 2.2 Schematic of an Aluminum Droplet Burning in a Mildly Convective Flow 
?????????????????????????????26 
FIGURE 2.3 Physical Stages of Aluminum Particle Ignition and Combustion...???.27 
FIGURE 2.4 Schematic of Aluminum Particle Combustion by Melcher???..???27 
FIGURE 2.5 Burning Times of Single Particles Using Beckstead?s Correlation...??...33 
FIGURE 2.6 Physical Stages in Boron Ignition and Combustion???????...?..34 
FIGURE 2.7 Elementary Heterogeneous Reactions Associated with Boron Particle 
Oxidation?????????????.???????????...36 
FIGURE 3.1 Schematic Diagram of McKenna Flat Flame Burner???????...?52 
FIGURE 3.2 Flat Flame Burning Test Configuration?????????.????..54 
FIGURE 3.3 Schematic Diagram of Sonicating Atomizer??..?????????..55 
 
 
 
x 
FIGURE 3.4 Schematic Diagram of Flat Flame Burning Experiment..???????55 
FIGURE 3.5 Example of Rotameter Calibration?????...??????????56 
FIGURE 3.6 Schematic Diagram of PIV Experiment?????????????...60 
FIGURE 3.7 Flat Flame Burner Thermocouple Configurations???????..??.61 
FIGURE 3.8 Curve Fit for Air Thermal Conductivity??????.???????.63 
FIGURE 3.9 Schematic Diagram of Reverse Cyclone Seeder???????.???.67 
FIGURE 3.10 Pilot Flame Components???...???????????????.69 
FIGURE 3.11 Schematic Diagram of Pilot Flame Nozzle...???????????.70 
FIGURE 3.12 Schematic Diagram of Pilot Flame Combustor??????????..70 
FIGURE 3.13 Schematic Diagram of Combustor Set Up?????...??????..74 
FIGURE 3.14 Assembled Combustor on Test Stand?????..????????...74 
FIGURE 3.15 Schematic Diagram of Combustor Mie Scattering Experiment...???..77 
FIGURE 4.1 TGA of 50 nm ALEX???????????????????...?83 
FIGURE 4.2 TGA of SB99???????????????????????....84 
FIGURE 4.3 SEM of 50 nm ALEX???????????????????...?86 
FIGURE 4.4 SEM of SB99 Boron Particles???..????????..?????..86 
FIGURE 4.5 Example of Image Pair???.?????????????????89 
FIGURE 4.6 Velocity Vectors Obtained from PIV Experiments???????..??89 
FIGURE 4.7 Axial-Component of Velocity Contour for Highest Injection Flow Rate and 
Crossflow???????????????..?????.????90 
FIGURE 4.8 Axial-Component of Velocity Contour for Intermediate Injection Flow Rate 
and Highest Crossflow ???????..????????????90 
 
 
 
xi 
FIGURE 4.9 Axial-Component of Velocity for Lowest Injection Flow Rate and Highest 
Crossflow????????????????????.????..91 
FIGURE 4.10 Axial-Component of Centerline Velocity for Intermediate Injection Flow 
Rate and Several Crossflows?????????????????.92 
FIGURE 4.11 Centerline Axial-Component of Velocity of Lowest Jet Flow Rate (0.61 
L/m)???????????????????????????94 
FIGURE 4.12 Centerline Axial-Component of Velocity of Intermediate Jet Flow Rate 
(0.84 L/m)?.??????????????????????.....94 
FIGURE 4.13 Centerline Axial-Component of Velocity of Highest Jet Flow Rate (1.15 
L/m)??????????.???.??.???????????.95 
FIGURE 4.14 TEM Image of Aerosolized SB99???.???????????.?99  
FIGURE 4.15 Close-up TEM Image of Aerosolized SB99???????????...99 
FIGURE 4.16 Schematic Diagram of a Typical DMA Set Up????..?????..102 
FIGURE 4.17 Agglomerate Size Distribution by DMA?????..???????103 
FIGURE 4.18 Summary of Seeder Calibration Testing???????????.?.106 
FIGURE 4.19 Seeder Carrier Gas Filling Time????...???????????107 
FIGURE 4.20 Schematic Diagram of Jet in Crossflow Problem???????..?...110 
FIGURE 4.21 Intensity Contour of Plume Jet for Crossflow of 40 m/s?...????...112 
FIGURE 4.22 Intensity Contour of Plume Jet for Crossflow of 55 m/s...?????...112 
FIGURE 4.23 Intensity Contours of Plume Jet for Crossflow of 70 m/s.?????...113 
FIGURE 4.24 Outer Jet Boundary Profiles for Selected Crossflows???????..113 
FIGURE 4.25 Determination of Jet Penetration Correlation????????..??.114 
FIGURE 4.26 Normal Combustor Operation with Pilot Flame On???????.....115 
 
 
 
xii 
FIGURE 4.27 Example of Flame Anchoring at Rearward Facing Step..??????116 
FIGURE 4.28 Blowoff Limits??????????????????????.116 
FIGURE 5.1 Image of Two-Stage Phenomenon in Nano-Boron Combustion (test 
condition 7)???????????????????????..121 
FIGURE 5.2 Filtered Image in Nano-Boron Combustion (TC7).????????...122 
FIGURE 5.3 Typical Background Image (no particles TC7)??????..????122 
FIGURE 5.4 Intensity Contours for XO2 = 0.3???????????????....124 
FIGURE 5.5 Intensity Contours for XO2 = 0.2?????.??????????...125 
FIGURE 5.6 Intensity Contours for XO2 = 0.15?????.??????????.126 
FIGURE 5.7 Intensity Contour for XO2 = 0.1????????????????..127 
FIGURE 5.8 Intensity Contours Using Different Filters (TC 6)???????.??128 
FIGURE 5.9 Normalized Intensity Profile Using Different Filters?...??????..129 
FIGURE 5.10 Example of t2 Burn Time Determination (TC2)?????.????..131 
FIGURE 5.11 t1 Burning Times for SB99?????????..????????.132 
FIGURE 5.12 Comparison of Ignition Time,  t1, with Other Data??..??????134 
FIGURE 5.13 Schematic Diagram of Energy Balance?????????????136 
FIGURE 5.14 Ignition Time Correlation??????????????????.138 
FIGURE 5.15 t2 Burning Times for SB99???..??????????????.140 
FIGURE 5.16 Total Burning Times (t1+t2) for SB99????.??????.???141 
FIGURE 5.17 t2 Burning Time Comparison to Larger Particles??????..??...143 
FIGURE 5.18 Arrhenius Burning Rate Law for t2??????????????..144 
FIGURE 5.19 t2 Burning Time Correlation???.??????????????146 
 
 
 
xiii 
FIGURE 6.1 Combustor Operating in Traditional Ramjet Mode with a) Gaseous Ethylene 
as Fuel b) Gaseous Ethylene with 50 nm Aluminum Particles and c) 
Gaseous Ethylene with Boron Particles.?.??????????.?153 
FIGURE 6.2 Typical Instantaneous Mie Scattering Images (from Test Condition 6a) ..155 
FIGURE 6.3 Average Intensity Contour of Mie Scattering Images (Test Condition 
6a)???????????????????????????157 
FIGURE 6.4 Average Intensity at Axial Locations (Test Condition 6a)?..????...157 
FIGURE 6.5 Average Intensity at Axial Locations (Test Condition 7a)..?????...158 
FIGURE 6.6 Instantaneous AlO Chemiluminescence Images??????????160 
FIGURE 6.7 AlO Chemiluminescence Contours (C2H4 only, 6a, 7a, 8a 9a, 10a) ?.....161 
FIGURE 6.8 AlO Intensity Profiles Versus Combustor Length???????.....?.162 
FIGURE 6.9 AlO Intensity Profiles Versus Time?????????..?????.162 
FIGURE 6.10 Instantaneous Images of BO2 Emission????????.????..164 
FIGURE 6.11 BO2 Chemiluminescence Contours (C2H4 only, 1b, 2b, 3b, 4b, 5b)...?.167 
FIGURE 6.12 BO2 Intensity Profiles Versus Combustor Length???????....?168 
FIGURE 6.13 BO2 Intensity Profiles Versus Time????..?????????...168 
FIGURE 6.14 Average Intensity as a Function of Adiabatic Flame Temperature?..?171 
FIGURE 6.15 Intensity-Velocity Coupling as a Function of Adiabatic Flame 
Temperature???????????????????????.171 
FIGURE 6.16 Typical Thermocouple Trace (Uncorrected For Radiation) From Test 
Condition 1b??????????????..????????..173 
FIGURE 6.17 Combustor Temperature Profiles for Selected Ethylene/ALEX Tests?.173 
FIGURE 6.18 Temperature Profiles for Test Conditions 1, 1b, 2, 2b?????..?...175 
 
 
 
xiv 
FIGURE 6.19 Temperature Profiles for Test Conditions 3, 3b, 4, 4b??..????...176 
FIGURE 6.20 Temperature Profiles for Test Conditions 5, 5B???????.??.176 
FIGURE 6.21 Temperature Profiles for Test Conditions 6, 6A??..??????....178 
FIGURE 6.22 Temperature Profiles for Test Conditions 7, 7a, 8, 8a, 9, 9a.????...178 
FIGURE 6.23 Temperature Profiles for Test Conditions 10, 10a??????.?......179 
FIGURE 6.24 Comparison of Ethylene and Boron Loaded Tests???????..?.181 
FIGURE 6.25 Comparison of Ethylene and Aluminized Tests?????????...182 
FIGURE 6.26 Thermal Ratios??..???????????????????...184 
FIGURE 6.27 Boron Particle Quenching????.?????????????...187 
FIGURE 6.28 Specific Impulse of Boron Loaded Fuels Considering Active Content...193 
FIGURE 6.29 Density Impulse of Boron Loaded Fuels Considering Active Content...193 
FIGURE 6.30 Specific Impulse of Aluminized Fuels Considering Active Content?...194 
FIGURE 6.31 Density Impulse of Aluminized Fuels Considering Active Content?...194 
FIGURE A.1 Photograph of Flat Flame Burner Experiment????..??????209 
FIGURE A.2 Components of Sonicating Atomizer??????????????209 
FIGURE A.3 Photograph of Thermocouple Configuration???????????210 
FIGURE A.4 Photograph of Seeder Components??????????????...210   
FIGURE A.5 Modified Orifice Used for Particle Injection???????????211 
FIGURE A.6 Specially Designed Adapter and Flow Straightener?.???????211 
FIGURE A.7 Combustor Components???????????????????212 
FIGURE A.8 Exploded View of Combustor Components???????????..212 
FIGURE A.9 Pilot Flame Combustor Aft End Cap?????.?????????213 
FIGURE A.10 Pilot Combustor Flame Main Body??????????????.214 
 
 
 
xv 
FIGURE A.11 Seeder Body?.??????????????????????215 
FIGURE A.12 Seeder Plug??.?????????????????????.216 
FIGURE A.13 Seeder Cap???????????????????????...217 
FIGURE A.14 Combustor Bottom Plate???..??????????????...218 
FIGURE A.15 Combustor Injector Plate (Mates With Pilot Flame Combustor).??...219 
FIGURE A.16 Combustor Top Plate???.????????????????..220 
FIGURE A.17 Alternate Top Plate?????????..???????????221 
FIGURE A.18 Left Window Holder Side View???????????.????.222 
FIGURE A.19 Left Window Holder End View???????????????...223 
FIGURE A.20 Right Window Holder Side View???????????????224 
FIGURE A.21 Right Window Holder End View???.???????????...225 
FIGURE A.22 Quartz Window????????????????????..?..226 
FIGURE A.23 Front Transition?????????????????????...227 
FIGURE A.24 Rear Transition?????..????????????????..228 
FIGURE C.1 Calibration for Oxygen???????????????????..231 
FIGURE C.2 Calibration for Air???.??????????????????231 
FIGURE C.3 Calibration for Methane????..??????????????...232 
 
 
 
 
 
 
 
 
xvi 
LIST OF TABLES 
 
TABLE 1.1 Lower Heating Values of Various Fuels??????????????15 
TABLE 4.1 Cold Flow PIV Test Matrix???????????????..???.88 
TABLE 4.2 Flat Flame Burner Test Matrix?????????????????...97 
TABLE 4.3 Pilot Flame Operating Conditions (Calculated) ??????????..109 
TABLE 5.1 Summary of SB99 Ignition and Combustion Results????????..120 
TABLE 5.2 Summary of Camera Settings?????????????????..123 
TABLE 5.3 Properties Used in Analytical Analysis of Ignition Time???????135 
TABLE 6.1 Summary of Combustor Test Conditions?????????????.153 
TABLE 6.2 Flow Conditions??..????????????????????.154 
TABLE 6.3 Mole Fractions of Oxidizing Product Species??????..????...154 
TABLE 6.4 Flight Conditions??...???????????????????..188 
 
 
 
xvii 
NOMENCLATURE 
 
A  Area 
A*  Critical Area for Choking Flow 
B  Spalding Transfer Number 
Cc  Cunningham Correction Factor 
dq  Quenching Diameter  
dt  Incremental Time 
dx  Incremental Axial Distance 
Da  Damkohler Number 
D  Particle Diameter 
Do  Initial Particle Diameter 
E  Charge of an Electron 
E  Electrical Field Strength 
Ea  Activation Energy 
f  Fuel to Air Ratio 
F  Thrust 
FD  Drag Force 
g  Acceleration due to Gravity 
h  Convective Heat Transfer Coefficient 
hi   Enthalpy of Formation of Species i 
H  Combustor Inlet Channel Height 
Isp  Specific Impulse 
 
 
 
xviii 
J  Jet Momentum Flux Ratio 
Je  Jet Initial Momentum Flow 
L  Lift Force 
k  Thermal Conductivity 
m  Mass 
mAl  Mass of Aluminum 
mf  Final Mass 
mi  Initial Mass 
mo2  Mass of Oxygen 
?m   Mass Flow Rate 
M  Mach Number 
MW  Molecular Weight 
n  Number of Charges on a Particle 
Ni  Number of Moles of Species i 
Nu  Nusselt Number 
pO2  Partial Pressure of Oxygen 
P  Pressure 
Po  Stagnation Pressure 
Q5  Heat Released by Surface Reaction 
r  Jet Incremental Radius 
R  Jet Radius 
Ru  Gas Constant 
Red  Reynolds Number Based on Sphere Diameter 
 
 
 
xix 
Sc  Schmidt Number 
Sparticle  Characteristic Particle Stopping Time 
t  Time 
tb  Burning Time 
tb,diff  Time Scale for Diffusion Limited Burning 
tb,kin  Time Scale for Kinetically Limited Burning 
t1  Stage 1 or Ignition Stage of Boron Combustion 
t2  Stage 2 or Full-Fledged Combustion Stage of Boron Combustion 
T  Temperature 
Tave_metal Average Measured Temperature for a Given Metallized Test 
Tave_ethylene Average Measured Temperature for a Given Hydrocarbon Only Test 
Tg  Gas Temperature 
To  Stagnation Temperature 
Tq  Quenching Temperature 
Ttc  Thermocouple Temperature Reading 
T?  Ambient Temperature 
To,?  Ambient Stagnation Temperature 
U  Velocity 
Uave  Average Velocity 
Ue  Jet Exit Velocity 
UJ  Jet Velocity 
Ux  Axial Component of Velocity 
U?  Freestream Velocity 
 
 
 
xx 
V  Particle Velocity 
VTE  Terminal Electrostatic Velocity 
w  Width of Pilot Jet Slot 
x  Axial Distance 
Xeff  Effective Oxidizer Concentration 
Xo,?  Oxygen Mole Fraction in the Far Field 
XO2  Oxygen Mole Fraction 
y  Vertical Distance 
YCG  Carrier Gas Mass Fraction 
Yo,?  Oxygen Mass Fraction in the Far Field 
Z  Electrical Mobility 
?s  Thermal Diffusivity 
?v  Evaporation Coefficient 
?tc  Thermocouple Emissivity 
?hr  Lower Heating Value on a Mass Basis 
?m  Change in Mass 
?  Thermal Ratio 
?  Ratio of Specific Heats 
?  Mean Free Path 
?  Kinematic Viscosity 
?? ? Density 
?l  Liquid Phase Density 
?J  Jet Gas Density 
 
 
 
xxi 
?p  Particle Density  
?s  Solid Phase Density 
???hr  Lower Heating Value on a Volumetric Basis 
??  Freestream Gas Density 
??Isp  Density Impulse 
?  Fluid Viscosity 
?s  Fluid Viscosity at Thermocouple Surface 
??  Fluid Viscosity in the Far Field 
?  Similarity Variable 
?  Boltzman Constant 
?  Equivelance Ratio 
 
 1
CHAPTER 1 
 
INTRODUCTION 
 
1. Introduction 
Airbreathing propulsion systems offer the potential of increasing vehicle range 
due to the fact that they are considerably more efficient than rockets.  For volume limited 
systems, the use of metals as fuel additives offers to increase performance even further by 
increasing the volumetric energy stored in the propulsion system.  However, traditional 
metals such as aluminum and boron require longer combustor residence times than 
hydrocarbons or hydrogen, which are typically used in airbreathing systems.  The 
development of nano-scale metals has opened up the possibility of utilizing metals in 
high-speed propulsion systems. 
The focus of this study was to determine the fundamental ignition and combustion 
behavior of both boron and aluminum nanoparticles for which very little data exists.  Due 
to the lack of available data, the combustion behavior of boron nanoparticles was first 
studied in a controlled flame environment to extract basic combustion parameters, such 
as ignition criterion and burning time. 
With an understanding of the fundamental behavior of both boron and aluminum 
nanoparticles, the next portion of this study was to determine whether or not either 
material could be considered a viable fuel/fuel supplement in a high speed airbreathing 
propulsion system.  The fundamental combustion behavior of both boron and aluminum 
 2
nanoparticles were studied extensively in an airbreathing combustor simulating ramjet 
conditions.   
Chapter 1 provides some basic background and motivation for the study, while 
Chapter 2 discusses relevant research available in the literature.  Chapter 3 describes in 
detail the experimental approach used in this study.  Chapter 4 outlines the basic 
characterization studies required to meet the goals of the current research.  Finally, 
Chapters 5 and 6 discuss in detail the results of this research.  
 
1.1 Jet Propulsion Classification 
Propulsion mechanisms provide a force that moves bodies that are initially at rest, 
changes a velocity, or overcomes retarding forces when a body is propelled through a 
medium1.  Jet propulsion is a means of locomotion whereby a reaction is imparted to a 
device by the momentum of ejected matter.  Two common forms of jet propulsion are 
rocket propulsion and duct propulsion.  Duct propulsion is a class of jet propulsion and 
includes turbojets and ramjets; these engines are commonly referred to as air breathing 
engines.  
  
1.2 Airbreathing Propulsion Systems 
Airbreathing propulsion systems differ from rockets in that as their name implies 
they use atmospheric oxygen together with a fuel carried onboard as the propellant 
system.  Rockets carry their oxidizing agent onboard in a variety of ways depending on 
the type of rocket.  By obtaining oxygen from the atmosphere, airbreathing systems are 
considerably more efficient than rockets.  Furthermore, most airbreathing engines are 
 3
reusable, not just refurbishable like rockets.  An additional feature for space access is 
short turn around time, with potential cost reduction of 10-100 times per pound payload2.  
However, airbreathing devices are limited to that portion of the atmosphere, where air is 
dense enough to provide sufficient thrust, typically around 100 km3.  Figure 1.1 provides 
a summary of the performance envelope of the different jet propulsion systems4.  As the 
figure shows, airbreathing systems have much higher specific impulses (Isp) than rockets.  
The range that a vehicle can attain is directly proportional to the vehicles Isp, this implies 
that airbreathing propulsion systems should be the vehicle of choice for long range 
missions.   
 
Fig. 1.1 Performance Envelope of Jet Propulsion systems4 
 
Airbreathing propulsion systems can be classified into two categories: 1) Engines 
with rotating machinery, and 2) engines without rotating machinery.  Engines with 
rotating machinery include turboprops, turbojets, and turbofans.  Engines without rotating 
 4
machinery include ramjets and scramjets.  Ramjet performance is the main focus of this 
thesis.  Typically turboprops are limited to speeds of about Mach 0.7, turbofans and 
turbojets are limited to speeds up to about Mach 3.0, ramjets operate between Mach 2 and 
5, while for higher speeds scramjets are employed3.    
 
1.2.1 Ramjets/Ducted Rockets 
Simple in concept, the ramjet uses fixed components to compress and accelerate 
intake air by ram effect.  Ramjets cannot operate from a standing start; they must be 
accelerated to high speeds by some other form of propulsion due to the fact that they rely 
on the ram effect for air compression, thus the name ramjet.  They have also been called a 
flying stovepipe due to the absence of rotating parts that characterize turbine engines2. 
The basic ideas behind the operation of a ramjet are fairly straightforward.  There 
are typically four components to the engine, an inlet, diffuser, combustor and an exit 
nozzle.  In most cases, supersonic air enters the inlet and is decelerated through the 
diffuser.  The diffuser is analogous to the compressor of a turbine engine.  Compression 
itself is dependent on velocity and it increases dramatically with flight speed.  After the 
diffuser, the air enters the combustor where it is mixed with injected fuel.  The mixture 
burns in the combustor, quite often with the aid of a flame holding device of some sort.  
The burning mixture imparts thermal energy to the gas, which expands to high velocity 
through an exit nozzle at speeds greater than the entering air.  This process produces the 
thrust necessary for ramjet flight.  
Although, ramjets are capable of operating at subsonic speeds, they are most 
suited to supersonic flight due to the higher pressure rise accompanying higher flight 
 5
speeds3.  For speeds above Mach 3, turbine based propulsion systems become 
impractical.  A ramjet engine is one alternative that is especially attractive due to its 
simplicity in the fact that it has no rotating parts.  Generally, ramjets can be broken into 
three different types: 1) Liquid Fuel, 2) Solid Fuel, 3) Ducted Rockets.  The first 
approach to estimating the performance of a ramjet engine is that of a heat engine or 
thermal engine closed cycle analysis, often referred to as a Brayton Cycle Analysis5.  A 
temperature-entropy diagram of the ideal ramjet is shown in Figure 1.2 to illustrate the 
cycle.  The stations correspond to those given in Figure 1.3.  Briefly, the cycle goes as 
follows.  Adiabatic compression from station 0 to station 2, which is accompanied by an 
increase in static temperature.  From station 2 to station 4, there is a constant static 
pressure heat addition due to the combustion of fuel and air.  Finally, from station 4 to 
station 6 there is an adiabatic expansion through the nozzle to the atmospheric static 
pressure.  This analysis is an idealized case and does not consider any losses or 
irreversibilities.      
 
 
 
 
 
 
Fig. 1.2 Brayton Cycle Temperature-Entropy Diagram for Ramjet 
 
T 
s 
0 
02 
04 
06 
 6
Figure 1.35 shows a schematic diagram of a typical liquid fuel ramjet engine.  
Supersonic air is decelerated through a diffuser after which it enters the combustion 
chambers at subsonic speeds.  Liquid fuel, usually a hydrocarbon, is injected and 
combusted with the atmospheric air.  Often flame holding devices are employed to 
encourage flame stabilization.  After the combustion chamber the combustion products 
are accelerated through an exhaust nozzle into the atmosphere to provide thrust.  Liquid 
fuels offer a number of advantages over solid fuels6. 
1) Better specific thrust 
2) Higher modulation ratios for liquids because of their incompressibility 
3) More compactness and optimization of available space compared to ducted 
rockets (storage tanks can be made in any shape) 
For these reasons, the use liquid fuels proves necessary when the specified missions 
involve long ranges or when the flight envelope entails considerable altitude and or Mach 
number variations6.   
 
 7
Exhaust Jet
Fuel Inlet
NozzleCombustionChamberDiffuser
Combustion Expansion
0 1 2 3 4 5 6
Supersonic
Compression
Subsonic
Compression  
Fig. 1.3 Schematic Diagram of a Liquid Fueled Ramjet5 
 
One variant of the LFRJ is the use of slurry, slush, or gelled fuels.   Typically 
solid particles such as aluminum or boron can be added to a liquid hydrocarbon fuel to 
produce a gel or slurry.  These additives are used to increase the energetic output of the 
fuel.  Although the theoretical output of the fuel increases with the addition of these solid 
particles, a new set of challenges arises.  Perhaps the largest challenge comes from the 
fact that the combustion efficiency in the chamber may significantly decrease because the 
solid particles require a minimum amount of energy to initiate their combustion and the 
residence time in the chamber may not permit complete combustion6. 
Solid Fuel Ramjets (SFRJ) have been considered for propulsion devices in 
missiles since the 1930?s7.  Figure 1.43 shows a schematic diagram of a typical solid 
fueled ramjet.  In a SFRJ, the fuel is stored in a solid matrix such as a polybutadiene, 
polyether, polyester, polyeurethane, polymethilmetacrylate, or plexiglass. There have 
 8
been numerous studies incorporating metals into the solid fuel matrix as a means of 
increasing the energetic content of the fuels8-11. 
Atmospheric air is decelerated and brought into the combustor section upon 
which it reacts with the solid fuel.  The solid fuel pyrolyses and creates a fuel rich 
diffusion flame near its surface.  Fuel regression rates depend primarily on the inlet air 
temperature and the air mass flux7.  Often a secondary combustion chamber is employed 
to achieve complete fuel/air mixing and combustion.  After the combustion chamber, just 
as in the liquid fuel ramjet, the combustion products are accelerated through an exhaust 
nozzle to the atmosphere to provide thrust.  SFRJ?s have several major advantaged over 
LFRJ?s. 
1. Storage as a solid fuel is more efficient. 
2. Storage as a solid fuel is typically safer since there is no possibility of fuel 
leakage 
3. Solid fuel provides insulation to the combustor walls allowing higher 
temperature operation.  
 
Fig. 1.4 Schematic Diagram of a SFRJ3 
 
Ducted rockets or air augmented rockets, differ form SFRJ?s in that they burn a 
fuel rich propellant/gas generator whose products are then combusted with the 
 9
atmospheric air.  The gas generator provides hot combustible products to the main 
combustion chamber.  This promotes faster and generally more complete combustion, 
however, since some oxidizer is carried on board in the fuel rich propellant some of the 
initial benefits of the air breathing system are lost.  The principal attribute of a ducted 
rocket is the simplicity, ruggedness, and handling ease it enjoys by virtue of its solid fuel.  
In tactical missile applications this factor can weigh heavily and can favor the selection of 
a ducted rocket even though its performance might be only one-half to two-thirds that of 
a LFRJ11.  Figure 1.53 shows a schematic of a ducted rocket. 
 
 
Fig. 1.5 Schematic Diagram of a Ducted Rocket3 
 
Ramjets have a number of advantages over other propulsion systems. 
1. Simple design (e.g. no moving parts). 
2. Higher speed operation than turbojets. 
3. Greater efficiency in terms of specific impulse than rockets ? generally 
translates into increased range. 
However there are a number of disadvantages with ramjet compared to the other jet 
propulsion systems. 
1. Can not generate static thrust ? need a booster system to operate 
 10
2. Generally lower thrust to weight than rockets 
3. Can only operate in the earth?s atmosphere 
4. Operation is limited to approximately 2 < Mach < 5  
 
1.3 Brief History of Ramjets 
Ramjet concepts have been around since the early 1900?s14.  Numerous countries 
have been involved with their development including the United States, Great Britain, 
France, Germany, Russia (or USSR for part of ramjet history), China, Israel, and South 
Africa.  Applications range from boost to main propulsion for aircrafts, gun projectiles, 
missiles, and space launch vehicles.  Rene Lorin of France is generally credited with first 
recognizing the possibility of using ram pressure in a propulsion device, however, Lake 
of the United States was issued the first patent for a subsonic ramjet cycle device in 
19092, and Albert Fono of Hungary holds the first patent for the use of a ramjet engine as 
a means of propulsion for supersonic flight in 19283,5.  Although the concept of a ramjet 
was around much earlier, the onset of World War II brought about the true beginnings of 
the ramjet, with a peak coming in the 1950?s.  
In Germany, Trommsdorff led a successful effort that began in 1935 to develop 
artillery shells powered by multiple-shock, conical-inlet, liquid-fueled ramjets.  These 
shells actually accelerated from Mach 2.9 to 4.2 in trials in the early 1940?s14.  The 
Germans fielded the first operational ramjet powered missile in the form of the V1 
?buzzbomb? powered by a subsonic flight speed pulsejet engine.  In 1939 under the 
direction of Merkulov, Russia successfully flight-tested a tandem boosted ducted rocket 
using a magnesium/aluminum solid fuel2.  The first ramjet powered airplane was 
 11
designed and tested by Leduc in France, with the first powered flight taking place in 
April 1949.  
Reid (United States) and Marquardt (England) joined the ramjet development 
effort in the early 1940?s in the form of aerial guided projectiles and aircraft performance 
augmenters, respectively.  These efforts continued after World War II and resulted in 
weapon systems such as the BOMARC (U.S. Air Force), Talos (U.S. Navy), and 
Bloodhound (United Kingdom).  In France a number of operational ramjet missiles 
(VEGA, CT-41, and SE 4400) were deployed in the last 1950?s and 1960?s2.  In the 
1990?s China and South Africa began ramjet development programs.   
 
1.4 Motivation 
While hydrogen is a popular choice for airbreathing propulsion systems 
particularly scramjets, due to its high gravimetric heating values, there are some major 
disadvantages with the use of hydrogen. 
1. Low density requires large storage tanks.  Large tanks suffer large drag 
penalties. 
2. Liquefaction requires the use of cryogenics, which complicates the system 
and restricts storage capabilities. 
3. More hazardous than hydrocarbons. 
Naturally, one alternative is the use of hydrocarbons.  Although the gravimetric heating 
value of typical hydrocarbons is less than half that of hydrogen, the volumetric heating 
value is very comparable.  The result is a more compact system capable of greater range.  
Furthermore, liquefaction and storage of hydrocarbons is quite simple.  Hydrocarbons are 
 12
typically stored as a liquid, but can also be stored easily as a gel, slurry, or as part of a 
solid matrix in a fuel rich solid propellant or solid fuel.  To further increase the 
performance of hydrocarbon fuels, it is possible to add metal additives which can 
increase overall performance and in particular the energy density of a system.  Aluminum 
has been used as an additive in solid propellant rocket motors for many years.  In 
addition, aluminum has been studied as a fuel additive in Solid Fueled Ramjets (SFRJ).  
Boron has been studied for many years as a fuel and or a fuel supplement because of its 
high gravimetric and volumetric heating values8.  Boron has been investigated as a fuel 
supplement in solid rocket motors, hybrid rocket engines, and SFRJ?s.  One major issue 
with previous studies in a traditional ramjet setting, is that all of the studies have been 
conducted using micron size particles or larger.  
Metal additives such as aluminum have been used in solid rocket motors for many 
years as a means of increasing performance since they have the ability to increase the 
volumetric heat release of the propellants.  However, one downfall in using these 
additives is that it in many cases incomplete combustion of the additives results in lower 
than expected performance.  Typically this is a result of large agglomerates forming and 
not fully reacting in the combustor.  With these problems evident in rocket motors, the 
use of metal additives has not been fully realized in air breathing combustors since 
residence times are typically much smaller.  Now with the creation of nanoscale particles, 
it is expected that combustion times for individual particles will decrease significantly 
and the potential benefits for air breathing systems may be realized.  Olsen15 provided a 
summary of micron-sized aluminum particle burn times in various environments.  
Typical burn times for particles from about 20 ?m to 100 ?m vary from about three 
 13
milliseconds to about thirty milliseconds with a fair amount of scatter attributed to 
variations in methods, test conditions, data reduction methods, and the use of relatively 
wide particle size distributions.  Parr16 reported burn times of agglomerated nano-sized 
aluminum particles of less than 1 millisecond, burning in a steam environment.  In 
addition, Parr found that the ignition temperatures of some types nanoaluminum are 
significantly decreased compared to micron-sized aluminum. In a series of shock tube 
measurements, Bazyn17 also found burn times of nanoscale aluminum to be less than one 
millisecond in environments of primarily oxygen or carbon dioxide in temperatures 
ranging from 1200-2200 K and pressures ranging from 4-32 atmospheres.  If we consider 
that typical residence times in a ramjet engine are on the order of milliseconds, it would 
seem that it would be possible to extract all or most of the energy from these aluminum 
particles within the engine and maximizing the performance provided that the particles 
are close to their original size within the combustor.  Furthermore, Tepper and Kaledin18 
found that the addition of ALEX particles can act as a combustion accelerant for RP-1 in 
oxygen when compared to gelled and neat RP-1 in terms of ignition delay.   
Traditionally, metals have been considered as additives to solid or gelled fuels, 
however, Goroshin et. al.19 point out through a series of thermochemical calculations that 
powdered metals alone, in particular boron and aluminum, can provide substantial 
performance increases over hydrocarbon fuels.  Palaszewski20 and Mordosky21 have 
demonstrated the ability to gel hydrocarbons with aluminum particles with up to 55% by 
weight aluminum.  Mordosky21 gelled nanoaluminum with RP-1, and tested it with 
gaseous oxygen in a rocket engine environment, with some c* efficiencies reported 
greater than 90%.  Furthermore, nano-sized aluminum and boron were investigated as a 
 14
fuel supplement in hybrid rocket engines22 and were proven to significantly increase 
regression rates of solid fuels.  More recently Palaszewski23 utilized nano-sized 
aluminum as an additive to gelled hydrocarbons in pulsed detonation engines.      
In airbreathing propulsion one of the most important parameters in determining a 
fuel?s potential is the heating value.  However for volume limited systems, the heating 
value per unit volume can become the driving parameter for performance.  Table 1.1 
shows the lower heating values of RP-1 fuels with and without the addition of aluminum 
or boron determined from a series of calculations.  The lower heating value corresponds 
to the case where none of the water in the products is assumed to condense.  The density 
of the fuels is also determined from calculations.  As Table 1.1 shows, the heating value 
per unit mass decreases with the addition of aluminum, however, the volumetric heating 
value increases significantly due to the high density of aluminum compared with that of 
RP-1.   
To illustrate further the benefits of these additives, figures 1.6-1.9 show calculated 
ideal ramjet specific impulses as a function of Mach number for a variety of additive 
loadings in RP-1.  Specific impulse is an important parameter for all forms of propulsion 
as it is directly related to the range that a vehicle can attain.  The specific impulse of a 
ramjet is dependent on many parameters such as fuel to air ratio and altitude.  In order to 
simplify the analysis the calculations were based upon ideal ramjet assumptions, and are 
made considering each air/fuel mixture is in stoichiometric proportions, which may not 
necessarily be ideal for the given fuel.  In addition, the calculations were performed with 
air properties at an altitude of 10 km (~60,000 ft).  While the standard specific impulse 
(Isp) actually decreases with increasing metal content, the density impulse (?Isp) increases 
 15
Fuel 
 
?hr (kJ/kg) 
 
? (g/cc) 
 
??hr (MJ/m3) 
 Hydrogen (H2) 
 
119,971 
 
0.07 
 
 8,398 
 RP-1 (CH2) 
 
39,688 
 
0.80 
 
 31,750 
 Aluminum 
 
31,643 
 
2.7 
 
 85,436 
 Boron 
 
58,497 
 
2.34 
 
 138,762 
RP-1/5%Al 
 
39,286 
 
0.829 
 
 32,568 
 RP-1/10%Al 
 
38,883 
 
0.861 
 
33,478 
 
RP-1/20%Al 
 
38,079 
 
0.931 
 
 35,452 
 RP-1/50%Al 
 
35,665 
 
1.234 
 
 44,011 
 RP-1/5%B 
 
40,628 
 
0.827 
 
 33,609 
 RP-1/10%B 
 
41,569 
 
0.856 
 
 35,598 
 RP-1/20%B 
 
43,450 
 
0.921 
 
 40,029 
RP-1/50%B 
 
49,092 
 
1.192 
 
 58,536 
 
significantly with increasing metal content.  In the case of boron, the addition of 50% by 
weight boron to RP-1 yields an improvement of approximately 50% in density impulse.  
For volume limited systems, such as missiles, this improvement in potential performance 
is undeniable.        
Table 1.1 Lower Heating Values of Various Fuels 
 
 
 
 
 
 
 
 16
 
 
 
 
 
 
 
 
 
Figure 1.6 Isp of Aluminum Loaded RP-1 Fuels 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 1.7 ?Isp of Aluminum Loaded RP-1 Fuels 
 
 
 
 
200
400
600
800
1000
1200
1400
1 2 3 4 5 6 7 8 9
RP-1
RP-1/5%Al
RP-1/10%Al
RP-1/20%Al
RP-1/50%Al
?I s
p 
(s*
g/c
m3
)
Mach Number
200
400
600
800
1000
1200
1400
1600
1 2 3 4 5 6 7 8 9
RP-1
RP-1/5%Al
RP-1/10%Al
RP-1/20%Al
RP-1/50%Al
I sp
(s)
Mach Number
 17
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 1.8 Isp of Boron Loaded RP-1 Fuels 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 1.9 ?Isp of Boron Loaded RP-1 Fuels 
 
 
200
400
600
800
1000
1200
1400
1600
1 2 3 4 5 6 7 8 9
rho RP-1
rho RP-1/5%B
rho RP-1/20%B
rho RP-1/10%B
rho RP-1/50%B
?I s
p 
(s*
g/c
m3
)
Mach Number
200
400
600
800
1000
1200
1400
1600
1 2 3 4 5 6 7 8 9
RP-1
RP-1/5%B
RP-1/10%B
RP-1/20%B
RP-1/50%B
I sp
 (S
)
Mach Number
 18
1.5 Objectives/Goals 
The overall goal of this research program was to determine whether metallic 
nanoparticles could be used as suitable fuels/supplements for airbreathing propulsion 
systems.  Meeting the overall goal was a multi-step process that required completion of a 
number of sub-tasks or objectives along the way. 
The first objective to be met was to develop an experimental approach/technique 
for evaluating the ignition/combustion characteristics of nanoparticles under various 
conditions.  Once the approach was developed the burning behavior of boron 
nanoparticles needed to be understood in a controlled flame environment so that 
fundamental ignition and combustion behavior could be extracted as functions of 
temperature and gas composition.  Parameters such as critical ignition temperature, gas 
composition, ignition times, and burning times were all to be studied. 
Finally, the behavior of both boron and aluminum nanoparticles was to be 
investigated in a hydrocarbon/air environment simulating ramjet conditions.  In order to 
meet this goal, an airbreathing combustor was to be designed and fabricated in which the 
performance of metallized fuels could be compared to non-metallized fuels.  This also 
included the design of a simple, reproducible particle injection system. 
 
1.6 Scope 
Within this study, the ignition and burning characteristics of boron nanoparticles 
were studied through both a two-dimensional airbreathing combustor and the post-flame 
region of a hydrocarbon/air/oxygen flat flame burner.  Parameters of interest in this 
portion of the study include: 
 19
 
1. Minimum temperature and oxygen concentration required for ignition  
2. Ignition times as a function of gas temperature and composition 
3. Burning times as a function of gas temperature and composition 
 
In addition, the combustion performance of metal nanoparticles laden flows of 
aluminum and boron were compared directly to a pure hydrocarbon/air analog conditions 
in order to determine whether or not these particles can be considered as potential fuel 
additives for high speed airbreathing engines.  The metallized fuels are compared directly 
to analogous ethylene/air tests for fuel performance assessment.  A two-dimensional 
airbreathing combustor was designed and fabricated with optical access for 
characterization of the combustor under a variety of conditions.  The following is a list of 
parameters under consideration during combustor testing. 
 
1. Combustor inlet velocities ranging from 40 ? 70 m/s. 
2. Equivalence ratios from 0.5 ? 0.7 
3. Adiabatic flame temperatures from 1600 K to 2100 K 
4. Particle loadings from ~10 ? 25% by weight of fuel 
5. Nano-sized aluminum versus nano-sized boron particles 
 
1.7 Benefits for commercial applications based upon results of this research 
This study assessed the viability of using nanoparticle additives in fuels for 
airbreathing propulsion systems.  The vast majority of studies included in the literature 
 20
only consider micron-sized particles as fuel additives.  It is quite possible that the use of 
nanoparticles could allow all or most of the energy to be extracted from the metal 
additives.  This result would be especially encouraging for military applications in which 
volume limited systems are nearly always utilized.   
In addition, this research provided fundamental data on the ignition characteristics 
nano-sized boron particles in an oxidizing environment containing air/H2O/CO2.  Burning 
times were extracted for boron nanoparticles.  This data did not previously exist in the 
literature for boron.  Previous studies conducted on nano-sized aluminum showed that 
nano aluminum behaves differently than traditional micron-sized aluminum particles.  
For instance, the reactivity of nano-aluminum increases as the particle size decreases.  
The data shows that nano-aluminum begins to react at lower temperatures than micron 
sized aluminum.  In the case of boron this would be an especially encouraging result 
given its history of inefficient combustion.   
 21
CHAPTER 2 
 
REVIEW OF RELEVANT PREVIOUS STUDIES 
 
2. Review of Relevant Previous Studies 
 Chapter 2 provides a review of studies relevant to the main theme of the current 
research.   
 
2.1 Brief Review of Droplet Combustion and the D2 Law 
In many instances of certain metals? combustion regimes, an analogy exists 
between a metal particle?s combustion behavior and that of a liquid droplet, such as a 
hydrocarbon.  If one considers the well known D2 evaporation law25 that is commonly 
used to estimate the diameter of a burning liquid droplet or particle it suggests that: 
D D to v2 2? ? ?     Eq. 2.1 
where: D = particle diameter at time t 
 Do = initial particle diameter 
 ?v = evaporation coefficient 
and the evaporation coefficient is defined as: 
 
     ? ?? ? ??v s s
l
B? ?8 1ln     Eq. 2.2 
where ?s = solid phase density 
 ?s = thermal diffusivity of the solid phase 
 ?l = liquid phase density 
 22
 B = Spalding transfer number 
The total burning time, tb, for the particle is determined by setting the final particle 
diameter = 0.  Therefore the burning time becomes: 
     t Db o
v
?
2
?      Eq. 2.3 
Thus, if a similar law could be applied to metal particles, one could easily see how 
changing the particle size by 1 order of magnitude would decrease burning times by two 
orders of magnitude, making nanoparticles very attractive when short combustor 
residence times apply. 
  
2.2 Metal Particle Combustion 
In Chapter 1, the great potential of using of metals in propulsion systems was 
demonstrated in terms of specific impulse and density impulse.  Metals possess desirable 
combustion characteristics such as high heats of combustion or energy release and good 
physical properties such as density26,27.  For example the relatively high densities of 
metals increase the performance of solid propellants and fuels, which, in turn increases 
the performance of a propulsion system for the same volumetric capacity.  Metals were 
first added to solid propellants in the 1950?s when it was found that the addition of 
aluminum particles substantially increased propellant performance26,28,29.  Around the 
same time period, metals were being considered for fuels in ramjets30.  Aluminum, boron, 
magnesium, zirconium, and beryllium have all been considered as fuels or fuel additives.   
Aluminum is perhaps the most widely used metal as an additive in propellants and 
fuels.  Boron provides the highest energy per unit volume of all elements.  However, the 
inability to extract all of its available energy in reasonable time frames has limited the use 
 23
of boron in propellants and fuels.  Although beryllium is second to boron in terms of 
energy per unit volume, the use of beryllium has not been seriously considered due to its 
extreme toxicity7.  Even though magnesium and zirconium have each been considered, as 
Figure 2.1 shows, they both have lower heats of combustion than aluminum in terms of 
both mass and volume.  Therefore, between the difficulty in fully combusting boron in 
reasonable combustor residence times, beryllium?s toxicity, and energetic considerations, 
it is easy to see why aluminum has been the most widely used metal in fuels and 
propellants.     
0
20
40
60
80
100
120
140
Aluminum Boron Beryllium Magnesium Zirconium
?HR (kJ/g)
??HR (kJ/cc)
 
Fig. 2.1 Comparison of Heats of Combustion of Selected Metals 
 
The combustion of metals in oxygen is typically classified by the way the metal is 
first oxidized to its smallest oxide31.  This process can either occur with the metal and 
oxidizer in the gas phase (vapor-phase reaction) or with the metal as a condensed phase 
(heterogeneous reaction)31.  Heterogeneous reactions occur at the interface of two 
 24
different substances for different phases such as gas-solid or gas liquid interfaces31.  The 
determination of which metals burn in a heterogeneous combustion mode can be made by 
evaluating the thermodynamic and physical properties of the metal and its oxide.  A 
unique feature of metal burning in oxygen is that the flame temperature developed is a 
specific known value ? the vaporization-dissociation or volatization temperature of the 
metal oxide product29.  This can be attributed to the physical fact that the heat of 
vaporization-dissociation or decomposition of the metal oxide formed is greater than the 
heat available to raise the condensed state oxide above its boiling point.  For a metal to 
burn in the vapor phase the oxide volatization temperatures must be greater than the 
temperature of the metal boiling point; this is known as Glassman?s Criterion29.   
Yetter and Dryer31 summarize the criteria for determining metal combustion 
classification as follows: 
1) Does the available energy exceed the energy required to heat and volatize 
the final metal oxide? 
2) Does the available energy exceed the energy required to heat and vaporize 
the metal itself?  With sufficient energy the combination of a volatile 
product and a volatile metal indicates that the metal will burn much like a 
hydrocarbon droplet. 
3) Intersolubility of the metal and its product is also relevant to combustion 
behavior ? for volatile metals, certain solubility combinations are known to 
lead to disruption and breakup of the original particle. For nonvolatile 
metals, purely condensed phase combustion may result. 
 
 25
Despite all of the particle combustion studies throughout the literature, studies of 
single nano-sized particle combustion are limited.  There is no data in the literature on 
single particle nano-sized boron.  Most of the studies involving nanoparticles utilize bulk 
sample techniques, which will be described in subsequent sections.       
 
2.2.1 Aluminum Combustion Fundamentals 
Aluminum is of great interest for propulsion applications and has been the subject 
of many studies since it was first discovered that the addition of aluminum increased the 
performance of solid propellants.  According to Glassman?s Criterion29, aluminum will 
burn in the vapor phase since its boiling point (2791 K) is lower than the oxide 
volitization temperature (4000 K).  As shown in figure 2.2, Price provides an excellent 
description of the fundamental processes in aluminum droplet combustion27.  In this 
particular portrayal, Price describes the process of an aluminum droplet combusting in a 
mildly convective flow.  Unless the particles undergo some special processing, all 
aluminum particles start out with an impervious oxide (Al2O3) layer around the solid 
aluminum core, which covers the entire exterior surface of the particle.  This oxide layer 
is typically between 2 to 5 nm.  Ignition of the particle is thought to occur when the oxide 
layer is broken down to expose pure aluminum and the oxidative reactions are vigorous 
enough to induce a progressive temperature rise.  The breakdown of the particle oxide 
can occur in two ways.  The first way is due to the fact that the melting temperature of 
aluminum oxide (2327 K) is much higher than that of pure aluminum (930 K).  As the 
particle is heated, the molten aluminum expands and creates thermal stresses inside the 
oxide shell, which could damage the shell exposing pure molten aluminum to the 
 26
oxidizing species.  The other possibility is that the oxide layer undergoes melting itself, 
which would require much higher temperatures for ignition.    
 
Fig. 2.2 Schematic of an Aluminum Droplet Burning in a Mildly Convective Flow27 
  
According to Price27, following ignition the particle temperature is at approximately 
2500 K, with a residue of molten ?retracted? oxide on the surface and a detached flame 
envelope.  The temperature is constrained by the oxide melting point and the aluminum 
boiling point.  Because the oxide melting temperature, 2327 K, is higher than the melting 
temperature of aluminum, 930 K, the aluminum is assumed to be fully melted upon 
ignition.   
The physical processes associated with the various stages of aluminum particle 
combustion are depicted in figure 2.332.  As Price27 points out, the aluminum particles 
have a characteristic oxide coating that must be melted or cracked before the particle can 
ignite.  Although the global oxidation reaction of aluminum produces alumina, Al2O3, 
 27
many suboxides such as AlO, AlO2, Al2O2 exist in the flame zone, with AlO as the most 
prominent28.  The condensation of alumina at its estimated boiling point, ~3800 K, sets 
the approximate temperature of the oxide cloud because Al2O3(l) condenses directly from 
the association of suboxides and the idealized gaseous form, Al2O3(g), immediately 
dissociates back into AlO and O233 as shown in figures 2.227 and 2.428.  Thermochemical 
calculations support the concept of a limiting or constant temperature29.    
 Fig. 2.3 Physical Stages of Aluminum Particle Ignition and Combustion32 
 
 
Fig. 2.4 Schematic of Aluminum Particle Combustion by Melcher28 
                 Liquid Metal                Aluminum Evaporates 
       Solid Metal       Core            Rapidly Above 
           Core              (Tm = 930 K)      2000 K 
 
 
 
Impervious           Egg-Like     Oxide Cap Forms 
Oxide Shell          Stresses Can Fracture      Above ~2300 K 
Oxide Layer < 2300 K 
Expose Raw Aluminum 
 28
The temperature of the system remains constant at the boiling temperature of 
aluminum oxide because enthalpy changes in the total system are absorbed in the 
condensation of the aluminum suboxides or in the decomposition of the liquid aluminum 
oxide31.  As the particle is heated to the melting temperature of aluminum the solid 
aluminum begins to change phase, which results in the particle becoming egg-like.  At 
the same time, the oxide shell restricts the molten aluminum from being exposed to the 
hot environment.  If the particle is heated slowly, the oxide shell will grow in thickness 
around the pure aluminum core32.  As the temperature of the oxide layer increases beyond 
2300 K, the molten aluminum oxide coalesces and forms a pool commonly known as an 
?oxide lobe?.  This exposes the pure aluminum, which can now readily evaporate and is 
available for reaction with the surrounding environment.  Even though the boiling point 
of aluminum is 2767 K, Beckstead32 suggests that aluminum readily evaporates above 
2000 K.  Above 2767 K, the molten particle material will boil violently and begin to 
fragment.  Studies have shown that the residuals from quenching aluminum particles 
form hollow spheres34.  The flame zone is characterized as homogeneous since the 
reactions are aluminum vapor and gaseous oxygen.  The actual aluminum/oxygen flame 
zone is characterized by a luminous detached flame of approximately 3800 K, which is 
expected since the saturation temperature of Al2O3 is about 3800 K.  Since the aluminum 
oxide decomposition process is endothermic, the flame temperature is not affected.   
 
2.2.1.1 Aluminum ? Single Particle Measurements 
There have been numerous studies involving the combustion of a single 
aluminum particle, nearly all of which dealt with micron-sized or larger particles.  Most 
 29
of the studies have sought to identify burn times and or ignition temperatures of single 
particles.  In general, there is a significant amount of scatter in the published data.  
Beckstead attributes this scatter to the different techniques used to obtain the data and the 
lack of a standard definition for the burning time34.  By and large, there were only a few 
types of experiments conducted; 1) propellant ignited particles, 2) gas burner ignited 
particles, 3) laser ignited particles, and 4) shock tube ignited particles.  Parameters of 
interest include oxidizer types/concentrations, pressure, and ambient temperature. 
Friedman and Macek35,36 and Macek37 provide some of the first reported burning 
time data for aluminum particles.  Burning times and ignition delays were deduced from 
particle streak images.  They used propane or carbon monoxide flat flame burners to 
ignite particles ranging from 15-70 ?m.  In these experiments the effect of fuel did not 
play a large role in burning times.  However, testing in both burners produced a fine 
oxide smoke and porous or hollow spheres, which were more numerous in the propane 
flame which contained water in the products as a potential oxidizer.  In addition they 
noted that the particles would fragment and combust vigorously in oxygen rich 
environments.  Friedman and Macek concluded that in their experiments, ignition of the 
aluminum particles only occurred at temperatures corresponding to the melting of the 
oxide layer (2300 K). 
There were many other experiments similar to those of Friedman and Macek?s 
including studies by Davis38, Drew39, Turns40, Olsen15, and Foelsche41 just to name a few.  
Davis38 used a carbon monoxide flame to ignite particles that ranged from 53 to 66 ?m.  
From a set of images, Davis found that the particle?s flame front was several diameters 
larger than the particles themselves, which was evidence of vapor-phase combustion.  In 
 30
addition, Davis witnessed a fine trail of smoke in the particles wake, which was 
determined to be alumina, further evidence of gas-phase oxidation.  Davis also 
determined that particle fragmentation increased for any oxygen concentration over 32% 
by volume, and concluded that the ambient oxygen concentration was more important to 
particle combustion rate than the ambient temperature. 
Turns40 created slurry droplets with diameters ranging from 500 to 1100 ?m by 
adding aluminum powder to JP-10 jet fuel.  The droplets were suspended on silicon 
carbide fibers and then suddenly exposed to the hot exhaust gases of carbon monoxide or 
methane flames.  Using high-speed cinematography, Turns observed that the vapor-phase 
flame front was smaller with the methane flame, which included vapor phase water in its 
products, when compared to the ?dry? environment of the carbon monoxide flame.  
Furthermore, Turns also witnessed the fragmentation of particles in the ?wet? 
environment. 
Olsen15 used a carbon monoxide/hydrogen/oxygen diffusion flame to ignite 
particles ranging from 40 to 80 ?m.  A photomultiplier tube was used to record the 
combustion event.  During some experiments, Olsen interrupted combustion and 
quenched the particles for examination using Scanning Electron Microscopy (SEM) and 
X-Ray analysis.  He found that the oxidizer concentrations in his study, H2O, CO2, and 
O2, played a large role in the burning time of the particle.  Olsen suggests that if one 
considers the particle burning rate to follow the famous D2-law, that the burning rate 
exponent changes during the combustion event, likely starting at 2 during ignition and 
decreasing towards burnout.  The decreasing burning rate is due to the increasing fraction 
of aluminum oxide covering the molten aluminum. 
 31
Foelsche41 used photodiodes to measure the burning time of a small cloud of 
aluminum powder (~22 ?m) inside a combustion bomb with the pressure ranging from 38 
to 145 atm, at 2630 K.  The combustion bomb used nitrogen diluted premixed H2/O2 
mixtures to create a high-pressure, constant volume environment in 20% excess oxygen 
concentration.  Foelsche41 monitored the emission of AlO, a gas-phase intermediate in 
aluminum oxidation as a means of monitoring the burning time.  Since Al2O3 was not 
expected to dissociate much at 2630 K, once all of the aluminum was consumed, the 
source of AlO(g) was also consumed.  The data obtained showed a trend of decreasing 
burning times with increasing ambient pressure.  Foelsche found that for the 22 ?m 
particles that the burning time as a function of pressure could be expressed by a power 
law with an exponent of ?0.55.  
Bucher42,43 used a 150 W laser to ignite aluminum particles (230 ?m) in pure 
N2O, pure CO2, and in mixtures of O2, N2, Ar, and He.  In-situ temperature and species 
distributions around individual burning particles were made using planar laser-induced 
fluorescence.  The study found that the flame diameter around a particle decreased with 
varying environment gas mixture in the following order: O2/Ar, O2/N2, CO2, and N2O.  
AlO was found to be an intermediate species in the combustion reaction while for the 
first time the presence of aluminum-oxy-nitrides was established.  The measurements 
confirmed the idea that the aluminum oxide?s boiling point limits the flame temperature. 
Olsen15 summarizes the majority of the available data on aluminum single particle 
experiments at atmospheric pressure.  Typical burn times range from about 2 to 30 
milliseconds for particles ranging from 10 ? 100 ?m. Aluminum combustion can be 
though of in much the same way as a hydrocarbon fuel droplet.  With that in mind, most 
 32
of the studies on aluminum combustion have attempted to correlate the particle burning 
times with the ?D2 Law?. Beckstead34 provides an excellent summary of the studies on 
aluminum combustion in which he attempts to correlate a significant portion of the data 
available in literature.  He proposes the following correlation for burning times of single 
particles: 
t aDX P Tb
n
eff o
? 0 1 0 2. .     Eq. 2.4 
where the pressure is given in atmospheres, temperature in Kelvin, diameter in ?m, time 
in milliseconds, and Xeff is an effective oxidizer concentration. Xeff takes into account the 
effects of concentration of carbon dioxide, water vapor, and oxygen: 
X C C Ceff O H O CO? ? ?
2 2 2
0 6 0 22. .    Eq. 2.5 
a and n are constants with the following possible values: 
a = 0.0244 for n = 1.5 or 
a = 0.00735 for n = 1.8 
Using this correlation figure 2.5 is obtained for a single particle burning in a 2000 K, 1 
atmosphere environment of air.  The data in the figure has been extrapolated far below 
the range of diameter from which the correlation came from.  This was done only as an 
exercise to provide an estimate for burning time of particles on the nano-scale.  If the 
correlation is extrapolated into the nanometer regime, the correlation gives burning times 
on the order of microseconds for a 100 nm particle. 
 33
10-5
0.001
0.1
10
1000
0.01 0.1 1 10 100 1000
a = 0.0244, n = 1.5
a = 0.00735, n = 1.8
Bu
rni
ng
 Ti
me
 (m
s)
Diameter (?m)
Correlation based on 
experimental data
Extrapolation 
of correlation
Oxidizer: Air
Temperature: 2000 K
Pressure: 1 atm
 
Fig. 2.5 Burning Times of Single Particles Using Beckstead?s34 Correlation 
 
2.2.2 Boron Combustion Fundamentals 
Nearly all of the studies involving boron combustion in current literature have 
been performed with micron-sized particles.  Boron has been studied for many years 
because of its theoretically high energetic potential as demonstrated in figure 2.1.  
However, in most cases the theoretical performance output of boron-based fuels has not 
been met.  The two major reasons why overall propulsion system performance may be 
reduced when using boron are; 1) the ignition of boron particles is significantly delayed 
due to the presence of an oxide layer on the particle surface, and 2) the energy release 
during the combustion process of boron particles in hydrogen containing gases is 
significantly lowered due to the formation of HBO244.  
Ulas and Kuo44 claim that the combustion of boron particles is defined by a two-
stage process.  The first stage of boron combustion was considered as the removal of the 
 34
oxide layer.  This oxide removal process is a slow, kinetic and/or diffusion controlled 
process, which constitutes a significant portion of the overall burning time of the particle.  
After removal of the oxide layer, the second stage begins with the actual combustion of 
the pure boron material.  The full energy release of boron occurs when the final products 
condense to B2O3(l), which is thermodynamically favored below 2400 K44.  In hydrogen 
containing environments, the formation of HBO2 severely inhibits boron combustion.  
The conversion of HBO2(g) to B2O3(l) in an oxygen and hydrogen atmosphere is slow; 
thereby reducing the rate of energy release.  Figure 2.6, adopted from Beckstead45, shows 
the basic physical processes for boron ignition and combustion.  Similar to aluminum, the 
boron particle has a thin oxide layer, which surrounds a pure boron core.  The boron 
oxide shell however, melts at a much lower temperature (722 K) than alumina.  
According to Glassman?s criterion29, since B2O3 has a volatization temperature of 2340 
K, and boron?s boiling temperature is 4139 K, boron will not burn in the vapor phase.   
Fig. 2.6 Physical Stages in Boron Ignition and Combustion45 
  
As the temperature of the particle rises, the oxide shell melts and forms a liquid 
layer on the surface; during this time some of the oxide vapor begins to diffuse into the 
gas phase.  The combination of the liquid and vapor phase oxide layer forms a significant 
 Solid Metal       Boron Oxide       Minimal Reaction 
     Core   Melts (Tm = 722 K)    Due to Low Vapor  
                  Pressure Below 4000 K 
 
       
 
         
 
 
   Pervious      Liquid Layer   Oxide Boils at ~2300 K 
 Oxide Shell      on Solid Core  Inhibiting Combustion 
 35
diffusive barrier between the oxidizing environment and the pure boron core, which 
inhibits oxidation of the pure boron.  This stage of combustion is diffusion limited since 
the diffusion rates are slow compared to the kinetic rates.  As the particle is heated further 
to ~2300 K the oxide layer becomes completely vaporized and the boron begins to melt.  
The low vapor pressure of boron at this temperature prevents the particle from 
evaporating below its boiling point.  Now the oxidizer can more easily diffuse to the 
surface than before because the oxide layer is entirely in the vapor phase.  This may 
reduce the oxidizer diffusion time to less than that of the kinetic reaction time, shifting 
the burning rate limiting mechanism from the diffusion regime into the kinetic regime 
depending on the environment and particle size.   
Because of the potential for a shift in the controlling mechanism once ignition is 
complete and full-fledged combustion is achieved, each regime may also establish its 
own distinct physical processes. The diffusion-limited regime tends to create pronounced 
species and thermal spatial gradients at the surface of the particle, in the oxide layer, and 
in the oxidizing environment surrounding the particle50.  The species gradients promote 
reactant diffusion towards the surface and product diffusion away from the surface, while 
the thermal gradients encourage heat transfer from the flame zone towards the particle. 
After the initial stage of combustion, the boron particles may end up in a 
kinetically controlled regime.  In this case, the spatial species and thermal gradients 
become less influential on the burning rate.  Now, the controlling mechanism is the 
reaction rate to the individual heterogeneous kinetic reactions at the particle surface.  The 
primary overall surface reaction occurring at this point is: 
  2B(l) + O2(g) ? B2O2(g) ? 110 kcal/mole 
 36
A large majority of the surface reactions follow one of two paths to the overall reaction 
depending on the water content in the oxidizing environment.  These are shown in figure 
2.745. 
            
 
 
 
In an O2 environment 
 
 
? ? ? ? ? ?BO O OBOg g g? ?
12
2  
? ? ? ? ? ?M OBO BO B O Mg g g? ? ? ?2 3  
 
 
 
 
 
 
  
Fig. 2.7 Elementary Heterogeneous Reactions Associated with Boron Particle 
Oxidation 
 
One way to determine the dominate regime during the particle combustion is 
through the Damkohler number, which is defined as the ratio of the diffusion burn time to 
the kinetic burn time46.  
    ? ?Da tt M kPd X
D iY
b diff
b kin
B o o
o
? ?
?
?
?
,
,
,
,ln
2
4 1?
   Eq. 2.6 
where MB is the molecular weight of boron, k is the thermal conductivity, P is the 
pressure, do is the initial diameter of the particle, XO2 is the ambient mole fraction of 
In an O2 and H2O environment 
? ? ? ? ? ?M H O OH O Mg g g? ? ? ?2  
? ? ? ? ? ?M BO OH HOBO Mg g g? ? ? ?  
? ? ? ? ? ? ? ?HOBO BO HBO OBOg g g g? ? ?  
? ? ? ? ? ?M HBO BO H Mg g g? ? ? ?  
? ? ? ? ? ?OBO BO B Og g g? ? 2 3  
? ? ? ? ? ?2 12 2 2H O H Og g g? ?  
 
? ? ? ? ? ?B O BOl g g? ?
12
2
? ? ? ?B B O BOl g? ?2 3 3
? ? ? ?BO BO B Og g? ? 2 2
 37
oxygen, ? is the density of boron, D is the diffusivity YO2 is the ambient mass fraction of 
oxygen.  The Damkohler number compares the controlling rate mechanisms and offers 
insight into there dependency.  For example, large particles combusted at high pressures 
(Da > 1) will tend to be diffusion rate limited, while small particles at lower pressures 
(Da < 1) will be kinetically dominated.  
   
2.2.2.1 Boron - Single Particle Experiments 
A number of studies have been conducted on single boron particles46-51.  Macek47-
49 provides some of the first single particle combustion measurements for boron.  Similar 
to their aluminum particle experiments, Macek47 examined single particle combustion of 
crystalline boron powder ranging from roughly 35-45 ?m in the products of carbon 
monoxide or propane and oxygen flames.  Macek found ignition temperatures to be 
consistently in the rage of 1850-2000 K and goes on to describe a two-stage combustion 
process.  During the first stage the particle ignites and burns brightly for a short period of 
time, then becomes extinguished, and finally reignites to burn to completion in the 
second stage.  During the second stage, the particle burns brighter and longer.  In addition 
they studied this two-stage phenomenon spectroscopically and found that both BO and 
BO2 were present in both stages.  They also found that boron particles required four times 
as much time to complete their combustion than similar sized aluminum particles in 
similar environments.  Macek47 also observed that the burning times decreased with 
increases in ambient gas temperature and oxygen mole fraction.  Finally they observed 
that particle burning times decreased in the presence of water vapor.  
 38
In another experiment, Macek48 investigates the effect of pressure and oxidizer on 
the burning time of 75 ?m crystalline boron particles at 2000 K.  The particles were 
ignited by laser ignition.  In this study they found that the burning time of a particle in air 
and in an oxygen/argon environment decreased from about 45 milliseconds to about 20 
milliseconds when the pressure was increased from 1 to 35 atmospheres.  In pure oxygen 
they deduced burn times of only 6.8 milliseconds.  Finally, Macek48 found that in a pure 
carbon dioxide environment and an oxygen depleted environment of 7% O2 and 93% N2 
that a steady state combustion is never achieved.  Furthermore, the particles could only be 
brought to a burning regime by an external energy heat flux, and that the burning regime 
could only be maintained for a brief period before the particle extinguished.  
Yeh and Kuo46, provide a very detailed summary of boron particle combustion 
testing up to 1996.  In addition, they conducted a series of experiments of 2-3 ?m 
amorphous boron powder and 3 ?m crystalline boron particles in the post flame region of 
a flat flame burner.  Similar to Macek47, they found that the particles burned in two-
stages.  In addition, Yeh and Kuo46 found through environmental scanning electron 
microscope (ESEM) images that at elevated temperatures the diffusion of dissolved 
boron into molten B2O3(l) dominates the diffusion of gaseous O2 through the B2O3(l) 
layer.  Dissolution of solid boron into the boron oxide layer caused liquification of boron 
particles at relatively low temperatures (940 oC).  X-ray diffraction analysis showed that 
the surface characteristics of boron particles changed from a crystalline to an amorphous 
structure after heating to 940 oC, meaning that the B2O3(l) layer lost its identity at 
elevated temperatures.  Finally, Yeh and Kuo constructed a theoretical model in which 
they demonstrated that the product of the pressure and particle diameter determined 
 39
which burning regime that the particles were under during the second stage of 
combustion.  For example, for Pd << 75 atm-?m, the second stage was dominated by 
chemical kinetics, for Pd ~ 75 atm-?m, both diffusion and kinetics play large roles, and 
finally for Pd >> 75 atm-?m, the second stage combustion was dominated mainly by the 
diffusion of oxygen.   
Ulas44 performed studies on 1 ?m amorphous and 3 ?m crystalline boron particles 
in the exhaust of a flat flame burner, similar to the aluminum studies discussed earlier.  
The flat flame burner combusted mixtures of methane, oxygen, nitrogen trifluoride, and 
nitrogen.  Ulas monitored particle streaks and deduced burning times.  He found that in 
the presence of fluorine, boron does not exhibit the two-stage burning phenomenon 
observed in non-fluorinated environments.  In addition, the presence of hydrofluoric acid 
(HF) increased the overall burning time of the particles, while atomic fluorine reduced 
burning times significantly compared to non-fluorinated environments. 
Foelsche et. al51 examined the effects of pressure (30-150 atm.), temperature 
(2440-2830 K), and excess O2 concentrations on the burning properties of 24 ?m boron 
crystalline particles in a constant volume combustion bomb using H2/O2/N2  mixtures to 
create the environement.  In this study, they found that particle ignition delays were 
reduced with increased pressure, decreased particle size, and increased temperature.  As 
an example, they found that the ignition delay could be correlated to a power law with an 
exponent of ?0.5. In addition Foelsche examined the effects of two proposed ignition 
enhancing agents and found that neither CO2 nor HF reduced ignition delay, and that HF 
actually increased the ignition delay.  Burn times were measured by monitoring the 
emission of BO2 which is a reactive intermediary gas-phase species formed through 
 40
particle ignition and combustion.  According to gas phase modeling studies conducted by 
Yetter52, BO2(g)  gets converted to B2O3(g) or HBO2(g) products quickly depending on 
how much water vapor is present.  Therefore, once the source of B/B2O3 was consumed, 
then so was the source of BO2(g) as well.  The BO2 spectrum is readily observable during 
both the ignition and combustion stages53.  
      
2.3 Metal Combustion in Air breathing Engines 
As was demonstrated in Chapter 1, metal additives can potentially provide great 
benefits for air breathing propulsion systems.  The benefits arise mainly in the form of 
energy density for volume limited systems.  Pure metals or the addition of metals to more 
conventional fuels generally increase the energy density of a fuel over non-metallized 
fuels.  For that reason, there have been numerous studies incorporating them into fuels 
mainly for solid fuel ramjet or ducted rocket applications.  Metals have been considered 
as fuels or additives to fuels in air breathing propulsion systems since the 1930?s when 
the then Soviet Union successfully flight-tested a tandem boosted ducted rocket using a 
magnesium/aluminum solid fuel2.   
Magnesium, aluminum, boron, and combinations of each have been considered 
most often for airbreathing engines.  For reasons described earlier aluminum and boron 
have received the most attention.  However, magnesium has drawn interest despite the 
fact that it cannot match the theoretical output of aluminum and boron because it is 
known to ignite at low temperature and to burn well even at low pressure and low 
air/propellant ratio54. 
 41
In the 1950?s30 NACA performed a research program utilizing aluminum as the 
sole fuel.  They tested a two inch diameter ramjet engine using both powdered aluminum 
and wire aluminum as the fuel source and measured combustion efficiencies of about 
75% over a wide range of fuel/air ratios at relatively low combustor inlet velocity of 35 
m/s (115 ft/s). 
Schadow and others55-57 experimented with boron-based fuels in ducted rocket 
applications with boron loadings up to 55% by weight.  Typical particle sizes ranged 
form sub-micron to 3 ?m.  In some cases55 the experiments were conducted with the 
boron in a fuel rich solid propellant matrix and in others, the boron was introduced to the 
combustor in particle form by a carrier gas56,57.   In these experiments relatively high 
combustion efficiencies (greater than 90%) were only achieved when local gas 
temperatures exceeded 2000 K.  In fact, Schadow57 suggests that a minimum temperature 
of 2300 K is required for very efficient combustion of boron particle laden flows.  This 
temperature corresponds to the melting temperature of boron and the boiling temperature 
of boron oxide. 
Gany et. al58,59 have examined the combustion characteristics of boron-based solid 
fuels in a typical SFRJ setting.  Gany58 suggests that part of the problem in achieving 
efficient combustion of boron in a SFRJ setting can be attributed to the manner in which 
the material is ejected from the condensed fuel surface to the gas flow.  More 
specifically, the ejection is often irregular with large fragments of uncontrollable 
dimensions and velocities entering the gas flow.  Furthermore, the gas flowfield in 
SFRJ?s is typically characterized by non-uniformities of temperature and species 
concentrations, which hamper the ignition, and sustained combustion of boron particles 
 42
within the combustor.  As a means of promoting more efficient ignition and combustion 
Gany proposes the use of bypass air58,59 and possibly the use of other additives such as 
titanium58.  By including an additive such as titanium, Gany58 suggests that one can take 
advantage of the exothermic reaction between titanium and boron to enhance particle 
preignition rates leading to shorter ignition delays.  
More recently, Gany60 provided a detailed account of the difficulties in achieving 
efficient boron combustion and their implications in air breathing engines.  There are a 
number of contradictory necessities to achieve efficient combustion.  For instance, high 
temperatures are required for ignition and sustained combustion of boron. However, as 
Gany points out, the most energetic output of boron combustion occurs when the final 
product is B2O3(l), yet at high temperatures above about 2300 K (the boiling point of 
B2O3) combustion would yield gaseous B2O3 whose heat of formation is about 1/3 that of 
liquid phase B2O3.  Furthermore, combustion in hydrogen containing environment 
ultimately leads to the formation of HBO2(g) whose heat of formation is about 60 times 
less than B2O3(l).  Conversion of HBO2(g) to B2O3(l) is very slow, so in combustors with 
relatively short residence times, formation of B2O3(l) is unlikely to occur. Gany60 calls 
this phenomenon an ?energy trap?.    
       
2.4 Nanoparticle Work 
The majority of work to date on nanoparticles has involved testing of bulk 
samples.  A few burning time measurements of nano aluminum have been made16,17, but 
the measurements do not guarantee single particles and are more likely agglomerates.  
Son61 studied the combustion of nano aluminum in air.  In one experiment, Son used a 
 43
combustion chamber made of Plexiglas, to combust a ?pile? of aluminum at atmospheric 
pressure in air, nitrogen, nitrogen and argon mixtures, argon and oxygen mixtures, and 
pure oxygen environments.  In another experiment, he studied the ignition requirements 
of nano aluminum using a laser.  He found that the nanoaluminum was much easier to 
ignite than micron-sized aluminum, and that it readily self deflagrated in air.  In addition, 
the nano aluminum exhibited a two-stage burning phenomenon in air.  Son determined 
the first stage to be surface burning and the second stage to be a bulk reaction.  They also 
found that when burned in air, nano aluminum reacts with oxygen and nitrogen to form 
oxides, nitrides, and aluminum oxynitrides.  Finally, Son found that it was possible to 
burn the nano aluminum in a nitrogen/argon environment where aluminum nitrides are 
formed.  
Parr16 studied the ignition delay, burning times, and ignition temperatures of 
several different types of particles including nanoaluminum in the post flame region of a 
hydrogen-oxygen-argon Hencken Burner, such that the oxidizing environment was water 
vapor alone.  They estimate that the nanoaluminum agglomerates to about 2 ?m as it is 
introduced into the test section.  Their results show that the ambient temperature has a 
pronounced effect on the burning time and ignition delay.  However, of interest is that 
below about 10 ?m the ignition temperature decreases with decreasing particle size, 
where as ?normal? micron sized aluminum generally ignites near 2200 K.  In fact, the 
lowest ignition temperature measured was only 1075 K.  Although the nanoaluminum 
was somewhat agglomerated, typical burning times were less than 1 millisecond at 
~2000K, which is significantly lower than burning times of micron sized aluminum, yet 
 44
much longer than one could extrapolate from the D2 type dependence implying that 
kinetic limited burning is beginning to dominate over diffusion limited burning. 
Bazyn17 investigated reaction times and temperatures of nanoaluminum particles 
burning in the reflected shock region of a shock tube.  He deduced burning times by 
monitoring the broadband emission of light because under most conditions molecular 
emission of AlO was not observed.  They found that reactions of aluminum occurred at 
temperatures as low as 1200 K, and that burning times were on the order of tenths of 
milliseconds. 
Park et. al. 24 provides the only true single particle oxidation data of 
nanoaluminum available in the literature today.  In this study, Park oxidized aluminum 
particles that were either created in situ by a DC arc discharge or laser ablation, or were 
obtained commercially in an aerosol flow reactor in temperatures up to 1100 oC.  They 
size selected particles with a Differential Mobility Analyzer (DMA), to send a 
monodisperse aerosol through the reactor and monitored changes in the particle with a 
Single Particle Mass Spectrometer (SPMS).  Park found that the reactivity of the 
aluminum particles was enhanced as the particle size decreased.  They also found that the 
reaction rates of single particles were much different than those measured by bulk 
techniques such as thermal gravimetric analysis (TGA) and that they were consistent with 
an oxide diffusion controlled rate limiting step.        
 
2.4.1 Nanoparticles in Propulsion 
A number of studies have been conducted on utilizing nanoparticles in propulsion 
systems21-23,62.  Mordosky21 studied the potential of nano aluminum for gel propellant 
 45
rocket engines.  The nano aluminum was gelled with RP-1 and combusted with gaseous 
oxygen.  The purpose of the study was to demonstrate a system similar to a liquid rocket 
engine, but with a high energy density fuel for greater performance in volume limited 
systems.  In addition, Mordosky21 suggests that by gelling a propellant one can increase 
the safety of the vehicle since the possibility of leakage would be negated.  Mordosky, 
incorporated the nano aluminum in the fuel in quantities up to 55% by weight.  Some 
measurements of c* (characteristic exhaust velocity) efficiencies of greater than 90% 
were made.  Furthermore, the c* efficiencies were typically higher than those achieved 
by Palaszewski20 using micron sized aluminum. This suggests great promise for the use 
of nano aluminum in gelled propellants. 
Risha22 investigated the effects of several different types of nano particles on the 
fuel regression rate in a hybrid rocket engine.  The nano particles studied include several 
different types of aluminum, boron, and boron carbide.  He created numerous fuels with 
varying amount of nano particles added to a hydroxyl terminated polybutadiene (HTPB) 
solid matrix and compared regression rates as well as c* efficiencies of the fuels to the 
baseline fuel which was just plain HTPB.  Risha found that the majority of the fuels 
tested with nano particles exhibited an increase in the linear regression rate of the fuel 
compared to the plain HTPB fuel.  This is a very important outcome since one of the 
major downfalls of the hybrid rocket motor is the typically low regression rate of the 
fuels.  He attributed the increase in regression rate to the increased energy release of the 
particle near the surface of the fuel.  In nearly all of the cases, the measured c* 
efficiencies were comparable to the baseline HTPB fuel, with boron and boron carbide 
typically showing the lowest efficiencies. 
 46
Palaszewski23 investigated the potential of nano aluminum gelled with JP-8 as a 
fuel for a pulse detonation engine.  Metal loadings between about 5 and 25% by weight 
were studied.  Palaszewski demonstrated combustion without the addition of oxygen with 
loadings between 12 and 18% by weight.  Again, this is an important finding because by 
creating a more volumetric energetic fuel, one can improve the performance of systems 
that are volume limited. 
In a set of strand-burning studies, Mench62 investigated the effects of nano 
aluminum on the burning rates of solid propellants.  The study essentially looked at two 
identical solid propellants with 18% by weight aluminum: one with micron-sized 
aluminum, and the other with 50% of the micron aluminum (i.e. 9% by weight of the 
overall formulation) replaced by nano aluminum.  The study found that addition of the 
nano aluminum increased the burning rate of the propellant by 100%.  Increasing the 
burning rate of a solid propellant can increase the thrust of a rocket motor, allow for a 
more compact design, or allow the grain design to be optimized in such a way as to 
improve the mass fraction.     
 
2.4.2 Advantages and Disadvantages of Nano-Sized Particles 
In combustion applications, the ignition and burning times of single particles is 
directly related to the initial particle size.  The smaller the particle, the faster the entire 
particle can heat up and for some metals reach its vaporization temperature.  This allows 
the metal vapors to react more readily with the surrounding oxidizer.  Some specific 
examples of the benefits of nano particles for propulsion related applications include: 
1) Shortened ignition delay 
 47
2) Significantly decreased burning times 
3) They can provide greater flexibility in terms of designing new fuels or 
propellants 
4) Can serve as a gelling agent, minimizing inert or low energy ingredients 
5) Increased burning rates and regression rates can improve thrust, and mass 
fraction 
Despite the apparent benefits of nanoparticles for propulsion systems, they do 
possess some undesirable characteristics.  For example, most aluminum or boron 
nanoparticles have an oxide layer.  Although the oxide layer is generally only a few 
nanometers, as the particles get smaller, the oxide layer begins to occupy a larger fraction 
of the volume of the particle.  This in effect removes some of the available energy from 
the particle. 
 
2.5 Laminar Jets Exhausting into a Quiescent Atmosphere  
As part of this study, a laminar jet is used to inject nanoparticles into the post 
flame region of a flat flame burner.  The details of the particular experiment is discussed 
in subsequent chapters.  However, a brief review is provided here to describe the 
behavior of a laminar jet as it exhausts into a quiescent atmosphere.  In particular, the 
centerline axial component of velocity is of interest.  A simple analysis is provided by 
Turns63 of a nonreacting-laminar jet of fluid flowing into a quiescent atmosphere.  In this 
analysis the following assumptions are made63: 
 
1) The velocity profile is uniform at the jet exit 
 48
2) Close to the jet exit there exists a region called the potential core.  Within the 
potential core, the effects of viscous shear and diffusion have yet to be felt 
3) The molecular weights of the jet and the atmosphere are equal, leading to 
constant density 
4) Species molecular transport is by simple binary diffusion governed by Fick?s 
Law 
5) Schmidt number, Sc D? ? , is unity.  Implies that momentum and species 
diffusivities are constant and equal 
6) Only radial diffusion of momentum and species is important, axial diffusion is 
neglected  
 
Throughout the flowfield, the initial jet momentum is conserved.  As the jet 
exhausts into the atmosphere, some of its momentum is transferred to the atmosphere.  
Therefore, the jet velocity decreases and the atmosphere is entrained into the jet as it 
proceeds downstream.  This can be represented by an integral form of momentum 
conservation: 
   ? ? ? ?2 2 2 2? ? ? ?r x u r x rdr u Rx e e, ,? ?    Eq. 2.7 
Where R is the jet radius, re, and ue refer to the jet exit density and velocity respectively.  
Similarly, the carrier gas concentration field is governed by the same processes as the 
velocity field, namely convection and diffusion and the mass of the fluid exhausting from 
the jet is conserved: 
 49
   ? ? ? ? ? ?2 2? ? ? ?r x u r x Y r x rdr u R Yx CG e e CG e, , , ,? ?  Eq. 2.8 
YCG refers to the mass fraction of the carrier gas.  The governing equations for this 
problem come from the boundary-layer equations, of mass, axial momentum, and species 
conservation: 
? ??
?
?
?
u
x r
ru
r
x r? ?1 0    Eq. 2.9 
u ux u ur r r r urx x r x x?? ?? ? ?? ??? ? ??? ???             Eq. 2.10 
       u Yx u Yr D r r r Yrx CG r CG CG?? ?? ?? ??? ? ??? ???1                       Eq. 2.11 
The boundary conditions for this problem are as follows: 
    ? ?u xr 0 0, ?               Eq. 2.12 
     ? ???ur xx 0 0, ?               Eq. 2.13 
    ? ???Yr xCG 0 0, ?              Eq. 2.14 
    ? ?u xx ? ?, 0              Eq. 2.15 
    ? ?Y xCG ? ?, 0                     Eq. 2.16 
    ? ?u r R ux e? ?,0                   Eq. 2.17 
    ? ?u r Rx ? ?,0 0              Eq. 2.18 
    ? ?Y r R YCG CG e? ? ?, ,0 1            Eq. 2.19 
    ? ?Y r RCG ? ?,0 0             Eq. 2.20 
 
 50
The velocity field can be obtained by assuming the profiles to be similar.  In this 
case, this implies that the radial distribution of ux(r,x), when normalized by the local 
centerline velocity ux(0, x), is a universal function that depends only on the similarity 
variable r/x.  The solution for the axial component of velocity is given by Schlicting64: 
    u Jxx e? ??
??
?
??
?3
8 1 4
2 2
? ?
?             Eq. 2.21 
where Je is the jet initial momentum flow, 
    J u Re e e? ? ?2 2                         Eq. 2.22 
and ? contains the similarity variable r/x, 
    ? ? ? ?? ??? ???316 1
12
e eJ r
x                        Eq. 2.23 
By substituting Eq. 2.22 and 2.23 into 2.21, the centerline axial component of velocity is 
obtained. 
     u u Rxx CL e, ? 38
2 2
?              Eq. 2.24 
 
 
 51
CHAPTER 3 
 
EXPERIMENTAL APPROACH 
 
3. Experimental Approach 
Chapter 3 discusses the details of the experiments utilized to meet the goals of this 
study.  There were two major experiments conducted, a flat flame burning experiment 
and an airbreathing combustor experiment, each of which had its own goals and 
challenges to be met.  Within the two main experiments there were several sub-
experiments to be conducted. 
 
3.1 Flat Flame Burning Experiment 
A series of experiments were conducted in order to determine critical parameters 
with regards to the ignition of the particles as well as their reaction times.  Only SB99 
Boron nanoparticles (described later) were examined in this portion of the study.  To date 
there have been several studies investigating these parameters in nano-aluminum16,17.    
 For this set of experiments a commercially available Mckenna Flat Flame burner 
using a methane/oxygen/air mixture was utilized.  The burner is designed to provide a 
stable laminar premixed flat flame for long durations.  It is water-cooled and constructed 
primarily of stainless steel.  During burner operation, the cooling water was run at 
approximately 0.8 liters per minute.  Figure 3.1 is a schematic diagram of the burner, 
while Figure 3.2 is a photograph of the burner on the test stand.  The stainless steel 
sintered plug model was used in this study. 
 52
 
 
Fig. 3.1 Schematic Diagram of McKenna Flat Flame Burner 
 
At the time of testing, the boron nanoparticles were placed in distilled water to 
form a solution with a fixed concentration of 1% by weight.  Typically. 25 ml of distilled 
water and 25 mg of boron was used in the experiments.  Ideally, one would like to match 
the total number of particles in the solution to the total number of droplets that are 
produced by the atomizer, such that each droplet would statistically contain only one 
particle.  However, once nanoparticles are aerosolized, the signature of particle 
combustion is no longer visible to the human eye, and diagnostic signal levels become 
too low.  Therefore, in order to achieve a reasonable signal to noise ratio, the mass 
 53
concentration of particles was held constant for all experiments rather than the particle 
number concentration.   
The solution was subjected to a sonicating-atomizer to create an aerosol using air 
as the carrier gas with a flow rate ranging from approximately 0.61 to 1.15 liters per 
minute.  Figure 3.3 is a schematic diagram of the components of the sonicating atomizer.  
A photograph of the actual components is included in Appendix A.  The carrier gas was 
regulated by a rotameter and the flow comes into the atomizer via tygon tubing from the 
side and leaves as an aerosol through the top.  The main body and the two end caps of the 
atomizer are made of plexi-glass.  The end cap located at the top of the atomizer was 
solid plex-glass with a hole in the center that had a tube attached, which served as the exit 
for the aerosol flow.  A stainless steel cylinder was placed on top of the atomizer to 
ensure that it stays closed and sealed.  The end cap at the bottom of the atomizer had a 
large hole cut out of it and utilized a thin film to seal the bottom portion of the atomizer.  
The entire atomizer with the solution was placed into the base of a standard commercial 
humidifier.  The humidifier unit was then turned on and set to approximately ? of the 
maximum power.  The humidifier causes the film to vibrate quickly, which created a mist 
of the solution that was carried away by the carrier gas in the form of an aerosol.        
Prior to injection in the post-flame region of the flat flame burner, the aerosol 
passed through 3 silica-gel dryers placed in series to remove all of the water content from 
the particles.  The aerosol was then injected into the post-flame region of the flat flame 
burner via a stainless steel tube with an inner diameter of 1 mm that was located 5 mm 
above the burner surface.  The tube was placed at the very edge of the burner such that 
 54
the particles would traverse the full diameter of the burner.  Figure 3.4 is a schematic 
diagram of the test set up.   
In the vast majority of previous studies using flat flame burners35,36,46,47 the 
particles were injected in the direction of the burner flow and the particles were assumed 
to be traveling at the flame velocity.  In this study, the particles are injected perpendicular 
to the direction of the flow.  This allows for adjustment of the particle velocity to allow 
for increased resolution of events happening in a short duration.  By increasing the 
particle velocity, the events occur over a larger spatial duration, thereby increasing the 
resolution.  
 
 
Fig. 3.2 Flat Flame Burning Test Configuration 
 
 55
 
Fig. 3.3 Schematic Diagram of Sonicating Atomizer 
 
 
Burner
Particle
Injection
Compressed Air
Atomizer
Silica Gel Dryers
O2 CH4
Rotameters
 
Fig. 3.4 Schematic Diagram of the Flat Flame Burning Experiment 
   
 
 56
3.1.1 Flow Controls for Flat Flame Burning Experiment 
In order to obtain meaningful data, the flame characteristics need to be understood 
in great detail.  As was mentioned, in this study the flat flame burner was operated on a 
mixture of oxygen, air, and methane.  All gas flows were controlled through the use of 
high precision rotameters.  All rotameters were calibrated for the gas being used by 
utilizing a calibrating gillabrator.  Figure 3.5 is an example of the calibrations obtained 
for the rotameters.  As the figure shows the gas flow rates behavior linearly with respect 
to the settings of the rotameter.  The calibrations for the other gases can be found in 
Appendix C. 
0
5
10
15
20
25
0 10 20 30 40 50 60 70 80
y = -1.32 + 0.36914x   R2= 0.99975 
Ox
yg
en
 Fl
ow
 R
ate
 (L
/m
in)
Rotameter Setting  
Fig. 3.5 Example of Rotameter Calibration  
 
3.1.2 Particle Image Velocimetry (PIV) 
Since the particles were injected perpendicular to the burner flow direction, the 
velocity of the particles must be measured in order to ensure that a reasonable 
measurement of ignition delay and burning times could be made.  Particle Image 
 57
Velocimetry (PIV) was employed in order to resolve the particle velocity at a number of 
conditions.  The basic principle of PIV is to record the light scattered by tracer particles 
that are a part of the flow or that have been added to the flow.  A typical PIV set up 
includes a high-powered laser and a digital camera.  The laser and camera are 
synchronized such that as the laser is pulsed, the camera collects an image.  There are 
also generally some optic components, which convert the laser beam into a laser sheet.  
The camera is placed perpendicular to the plane of the laser sheet.  The laser sheet is 
pulsed, switched on and off very quickly, twice at a know interval.  The light from the 
laser sheet is scattered by the particles and collected by the camera as two separate 
images. The first image collected shows the initial position of the seeding particles, and 
the second image shows their final positions.  The images are broken into interrogation 
regions from which displacement vectors may be determined through a correlation 
technique.  With the knowledge of the laser pulse interval, the displacement vector can be 
converted into a velocity vector. 
PIV has many advantages over other velocity measurement techniques.  For the 
most part, as long as the seed particles are properly chosen65 the technique is virtually 
non-intrusive.  It is important to choose appropriate seed particles such that they:  
1) Are sufficiently small that they ?follow? the flow correctly 
2) Are large enough that they provide reasonable scattering signals 
To determine if the particles are small enough, one needs to take into consideration the 
Stokes Number, which will be discussed later.   
The fact that PIV is virtually non-intrusive with appropriate seed particles alone is 
a large advantage over other measurements such as pitot tubes or hotwires.  Furthermore, 
 58
since the laser provides a sheet, velocity measurements are obtained for an entire plane 
rather than just point measurements. 
 A total of 3 different injection velocities were measured corresponding to three 
different volumetric flow rates controlled by a rotameter.  All measurements were made 
in ?cold flow?, i.e. non-reacting over a wide range of crossflows, and momentum flux 
ratios.  For this experiment, an Nd:YAG Laser operating at wavelength of 532 nm was 
synchronized with an Intensified Charge Coupled Device (ICCD) camera.  The PIV 
system used in this study was from LaVision.  In some cases PIV can be sensitive to the 
correlaion technique used during the processing procedure.  During this study, for all 
cases, the standard PIV sum of correlation technique was used to process the data.  Figure 
3.6 is a schematic diagram of the PIV setup. 
In order to conduct the PIV experiments in a simple manner that required the least 
modification to the test rig, the sonicating-atomizer was utilized with no particles placed 
in solution.  Only 25 ml of distilled water was placed into the atomizer, and the silica gel 
dryers were removed from the injection lines.  The water droplets created by the atomizer 
were then used as the seed particles for the PIV experiments.  An important parameter to 
consider when using seed particles in any light scattering technique is the Stokes Number 
which provides a measure of how well the particles follow the flow.  The Stokes Number 
is the ratio of the characteristic particle stopping time or distance to the characteristic 
fluid time or distance.  The characteristic particle stopping distance is given by: 
    S D Uparticle p? ? ?
2
18        Eq. 3.1 
where ? is the viscosity of the fluid, ?p is the seed particle density, U is the velocity of the 
fluid, and D is the seed particle diameter.  The characteristic fluid time in this case is 
 59
scaled with the jet core length.  In this case the jet core length was taken to be six times 
the diameter of the injection tube.  The average seed particle diameter (water droplets) 
created by the sonicating atomizer is 5 ?m.   
The water droplet size was determined by placing a known mass of salt in a 
known mass of water.  The atomizer was then turned on to begin spraying the water 
droplets assuming that the mass concentration of salt in a droplet was the same as the 
total mixture.  The aerosol was then passed through a series of silica-gel dryers as shown 
in figure 3.4 to ?dry off? the particles.  At this point, the aerosol only contained the salt 
particles.  The aerosol was then passed through a Differential Mobility Analyzer (DMA) 
to determine the size of the salt particle.  Since the size of the salt particle was now 
known, and the mass fraction of salt in the water was known, solving for Dwater in the 
following set of equations leads you to the size of the water droplet.   
m V Dsalt salt salt salt salt? ?? ? ?* *
3
6    Eq. 3.2 
m D Dwater water water salt salt? ?? ? ? ?
3 3
6 6    Eq. 3.3 
 
Given these parameters the Stokes Number in these experiments ranged from 0.16-0.31.  
Since the Stokes Number was less than unity, the water droplets did in fact follow the 
gaseous flow sufficiently for the PIV measurements.  
 60
PC
Nd:YAG Laser
Synchronizer
ICCD
Camera
Burner
Particle
Injection
Vertical
Laser
Sheet
Beam
Optics
 
Fig. 3.6 Schematic Diagram of PIV Experiment 
 
3.1.3 Temperature Measurements 
In order to know the temperature of the environment the particles were exposed to 
B-type (Pt-6%Rh/Pt-30%Rh) thermocouples with bead sizes of approximately 200 ?m 
were used for the measuring the temperature.  B-type thermocouples are suitable for 
oxidizing or inert environments and have a range of 0 ? 1700 oC.  The typical error 
associated with a measurement is 0.5% above 800 oC.  The time constant associated with 
an exposed thermocouple of 200 ?m is approximately 0.2 seconds.  The time constant is 
defined as the time required to reach 63.2% of an instantaneous temperature change.   
A thermocouple was placed on a ?slide? that was marked at 10 mm, 30 mm, and 
50 mm axially from the injection location, at a location of 5 mm above the burner 
surface.  The thermocouples were placed in the center of the burner as shown in figure 
3.7, which is a schematic diagram of the thermocouple measurement configuration.  A 
detailed photograph of the configuration can be found in Appendix A.    
 61
 
Figure 3.7 Flat Flame Burner Thermocouple Configurations 
 
Since the thermocouples were subjected to very high temperatures, all 
measurements were corrected for radiation primarily based upon the recommendations of 
Shaddix66.  The effects of conduction through the wire and catalysis were neglected.  In 
addition, the temperature readings of interest were taken at what was considered to be a 
?steady-state?.  With these assumptions the energy equation for the system reduces to a 
balance of convection and radiation given by equation 3.2, which can be solved for the 
gas temperature in equation 3.3. 
 
? ? ? ?h T T T Tg tc tc tc? ? ? ?? ? 4 4          Eq. 3.2 
 
? ?T T T T
hg tc
tc tc? ? ? ?? ?
4 4
            Eq. 3.3 
 
where Tg, Ttc, T?, ?, and ? are the gas temperature, thermocouple surface temperature, 
ambient temperature, the thermocouple emissivity, and the Stefan Boltzman constant 
 62
respectively.  h, the convective heat transfer coefficient was determined from an 
empirical Nusselt number correlation given by Whitaker67.  The correlation was obtained 
for a sphere with a Prandtl Number range of 0.71 < Pr < 380, and Reynolds number range 
of 3.5 < Re < 76,000 where the Reynolds number is the standard Reynolds number for a 
sphere.  All gas properties are taken at Tg, except for ?s which is the gas viscosity taken at 
the thermocouple surface temperature.  Since Tg is unknown, this requires an iterative 
procedure.  Gas properties for air were used at the temperatures of interest and were taken 
from Incropera and Dewitt68.  For some properties, such as the thermal conductivity of 
air, k, the data from Incropera and Dewitt68 was plotted and a curve fit was applied to the 
data and used for data reduction in order to expedite data reduction.  Figure 3.8 is an 
example of such a curve fit.      
 
h Nukd?                     Eq. 3.4 
 
? ?Nu d d
s
? ? ? ??? ????2 0 4 0 0612 2 3 0 4. Re . Re Pr . ??           Eq. 3.5 
 
 63
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0 500 1000 1500 2000 2500
Th
er
ma
l C
on
du
cti
vit
y, 
k, 
[W
/(m
*K
)]
Temperature (K)
k [W/(m*K)] = 0.11662 + 7.203E-5 * T(K) -3.305E-8 * T(K)2 
+ 1.848E-11 * T(K)3 - 2.068E-15 *T(K)4 
 
Fig. 3.8 Curve Fit for Air Thermal Conductivity 
 
3.1.4 Chemiluminescence for Flat Flame Burning Experiments 
There are a number of sources of optical radiation that are emitted from a flame or 
combustion system that can be used to study specific combustion characteristics of a 
given system.  The source most directly connected to the combustion reactions, however, 
is chemiluminescence69. Chemiluminescence appears when a certain chemical reaction in 
a chain of reactions mainly produces some molecules in an electronically excited level70.  
These molecules undergo transitions from higher to lower energy states that result in 
fluorescent emissions at specific frequencies depending on the molecule that has been 
excited. Such excited molecules or atoms have radiative lives on the order from 10-8 to 
10-6 seconds69. The intensity of the resulting emissions is proportional to the production 
 64
rate of the excited state species. For this reason chemiluminescence has been used as a 
rough measure of reaction time and heat release rate previously69. 
Chemiluminescent images can be obtained at visual wavelengths with standard 
optics using an appropriate narrow band interference filter to isolate the wavelength of 
interest.  Should the wavelength of interest be outside of the visible range, special lenses 
may be required.  In the boron nanoparticle flat flame burner experiments BO2 
chemiluminescence was monitored by utilizing a narrow bandpass filter centered at 546 ? 
2 nm and an Intensified Charge Coupled Device (ICCD) camera.  The bandpass filter has 
a 10 nm full width half maximum value.   
The BO2 molecule is a reactive intermediary gas-phase species formed throughout 
boron particle ignition and combustion51.  According to Spalding72, the BO2 spectrum is 
readily observable during both the combustion and ignition stages of boron combustion.  
Gas phase modeling studies conducted by Yetter52 indicate that BO2(g) gets converted to 
B2O3(g) or HBO2(g) products quickly depending on the amount of water vapor present.  
Therefore, once the source of B/B2O3 was consumed, then so was the source of BO2(g) as 
well51.  As a result, monitoring the emission of the BO2 molecule was a reasonable 
approach towards measuring the reaction/burning time of boron.      
 
3.2 Airbreathing Combustor Experiments 
After determining the fundamental ignition and combustion behavior of the 
nanoparticles in the flat flame burner experiments, the next set of experiments in this 
study provided a demonstration of the ability or inability to extract the energy from 
metallized nanoparticles within reasonable residence times and realistic combustor 
 65
conditions (reactant/product species).  A combustor simulating a ramjet engine was 
designed and fabricated for this purpose.  The combustor utilized in this study was 
located on the reacting test stand at the Advanced Propulsion Research Laboratory at the 
University of Maryland.   
 
3.2.1 Overall Design Strategy 
 The overall goals of this research effort were described earlier, with a large 
component based on a demonstration of the use of metallic nanoparticles in an 
airbreathing combustor, under realistic combustor conditions, namely residence time.  
With this in mind, the overall design strategy for this portion of the study was to design 
and test a combustor capable of providing a wide array of diagnostic capability.  
Specifically, the combustor needed to provide optical access so that any number of 
optical or a variety of laser diagnostics could be employed.  In addition, the combustor 
needed to be capable of providing a means of using other diagnostics such as probes 
should they be desired.  All testing conducted in this study was with the combustor 
operating in a ramjet-simulating mode.  As will be discussed, the combustor provided 
means of operating in other modes such as a ducted rocket configuration.  Finally, this 
study called for a simple, yet reproducible means of delivering the particles into the 
combustor, which will be discussed in detail in a subsequent section.  All design 
drawings not presented here are included in Appendix A. 
 
 
 
 66
3.2.2 Flow Facilities 
 The airflow for the combustor experiments was provided by an Atlas Copco 
Compressor that can provide up to 358 cubic feet per minute at delivery pressures up to 
150 psig.  The compressor delivers clean dry air through a series of filters and dryers to 
the laboratory supply lines.  The air is then delivered to an individual test rig by a 2 inch 
diameter steel pipe. 
 
3.2.3 Particle Delivery System Design 
The particles were introduced into the combustor via a reverse cyclone seeder 
similar to the technique used by Glass73.  The seeder consists of three main parts, the 
main body, and two end caps.  One of the two end caps contained a port in which high 
pressure stainless steel tubing was attached to the seeder as a means of exhaust for the 
particle laden flow.   
The main body of the seeder was constructed from 304 Stainless Steel.    The 
inner diameter of the main body was 6.35 cm (2.5 in).  The overall length of the main 
body was 20.3 cm (8.0 in), although 2.54 cm (1 in.) on each side was left for insertion of 
the end caps.  The seeder was sealed with O-rings at the main body and end cap 
interfaces.  A stainless steel tube was welded to the outside of the main body near the top 
of the seeder.  This was the location of the gas entrance into the seeder.  The tube was 
welded such that the carrier gas enters the seeder tangentially with respect to the walls 
and promotes circulation within the seeder.  Similarly stainless steel tubing was used for 
the exhausting particle laden flow.  The seeder was designed to be intentionally heavy 
 67
walled in the event that the combustor experienced flashback into the seeder.  A 
schematic diagram of the seeder is provided in figure 3.9.   
The stainless steel exhaust tubing was located in the center and extended down 
into the seeder 4 inches (10.16 cm).  As the gas entered the top of the seeder it was forced 
down towards the bottom of the seeder where the particles were loaded.  Before the gas 
left the seeder it formed a ?cyclone? which entrained the particles into the gas flow.  A 
particle-laden flow was established and could be delivered to the combustor for testing.  
Of the available volume, the length available for the particles was 15.24 cm (6 in), 
however the seeder was never more than about 1/3 full of particles.  A photograph of the 
seeder components can be found in Appendix A. 
 
 
 
Fig. 3.9 Schematic Diagram of Reverse Cyclone Seeder 
 
A fuel was chosen as the carrier gas since the metallic particles are also fuels.  
This was done in order to limit the chance of flashback once the combustor was in 
operation.  An inert gas could have been selected, however by using a fuel, which is the 
Tangential Carrier 
Gas Inlet 
Central Exhaust Tube 
Metal Powder 
 68
same as the main fuel in the combustor, the combustor is still operating on air and fuel 
alone with no diluents.   
For actual combustor testing an O?Keefe Choke Orifice (0.120 inches, 0.305 cm) 
with a small portion (0.75 inch, 1.91 cm) of stainless steel tubing welded onto the end 
was used to connect the seeder to the combustor.  The particle-laden flow was not choked 
at this point, this configuration was simply used as a simple means of connecting the 
seeder to the combustor.  In addition the tubing ended at the center of the combustor, 
meaning the particles were injected at the center of the combustor and could mix freely.  
In early testing without the tubing, the particle-laden flow was unable to substantially 
penetrate the main airflow of the combustor.  The tubing therefore provided a means of 
obtaining meaningful data in combustor testing.  A photograph of the modified orifice 
can be found in Appendix A.  
    
3.2.4 Pilot Flame - Design  
 The main combustor was ignited by a separate ethylene/oxygen pilot flame.  The 
pilot flame contains two main components, a combustor constructed from 303 Stainless 
Steel and an exhaust nozzle constructed from graphite.  Figure 3.10 is a photograph of the 
two components together.  The nozzle was constructed of graphite as a separate piece so 
that multiple pilot flame injection schemes could be employed if so desired.  For instance 
the shape of the pilot injection could be changed very easily and at very little cost.  A 
schematic diagram of the nozzle used in this study is included in figure 3.11.   
The pilot flame combustor was heavy weight and contained an initial channel and 
a rearward-facing step that served as the primary mixing and combustion zone prior to 
 69
exhausting into the main combustor.  The details can be seen in figure 3.12 (all units are 
in inches unless otherwise noted).  The pilot flame was ignited by a standard reusable 
automotive spark plug, and had four ports available for injection of fuel(s) and 
oxidizer(s).  It was designed this way should the need arise for the entire airbreathing 
combustor to be utilized in a ducted rocket mode rather than a traditional ramjet.  For all 
testing in this study a nozzle with a slot shape was chosen.  The slot was 0.125 inches in 
width and 0.418 inches in length.  The pilot flame was designed to thread directly into the 
main combustor components at the aft end.  The exit nozzle then mated flush with the 
surface of the main combustor wall.  In addition a separate aft end enclosure was 
designed and fabricated in the event that any testing would need to be conducted on just 
the pilot flame. 
 
Fig. 3.10 Pilot Flame Components 
Spark Plug 
Oxygen Injection 
Location 
Ethylene Injection 
Location 
Nitrogen Purge 
Injection Location 
Additional Port 
Graphite Exhaust 
Nozzle  
Threads to 
Mate with 
Combustor 
 70
 
Fig. 3.11 Schematic Diagram of Pilot Flame Nozzle 
 
Gas Generator Body
Quantity (1 ea)Greg Young
Drawing: NB-01
Nano-Burner
inches Scale: 1 to 1Tol: ?0.005
303 S.S.
Material:
Approved By:
15-FEB-05  
Fig. 3.12 Schematic Diagram of Pilot Flame Combustor 
  
 71
As was mentioned earlier, the pilot flame operated on a mixture of ethylene and 
oxygen which was spark initiated, through a standard automotive spark plug.  In all main 
combustor testing, the pilot flame conditions were held constant in order to minimize the 
amount of variables affecting the nanoparticle laden flow combustion.  The pilot flame 
combustor had 6 total ports, 4 for gas flows and 2 for the spark plug depending on test 
selection.  For all of the tests in this study, the spark plug port at the head end of the pilot 
flame was used.  In addition to the oxygen and ethylene, nitrogen was used at the 
conclusion of each test as a purge gas in order to quench any remaining combustibles and 
to help cool the system.  
 
3.2.5 Main Combustor Design 
An ethylene/oxygen pilot flame was used to ignite the main combustor, which 
contained air and fuel with or without the addition of aluminum or boron powder.  The 
pilot flame was described in the previous section.  Prior to reaching the combustor, the 
airflow is choked by an orifice with a diameter of 1.00 cm (0.395 in) to ensure a steady 
mass flow rate for testing.  A Wilkerson Regulator controlled the flow and the static 
pressure upstream of the choke point was recorded with a Setra Model 206 Static 
Pressure Transducer.  The powder was injected approximately 4.3 cm (1.7 in) upstream 
of the pilot flame.  The powder was introduced through a particle seeder (described 
earlier) using gaseous ethylene as the carrier gas. In addition, ethylene was injected 2.1m 
(7ft) upstream of the combustor such that it was essentially premixed.  During typical 
combustor testing between 40-60% of the fuel by mass was premixed depending on the 
 72
test condition.  Lastly, an injection port for additional fuel was added just downstream of 
the step should the desire to investigate the effects of injection location arise.  
The primary combustor in the experiments was two-dimensional with a rearward-
facing step.  The combustor was constructed entirely from 303 stainless steel.  The 
combustor had a width of 4.45 cm (1.75 in) and prior to the step, a height, H, of 3.81 cm 
(1.5 in).  The step, which was one half of a channel height, was located two channel 
heights downstream of the pilot flame.  The combustor was designed to provide optical 
access via quartz windows from three sides.  Figure 3.13 provides a schematic diagram of 
the combustor set up. 
All gas flows besides the air came from pressurized bottle sources and were 
choked with O?Keefe orifices and the static pressures upstream of the choke points were 
recorded by Setra Model 206 Static Pressure transducers.  The gas flows were carefully 
controlled by a Lab View Virtual Interface (VI), which controlled a series of solenoid 
valves, which dictated the timing of the test according to the user?s specifications.  This 
control allowed for repeatable and consistent testing.  Following each test, the pilot flame 
was purged with gaseous nitrogen (N2), to quench any residual fuels left in the pilot flame 
combustor.  The VI also recorded the static pressure settings of the gas flows as well as 
the thermocouple data obtained for each test for post processing.  The typical duration of 
a combustor test was three to four seconds.     
At the exit plane of the combustor (41.3 cm (16.25 in) from the pilot flame 
location) there were four ports, which were used in this study to provide access for 
thermocouple measurements.  Two different tops were designed and fabricated, one, 
which had a slot for a quartz window for optical access and the other with an array of 
 73
ports, which could be used for any number of diagnostics.  In this study only the top with 
the quartz window was utilized.  High temperature rubber gasket was used to seal most 
surfaces.  In some cases O-rings were used instead of the gasket in order to maintain a 
gas seal.   
Because the air delivery system (described earlier) delivered air from a round 
pipe, a special adapter was designed to convert the flow from a round pipe to a 
rectangular cross section.  Within this adapter, a honeycomb was placed and used as a 
flow straightener.   The adapter is made up of two components plus the honeycomb, 
which can be seen in a photograph in Appendix A.  Figure 3.14 is a photograph of the 
assembled combustor on the test stand.  All design drawings are supplied in Appendix A. 
 74
 
Fig. 3.13 Schematic Diagram of Combustor Set Up 
 
 
Fig. 3.14 Assembled Combustor on Test Stand 
  
 75
3.3 Combustor Diagnostics 
 As was previously mentioned the combustor was designed such that many 
different diagnostics could be applied to it very easily.  It was optically accessible 
through quartz windows from three sides so that any number of laser diagnostics can be 
relatively easy to apply.  In addition there was an array of ports at the exit plane of the 
combustor, which can be used in a variety of ways for any sort of probe measurement.  
Similarly, a second ?top? portion of the combustor was designed and fabricated with an 
array of ports for probe measurements.  The subsequent sections outline the diagnostics 
applied to this study, however, the combustor has far greater capabilities depending on 
the goal of a particular study. 
 
3.3.1 Flow Control for Airbreathing Combustor 
 Aside from the main airflow, all of the gases were supplied by high-pressure 
bottles.  The flow rates for each individual gas were controlled through the use of a 
combination of regulators, Setra Model 206 Static Pressure Transducers, and O?Keefe 
High Presicions Choke Orifices.   
Regulators were connected to the gas bottles and the gases were transported to the 
test section by high-pressure stainless steel tubing.  A solenoid valve controlled by a Lab 
View VI was placed in each gas line for accurate sequencing during testing.  After the 
solenoid valve a check valve was placed in each line for safety purposes.    The Setra 
Pressure Transducers were placed after the check valves just prior to injection into the 
combustor system.  Finally the O?Keefe choke orifices were placed right at the combustor 
injection points.  The Lab View VI also recorded the static pressure in the gas lines 
 76
during combustor testing.  With the knowledge of the orifice dimensions, and the static 
pressures, mass flow rates were obtained for a given test through isentropic flow 
relationships.      
 
3.3.2 Mie Scattering 
Mie scattering can be used as a tool for flow visualization in many applications.  
In general a laser is used to create a beam or planar sheet in given area of interest.  The 
light from the laser is scattered by seed particles, and the scattered light is collected, 
usually by a camera, in order to provide information about the flow field.  Mie scattering 
differs from other techniques such as Raleigh scattering in that larger seed particles are 
employed.  This results in much higher light scattering intensities, which increases the 
signal to noise ratio and eliminates the need for a light intensifier in the receiving optics.  
As a result, the implementation of this technique is easier and the cost is lower74.    
 A series of Planar Mie Scattering experiments were conducted during this study in 
order to ensure that the dispersion of the particles was uniform.  An Nd:Yag laser 
operating at about 12 hz and 532 nm wavelength was synchronized with a FASTCAM-
Ultima1024 model 16K high speed camera high speed camera to capture the images.  The 
laser sheet was passed through the window that was located on the top of the combustor 
with the aid of an ?optical arm?.  The images were then recorded looking directly through 
the window located on the side of the combustor, such that the laser sheet and camera 
were perpendicular to each other.   
The Mie Scattering experiments were conducted in the fully reacting flow using 
ALEX nanoaluminum particles as the scattering source.  Since the images were taken in a 
 77
reacting flow, a narrow bandpass filter centered at 532 ? 2 nm was used to filter out all of 
the excess radiation coming from the reacting flow.  The filter only allowed light with 
wavelengths near 532 nm to be viewed by the camera, such that it minimized the excess 
radiation from the reacting flow, and the main source of light captured was the laser light 
being scattered by the particles in the flow field.  Figure 3.15 is a schematic diagram of 
the Mie Scattering experiment. 
PC
Nd:YAG Laser
Synchronizer
High
Speed
Camera
Optical Arm
Laser Sheet
Combustor Test Section
Window for Optical Access
 
Fig. 3.15 Schematic Diagram of Combustor Mie Scattering Experiment 
 
3.3.3 Chemiluminescence 
In the airbreathing combustor portion of this study, the oxidation of the particles 
was studied through the natural chemiluminescence of selected particle oxides, AlO and 
BO2 for aluminum and boron respectively.  AlO is a known gas phase intermediate in the 
oxidation of aluminum43.  The BO2 molecule is a reactive intermediary gas-phase species 
 78
formed throughout boron particle ignition and combustion51.  In both cases images were 
captured using a Cooke Dicam Pro Intensified CCD camera (ICCD) with the appropriate 
narrowband interference filters in place.  The camera was placed so that it was imaging 
through the side windows, similar to the position of the high speed camera in the Mie 
Scattering experiment.   
The filters were centered at 488 ? 2 nm for AlO and 546 ? 2 nm for BO2.  Both 
filters had full width half maximum values of 10 nm.  Since the primary fuel was a 
hydrocarbon (Ethylene), the potential for hydrocarbon emission overlapping the BO2 
band existed.  In particular, C2 exhibits several bands in the vicinity of the filter.  
However, according to Gaydon75, the C2 bands near 546 nm are characterized as weak or 
very weak.   
Based upon information from Gaydon75, there are three possible AlO bands that 
could be detected with the 488 nm filter.  He showed that the band at 484.23 nm has very 
strong emission characteristics, the band at 486.64 nm has strong characteristics, and the 
band at 488.88 nm has medium strength emission characteristics.  In the wavelengths 
allowed by the 488 nm band pass filter there were some bands of emission due to 
hydrocarbon combustion as well.  In particular, C2*, CH*, and CO2 exhibit bands that 
overlap the filter.  However, Gaydon75 lists the C2* and CH* emission as either weak or 
very weak near 488 nm.   
 
3.3.4 Temperature Measurements 
 Temperature measurements were made at the exit plane of the airbreathing 
combustor using a combination of 1000 ?m K-type (Chromel/Alumel) and 200 ?m B-
 79
type (Pt-6%Rh/Pt-30%Rh) thermocouples.  The B-type thermocouples were described in 
section 3.1.3.  K-type thermocouples are suitable for use in clean oxidizing and inert 
environments, and have a temperature range of ?200 ? 1250 oC.  The typical error 
associated with a K-type thermocouple temperature measurement is the larger of 2.2 oC 
or 0.75%.   
Starting from the bottom of the combustor, the first thermocouple was located at 
0.318 cm (1/8 in) and they were spaced 1.693 cm (2/3 in) apart.  The K-type 
thermocouples were placed near the combustor walls, and two B-type thermocouples 
were located in the bulk flow of the combustor. 
 The temperature measurements were used to provide combustor temperature 
profiles and also to provide a means of comparing the thermal output of the different 
fuels.  Once again the thermocouple temperature measurements were corrected for 
radiation losses using the procedure described in section 3.1.3.  
 80
CHAPTER 4 
 
BASIC CHARACTERIZATION STUDIES 
 
4. Basic Characterization Studies 
 Within this study there were a number of intermediate steps that needed to be 
taken.  First was to understand some of the basic properties of the materials under 
investigation.  Next, since a new experiment was developed (flat flame burning with 
horizontal particle injection), some basic characterization of the individual components of 
the experiment was required.  Finally, with a newly designed combustor, its behavior 
under well-defined conditions needed to be understood.  This chapter discusses these 
basic characterization studies that needed to take place prior to obtaining the results that 
were the main focus of the study.  It is broken into three main sections, particle 
characterization, flat flame burner experiment characterization, and airbreathing 
combustor characterization. 
 
4.1 Particle Characterization 
This study focuses on two different types of nanoparticles.  The first is designated 
as 50 nm ALEX and was purchased from the Argonide Corporation.  ALEX is a trade 
name for aluminum particles formed from the explosion of aluminum wires76.  The 
second particles type to be investigated was SB99.  SB99 is a boron nanoparticle that was 
supplied by the Naval Surface Warfare Center ? Indian Head Division who purchased the 
 81
material from the SB Boron Corporation.  The technique used for manufacturing the 
particles is unknown to the author.    
 
4.1.1 Thermal Gravimetric Analysis (TGA) 
A Thermal Gravimetric Analysis was performed to investigate the reactivity of 
the nanoparticles involved in this study.  TGA is a thermal analysis technique used to 
measure changes in the mass of a sample as a function of temperature and/or time77.  This 
technique is commonly used to determine a samples? inorganic material content, 
degradation temperatures, moisture, etc.  In short, to perform a TGA analysis, a sample 
was tared in a sample pan attached to an ultra-senistive microbalance apparatus inside of 
a high temperature furnace77.  The temperature of the furnace was gradually increased at 
a specified rate.  Simultaneously, the mass addition or loss was continually monitored 
until the desired upper temperature limit was achieved.  The environments are commonly 
argon, helium, nitrogen, air, or even pure oxygen. 
 
4.1.1.1 Determination of Particle Active Content by TGA 
The amount of mass gained due to the oxidation process of the particles was 
found by performing a TGA analysis in a 50/50 oxygen/argon environment.  This allows 
one to deduce the active content of particle.  Active content is defined as the amount of 
pure material in a given particle.  This analysis assumes that the samples undergo 
complete oxidation; therefore the mass added to the sample during oxidation is 
exclusively a result of oxygen reacting with the particles.  The stoichiometric ratio and 
the amount of oxygen gained during oxidation was used to determine the amount of 
 82
active content in the particles.  The mass of the oxygen due to oxidation can be expressed 
as: 
   m m m mox f i? ? ??  
where 
   mi: initial mass of the sample 
   mf: final mass of the sample 
For the complete oxidation of aluminum in an oxygen environment, the 
stoichiometric equation can be written as: 
   4Al + 3O2 ? 2Al2O3 
therefore, 
  moles of Al = 4/3 moles of O2 
Converting moles of aluminum and moles of oxygen to their respective masses, we get 
   mMW mMWAl
Al
O
O
? 43 2
2
 
Finally rearranging and solving for the mass of aluminum we obtain: 
   m MWMW mAl Al
O
O?
4
3
2
2
 
Calculating the mass of aluminum based on the mass of oxygen due to full oxidation, we 
have 
   m mAl O? 1124
2
.  
As an example, using the results obtained from the 50 nm ALEX particles, the percentage 
mass increase, ?i, due to full oxidation of aluminum is 81.95% determined by subtracting 
the baseline value from the peak value as shown in Figure 4.1.   
 83
Figure 4.2 shows the mass gain plot for the SB99 material.  The slight dip below 
100-weight% in the low temperature regime is due to the evaporation of moisture that 
was collected in the powder.  A similar analysis to the one performed for the ALEX 
particles reveals that the SB99 material was 72.2% pure boron.  The remaining volume of 
each material was assumed to be the metal oxides Al2O3 and B2O3 for each material. 
100
120
140
160
180
0 200 400 600 800 1000 1200
W
eig
ht 
%
Temperature [oC]
173%
100%
HR=5oC/min in
50% O2/50% Argon Environment
Tmelt = 660 oC
 
Fig. 4.1 TGA of 50nm ALEX 
 
 84
120
160
200
240
280
0 200 400 600 800 1000 1200
W
eig
ht 
%
Temperature [oC]
257%
97%
HR = 5oC/min in 
50% O2/50% Argon Environment 
 
Fig. 4.2 TGA of SB99 
 
4.1.2 Scanning Electron Microscope Imaging (SEM) 
Two major uses of a Scanning Electron Microscope (SEM) are to produce high-
resolution images and to provide a visual record of the surface of sub-micron materials78.  
The resolution of a typical SEM is about 0.01 ?m79.  Its great depth of field and the 
accurate rendition of surface features make the SEM well suited to particle sizing and 
morphology studies79.  However, when imaging with an SEM, one needs to take 
precautions against the build up of electric charge on the sample surface, which can 
deflect the electron beam and cause distortion of the image.  Generally this is 
accomplished by applying a very thin coating of gold or carbon.     
 85
Figures 4.3 and 4.4 are SEM Images of the ALEX and SB99 particles 
respectively.  In order to achieve an image with reasonable resolution, the particles were 
coated with a 60/40 gold/palladium coating.  The coating was necessary to eliminate 
charging of the particles since the materials were conducting.  The coating adds 
approximately 7.5 nm radially to the particles.  The images show that both sets of 
particles are highly spherical in nature.   
The manufacturer specifies the ALEX particles as 50 nm, but Figure 4.3 shows 
that although there were some ALEX particles that were approximately 50 nm in size, the 
overall size distribution was quite wide with some particles close 200 nm.   
Figure 4.4 shows that the SB99 particles were highly agglomerated, but majority 
of the primary particles were near 50 nm in diameter or less.  Reference 80 suggests that 
the average particle size of the SB99 particles is 62 nm.  The images also show that both 
particle types suffer from some agglomeration.  That is, the particles were stuck together 
to create larger particles.  This is important because this is the likely condition that the 
particles enter the airbreathing combustor and not in their primary particle size.   
 86
 
Fig. 4.3 SEM of ?50 nm? ALEX 
 
Fig. 4.4 SEM of SB99 Boron Particles 
 
  
 87
4.2. Flat Flame Burner and Aerosol Injection Characterization 
 The next portion of this study involved characterizing the McKenna Flat Flame 
Burner at specified test conditions and the aerosol injection system for the particle 
ignition and burn time experiments. 
  
4.2.1 Velocity Measurements 
 A total of 3 different injection velocities were measured corresponding to three 
different volumetric flow rates controlled by a rotameter.  All measurements were made 
in ?cold flow?, i.e. non-reacting over a wide range of crossflows, and momentum flux 
ratios.  The goal of the experiment was to cover the range of momentum flux ratios that 
would be covered in reacting flow, such that the force acting on the jet by the crossflow 
would be similar.  The definition of momentum flux ratio is given by equation 4.1.   
     J J jUU?
? ?
?
?
2
2      Eq. 4.1 
where the ? and U represent density and velocity respectively, and the subscripts j and ? 
represent the jet and crossflow respectively. 
Table 4.1 provides a summary of the PIV test conditions.  As the table indicates, 
for all of the injection conditions, the jet remained below the critical Reynolds Number 
(2300) for transition to turbulence, indicating that at the jet exit the flow was laminar for 
all test conditions.  The Reynolds number was based on the injection tube inner diameter 
(1 mm).  The volumetric flow rates were specified by rotameters for all gases, and the 
average velocities are based upon the results of the one dimensional continuity equation, 
based on the cross-sectional area of the injection tube.  The details of the test set up are 
provided in Chapter 3. 
 88
Table 4.1 Cold Flow PIV Experiment Test Matrix 
Burner (Crossflow) Properties Jet Properties  
Volumetric 
Flow Rate 
(L/min) 
Average 
Velocity (m/s) 
Volumetric 
Flow Rate 
(L/min) 
Average 
Velocity (m/s) 
Reynolds 
Number 
Momentum 
Flux Ratio 
2.60 0.0153 0.61, 0.84, 1.15 12.89, 17.93, 24.30 811, 1128, 1529 842, 1171, 1587 
6.63 0.0391 0.61, 0.84, 1.15 12.89, 17.93, 24.30 811, 1128, 1529 330, 458, 621 
10.67 0.0629 0.61, 0.84, 1.15 12.89, 17.93, 24.30 811, 1128, 1529 205, 285, 386 
14.71 0.0867 0.61, 0.84, 1.15 12.89, 17.93, 24.30 811, 1128, 1529 149, 207, 280 
18.74 0.1105 0.61, 0.84, 1.15 12.89, 17.93, 24.30 811, 1128, 1529 117, 162, 220 
22.78 0.1343 0.61, 0.84, 1.15 12.89, 17.93, 24.30 811, 1128, 1529 96, 134, 181 
26.82 0.1581 0.61, 0.84, 1.15 12.89, 17.93, 24.30 811, 1128, 1529 82, 113, 154 
30.85 0.1819 0.61, 0.84, 1.15 12.89, 17.93, 24.30 811, 1128, 1529 71, 99, 134 
34.89 0.2057 0.61, 0.84, 1.15 12.89, 17.93, 24.30 811, 1128, 1529 63, 87, 118 
38.93 0.2295 0.61, 0.84, 1.15 12.89, 17.93, 24.30 811, 1128, 1529 56, 78, 106 
42.96 0.2533 0.61, 0.84, 1.15 12.89, 17.93, 24.30 811, 1128, 1529 51, 71, 96 
47.00 0.2771 0.61, 0.84, 1.15 12.89, 17.93, 24.30 811, 1128, 1529 47, 65, 88 
 
  
Figure 4.5 is an example of an image pair collected during an experiment.  From a 
total of 50 image pairs collected for a given experiment, velocity vectors were obtained 
and an example can be seen in Figure 4.6.  The image in Figure 4.6 corresponds to the 
case of highest crossflow (47 L/min) and the intermediate injection flow rate (0.84 
L/min).  Contours of the axial-component of velocity are given for all three injection flow 
rates at the highest crossflow condition in figures 4.7-4.9.  Qualitative comparison of 
figures 4.7-4.9 shows that as the injection flow rate is decreased the jet is displaced more 
by the crossflow as would be expected.    
 
 89
 
Fig. 4.5 Example of Image Pair 
X (mm)
Y
(m
m)
0 10 20 30 40-5
0
5
10
15
Jet Flow Rate = 0.84 L/min
Cross Flow Flow Rate = 47 L/min
 
Fig. 4.6 Velocity Vectors Obtained from PIV Experiments 
 
Injection Tube 
Burner Surface 
Water Droplets 
 90
 
Fig. 4.7 Axial-Component of Velocity Contour for Highest Injection Flow 
Rate and Crossflow 
 
Fig. 4.8 Axial-Component of Velocity Contour for Intermediate Injection 
Flow Rate and Highest Crossflow 
 91
 
Fig. 4.9 Axial-Component of Velocity Contour for Lowest Injection Flow Rate and 
Highest Crossflow 
 
Using the commercially available Tecplot computer software, the centerline axial-
component of velocities were extracted for use in determining the burning times and 
ignition delays of the boron nanoparticles.   Figure 4.10 shows a collection of a wide 
range of the data taken in the intermediate jet flow rate case.  The numbers in the legend 
correspond to increasing crossflow, for instance 1a was the lowest crossflow condition 
(2.60 L/min), and 12a was the highest crossflow condition (47 L/min).  As figure 4.10 
indicates, the axial-component of the centerline velocity was not greatly influenced by 
the crossflow over the test conditions.  For the actual testing of the boron nanoparticles in 
the reacting flow, the intermediate injection flow rate was selected. 
 92
 
Fig. 4.10 Axial-Component of Centerline Velocity for Intermediate Injection 
Flow Rate and Several Crossflows 
 
As figure 4.10 shows, for a given injection velocity, the axial-component of the 
centerline velocity was relatively unaffected by the change in crossflow.  Applying the 
analysis described in Section 2.6 of this thesis of a nonreacting-laminar jet of fluid 
flowing into a quiescent atmosphere can reasonably explain the results of the PIV 
experiments.   
 In this study, the jet actually exhausted into a crossflow, however, a correlation 
(Eq. 4.2) similar to that of equation 2.24 still represented the data well after a spatial 
duration of approximately 6 mm, particularly for the two higher injection flow rates. 
        ? ?u m s a u xx CL e b, ( / ) * *? 2                Eq. 4.2 
 93
 In this case the terms involving viscosity and the tube radius have been combined 
into a single constant, a.  ue represents the average of the maximum velocities measured 
for each crossflow for a given injection flow rate.  a, and b where solved for by applying 
a curve fit in the form of equation 4.2 in the commercially available computer software 
Kaleidagraph, the results of which can be seen in figures 4.11-4.13.  The red line in the 
plots is the curve fit itself, which follows the data extremely well, especially for the two 
higher injection flow rate conditions. In general, the value of the exponent b, is in good 
agreement with the theory of a laminar jet exhausting into a quiescent atmosphere at 
these levels of crossflow.  The exponent, b, ranges from -0.82 to -1.12 compared to ?1 
which is dictated by theory.  Finally since measurements very close to the injection 
source were not very reliable, the maximum velocity was taken to represent the core 
centerline velocity up until it reached six millimeters at which point the correlation 
represents the velocity well.  Therefore, when using the measurements to determine 
burning time and ignition time data, a piecewise function for velocity is used which is 
given in figures 4.11-4.13.  During all of the particle testing, the intermediate jet flow rate 
was selected (0.84 L/m), therefore the velocity was represented by equation 4.3. 
   ? ?? ?
0 6 32 79
6 0 208 32 79 2 1 11
? ? ? ?
? ? ? ?
x mm u m s
x mm u x
x cl
x cl
,
,
.
. /
. .    Eq. 4.3 
 
 94
 
Fig. 4.11 Centerline Axial-Component of Velocity of Lowest Jet Flow Rate (0.61 
L/m) 
 
 
 
 
Fig. 4.12 Centerline Axial-Component of Velocity of Intermediate Jet Flow Rate 
(0.84 L/m) 
 95
 
Fig. 4.13 Centerline Axial-Component of Velocity of Highest Jet Flow Rate (1.15 
L/m) 
 
 With an expression for the particle velocity now known, and the spatial locations 
(x1 (start of burning) and x2 (end of burning)) of the particle burning obtained from the 
ICCD camera chemiluminescence images, the ignition and burning times were calculated 
using the following procedure. 
? ? ? ?t t x t xb ? ?2 1                Eq. 4.4 
where tb was the burning time, t1 and t2 were the times associated with reaching the spatial 
locations x1 and x2.  The times, t1 and t2 are determined utilizing the correlation developed 
and shown in equation 4.3.  
 As an example, the location where t2 took place, i.e. x2, was found by a technique 
described in Chapter 5.  Since x2 was now known, and the velocity at a given spatial 
 96
location was known from figure 4.12 the time it for the particles to reach x2 was 
determined by:     
     t t dti i i
i
i
? ??? 1                Eq. 4.5 
where the subscript i refers to a specific location and the subscript i-1 refers to its nearest 
neighbor for which spatial velocity data was available.  The incremental time dti was 
found by:  
     dt dxUi i
avei
?                 Eq. 4.6 
where dxi is the incremental spatial resolution of the velocity data and Uave,i was 
determined by: 
     U U Uave i i
i
? ? ?12                   Eq. 4.7 
where once again, the subscripts i and i-1 refer to the particular data point (1 or 2) and its 
nearest neighbor for which spatial velocity data was available. 
  
4.2.2 Temperature Measurements 
 As was outlined in Chapter 3, temperature measurements of the flat flame burner 
were taken using a 200 ?m B-type thermocouple at three locations.  After correction for 
radiation loss, the average of the three measurements was taken and used for data 
analysis purposes.  Table 4.2 gives the test matrix for the flat flame burner experiments as 
well as the results of the temperature measurements.  As the table indicates the measured 
temperatures are well below the adiabatic flame temperatures at all conditions.  This was 
mainly due to heat loss to the burner through the cooling water.  In fact other 
 97
researchers81,82 have demonstrated that the burner is relatively insensitive to equivalence 
ratio.  Instead temperature is driven more by fuel flow rate (i.e. thermal output).   
 
Table 4.2: Flat Flame Burner Test Matrix 
Volumetric Flow Rates (L/min) Product Mole Fractions Test Condition 
O2 Air CH4 O2 H2O CO2 N2 
Average Burner 
Temperature (K) 
1 16.8 30.2 5.8 0.200 0.234 0.117 0.450 1684 
2 21.0 29.8 7.3 0.200 0.266 0.133 0.400 1808 
3 24.8 27.5 8.7 0.200 0.300 0.150 0.350 1854 
5 28.6 24.5 10.1 0.200 0.234 0.117 0.450 1637 
6 23.2 30.2 5.6 0.300 0.200 0.100 0.400 1630 
7 28.9 29.8 7.3 0.300 0.233 0.117 0.350 1797 
8 33.2 26.8 8.8 0.300 0.267 0.133 0.300 1872 
9 19.2 24.9 4.6 0.300 0.200 0.100 0.400 1578 
10 14.0 26.5 5.8 0.150 0.267 0.133 0.450 1718 
11 17.7 33.4 7.3 0.150 0.267 0.133 0.450 1814 
13 11.0 20.7 4.5 0.150 0.267 0.134 0.449 1596 
14 8.5 28.9 4.5 0.100 0.233 0.117 0.550 1614 
15 11.0 28.3 5.6 0.100 0.267 0.134 0.499 1712 
16 13.7 35.3 6.9 0.100 0.267 0.134 0.499 1791 
 
4.2.3 Particle Size Measurements 
 Although the primary particle size of the nanosized boron used in the flat flame 
burner experiments was near 50 nm, agglomeration of the particles was inevitable even 
with the sonicating atomizer.  In order to understand the size distribution of the 
agglomerated particles two separate measurements were taken and described in the 
subsequent sections. 
   
4.2.3.1 TEM 
 A Transmission Electron Microscope (TEM) was used as a visual aid to 
understand the agglomerate size.  The limit of resolution for a typical TEM is less than 
 98
0.001 ?m79.  In a TEM electrons are generated by thermoioninc emission from a heated 
tungsten filament and are focused by magnetic lenses.  Particles in the electron beam 
absorb and scatter electrons to produce a two-dimensional silhouette image.  Samples 
must be deposited on a special grid for imaging.  Typically the grids are 200 mesh 
electrodeposited screen with a thin film of carbon or Parlodion (1-4% nitrocellulose in 
amyl acetate covering the screen openings79.  The film is sufficiently thin compared to 
the particles such that it causes only slight attenuation of the electron beam, whereas the 
samples (particles) scatter and absorb the electron beam significantly to form a high 
contrast image.    
A small section of tubing was attached to the end of the aerosol injection system, 
which led to a TEM grid.  The aerosol was allowed to flow for several minutes to ensure 
sufficient coverage of the grid.  Once samples were applied to the grid, images using the 
TEM were taken and can be seen in Figures 4.14 and 4.15.  The images show that the 
particles are in fact large non-spherical agglomerates of small primary particles.  The 
agglomerates in the images are near 200 nm in size. 
 
 99
 
Fig. 4.14 TEM Image of Aerosolized SB99 
 
Fig. 4.15 Close-up TEM Image of Aerosolized SB99 
 
4.2.3.2 DMA 
 As a second and more quantitative measure of agglomerate size a Differential 
Mobility Analyzer (DMA) was employed.  DMA?s are commonly used to measure 
 100
particle size distribution with high resolution in the submicrometer range79.  Figure 4.16 
is a schematic diagram of a typical DMA set up.  The DMA takes advantage of certain 
properties of a charged particle.  Essentially when a charged particle is placed in an 
electric field it is acted on by a force, FE.  The force on the particles results in the 
particles traveling at the terminal electrostatic velocity, VTE.  For particles in the Stokes 
region, VTE is obtained by equating (Eq. 4.8) the electrostatic force to the Stokes Drag 
(Eq. 4.9) and solving for the velocity (Eq. 4.11). 
     neE VdC
c
? 3??                Eq. 4.8 
     F VdCD
c
? 3??                   Eq. 4.9 
where ? is the fluid viscosity, V is the velocity of the particle, d is the particle diameter, n 
is the total number of elementary units of charge on the particle, e is the charge of an 
electron, E is the electric field strength, and Cc is the Cunningham Correction factor 
given by: 
     C dc ? ?1 2 52. ?              Eq. 4.10 
where ? is the mean free path. 
       V neECdTE c? 3??              Eq. 4.11 
One can relate a particles ability to move in an electric field in terms of the particle?s 
electrical mobility, Z, the velocity of the particle with a charge ne in an electric field of 
unit strength.  Electrical mobility is defined by: 
Z VE neCdTE c? ? 3??                Eq. 4.12 
 101
In a DMA, an aerosol is passed through a neutralizer to the Boltzmann 
equilibrium charge distribution before entering the electrostatic classifier section.  A 
laminar flow of clean air is surrounded by a thin annular layer of aerosol as the two fluids 
travel axially between the central rod and the coaxial tube.  The voltage is controlled on 
the central rod, and the tube is grounded.  Near the bottom of the central rod is a gap that 
allows a small airflow to leave the classifier and exit through a central tube.  Only 
particles with a narrow range of mobilities can enter the gap.  Particles with larger 
mobilities reach the central rod before entering the gap, while those with lower mobilities 
go beyond the gap and are filtered out.  Therefore, the exiting aerosol is nearly 
monodisperse.  By adjusting the input voltage over a wide range and connecting the 
DMA to a Condensation Nuclei Counter (CNC) a size distribution can be obtained for a 
given aerosol. 
 102
CNC and
OUtput
High Voltage
Power Supply
Central Tube
Outer Tube
Charge
Neutralizer
Impactor
Clean Air In
Excess Air Out
Aerosol In
Exit Slot
 
Fig. 4.16 Schematic Diagram of a Typical DMA Set Up77 
  
Similar to the TEM sampling a tube was attached to the end of the aerosol 
injection system and the aerosol was sent to a DMA for analysis.  Figure 4.17 shows the 
size distribution given by the DMA in terms of a mobility diameter.  The mobility 
diameter is defined as the diameter of a sphere with the same friction coefficicent as an 
agglomerate under similar dynamic conditions83.  As figure 4.17 shows, the size 
 103
distribution is fairly wide, ranging from about 20 nm to about 400 nm with a mean of 200 
nm. 
0
5000
1 104
1.5 104
2 104
2.5 104
3 104
3.5 104
4 104
11.3 15.7 21.7 30 41.4 57.3 79.1 109 151 209 289 400
Pa
rtic
le 
Co
un
t
Mobility Diameter (nm)
Median = 183 nm
Mean = 200 nm
Geometric Mean = 165 nm
Mode = 209 nm
 
Fig. 4.17 Agglomerate Size Distribution by DMA 
 
4.3 Airbreathing Combustor Characterization 
 Prior to combustor operation with metallized fuels, a number of experiments were 
conducted to understand the operation of individual components as well as how the 
combustor behaved with a standard baseline fuel.  The subsequent sections outline these 
experiments. 
  
 
 
 104
4.3.1 Gas Flow Calculation 
 In section 3.3.1, the details of how the gas flows were controlled were outlined.  
Since the static pressure data was recorded for all of the gas flows, an accurate estimate 
of mass flow rates for each individual experiment was possible.  For each experiment, the 
data from the last second of combustor operation was averaged to give the static pressure 
used in flow rate calculations.  In all cases, the measured static pressure was always 
sufficient to achieve a choked flow condition.  Depending on the gas, this is roughly at 
least twice atmospheric pressure.  All gas mass flow rates were calculated from equation 
4.13. 
? ? ? ?
?
*
m p ARTo
o
? ???? ???
?
?
? ?
?
?2
1
1
2 1
             Eq. 4.13 
 Where, po, A*, R, To, and ? refer to the stagnation pressure, the throat area of the orifice, 
the gas constant, the stagnation temperature, and the ratio of specific heats respectively. 
 
4.3.2 Particle Seeder Characterization 
In order to determine the mass flow rate of particles a series of experiments were 
under taken to characterize the particle seeder.  Initially, 3M Glass Bubbles, with an 
average size of 50 ?m were used for debugging the system.  The carrier gas used in all of 
the experiments was ethylene and was controlled by a LabView Virtual Interface (VI), a 
solenoid valve, and Setra Model 206 Static Pressure Transducer, and an O?Keefe choked 
flow orifice (0.040 inches, 0.102 cm). 
After a carrier gas flow rate was selected, tests were run for different lengths of 
times and the particles were captured in a large bag and weighed on a high precision scale 
 105
between each test.  In addition, the amount of particles remaining in the seeder was also 
weighed as a second measurement.  The amount of particles placed into the seeder prior 
to testing was measured and kept constant to within ?0.03 grams because it was found 
that the seeding rate was very sensitive to the initial amount of particles in the seeder.  
Specifically, the seeding rate was very sensitive to the distance between the central 
exhaust tube and the particle interface.  In the cases of the aluminum and boron testing, 
25 grams of material was put into the seeder for testing.  Typically, three tests per time 
duration were conducted in order to determine the repeatability of the seeder.  After 
debugging, tests proceeded with ALEX/SB99 particles.   
Figure 4.18 shows the results from four different test series; the error bars in the 
figure represent the measurement error for a particular test.  The standard deviation for a 
particular duration ranges from about 0.03 grams to 0.14 grams.  As figure 4.18 shows, 
for the duration of these tests, the total mass collected varies linearly with time.  It was 
not expected that the linear trend would continue for long durations.  In fact it was found 
that after some time the seeding rate approaches zero.  During combustor testing, the 
slope of these lines was used as the mass flow rate of ALEX/SB99 particles.  Following 
each combustor test, the particles remaining in the seeder were weighed to ensure that the 
total amount used was consistent with the characterization testing.  Finally, the seeder 
was completely cleaned out between each test in order to ensure that each test was 
occurring under the same conditions.  
In addition, since there was a significant amount of free volume in the particle 
seeder several experiments were conducted in order to determine the amount of time for 
the seeder to achieve a steady state gas flow rate.  In this experiment, a Setra Model 206 
 106
Static Pressure transducer was placed just downstream of the seeder to determine when 
the gas flow after the seeder could be considered constant.  An aluminum rod occupying 
approximately the same volume as the metallic nanoparticles was placed inside the seeder 
to simulate the real test condition free volume.  Figure 4.19 shows the pressure traces of 
the gas flows during a combustor test for several of the gases used in the test as well as 
the pilot flame.  Comparing the green line (downstream) to the purple line (upstream) it 
can be seen that after approximately 1.5 seconds the gas reached a steady state for this 
particular test condition.  At 1.0 seconds, the gas reached about 75% of its steady state 
pressure.   
0
0.5
1
1.5
2
2.5
3
3.5
0 1 2 3 4 5 6 7 8
ALEX Carrier Gas 1 g/s
ALEX Carrier Gas 1.5 g/s
SB99 Carrier Gas 1 g/s
3M Glass Bubbles carrier Gas 0.5 g/s
m(t) = 0.44 + 0.39t   R2= 0.99 
m(t) = -0.15 + 0.87t   R2= 0.98
m(t) = -0.61 + 0.81t   R2= 0.96 
m(t) = 0.13 + 0.50t   R2= 0.96 
Ma
ss
 (g
)
Time (s)  
Fig. 4.18 Summary of Seeder Calibration Testing 
 107
0
20
40
60
80
100
120
0 1 2 3 4 5 6 7 8
Oxygen
Pilot Flame Fuel
Seeder Fuel (Downstream)
Seeder Fuel (Upstream)
Pilot FlameP
re
ss
ur
e (
ps
ig)
Time (sec)  
Fig. 4.19 Seeder Carrier Gas Filling Time 
 
4.3.3 Pilot Flame Plume Measurements 
As was mentioned earlier, the pilot flame operated on a mixture of ethylene and 
oxygen which was spark initiated, through a standard automotive spark plug.  In all main 
combustor testing, the pilot flame conditions were held constant.  In addition to the 
oxygen and ethylene, nitrogen was used at the conclusion of each test as a purge gas in 
order to quench any remaining combustibles and to help cool the system.  In fact the 
small spike seen at the end of the test (light blue line seen in Figure 4.19) was a result of 
the purge gas turning on to conclude the test.  Similarly the spike at the beginning of the 
test (t = ~1 second) was a result of the ignition of the flammable mixture.  
 108
Under normal operating conditions, the pilot flame operated in a steady manner at 
4 psig.  Table 4.3 provided a summary of the pilot flame test conditions as calculated 
using the NASA CEA 2000 Chemical Equilibrium computer code84 based upon a 
measured combustor operating pressure.  The exit temperature, and density were 
calculated based upon an isentropic flow assumption and the following equations 
considering that the exit pressure was atmospheric.  The velocity was calculated by the 
one dimensional continuity equation.   
P
P
T
T
2
1
2
1
1
? ??? ???
?
?
?
              Eq. 4.14 
P
P
2
1
2
1
? ??? ?????
?
              Eq. 4.15 
 
 109
Table 4.3 Pilot Flame Operating Conditions (Calculated) 
Measured Pressure (psig) 4 
Adiabatic Flame Temperature (K) 3213 
Gas Temperature (K) at Exit 3137 
Equivalence Ratio ? 1.07 
Gas Exit Velocity (m/s) 800 
Gas Density (kg/m3) at Exit 8.84*10-2 
Specific Heat Ratio ? 1.11 
Mole Fractions at Exit 
CO 0.25466 
CO2 0.13798 
H 0.06121 
H2 0.06209 
H2O 0.25524 
O 0.05329 
OH 0.08937 
O2 0.08613 
 
 
In addition to the parameters described above the pilot flame plume and crossflow 
interaction was studied to understand the effects of the crossflow on the plume 
penetration.  A schematic diagram of the basic jet in a crossflow problem is provided in 
figure 4.20.  Basically, as a jet exhausts into a crossflow, the jet trajectory gets deflected 
 110
as a result of the force of the crossflow hitting it.  Therefore, as the crossflow is increased 
the jet penetration depth decreases.    
 
Fig. 4.20 Schematic Diagram of Jet in Crossflow Problem 
  
Unfiltered images were taken with the use of an Intensified Charge Coupled 
Device (ICCD) Camera of the pilot flame operating in the combustor under airflow 
conditions selected for fulfilling the main goals of this study without any additional fuel 
to determine the penetration characteristics of the pilot flame.  Figures 4.21-4.23 show 
light intensity contours of the three crossflows considered.  In this case, the crossflow 
velocity, or combustor inlet velocity is calculated by the one dimensional continuity 
equation.  As would be expected and is clear from the images, the penetration distance 
(y-axis) decreases from about 2 cm at 40 m/s to about 1.5 cm at 70 m/s.     
From these contours the outer boundary was extracted using the commercially 
available computer software Tecplot to give the profile of the plume jets as seen in Figure 
4.24.  This figure is a direct comparison of the outer boundaries of the different 
 111
conditions and also shows the decreasing penetration depth with increasing crossflow.  In 
this case, the jet exit was considered the origin for both the x and y locations.   
Of primary importance in jet-crossflow interaction is the momentum flux ratio, 
which was given in equation 4.1, and can be considered the principal controlling 
parameter in determining the jet penetration85.  Keffer and Baines86 found that the 
penetration profile data obtained from a subsonic jet in crossflow collapsed to a single 
curve in the near injector region when the coordinate axis were scaled by the momentum 
flux ratio, J.  After these profiles were collected, a correlation (Eq. 4.16) based upon the 
findings of Keffer and Baines86 for the penetration of the plume jet was developed as a 
function of the momentum flux ratio, J, and physical dimensions such as the x and y 
locations of the jet, and the width of the slot, w.  The results of this correlation can be 
seen in Figure 4.25, which shows that the data does in fact collapse to a single curve 
within experimental error. 
y
wJ a
x
wJ
n
? ??? ???              Eq. 4.16 
For crossflow ratios ranging from 3.06 to 5.36, the exponent, n, is 0.224. 
 112
 
 
Fig. 4.21 Intensity Contour of Plume Jet for Crossflow of 40 m/s 
 
 
Fig. 4.22 Intensity Contour of Plume Jet for Crossflow of 55 m/s 
 
 113
 
Fig. 4.23 Intensity Contour of Plume Jet for Crossflow of 70 m/s 
 
 
Fig. 4.24 Outer Jet Boundary Profiles for Selected Crossflows 
 114
 
Fig. 4.25 Determination of Jet Penetration Correlation 
 
 
4.3.4 Blowoff Limits 
 For the three crossflows selected a series of experiments were conducted to 
determine the blowoff limits for this combustor.  These tests were conducted by 
premixing ethylene far upstream (2.1 m) of the combustor and conducting experiments in 
which the pilot flame was turned off to see whether or not the mixture could sustain its 
own combustion.  Typically during an experiment, a flame was established at the location 
of the pilot flame (see figure 4.26), but in order to determine the blowoff limits of the 
combustor the pilot flame was turned off.  Blowoff is a condition in which a stable and 
self-sustaining flame is not occurring.  Blowoff can occur if a flammable mixture is too 
fuel lean, or if the gas velocity is well above flame speeds.  One resolution to blowoff in 
high-speed flows is the use of a flame holding device.  A flame holding device typically 
 115
creates an area of low speed flow where a flame can attach and sustain itself from.  In the 
case of this study, the rearward facing step provides a low speed recirculation zone in 
which a flammable mixture can potentially sustain itself under appropriates conditions.    
In this study, if the equivalence ratio was high enough (yet still fuel lean), the 
flame would anchor itself at the base of the rearward facing step as shown in figure 4.27.  
These images were taken with a video recorder with no filtering applied.  The flame was 
considered sustained if it stayed anchored at the base of the step for a period of at least 4 
seconds.   
Figure 4.28 shows the results of these experiments.  As would be expected, as the 
combustor inlet velocity was increased, the amount of fuel necessary in order to achieve 
an anchored flame was increased.  The equivelance ratio increase from about 0.58 at a 
velocity of 25 m/s to approximately 0.68 at a velocity of 70 m/s.  
 
Fig. 4.26 Normal Combustor Operation with Pilot Flame On 
 
 116
 
Fig. 4.27 Example of Flame Anchoring at Rearward Facing Step 
 
 
Fig. 4.28 Blowoff Limits 
 
  
 117
CHAPTER 5 
 
FLAT FLAME BURNER RESULTS AND ANALYSIS 
 
5.  Flat Flame Burner Results and Analysis 
This Chapter outlines and discusses the results of the flat flame burner 
experiments for which measurements of ignition and burning times of boron 
nanoparticles were made.  In addition, a brief discussion on recent relevant studies 
conducted by other researchers is given.     
 
5.1 Brief Discussion of Boron Combustion Physics 
 Chapter 2 provides a detail discussion of the physics of boron combustion.  
However a brief summary is provided here to identify some key parameters to be 
investigated in this study.  Early studies on boron particle combustion by Macek47-49 
describe a two-stage combustion process.  The first stage, sometimes termed the ignition 
stage, is associated with particle burning while the particle is still coated with an oxide 
layer.  As the particle heats up to above the B2O3 boiling point and the oxide layer is 
completely removed by evaporation, the second stage of boron combustion begins.  Since 
pure boron has relatively high melting and boiling temperatures, 2350 K and 4139 K 
respectively, the initial reactions in the consumption of the boron particle involve 
heterogeneous reactions between the gas phase oxidizer and solid or liquid phase boron.  
The second stage is essentially the full-fledged combustion of the bare boron particle.     
 118
According to Yeh and Kuo46, the burning time of a boron particle can be 
described by the following: 
     t t tb b dif b kin? ?, ,     Eq. 5.1 
where tb,dif and tb,kin refer to the combustion time due to species diffusion and the 
combustion time due to chemical kinetic mechanisms respectively.  They are defined by: 
    ? ?t d
D iYb dif
B
O
,
,ln
?
? ?
?
?
0
2
8 1
2
    Eq. 5.2 
and, 
    t dM kPXb kin B
B O
,
,
?
?
? 0
2
2
     Eq. 5.3 
where ?B, do, ?, D, YO2,?, MB, k, P, XO2,? refer to boron density, the initial diameter of the 
particle, the gas density, the gas diffusivity, oxygen mass fraction in the freestream, 
molecular weight of boron, a kinetic rate constant, pressure, and oxygen mole fraction in 
the freestream respectively.  It can be seen from these two equations that the diffusion-
controlled combustion is consistent with the D2-law (particle diameter), while the kinetic 
regime is consistent with a D1-law.  In order to determine which mechanism is dominant 
one can inspect the Damkholer number (Da) for surface reactions, which is defined as a 
ratio between the diffusion and kinetic time scales: 
    Da tt M kPd XD iYb dif
b kin
B O
O
? ? ? ?
?
,
,
,
,ln( )
0 2
2
4 1?    Eq. 5.4 
After substitution of parameters Yeh and Kuo46 determined that the Damkholer number 
reduces to: 
     Da Pd? 075      Eq. 5.5 
 119
where P is in atmospheres and d0 is in ?m.  When Da is much larger than unity, the 
diffusion time scale is much longer than the kinetic time scale.  Consequently, the 
burning time of a particle should follow a D2-law.  Conversely, if Da is much less than 
unity, the combustion is controlled by kinetics, or the D1-law.  In the present study 
involving nanoparticles combustion time scales dependent on a D1-law would be 
expected because Da is only about 0.003.    
 
5.1.1 Ignition and Combustion Data for SB99 Nano-boron 
 When the boron particles were subjected to environments with high enough 
temperatures and oxidizer concentration, a two-stage phenomenon was observed.  After a 
duration where there was no visible event, a yellow/orange emission zone appeared after 
which a sharp transition to a white glow zone appeared.  This is consistent with the 
observed behavior in the works of Li and Williams87.  Li and Williams87 concluded that 
the yellow zone was the ignition stage, sometimes referred to as stage 1 combustion, 
concluding with the complete removal of the oxide layer, while the white glow zone was 
the clean combustion stage of burning, often referred to as stage 2 combustion.  From this 
point forward, time scales associated with the ignition stage (stage 1) will be designated 
t1, while time scales associated with stage 2 combustion will be designated as t2.  
During this study, all of the test conditions resulted in at least the beginning of the 
ignition stage.  However, not all of the test conditions provided a suitable environment to 
reach and complete stage 2 combustion.  All of the experiments that had oxygen mole 
fractions (XO2) greater than or equal to 0.2 reached and completed stage 2 combustion.  
For the experiments with XO2 equal to 0.15, stage 2 combustion was reached for all tests, 
 120
however at the lowest temperature condition, stage 2 was not completed within the burner 
environment.  At XO2 equal to 0.1, only the highest temperature test resulted in achieving 
stage 2 of boron combustion.  Table 5.1 provides a summary of the results for the ignition 
and combustion of the SB99 particles. 
 
Table 5.1 Summary of SB99 Ignition and Combustion Results 
 
Figure 5.1 is an example of the two-stage phenomenon.  This is an inverted 
?instantaneous? image of visible wavelengths emission taken with an ICCD camera at an 
exposure time of 150 ?s.  Typically 50 images per test condition were taken and averaged 
together during image processing.  Upon filtering with a 546 nm (which corresponds to a 
band of BO2 emission) narrowband filter, the yellow ignition zone is no longer visible in 
the images as seen in figure 5.2.  BO2 is known to be prevalent during both the ignition 
and full-fledged combustion stages of boron combustion72.  The fact that it was not 
observed here does not mean that it is non-existent, but rather that for the camera settings 
used in the experiment (namely exposure time) that the intensity was not significant 
O2 H2O CO2 N2
1 0.200 0.234 0.117 0.450 1684 yes yes yes
2 0.200 0.266 0.133 0.400 1808 yes yes yes
3 0.200 0.300 0.150 0.350 1854 yes yes yes
5 0.200 0.234 0.117 0.450 1637 yes yes yes
6 0.300 0.200 0.100 0.400 1630 yes yes yes
7 0.300 0.233 0.117 0.350 1797 yes yes yes
8 0.300 0.267 0.133 0.300 1872 yes yes yes
9 0.300 0.200 0.100 0.400 1578 yes yes yes
10 0.150 0.267 0.133 0.450 1718 yes yes yes
11 0.150 0.267 0.133 0.450 1814 yes yes yes
13 0.150 0.267 0.134 0.449 1596 yes yes no
14 0.100 0.233 0.117 0.550 1614 yes no no
15 0.100 0.267 0.134 0.499 1712 yes no no
16 0.100 0.267 0.134 0.499 1791 yes yes yes
Test 
Condition
Average Burner 
Temperature (K)
Stage 1 
Reached
Stage 2 
Reached
Stage 2 
Completed
Product Mole Fractions
 121
enough to detect.  This result is actually beneficial as it provides a clear boundary for the 
ignition and full-fledged combustion stages.   
For the purposes of image processing, a background image is taken and averaged 
together with 50 similar images, and then ?subtracted? from the original averaged image.  
The background image is taken with no particles injected into the post flame region of the 
burner, but with the carrier gas still flowing.  Figure 5.3 is an example of a background 
image.  During image processing, an averaged version of figure 5.3 would be subtracted 
from an averaged version of figure 5.1 and the result is then processed for further 
analysis.      
 
 
 
Fig. 5.1 Image of Two-Stage Phenomenon in Nano-Boron Combustion (XO2 = 0.3, T 
= 1797 K) 
 
Yellow 
Ignition 
Region 
(t1) 
White Glow 
Region (t2) 
Transverse  
Particle Injection 
Burner Flow 
Direction 
 122
 
 
Fig.5.2 Filtered Image in Nano-boron Combustion (XO2 = 0.3, T = 1797 K) 
 
 
 
Fig. 5.3 Typical Background Image (no particles, XO2 = 0.3, T = 1797 K) 
 
Figures 5.4-5.7 are BO2 intensity contours plotted for all of the different test 
conditions.  These contours were created by subtracting the filtered averaged images with 
images taken for a given test condition with no particles flowing as a background 
Yellow Ignition Region 
Not Observed 
Burner Flow 
Direction 
Burner Flow 
Direction 
Transverse  
Particle Injection 
Transverse Injection 
(No Particles) 
 123
subtraction as was previously mentioned.  These images represent second stage boron 
combustion (bright white zone). 
In all cases the scale of the color bar represents intensity, which has arbitrary 
units.  For each oxygen mole fraction test condition the camera exposure time was held 
constant, along with the looping and binning settings of the camera.  However, the 
intensity levels between the different oxygen mole fraction conditions cannot be 
compared directly because the camera exposure was different in each case.  Table 5.2 
provides a summary of the camera settings for the experiments to obtain the filtered 
images. 
Figure 5.4 shows the contours for all of the tests conducted having an oxygen 
mole fraction of 0.3.  It is clear from the figure that as the temperature is increased, the 
location of the BO2 emission is shifted closer to the origin (X = 0).  This is a clear 
indication of faster ignition, which would be expected in higher temperature 
environments.  In addition, it is evident that the higher temperature test conditions exhibit 
higher peak intensity levels.  In the case of XO2 = 0.3 all of the test conditions attempted 
resulted in and completion of second stage boron combustion.   
Table 5.2 Camera Settings for Filtered Experiments 
XO2 Exposure (ms) Binning Gain (%) Loops 
0.1 2.0 4  x 4 10 2 
0.15 1.55 4  x 4 10 2 
0.2 1.35 4  x 4 10 2 
0.3 1.2 4  x 4 10 2 
 
 124
20 30 40 50 60 70 80 90 100  
 
 
Fig. 5.4 Intensity Contours for XO2 = 0.3 
  
Figure 5.5 is a summary of the second stage combustion for the test conditions 
with an oxidizer mole fraction of 0.2.  Once again, as one would expect as the 
temperature is increased, the ignition location shifts toward the origin, and the higher 
Transverse  
Particle Injection 
Transverse  
Particle Injection 
Transverse  
Particle Injection 
Transverse  
Particle Injection 
Burner Flow Direction 
 125
temperature tests clearly exhibit higher peak intensities.  In addition all of the test 
conditions with XO2 = 0.2 resulted in and completion of t2 combustion.  
20 30 40 50 60 70
 
 
 
Fig. 5.5 Intensity Contours for XO2 = 0.2 
  
For the XO2 = 0.15 conditions, all of the experiments resulted in second stage 
combustion. However, at the lowest temperature condition, although second stage 
Transverse  
Particle Injection 
Transverse  
Particle Injection 
Transverse  
Particle Injection 
Transverse  
Particle Injection 
Burner Flow Direction 
 126
combustion started, the particles left the post flame region prior to completion of 
combustion.  Figure 5.6 shows the results of the XO2 = 0.15 condition for the filtered 
images for all of the test conditions in which stage 2 combustion was completed.  Unlike 
the previous test conditions, after the bright white zone there appears to be some BO2 
intensity still being detected as shown in the contours.  The intensity levels following the 
white zone are much lower than within the white zone.  It is possible that these are simply 
a result of residual oxides15.  
 
 
 
 
Fig. 5.6 Intensity Contours for XO2 = 0.15 
 
Transverse  
Particle Injection 
Transverse  
Particle Injection 
Burner Flow 
Direction 
 127
 Finally, for the XO2 = 0.1 test conditions, figure 5.7 shows the contour of the only 
test condition which clearly achieved stage two burning.  At temperatures of 1712 K and 
1614 K, stage two combustion was not achieved in this experiment, where as for all of 
the other oxygen mole fraction conditions at temperatures as low as 1578 K stage two 
burning was achieved.  This indicates that temperature and oxygen mole fraction play an 
important role in the ignition stages of boron combustion.     
 
 
Fig. 5.7 Intensity Contour for XO2 = 0.1 
 
 In order to be certain that what was being detected was BO2, a series of 
experiments were conducted in which a different band pass filter, centered at 532 nm, 
replaced the 546 nm filter.  BO2 does not have any known bands that emit near 532 nm.  
Figure 5.8 shows a comparison of intensity levels when changing the filter from 546 nm 
to 532 nm.  In this case the camera settings for exposure, binning, and gain were set 
identically.  It is clear from figure 5.8 that the image collected with the 546 nm filter has 
Transverse  
Particle Injection 
Burner Flow 
Direction 
 128
a higher overall intensity than the image collected with the 532 nm filter.  In fact the 
image collected with the 546 nm filter has a total intensity that is 53% higher than the 
image collected with the 532 nm filter.  This indicates that clearly is being detected in 
levels beyond what ordinary black body radiation would produce.  It should be noted that 
the filters have a 10 nm full width half maximum value, meaning that they would have 
some overlap, therefore even without black body radiation, some level of emission could 
still be expected with the 532 nm filter. 
X (mm)
Y
(m
m)
30 35 40 45 50 55 600
5
10
15
20
25
30
INTENSITY
26
24
22
20
18
16
14
12
10
8
6
4
2
532 nm Filter
 
Y (mm)
Y
(m
m)
30 35 40 45 50 55 600
5
10
15
20
25
30
INTENSITY
26
24
22
20
18
16
14
12
10
8
6
4
2
546 nm Filter
 
 
Fig. 5.8 Intensity Contours using Different Filters (XO2 = 0.3, T = 1630 K) 
Transverse  
Particle Injection 
Transverse  
Particle Injection 
Burner Flow 
Direction 
 129
 Figure 5.9 shows the one-dimensional (i.e. columns of the gray-scale image have 
been summed) intensity profiles that were obtained from figure 5.8.  In this figure, both 
profiles have been normalized by the maximum value attained in the image collected 
with the 546 nm filter.  Again, clearly the image collected with the 546 nm exhibits 
higher intensity levels.  In fact, the peak intensity levels are about 25% higher, with 
integrated intensity levels more than 50% higher than the images collected with the 532 
nm filter.   
 
Fig. 5.9 Normalized Intensity Profile Using Different Filters 
 
During data reduction for all of these test conditions, the second-stage or t2 
burning time (filtered images) was determined first.  Once the spatial locations of the 
second-stage combustion were determined the starting location of the second-stage 
combustion was designated as the ending of stage 1 combustion.  However, the results of 
the ignition stage or first stage of boron combustion, with timescale designated as t1, is 
presented first.   
 130
In order to determine the spatial locations of the boundaries for second-stage of 
combustion and the ending of first stage of combustion, the problem was converted to a 
one-dimensional problem.  This was accomplished by summing the intensity in the y-
direction and plotting it versus the x-dimension of distance.  One reason to conduct the 
analysis this way is that the shape of the second stage combustion as seen in Figures 5.4-
5.7 is not necessarily consistent from test condition to test condition.  By making the 
analysis one dimensional, a systematic approach that did not depend on the ?shape? of 
the combustion was attained.     
With a 1D intensity profile, and velocity measurements described in Chapter 4, 
the ignition delay, ignition time, and burn time were obtained.  Generally in this type of 
problem there are three methods that have been used by other researchers to determine 
burn time15.  They include a constant intensity cutoff method, percent peak height 
method, and a percent total area method.   
Since area based methods provide the most unambiguous determination of burn 
time88, an area-based method was adopted in this study.  In this case the burning time is 
defined as occupying 95% of the area of the original profile.  Unfortunately as 
Beckstead34 pointed out, there is no accepted standard area fraction for determining 
burning times.  For the present study, the 95% area rule was adopted to ensure that any 
estimate of burning time would be relatively conservative.  Figure 5.10 is an example of 
the 1D intensity profile for determining the burning time.  In this case the blue line 
represents the original profile, while the red line represents 95% of the total area, and 
therefore the burning time. 
 131
 
Fig. 5.10 Example of t2 Burn Time Determination (XO2 = 0.2, T = 1808 K) 
 
5.1.2 Ignition Stage, t1 Results 
The starting location for the first stage or ignition stage was determined by 
examining the averaged unfiltered images.  The time scales associated with the first stage 
of boron combustion have been designated as t1.  The starting point was selected by 
locating the first sign of light emission in the images, while the ending point was 
determined from analysis of the filtered images as described above.  Figure 5.11 shows 
the results of the ignition stage, burning times for all of the test conditions that reached 
and completed second stage combustion.  The data shows that for these conditions, the 
ignition stage is relatively insensitive to oxygen mole fraction, within experimental error.  
t1 times range from about 5 milliseconds at the lower temperature levels all the way down 
to about 1.5 milliseconds at the higher temperature conditions.  The ignition stage time is 
reduced by a factor of about 1/3 over a temperature range of about 200 K. 
 132
0
0.001
0.002
0.003
0.004
0.005
0.006
1550 1600 1650 1700 1750 1800 1850
XO2=0.3
XO2=0.2
XO2=0.15
XO2=0.1Ig
nit
ion
 S
tag
e T
im
e, 
t 1,
 (s
ec
)
Average Burner Temperature (K)  
Fig. 5.11 t1 Burning Times for SB99 Boron 
 
 The ignition data obtained in this study is compared directly to the data obtained 
in the study by Yeh and Kuo46 for 2-3 ?m boron particles in figure 5.12.  The Yeh and 
Kuo46 study was similar to this one in nature in that they investigated the burning times of 
larger boron particles in the post flame region of a flat flame burner.  Interestingly the 
data for which there is an overlap in burner temperature is quite comparable, suggesting 
that the boron nanoparticles do not necessarily behave significantly different that the 
larger micron-sized particles during the ignition stage.  Another interesting note when 
comparing the data is that the ignition stage appears to become less sensitive to 
temperature as the temperature exceeds approximately 1750 K.   
It is important to note that although the data from the two different studies are 
being compared, that great care should be taken when interpreting the results.  First, the 
 133
same exact methods or criteria may not have been used in quantifying the burning time 
data or even the temperature data for that matter.  Beckstead34 considers this the largest 
source of data scatter when comparing all of the particle burning experiments conducted 
by different researchers.  Second, the test conditions between the two studies are not 
necessarily identical, although in each study, the results of the ignition stage studies 
appear to be relatively insensitive to the product gas species.  Finally, another important 
distinction between this study and all others is that this study obtains ensemble averages 
of the burning times for a size distribution of agglomerated particles.  This could 
potentially lead to slight biases in the burning and ignition times, whereas in the other 
studies ?single? particles were analyzed.  At any rate, the results of this study suggest that 
there is not a significant difference in the ignition stage of boron combustion between 
agglomerates of nanoparticles (with mean ?particle? sizes of 200 nm), and larger 2-3 ?m 
sized particles, at least where temperature ranges are similar.     
 
 134
0.001
0.002
0.003
0.004
0.005
0.006
1500 1600 1700 1800 1900 2000 2100
This Study (200 nm Agglomerates)
Yeh & Kuo (2-3 ?m)
Ign
itio
n S
tag
e T
im
e, 
t 1,
 (s
) 
Average Burner Temperature (K)  
Fig. 5.12 Comparison of Ignition Time, t1, with Other Data 
  
Since the ignition time, or t1, was primarily dependent on temperature, an analysis 
was performed to develop an analytical expression for the particle ignition time.  Table 
5.3 provides the values of the variables used in this analysis.  The analysis assumed that 
the lumped capacitance assumption was appropriate, in other words, the temperature of 
the particle is spatially uniform at an instant during a transient process.  In order to verify 
the validity of the lumped capacitance method, the Biot number, Bi, should be analyzed.  
The Biot number provides a measure of the temperature drop across a solid relative to the 
temperature difference between the surface and fluid.  If Bi << 1, the resistance to 
conduction within the solid is much less than the resistance to convection across the fluid 
boundary layer, therefore, the assumption of a uniform temperature distribution is valid68.   
 
 135
Table 5.3 Properties Used in Analytical Analysis of Ignition Time 
Boron Thermal Conductivity, k, W/[m*K] 27.6 
Particle Diameter, Dp, nm 200 
Thermal Conductivity of Freestream (air @ 1700 K), k?, W/[m*K] 0.113 
Convective Heat Transfer Coefficient, h?, W/[m2*K] 1,130,000 
 
The Biot number is defined as: 
    Bi h Lk c? ?         Eq. 5.6 
where, h? is the convective heat transfer coefficient, k is the thermal conductivity of the 
particle, and Lc is the characteristic length defined as: 
     L VAc
s
?      Eq. 5.7 
where V refers to the particle volume, and As refers to the particle surface area.  The 
convective heat transfer coefficient can be determined from the definition of the Nusselt 
number: 
     Nu h DkD p? ?
?
     Eq. 5.8 
where Dp is the particle diameter and k? is the freestream thermal conductivity.  For 
laminar flow, i.e., as the Reynolds number approaches zero, the Nusselt number 
approaches a value of 2 as equation 3.5 would suggest.  Substituting in the appropriate 
parameters yields a typical Biot number of 0.005, substantiating the assumption of 
lumped capacitance.  Since the lumped capacitance method was valid for this analysis, an 
 136
 
?VC dTdtp
Dp 
? ?h A T T? ? ?
energy balance on the particle could be applied.  Figure 5.13 is a schematic diagram 
showing the details of the energy balance.   
 
 
 
 
 
Fig. 5.13 Schematic Diagram of Energy Balance 
 
The energy balance for the particle is given by: 
   ? ? ? ?E E E Est in out gen? ? ?     Eq. 5.9 
where, ?E refers to the rate of change of energy, and the subscripts st, in, out, and gen 
refer to stored, in, out, and generated respectively.  In this analysis chemical reactions 
have been neglected, therefore the rate of energy generated is negligible.  Similarly, 
radiation losses have been neglected; therefore the rate of energy leaving the system is 
also negligible.  With those simplifications, the energy balance reduces to only the 
storage term and a convective term as a energy input: 
   ? ? ( )E E VC dTdt h A T Tst in p p s? ? ? ?? ??             Eq. 5.10 
Introducing the temperature difference, ? ? ? ?T T , where ddt dTdt? ? and substituting into 
equation 5.10 we get: 
     ? ? ?p p
s
VC
h A
d
dt? ?              Eq. 5.11 
 137
Separating variables and integrating from the initial condition, t = 0 and T = Ti, we 
obtain: 
     ? ????p p
s o
tVC
h A
d dt
i?
? ??              Eq. 5.12 
Evaluating the integral and substituting in the variables results in the solution for the 
ignition time of the particles: 
    t D Ch T TT Tp p p initial
c
? ????? ???
?
?
?
?
6 ln              Eq. 5.13 
From this analytical solution a correlation was developed for the ignition times for the 
two highest oxygen mole fraction conditions, XO2 = 0.2, 0.3, in which the constants were 
lumped together into one ?time constant?, ?c, and Tc is the critical temperature to 
complete ignition in the burner.  Figure 5.14 shows the results of the correlation.  Critical 
temperatures were 1610 K and 1550 K for XO2 = 0.2, 0.3 respectively. 
       t T TT Tc initial
c
? ????? ????
?
? ln              Eq. 5.14 
 
 138
0.0015
0.002
0.0025
0.003
0.0035
0.004
0.0045
0.005
1550 1600 1650 1700 1750 1800 1850
t1 XO2 = 0.3
t1 XO2 = 0.2
Ign
itio
n T
im
e, 
t 1,
 (s
)
T (K) 
t1 = 0.00128 * ln((Tamb - Tinf)/(Tc - Tinf))
R2 = 0.92 Tc = 1550 K
t1 = 0.00123 * ln((Tamb - Tinf)/(Tc - Tinf)) 
R2 = 0.88 Tc = 1610 K
 
Fig. 5.14 Ignition Time Correlation 
 
5.1.3 Second Stage/Full-Fledged Combustion, t2 Burning Time Results 
A summary of the second stage combustion times corresponding to the white 
glow zone described earlier is provided in figure 5.15.  Typical burning times range from 
about 3 milliseconds down to about 1.5 milliseconds.  Oxygen mole fraction, XO2, had a 
clear effect on the burning time particularly at the lower range of the temperatures 
examined in this study.  For example the burning time at a temperature of about 1625 K 
was approximately 50% longer for XO2 = 0.2 than for XO2 = 0.3.  As the temperature was 
increased, the effect of XO2 was less pronounced although there was still a significant 
difference between the two highest oxygen mole fractions, 0.2 and 0.3.  For XO2 ? 0.2 
there was no clear effect of oxygen mole fraction on the second stage burning time.  
However, the effect of temperature became more pronounced at low XO2.  In the case of 
 139
XO2 = 0.15 at the lowest temperature condition second stage combustion was not 
completed before the flow left the burner area.  For XO2 = 0.1, second stage combustion 
was not even reached at all but the highest temperature condition.   
These are important results to consider for airbreathing propulsion systems whose 
lone oxidizer was air, which has an oxygen mole fraction of 0.21.  In most practical 
systems, if a metal were to be used as part of the fuel it would be stored in either a solid 
hydrocarbon matrix, or in a hydrocarbon gel or slurry.  In a best-case scenario with 
nanoparticles, the total fuel could possibly be comprised of about 50% by weight of the 
metal.  Since the burning times of the hydrocarbon would be much shorter than that of 
the boron in the fuel, the boron would be left to combust in an environment that only has 
whatever oxygen content was leftover from the hydrocarbon combustion, i.e. less than 
XO2 = 0.21.  Therefore, the hydrocarbon combustion must provide an environment that 
has a temperature between 1712 K and 1790 K if XO2 = 0.1 just to ensure that second 
stage combustion could take place. 
 140
0.001
0.0015
0.002
0.0025
0.003
0.0035
0.004
1550 1600 1650 1700 1750 1800 1850 1900
XO2=0.3
XO2=0.2
XO2=0.15
XO2=0.10
t 2 
Bu
rn
ing
 T
im
e (
s)
Average Burner Temperature (K)  
Fig. 5.15 t2 Burning Times for SB99 B 
 
 Figure 5.16 shows the total burning time, t1 + t2, for the SB99 particles.  The 
effect of oxygen mole fraction was clear at the lower temperature conditions, which 
comes directly from the fact that at the lower temperatures the t2 burning exhibits 
significant differences in burning time when comparing the XO2 quantities.  At the lower 
temperatures, the ignition stage was significantly longer than the full-fledged combustion 
stage.  Therefore the total burning times at the lower temperatures was dominated by the 
ignition stage.  As the temperature was increased, the ignition stage and full-fledged 
combustion stage became quite comparable. 
 141
0.003
0.004
0.005
0.006
0.007
0.008
1550 1600 1650 1700 1750 1800 1850
XO2 = 0.3
XO2 = 0.2
XO2 = 0.15
XO2 = 0.1
To
tal
 B
urn
ing
 Ti
me
, t 1
 + 
t 2,
 (s
)
Average Burner Temperature (K)  
Fig. 5.16 Total Burning Times (t1+t2) for SB99 B 
 
 Figure 5.17 is a comparison of the full-fledged combustion data collected in this 
study with data collected in other studies.  The data collected by Macek47,49 used particles 
ranging from about 30 ?m to 100 ?m.  The data from Li and Williams87 came from 7 and 
10 ?m boron particles, while the data from Yeh and Kuo46 came from particles that were 
2-3 ?m.  The data collected in this study was plotted considering its mean mobility 
diameter, 200 nm.   
Rather than plot just the burning time versus particle size, the product of burning 
time and oxygen mole fraction (XO2*t2) was plotted against particle size.  These 
coordinates were chosen since both the chemical reaction time, tb,kin, and the diffusion 
time, tb,diff, in equation 5.1 are inversely proportional to XO2, therefore the product of 
XO2*t2 is approximately independent of XO2, and should only be dependent on particle 
 142
size and pressure87.  From figure 5.17 it can be seen that the burning time of the largest 
particles studied by Macek47,49 are dependent on the particle diameter by a factor of 
approximately 2, which would be consistent with diffusion limited burning, or the D2-
law.  The intermediate sized particles studied by Li and Williams87, and Yeh and Kuo46 
have burning times that approach the kinetic limited regime where the burning time 
follows a D1-law.  As the size of the particles continue to decrease into the nanoparticle 
range, the particle size dependence continues to decrease, to well below unity (tb ~ D0.22), 
beyond what traditional theories have explained.  Even if the maximum size particle were 
considered rather than the mean size, i.e. 400 nm instead of 200 nm, the size dependence 
only reaches t2 ~ D0.31, still well below anything that traditional theory can explain.  This 
result certainly brings into question any benefit achieved from using boron nanoparticles 
in comparison to micron-sized particles for airbreathing applications.  If the burning 
times are comparable and the nanoparticles suffer from the active content issues 
described in Chapter 4, then the benefits are likely to be very minimal compared to the 
downfalls. 
 143
0.0001
0.001
0.01
0.1
0.1 1 10 100 1000
Macek
This Study
Yeh & Kuo
Li & Williams
XO2*t2= 4.9997e-07 * D2.42   R2= 0.97038 
XO2*t2 = 0.00049118 * D0.22   R2= 0.67147 
XO2*t2= 0.00033371 * D0.74   R2= 0.89596 
X O
2 
* t
2 
(s)
Particle "Size" (?m)
Macek DataLi & Williams
Yeh & Kuo
This Study
 
Fig. 5.17 t2 Burning Time Comparison to Larger Particles 
 
 Svante Arrhenius (1859-1927) stated that only those molecules which possess 
energy greater than a certain amount Ea, the activation energy, will react29, and these 
high-energy, active molecules lead to products.  The so-called Arrhenius law (Eq. 5.15) 
to calculate chemical reaction rates was named for Svante Arrhenius. 
     k A ER Ta
u
? ???? ???exp                Eq. 5.15 
where A is assumed to include the effect of the collision terms, the steric factor associated 
with the orientation of the colliding molecules, and the mild temperature dependence of 
the preexponential factor25.  Ru is the universal gas constant, and T is the temperature.  A 
form of this Arrhenius law was applied to the data collected in this study for the full 
 144
fledged combustion regime of the boron particles.  In this case the Arrhenius law took the 
form of equation 5.16. 
     1
2 2
X t A
E
R TO
a
u
? ???? ???exp              Eq. 5.16 
The results of applying the modified form of the Arrhenius law can be seen in figure 
5.18.  The data collected by Yeh and Kuo46 is also included for comparison.  In this case, 
only the data collected for XO2 = 0.2 and 0.3 so that the conditions could be most 
comparable to the Yeh and Kuo46 data.  In this range, the activation energies between the 
two studies, and two very different particle sizes, are comparable considering 
experimental error and data scatter.  This is in great contrast to similar data for 
nanoaluminum.   Park24 found that the activation energy of aluminum significantly 
decreases with decreasing particle size. 
1000
104
0.0005 0.00052 0.00054 0.00056 0.00058 0.0006 0.00062 0.00064
This Study
Yeh and Kuo
1/[
t 2*
X O
2]
 (1
/s)
1/T (K-1)
Nanoparticles (This Study):
1/[t2*XO2] (1/s) = A * exp[-Ea/(Ru * T(K))] 
A = 21058 (1/s)
Ea = 33.45 KJ/mol 
R2 = 0.78
2 Micron Particles (Yeh and Kuo):
A = 10460 (1/s)
Ea = 27.5 KJ/mol 
R2 = 0.59
 
Fig. 5.18 Arrhenius Burning Rate Law for t2  
 145
 Since the Damkholer analysis suggests that the burning of these boron 
nanoparticles should be kinetically controlled, a burning time correlation based on a 
kinetically controlled system was determined89 using the Arrhenius parameters obtained 
above.   
t aD
X A ER T
n
O
a
u
2
2
? ??
??
?
??exp
                Eq. 5.17 
The correlation takes the activation energy determined for the nanoparticles in this study, 
and ranges from the nanoparticle data collected in this study up through the data collected 
by Yeh and Kuo46 and can be seen in figure 5.19.  This correlation reduces some of the 
scatter seen in the correlations developed earlier.  With the Arrhenius parameters 
included, the size dependence decreases even further to tb ~ D0.14.  Therefore, only a 
small benefit in burning time was obtained when going from 2-3 ?m down into the 
nanometer range for boron.  
 146
0.1
1
10
0.1 1 10
t 2 *
 X
O2
 * 
A 
ex
p[-
E a
/(R
u*
T)
]
 Particle "Diameter" (?m)
t2 = aDn/[XO2 * A exp[-Ea/(Ru*T)]
a = 1.249
n = 0.14
Ea = 33.45 KJ/mol
R2 = 0.8
Yeh and Kuo
This Study
 
Fig. 5.19 t2 Burning Time Correlation 
 
5.2 Studies on Nanoaluminum Combustion 
 As was discussed in detail in Chapter 2, a number of other researchers have 
studied nanoaluminum in great detail.  Bazyn17 made estimates of burning times for 
nanoaluminum at high temperatures and pressures in a shock tube.  Parr16 estimated the 
ignition temperatures, and ensemble average burning times similar to this study of 
nanoaluminum in a hydrogen/oxygen/diluent Henken burner.  Parr?s16 studies revealed 
that the ignition temperature of some types of nanoaluminum was much lower than that 
of traditional micron sized aluminum.  In fact he found ignition temperatures less than 
1100 K for several types of nanoaluminum including ALEX.  Parr16 also demonstrated a 
larger dependence of burning time on temperature for nanoaluminum which had not 
previously been established for micron-sized particles.  Ordinarily aluminum particle 
 147
combustion times are only weakly dependent on temperature once ignition is achieved, as 
Beckstead?s correlation34 suggests tb ~ T-0.2.   This is due to the fact that the larger 
aluminum particles establish their own temperature field by burning with a detached 
diffusion flame whose temperature is predetermined by the vaporization-dissociation 
temperature of the aluminum oxide85.  Therefore, the surrounding gas temperatures do 
not play a large role in the particle burning time of micron-sized aluminum. 
 As a basis for comparison between the burning times of boron and aluminum 
nanoparticles, at a temperature of about 1600 K, Parr16 found that aluminum 
nanoparticles had burning times between roughly 0.8 and 1 millisecond and at 1800 K the 
measured burning times ranged from about 0.7 to 0.9 milliseconds.  In this study at 
similar temperatures the t2 burning times of the boron nanoparticles ranged from about 
1.7 to 3 milliseconds at roughly 1600 K and 1.5 to 2 milliseconds at about 1800 K.  In 
general the boron nanoparticles had burning times that were 2 to 3 times longer than the 
aluminum nanoparticles.  It should be noted that the environments in the two studies were 
quite different.  In Parr?s16 study, the nanoaluminum was subjected to water vapor as its 
only oxidizer, whereas in this study the boron was subjected to oxygen, water vapor, and 
carbon dioxide.  As Beckstead?s correlation34 points out, oxygen is a better oxidizer for 
aluminum than water vapor, therefore in similar environments it would be expected that 
the burning times of aluminum and boron nanoparticles would be separated by more than 
a factor of two or three.  As was previously mentioned for larger particles boron has 
burning times that are roughly four times that of aluminum in similar environments.         
 In a separate study, Huang89 correlated the data obtained by Parr16 in a similar 
form as equation 5.17, which can be seen in equation 5.18.   
 148
    t D
X C ERT
b
eff
b
? ??
??
?
??
0 3.
* exp
             Eq. 5.18 
where D is in cm, Xeff is the same as given by Beckstead?s correlation34, C = 5.5 * 10-4 , 
and Eb = 73.6 kJ/mol.  Interestingly, upon doing so, Huang88 found that the burning time 
dependence on particle size was only tb ~ D0.3 once again showing that the burning time 
is only weakly dependent on particle size in the nanoparticle regime.  However, 
nanoaluminum still has other advantages over micron-sized aluminum, such as the 
greatly reduced ignition temperatures.  Therefore in an airbreathing propulsion system 
where the aluminum would be carried in a solid matrix, gel, or slurry, one would not need 
nearly as much heat supplied for ignition of the aluminum compared to boron.  That 
provides a several advantages: 
1) The aluminum can be a larger percentage of the fuel (increasing density impulse) 
2) The combustor performance would be less sensitive to off design conditions.  The 
combustor could operate in a very fuel lean condition and still provide enough 
heat to ignite and combust the aluminum.   
  
5.3 Implications on High-Speed Airbreathing Combustors 
 As was established in section 5.1.3 and 5.2, the burning times of both boron and 
aluminum nanoparticles are only weakly dependent on the initial particle size.  This 
would tend to imply that there might not be a large benefit in employing them as fuels or 
fuel supplements.  However, any reduction in burning time certainly improves the ability 
to extract the available energy of a particle in a time sensitive environment such as a 
ramjet or scramjet combustor.   
 149
For boron, the total burning time data for realistic airbreathing conditions (i.e. XO2 
< 0.2 for a traditional ramjet or scramjet) shown in figure 5.14 shows that the shortest 
total burn times measured were about 4 milliseconds at about 1800 K.  Equation 5.3, 
which describes the burning time for the kinetically dominated regime implies that an 
increase in pressure will reduce the burning times.  For the time durations that the 
particles were exposed to the flat flame burner products, below about 1700 K (for XO2 = 
0.1) stage 2 combustion was never achieved.  If the particles had been subjected for 
longer durations it is possible that they could have completed the ignition stage.  
However, there was no indication in this study that the nano boron would have fully 
ignited (reached stage 2) at a lower temperature than what other researchers47, 90, 91 have 
observed in micron-sized particles.  Boron ignition has been observed to occur from 
1500-1900 K depending on the type of test and its conditions.  
For aluminum, the studies of Parr16 and Bazyn17 indicated burning times in the 
100?s of ?s which certainly suggests the possibility of complete combustion of a particle 
with a high-speed combustor.  Perhaps of greater importance was that Parr16 determined 
that the ignition temperature for some types of nanoaluminum is lower than traditional 
micron-sized particles.  This could allow the combustor to operate at fuel lean conditions 
(lower temperatures) yet still provide efficient combustion, which could help reduce the 
thermal protection requirements, improve specific impulse, and improve density impulse 
by allowing more of the fuel to be made up of metal.   
Unlike boron combustion, H2O and CO2 are reasonably effective oxidizers for 
aluminum.  H2O in particular can inhibit boron combustion, but according to Beckstead34 
is about 60% as effective of an oxidizer as O2.  This is an important since the metals are 
 150
essentially burning in the products of hydrocarbon combustion, even though the O2 
concentration is decreasing, H2O and CO2 still contribute significantly to the combustion 
of the aluminum particles.           
 151
CHAPTER 6 
 
AIRBREATHING COMBUSTOR RESULTS 
 
6. Airbreathing Combustor Experiments 
One of the major goals of this study was to determine whether or not the great 
potential of metallized fuels can be realized under realistic combustor conditions.  In 
particular, can all of the available energy of a metallized fuel be extracted in the short 
residence times typical of a high speed airbreathing combustor.  Chapter 6 discusses the 
results of this portion of the study. 
 
6.1 Test Conditions 
Once the combustor was designed, fabricated, and some basic characterization 
studies were conducted, testing under more realistic operating conditions commenced.  
Initially, ethylene (C2H4), a hydrocarbon was used as the sole fuel in order to create a 
baseline for fuel evaluation.  Following the baseline testing, testing started with 
metallized fuels.  The metallized fuels were mixtures of gaseous ethylene with either 
aluminum or boron added in as a particle laden flow.   
Figure 6.1 is a series of images recorded with a Sony Video Recorder showing the 
combustor operating with and without the addition of aluminum or boron nanoparticles.  
The ethylene only tests were characterized by the visible blue light as seen in the figure, 
the tests with aluminum addition appeared white in color, and the tests with boron 
appeared green or yellowish in color.  No filtering was applied in obtaining these images.  
 152
Table 6.1 provides a summary of all of the test conditions.  The test conditions given are 
nominal values and all data reduction was conducted on a test-by-test basis, in which the 
conditions could vary slightly from the given conditions.  A total of 20 different types of 
test were conducted.  Test conditions designated by a numeral were ethylene only tests.  
Conditions designated with a numeral followed by a letter ?a? or ?b? represent 
aluminized or boron loaded tests respectively.  Tests with the same numeral designation 
have identical ethylene mass flow rates, such that the only difference between the two is 
the addition of the metal nanopowders.   
The combustor inlet velocity was calculated from a one dimensional mass flow 
rate calculation knowing the flow area, ambient pressure and the temperature.  The metal 
mass flow rates came from the particle seeder calibration and represent the total mass 
flow rate of the particles.  However, the weight percentages given represents the amount 
of pure metal powder (i.e. the active content) material in the fuel.  The metal oxides were 
lumped into the fuel for the purpose of the calculations.  The adiabatic flame 
temperatures were calculated with the NASA CEA 200084 chemical equilibrium code.  
The flame temperature calculations take into account the active content of the particles 
and assumed that the remain particle material was the metal oxide.  The equivalence ratio 
also included both the fuel and oxidizer contributions from the pilot flame. 
Table 6.2 shows some of the flow parameters of interest for the different inlet 
velocities.  The Reynolds number was calculated based upon flat plat assumptions, where 
the average velocity was calculated by the one dimensional continuity equation, the 
length scale was taken as the distance between the flow straightener and the center of the 
pilot flame.  The upper limit for laminar flow on a smooth flat plate is 3 x 106, therefore 
 153
under these assumptions the flow remained laminar up until the pilot flame location.  The 
residence time was calculated using the average velocity and the difference in the axial 
location of the pilot flame and the thermocouples. 
 Finally, table 6.3 shows the mole fractions of the main oxidizing species that 
occur as a result of the gaseous ethylene combustion.  Essentially the mole fractions 
shown are for the test conditions designated by a numeral only.  This is shown since the 
time-scale for the gaseous hydrocarbon combustion is much smaller than that of the 
metals; therefore, the metals were considered to be burning in the product species of the 
ethylene combustion.  
 
Table 6.1 Summary of Combustor Test Conditions  
Test 
Condition 
Inlet Velocity 
(m/s) ?mMetal  (g/s) 
Wt % Metal  Weight % 
Premixed C2H4 
Overall ? Adiabatic Flame 
Temperature (K) 
1, 1b 40 ? 1 0, 0.8 0, 12.7 ? 0.6 60, 49 0.64, 0.69 1960, 2090 
2, 2b 40 ? 1 0, 0.8 0, 15.2 ? 0.8 49, 38 0.50, 0.57 1685, 1850 
3, 3b 55 ? 2 0, 0.8 0, 12.7 ? 0.6 59, 49 0.48, 0.52 1596, 1712 
4, 4b 55 ? 2 0, 0.8 0, 9.7 ? 0.5 70, 61 0.64, 0.69 1946, 2065 
5, 5b 70 ? 2 0, 0.8 0, 9.7 ? 0.5 70, 60 0.51, 0.54 1660, 1746 
6, 6a 40 ? 1 0, 0.39 0, 7.4 ? 0.4 62, 57 0.69, 0.70 2056, 2120 
7, 7a 55 ? 2 0, 0.39 0, 7.2 ? 0.4 62, 57 0.51, 0.52 1680, 1744 
8, 8a 55 ? 2 0, 0.87 0, 14.8 ? 0.7 48, 40 0.50, 0.52 1645, 1771 
9, 9a 55 ? 2 0, 0.87 0, 11.4 ? 0.6 62, 54 0.69, 0.70 2031, 2120 
10, 10a 70 ? 2 0, 0.87 0, 11.5 ? 0.6 62, 54 0.54, 0.56 1740, 1825 
 
 
Fig. 6.1 Combustor Operating in Traditional Ramjet Mode with a) Gaseous 
Ethylene as Fuel b) Gaseous Ethylene with 50 nm Aluminum Particles and c) 
Gaseous Ethylene with Boron Particles 
 
a) b) c) 
 154
Table 6.2 Flow Conditions 
Inlet Velocity (m/s) Inlet Reynolds Number, Re Residence Time (s) 
40 64,168 0.0103 
55 88,232 0.0075 
70 112,295 0.0059 
 
Table 6.3 Mole Fractions of Oxidizing Product Species 
Test Condition XO2 XH2O XCO2 
1 0.075 0.091 0.091 
2 0.107 0.072 0.072 
3 0.110 0.068 0.068 
4 0.074 0.089 0.090 
5 0.104 0.073 0.073 
6 0.064 0.097 0.098 
7 0.103 0.072 0.072 
8 0.105 0.071 0.071 
9 0.063 0.095 0.096 
10 0.095 0.075 0.076 
 
 
6.1.1 Mie Scattering Results 
Planar Mie Scattering was employed to determine the particle dispersion patterns.  
The test configuration and specific diagnostics were described in Chapter 3.  The images 
were recorded looking directly through the window located on the side of the combustor, 
such that the laser sheet, which was passed through the top of the combustor, and camera 
 155
were perpendicular to each other.  Figure 6.2 shows an example of typical instantaneous 
Mie Scattering Images.  These images were taken for test condition 6a. 
 
pilot
nano-particle
injection
laser sheet
 
Fig. 6.2 Typical Instantaneous Mie Scattering Images (Inlet Velocity = 40 m/s, wt% 
Al = 7.4, ? = 0.7, Tad = 2120 K) 
 
The images show that in the vicinity of the pilot flame the intensity levels 
fluctuate.  This was due to particles being entrained into the pilot jet.  Further 
downstream the scattering intensity appeared to become more uniform.  Figure 6.3 
illustrates the average intensity contour from all of the images collected from one test.  
Downstream of the step, the contour shows a fairly uniform scattering intensity, 
suggesting fairly uniform particle dispersion although the lower portion of the combustor 
appeared to have a slightly greater intensity.   
In order to investigate this further, figure 6.4 shows the average intensity profile 
from the averaged image at axial positions of 10, 15, and 20 cm for test condition 6a.  
The spikes near the bottom and top of the combustor were due to reflections from the 
surfaces of the test rig and should be ignored when interpreting the profiles.  At the 10 cm 
 156
axial location there was a relatively large spike between the 1 and 2 cm combustor height 
positions.  The spike was associated with the combustor step.  Up until the 10 cm axial 
position, there was a relatively large concentration of particles in the approximate height 
of the step.  A little further downstream (15 and 20 cm) the particles appear to be mixing 
more thoroughly into the combustor.   Inspection of figure 6.4 at the downstream 
locations indicated that the average intensity was relatively constant up to about the 3 cm 
vertical location point.  At that point the average intensity slowly decreased as the 
vertical location was increased towards the top portion of the combustor.  Near the top of 
the combustor the average intensity was about 70% of the ?constant intensity? region.  
This indicates that at these axial locations, the majority of the ALEX particles were in the 
bottom portion of the combustor, which would suggest that for these axial locations the 
equivalence ratio was not entirely uniform throughout the combustor, although the trend 
suggests that it was becoming more uniform. 
Figure 6.5 is the average intensity profile from the averaged image at axial 
positions of 10, 15, and 20 cm for test condition 7a.  Again there was a large spike at the 
10 cm axial location between the 1 and 2 cm vertical locations that was associated with 
the rearward facing step.  At the other axial locations, the intensity was relatively 
constant through the majority of the combustor, however, the bottom portion of the 
combustor did tend to exhibit higher intensity levels.    
 157
 
Fig. 6.3 Average Intensity Contour of Mie Scattering Images (Inlet Velocity = 40 
m/s, wt% Al = 7.4, ? = 0.7, Tad = 2120 K) 
 
 
Fig. 6.4 Average Intensity at Axial Locations (Inlet Velocity = 40 m/s, wt% Al = 7.4, 
? = 0.7, Tad = 2120 K) 
 158
 
Fig. 6.5 Average Intensity at Axial Locations (Inlet Velocity = 55 m/s, wt% Al = 7.2, 
? = 0.52, Tad = 1744 K) 
 
6.1.2 Chemiluminescence Results 
6.1.2.1 AlO Emission 
AlO is a known gas phase intermediate in aluminum oxidation42,43.  Figure 6.6 is 
series of examples of typical instantaneous images that were obtained with an ICCD 
camera and a narrow-band interference filter, centered at 488?2 nm.  A camera shutter 
speed of 5 ?s and a framing rate of 12 Hz was used to capture the images.  The 
instantaneous images show some fluctuation but qualitatively show that the most intense 
regions of AlO emission do not extend much beyond the combustor step.     
In the wavelengths allowed by the band pass filter there were some bands of 
emission due to hydrocarbon combustion.  In particular, C2*, CH*, and CO2 exhibit 
bands that overlap the filter.  However, a series of ethylene/air only tests were conducted 
 159
with this filter in place, and the emission due to hydrocarbons was minimal which can be 
seen in figure 6.7.  Testing showed that peak background emission of the hydrocarbon 
only tests was less than 1/20th of the peak emission of the aluminized flows.   
Figure 6.7 is a series of AlO intensity contours for all of the aluminized tests 
obtained by averaging together the images collected in a test.  Typically, the images 
collected in the first second of the test are not used to allow the particle seeder to 
approach a steady state condition.  Therefore in the case of the aluminum testing, in 
which a typical experiment was four seconds in duration, a total of 36 images were 
averaged to create the contours.  It is clear from the figure that the majority of AlO 
emission took place within the first 15 cm of the combustor for all conditions.  However, 
there is a clear trend of decreasing axial distance of AlO emission with decreasing 
combustor inlet velocity.   
Comparing the tests with similar inlet velocities and aluminum flow rates (8a and 
9a), we see that the test with a higher flame temperature (9a) exhibits higher overall 
intensities (see figure 6.7).  Although the majority of the AlO emission takes place within 
the first 15 cm of the combustor, the highest temperature tests (6a and 9a) did exhibit 
some significant emission much further downstream.  This could be a result of larger 
agglomerates igniting and or combusting further downstream compared to the smaller 
agglomerations, or as Olsen pointed out residual oxides/suboxides15.   
  In order to make an estimate for the ignition delay, the intensity was plotted 
along the length of the combustor starting from the pilot flame (figure 6.8) at a combustor 
height of 3.75 cm and then converted to time (figure 6.9), by considering the combustor 
inlet velocity.  The ignition delay was estimated by taking the first instance when the 
 160
intensity reached half of its maximum value. The delay was virtually identical in all of 
the tests of a given additive.  For aluminum the delay was consistently between 0.3 to 0.4 
milliseconds.  Typical uncertainty in this measurement is 0.1 millisecond.   
 
 
Fig. 6.6 Instantaneous AlO Chemiluminescence Images 
 161
 
 
Fig. 6.7 AlO Chemiluminescence Contours (C2H4 only, 6a, 7a, 8a, 9a, 10a) 
 
Tf = 2120 K 
V = 40 m/s 
Tf = 1744 K 
V = 55 m/s 
Tf = 1771 K 
V = 55 m/s 
Tf = 2120 K 
V = 55 m/s 
Tf = 1825 K 
V = 70 m/s 
C2H4 Only 
 162
 
Fig. 6.8 AlO Intensity Profiles Versus Combustor Length 
 
Fig. 6.9 AlO Intensity Profiles Versus Time 
 
 
 163
6.1.2.2 BO2 Emission 
The BO2 molecule is a reactive intermediary gas-phase species formed throughout 
boron particle ignition and combustion51.  According to Spalding72, the BO2 spectrum is 
readily observable during both the combustion and ignition stages of boron combustion.  
Figure 6.10 is a series of instantaneous BO2 emission images.  A camera shutter speed of 
4 ?s and a framing rate of 12 Hz was used to capture the images.  A Narrow-band 
interference filter, centered at 546?2 nm was used to capture images of the BO2 emission.  
Since the primary fuel was a hydrocarbon (Ethylene), the potential for hydrocarbon 
emission overlapping the BO2 band existed.  In particular, C2 exhibits several bands in 
the vicinity of the filter.  However, according to Gaydon75, the C2 bands near 546 nm are 
characterized as weak or very weak.  In order to conclude without question that the 
emission detected was BO2, several tests with ethylene only were conducted with the 546 
nm filter in place.  In each case any emission detected was negligible.  An example of the 
testing for background hydrocarbon emission can be seen in figure 6.11, which shows 
that no emission was detected at the previously stated camera settings for tests that did 
not include boron.  
 164
 
Fig. 6.10 Instantaneous Images of BO2 Emission 
 
Figure 6.11 provides a summary of intensity contours for all of the tests that 
contain boron particles.  Just as in the ALEX experiments, the contours were created by 
averaging all of the images collected during an experiment except for the first second.  In 
the case of the boron experiments, the typical test duration was 3 seconds.  Therefore, 24 
images were averaged to create the contours.   
The top contour is the case in which no boron was injected to determine the extent 
of hydrocarbon emission interference.  The second and third contours are for the slowest 
speed cases and show the largest intensities of all of the cases.  The fourth and fifth 
contours are the middle speed cases, and the last contour is the highest speed case.    The 
presence of BO2 clearly indicates the ignition of the boron particles.  It is interesting, but 
not surprising to note that all of the contours show that the highest intensity regions are 
significantly downstream of the pilot flame (X=0 cm) region suggesting a delay in the 
 165
ignition of the boron particles.  Qualitatively, comparing the tests with similar velocities 
(1b and 2b, and, 3b and 4b), it is clear that the tests with higher adiabatic flame 
temperatures (1b and 4b) exhibited higher intensities of BO2 emission and for longer 
spatial durations.   
     In order to make an estimate for the ignition delay, the intensity was plotted along 
the length of the combustor starting from the pilot flame (figure 6.12) at a combustor 
height of 3.75 cm and then converted to time (figure 6.13), by considering the combustor 
inlet velocity.  Again, the ignition delay was estimated by taking the first instance when 
the intensity reaches half of its maximum value. The delay was virtually identical in all of 
the tests, ranging from 0.9 to 1.1 milliseconds.  Typical uncertainty in this measurement 
is 0.1 millisecond.   
Of interest, it should be noted that the tests corresponding to similar velocities had 
nearly identical profiles in both space and time during the first 15 to 20 cm of the 
combustor, regardless of boron concentration, equivalence ratio, or temperature.  This 
implies that the pilot flame is once again the dominant source of ignition of the particles.  
The major difference in the tests having the same inlet velocity came further downstream 
in the combustor (or later in time).  The higher temperature tests (TC1b and TC4b) 
exhibit BO2 emission for longer durations than the tests of the same velocity but lower 
temperature (TC2b and TC3b).  This suggests an increase in the oxidation of boron with 
increased temperature, which is not in itself a surprising result; it does however begin to 
provide boundaries for efficient combustor operation.  In addition, the peak intensity 
values while consistent for a specific inlet velocity do not follow an increasing trend as 
might be expected with increasing temperature.  This implies that inlet velocity, or more 
 166
specifically, residence time, may be as important, or perhaps more important than 
temperature alone. 
 In some test conditions (1b) the intensity level did not reach background levels 
even at the end of the combustor.  Figures 6.12 and 6.13 imply that the boron may 
actually extinguish in some or all of the conditions.  In Figure 6.12, the lower temperature 
cases reach the background intensity before the end of the combustor whereas the higher 
temperature cases still show significant intensity levels.  Similarly, in the time analog 
(figure 6.13) the lower temperature conditions reach background levels in a shorter 
duration.  This result indicates that the particles are potentially extinguishing, at least in 
the lower temperature conditions.  In fact King92 showed that temperature boundaries for 
boron particle extinguishment are a strong function of pressure and oxygen concentration.  
King91 also showed that smaller particles required higher temperatures to prevent 
extinguishments.  These tests were conducted at essentially at atmospheric pressure 
requiring higher temperatures to ensure that the particles do not extinguish.    
 167
 
 
Fig. 6.11 BO2 Chemiluminescence Contours (C2H4 only, 1b, 2b, 3b, 4b, 5b) 
 
Tf = 2090 K 
V = 40 m/s 
Tf = 1850 K 
V = 40 m/s 
Tf = 1712 K 
V = 55 m/s 
Tf = 2065 K 
V = 55 m/s 
Tf = 1746 K 
V = 70 m/s 
C2H4 Only 
 168
 
Fig. 6.12 BO2 Intensity Profiles Versus Combustor Length 
 
 
Fig. 6.13 BO2 Intensity Profiles Versus Time 
 
 
 
 
 169
6.1.2.3 Chemiluminescence Summary 
The presence of BO2 and AlO clearly indicate that the particles were ignited and 
oxidation to some extent took place.  Several observations can be made about the 
differences between the BO2 and AlO contours.  First, it is obvious that the majority of 
emission of BO2 took place over a longer portion of the combustor than that of AlO under 
all conditions.  Another observation is that the peak areas of AlO emission were clearly 
closer to the pilot flame ignition source (X=0) than the BO2 emission, suggesting larger 
ignition delays for the boron particles. 
The average intensity over the entire image of a given test condition is shown as a 
function of adiabatic flame temperature in figure 6.14, which shows all of the data for the 
test conditions summarized in Table 6.1.  The aluminized data shows a clear trend for 
each aluminum mass flow rate of increasing average intensity with increasing flame 
temperature as would be expected, since the intensity is heavily dependent on flame 
temperature.  However, the data collected in the boron loaded tests have no real trend to 
speak of, suggesting that temperature is not the lone influence on chemilumiscent 
intensity in these tests.   
Figure 6.15 shows the product of average intensity and inlet velocity as a function 
of adiabatic flame temperature.  Using this parameter, the boron-loaded tests now begin 
to show a smoother trend, suggesting that residence time as well as temperature is 
affecting the chemiluminescent intensity. The trend also shows that there is some 
temperature threshold that needs to be attained before significant amounts of boron 
combustion occurs.  This threshold appears to be for experiments in which the adiabatic 
flame temperature exceeds 1800 K.  The data obtained and presented in Chapter 5 shows 
 170
that there was a significant temperature dependence on burning time, particularly in the 
t1, ignition stage, of boron combustion.  The total burning times for XO2 = 0.15 at about 
1700 K in the flat flame burning experiments was about 6 milliseconds, and was reduced 
to about 3.5 milliseconds at about 1800 K.  For the two highest speed conditions, this 
would imply that a minimum temperature of about 1800 K would be required in order to 
possibly complete the boron particle combustion. At 1700 K, for XO2 = 0.1 t2 combustion 
was not even achieved in the flat flame burner which has a typical residence time of 
about 12 milliseconds.  As table 6.2 shows, the typical residence time in the airbreathing 
combustor testing was a maximum of 10 milliseconds, which implies that t2 combustion 
would not be reached for test conditions at or below 1712 K and XO2 = 0.1.  
The aluminized tests do not show an obvious trend with the product of velocity 
and intensity, implying that residence time is not affecting the chemiluminescent 
intensity.   The resulting conclusion is that under these conditions the ?same? amount of 
aluminum was being consumed in all of test conditions. 
 
 171
 
Fig. 6.14 Average Intensity as a Function of Adiabatic Flame Temperature 
 
Fig. 6.15 Intensity-Velocity Coupling as a Function of Adiabatic Flame Temperature 
 172
6.1.3 Temperature Measurements 
Combustion gas temperature measurements were made using thermocouples at 
approximately the exit plane of the combustor.  In total, four thermocouples were placed 
at the exit plane of the combustor to form a ?rake?, which can provide temperature 
profiles for different test conditions.  In general the two outermost thermocouples were 
K-types (Chromel-Alumel) with bead diameters of approximately 1000 ?m, while the 
inner thermocouples were B-types (Pt-6%Rh/Pt-30%Ph) with bead diameters of 
approximately 200 ?m.  
Figure 6.16 shows a typical set of thermocouple traces prior to radiation 
correction.  For analysis purposes, a period of 0.2 seconds near the end of the test was 
averaged and then corrected for radiation to provide temperature profiles for each 
condition.  Figure 6.20 shows some typical profiles from aluminized tests.  In general, the 
tests were very repeatable.   
 
 173
0
500
1000
1500
2000
2500
0 0.5 1 1.5 2 2.5 3 3.5
H = 0.318 cm
H = 2.011 cm
H = 3.704 cm
H = 5.397 cm
Te
mp
era
tur
e (
K)
Time (s)
Average Temperature Taken 
for Combustor Profiles
 
Fig. 6.16 Typical Thermocouple Trace (Uncorrected for Radiation) from Test 
Condition 1b (Inlet Velocity = 40 m/s, wt% B = 12.7, ? = 0.69, Tad = 2090 K) 
 
Fig. 6.17 Combustor Temperature Profiles for Selected Ethylene/ALEX Tests 
 174
Figures 6.18-6.20 show the results of the boron loaded experiments and their 
ethylene only analogs.  Figure 6.18 shows the lowest speed tests (40 m/s inlet velocity), 
TC1b and TC2b, which had adiabatic flame temperatures of 2090 K and 1850 K 
respectively.  For TC1b, the highest peak temperature measured was 2086 K, while for 
TC2b the highest peak temperature measured was 1820 K.  For the most part, at each 
location, the boron loaded experiments had temperatures meeting or exceeding those of 
the ethylene only analogs.   
Similarly figure 6.19 shows the intermediate speed cases (55 m/s inlet velocity), 
TC3b and TC4b, which had adiabatic flame temperatures of 1712 K and 2065 K 
respectively.  The highest measured temperature for TC3b was 1592 K while the peak 
measured temperature for TC4b was 2069 K.  Since the combustible mixture was not 
entirely uniform, it is possible that the highest measured temperature could exceed the 
adiabatic flame temperature.  Interestingly for TC3b, the temperatures approximately 
meet or are significantly lower than its ethylene only analog.  In particular, the top 
portion of the combustor had measured temperatures that are significantly lower than the 
ethylene only analog.  Conversely, for TC4b, in most cases the temperatures 
approximately met or exceeded the ethylene only analog.   
Finally, figure 6.20 shows the result of the highest speed (70 m/s inlet velocity), 
TC5b, experiment whose nominal adiabatic flame temperature was 1746 K.  The highest 
measured temperature for this condition was 1688 K.  Once again, in this experiment, the 
measured temperatures were at or below the ethylene analog.   
In summary, experiments with nominal adiabatic flame temperatures of 1850 K or 
greater all showed increased combustor temperature than its ethylene analog.  On the 
 175
other hand, experiments with adiabatic flame temperatures below 1850 K show either no 
temperature improvement, or in some cases a temperature reduction.  This is not 
surprising in light of the results of the flat flame burning experiments described in 
Chapter 5, which showed that for XO2 = 0.1 second stage combustion was not even 
achieved until temperatures approaching 1800 K were reached for residence times of 
about 12 milliseconds.       
 
0
1
2
3
4
5
6
600 800 1000 1200 1400 1600 1800 2000 2200
Test Condition 1b
Test Condition 1
Test Condition 2b
Test Condition 2
Co
mb
us
tor
 H
eig
ht 
(cm
)
Temperature (K)  
Fig. 6.18 Temperature Profiles for Test Conditions 1, 1b, 2, 2b (Inlet Velocity = 40 
m/s) 
 
 176
0
1
2
3
4
5
6
600 800 1000 1200 1400 1600 1800 2000 2200
Test Condition 3b
Test Condition 3
Test Condition 4b
Test Condition 4
Co
mb
us
tor
 H
eig
ht 
(cm
)
Temperature (K)  
Fig. 6.19 Temperature Profiles for Test Conditions 3, 3b, 4, 4b (Inlet Velocity = 55 
m/s) 
 
 
0
1
2
3
4
5
6
600 800 1000 1200 1400 1600 1800
Test Condition 5b
Test Condition 5
Co
mb
us
tor
 H
eig
ht 
(cm
)
Temperature (K)  
Fig. 6.20 Temperature Profiles for Test Conditions 5, 5b (Inlet Velocity = 70 m/s) 
 
 
 177
Figures 6.21-6.23 show the aluminized experiments and their ethylene only 
analogs.  Figure 6.21 shows the results from the lowest speed case (40 m/s inlet velocity), 
which had an adiabatic flame temperature of 2120 K.  The top three thermocouple 
locations show that the aluminized version provide similar or lower temperatures that the 
ethylene analog.  However, the thermocouple located at the bottom of the combustor 
shows a dramatic increase in temperature ~350K.  A similar result occurs for the 
intermediate (55 m/s inlet velocity) cases as seen in figure 6.22.  For these cases the 
aluminized version has a temperature from 200 K to 400 K higher than the ethylene 
analog at the lowest thermocouple location.  This result also occurs for the highest speed 
(70 m/s) case, TC10a as seen in figure 6.23. 
  This behavior can have several explanations, or a combination of explanations.  
First, the Mie scattering data provided in section 6.1.1 showed that although particle 
dispersion was fairly uniform, it did favor the bottom of the combustor.  Further, the 
shape of the temperature profiles, namely the two innermost thermocouples, also 
indicates that the mixture is not entirely uniform.  Otherwise, one would expect the two 
inner thermocouples to have approximately the same temperature measurements.  If more 
aluminum were located in the bottom portion of the combustor than the top, then the 
equivalence ratio would be even more non-uniform than in the ethylene only analogs.  In 
addition, the equivalence ratio would be higher in a relative sense than it would for the 
ethylene only analogs.  This behavior is also seen in the boron-loaded tests, but not nearly 
as drastic.  This is due to the fact that the nanoaluminum ignition temperature 
requirements are not nearly as high as the nano boron, therefore the boron along the ?cool 
 178
wall area? does not necessarily achieve vigorous burning to the degree that the 
nanoaluminum would. 
0
1
2
3
4
5
6
1300 1400 1500 1600 1700 1800 1900 2000 2100
Test Condition 6a
Test Condition 6
Co
mb
us
tor
 H
eig
ht 
(cm
)
Temperature (K)  
Fig. 6.21 Temperature Profiles for Test Conditions 6, 6a (Inlet Velocity = 40 m/s) 
0
1
2
3
4
5
6
600 800 1000 1200 1400 1600 1800 2000 2200
Test Condition 7a
Test Condition 7
Test Condition 8a
Test Condition 8
Test Condition 9a
Test Condition 9
Co
mb
us
tor
 H
eig
ht 
(cm
)
Temperature (K)  
Fig. 6.22 Temperature Profiles for Test Conditions 7, 7a, 8, 8a, 9, 9a (Inlet Velocity 
= 55 m/s) 
 
 179
0
1
2
3
4
5
6
500 1000 1500 2000 2500
Test Condition 10a
Test Condition 10
Co
mb
us
tor
 H
eig
ht 
(cm
)
Temperature (K)  
Fig. 6.23 Temperature Profile for Test Conditions 10, 10a (Inlet Velocity = 70 m/s) 
 
 
When comparing the aluminized tests, it can be seen that the basic shape of the 
temperature profile was similar for tests 6a and 7a, but changes and was then the same for 
8a ? 10a.  In test condition 6a and 7a the highest measured temperature comes from the 
second thermocouple from the top of the combustor, while in the other test conditions, 
the highest measured temperature shifts to the thermocouple located at the second from 
the bottom.  This was a result of the increased mass flow rate of the seeder carrier gas and 
total seeder flow rate.  For conditions 6a and 7a the carrier gas flow rate was 1 g/s and it 
was increased to 1.5 g/s for the other conditions.  The increased flow rate of the carrier 
gas pushed more fuel toward the bottom of the combustor.  This resulted in a shift in the 
peak of the local equivalence ratio.  
 180
In order to determine the effectiveness of the metal additives in the thermal 
output, the metallized cases were compared directly to hydrocarbon only cases.  In this 
comparison, the ethylene mass flow rate was held approximately constant and the only 
difference in the test is the addition of the metal additive.  The temperature at all four 
locations was averaged for a given test and normalized by their ethylene only analog. The 
ratio of adiabatic flame temperatures of metallized to ethylene only tests were also 
plotted against the adiabatic flame temperature of the metallized tests to give a boundary 
for the maximum attainable contribution from the metal addition.  This can be seen in 
figures 6.24 and 6.25.  
As the earlier inspection of the temperature profiles showed for the boron loaded 
experiments, the only experiments which showed positive contribution from boron 
addition were the three highest temperature cases, 1b, 2b and 4b.  This can be seen more 
clearly in figure 6.24.  Conditions 1b and 4b both have adiabatic flame temperatures 
calculated using the NASA CEA 200084 code, greater than 2000 K.  These tests were the 
only ones conducted either on the upper end or well above the critical ignition 
temperature (1500-1950 K) given by other researchers47,90,91.  Test 3b and 5b show that 
boron addition is not beneficial under those conditions.  The BO2 emission data does 
prove that boron ignition started, but it was possible that the temperatures were not high 
enough to promote sustained combustion or possibly start stage two combustion.  So 
while ignition was achieved in all cases, it was possible that sustained or complete 
combustion of boron was never achieved in all of the conditions tested.  Figure 6.24 
provides further evidence of a critical temperature for sustained combustion of the boron 
particles.  All the tests with measured peak temperatures below ~1700 K (adiabatic flame 
 181
temperature less than 1750 K) showed no positive contribution for boron addition, while 
all tests with measured peak temperatures above ~1770 K (adiabatic flame temperature 
greater than 1850 K) showed positive contributions from boron combustion.  
In the case of aluminum addition (figure 6.25), all of the tests conducted showed a 
positive contribution from aluminum to the thermal output.  This was primarily a result of 
the large temperature improvement on the lower portion of the combustor.  The results 
indicate that efficient combustion even at very fuel lean conditions (? ~ 0.5) was likely 
to be occurring, which is important in achieving maximum specific impulse.  
 
0.9
0.95
1
1.05
1.1
1600 1700 1800 1900 2000 2100 2200
Tave_boron/Tave_ethylene Adiabatic
Tave_boron/Tave_ethylene Test Data
T a
ve
_b
oro
n/
T a
ve
_e
thy
len
e 
Adiabatic Flame Temperature of Boron Loaded Tests (K)  
Fig. 6.24 Comparison of Ethylene and Boron Loaded Tests 
 
 
 182
0.9
0.95
1
1.05
1.1
1600 1700 1800 1900 2000 2100 2200
Tave_aluminum/Tave_ethylene Adiabatic
Tave_aluminum/Tave_ethylene Test Data
T a
ve
_a
lum
inu
m
/T a
ve
_e
thy
len
e
Adiabatic Flame Temperature of Aluminized Tests (K)  
Fig. 6.25 Comparison of Ethylene and Aluminized Tests 
 
 
 In order to get a sense for how well the fuels were performing compared to their 
theoretical thermal output a thermal ratio was defined as: 
? ?
? ?
T T
T T
ave metal ave ethylene Test Data
adiabatic metal adiabatic ethylene
_ _ _
_ _
   Eq. 6.2 
where the average temperatures have been normalized by the ratio of the adiabatic flame 
temperatures.  The results are presented in figure 6.26, which shows that all of the 
aluminized tests come very close to theoretical performance limitations.  Therefore at 
these conditions, temperature, time, and gas composition have no significant effect on the 
performance of the fuels loaded with aluminum nanoparticles.   
 183
Figure 6.26 shows that the boron performance improves as the temperature is 
increased, in particular there is a large jump in performance above 1800 K.  However, 
and perhaps more telling is that the two conditions showing the highest thermal ratios, 
1850 K and 2090 K, come from the two lowest speed conditions (40 m/s inlet velocity).  
At 2065 K (55 m/s inlet velocity) one data point is in the range of the lower speed cases, 
however, there are more data points that are several percent lower.  On the other hand, 
the other data points for the 55 m/s inlet condition (1712 K) are significantly lower than 
the 2065 K condition.   
The fact that the lower temperature, lower velocity data (1850 K and 40 m/s) was 
comparable or better than the higher temperature data and higher velocity data (2065 K 
and 55 m/s) indicates that time and oxygen mole fraction are the most important 
parameters in determining performance.  At the same time the data points for the 55 m/s 
inlet condition (1712 K and 2065 K) show that even at lower oxygen concentration, the 
higher temperature condition had better performance, thus suggesting that time is the 
most dominant parameter when overall performance is considered for boron in this 
combustor.   
 184
0.9
0.92
0.94
0.96
0.98
1
1700 1800 1900 2000 2100 2200
Aluminum
Boron
?
Adiabatic Flame Temperature (K)  
Fig. 6.26 Thermal Ratios 
 
6.2 Data Interpretation 
One possible interpretation of Figures 6.12 and 6.13 is that the boron may actually 
extinguish within the combustor for some of the test conditions.  In figure 6.12, the lower 
temperature cases reach the background intensity before the end of the combustor 
whereas the higher temperature cases still show significant intensity levels.  Similarly, in 
the time analog (figure 6.13) the lower temperature conditions reach background levels in 
a shorter duration.  This is in direct contrast to what was expected, namely that at higher 
temperatures, the particles should combust quicker.  This result would indicate that the 
particles are likely extinguishing, at least in the lower temperature conditions.  In fact 
King92 showed that temperature boundaries for boron particle extinguishment are a strong 
function of pressure and oxygen concentration.  King92 also showed that smaller particles 
 185
required higher temperatures to prevent extinguishments.  Essentially once the particles 
become very small they may not be able to sustain their own temperature field as they 
begin to lose more heat to the surroundings than they can generate themselves.  Li and 
Williams87 performed an analysis to examine the extinction characteristics of boron 
particles in which they applied an energy balance between the rate of heat loss of the 
particle and the rate of heat generation by the particle.  The rate of heat generation is 
balanced by the heat loss due to radiation and convection to the surrounding as described 
by Equation 6.2 (using Li and Williams?87 notation).   
       ? ? ? ?Q
T T
d T T
q
q
q5 5
4 42? ? ??? ? ? ??
?     Eq. 6.2 
Here, Q5, ?5, ?, Tq, T?, dq, ?, ?, are the heat released by the surface reaction given in 
equation 6.2, the kinetic rate constant, the thermal conductivity, the quenching 
temperature, the ambient temperature, the quenching diameter, the emissivity of boron, 
and the Boltzman constant respectively.  The heat released is a result of the surface 
reaction described by equation 6.3. 
    2B(s) + O2(g) ? B2O2    Eq. 6.3 
With some manipulation and substitution of parameters provided in reference 87 
including the selection of the melting temperature of pure boron as the particle surface 
quenching temperature, Tq, they arrived at an equation relating the particle-quenching 
diameter, dq, to the surrounding temperature T?, and the oxygen partial pressure, pO2. 
? ?
? ?d
T T
Q T Tq
q
q
?
?
? ?
?
?
2
5 5
4 4
?
? ??
           Eq. 6.5 
Where ?5 is related to the partial pressure of oxygen by a constant: 
 186
?5
2
? cpO      Eq. 6.6   
Applying this result to the current set of test conditions (1 atm); assuming that the 
hydrocarbon is completely consumed before the boron, the oxygen mole fraction after 
complete combustion of the hydrocarbon is typically near 0.1 figure 6.27 shows that 
particles under the size of 11 ?m on the lower temperature tests and 6 ?m in the higher 
temperature tests would extinguish.  Clearly the primary particle sizes used in this study 
are much smaller than the predicted quenching diameter.  Even considering extreme 
amounts of agglomeration, this analysis would indicate that particle extinguishments is 
quite likely under the test conditions in this study.  The other curves generated in figure 
16 consider higher combustor pressures but still consider the mole fraction of oxygen to 
be 0.1.  Although the figure shows that increasing pressure reduces the quenching 
diameter, the greatest reduction in quenching diameter occurs once the surrounding 
temperature exceeds 2000 K.  
 187
 
Fig. 6.27 Boron Particle Quenching 
 
Under most practical situations, if boron were to be used as a fuel additive in an 
airbreathing combustor it would most likely be stored within a hydrocarbon, either gelled 
or as part of a solid fuel.  At atmospheric pressure most hydrocarbons have maximum 
flame temperatures between 2200-2400 K near stoichiometric conditions.  Since 
temperatures in this range are required by the analysis of Li and Williams87 to sustain 
combustion of small boron particles, the mole fraction of oxygen available for 
consumption of boron would be quite low.  However in a more realistic condition, the 
combustor pressure would be significantly higher than discussed here and the air 
temperature entering the combustor would be much higher.  As an example Table 6.4 
provides a brief summary of combustor conditions for a ramjet flying at an altitude of 
8,000 m (~60,000 ft) if the airflow is decelerated isentropically to a Mach number of 0.3 
 188
as it enters the combustor.  Ideally a ramjet follows a Brayton Cycle analysis, therefore 
the combustion process is considered isobaric.  The adiabatic flame temperatures were 
calculated using the NASA CEA 2000 computer code84, considering RP-1 as the 
hydrocarbon fuel, with combustor inlet conditions obtained from the isentropic flow 
calculations.  Table 6.4 shows that even at higher flight Mach Number conditions, a 
majority of the fuel would still need to be a hydrocarbon in order to be certain that the 
temperature of the surroundings were high enough to sustain combustion of very small 
boron particles according to the analysis of Li and Williams87.  At this point the mole 
fraction of oxygen remaining is quite small further complicating the problem. 
Although the analysis by Li and Williams87 suggests that particle extinction is 
possible, the results of Chapter 5 found no evidence of particle extinction.  However, 
only one data point for XO2 = 0.1 was found to produce stage two combustion of the 
boron particles, and that data point corresponded to a burner temperature of 1791 K.  For 
burner temperatures of 1614 K and 1712 K, stage two combustion was not achieved in 
the flat flame burner experiments.  During the airbreathing combustor experiments the 
only boron loaded experiments to show a positive thermal contribution were the 
experiments with measured peak temperatures greater than approximately 1780 K.  
Table 6.4: Flight Conditions 
Flight Mach 
Number 
Inlet Air Static 
Temperature (K) 
Inlet Air Static 
Pressure (atm) 
? for Tf = 2200 
or Tfmax (K) 
XO2 for Boron 
Consumption 
Quenching 
Diameter (?m) 
2 382 0.54 1.05, (2188 K) 0.0018 N/A 
3 594 2.56 0.87 0.0244 4.25 
4 891 10.57 0.72 0.0525 0.45 
 
 189
Therefore, another possible explanation of the results shown in figures 6.12 and 6.13 is 
that the particles never fully achieved stage two combustion under the lower temperature 
conditions.  In fact Macek47 found that even at very high temperatures and at low oxygen 
mole fraction, XO2 = 0.08, that it was impossible to distinguish between the ignition and 
full-fledged combustion stages for larger boron particles.   
The presence of water vapor is known to inhibit boron combustion through the 
formation of HBO2.  In two of the test conditions, 1b and 4b, water vapor has the highest 
molar content of any of the main oxidizing species remaining after hydrocarbon 
combustion (see table 6.3).  In the other cases water vapor is quite comparable to the 
oxygen mole fraction, which would serve to severely inhibit the performance of the boron 
loaded fuels should they actually achieve stage 2 combustion.  The competition between 
the available oxygen and water vapor leads to a condition in which one would expect 
relatively poor performance for the boron loaded fuels when oxygen and water vapor are 
comparable.   
 The data collected for the aluminum nanoparticles shows great potential for the 
use of the particles as fuels or supplements.  The chemiluminescence data showed that 
the aluminum completed combustion well within the combustor.  The temperature data 
also showed that under all conditions, the aluminum provided additional thermal energy 
to the flow. 
 
 6.3 Additional Considerations   
 As was pointed out in Chapter 4, unless specially treated aluminum and boron 
nanoparticles suffer from the formation of an oxide layer that essentially removes a 
 190
substantial part of the available energy of a particle.  For instance in the case of the SB99 
particles, only about 72% was active boron, while for the ALEX particles only about 82% 
was active aluminum.  The inactive content results in considerable performance loss from 
a theoretical standpoint.  As an example, by calculating the lower heating values of a 
50% metal/50% RP-1 mixture and comparing a pure metal to the real case of reduced 
active content one finds that the heating value of the aluminized mixture is reduced by 
about 8% and the boron loaded mixture is reduced about 17% on a mass basis.  In this 
case the heating value is calculated in the following manner: 
    ? ?H H H Hc R reac prod? ? ? ?     Eq. 6.7 
where: 
    H N hreac i i
reac
? ?   and H N hprod i i
prod
? ?   Eq. 6.8 
N refers to the number of moles of a reactant or product, hi is the enthalpy of formation 
and the subscripts reac and prod refer to the reactants and products of the reaction.  
Applying these parameters to an ideal ramjet case under the following assumptions: 
1) The compression and expansion of gases in the engine are both reversible and 
adiabatic 
2) The combustion process is represented by a constant-pressure heat and mass 
addition process 
the specific impulse and the density impulse were calculated by first specifying a flight 
Mach number.  Once the Mach number was specified, the specific impulse, or Isp was 
calculated by the following procedure: 
     U M RT? ?? ?     Eq. 6.9 
 191
where M is the flight Mach number, U? is the freestream velocity, ? is the ratio of 
specific heat of air, R is the gas constant for air, and T? is the freestream static 
temperature.  The freestream stagnation temperature To, ?, was calculate from: 
    T T Mo,? ?? ? ???? ???1 12 2?              Eq. 6.10 
With this information the gas velocity exiting the ramjet combustor, Ue, was calculated, 
where Tf was the adiabatic flame temperature of the combustion reactants in a constant 
pressure and enthalpy calculation, which was computed using the NASA CEA 2000 
Chemical Equilibrium computer code. 
     U U TTe f
o
? ?
?,
             Eq. 6.11 
The thrust per unit mass flow of air was calculated by: 
     ? ?Fm f U U
o
e? ? ? ? ?1              Eq. 6.12 
where, f was the fuel to air ratio.  In all of these calculations an equivalence ratio of one 
was assumed.  Finally the specific impulse was calculated by equation 6.13, where g was 
the acceleration due to gravity.  
     I gf Fmsp
o
? ??? ???1 ?              Eq. 6.13 
The results of this analysis can be seen in figure 6.28-6.31.  In the case of boron, the 
specific impulse (see figure 6.28) was typically about 12% lower for the fuels considering 
active content than for the theoretical values of the pure metal fuel.  Similarly the density 
impulse (see figure 6.29) suffers about a 12% loss compared to the pure metal fuel.  The 
 192
active content boron loaded fuel still however had density impulse about 20% greater 
than the hydrocarbon only fuel. 
 For the aluminized fuel, the specific impulse (figure 6.30) was about 5% lower 
than the pure metal fuel, while the density impulse was about 4% lower (figure 6.31) 
when active content was considered.  Interestingly, in the aluminized version, the density 
impulse recovers some of the performance loss due to the relatively high density of 
alumina, which is about 47% higher than that of pure aluminum.  While both metals still 
outperform the baseline hydrocarbon fuel, the analysis shows that the active content of 
the nanoparticles plays a large role in the final theoretical performance of the fuel and 
cannot be ignored.  As seen by equation 6.14, the Breguet Range equation, a loss in Isp 
results in a direct loss of vehicle range. 
     R LD mm I Usp? ln 1
2
             Eq. 6.14  
where L, D, m1, m2 refer to the lift, drag, the initial mass of the vehicle, and the final mass 
of the vehicle.  
 
 193
200
400
600
800
1000
1200
1400
1600
1 2 3 4 5 6 7 8 9
Isp RP-1
Isp 50%RP-1/50% B
Isp 50%RP-1/36% B/14% B2O3
I sp
 (s
ec
)
Flight Mach Number  
Fig. 6.28 Specific Impulse of Boron Loaded Fuels Considering Active Content 
 
200
400
600
800
1000
1200
1400
1600
1 2 3 4 5 6 7 8 9
?Isp RP-1
?Isp 50%RP-1/50% B
?Isp 50%RP-1/36% B/14% B2O3
?I s
p 
(s*
g/c
m3
)
Flight Mach Number  
Fig. 6.29 Density Impulse of Boron Loaded Fuels Considering Active Content 
 
 194
200
400
600
800
1000
1200
1400
1600
1 2 3 4 5 6 7 8 9
Isp RP-1
Isp 50%RP-1/50% Al
Isp 50%RP-1/41%Al/9%Al2O3
I sp
 (s
ec
)
Flight Mach Number  
Fig. 6.30 Specific Impulse of Aluminized Fuels Considering Active Content 
200
400
600
800
1000
1200
1400
1 2 3 4 5 6 7 8 9
?Isp RP-1
?Isp 50%RP-1/50% Al
?Isp 50%RP-1/41% Al/9% Al2O3
?I s
p 
(s*
g/c
m3
)
Flight Mach Number  
Fig. 6.31 Density Impulse of Aluminized Fuels Considering Active Content 
 
 195
CHAPTER 7 
 
CONCLUSIONS AND FUTURE WORK 
 
7. Conclusions and Future Work 
 Metals offer the potential of improving the performance of volume-limited 
airbreathing propulsion systems by increasing the energy density of the fuels utilized.  
Unfortunately due to small combustor residence times in high-speed applications, 
theoretical performance levels are difficult to achieve since burning times of metal 
particles are relatively long.  However, with the development of metallic nanoparticles 
such as aluminum and boron, new hope has been raised for the performance potential of 
metals as fuels in high-speed airbreathing combustors.   
This research effort was conducted in order to determine the viability of utilizing 
metallic nanoparticles as fuels or fuel supplements for high-speed airbreathing 
applications.  The study was broken into two main parts; the first was a fundamental 
investigation into the combustion behavior of boron nanoparticles in a controlled setting, 
while the second was a demonstration of the performance potential of the nanoparticles in 
a realistic airbreathing combustor configuration.  The following is a list of the major 
accomplishments achieved during this research effort. 
     
?? This study provided the first experimental data on boron nanoparticles burning 
times, which were obtained in the post-flame region of a flat flame burner.   
 196
?? The study extended the range of control parameters from previous ignition time 
studies87 involving nano-sized boron particles. 
?? The ignition and burning time data for the boron nanoparticles was obtained in a 
transverse flow configuration.  Previous studies utilizing the post flame region of 
a premixed burner injected either micron or nano-sized particles axially. 
?? Examined the dependence of metallic nanoparticles? combustion behavior on 
surrounding temperature and oxygen concentration. 
?? Compared the experimental trend of boron nanoparticles combustion with respect 
to both kinetic and diffusion limited theoretical behavior. 
?? Provided nanoparticle combustion data in a controlled open flame environment, 
as well as in an enclosed combustor configuration that simulated the reacting 
flowfield in a high-speed airbreathing propulsion system. 
?? Demonstrated the potential of metallic nanoparticles for enhancing propulsion 
performance in volume-limited systems. 
 
The subsequent sections provide a more detailed summary of the individual 
accomplishments achieved in each portion of the study, as well as direction for future 
studies.    
 
7.1 Particle Characterization 
?? Two different types of nanoparticles were investigated during this study, SB99 
boron nanoparticles, and 50 nm ALEX nanoaluminum.  Scanning Electron 
Microscope (SEM) images revealed that the SB99 particles had primary particles 
 197
sizes of about 50 nm, but were highly agglomerated.  SEM images of the ALEX 
particles showed a wide size distribution of spherical particles ranging from about 
50 nm up to about 200 nm.  The ALEX particles were also highly agglomerated. 
?? A Thermal Gravimetric Analysis (TGA) was performed on the particles to 
determine the active content of the particles.  The SB99 particles had an active 
content of 72% by mass, while the ALEX particles had an active content of 82% 
by mass.  In each case, the remaining content was assumed to be the particle 
oxides B2O3 and Al2O3 respectively.   
 
7.2 Flat Flame Burning Experiment 
?? A new experiment was developed as a means of determining the ignition and 
burning times of boron nanoparticles.  A standard McKenna flat flame burner was 
used for the study with a premixed mixture of CH4/Air/O2.  This experiment 
differs from previous particle burning experiments16,35,36,3744,46,47 in the post flame 
region of a flat flame burner, in that the particles were injected across the cross 
section of the flame rather than in the direction of the gas flow.  This technique is 
advantages because the gas properties in the transverse direction are more 
uniform than in the axial direction.  The results is that the particles are subjected 
to relatively constant gas properties. 
?? Since observation of a single nanoparticle is impossible at the moment, this 
technique allowed for the measurement of the ensemble average of the ignition 
and burning times of the particles. 
 198
?? Temperature measurements were made for a variety of burner conditions and a 
range of product species compositions including oxygen mole fraction ranging 
from 0.1-0.3.  Average measured temperatures ranged from 1578 ? 31 ? 1872 ? 
47 K.  The burner temperature was relatively insensitive to equivalence ratio, but 
was influenced more by total flow rate due to heat losses to the burner hardware. 
?? Particle Image Velocimetry (PIV) experiments revealed that the centerline 
particle velocity follows laminar flow theory for a round jet very closely.  For the 
PIV experiments, 5 ?m water droplets from a sonicating atomizer were used as 
the seed particles.  Three different injection flow rates were studied with twelve 
different crossflows and correlations for the centerline velocity were developed 
for all flow rates. 
?? The boron nanoparticles were injected into the post flame region of the flat flame 
burner via a sonicating atomizer.  ?As injected? particles were collected for 
Transimission Electron Microscopy (TEM) imaging and revealed that the 
particles were entering the post flame region as large non-spherical agglomerates 
of smaller primary particles.   
?? The ?as injected? particles were passed through a Differential Mobility Analyzer 
(DMA), which revealed that the particles had a size distribution ranging from 50 
nm to 400 nm in mobility diameter, with a mean mobility diameter of 200 nm.     
  
7.2.1 Ignition and Burning Time Conclusions 
?? Under appropriate temperature and oxygen mole fraction, XO2, conditions a two-
staged combustion phenomenon was observed for the SB99 particles.  Just after 
 199
injection there was no visible signature of combustion.  After some delay, which 
was dependent on the test condition, a yellow/orange glow became visible and 
was attributed to the first stage, t1, of boron combustion.  The first stage of boron 
combustion was called the ignition stage, because according to reference 44 this 
stage is related the complete removal of the oxide (B2O3) layer.  Following the 
first stage, a bright white glow zone was observed and attributed to the second 
stage, t2, of boron combustion.  The second stage combustion was considered full-
fledged combustion of the ?clean? pure boron particle.   
 
7.2.1.1 Stage 1 Combustion of SB99  
?? The ignition stage time of nano boron combustion, t1, was determined for a wide 
range of parameters, such as oxygen mole fraction XO2 ranging from 0.1-0.3, and 
temperatures ranging from 1580 K to 1810 K.   
?? Stage 1 combustion was found to be a strong function of temperature, but 
relatively insensitive to oxygen mole fraction in the ranges studied.   
?? Stage 1 burning times ranged from 1.5 milliseconds at the highest temperatures, to 
about 6 milliseconds at the lowest temperatures.   
?? When compared to available data from other researchers46 employing micron-
sized particles, the ignition stage of the boron nanoparticles was not substantially 
different from particles of approximately 2-3 ?m in size. 
 
 
 
 200
7.2.1.2 Stage 2 Combustion of SB99 
?? Stage 2 combustion of boron nanoparticles was studied for 0.1 ? XO2 ? 0.3, and 
temperatures ranging from 1580 K to 1870 K.   
?? At the lowest temperatures, the oxygen mole fraction played a large role in the t2 
burning time and affected the ability of the particles to achieve stage 2 
combustion.   
?? For XO2 ? 0.2 stage 2 combustion was achieved for temperatures as low as 1578 ? 
31  K.   
?? At the lowest temperature condition, and XO2 = 0.15, stage two combustion was 
achieved but was not completed while the particles remained in the post flame 
region of the burner.   
?? For XO2 = 0.1, and the two lowest temperatures conditions, second stage 
combustion was not achieved while the particles were in the post flame region.   
?? t2 burning times ranged from about 3 milliseconds to about 1.5 milliseconds over 
the test conditions. 
o As a comparison, Parr16 measured burning times of 24 nm aluminum in 
water vapor that were between 2 and 3 times faster than the measured 
burning times of boron nanoparticles in this study. 
??A direct comparison is impossible because of the different 
oxidizing species. 
?? Comparison of the available data for larger particles showed that 
aluminum burned about 4 times faster than boron under similar 
conditions. 
 201
?? Comparison with other researchers data revealed that the t2 burning times did not 
follow a diffusion limited D2-law or a kinetic limited D1-law.  Instead the t2 
burning time was found to be far less dependent on particle size than observations 
for larger particles.  In fact the dependence on sized was t2 ~ D0.2.   
?? An Arrhenius Burning Rate law was obtained which showed that the measured 
activation energy of the particles was consistent with other data46 employing 
micron-sized particles.  From the Arrhenius parameters, a burning time 
correlation based on a kinetically limited process was developed which resulted in 
showing a size dependence of t2 ~ D0.15.  A similar result was found by other 
researchers88 when correlating nanoaluminum burning time data. 
   
7.3 Airbreathing Combustor Conclusions 
An optically accessible airbreathing combustor was designed and fabricated to 
investigate the possibility of utilizing nanoparticles as fuel additives in a high-speed 
airbreathing propulsion system.  The combustor is fully capable of utilizing a wide 
variety of diagnostics.  In addition to the combustor a reverse cyclone particle seeder was 
also designed, fabricated, and characterized in order to provide a simple, reliable, and 
reusable means of delivering particles into the combustor.  
Gaseous ethylene was selected as the baseline fuel for comparison with metallized 
fuels.  The combustor was characterized to understand its behavior prior to utilization.  
The blowoff limits were obtained for combustor inlet velocities ranging from 25 m/s to 
70 m/s.  In addition, the pilot flame trajectory was analyzed as a function of inlet 
velocity, and a correlation was developed to describe its behavior. 
 202
Once combustor characterization was completed, testing commenced in which the 
performance of a hydrocarbon (ethylene) fuel with the addition of a metal nanoparticle, 
either aluminum or boron, was compared directly to an ethylene only analog.  Testing 
proceeded under three different combustor inlet velocities, 40, 55, and 70 m/s and 
equivalence ratios of approximately 0.5 ? 0.7.  The active content of the metals made up 
from about 7% by weight to 15% by weight of the total fuel. 
Planar Mie scattering was employed to study the particle dispersion patterns within 
the combustor.  In general, the dispersion was fairly uniform with a slight bias towards 
the bottom of the combustor. 
 
7.3.1 Boron Combustor Conclusions 
?? BO2 emission was monitored with the use of an ICCD camera and a 546 nm 
bandpass filter and provided a measure of ignition delay for the particles.  The 
emission data indicated that ignition of the particles was clearly achieved, with 
delays ranging from 0.9 to 1.1 milliseconds.  Ignition delay was independent of 
combustor test condition, indicating that combustor flame temperature and 
chemical species played no role under the conditions tested.  The BO2 emission 
data also implied that a critical temperature existed for sustained combustion of 
boron.  This was further verified by temperature measurements. 
o The BO2 emission data showed that the product of the average intensity 
and the inlet velocity correlates well with flame temperature suggesting 
that residence time was as important of a parameter as temperature alone. 
 203
?? Combustor temperature measurements indicated that the addition of boron only 
added to the combustor heat release in the highest temperature cases.  In fact, in 
some cases measured temperatures were lower than the hydrocarbon analog tests.  
The experiments which achieved positive heat addition from the boron particles 
were the only tests where the flame temperatures were either on the upper end or 
well above the critical ignition temperatures (1500 ? 1950 K) that were defined by 
other researchers47,89,90.  The other tests were all well within the range of the 
defined ignition temperatures but the ethylene analog tests showed that the 
majority of the combustor temperature profile was either below the ignition 
temperatures or on the lower end.   
?? Although the BO2 emission data clearly shows ignition of the boron particles, it is 
clear that the boron was not completely reacted.  A critical peak combustor 
temperature for sustained boron combustion was identified.  For all tests with 
measured peak temperatures below 1700 K (adiabatic flame temperature less than 
1750 K) no positive contribution was achieved, while tests with peak 
temperatures above 1770 K (adiabatic flame temperature greater than 1850 K) all 
exhibited beneficial contributions from the addition of boron.  In terms of the 
definition of a thermal ratio, only tests with flame temperatures above 1800 K 
achieved ratios above 96%, with some as high as 99%.  Tests with flame 
temperatures below 1800 K had ratios near 0.91. 
o The highest thermal ratios for the boron testing were obtained for the two 
lowest inlet velocity conditions (40 m/s), with the higher temperature 
condition having the best thermal ratio. 
 204
?? For similar test conditions, i.e. temperature and oxygen concentration, the flat 
flame burning experiments indicated that second stage combustion may not be 
achieved in the residence time for the two highest inlet velocity conditions of the 
combustor.   
?? At low oxygen mole fractions that would be prevalent in a hydrocarbon assisted 
airbreathing combustion, a critical temperature of about 1800 K exists for 
reasonably efficient boron combustion.   
o The airbreathing combustor experiments with XO2 ~ 0.1 showed no 
improvement in thermal output for temperatures with peak temperatures 
below about 1800 K 
o Similarly, the flat flame burner experiments showed that for XO2 = 0.1 
second stage combustion was not even achieved until average burner 
temperatures reached almost 1800 K.  
 
7.3.2 ALEX Combustor Conclusions 
?? Testing with ?50 nm? ALEX nanoaluminum particles showed great promise for 
air-breathing applications.   
?? AlO emission detected with an ICCD camera and a bandpass filter centered at 488 
nm indicated that the majority of the aluminum oxidation took place within the 
combustor under all conditions.  The aluminum was consumed within the first 3 
milliseconds of residence time in the combustor. 
o The average AlO intensity increased with increasing temperature alone, 
which would be expected for an event that is not dependent on time. 
 205
??Indicates aluminum is being fully consumed.    
?? Ignition delays were about one third that of boron, typically being 0.3 
milliseconds.   
?? Temperature measurements indicated that positive thermal output was achieved 
under all conditions with thermal ratios in the 99% range even at low equivalence 
ratios (~0.5).  This is especially promising, since efficient operation at fuel lean 
conditions would improve specific impulse.        
 
7.4 Overall Study Conclusions 
?? This study, as well as others16,88 showed that nanoparticles burning times are 
indeed lower than micron-sized particles of similar materials as would be 
expected.  However, the size dependence is not nearly as great as would be 
predicted by traditional diffusion or kinetic limited theories.  The results of these 
studies show that tb ~ D0.2-0.3.  This result combined with the effects of active 
content certainly call into question the benefit of using nanoparticles as fuel 
additives in high-speed airbreathing propulsion. 
o However, any burning time improvement is certainly beneficial for time-
limited systems. 
o Active content may potentially be addressed through improved processing 
and coating of the materials in order to keep them from forming an oxide 
layer.   
?? The results of this study did not show any effect of increased reactivity of nano-
sized boron compared to micron-sized boron.  For example, ignition temperatures 
 206
were well within range of that demonstrated for larger particles.  In addition, the 
measured activation energy was quite comparable to that of larger particles. 
?? Only a small window exists for efficient combustion of boron nanoparticles in 
high-speed airbreathing applications since both temperature and oxygen mole 
fraction are important parameters in the combustion behavior of boron 
nanoparticles. 
?? Thermochemical calculations showed that fuel performance improved with 
increasing boron content.  However, boron requires significant thermal assistance 
in reaching its full-fledged combustion stage.  Therefore, in a practical system 
utilizing a hydrocarbon as a matrix for loading boron, the hydrocarbon being 
utilized must supply enough heat to trigger full-fledged combustion of boron.  
This fact limits the amount of boron that can realistically be used as a fuel 
additive. 
o This limit however, is dependent on a number of variables including flight 
conditions. 
o This limit can also be improved through fuel formulation.  For instance, 
coating the boron particles with an ignition accelerant such as magnesium 
can help supply much of the ignition requirements.  
?? The results of this study indicate that aluminum nanoparticles show great promise 
for use in high-speed airbreathing applications. 
 
 
 
 207
7.5 Benefits of this Study to the Scientific Community 
?? This study provided the first available data on burning times of boron 
nanoparticles.  It also extended the range of control parameters for the ignition 
time data available for boron nanoparticles. 
?? This study was the first to examine the possibility of utilizing both aluminum and 
boron nanoparticles in a traditional ramjet and began to provide boundaries for 
efficient combustion of boron nanoparticles as fuel supplements in airbreathing 
propulsion systems. 
?? This study showed that aluminum nanoparticles have great promise for use in 
high-speed airbreathing engines and merits consideration for future studies and 
implementation in volume limited systems.  
 
7.6 Recommendations for Future Work 
?? Study the effect of particle size on burning times of both boron and aluminum 
nanoparticles.  As this study and another study89 found, the size dependence on 
burning time cannot be explained by current diffusion limited (D2-law) or kinetic 
limited (D1-law) theories. 
?? Study the effect of pressure on the burning properties of boron nanoparticles.  All 
testing under this study was conducted at atmospheric pressure.  In a realistic 
propulsion system, combustion of the particles would take place under significant 
combustor pressure.  Both diffusion limited and kinetic limited theories suggest 
that an increase in pressure could lead to significant reductions in burning times. 
 208
?? Study the potential of boron nanoparticles in a ducted rocket environment.  This 
study showed that the combustion behavior of boron nanoparticles is very 
sensitive to both temperature and oxygen mole fraction concentration.  In a 
traditional ramjet, the available mole fraction of oxygen is limited by the amount 
of oxygen in the atmosphere as well as the amount of oxygen that remains after 
the hydrocarbon carrier completes combustion.  In a ducted rocket environment, 
additional oxidizer is provided in the fuel.  While some performance is lost, a 
greater chance for energy extraction exists. 
?? A study should be conducted in which mixtures of nano-sized aluminum and 
nano-sized boron are used as composite particles to create fuel additives.  Since 
nanoaluminum has demonstrated lower ignition temperatures than traditional 
micron-sized aluminum, it can potentially be used as an additional ignition source 
for the boron particles.    
?? Conduct ramjet studies with nanoaluminum included in either a solid matrix such 
as hydroxyl-terminated polybutadiene (HTPB) or in a liquid such as RP-1 to 
create a gel or slurry.  The initial screening this study provided suggests great 
promise for the use of aluminum nanoparticles in high-speed airbreathing 
applications.  However, in a practical system, the nanoaluminum must be stored 
in a solid matrix, a gel, or slurry. 
 209
APPENDIX A ? DESIGN DRAWINGS 
 
Fig. A.1 Photograph of Flat Flame Burner Experiment 
 
 
Fig. A.2 Components of Sonicating Atomizer 
 
 
McKenna Flat 
Flame Burner 
Aerosol 
Injector 
 210
 
Fig. A.3 Photograph of Thermocouple Configuration 
 
 
 
 
Fig. A.4 Photograph of Seeder Components 
 
 
Aerosol Injector 
B-Type 
Thermocouple 
Inlet Exhaust Tube 
 211
 
Fig. A.5 Modified Orifice Used For Particle Injection 
 
 
 
Fig. A.6 Specially Designed Adapter and Flow Straightener 
 212
 
Fig. A.7 Combustor Components 
 
Fig. A.8 Exploded View of Combustor Components
 213
 
Fig. A.9 Pilot Flame Combustor Aft End Cap 
 214
  
Fig. A.10 Pilot Flame Combustor Main Body 
 
 215
 
Fig. A.11 Seeder Body 
 
 216
 
Fig. A.12 Seeder Plug 
 217
  
Fig. A.13 Seeder Cap 
 
 218
 
Fig. A.14 Combustor Bottom Plate 
 
 219
 
 
Fig. A.15 Combustor Injector Plate (Mates With Pilot Flame Combustor) 
 220
 
Fig. A.16 Combustor Top Plate 
 221
 
Fig. A.17 Alternate Top Plate 
 222
 
Fig. A.18 Left Window Holder Side View 
 223
Tol: ?0.005
Left Window Holder
Nano-Burner
Drawing: NB-12
Greg Young
inches 25-APR-05
Material:
303 S.S.
Scale: 1 to 1
Approved By:
 
Fig. A.19 Left Window Holder End View 
 224
 
Fig. A.20 Right Window Holder Side View 
 225
 
Tol: ?0.005
Right Window Holder
Nano-Burner
Drawing: NB-13
Greg Young
inches 25-APR-05
Material:
303 S.S.
Scale: 1 to 1
Approved By:
 
Fig. A.21 Right Window Holder End View 
 226
 
Fig. A.22 Quartz Window 
 227
 
Fig. A.23 Front Transition 
 
 228
 
Fig. A.24 Rear Transition 
 229
APPENDIX B ? UNCERTAINTY ANALYSIS 
  
The uncertainty analysis in this study followed the procedures given by Kline and 
McKlintock93.  The procedure dictates that for a given variable, y (see equation B.1), that 
the uncertainty, ?y, in y, can be determined by equation B.2.    
? ?y f x x xn? 1 2, ,...,               Eq. B.1 
? ?? ? ?? ? ?? ?y fx x fx x fx x
n
n?
?
??
?
?? ?
?
??
?
?? ? ?
?
??
?
??1 1
2
2
2
2 2
....             Eq. B.2 
 
As an example, the uncertainty in the radiation correction was examined.  First, all of the 
variables and sub-variable involved were identified and can be seen in equations B.3-B.6. 
Red Ud? ?                  Eq. B.3 
Nu d? ?0 42 0 570 2 13 12. * Pr . * Pr * Re.              Eq. B.4 
h Nu kd? *                         Eq. B.5 
? ?T T T T
hg tc
tc tc? ? ? ?? ?
4 4
                         Eq. B.6 
Next, the uncertainty in each one of the variables was determined by the procedures 
outlined by Kline and McKlintock93 as seen in equations B.7-B.9.  The uncertainty in 
thermocouple diameter was taken to be a typical machining tolerance of 0.005 inches, 
while the uncertainty in all gas properties was taken as 10%.  The emissivity values were 
taken from Shaddix64 as 0.22 and 0.25 respectively for B-type and K-type thermocouples 
respectively. 
 230
? ? ? ? ? ? ??Red d U U d Ud? ??? ??? ? ??? ??? ? ??? ???
2 2
2
2
             Eq. B.7 
 
? ? ?Nu d d d? ???? ?????? ??? ? ??? ???? ? ?0 2 0 42 13 057 12 0 570 8 23 12
2
13 12 2. * . * Pr * . Pr * Re * Pr * . * Pr * Re * Re.
 
                     Eq. B.8 
? ? ?h kd Nu Nu kd d? ??? ??? ? ??? ???
2
2
2*
             Eq. B.9 
and finally the uncertainty in the actual gas temperature is found by equation B.10. 
? ?? ? ??? ?? ?? ? ?? ?T TT T T TT T Th hg g
tc
tc
g g g? ?
??
?
?? ?
?
??
?
?? ?
?
??
?
?? ?
?
??
?
??? ?
2 2 2 2
        Eq. B.10 
where each term was evaluated in the following manner: 
?
?
??T
T
T
h
g
tc
tc? ?1 4
3
            Eq. B.11 
? ??
??
?T T T
h
g tc? ? ?
4 4
            Eq. B.12 
?
?
??T
T
T
h
g
?
?? ? 4
3
            Eq. B.13 
? ??
?
??T
h
T T
h
g tc? ? ? ?
4 4
2            Eq. B.14 
The same procedure was used for determining the uncertainty for every measurement 
made in this study.  
 231
APPENDIX C ? ROTAMETER CALIBRATIONS 
0
5
10
15
20
25
0 10 20 30 40 50 60 70 80
y = -1.32 + 0.36914x   R2= 0.99975 
Ox
yg
en
 Fl
ow
 R
ate
 (L
/m
in)
Rotameter Setting  
Fig. C.1 Calibration for Oxygen 
5
10
15
20
25
30
35
10 20 30 40 50 60 70 80 90
y = -1.4399 + 0.40368x   R2= 0.99982 
Ai
r F
low
 R
ate
 (L
/m
in)
Rotameter Setting  
Fig. C.2 Calibration for Air 
 232
2
3
4
5
6
7
8
9
20 40 60 80 100 120 140
y = -0.11367 + 0.074964x   R2= 0.99974 
Me
tha
ne
 Fl
ow
 R
ate
 (L
/m
in)
Rotameter Setting  
Fig. C.3 Calibration for Methane 
 
 233
References 
1. Sutton, G.P., Rocket Propulsion Elements: An Introduction to the Engineering of 
Rockets, 6th Edition, John Wiley and Sons, Inc., 1992 
2. Fry, R.S., ?A Century of Ramjet Propulsion Technology Evolution?, Journal of 
Propulsion and Power, Vol. 20, No. 1, January-February, pp. 27-58 
3. Timnat, Y.M., Advanced Airbreathing Propulsion, Krieger Publishing Company, 
1996 
4. Harris, R.V., ?On the Threshold ? The Outlook for Supersonic and Hypersonic 
Aircraft?, Journal of Aircraft, January-February 1992, pg. 10-19 
5. Heiser, W.H., Pratt, D.T., Hypersonic Airbreathing Propulsion, American 
Institute of Aeronatics and Astronautics, 1994 
6. Cazin, P., Laurent, J.M., ?Liquid Fuel Ramjet Engine?, Tactical Missile 
Propulsion, AIAA Volume 170, Gorden E. Jensen, David W. Netzer, Eds., pp. 
423-446, Progress in Astronautics and Aeronautics, 1996 
7. Leisch, S., Netzer, D., ?Solid Fuel Ramjets?, Tactical Missile Propulsion, AIAA 
Volume 170, Gorden E. Jensen, David W. Netzer, Eds., pp. 469-496, Progress in 
Astronautics and Aeronautics, 1996 
8. Gany, A., Netzer, D.W., ?Fuel Performance Evaluation of the Solid Fuel Ramjet?, 
International Journal of Turbo and Jet Engines, 2, 1985, pp.157-168 
9. Gany, A., ?Combustion of Boron Containing Fuels in Solid Fuel Ramjets?, 
Combustion of Boron-Based Solid Propellants and Solid Fuels, K. K. Kuo, R. 
Pein, Eds., pp. 91-112, CRC Press Begell House, Boca Raton, 1993  
 234
10. Ciezki, H.K., Sender, J., Claub, W., Feinauer, A., Thumann, A., ?Combustion of 
Solid-Fuel Slabs Containing Boron Particles in Step Combustor?, Journal of 
Propulsion and Power, Vol. 19, No. 6, November-December 2003 
11. Wimmerstrom, P., Lundgren, R., ?Studies of Fuel-Rich Magnesium Propellants in 
a Small Solid Fuel Ramjet Combustor?, 22nd International Annual Conference of 
ICT, Karlsruhe, pp. 96-1 ? 96-7, Germany, July 2-5, 1991 
12. Leingang, J.L., Petters, D.P., ?Ducted Rockets?, Tactical Missile Propulsion, 
AIAA Volume 170, Gorden E. Jensen, David W. Netzer, Eds., pp. 447-468, 
Progress in Astronautics and Aeronautics, 1996 
13. Hallion, R.P., ?The History of Hypersonics: or, ?Back to the Future ? Again and 
Again??, AIAA 2005-0329, 43rd AIAA Aerospace Sciences Meeting and Exhibit, 
Reno, Nevada, January 10-13, 2005 
14. Waltrup, P.J., White, M.E., Zarlingo, F., Gravlin, E.S., ?History of U.S. Navy 
Ramjet, Scramjet, and Mixed-Cycle Propulsion Development?, Journal of 
Propulsion and Power, Vol. 18, No. 1, January-February 2002, pp. 14-27 
15. Olsen, S.E., Beckstead, M.W., ?Burn Time Measurements of Single Aluminum 
Particles in Steam and CO2 Mixtures?, Journal of Propulsion and Power, Vol. 12, 
No. 4, pp. 662-671, 1996 
16.  Parr, T., Johnson, C., Hanson-Parr, D., Higa, Kelvin, and Wilson, K., ?Evaluation 
of Advanced Fuels for Underwater Propulsion?, 39th JANNAF Combustion 
Subcommittee Meeting, December, 2003 
 235
17. Bazyn, T., Krier, H., Glumac, N., ?Shock Tube Measurements of Combustion of 
Nano-Aluminum?, AIAA 2006-1157, 44th AIAA Aerospace Sciences Meeting 
and Exhibit, Reno, Nevada, January 9-12, 2006 
18. Tepper, F. and Kaledin, L.A., ?Combustion Characteristics of Kerosene 
Containing Alex Nano-Aluminum,? 5th International Symposium on Special 
Topics in Chemical Propulsion, Stressa, Italy, June 18-22, 2000 
19.  Goroshin, S., Higgins, A.J., and Kamel, M., ?Powdered Metals as Fuel for 
Hypersonic Ramjets?, AIAA-2001-3919, 37th AIAA/ASME/SAE/ASEE Joint 
Propulsion Conference and Exhibit, Salt Lake City, Utah, July 8-11, 2001 
20. Palaszewski, B., ?Metallized Gelled Propellants: Oxygen/RP-1/Aluminum Rocket 
Engine Calorimeter Heat Transfer Measurements and Analysis?, AIAA-97-2974, 
33rd AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit, Seattle, 
Washington, July 6-9, 1997 
21. Mordosky, J.W., Zhang, B.Q., Kuo, K.K., Tepper, F., Kaledin, L.A., ?Spray 
Combustion of Gelled RP-1 Propellants Containing Nano-sized Aluminum 
Particles in Rocket Engine Conditions?, AIAA-2001-3274, 37th 
AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit, Salt Lake 
City, Utah, July 8-11, 2001 
22. Risha, G.A., Evans, B. J., Boyer, E., Wehrman, R. B., and Kuo, K. K., ?Nano-
Sized Aluminum- and Boron-Based Solid-Fuel Characterization in a Hybrid 
Rocket Engine,? AIAA Paper No. 2003-4593, 39th AIAA/ASME/SAE/ASEE 
Joint Propulsion Conference and Exhibit, Huntsville, Alabama, July 2003 
 236
23. Palaszewski, B., Jurns, J., Breisacher, K., Kearns, K., ?Metallized Gelled 
Propellants Combustion Experiments in a Pulse Detonation Engine?, AIAA-2004-
4191, 40th AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit, Ft. 
Lauderdale, Florida, July 11-14, 2004 
24. Park, K., Lee, D., Rai, A., Mukherjee, D., Zachariah, M.R., ?Sized-Resolved 
Kinetic Measurements of Aluminum Nanoparticle Oxidation with Single Particle 
Mass Spectrometry?, Journal of Physical Chemistry B, 109 (15), 2005, pp. 7290-
7299   
25. Kuo, K.K., Principles of Combustion, Wiley & Sons, New York, 1986 
26. Price, E.W. and Sigman, R.K., ?Combustion of Aluminized Solid Propellants,? 
Solid Propellant Chemistry, Combustion, and Motor Interior Ballistics, AIAA 
Volume 185, Vigor Yang, Thomas B. Brill, and Wu-Zhen Ren, Eds., pp. 663-687, 
Progress in Astronautics and Aeronautics, 2000 
27. Price, E.W., ?Combustion of Metallized Propellants,? Fundamentals of Solid 
Propellant Combustion, Editors Kuo, K.K., and Summerfield, M., AIAA Volume 
90, pp. 479-513 Progress in Astronautics and Aeronautics, 1983 
28. Melcher, J.C., Burton,R.L., Krier, H., ?Combustion of Aluminum Particles in 
Solid Rocket Motor Flows,? Solid Propellant Chemistry, Combustion, and Motor 
Interior Ballistics, AIAA Volume 185, Vigor Yang, Thomas B. Brill, and Wu-
Zhen Ren, Eds., pp. 723-747, Progress in Astronautics and Aeronautics, 2000 
29. Glassman, I., Combustion, 3rd Edition, Academic Press, San Diego, California, 
1996 
 237
30. Branstetter, J.R., Lord, A.M., Gerstein, M., ?Combustion Properties of Aluminum 
as a Ram-Jet Fuel,? NACA Research Memorandum, NACA RM E51B02, March 
28, 1951 
31. Yetter, R.A., and Dryer, F.L., ?Metal Particle Combustion and Classification?, 
Microgravity Combustion: Fire in Free Fall, Academic Press, 2001 
32. Beckstead, M.W., ?Lectures on Metal Combustion?, February 25, 1999, China 
Lake, CA 
33. Liang, Y., Beckstead, M.W., ?Numerical Simulation of Unsteady, Single 
Aluminum Particle Combustion in Air?, AIAA Paper 98-3697, June 1998 
34. Beckstead, M.W., Newbold, B.R., Waroquet, C., ?A Summary of Aluminum 
Combustion?, 39th JANNAF Combustion Subcommittee Meeting, December, 
2003 
35. Friedman, R., Macek, A., ?Combustion Studies of Single Aluminum Particles?, 
Ninth Symposium (International) on Combustion, 1963, pp. 703-709 
36. Friedman, R., Macek, A., ?Ignition and Combustion of Aluminum Particles in 
Hot Ambient Gases?, Combustion and Flame, Vol. 6, 1962, pp. 9-19 
37. Macek, A., ? Fundamentals of Combustion of Single Aluminum and Beryllium 
Particles?, Eleventh Symposium (International) on Combustion, 1967, pp. 203-
217  
38. Davis, A., ?Solid Propellants: The Combustion of Particles of Metal Ingredients?, 
Combustion and Flame, Vol. 7, 1963, pp. 359-367 
 238
39. Drew, C.M., Gordon, A.S.,  Knipe, R.H., ?Study of Quenched Aluminum Particle 
Combustion?, Heterogeneous Combustion, AIAA, Progress in Astronautics and 
Aeronautics Series, Vol. 15, Academic Press, New York, NY, 1964, pp. 17-39 
40. Turns, S.R., Wong, S.C., Ryba, E., ?Combustion of Aluminum-Based Slurry 
Agglomerates?, Combustion Science and Technology, Vol. 54, 1987, pp. 299-318 
41. Foelsche, R.O., Burton, R.L., Krier, H., ?Ignition and Combustion of Aluminum 
Particles in H2/O2/N2 Combustion Products?, Journal of Propulsion and Power, 
Vol. 14, No. 6, 1998, pp. 1001-1008 
42. Bucher, P., Yetter, R.A., Dryer, F.L., Parr, T.P., Hanson-Parr, D.M., ?PLIF 
Species and Ratiometric Temperature Measurements of Aluminum Particles in 
O2, CO2, and N2O Oxidizers, and Comparison with Model Calculations?, Twenty-
Seventh Symposium (International) on Combustion, The Combustion Institute, 
Pittsburgh, PA, 1998, pp. 2421-2429 
43. Bucher, P., Yetter, R.A., Dryer, F.L., Parr, T.P., Hanson-Parr, D.M., Vincenzi, 
E.P., ?Flame Structure of Single, Isolated Aluminum Particles Burning in Air?, 
Twenty-Sixth Symposium (International) on Combustion, The Combustion 
Institute, Pittsburgh, PA, 1996, pp. 1899-1908 
44. Ulas, A., Kuo, K.K., ?Ignition and Combustion of Boron Particles in Fluourine-
Containing Environments?, Combustion and Flame, Vol. 127, 2001, pp. 1935-
1957 
45. Beckstead, M., ?Lectures on Comustion,? Short Course on Metal Combustion, 
Applied Research Laboratory, The Pennsylvania State University, August 20, 
2002 
 239
46. Yeh, C.L., Kuo, K.K., ?Ignition and Combustion of Boron Particles?, Progress in 
Energy Combustion Science, Vol. 22, pp. 511-541 
47. Macek, A., Semple, J.M., ?Combustion of Boron Particles at Atmospheric 
Pressure?, Combustion Science and Technology, Vol. 1, 1969, pp. 181-191 
48. Macek, A., Semple, J.M., ?Combustion of Boron Particles at Elevated Pressures?, 
Thirteenth Symposium (International) on Combustion, The Combustion Institute, 
Pittsburgh, PA, 1971, pp. 859-868 
49. Macek, A., ?Combustion of Boron Particles: Experiment and Theory?, Fourteenth 
Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, 
PA, 1972, pp. 1401-1411 
50. King, M.K., ?Boron Particle Ignition in Hot Gas Streams?, Combustion Science 
and Technology, Vol. 8, 1974, pp. 255-273 
51. Foelsche, R.O., Burton, R.L., and Krier, H., ?Boron Particle Ignition and 
Combustion at 30-150 ATM,? Combustion and Flame, No. 117, pp. 32-58, 1999 
52. Yetter, R.A., Rabitz, H., Dryer, F.L., Brown, R.C., Kolb, C.E., ?Kinetics of High-
Temperature B/O/H/C Chemistry?, Combustion and Flame, No. 83, pp. 43-63, 
1991 
53. Pearse, R.W.B., Gaydon, A.G., The Identification of Molecular Spectra, 4th 
edition, Chapman and Hall, London, 1976, p. 59 
54. Wimmerstein, P., Lundgren, R., ?Studies of Fuel-Rich Magnesium Propellants in 
a Small Solid Fuel Ramjet Combustor?, 22nd International Annual Conference of 
ICT 1991, Karlsruhe, Germany, July 2-5, 1991 
 240
55. Abbott, S.W., Smoot, L.D., Schadow, K., ?Direct Mixing and Combustion 
Efficiency Measurements in Ducted, Particle-Laden Jets?, AIAA Journal, Vol. 12, 
No.3, March 1974, pp. 275-282 
56. Schadow, K., ?Experimental Investigation of Boron Combustion in Air-
Augmented Rockets?, AIAA Journal, Vol. 7, No. 10, October 1969, pp. 1870-
1876 
57. Schadow, K.C., Wilson, K.J., Gutmark, E., Smith, R.A., ?Effect of Gaseous Fuel 
Mixing on Boron Combustion in Ducted Rocket with Side Dump?, Combustion of 
Boron-Based Solid Propellants and Solid Fuels, Editors Kuo, K.K., and Pein, R., 
CRC Press, Begell House, Boca Raton, 1993, pp. 402-411 
58. Gany, A., ?Combustion of Boron-Containing  Fuels in Solid Fuel Ramjets?, 
Combustion of Boron-Based Solid Propellants and Solid Fuels, Editors Kuo, 
K.K., and Pein, R., CRC Press, Begell House, Boca Raton, 1993, pp. 91-112   
59. Natan, B., Gany, A., ?Combustion Characteristics of a Boron-Fueled Solid Fuel 
Ramjet with Aft-Burner?, Journal of Propulsion and Power, Vol. 9, No. 5, Sept.-
Oct. 1993, pp. 694-701 
60. Gany, A., ?Comprehensive Consideration of Boron Combustion in Airbreathing 
Propulsion?, AIAA-2006-4567, 42nd AIAA/ASME/SAE/ASEE Joint Propulsion 
Conference and Exhibit, Sacramento, California, July 9-12, 2006 
61. Son, S.F., Dye, R.C., Busse, J.R., Sandstrom, M.M., Oschwald, D.M., Janicke, 
M.T., ?Combustion of Nanoaluminum in Air?, 39th JANNAF Combustion 
Subcommittee Meeting, December, 2003 
 241
62. Mench, M. M., Yeh, C.L., Kuo, K.K., ?Propellant Burning Rate Enhancement and 
Thermal Behavior of Ultra-fine Aluminum Powders?, Energetics Materials 
Production, Processing and Characterization, 29th International ICT Conference, 
June 30-July 3, 1998, pp. 3-01-3-05 
63. Turns, S.R., An Introduction to Combustion, McGraw-Hill, Inc., New York, 
1996,  
64. Schlichting H., Boundary-Layer Theory, 6th Edition, McGraw-Hill, New York, 
1968 
65. Raffel, M., Willert, C., Kompenhans, J., Particle Image Velocimetry, Springer, 
New York, 1998 
66. Shaddix, Christopher R., ?Correcting Thermocouple Measurements for Radiation 
Loss: A Critical Review,? NHTC99-282, Proceedings of the 33rd National Heat 
Transfer Conference, Albuquerque, New Mexico, August 15-17, 1999 
67. Whitaker, S., (1972), AIChE J. 18:361 
68. Incropera, F.P., Dewitt, D.P., Introduction to Heat Transfer, John Wiley & Sons, 
New York, 1996 
69. Muruganandam, T.M., Kim, B., Olsen, R., Patel, M., Romig, B., and Seitzman, 
J.M., ?Chemiluminescence Based Sensors for Turbine Engines,? 39th  
AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit, AIAA 2003-
4490, 2003 
70. Mayinger, F., ?Optical Measurements: Techniques and Applications,? Springer- 
Verlag, 1994, pp. 356 
 242
71. Gaydon, A.G., Wolfhard, H.G., Flames: Their Structure, Radiation, and 
Temperature, John Wiley & Sons, New York, 1979 
72. Spalding, M.J., Krier, H., Burton, R.L., ?Emission Spectroscopy During Ignition 
of Boron Particles at High Pressure,? 35th AIAA Aerospace Sciences Meeting and 
Exhibit, January, 1997, AIAA 97-0119 
73. Glass, M., Kennedy, I.M., ?An Improved Seeding Method for High Temperature 
Laser Doppler Velocimetry,? Combustion andFlame, No. 29, pp. 333-335, 1977 
74. Yu, K.H., Parr, T.P., Schadow, K.C., ?Planar Mie Scattering Visualization of 
Reacting and Nonreacting Supersonic Coaxial Jets,? Non-Intrusive Combustion 
Diagnostics, Eds. K.K. Kuo and T.P. Parr, Begell House, Inc., New York, 1994, 
pp. 504-517  
75. Gaydon, A.G.,  The Spectroscopy of Flames, John Wiley & Sons, New York, 
1974  
76. Katz, J., Tepper, F., Ivanov, G.V., Lerner, M.I., Davidovich, V., ?Metastable 
Nanosize Aluminum Powder as a Reactant in Energetic Formulations,? JANNAF 
Propulsion  Meeting, CPIA Publication 675, Vol. 3, 1998, pp. 343-349 
77. Impact Analytical Documentation for Thermal Analysis Website, 
http://www.impactanalytical.com, 2002 
78. Verhoeven, J.D., ?Scanning Electron Microscopy,? ASM Handbook, Vol. 10, 4th 
Edition, January 1996, pp. 490-515 
79. Hinds, W.C., Aerosol Technology, John Wiley & Sons, New York, 1999 
 243
80. Yang, Y., Wang, S., Sun, Z., Dlott, D., ?Near Infra-red and Visible Absorption 
Spectroscopy of Nano-Energetic Materials Containing Aluminum and Boron,? 
Propellants, Explosives, and Pyrotechnics, Vol. 30, 2005, No. 3, pp. 171-177 
81. Mihalcea, R.M., Baer, D.S., Hanson, R.K., ?Diode Laser Sensor for 
Measurements of CO, CO2, and CH4, in Combustion  Flows,? Applied Optics, 
Vol. 36, No. 33, November 1997 
82. Mihalcea, R.M., Baer, D.S., Hanson, R.K., ?A Diode-Laser Absorption Sensor 
System for Combustion Emission Measurements,? Measurement Science and 
Technology, 9, 1998, pp. 327-338 
83. Friedlander, S.K., Smoke, Dust, and Haze, 2nd Edition, Oxford University Press, 
New York, 2000 
84. B.J. McBride, S. Gordon, ?Computer Program for Calculation of Complex 
Chemical Equilibrium Compositions and Applications,? NASA Reference 
Publication 1311, June 1996 
85. Gruber, M.R., Nejad, A.S., Chen, T.H., Dutton, J.C., ?Mixing and Penetration 
Studies of Sonic Jets in a Mach 2 Freestream,? Journal of Propulsion and Power, 
Vol. 11, No. 2, March-April 1995 
86. Keffer, J.F., and Baines, W.D., ?The Round Turbulent Jet in a Cross-Wind,? 
Journal of Fluid Mechanics, Vol. 15, 1963, pp 481-496 
87.  Li, S.C., Williams, F.A., ?Ignition and Combustion of Boron Particles?, 
Combustion of Boron Based Solid Propellants and Fuels, Kuo, K.K., and Pein, 
R., Eds., CRC Press, Boca Raton, 1993, pp 248-271 
 244
88. Bazyn, T., Krier, H., Glumac, N., ?Oxidizer and Pressure Effects on the 
Combustion of 10-?m Aluminum Particles,? Journal of Propulsion and Power, 
Vol. 21, No. 4, July-August 2005, pp. 577-582 
89. Huang, Y., Risha, G.A., Yang, V., Yetter, R.A., ?Combustion of Bimodal 
Nano/Micron-sized Aluminum Particle Dust in Air,? Proceedings of the 
Combustion Institute 31, 2007  
90. Yuasa, S., Isoda, H., ?Ignition and Combustion of Small Boron Lumps in an 
Oxygen Stream,? Combustion and Flame, Vol. 86, 1991, pp. 216-222 
91. Pen?kov, S.N., ?Investigation of Boron Ignition and Combustion by Means of 
Fast Cinespectrography?, Combustion of Boron Based Solid Propellants and 
Fuels, Kuo, K.K., and Pein, R., Eds., CRC Press, Boca Raton, 1993, pp 218-231 
92. King, M.K., ?A Review of Studies of Boron Ignition and Combustion Phenomena 
at Atlantic Research Corporation Over the Past Decade?, Combustion of Boron 
Based Solid Propellants and Fuels, Kuo, K.K., and Pein, R., Eds., CRC Press, 
Boca Raton, 1993, pp 1-80 
93. Kline, S.J., and McClintock, F.A., ?Describing Uncertainties in Single Sample 
Experiments,? Published by the American Society of Mechanical Engineers, 1953