Combustion Characteristics and Flame Structure
Combustion Characteristics and Flame Structure
Combustion Characteristics and Flame Structure
College of Engineering
A Thesis in
Mechanical Engineering
by
J. Eric Boyer
Doctor of Philosophy
December 2005
The thesis of J. Eric Boyer was reviewed and approved* by the following:
Kenneth K. Kuo
Distinguished Professor of Mechanical Engineering
Thesis Advisor
Co-Chair of Committee
Paul W. Brown
Professor of Materials Science and Engineering
Thesis Co-Advisor
Co-Chair of Committee
Thomas A. Litzinger
Professor of Mechanical Engineering
Stefan T. Thynell
Professor of Mechanical Engineering
Vigor Yang
Distinguished Professor of Mechanical Engineering
ABSTRACT
The push for higher performance and reduced toxicity monopropellant has led to
the search for a hydrazine replacement. One promising candidate, nitromethane, was
investigated in this study. To assist the design and evaluation of future systems,
measurement of basic propulsion parameters such as burning rate were made over a wide
range of conditions for both static (tube) and feeding tests. Three pressure regimes were
measured using microthermocouples showed a very thin reaction zone and low-
A comprehensive detailed model for linear regression was developed and exercised.
The model considered one-dimensional behavior with surface vaporization and detailed
gas-phase kinetics based on the RDX mechanism of Yetter, et al. combined with the
using a custom FORTRAN code wrapping the CHEMKIN PREMIX gas-phase code
coupled with the condensed-phase model. Predicted burning rates using the model
iv
showed good agreement with measured rates over the subcritical range of 3 to 6 MPa,
based on review of recent literature, two rates were adjusted slightly to improve the
match. Calculated species and temperature profiles showed three regions that could be
distinguished based on species. CH4 and NO were the most important intermediate
species, with smaller percentages of CH2O, N2O, HNO, and HONO present. The first
intermediate species except CH4 and NO, and final by rise to final temperature and
to the temperature profile, and therefore burning rate. Although the absolute levels of
NH and HCO were low, they served as an important intermediate species transporting
such as nitramines, the combustion characteristics (no stable combustion at low pressure
and very thin reaction zone) made it more difficult to study. Many of the techniques that
have been applied to nitramines cannot be used with nitromethane because of its nature as
a clear liquid. Future refinement of the current model can be aided by the development
reaction zone so that current techniques can more easily resolve species and temperature
profiles.
v
TABLE OF CONTENTS
NOMENCLATURE........................................................................................................ xii
REFERENCES.............................................................................................................. 211
vii
LIST OF FIGURES
Figure 3.2 Cross-sectional view of the windowed LPSB chamber with feeding base
installed. .....................................................................................................................64
Figure 3.3 Cross-sectional diagram of LPSB chamber base and feed system outlet........65
Figure 3.4 Schematic diagram of the UHPSB and control system. ..................................74
Figure 3.5 Schematic diagram of the optical arrangement for UV/Visible absorption
spectroscopy measurements.......................................................................................79
Figure 4.1 Pictures of near-surface behavior for nitromethane at a stable free surface in
LPSB feeding tests at different pressures. .................................................................83
Figure 4.2. Images of regressing surface into quiescent propellant in square tubes.........84
Figure 4.3 Density and temperature contours for nitromethane combustion plume into a
slow purge flow (0.5 cm/s). .......................................................................................87
Figure 4.4 Density and temperature contours for nitromethane combustion plume into a
medium purge flow (5 cm/s)......................................................................................88
Figure 4.5 Density and temperature contours for nitromethane combustion plume into a
fast purge flow (50 cm/s). ..........................................................................................89
Figure 4.7 High-pressure nitromethane burning rate data of Rice and Cole37 compared to
current study...............................................................................................................94
Figure 4.8 Data of Kelzenberg, et al. and Raikova, et al. compared to correlations
obtained from current study. ......................................................................................95
Figure 4.9 Nitromethane burning rate sensitivity as a function of initial temperature .....97
viii
Figure 4.13 Image of nitromethane flame immediately above the surface of the sample
contained in a quartz tube (8 x 8 mm2) obtained from the UV/Visible test rig at 5.6
MPa (800 psig).........................................................................................................104
Figure 4.14 Absorption spectra of nitromethane flame at 5.6 MPa (800 psig). .............104
Figure 5.1 Physical model and axial temperature distribution for steady-state combustion
of liquid nitromethane fed into the burning zone under a controlled constant-pressure
condition. .................................................................................................................107
Figure 5.3 Schematic diagram of physicochemical processes for the bubbles within the
RDX two-phase reaction zone (from Ref. 186) .......................................................113
Figure 5.4 Comparison of two different methods of calculating enthalpy change for
initial temperature of 273 K to gas phase at specified surface temperature. ...........118
Figure 5.6 Flow chart for PREMIX gas-phase driver code and CHEMKIN system (from
Ref. ). .......................................................................................................................127
Figure 5.7 Flow chart of program implementing gas- and condensed-phase coupling. .130
Figure 5.8 Fall-off curves for CH3NO2 dissociation (from Glarborg, Bendtsen, and
Miller94)....................................................................................................................134
Figure 5.9 Curve fit to center-broadening parameter (Fcent) of Glarborg, Bendtsen, and
Miller for Troe form of falloff reaction. ..................................................................136
Figure 5.10 Comparison of literature values (DIPPR), curve fit, and CHEMKIN
polynomial approximation for nitromethane vapor viscosity. .................................145
Figure 6.1 Comparison of calculated and measured burning rates for nitromethane at 3 to
6 MPa. ......................................................................................................................153
ix
Figure 6.3 Variation of nitromethane equilibrium flame temperature and product species
with pressure. ...........................................................................................................156
Figure 6.7 Comparison of calculated temperature profiles for a range of pressures. .....165
Figure 6.8 Temperature and species profiles calculated for 150 MPa............................167
Figure 6.12 Calculated burning rate as a function of initial temperature and pressure. .178
Figure 6.13 Comparison of measured and calculated values for burning rate temperature
sensitivity. ................................................................................................................179
x
LIST OF TABLES
Table 2.4 Antoine Equation parameters for boiling point of nitromethane (coefficients
calculated by NIST12 from other authors data).........................................................14
Table 2.5 Coefficients for nitromethane vapor pressure correlation (from DIPPR6). ......14
Table 2.6 Coefficients of the Simon-Glatzel equation for the high-pressure fusion curve
of nitromethane (from Ref. 20). .................................................................................16
Table 2.7 Mechanisms and reaction parameters for the thermal decomposition of
nitromethane. .............................................................................................................42
Table 3.1 D-type thermocouple inverse polynomial for calculating temperature from
voltage........................................................................................................................78
Table 4.1 Input and intial properties and conditions for FLUENT simulation of flame
plume in cold purge flow. ..........................................................................................86
Table 5.1 Coefficients for nitromethane density correlation (from DIPPR6) .................137
Table 5.2 Coefficients for nitromethane liquid heat capacity correlation (from DIPPR6)138
Table 5.3 Coefficients for approximation of the reduced collision integral ij(2,2)* ........143
Table 5.4 Coefficients for vapor viscosity correlation (from Ref. 6). ............................144
Table 5.5 Coefficients for CHEMKIN nitromethane vapor viscosity 3rd-order polynomial
correlation. ...............................................................................................................146
Table 5.6 Coefficients for DIPPR nitromethane vapor thermal conductivity correlation
(from Ref. 6). ...........................................................................................................148
Table 5.7 Coefficients for CHEMKIN nitromethane vapor thermal conductivity 3rd-order
xi
NOMENCLATURE
T temperature (K)
Greek Symbols
polarizability (3)
reaction rate
Subscripts
cond conduction
init initial
i, j indices
rad radiation
reac reaction
s, sur surface
xiv
ACKNOWLEDGEMENTS
I would like to thank Prof. Kenneth K. Kuo for his invaluable advice and support
throughout my graduate career, and acting as advisor and committee Co-Chair for this
study. I would also like to thank Paul W. Brown for serving as the committee Co-Chair.
The time and effort of Profs. Thomas A. Litzinger, Stefan T. Thynell, and Vigor Yang are
Discussions with Prof. Richard Yetter and use of his RDX mechanism were very
valuable to this work. Mr. Donald Kochs help in the setup of the ultra-high pressure
strand burner, and the assistance of Dr. Yi-Ping Chang and Dr. Abdullah Ulas with
like to thank Prof. BaoQi Zhang for his valuable translations of Russian-language works,
Portions of this material are based upon earlier work supported by the U. S. Army
under the management of Dr. Robert W. Shaw and Contract No. DAAH04-96-1-0054
Finally, I would like to thank my parents for their unflagging support and
encouragement.
1
CHAPTER 1
Introduction
There are three general types of chemical rockets, with the name a rough
description of the propellant storage phase in the system: liquid, solid, and hybrid.
Liquid rockets depend on liquid propellants being fed under pressure from tanks into the
thrust chamber; either pressurized tanks or pumps can be used. In solid rockets, the
container holding the solid propellant is the motor chamber. Hybrid rockets combine
aspects of both solids and liquids, while separating the fuel and oxidizer by both phase
and space (usually with liquid oxidizer and solid fuel).1 In this study, nitromethane liquid
In general, liquid rockets offer the highest performance and best control over
engine operation, although the pumping/pressurization and feed control systems are
heavier and more complex than the support components needed for solids or hybrids.
separate from the thrust chamber, tank locations and shapes are not tightly prescribed,
leading to design flexibility such as the use of conformal tanks. However, the liquid can
be subject to spills and leaks, leading to handling difficulties when using toxic or
hazardous propellants. Since only a portion of the system is subjected to the heat and
A sub-classification of liquid rockets depends on the form of the oxidizer and fuel
in the propellant. The fuel and oxidizer may be separate compounds in the mixture or in
monopropellant. The majority of liquid rocket systems utilize separate fuel and oxidizer,
monopropellants are very useful for attitude and trajectory control rockets. Another
storage) but easily decompose and have good combustion properties.1 Physical
properties of the liquid (e.g. density, boiling point, freezing point, etc.) must also be
suited to the application. For example, cryogenic or high freezing point propellants
requiring temperature control would not be appropriate for space or tactical applications
due to the adverse environments encountered by the systems. The material must also be
user-friendly for handling. Because of the above requirements and other storage and
bipropellants. Also, because the material must be somewhat insensitive for safety up
until the point it is injected into the chamber, high-power igniters or catalysts are often
can be lighter and less complex because only a singly storage and feed system is required.
3
Monopropellant engine supply systems often consist of just a pressurized tank and inline
relatively stable material, they often require the use of a catalyst in the chamber. In some
cases, just an atomizing injector for supplying monopropellant sprays into the combustor
is sufficient.
H2O).3 Nitromethane also shows almost 50% improvement over 98% hydrogen peroxide
A.
There is also potential for the use of liquid monopropellant in large-bore guns
such as howitzers. Compared to current solid stick propellants, a liquid system can allow
greater control over ballistic performance by metering a precise load of propellant into
chamber, allowing the range and velocity of the round to be tailored. However, the
battlefield environment.
4
300
P = 1000 psia
c
282.1
= 40 266.4 269.4
250
Vacuum Specific Impulse [sec]
200
188.3
150
100
50
0
HTP 98% Hydrazine HAN 269 MEO Nitromethane
Hydrogen Peroxide
monopropellant has led to the search for a hydrazine replacement. The material
examined in this study, nitromethane, offers potential to meet that need with better
basic propulsion parameters such as burning rate over a wide range of conditions is useful
in evaluating these materials for future designs. Previous measurements available in the
literature were over a limited range and performed using crude instruments leading to
predictive capability for burning behavior will aid in development of new systems
As part of the renewed interest, old problems must be also considered and
addressed. Previously, shock sensitivity was found to be an issue. However, there are
known additives (especially alcohols such as butanol) that can reduce the risk to
acceptable levels. In early rocket tests, combustion instability was also noted during
nitromethane engine firings. There are a number of possible solutions to this negative
aspect. Although the solution depends on the precise mode, with great advances in
chamber and injector design and materials over the past half-century, it is likely this
approaches are available: increasing viscosity of the propellant was shown to reduce
The overall goal of this research was to achieve a deeper understanding of the
especially rocket engines. In order to achieve this goal, major objectives of this
Observe surface burning behavior and flame structure as well conduct other
nitromethane; and
Experimental, theoretical, and literature studies have been undertaken to meet the
objectives listed in the previous section. The complementary connection between these is
state of knowledge in the field and avoided duplication of work. Properties of materials
and specific reaction rate data were obtained from literature and standard references.
Chapter 3, with results presented and analyzed in Chapter 4. Experimental work was
selected to generate basic combustion parameters, characterize the physical and chemical
structure of the surface region and luminous flame zone, and to provide inputs to the
modeling activities. Regression rate data allowed the model results to be tested against
combustion were developed to provide a predictive tool that allowed linear regression
determined based on simulation of the chemical and physical processes involved. Results
of the model, given in Chapter 6, also allowed trends in other calculated intermediate
Finally, the findings are summarized and suggestions for the direction of future
In situ Observation of
Flame and Near-surface
Zones
Reaction Data
Limited Species from Literature
Data from
UV/Vis
Absorption Model and
Numerical Code
CHAPTER 2
subject to best allocate time and effort. This chapter provides some general information
and background on the substance of interest, and detailed literature review of previously-
applications. It is commonly used as a solvent for chemical processing and analysis. The
energetic nature of the compound makes it useful as a liquid explosive and as a high-
performance additive to fuel for internal combustion (IC) engines.4 As a storable high-
energy monopropellant, it has potential to be used in attitude control rockets where its
lower toxicity makes it an attractive replacement for hydrazine, and when mixed with
other components such as alcohol has been investigated for liquid propellant guns.
shown in Figure 2.1. As the name suggests, the molecule is essentially methane with one
of the four hydrogens replaced by a nitro group. In essence, it is the simplest of possible
energetic CHON molecules that contain nitro groups, which is why it is often used in
9
H C NO2
classified as a flammable liquid for storage and shipping, it may detonate when
shocked with sufficient energy. It can also be explosion sensitized when contaminated
with amines, strong acids, or alkalis. Some metal ions are incompatible with
containing either metal. Other materials such as some plastics, rubbers, and coatings can
react directly with nitromethane. Contact with the liquid or breathing large amounts
vapor can cause irritation, but it is not considered a carcinogen.4 Because the activation
energy of its weakest bond is well over the rule-of-thumb value of 170 kJ/mol, the pure
More detailed discussion of thermal and transport properties and their variation with
Oxygen balance is a way to measure the surplus (or deficit) of oxygen available in
the combustion of a material. It can be defined as the mass fraction of oxygen in excess
of that needed for complete combustion (i.e. to CO2 and H2O), expressed as a percentage.
values mean oxygen is available to combine with other materials in the propellant
mixture; negative values mean the material itself cannot achieve complete combustion.
Because rocket performance is a function of more than just molecular formula, oxygen
balance is most valuable when used to indicate relative trends, not absolute quality of a
particular material. For comparison and reference, oxygen balance of nitromethane and a
chemical isomer of nitromethane, and therefore possessing the same molecular weight
and oxygen balance, methyl nitrite is not usable as a storable propellant due to its low
boiling point. Nitroethane, a two-carbon chain with an attached nitro group, has more
fuel (C and H) compared to oxygen, and therefore is more appropriate as a liquid fuel
compounds with multiple nitro groups, and more stoichiometric than those with longer
carbon chains. This is demonstrated with dinitromethane (containing two opposing nitro
groups), which tends to explode near the boiling point. Butyl alcohol can be used as a
In designing and analyzing practical systems, temperature and pressure data for
nitromethane was found in the literature for conditions encountered in typical rocket
12
the phase behavior in the gigapascal range, and engineering tables have been generated
for the phase change behavior at industrial conditions, but no source unified the data into
a single diagram. To rectify this, a phase diagram was generated, using data collected
from a variety of sources and approximating curves where data were not available. Two
of the most important constants on the phase diagram, the triple and critical points, are
noted in Table 2.3. Subsequent sections document the source of these values and the
One of the most important points on the phase diagram is the critical point,
indicating where differences between vapor and liquid phases vanish. Although the
for critical pressure appear in the literature. One of the best-known standard references,
the CRC Handbook of Chemistry and Physics,11 lists a value (5.87 MPa) which is lower
than that found in other sources. The National Institute of Standards and Technology
(NIST) Webbook12 cites two studies with contradictory values. The value given by the
Design Institute for Physical Properties (DIPPR) project6 (6.3126 MPa) was compiled
from multiple sources and critically evaluated, so it will be accepted as the reference
13
At the triple point, all three phases (solid, liquid, and vapor) exist in
thermodynamic equilibrium. The location of this point for nitromethane is much better
established than that of the critical point, with the value of Jones and Giauque (244.73 K)
cited by the NIST Webbook12 and the value predicted by the DIPPR project staff6
(244.60 K) differing by less than 0.1 percent. From the actual manuscript of Jones and
Giauque,8 it is not clear that they measured the triple point; rather it appears to be the
melting point at one atmosphere. In any case, since the melting point is relatively
documented for nitromethane. Studies by McCullough et al.13 and Berman and West14
temperature ranges. McCullough et al. obtained data over the range of 55 to 136 C and
Berman and West investigated higher temperatures between 130 and 200 C. Although
Berman and West found discoloration at temperatures above 140 C, using mass
spectroscopic analysis they could find no change in composition, indicating that there
was virtually no reaction at these temperatures. As compiled in the NIST Webbook,12 the
B
log P = A (2.1)
T+C
14
Table 2.4 Antoine Equation parameters for boiling point of nitromethane (coefficients
calculated by NIST12 from other authors data).
Temperature (K) A B C Source of Data
404.9 476.0 4.11350 1229.574 -76.221 Berman and West
328.86 - 409.5 4.40542 1446.196 -45.633 McCullough, et al.
temperature, limiting its usefulness in combustion modeling. Gallant expanded the range
using the Kharbanda nomograph,15 covering -20 to 240 C, but published his results only
in graphical form.16 More recently, the DIPPR project6 developed a correlation over the
entire temperature range where the liquid-vapor phase transition exists (between the triple
and critical points) based on regression analysis of six different studies. The following
B E
A + + C ln T + DT
P = e T
(2.2)
With the coefficients listed in Table 2.5, it was deemed to be accurate to within 3% over
Table 2.5 Coefficients for nitromethane vapor pressure correlation (from DIPPR6).
Temperature (K) A B C E E
244.60-588.15 8.7411E1 -7.1332E3 -9.7786 7.9061E-6 2.0000
has been investigated by a number of researchers using a diamond anvil cell. The study
100 to 160 C, no transition was observed from the crystalline phase. A more recent and
the equilibrium melting point at very high pressures and temperatures. Condensed-phase
in the investigation of Courtecuisse, et al.19 Five distinct solid phases were identified
Thompson20 studied melting specifically over the range of 1 to 30 kbar (0.1 to 3 GPa).
Computed values were correlated by them using the Simon-Glatzel equation,21 which has
been shown to work well for non-metallic and non-ionic substances with normal melting
curves:22,23
P = aT b + c (2.3)
Coefficient values (yielding pressure in kilobars) found by them are shown in Table 2.6.
Block, and Miller18 at high pressure, extrapolation of the curve to low pressure (e.g. 1
atm) showed some differences from measured values. For the purposes of Figure 2.2, a
smooth curve was interpolated from the triple point, through the empirical melting point
Thompson.
16
Table 2.6 Coefficients of the Simon-Glatzel equation for the high-pressure fusion curve
of nitromethane (from Ref. 20).
Coefficient Value
a 1.597E-5
b 2.322
c -6.74
Although solid-vapor phase transition data was not found in the literature, the
sublimation line can be approximated with knowledge of the enthalpy of sublimation and
P H 1 1
ln = (2.4)
Po R To T
where Po is the vapor pressure at a known temperature To, H is the enthalpy of the phase
transition (sublimation in this case), and T is another temperature at which the vapor
pressure, P, is to be determined (or vice versa). To find the unknown pressure of phase
H 1 1
R To T
P = Po e (2.5)
The triple point temperature and pressure as given by the DIPPR project6 and listed in
Table 2.3 were used as the reference state since, by definition, all phases must be in
were found. However, using molecular dynamics simulation, Alavi and Thompson24
253.5 to 233.1 kJ/mol over that range. As the Clausius-Clapeyron relation as expressed
17
in Equations (2.4) and (2.5) assumes a constant value for the enthalpy of sublimation, in
addition to ideal gas behavior of the vapor, an intermediate value of 243 kJ/mol was
assumed to approximate the sublimation line. Some error was introduced by this
approximation, but it would be most significant at the lowest temperatures which are least
likely to be found in any practical application. In addition, the vapor pressure is already
so low that the error introduced by assuming a single constant property value will have
virtually no engineering significance. For example, the calculated vapor pressure at 100
K changed from 1.46E-73 to 7.31E-80 kPa as enthalpy of sublimation varied from 233.1
to 253.5 kJ/mol. Although this is a nearly a seven order of magnitude change, the
phase diagram has been developed for nitromethane and is shown here as Figure 2.2.
The temperature range spans the triple and critical points (roughly 200 to 600 K), over a
pressure of 10-5 to 107 kPa. Phase behavior in most conceivable practical applications
5
10
LIQUID
SOLID
3
10
Pressure (kPa)
VAPOR
1
10
Temperature (K)
Looking at the phase boundaries, one can see that nitromethane can be considered
storable at ambient conditions for most applications, though the freezing point may be an
issue at extreme conditions. The critical point lies within the temperature and pressure
1930s to early 1950s), interest in nitromethane waned when hydrazine and higher-
performance bipropellants were discovered and implemented. Low level interest with
more of a focus on fundamental chemistry and detonation phenomena continued for the
next few decades, while experimental instruments and analytical techniques grew
nitramines like RDX and HMX, due to the fact that as a simpler, but similar molecule,
nitromethane can be used to validate methods or techniques for use with more energetic
systems. With the continuing development of algorithms and numerical methods coupled
recent studies have again considered nitromethane (or mixtures containing nitromethane)
for use in propulsion systems. This seems to have been prompted by a number of factors,
replacement of hydrazine) and the perceived advantages that may be gained by use of
and sensitivity results at both atmospheric and high (60-90 atm) pressures were deemed
firings showed promise, but the study ended due to lack of funding and there was little
further activity recorded anywhere until the World War II time frame in the United
States.2
By the end of World War II, the Aerojet Corporation,27,28 the Jet Propulsion
Laboratory (JPL) at the California Institute of Technology,29 and the Navy Engineering
Experiment Station (EES) at Annapolis2 had all investigated the use of nitromethane as a
toxic, and storable propellant was widely recognized, and it was believed that the
sensitivity issues, while significant, could be overcome.30,31,32 Aerojet found that the
detailed injector and chamber design studies immediately following the war.33
rockets to be used in launching of buzz bombs. Ignition difficulties were overcome with
the use of a torch-type igniter. Although detonation sensitivity was found to be reduced
significantly with the use of certain additives such as gasoline, methanol, or amyl acetate,
unwanted initiation in the injector system under certain conditions remained an issue.
Rocket testing was extensive, with over 200 runs made at chamber pressures ranging
Aerojet also studied the kinetics of thermal decomposition and evaluated catalytic
methods of lowering the L* (ratio of chamber volume to nozzle area) necessary for
21
effective combustion in a rocket motor. It was found that stainless steel served as a
catalyst, increasing the reaction rate by a factor of 40. A number of materials, including
solve ignition and combustion stability concerns, resulting in the eventual development
At approximately the same time, liquid strand tests were conducted over a limited
range of conditions by Rice and Cole37 at the Naval Ordnance Laboratory (NAVORD) to
characterize the burning behavior. According to their report, these were the first reported
data on burning rate as a function of pressure. They used several horizontally mounted
lead break wires across a 6-mm ID Pyrex tube at different axial positions. The burning
rate of nitromethane was deduced from recorded break-wire signals in the pressure range
from 1,080-2,076 psig. It was postulated that the hump observed in the burning rate
curve near 1500 psi in Figure 2.3 (replotted from a poor quality microfiche reproduction)
may have been the cause of observed erratic behavior in contemporary rocket motor tests.
Burning rate data at higher pressures (up to 25,000 psig) were also obtained, but the
combustion chamber due to the mass addition of the product gases. Also, the timing
instruments used led to significant sources of error at higher pressures and burn rates.
Prior to the current work, the NAVORD report represents the first and only published
burning rate data for nitromethane, to the authors knowledge. Their results are directly
22
0.150
Burning Rate (in/s)
0.100
0.050
1000 1250 1500 1750 2000
Pressure (psi)
Causes and possible methods to reduce the high required L* were also
influenced the overall reaction rate, and the addition of catalysts which would promote
Koch, et al.39 and Boyer, et al.40 previously determined burn rate and temperature
sensitivity using the constant-feed liquid propellant strand burner described in this
Other than this work, the only reported attempt to model the linear regression of
simplified model of nitromethane ignition and combustion with a very reduced overall
23
reaction mechanism consisting of two consecutive first order steps. This mechanism
(described in greater detail in a later section) was combined with the CTEM one-
dimensional code of Zarko, et al.43 to find calculated transient and steady-state regression
rates. The results were stated to compare well with measured strand burner rates at their
observed low pressure limit (~0.02-0.03 cm/s below 1 MPa). However, no other data or
Only by adding a stoichiometric amount of air or oxygen could they obtain a stable
flame. In this configuration, they observed two relatively thin, distinct regions: an
orange/yellow zone followed by a violet region characterized by C2, CH, CN, NH, NO,
and OH emission. The flame structure lacked the blue initial zone with formaldehyde
de Jaegere and van Tiggelen.45 Only a single stage of combustion was found in their
study.
Eckl, Weiser, and Eisenreich (with co-workers) measured infrared (IR)46 and
decomposition of NM. The earliest investigations were focused on applications, but most
later ones investigated the just the chemical characteristics and properties. Many studies
group and a single nitro group) to find techniques and trends that could be applied to
availability, energetic nature, and other properties, nitromethane has been studied around
the world, including significant work found in Russian literature. Table 2.7 at the end of
the section summarizes conditions, findings, and kinetic parameters of the most of the
significant studies. Rates in the table have been standardized to use the three parameters
E
k = AT exp a (2.6)
RuT
where A is the pre-exponential factor in the appropriate units containing cm, mol, and s
performed by Taylor and Vesselovsky at New York University in 1934.49 In their study,
the pressure increase in a small sample volume containing nitromethane was monitored
as a function of time. Initial pressures examined ranged from 24.5 to 307 mm (3.27 to
40.9 kPa), with most cases below 200 mm. Initial temperatures of 380 to 420 C were
25
chosen to yield convenient decomposition times on the order of minutes. In all cases, the
final pressure was found to be about 130 percent of the initial value. By fitting an
Arrhenius expression to the elapsed times until one-fourth of the total pressure rise was
associated with the loss of an oxygen atom from the nitromethane molecule, creating
nitrosomethane:
CH 3 NO 2 CH 3 N O + O (2.7)
Subsequent reactions were deemed too complex to analyze with the limited empirical
data.
Some years later, Cottrell and Reid50 noted inconsistencies in the work of Taylor
and Vesselovsky,49 and performed a similar set of static experiments, but over a higher
initial temperature and pressure range of 380 to 430 C and 200 to 350 mm. They also
52, 700
k = 1014.36 exp sec
-1
(2.8)
RT
with a significantly lower activation energy than that previously determined by Taylor
and Vesselovsky. At pressures less than 200 mm, the rate constant increased with
pressure. Cottrell and Reid postulated that the initial decomposition step resulted from C-
CH 3 NO 2 CH 3 + NO 2 (2.9)
with activation energy of approximately 52 to 53 kcal. Further expanding their test
matrix in the continuing study, Cottrell, Graham, and Reid51 looked at slightly higher
26
pressures (up to 400 mm) and refined the result reported in Ref. 50. Here, the first order
53, 600
k = 1014.6 exp sec
-1
(2.10)
RT
They also re-analyzed the earlier data of Taylor and Vesselovsky and found it yielded a
similar fit. Using chemical analysis along with mass and infrared spectroscopy,
intermediate products were measured and a set of five likely intermediate reactions
identified.
pressure (4-40 mm Hg) and similar temperatures. The effect of other gases on
decomposition rates was also determined. At this much lower pressure, a significantly
and Kilpatrick used a flow reactor with nitrogen as a carrier.53,54 The reactor pressure
was atmospheric, and temperature was maintained at a constant value between 420 and
480 C. The rate expression obtained in this study was similar to that of Cottrell,
Graham, and Reid, and agreed with their conclusion that the activation energy of the rate-
limiting step in thermal decomposition is consistent with C-N bond scission. However,
Along with rocket motor tests, Kindsvater, et al.36 at Aerojet also studied the
effect of additives, they designed a glass reaction vessel that was placed in a furnace at a
27
specified temperature. The evolved gas was then measured and sampled. Among the
more notable results, it was found the decomposition reaction was first-order, with
activation energy of 53 kcal/mole, and stainless steel served as a catalyst, increasing the
Gray, Yoffe, & Roselaar55 obtained only limited quantitative results, but they
were consistent with a unimolecular decomposition resulting from C-N bond fission.
They interpreted the formaldehyde seen by Cottrell, Graham, and Reid51 to be formed
from multiple pathways involving reactions of the decomposition products CH3 and NO2;
elevated pressure was Mueller at Aerojet;56 although early studies in Italy had measured
the gas evolution rate from high-pressure reactions,25,26 no detailed chemistry was
minutes to 24 hours), then quenched at room temperature and the remaining species
analyzed. Initial pressures ranged from 180 to 300 psia. Based on the variations of
measured species with pressure, a brief reaction scheme was proposed that included both
N-O and C-N bond scission as parallel pathways for unimolecular decomposition; the
former was considered prevalent at high pressure, and the latter at low pressure.
results with previous low-pressure studies, conducted a series of tests using a similar
method, but at higher pressure (40 atm). A set of lower pressure (0.714 8.2 atm) runs
28
were also conducted to closely examine the effect of pressure on HCN formation. By
methyl radicals and nitric oxide to form nitrosomethane (an HCN precursor),
CH 3 + NO CH 3 NO (2.11)
they found the inclusion of a decomposition reaction involving N-O bond fission was no
longer necessary to explain the pressure dependence of observed results. Although only a
single rate constant at 40 atm was given, it was observed that the first-order
decomposition reaction was about half as fast at 1 atm. Again, the increase in rate with
low pressure in a shock tube using time-of-flight mass spectroscopy. Activation energy
(when calculated with first-order rate constants) was found to be a very strong function of
temperature (decreasing by almost three times) over the range examined, 872-1187 C.
k
ZZZ
NM + Ar YZZ1
X
Z NM* + Ar
k 1
k2
ZZZX
NM* + Ar YZZZ NM** + Ar (2.12)
k 2
NM** products
He also cited the pressure-dependent behavior seen by Cottrell, et al.51 at pressures below
200 mm, implying third-body interactions were important below this pressure limit. The
actual form of the resulting decomposition reaction was not inferred absent further data.
29
temperature shock waves were consistent with C-N bond fission, but no quantitative
detonation reaction zone as a function the velocity profile of stepped velocity driving
utilized with accurate equations of state. The reaction was modeled using a unimolecular
hydrodynamic equations were extended to two-dimensional forms, and both Eulerian and
Lagrangian formulations were examined, with no effect on stability found between the
two. The sensitivity of detonation reaction zone stability to different activation energies
Similar to Bradley,58 Borisov, Kogarko, and Skachkov62 also saw change of the
reaction rate at higher temperature. By correlating data from Refs. 49, 51, 54, 57, and 58
along with those from their own study, they found the reaction rate could be expressed
well for temperatures between 327 and 927 C by a rate constant expression containing a
2.7
57, 000 57, 000
k = 1011.2
exp sec
-1
(2.13)
RT RT
At temperatures over 927 C, the deviation from this expression was significant. They
also noted that when correlated with the other studies, Bradleys data appeared
30
anomalously low. High measured rates at high temperatures were attributed to the effects
of increased rates of secondary reactions consuming the CH3 and NO2 decomposition
measurements were not able to directly identify species, but times and optical density
CH 3 NO 2
k1
CH 3ONO
CH 3ONO
k2
CH 3O + NO
CH 3O + CH 3 NO 2
k3
NO 2 + CH 3OCH 3 (2.14)
NO + CH 3O
k4
HNO + CH 2 O
NO 2 + CH3O
k5
CH 2 O + HNO 2
where k1 was expressed by Eq. 2.10 (determined previously62) and k2 could be found
from methyl nitrite decomposition. At temperatures below 827 to 927 K, k1 was the
limiting rate, above that, methyl nitrite decomposition limited the overall decomposition,
leading to the shift in overall observed decomposition rate with temperature. It was noted
that the isomerization reaction to methyl nitrite could proceed either by direct molecular
recombination to CH3ONO; the two paths were would be indistinguishable under the
in static tests at the relatively low temperature of 375-440 C. They found pressure
dependence of the rate at pressures below 150 mm, and approximately first-order
31
behavior above. The decomposition reaction was assumed to consist of C-N bond
The unimolecular decomposition rate was calculated from kinetics of the back
decomposition rate, they used the reverse rate of Phillips and Shaw,66 confirmed by
Similarly to what has been done in this document for Table 2.7, Benson and ONeal
standardized the reaction rates cited by various researchers into two Arrhenius
parameters: the pre-exponential factor, log A, and activation energy, E. However, the rate
dependence outside the exponential that could not be accounted for with the two terms
available, leading to significant error between the original expression and that tabulated
including a temperature exponent with the pre-exponential factor and activation energy,
rate, although they observed that it continued above 400 mm. Their activation energy
of the seminal studies on nitromethane decomposition, they found the reaction rate to be
CH 3 NO 2 ( + M ) CH 3 + NO 2 ( + M ) (2.15)
Using the Kassel integral to fit the fall-off curve, the low- and high-pressure limiting rate
42, 000
k0 = [ Ar ] 1017.1 exp 3 -1 -1
cm mol s
RT
(2.16)
58,500 -1
k = 1016.25 exp s
RT
decomposed through C-N bond fission into CH3 + NO2, which then rapidly recombined
into CH3ONO and CH3O. Relatively low activation energy for decomposition (48.1
kcal/mol) was obtained from fitting to data from both literature and their study. The
techniques was performed by Alster, et al. in 1979.71 Using a modified version of the
the activation energy of the C-N bond breaking decomposition reaction to be 62.6
kcal/mol. Somewhat later, Dewar, Ritchie, and Alster72 performed a more detailed
(as low as 42.8 kcal/mol) reported in the literature reviewed by them to the calculated C-
N bond energy (60.0 kcal/mol) led them to believe that nitromethane decomposition did
33
not occur through simple dissociation into radicals, but rather via a rearrangement to
methyl nitrite, which has a much lower activation energy of dissociation, calculated by
them to be 33.0 kcal/mol. They found the activation energy of the rearrangement itself to
be 47.6 kcal/mol. The ratio of rearrangement to C-N bond fission at 700 C was
importance of the C-N bond fission. It should be pointed out that the literature cited by
Dewar, Ritchie, and Alster was very incomplete and did not include a number of studies
that reported higher activation energies at low temperature, bringing their conclusions
into question.
Based on their low temperature (400 to 500 C) static tests, Perche, Tricot, and
reaction with C-N bond fission using an activation energy of 57 kcal/mol. The effect of
pressure does not seem to have been examined. The pyrolysis process was observed to
findings of Dewar, Ritchie, and Alster,72 McKee found the rearrangement barrier was
higher than the C-N bond energy, and therefore concluded that reactions must either be
bimolecular or proceed through dissociation to CH3 and NO2. He also pointed out the
surprising choice of the MINDO/3 method by Dewar, Ritchie, and Alster since more
nitromethane.
34
Kelzenberg, et al.42 Using a heatable infrared cell, they found activation energy of 55
kcal/mol.
to yield rapid heating and temperature conditions similar to those in shock tubes.
Nitromethane diluted in both N2 and Ar at mole fractions between 0.064 and 0.112 was
compressed to final temperatures between approximately 700 and 1100 K. The results
using Ar as a diluent were judged to be not reliable; the much lower compression ratio
required to reach a given temperature point due to higher specific heat ratio of monatomic
gas led to large experimental uncertainties. At a given comparison condition (952 K), the
observed decomposition rate for nitromethane in nitrogen was consistent with the first-
order C-N fission decomposition reaction rate as calculated from Benson and ONeal,65
five times above that of Crawforth & Waddington64 and one-fifth that of Glnzer and
Troe.69 Although reaction rates fell within the range seen in the literature, product
fractions, especially high levels of methane, were at variance with studies at similar
conditions (i.e. shock tubes) in literature which showed virtually no methane. Possible
that there were three possible decomposition pathways, in contrast with only two
35
identified by experiment at that time. One was the classic C-N bond scission, while two
went through intermediate formation of methyl nitrite (CH3ONO), which then split into
either CH3 and NO2 or CH3O and NO. However, it was recognized that energy
predictions for the transition states did not match those made by ab initio methods, and
since the reaction products of the newly-identified methyl nitrite pathway with a
relatively low energy barrier (47.6 kcal/mol) resembled those of the C-N scission
Roszak and Kaufman80 computed the C-N bond dissociation energy using a large
initio calculation. Although they stated the calculated value of 64.31 kcal/mol was in
good agreement with experimental values, it exceeded the cited measurements tabulated
determine bond dissociation energies. Unimolecular C-N bond fission was assumed to be
the initial decomposition step, with the dissociation energy calculated to be 58.9
kcal/mole.
Zeman82 correlated the empirical pyrolysis activation energy and oxygen balance
for a large number of nitroparaffins, classified into groups based on molecular structure.
From the correlations, interpolated Arrhenius parameters for reaction were determined.
energies to intermolecular steric strain energy was also found for aliphatic nitro
compounds that could allow simpler approximation of activation energy for similar, but
temperature in a flow reactor. They found no reaction at temperature below 440 C, and
was directly derived from the experimental measurements, but results were successfully
modeled using a mechanism created by adding two reactions (nitromethane C-N bond
Unfortunately, the reference as cited by the authors for the first two reaction rates does
not exist, and it is unclear what expression they actually used for the rate of the initiation
conditions near those of detonation. In the pressure range (170 to 880 kPa) used, they
considered the decomposition reaction to behave as bimolecular with the carrier gas
(similar to Hsu and Lin87) and calculated the rate constant accordingly from their
measured data.
studied by Zaslonko, Petrov, and Smirnov88 in a shock tube over a wide range of
pressures using a spectrophotometric technique. The measured reaction rates were found
37
diluent gas. Limiting rates and parameters of the fall-off reaction were found, using the
modified Kassel integral89 to construct the transition curves between low- and high-
pressure reaction limits. The rates were similar to those of Glnzer and Troe,69 although
the parameters of the Kassel integral fall-off curve differed significantly. This was
attributed to a lower number of data points and simplified fitting procedure in the original
work of Glnzer and Troe. Zaslonko, Petrov, and Smirnov also determined the
effectiveness of Ar, He, CO2 and CF4 as collision partners, finding relative efficiency of
degree of pyrolysis based on UV absorption of an axial beam through the shock tube. A
fit of the experimental data found a relatively low reaction rate not consistent with that of
Glnzer and Troe69 However, significant data scatter was noted, and in the proposed
mechanism, the high-pressure (unimolecular) rate of Glnzer and Troe was adopted. The
reaction, R2, printed in their article appears to be an odd choice. The explanation in the
text does not make clear where the parameters are obtained; in fact, the purportedly
explanatory figure (their Fig. 2) makes the choice appear rather arbitrary. Looking just at
the rate, the pre-exponential value given (1.26E14) may be a misprint, and should
probably correspond to that of Glnzer and Troe69 (1.26E17), which is how it was listed
as in a subsequent publication by the same authors.91 The later paper also added a
38
T T
Fcent = 0.3732 exp + 0.6268exp (2.17)
250 901
It was calculated from the experimental data of Glnzer and Troe to better approximate
rates in the fall-off region, since techniques that used this parameter were not available at
The work of Manaa and Fried93 showed calculated C-N bond energy was very
Glarborg, Bendtsen, and Miller94 analyzed experimental shock tube data of other
researchers (Glnzer and Troe69 and Zaslonko, Petrov, and Smirnov88) and refined the
rate constant of the decomposition reaction based on this analysis. They determined that
the Glnzer and Troe low- and high-pressure limits were consistent with all the shock
with two different basis sets to find C-N bond dissociation energies for a set of
nitroalkanes. The nitromethane C-N bond dissociation energy at 298 K was found to be
55.827 kcal/mol with the 6-31G(d) basis set and 53.920 kcal/mol using the 6-
hopes of resolving the issues raised by previous studies and pointed out that the pyrolysis
kinetics are actually quite complex for the seemingly simple molecule. It was noted that
the analysis of experimental work by Wodtke, Hintsa, and Lee97,98 relied on frequency
making their conclusion that isomerization to methyl nitrite had a lower barrier than C-N
bond breaking (contradicting the results of McKee76) very suspect. (A later work by
McKee99 using improved computational methods, once again found that the C-N bond
fission was favored.) Using a different quantum mechanical model than previous
computational studies, Nguyen, et al. found that the C-N bond dissociation energy was
602 kcal/mol, consistent with many experimental findings. This did not match the
MCSCF results of McKee99 or Saxon and Yoshimine100, but partially agreed with Hu, et
al.101 Nguyen, et al. concluded that although their results showing that C-N bond fission
was preferred were consistent with experimental data where available, much additional
the same issue, namely the balance between C-N bond breaking and isomerization to
methyl nitrite. Calculations using two different levels of theory showed that in all cases,
dissociation was the preferred reaction. The isomerization transition state found
resembled that of McKee (1986),76 while the lower energy transition states reported by
McKee (1989)99 and Saxon and Yoshimine100 appear to have been computational artifacts
theory techniques to compare the energies needed for C-N bond breaking to that
necessary for rearrangement of the nitro group to a nitrite form for a selection of
be the most reliable, giving a range of 65-65.7 kcal/mol for rearrangement. Comparing
these values to those of C-N dissociation, they concluded that the probability of
From the studies reviewed here, it can be seen that historically there is no
results have been reported from both empirical and computational investigations.
C-N bond fission with activation energy around 55-60 kcal/mol, but measures of pressure
dependence vary. The rates of Glnzer and Troe69 seem to be fairly widely accepted, but
they provided no definitive accounting for the fall-off region, although characterizing its
decomposition, both Zhang and Bauer91 and Glarborg, Bendtsen, and Miller94 calculated
methyl nitrite, and the conditions under which that isomerization could be important.
Computational studies have shown varied outcomes based on technique and level of
theory. Low measured apparent activation energy under certain conditions has been
41
separate reaction pathway other than C-N bond fission (which could have a lower
energy). Recent results seem to indicate the C-N bond fission is energetically preferred
over rearrangement, but questions remain. These issues can only be addressed with
further computational and experimental studies on the behavior and properties of the
nitromethane molecule.
Table 2.7 Mechanisms and reaction parameters for the thermal decomposition of nitromethane.
Temperature Pressure Ea
Investigators Year Method Mechanism A* Notes
(C) (Initial) (cal/mol)
Taylor and 1935 380-420 25-300 mm static N-O bond fission 61,000
Vesselovsky49
Cottrell and Reid50 1950 380-430 200-350 mm static C-N bond fission 2.29E14 52,700
Frjacques52 1950 310-440 4-40 mm static C-N bond fission 2.5E11 42,800
Cottrell, Graham, 1951 380-430 200-400 mm static C-N bond fission 3.98E14 53,600
and Reid51
Hillenbrand and 1951 420-480 45 and 135 mm flow C-N bond fission 4.07E13 50,600 pressure of NM + N2
Kilpatrick53 carrier gas was 745 mm
Kindsvater, et al.36 1951 391-418 77-150 mm static unspecified first-order 7.8E15 53,000 several effective
catalysts identified
Hillenbrand and 1953 420-480 45 and 135 mm flow C-N bond fission, with 2.70E13 50,000 pressure of NM + N2
Kilpatrick54 some intramolecular 1.35E13 580 carrier gas was 740-750
rearrangement mm
Gray, Yoffe, and 1955 430-485 ~4 mm flow C-N bond fission 50,000 limited quantitative
Roselaar55 (0.04 mole% data, NM with He
NM) carrier gas at
atmospheric pressure
Mueller56 1955 355 180-300 psia static parallel N-O and C-N N-O bond fission
bond fission thought to predominate
at higher P, C-N at
lower pressure
Makovky and 1959 312-340 40 atm static C-N bond fission 5.37E13 49,200
Gruenwald57
Bradley58 1961 872-997 6-10 mm of 10 shock tube C-N bond fission 35 lower Ea at higher T,
997-1187 % NM/90% Ar 13,000 explained with multiple
7,000 activation steps
Hiraoka and 1963 1147-1587 0.0091 to 0.0133 shock tube C-N bond fission No quantitative
Hardwick59 mole fraction of determination of kinetic
NM in Ar parameters
42
Temperature Pressure Ea
Investigators Year Method Mechanism A* Notes
(C) (Initial) (cal/mol)
Borisov, Kogarko, 1966 427 -1027 0.11.5 atm NM static, adiabatic C-N bond fission 1.72E23 -2.7 57,000 parameters calculated
and Skachkov62 in Ar at Po = 1 compression, from given correlation,
15 atm shock tube rate is higher at
T > 927 C
Borisov, Zaslonko, 1968 777-1027 0.7-14% NM in shock tube Isomerization to 23,000 Ea given is for high T
and Kogarko63 Ar at 10-50 mm CH3ONO, then 4000 reaction (lower Ea at
initial P decomposition to lower T)
CH3NO and NO
Crawforth and 1969 305-440 50 - >200 mm static C-N bond fission 1.26E14 55,200 pressure-dependent A &
Waddington64 900 Ea (lower at P < 150
mm)
Benson and 1970 calculation C-N bond fission 3.98E15 59,000 calculated from reverse
ONeal65 reaction data of Refs. 66
and 67
Dubikhin, Nazin, 1971 360 and 390 200-500 mm static C-N bond fission 2.19E+14 54,300
and Manelis68
Glnzer and Troe69 1972 627-1227 0.01-0.3% NM C-N bond fission; fall- ko: 1.26E17 42,000 conditions in falloff
in 1.5E-6 off reaction k: 1.78E16 58,500 range, Kassel integral
3.5E-4 mol/cm3 500 for curve
Ar
Zaslonko, et al70. 1972 757-1307 70 and 120 mm shock tube C-N bond fission 6.31E12 48,100 some pressure and
initial, 1-2.8 atm temperature dependence
after shock seen
Perche, Tricot, and 1979 400-500 static C-N bond fission 6.62E14 57,000
Lucquin75
Alster, et al.71 1979 computational C-N bond fission 62,600 modified MINDO/3
method
Dewar, Ritchie, and 1985 computational mostly rearrangement 2.00E13 47,600 MINDO/3 method
Alster72 to CH3ONO
43
Temperature Pressure Ea
Investigators Year Method Mechanism A* Notes
(C) (Initial) (cal/mol)
McKee76 1986 computational C-N bond fission or 57,450 bond dissociation
bimolecular reaction energy found using
MP2/6-31G*+ZPC//6-
31G* method
Krause, et al.77 1987 1 atm(?) heated cell C-N bond fission 1.00E14 55,000 from summary in Ref.
42
Cohn78 1990 638-812 0.064 to 0.100 adiabatic C-N bond fission k varied from 29-600 s-1
(Ti = 39) mole fraction in compression at T=922-1085 K; at
N2 at 0.842 bar 952 K, k=120 s-1
initial (similar to Ref. 65)
Melius81 1990 computational C-N bond fission 58,900 bond dissociation
energy found using
BAC-MP4 method
Rice and 1990 computational C-N bond fission and multiple pathways for
Thompson79 two pathways for decomposition are
isomerization to possible
CH3ONO
Roszak and 1991 computational C-N bond fission 64,310 bond dissociation
Kaufman80 energy found using
MRD-CI method
Zeman82 1995 calculated C-N bond fission 2.00E14 54,500 from correlation to
oxygen balance of
similar compounds
Khrapkovskii, et 1996 computational C-N bond fission 56,900 MINDO/3 method
al.83
Djebaili-Chaumeix, 1997 787-1057 0.2-1.5% NM in shock tube bimolecular (with Ar); 1.33E18 52,160 at conditions of study,
Abid, and Paillard86 Ar, 170-880 kPa C-N bond fission decomposition acts as
after shock bimolecular reaction
44
Temperature Pressure Ea
Investigators Year Method Mechanism A* Notes
(C) (Initial) (cal/mol)
Zaslonko, Petrov, 1997 777-1127 0.01-0.5% NM shock tube C-N bond fission ko: 2.00E17 41,990 Kassel integral falloff,
and Smirnov88 in Ar, He, CO2, k: 1.58E16 59,700 3rd-body efficiencies
and CF4, 0.2-40 Ar:He:CO2:CF4 =
atm after shock 1:2:3:3
Zhang and Bauer90 1997 727-827 0.6% NM in shock tube C-N bond fission 1.977E12 38,000 fit to experimental data;
300-400 Torr Ar 3,000 for modeling used
Glnzer and Troe69 rates
Manaa and Fried93 1998 computational C-N bond fission 49,900 multiple computational
56,200 techniques
Glarborg, Bendtsen, 1999 727-1127 0.5-6 bar reinterpretation C-N bond fission ko: 1.3E17 42,000 Glnzer and Troe69 rates
and Miller94 of literature k: 1.8E16 58,500 with Fc=0.183
data
Zhang and Bauer91 1999 627-1227 correlation to C-N bond fission ko: 1.26E17 42,000 Glnzer and Troe rates
literature data k: 1.78E16 58,500 with Fcent = 0.3732
exp(-T/250) + 0.6268
exp(-T/901)
Khrapkovskii, 2001 -273.15 computational C-N bond fission 54,775 bond dissociation
Chachkov, and 25 55,827 energy found using
Shamov95 B3LYP method
Hu, et al.101 2002 25 computational C-N bond fission 61,900 G2MP2//B3YLP/6-
311++G(2d,2p) method
Nguyen, et al.96 2003 computational C-N bond fission 60,000 multiple computational
2000 methods, computational
discrepancies persist re
rearrangement
Arenas, et al.102 2003 computational C-N bond fission 59,000 CASPT2 asymptotic
limit
*Units of A are cm, mol, and s as appropriate for reaction order.
Note: Where required, pre-exponential A-factors have been converted to standard scientific notation and activation energies to
cal/mol for comparative purposes.
45
46
their studies, as the actual measured data were often limited and not directly quantitative
for many species and reactions. These models were generally tailored to the particular
conditions of their studies (especially in rate and choice of expression for the
outside of those investigated. Some of the more comprehensive models were used as
stepping stones in the development of a more complete nitramine mechanism. Only one
decomposition), but it was extremely simple. A list of the mechanisms found in the
literature and a brief summary of their features are shown in Table 2.8, with more
decomposition of nitromethane into CH3 and NO2 was chosen as an intermediate between
the Glnzer and Troes low- and high-pressure limits of 42 and 59 kcal/mole,
respectively.69 The resulting rate constant was verified against their earlier static
overall species formation rates. Although discrepancies were seen in some species
concentrations and profiles with time, overall agreement was good, and the rate constants
of some of the rate-limiting steps (such as reaction of CH3NO2 with OH) were refined.
In modeling their shock tube study, Hsu and Lin87 calculated a single rate constant
Over the limited pressure and temperature range of their experimental data, the
step of Glnzer and Troe69 with parameters modified for conditions of the study:
CH 3 NO 2 + M
k1
CH 3 + NO 2 + M
44, 000 (2.18)
k1 = 6.0x1016 exp
RT
The 37-step mechanism was formulated specifically to model the NO and CO species
measured in the study; other species were utilized only as computational intermediates.
The calculated species profiles were NO and CO found to fit measured values very well.
Using the proposed mechanism, most NO was found to form quickly from the
recombination of CH3 and NO2 fragments, while CO came from later reactions via
CH2O.
nitromethane at 1-10 atm using a shock tube with mass and infrared spectroscopy.
Special emphasis was placed on understanding the role of the NO2 group in ignition. A
detailed mechanism of 48 reactions was developed to model the process, and calculated
induction times showed good agreement with measured values. Although Glnzer and
Troes69 low-pressure reaction was used for the initiation step, a fit of the calculated
induction times yielded a global Arrhenius parameter of 38.6 kcal/mol, implying that the
the carbon chain reaction took two major parallel pathways to CH2O, through methoxy
(CH3O) and CH2NO2 radicals. Formaldehyde was converted to HCO and finally CO. It
49
was noted that the endothermicity of breaking the C-N bond prevented ignition from
taking place in the adiabatic system until all the nitromethane was consumed. The chain-
NO 2 + H OH + NO (2.19)
decomposition. The list of reactions was developed through review of the literature and
included, but were not limited to, earlier nitromethane decomposition studies,74-87
validated using CHEMKIN115 against the shock tube results of Hsu and Lin87 and
Guirguis, et al.,105 and the static study of Tricot, Perche, and Lucquin73,104 The effect of
adding the isomerization reaction of nitromethane to methyl nitrite was examined. Using
parameters from the infrared multi-photon dissociation study of Wodtke, Hintsa, and
Lee116 (N.B. the parameters calculated by Wodtke, Hintsa, and Lee were later questioned
by Nguyen, et al.96), the influence on the overall reaction was found to be small, but
noticeable. At the conditions of Hsu and Lin,87 addition of this reaction accelerated the
abstraction reactions,
CH 3 NO 2 +H CH 3 NO 2 +H 2 (2.20)
CH 3 NO 2 +OH CH 3 NO 2 +H 2 O (2.21)
Differences from ideal gas behavior were found to be significant, although the rate-
controlling reactions were the same. A sensitivity analysis found that reaction between
CH 3 NO 2 + CH 3 NO 2 CH 3 NO + CH 2 O + HNO 2 (2.22)
comprehensive.107
were examined.
51
Initial studies118,119 were able to adequately match the characteristic delay times
for the first stage of high-temperature ignition (data of Guirguis, et al.105), but resulted in
times that were orders of magnitude too slow for second stage. Much better agreement
was obtained after updating a number of reactions, especially those of NO, HNO, NO2,
and HCN based on evolving knowledge of reaction including those species. With the
improved mechanism, both ignition stages matched well.107 However, Guirguis, et al.
only reported multiple (i.e. two-stage) ignition pressure spikes in two out of seventeen
experiments, raising the possibility that multiple stages may not be distinctive (at least at
the level of sensitivity of the measurement techniques used) for most of the conditions
Melius concluded that although the global ignition process could be separated into
ignition), the actual chemical process was very complex. The slow preliminary reactions
reactions involving HNO. The net result after the first-stage ignition was the overall
pathways, with limited production of N2 contributing heat release. All nitromethane was
consumed by the end of the first stage ignition. The exothermic intermediate stage was
characterized by the slow consumption of CH4 and NO, with generation of HCN. When
the temperature became high enough, reaction rates rapidly accelerated in a runaway
fashion, creating the second-stage ignition. Here, the majority of heat release occurred,
converting NO to N2.
52
conditions was also conducted. When compared to the experimental ignition delay data
of Lee, et al.108 at 10 and 50 kbar, and low temperatures (400 to 600 K), significant
differences were found. Calculated delay times at those conditions were longer than
measured, while extrapolating the data of Lee, et al. (using an Arrhenius expression) to
higher temperatures (i.e. 1100 to 1300 K) yielded calculated rates much shorter than the
trend of experimental values would show. Melius postulated that this indicated a change
gas effects (as examined by Yetter and Rabitz117) were not discussed, and no further
More recently, Zhang and Bauer90 studied the high-temperature decomposition (at
1000-1100 K) using shock heating in a shock tube. Building on earlier works, they
developed a mechanism consisting of 99 reactions and 41 species to match both their data
and published results of other research groups. This mechanism was based on ones
53
previously proposed by Hsu and Lin87 and Perche, et al.,74 combined with other rate
expressions from the NIST Chemical Kinetics Database.123 The resulting set of reactions
and parameters was validated and adjusted based on static vessel experimental data from
Perche, et al.73 and shock tube data of Hsu and Lin87 and Glnzer and Troe.69 An
secondary reactions involving free radicals such as H, OH, and CH3. Semi-quantitative
results were reported. Also, a number of important reactions with unreliable rate
mechanism and eliminated 20 species and 49 reactions. However, they indicated that
concentrations calculated with the reduced model varied up to 5% from the more detailed
reaction kinetics. Examination of the reaction set shows that it was developed solely for
reactions.
specific subset consisting of 19 reactions (one of them duplicating an earlier reaction, but
with a modified rate constant) were added to their previously-developed mechanism for
SENKIN121 to experimental shock tube data of other researchers (Glnzer and Troe,69
54
Hsu and Lin,87 and Zhang and Bauer90) that covered relatively high temperatures (1000-
Based on comparison with the NO2 profile of Glnzer and Troe, they adjusted the
rate constant of the CH3 and NO2 recombination reaction (R14 in their manuscript):
CH 3 + NO 2 CH 3O + NO (2.24)
to better match the experimental data. This reaction was one of the ones recommended
important, as well. However, unlike the conditions of Guirguis, et al.,105 at the highly-
diluted conditions of the shock tube studies evaluated, no ignition was seen and the
reactions did not proceed to final products. Good agreement was seen between
calculations and the time profile of NO and CO concentrations measured by Hsu and Lin.
The nitromethane and methane concentrations of Zhang and Bauer were matched very
while, while the agreement with individual C2 hydrocarbons (C2H4 and C2H6) was not as
good, although the combined levels were similar. At low temperatures and atmospheric
CH 3 + NO 2 (+ M) CH 3 NO 2 (+ M) (2.25)
was determined to be competitive with Eq. (2.24), although its rate decreased rapidly
k1 k2
A BP (2.26)
and P the products. The initial step was considered to be controlled by C-N bond fission,
so the kinetic parameters for the unimolecular decomposition found in a previous study77
were used for the first step. Parameters for the second step were obtained by choosing
values that would match the final results of CHEMKIN II when run for adiabatic
combined with the gas phase CH4/NO2/O2 mechanism of Miller and Bowman.109 Kinetic
parameters for the two reactions are listed in Table 2.9. It was noted that profiles of the
mechanism, and postulated that NO consumption was the rate-determining step for the
Since the combustion processes examined in this study occur on relatively slower time
scales, larger length scales, and lower pressures than shocks, this shock literature was not
reviewed in detail. A few of the investigations are cited here as an example of the some
The work of Shaw, et al.122 showed that the thermal explosion temperature of
liquid nitromethane under static conditions was a strong function of pressure; 354 C at 1
kbar, 276 C at 10 kbar, and 184 C at 50 kbar. The apparent activation energy was also
inversely pressure dependent; Ea = 127, 22, and 20 kcal/mole at 1, 10, and 50 kbar,
respectively. At very high pressures, hydrogen transfer was suggested as the rate-
controlling step, instead of C-N bond fission, since the A-factor rate constant at 50 kbar
was more than 107 times greater than that of C-N bond fission (as given by Benson and
ONeal65). A significant isotope effect was noted when nitromethane values were
authors assertion that C-N fission is not the rate-determining step for high-pressure
liquid decomposition.
reaction zone at high detonation wave pressures up to 14 GPa. The coupling of a time-
giving a record of the chemistry as a function of time and space. Both neat and sensitized
13
(up to 5% diethylenetriamine) nitromethane were tested in normal, CH3NO2, and
CD3NO2 forms. Although the authors did not feel the data was sufficient to conjecture on
the nature of detailed reaction mechanisms, they noted that the first step in the fast
containing numerous nitrogen atoms. At the high densities found in the experiments,
faster reactions involving radicals could be suppressed by high pressure and limited
diffusion.
Winey and Gupta126 also examined shock-induced decomposition. Very high (up
to 19 GPa) shock pressures were used to initiate both neat and ethylenediamine-sensitized
nitromethane. A mechanical shock was created by the impact of a projectile fired from a
light-gas gun. Subsequent reflections of the shock from opposing windows bounding the
spectral changes in the UV/Visible absorption spectrum between 300 and 550 nm were
used to infer induction times and changes in chemical species. Measured induction times
as a function of pressure and initial temperature were consistently shorter than those
attributed this to the greater sensitivity of the absorption technique to earlier stages of the
was found.
58
They found the decomposition to be consistent with unimolecular processes, with C-N
bond breaking to form CH3 and NO2 dominant at near-threshold energies and
higher energies.
test surface in a vacuum. Adsorbed species and gas-phase fragments were sampled. It
was found that the rate-limiting step for the catalyzed decomposition was not the C-N
bond breaking, but instead the N-O bond scission. Therefore, it was determined that a
material that would weaken the N-O bonds when adsorbed could yield a more effective
catalyst.
Since first being considered in early rocket studies, nitromethane has been
large-caliber guns to IC engines. Theoretical and applied research has been conducted
for use in pulse detonation engines,129 pulse combustion rockets,130 and meso- or micro-
scale thrusters131,132 and propulsion systems.133,134 It has also been considered as fuel for
When the US Army was considering using liquid propellant for its next-
59
generation 155-mm gun, a study was performed by Knapton, et al.136 at the Ballistic
below. Required ignition energy was found to be high, but acceptable, and some
40-mm gun.138
that of gasoline or methanol, leading to usage in applications where high energy output is
of paramount importance such as drag racers (Top Fuel), pulling tractors, and racing
output hobby engines for model planes and vehicles. It can be run at a significantly
richer air-fuel ratio than gasoline because of the oxidizer included in the nitromethane
60
engines, engine lifetime is often limited to tens of seconds, at best.139 Although much
anecdotal evidence can be found, few published scientific studies have been conducted
associates140,141 in which they examined the power increase generated by the use of up to
20% nitromethane as a fuel additive for piston engines. Noting that the limiting factor
was knock (as apparent octane number of fuel mixture decreases markedly with
than with hydrocarbon fuels due to the higher optimum fuel-air ratio of the alcohol.
Bush, Germane, and Hess142 closely examined the effect of engine operating conditions
and fuel blends on knock. With the use of carefully-chosen engine conditions, up to 80%
NM in methanol was run successfully and conditions were identified that would retain
power output while lowering combustion stress. A recent patent143 has also made use of
In the last few years, there have been a number of studies on the physical
miscibility,144 densities and viscosities,145 and densities and speed of sound.146 Although
no application to propulsion was claimed by the authors, the inclusion of butanol in these
To modify physical and energetic properties for use in new or novel systems, a
variety of additives have been examined as gelling agents. In earlier works, various
gelling agent (with no inert gellant needed) has become of interest,155 and this technology
CHAPTER 3
In this chapter, experimental equipment and methods used for this study are
variety of strand burners available at the High Pressure Combustion Lab were employed
combust the propellant at desired pressure and environmental conditions. Burning rates
were measured in two ways: with the propellant feed to a stationary free surface, and
with the burning surface regressing into a quiescent fluid contained in a tube.
psig.
At lower pressures (below 5,000 psig), tests could be run with either regression
into quiescent propellant in a tube or propellant fed to a stationary burning surface. The
feeding system built into the Liquid Propellant Strand Burner (LPSB) allowed optical
access and a long burn time (on the order of minutes). In addition, initial propellant
Ultra-High Pressure Strand Burner (UHPSB) was designed and constructed for static
tests.
63
utilized to study the burning behavior of nitromethane. After being developed for the
nitromethane study, this device was later used to investigate flashback behavior of XM46
system is shown in Figure 3.1, and details of the chamber and chamber base with feed
Figure 3.2 Cross-sectional view of the windowed LPSB chamber with feeding base
installed.
65
Figure 3.3 Cross-sectional diagram of LPSB chamber base and feed system outlet.
The optically accessible windows of the LPSB allowed visual observation of the
whole combustion event and the usage of non-intrusive diagnostic techniques. Although
the test chamber itself could sustain pressure up to 70 MPa, all tests were conducted in
66
this chamber with pressures less than 35 MPa (5,000 psig) due to the limitations of the
optically accessible windows. Tests at higher pressures were performed in the ultra-high
Although the feeding system itself was installed as part of a previous project,
operational procedures and extensive modifications to fully utilize it were developed for
the current study. The temperature conditioning system and flame extinguisher were also
LPSB. In static tests, experiments were conducted by filling liquid propellant in a quartz
paper. Each method had its own advantages and disadvantages. By using the quartz
tube, the propellant flame was confined in a one-dimensional reaction zone and the
combustion product gases forced to accelerate in the vertical direction only. On the other
minimized heat loss from the propellant flame to the tube wall, thereby allowing the
liquid propellants to burn at their intrinsic burning rates. For nitromethane, both were
utilized and negligible difference was noted between the two types of containment.
To limit the amount of liquid propellant (LP) in the test cell, only the LP reservoir
cylinder and a feeding tube leading to the chamber were filled (see Figure 3.1) during the
test. Except when loading, bulk liquid propellant was stored away from the LPSB in a
into the feed system for each test. At nitromethane regression rates, this was sufficient to
During tests, the linear actuator drove the piston into the cylinder, forcing LP up
through the feeding tube into the chamber. A section of quartz tube about 2.5 cm long
with a 4.83 mm ID was connected to the tip of the feeding tube to reduce heat loss to the
metal feeding tube and assist in flame stabilization. In addition, this quartz tube enabled
visual tracking of the LP surface position when the LP was pushed upward prior to
ignition as well as during burning. Once the stable flame was established, one could
To achieve ignition of the LP, a nichrome wire was placed across the exit plane of
the quartz tip prior to each test. The nichrome wire was heated by supplying AC electric
energy to the wire; ignition of LP was achieved through heat transfer from the electrically
heated wire to the LP. To eliminate any interaction with the LP flame and to avoid
obstruction of the viewing area, the nichrome wire was stretched by springs from both
ends. When ignition was accomplished, the springs pulled the ends of the broken
nichrome wire away from the flame region and viewing ports. In the case that the
temperature near the pyrolyzing propellant surface region was not high enough to break
the nichrome wire, the electric energy supplied to the wire was quickly increased after
ignition was achieved. Increased temperature resulting from the electrical energy surge
melted the nichrome wire and caused it to break. The springs then retracted the two
utilized to determine the effect of initial temperature on the burning rate. This
surrounding the propellant feeding tubes. Conditioned fluid from a regulated temperature
bath/circulator flowed through the coolant jackets thus regulating the initial temperature
of the propellant. The cooling bath used for this system was the NESLAB ULT-80 Low-
Temperature Bath/Circulator. The stabilized temperature range for this circulator bath is
-80 to +18 oC with temperature stability of 0.03 oC. Although the unit can circulate
higher temperature fluids (up to 80 oC), the internal heating and control system cannot
maintain liquid at those temperatures. In order to condition the propellant at the ambient
and high temperature ranges, a supplemental heat exchanger system was constructed and
placed in the NESLAB liquid bath. With this feature added to the NESLAB liquid bath,
the temperature range for the test sample could be controlled between 80 and +80 C.
When Dow Chemical Syltherm XLT silicone heat transfer fluid was utilized as the
working fluid, the practical temperature range for nitromethane varied from -30 C
The liquid propellant would have regressed into the feeding tube after ignition if
no further efforts were made to prevent this from occurring. The unique feature of this
LPSB setup is that the LP burning surface can be stabilized at the exit plane of the quartz
tube during burning. To achieve this, the LP was fed into the tube at its regression rate
69
Two methods of setting the feed rate were available: a pulsed mode that varied
the speed depending on an external input and continuous run at a preset rate. Since the
burning rate of the LP at a particular pressure was not known a priori, adjustment of the
feeding rate during testing was necessary. This was achieved by programming the
voltage input. The voltage was controlled manually by the operator turning a
potentiometer connected to a 9-Volt battery. The operator adjusted the voltage output
according to the real-time video images of the LP surface location. The voltage output
from the potentiometer was connected to an I/O module on the controller that read in
analog data at a fixed rate of 10 Hz. Through this interactive operation procedure, a
It should be noted that although the above procedure showed a relatively steady
position for the LP burning surface, it was found that this method resulted in a
"pulsating" liquid motion near the tip of the quartz tube. This was due to a limitation in
the feeding systemthe linear actuator and controller could not operate at an infinite
stage was involved. Every time the I/O module sampled the user input (sampling rate =
voltage input for a preset time, then decelerated while waiting for the next input cycle.
The pulsating motion became more evident at higher pressures and feed rates.
therefore the intrinsic burning rate at a particular chamber pressure, a procedure which
involved two series of tests was created and implemented. In the initial set of tests,
burning rate measurements using the user-input stepping motion at the specified pressure
were conducted. Obviously, the constant speed read from the user input did not represent
the true burning rate; however, the results served as a guideline for more accurate
In the continuous feed tests, the chamber pressure was initially set equal to that of
programmed to move at a constant speed rather than a stepping motion. The speed was
usually set at a level lower than the steady-state level obtained in the variable-speed tests,
based upon the controller sampling procedure described earlier. Once the ignition wire
was energized, the motion of the linear actuator was initiated. The regression rate of the
LP at the initial chamber pressure may not have matched the set feeding rate of LP
immediately. Since the linear actuator did not allow speed adjustment during motion,
chamber pressure was adjusted to match the pressure-sensitive LP burning rate to the
actuator feeding rate. When the observed burning surface position started to move
downward into the clear quartz tip, chamber pressure was reduced to decrease the LP
burning rate. If LP started to bulge above the quartz tip exit plane, chamber pressure was
raised to increase the LP burning rate. The chamber pressure was manually adjusted by
controlling the purge flow into the chamber and out of the exhaust line shown in Figure
3.1. When the LP burning rate at the adjusted chamber pressure matched the LP feeding
rate of the linear actuator, the burning surface was stable at the exit plane of the quartz
71
tube. The pressure level at that condition was then paired with a corresponding burning
rate deduced from the linear actuator velocity setting. Once a propellant was partially
characterized, the burning rate could be extrapolated from previous data with a standard
power-law expression to provide a better initial prediction for the initial continuous feed
The LP flame was extinguished before the motion of the linear actuator was
terminated; otherwise the flame could propagate into the feeding tube. Flame
extinguishment was achieved by blowing out the LP flame with nitrogen gas jet injection
aimed toward the LP burning surface region. Just prior to desired end of the test run, the
chamber pressure was lowered, causing a liquid bulge to form above the quartz tube exit,
raising the flame position. Since the LP mass feeding rate was constant, the burning
surface area increased to compensate for the lower linear regression rate. When the
extinguisher gas stream was initiated, it drove the flame away from the surface region,
causing drastic reduction of heat feedback from the gas phase to the LP surface, thereby
extinguishing the flame. The linear actuator was then stopped and the chamber
depressurized.
One goal of this study was to expand the pressure range of nitromethane burning
burners generally have pressure limits of up to 10,000 psi. However, pressures generated
72
inside gun chambers or future propulsion devices could be much higher, possibly
can also be useful in understanding detonation initiation and reaction processes for
The UHPSB is a unique facility that allows the regression rates of strand
propellants to be measured at pressures up to 207 MPa (30,000 psi). Two solid or liquid
strands up to 65-mm long could be tested simultaneously to minimize gas usage and
maximize data return. Liquid or gelled propellant strands could be contained in either a
quartz tube (for simplicity) or combustible straws to minimize tube effects. To pass
ignition power and diagnostic signals through the high-pressure chamber wall, a custom-
made 8-wire electrical feedthrough was utilized. A total of three thermocouples (TCs), or
up to four break wires per sample can be sampled using this setup. Also, the number of
break wires could be increased with simple modifications to the sample base. The
control panel allowed operation and observation of a number of important functions from
trigger output from ignition switch for data acquisition system activation
Either glass tubes or combustible straws can be used to contain the liquid
73
75 mm (3 inches) in length and 6.4 mm (0.25 inches) inner diameter. Up to 4 holes were
drilled or punched at known intervals, and segments of Bussman 0.25-ampere fuse wire
were utilized as break wires threaded through the propellant strands. Break wire holes in
combustible straws or quartz tubes were sealed with a dot of epoxy to contain the liquid
propellant. Burning rate was deduced by measuring the time delay between burn-through
of adjacent wires.
electrical current through the wire. An ignition booster charge consisting of a small piece
(~0.05 to 0.10 gram) of an easily ignitable solid propellant (JA-2) was attached to the
nichrome wire to ensure uniform ignition. Electrical energy was supplied by either a
Variac variable AC transformer or DC power source at the control panel. In addition, the
system can be easily reconfigured to use any other type of initiator or electrically-
actuated igniter.
Since the test chamber was adapted from a high-pressure compressor storage tank
with a length of approximately 6 feet and a 4-inch inside diameter, the internal volume is
very large. Therefore, there was negligible pressure change due to product gases released
during the combustion event. The atmosphere for combustion was determined by the gas
input source to the compressor. Pressurized nitrogen was used for the environment in all
tests in this study, but other gases such as air or argon could easily be substituted.
Because of the chamber design and horizontal orientation, there was no provision for an
axial purge gas flow during a test. Although the lack of purge could lead to flame
74
spreading down the sides of solid propellant strands, for liquid propellants contained in
A schematic diagram of the UHPSB is shown in Figure 3.4. It can be seen that
for safety, the controls and data acquisition system are located in another building apart
from the one containing the UHPSB test chamber. Both the compressor and the exhaust
solenoid valve (SV1) could be remotely operated to increase or decrease the chamber
pressure to a desired level. To run an experiment, after pressurization, the igniter power
Compressor Building
(P1) (Test Cell)
High Pressure HV1
Solenoid
Valve Test Sample
Purge gas flows up through the chamber base and exits through a perforated plate
to uniformly distribute the purge flow field around the propellant strand. This gas flow
serves the purpose of cooling the propellant product gases and sweeping any smoke or
colored product gases away from the viewing ports and into the exhaust system.
75
temperature data with burning rate and observed position, a temperature profile as a
function of location can be determined. The subsurface thermal profile will help in
surface temperature, and supplying initial condition and validation data to the gas-phase
reaction and flame simulations. Flame location and structure can be determined by
examining the profile for rapid temperature rise and plateau features. These data can be
used as a guide during formulation of the model to determine what phenomena need to be
considered and which ones can safely be neglected while still capturing important
behavior. Using measured values as an initial guess for the gas-phase solution can reduce
relatively small thickness of the thermal features in the regions of interest for burning
profile, the thermal wave must be more than 10 times the thermocouple thickness to
ensure sufficient spatial and temporal response. Currently, the practical limit on
Basically, the thermal wave is the depth of heat penetration conducted into the
condensed phase from the regressing burning surface. At higher burning rates, the
preheated zone is thinner than those at lower burning rates. For a homogeneous
condensed phase, the thermal wave thickness (th) for 99% of the total temperature rise
th = 4.6 (3.1)
rb
where is the thermal diffusivity of the liquid propellant and rb is the regression rate.
For nitromethane at ambient temperature, the burning rate at a low pressure condition (P
= 3 MPa) is about 0.93 mm/s. Calculated from values found in Ref. 16, the value of
liquid thermal diffusivity, , is approximately 1.0x10-3 cm2/s. Using Eq. (3.1), the
thermal wave thickness is about 500 m. This value decreases rapidly with pressure
(inversely proportional to burning rate), to less than 250 m at 6 MPa. Flame zone
expected to be a strong function of reaction scheme and rates, and therefore can vary
flame, while others include multiple stages and even dark induction zones with nearly-
and D-type thermocouples can reach a maximum temperature near 3100 C, but they are
not recommended for below 400 C. Both of these types are normally made from
77
the junction section and surrounding wire to the desired thickness. The procedure for
Combustion Lab. The junction can be formed in one of two manners, either using a spark
welding process or by flowing current through two wires in contact with each other at
their tips. By regulating the voltage and current, the temperature can be controlled to
form a welded junction at the point of contact. Both of these processes result in a
junction bead of approximately the same size as the wire diameter. Using the typical
values shown above for the thermal wave thickness and expected flame dimensions, for
expected that only low burning rate (i.e. low pressure) conditions can be successfully
investigated.
supported by a thin dielectric post was installed through a hole drilled in the side of a
tube. The post limited the motion of the thermocouple junction during the combustion
event, ensuring that the junction was at a known location, generally near the centerline of
the tube.
same type in an ice-water bath and connecting it in series to the thermocouple mounted in
the sample. The zero-referenced thermoelectric voltage was amplified 100 times near the
point of measurement to reduce the effect of electrical noise introduced on the cabling
inverse polynomial. For the S-type thermocouples, the ninth-order ITS-90 polynomial
given in the Omega Temperature Handbook159 was used. For the D-type, no standard
was given, so a seventh-order polynomial curve fit was obtained from the direct
polynomial given in the Omega Handbook. For temperature in degrees C and voltage in
7
T = cn E n (3.2)
n =0
For greater accuracy, the curve was broken into two temperature ranges, resulting in the
Table 3.1 D-type thermocouple inverse polynomial for calculating temperature from
voltage.
T < 783 C T > 783 C
(EMF < 13.822 mV) (EMF > 13.822 mV)
Coefficient Value Coefficient Value
c0 0.35788 c0 -2561
c1 100.67 c1 876.97
c2 -14.888 c2 -107.9
c3 3.0624 c3 7.6642
c4 -0.41425 c4 -0.32134
c5 0.033629 c5 0.0079706
c6 -0.0014708 c6 -0.00010833
c7 2.6553e-05 c7 6.2345e-07
using the apparatus in this study were previously reported in detail by Lu, et al.160,161
Figure 3.5 shows a schematic diagram of the optical arrangement. A xenon arc lamp was
79
used as a high-intensity source for a light beam passing through the flame zone; a portion
of the light energy was absorbed by OH molecules present within the flame zone. The
beam was focused on a pinhole to set the spatial resolution of the measurement, then
directed into the spectrometer. Least-squares fitting of a model spectra to the measured
Figure 3.5 Schematic diagram of the optical arrangement for UV/Visible absorption
spectroscopy measurements.
The use of liquid propellant presented unique problems. All previous uses of this
device were with solid propellant sticks. However, unlike a solid-propellant strand,
liquid propellant needed to be confined in some sort of container in order to retain its
shape. In addition, the container had to interact as little as possible with the white light
80
beam passing through the combustion zone of interest. For this series of tests, an 8-mm
ID square quartz tube approximately 75 mm long was used. The parallel sides of the
square tube minimized distortion in the light beam passing through the flame zone that
could be caused by a cylindrical sample tube. The tube also confined the flame and
above the burning liquid surface. Quartz was selected because of its high transmissivity
measurements, ignition was achieved through the use of a 100-W CO2 laser incident on
the top surface of the propellant. At low pressures in the N2 atmosphere, ignition was
very difficult both due to the N2 gas environment and in-depth absorption of the infrared
laser energy. A small piece of solid propellant suspended just above the liquid surface
CHAPTER 4
engineering data for system design, as well as serve to guide development of a model and
provide data for model validation. Surface region observations and temperature profile
Regression rate data allowed the model results to be matched against real-world
measurements.
Tests over a wide range of pressures indicated that nitromethane is very difficult
feeding tests, there was no successful ignition for nitromethane in nitrogen at pressures
up to the highest level tested at 19.4 MPa (2800 psig) using an electrically-heated
nichrome wire. In air, ignition even in air was difficult when the chamber pressure was
below 4.24 MPa (600 psig). However, tube tests in a nitrogen atmosphere (in the
UHPSB or UV/Visible chamber) ignited reliably when using a small piece of solid
propellant as a booster.
Close-up observations of the liquid propellant burning surface were very useful in
phenomena (such as bubbling, surface instabilities, etc.) and visible flame structure, high-
pressure, windowed strand burners were used for close visualization of the surface
were captured using either an S-VHS VCR coupled with a PC-based frame grabber or a
high-speed digital camera system. The Photron FastCam PCI high-speed camera system
used was capable of capturing up to 2000 24-bit color frames per second, although the
achievable rate was lower due to relatively low luminosity of the nitromethane flame and
limited dynamic light sensitivity of the CCD imager. In both cases, a macro zoom lens
observations were made: feeding liquid in a stationary open tube in the LPSB, and
regressing liquid in a static tube in a windowed strand burner with CO2 laser ignition.
At low pressures (e.g. 3.1 MPa) in the LPSB feeding mode, nitromethane
exhibited distinct bubbling near the liquid surface, as seen in Figure 4.1. The bubble
sizes were relatively large, up to about 800 m. Bubble size and frequency decreased
with elevated pressure. At 13.3 MPa, the burning surface was quite uniform with no
visible bubbles.
In the 3.1 MPa image of Figure 4.1, a large, unstable, contracting flame plume
can be seen. Just above the surface, a thin luminous reaction zone is distinct. This thin,
uniform flame is assumed to be the monopropellant flame, while the large plume is due to
83
the H2 and CO products could react to water and carbon dioxide if additional oxidizer
was supplied. As the pressure increased, the plume elongated and became more
luminous and longer. Eddies can be seen at the flame boundaries from interaction of the
For tests with a static strand, the regressing surface was observed to be very
smooth and uniform, with only a few occasional surface disturbances visible. Typical
images of the regressing surface at 3 and 6 MPa in an 8x8 mm square tube are shown in
Figure 4.2. The liquid meniscus against the tube wall gives the appearance of slight
curvature near the edges. The surface is illuminated from below; the glow near the
surface is believed to be a combination of self-illumination from the faint flame and light
84
refracted or reflected by the interface from the external illumination. At the conditions
3 MPa 6 MPa
Figure 4.2. Images of regressing surface into quiescent propellant in square tubes.
Although the bubbles seen in feeding tests were believed at first to be uniformly
distributed across the surface, continued observations and comparison to tube tests at
similar conditions with no evidence of bubbles led to the conclusion that the visible
bubbles were forming at the walls only. The presence of bubbles near the burning
surface is believed to have been caused by the combined effects of long liquid propellant
residence time associated with low burning rate and tube walls heated from the adjacent
stationary flame. At the exit plane, the tube walls were exposed to a steady high-
higher pressures was associated with shorter residence time and increased boiling point.
Higher burning rate (and hence convective cooling of the walls) also served to reduce the
effect.
85
Other than the bubbling near the heated exit tip wall for feeding tests at lower
pressures, the images displayed in the previous section for both feeding and static tests
showed a uniform surface with no indication of instability over the entire pressure range
observed. This was in stark contrast to XM46 HAN-based liquid propellant, which when
tested under similar conditions on the same equipment showed extreme surface instability
and waves.156 The flat nitromethane monopropellant flame was of low luminosity, and
very thin (< 1 mm) with no apparent induction zone or distinct stage structure at any
reaction zone was judged to be reasonable for model development, and the thin flame
structure was used to guide the initial guess for the gas-phase solution.
In the free-surface LPSB feeding tests, the flame plume showed both contraction
and eddies at the boundary. These details were examined to understand their existence
and magnitude. It was believed the flame contraction was caused by buoyancy affecting
the very hot flame plume emerging into the relatively cool purge environment. To
investigate this behavior, an axisymmetric simulation of the plume was done using
variable properties with temperature were considered. The gases were assumed to be
"incompressible ideal gases," meaning that the density was only a function of
temperature. For one simulation, this assumption was relaxed to ideal gases and no
changes were noticed in the solution except for much slower convergence. The
conservation equations were solved by a Newton iteration scheme where the equations at
method. The initial grid size used was 3000 cells; further refinement showed no change
in the solution.
velocity at the completion of combustion at 5 MPa calculated using the gas-phase model
in described in Chapter 5. From the observations and measurements, most of the plume
effects occurred above the thin reaction zone. In addition, the effects were believed to
result from the composition change and energy release generating the hot plume.
Important initial and boundary conditions are shown in Table 2.1. All thermodynamic
Table 4.1 Input and intial properties and conditions for FLUENT simulation of flame
plume in cold purge flow.
Gas Property Value Units
Pressure 49.34614 atm
Density, 0.004970601 g/cm3
Axial velocity 38.06179 cm/s
Temperature 2459.512 K
Mass fraction H2 0.0220551
Mass fraction H2O 0.2447451
Mass fraction CO 0.381092
Mass fraction CO2 0.1221842
Mass fraction N2 0.2293859
Since the actual purge flow rate was not measured for the images above, a range
of typical values for the equipment used was estimated. For the simulation, three
different cases were run with the concentric flow of N2 purge gas set at 0.005, 0.05, and
0.5 m/s, which were labeled as slow, medium, and fast respectively.
Flow streamlines, and temperature and density contours for the three purge
conditions are shown in Figs. 4.3 to 4.5. Looking at these figures, it can be seen that the
87
hot, low density flame plume was rapidly accelerated against the gravity vector by
buoyancy in all cases. Because of continuity, the flow tube must contract, resulting in the
characteristic plume shape seen in the previous images. This is illustrated by the
streamlines bending toward the centerline of the flow just above the zero position. Not
surprisingly, the higher mass and momentum flow at the fast condition was affected least
by the flame plume, while the slow purge showed very significant entrainment.
Figure 4.3 Density and temperature contours for nitromethane combustion plume into a
slow purge flow (0.5 cm/s).
88
Figure 4.4 Density and temperature contours for nitromethane combustion plume into a
medium purge flow (5 cm/s).
89
Figure 4.5 Density and temperature contours for nitromethane combustion plume into a
fast purge flow (50 cm/s).
Eddies seen at the flame boundaries in Figure 4.1 are manifestations of a different
phenomenon. The flame above the free surface can also be viewed as a thermal plume or
jet into a cooler, denser environment. Periodic instabilities in buoyant jets have been
plume and environment are in general still not well-understood, but observations have
columnar flame, and a varicose mode with larger-scale low frequency vortex formation
90
near interface.169 Most work for flames has been done on either unreactive heated
plumes (e.g. He) or diffusion flames. However, essentially isothermal plumes from
observed in the nitromethane plume as seen in Figure 4.1, although some sinuosity is
seen at the lower pressure (3.1 MPa) condition, also. Because the situation is
mathematically very challenging due to complex interactions of different flow fields and
chemical reactions, a variety of parameters to characterize the flow and methods to define
stability have been developed. These include the jet-to-environment density ratio,
V2
Fr = (4.1)
gD
gD 1 j
Ri = (4.2)
V2 j
With a Reynolds number of about 135 and a Richardson number of approximately 3.4,
the nitromethane plume at 5 MPa resembles very closely a helium jet at similar
conditions as imaged by Subbarao and Cantwell.165 The density ratio (~0.09) and
Reynolds number (~135) for the conditions discussed here also place it within the
A series of tests were conducted in both the LPSB and the UHPSB at pressures up
to 170 MPa in order to better characterize the burning rate of nitromethane. HPLC-grade
nitromethane obtained from Sigma-Aldrich was used for all tests conducted in this
investigation (including LPSB and tube data). Two methods of measuring the
nitromethane burning rate were used: the burning rate at a stationary surface determined
tube tests and feeding tests in the LPSB, combustible straw tests in the UHPSB, and
quartz tube data from NAVORD37 is shown in Figure 4.6. The data appeared to break
into three regimes based on the pressure exponent. For feeding tests at low pressures in
the LPSB, a pressure exponent of 1.03 was found. Quartz tube test data at the same
conditions showed a slightly lower burning rate compared to the feeding tests, but a
slightly higher pressure exponent of 1.17. These two sets of burning rates converged as
the pressure increased. A slope break to much higher exponent (n=2.33) appeared around
15 MPa. The abrupt change in the burning rate versus pressure slope was also noted by
Rice and Cole37 of NAVORD in this range. A second slope break occurred near 70 MPa;
where the pressure exponent flattened significantly to n=0.86. In summary, the burning
rate of nitromethane could be described by three separate power-law equations for their
The close agreement between the measured burning rates from the present quartz tube
tests and NAVORD quartz tube test results can be seen from this figure. This is expected
1.07
Quartz Tube (NAV ORD): r = 0.226P
b
10
0.1
1 10 100
Pressure (MPa)
For both LPSB feeding tests and low-pressure quartz tube tests, an air
environment was used to achieve ignition. Although tube tests were conducted in air, as
soon as the flame passed below the tube mouth, the tube walls blocked radial diffusion of
the surrounding gas to the flame zone, hence the propellant burned in a monopropellant
93
mode in the tube tests. The nitromethane burning rate from LPSB could be enhanced
slightly by reducing heat loss to the tube wall and the oxygen in the purge gas (air).
However, as pressure was increased, the difference between the two burning rates
became less, with the two curves nearly converging. Although the quartz tube data could
be slightly affected by the heat loss from the flame zone to the tube walls, from the
captured images the reaction takes place in a very concentrated zone. Based on the
observed wall bubbling in the feeding tests, it is believed the tube test results are closer to
Rice and Cole37 also reported high-pressure strand burning rate results. However,
due to limitations of their equipment including the use of electromechanical timers and
large pressure excursions due to small chamber size (pressure increase of up to 5000 psi),
the data are much less reliable than lower-pressure values. Their reported values are
reproduced in Figure 4.7 below and compared to the correlations derived in the current
study. It can be seen that for the middle pressure range, their measured burning rates are
higher than those of the current study, but nearly identical after the slope break (other
than a single outlier). The distinct change in slope is also present in their data.
94
1000
NAVORD high pressure data
Curve fit of present study
Burning Rate (mm/s)
100
10
1
10 100 1000
Pressure (MPa)
Figure 4.7 High-pressure nitromethane burning rate data of Rice and Cole37 compared to
current study.
found in a recent paper by Kelzenberg, et al.42 Although the cited Ph.D. thesis173 and
data from the Mendeleev University of Chemical Technology174 were not acquired,
values could be estimated graphically from the plot given in the paper. Data of
Kelzenberg, et al. was over the pressure range of 0.5 to 4 MPa, and Raikova, et al.
covered 6.5 to 30 MPa. These data along with the lines showing the correlations derived
in this study are shown in Figure 4.8. It can be seen the data of Kelzenberg, et al. shows
a lower pressure limit. To achieve ignition and sustained combustion, they used an air
environment, and at such a low burning rate (less than 0.2 mm/s), the combustion may
Kelzenberg, et al. and Raikova, et al. match the low pressure correlation from this study
very well. The slope break near 15 MPa also appears, but above that pressure, the limited
data of Raikova, et al. appears to show a somewhat higher pressure exponent. Since no
details were given on experimental equipment or technique used to obtain this data, it is
1000
Kelzenberg, et al.
Raikova, et al.
100 Correlation of Current Study
Burning Rate (mm/s)
10
0.1
0.1 1 10 100
Pressure (MPa)
Figure 4.8 Data of Kelzenberg, et al. and Raikova, et al. compared to correlations
obtained from current study.
Intrinsic burning rates of nitromethane were found using the LPSB over a range
of 2.5 to 15 MPa (350 to 2200 psig) and initial temperatures of -30, 15, 25, and 45 C. A
96
small variation in burning rate was caused by change in initial temperature. The
following power-law equations provide a close fit to the measured intrinsic burning rates:
ln rb
p (4.10)
Ti P
Using this relationship, p indicates the fractional variation in burning rate with respect to
the initial temperature change at a given pressure. Figure 4.9 shows the logarithm of the
burning rate plotted versus propellant initial temperature for a number of chamber
pressure conditions. The p for each pressure can then be deduced from the slope of a
linear curve fit. Values of p obtained are shown in Figure 4.10 as a function of pressure.
It can be seen that nitromethane burning rate is most sensitive to initial temperature
change at low pressures, with p = 2.9 x 10-3 K-1 at 2.51 MPa (350 psig). The p
decreases significantly with increasing pressure, dropping to a very low value of 0.6 x 10-
3 K-1 at a pressure of 9.96 MPa (1430 psig). Normal homogeneous solid propellant has
p on the order of 2 x 10-3 K-1 at very high pressures such as 180 MPa.175 It was found
97
that a second-order polynomial given by Eq. (4.11) can fit the deduced p data at
different pressures.
where P is in MPa and p in K-1. The decreasing trend of p with increasing pressure
indicates even lower temperature sensitivity at higher pressures. This behavior is highly
desirable for nitromethane to burn according to the steady state burning rate with very
combustor. The physical interpretation of the extremely low p is that since the thermal
wave thickness of nitromethane is very thin at high pressures, the relaxation time needed
to respond to a pressure excursion is very short. The energy stored in the thermal wave is
1.5
1.0
ln(r b )
0.5
0.0
-0.5
240 250 260 270 280 290 300 310 320
Ti (K)
2.51 MPa (350 psig): ln(r b) = -1.091 + 0.0029T i
3.41 MPa (480 psig): ln(r b) = -0.622 + 0.0024T i
5.06 MPa (720 psig): ln(r b) = -0.013 + 0.0017T i
7.68 MPa (1100 psig): ln(r b) = 0.629 + 0.0010T i
9.96 MPa (1430 psig): ln(r b) = 1.028 + 0.0006T i
Figure 4.9 Nitromethane burning rate sensitivity as a function of initial temperature
98
3.0 10 -3
-3 -4 -5 2
= 4.38x10 - 6.73x10 P + 2.97x10 P
p
2.5 10 -3
2.0 10 -3
p (K-1 )
1.5 10 -3
1.0 10 -3
5.0 10 -4
2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
Pressure (MPa)
Figure 4.10 p sensitivity as a function of chamber pressure
sample tubes. A set of typical traces at 4 and 6 MPa are plotted in Figure 4.11. Although
some variation is expected between tests due to uncertainty of the thermocouple junction
position and possible slight motion from surface tension effects, the trends seen in the
figure are representative of the pressure effect. The time axis has been converted to
distance through use of the regression rate. The trace can be seen to consist of four
regions.
99
4 MPA
6 MPa
Temperature (K) 2000
1500
1000
500
The first is the undisturbed liquid, seen as the constant-temperature section at the
left side of the plot. Most of this region was truncated in the plot shown in Figure 4.11.
Although it was expected that the temperature should remain constant at the initial value,
in all tests performed, a low-rate, small-magnitude temperature rise (on the order of 15-
20 C) was noted starting after ignition. Since the incident laser was shut off
immediately after ignition (less than 1 second of irradiation) and this portion of the trace
was far away from the main thermal wave, the indicated temperature rise may be due to
radiation from the approaching flame front penetrating the transparent nitromethane and
heating the exposed thermocouple junction. The liquid propellant temperature is believed
The second region is the arrival of the thermal wave in the liquid subsurface
100
region, and is just under 1-mm thick for the 4 MPa case, and about 600-m thick for the
6 MPa case. As expected, the thermal wave for the lower pressure, lower burning rate
A change in slope and rapid temperature rise indicates the liquid/gas interface and
reactions associated with the flame in the third zone. The flame zone is very thin, with
the distance as shown less than 50 micrometers (if the plot is significantly stretched in
scale). As this is too thin to be accurately resolved by the thermocouple used, the true
The fourth and final region is the product gases with a plateau near the flame
temperature. For the 4 MPa case, after the initial jump just above the surface, the more
gradual rise to this final temperature occurs over approximately 1.5 mm. At higher
pressure (P = 6 MPa), the reactions leading to the plateau temperature occur faster, with
the peak temperature found at 0.75 mm from the surface. The temperature decrease after
the peak is most likely due to heat loss to the surrounding walls, since they are exposed to
a cold purge flow on their outside surface. This heat loss to the surroundings is
illustrated by the thermal boundary layer seen in the purge flow on the outside of the tube
above the flame location shown in the images of Fig. 2. Measured peak flame
temperatures varied from about 2080 K at 3 MPa to 2120 K at 6 MPa; these were
significantly lower than the equilibrium flame temperature of 2460 K as calculated by the
NASA CEA code.176 The traces as shown are truncated after 2.5 mm due to burnout of
the microthermocouple and associated loss of signal soon after that point.
can be distinguished in the curve where the temperature gradient jumps to a very steep
slope. The values of this takeoff temperature (which may not be the true surface
temperature) are near 230 C, with higher temperature at higher pressure. By plotting the
apparent activation energy can be found. Although not a true surface temperature, this
takeoff temperature should behave similarly. In Figure 4.12, two sets of data are plotted.
One set corresponds to the measured takeoff temperatures at pressure levels of 4 and 6
MPa. The other set is the calculated surface temperatures based on phase equilibrium.
Although the pre-exponential factor varies, the slopes (i.e. activation energies) are very
similar. This suggests that the physical processes occurring are similar; therefore surface
0.5
Measured "Takeoff"
Temperatures
0.2
0.1 Calculated T
surface
0.09
0.08
0.07
0.06
0.05
0.001 0.002 0.003
-1
1/T (K )
s
Previously, UV-visible absorption spectroscopy has been used for flame zone
diagnostics for solid propellants.177 Vanderhoff, et al.178,179 measured both species (OH
and NO) and temperature profiles in a variety of solid propellant flames. More recently,
Lu, et al.160,161 successfully measured species (OH and NO) and temperature profiles for
Transmission spectra covering the wavelength range from 307 to 311 nm were
measured just above the nitromethane burning surface, through the concentrated reaction
zone. This was achieved by mounting the sample cell so that the light beam passed
103
through the unburned liquid. Then, the spectrometer was set to continuously gate and
sample spectra. The sample was ignited and began to regress, with the surface
eventually passing through the probing location. Since the attenuation was much
different through liquid than through the gas phase, it was obvious when the burning
surface passed the sampling volume. With the spatial resolution near the dimension of
the reaction zone (approximately 200 m), it was assumed that the first good spectrum
was just above the propellant surface and thus reflected the average composition in the
reaction zone.
nitromethane flames over the pressure range from 3.5 to 10.3 MPa (500 to 1500 psig).
An image of combustion in a square quartz sample cell and absorption spectrum at 5.6
MPa (800 psig) are shown in Figure 4.13 and Figure 4.14, respectively. In Figure 4.13,
the concentrated reaction just above the surface (about 25% from the bottom of the
image) can be seen, with the luminous flame situated on top of the surface. The two very
large bright regions are caused by scattering of the UV/Visible light beam from the
sample cell walls. Small carbonaceous deposits can be seen on the quartz walls. The
spectrum in Figure 4.14 was actually taken at an earlier time than is indicated in Figure
4.13; as shown, the surface has regressed significantly below the sampling region.
When the captured spectra were analyzed, levels of OH were too low (< 1%) to be
accurately determined using the data reduction program, but the strong absorption peak
White light
beam from
Oriel arc lamp
Square quartz
Flame location tube containing
Unburned nitromethane
Nitromethane liquid propellant
Figure 4.13 Image of nitromethane flame immediately above the surface of the sample
contained in a quartz tube (8 x 8 mm2) obtained from the UV/Visible test rig at 5.6 MPa
(800 psig).
1.00
Transm ittance
Transmittance
0.95
0.90
Sharp reduction in
transmissivity
represents absorption
due to OH presence
0.85
307 308 309 310 311
Wavelength (nm)
Figure 4.14 Absorption spectra of nitromethane flame at 5.6 MPa (800 psig).
105
Because of the small length scale of the reaction zone compared to the resolution of the
instrument, even at low pressures, it was impossible to resolve the profile or temperature
of active species (e.g. OH, NO, CN). The instrument resolution of approximately 150
m was larger than the gas-phase reaction zone in which the intermediate radicals formed
and were consumed. Therefore, the amount of information that could be obtained from
this technique was very limited. The only conclusions that could be drawn were that OH
was present in the reaction zone, and that the zone thickness was similar to or smaller
than the resolution of the instrument. The thin gas-phase reaction zone with low
discussed in Chapter 6.
106
CHAPTER 5
process in steady regression had been made. One of the main tools that can be applied to
address the problem is a detailed model of the combustion process, incorporating both
temperature and gasification at the surface. Although gas-phase reactions have been
addressed by previous researchers, no detailed model has been used for a steady
deflagration situation. By coupling the two regions at the burning surface and using
selected empirical parameters as inputs, a predictive model can be developed. For this
study, only behavior at ambient pressures below the critical point was considered.
Two types of experimental measurements were made for burning rate: regression
into quiescent liquid in a tube, and feeding tests where the liquid was fed to a free surface
at its regression rate to ensure a stationary surface. In principle, both of these situations
can be treated as a stationary surface fed by a steady flow. In its most general form, the
stable combustion of liquid propellant can be broken into a series of adjacent regions
107
Hot Plume
Region
Gas Phase
Tflame
Combustion Products
Two-Phase Region
Preheated Region
Uniform
Temperature
Region
Unreacted Liquid
Ti
T Feeding Velocity
Figure 5.1 Physical model and axial temperature distribution for steady-state combustion
of liquid nitromethane fed into the burning zone under a controlled constant-pressure
condition.
nitromethane. Next is a complex preheated region where heat transfer from the surface
and possibly the walls affects the fluid. This is largely driven by energy transfer from the
flame zone to the liquid surface and subsequent conduction in the steep thermal wave
zone. Near the surface of the preheated region, bubble formation is possible due to
108
homogeneous nucleation phenomena at the boiling point. As the temperature rises, the
fluid may bubble giving rise to a two-phase region. At the gas/liquid interface, if bubbles
exist, they burst to release their internal chemical species and bulk gasification occurs
through either pyrolysis or evaporation. Finally, in the gas phase region, many
A block diagram of the general structure of the model and its relationship to input
data and anticipated output information is shown below in Figure 5.2. Experimental
observations and measurements can contribute to the model development and its
validation. The test results can also suggest simplifying assumptions, and provide initial
values for certain unknowns such as axial temperature profile. Some of the properties
and the reaction mechanism can be taken from the literature. By combining all these
parameters and solving the governing equations for the zone of interest, burning rate,
species and temperature profiles, and other useful values can be determined.
The challenge of an effective and efficient model is to simplify the true situation
order to achieve this, the theoretical approach to this investigation is divided into two
major parts: a condensed-phase model including all liquid processes and surface
phenomena, and a gas phase model incorporating chemical reactions and heat release.
thin region in the gas phase just above the surface. The high-temperature flame feeds
109
Assumed Values
for Unknowns Initial
Conditions Boundary
Empirical Data Conditions
and Correlations
Thermal
Simplifying Properties
Assumptions Governing Equations
Conservation
Species Considered and Equations Transport
Reaction Mechanism Equation of State Properties
Output
Flame Structure
- Species Profiles
- Temperature Profile
Burning Rate
Bubble
Characteristics, Etc.
in a tube. Because fluid below the thermal wave and combustion zone is quiescent, the
applied to both condensed and gas-phase models. Conduction through and heat loss to
the quartz tube walls were neglected here; the effect could be significant for the feeding
tests where the burning surface and flame are stationary just above a particular segment
of tube.
110
known. In some cases, these can be approximated or calculated, but measured values
will be used wherever available. Many of these properties are mentioned in Chapter 2;
To adequately represent subsurface region, a way must be found to simplify the complex
phenomena that are occurring. A number of different techniques have been proposed for
a similar situation found near the surface of burning nitramine propellants. Although
surface, nitramine combustion differs in that the bulk material is in the solid phase and
has a melting process. Nitromethane starts in a pure liquid phase and is converted to gas,
What is referred to as the condensed phase may in reality consist of some degree
decomposition processes. The most comprehensive technique is to track both gas and
condensed phases separately in a two-phase zone. However, this adds much complexity
to the formulation and solution. This two-phase region has been modeled in different
(RDX and HMX) combustion which was later extended by Li, Williams, and Margolis181.
111
This model treated the gas and condensed (liquid) phases separately, and related them
using a void fraction variable. A global reaction scheme was used, with three simple
account for the presence of bubbles and droplets. Two zones were found in the two-
phase region. At higher liquid volume fractions, evaporative equilibrium was present,
Solutions showed good agreement with measured burning rates and pressure and
temperature sensitivities.
Another method, used by Liau and Yang182 for RDX steady-state burning, was
to nitromethane due to the thick bubbling two-phase layer at the surface. Their technique
used spatial averaging over gas bubbles randomly distributed in the condensed phase.
Separate analyses and conservation equations for gas bubbles and condensed phase were
Ag = A (5.1)
where Ag is the fractional cross-sectional area of the gas bubbles and A is the total cross-
sectional area of the propellant. A reduced set of condensed phase decomposition and
Kuo and Lu185,186 used a statistical technique to extend the Lagrangian treatment
distribution of bubble sizes throughout the zone, average parameters (e.g. density and
and evaporation phenomena, this technique has not been computationally solved to date.
Before applying any of the two-phase flow models described above or similar
two-phase effects had any significant influence. Although tests with feeding showed
bubbling near the surface, static tube tests had no indications of any subsurface
bubbles seen in feeding tests were from wall effects. Simulations of gas bubble
formation and growth using the single-bubble model combined with observations allowed
model.
To examine bubble growth, a single bubble was modeled. The Lagrangian model
following the growth of a single bubble was originally developed by Kuo and Lu185 and
implemented by Kuo, Lu, and Tzeng186 for an RDX foam layer (shown in Figure 5.3).
Since there were many similarities between the RDX foam layer and nitromethane near-
surface region, the adaptation of this model could provide insight into the processes
taking place during nitromethane combustion. It was used to investigate bubble growth
nitromethane occurred in the subsurface region. A brief summary of this model and
113
Figure 5.3 Schematic diagram of physicochemical processes for the bubbles within the
RDX two-phase reaction zone (from Ref. 186)
For a spherical bubble, the mass balance at the bubble surface (boundary of control
D 3
g DB = DB l rl g
2
(5.2)
Dt 6
with rl g representing the liquid-gas conversion rate, expressed as the radial growth
velocity of the bubble. Species concentration tracked the balance of chemical species at
D 3
g DBYgi = DB l rl g i + DB i
3 2
(5.3)
Dt 6 6
where i is the volumetric production rate of the ith species from chemical reactions. It
114
was assumed that there were no dissolved gases in the liquid, so only the single
component liquid species passed across the boundary in the first right-hand side term.
Therefore, i was equal to 1 for the liquid component and 0 for all other species. Energy
D DV
g DB eg = DB l rl g hRDX + hc (Tl Tg ) p
3 2
(5.4)
Dt 6 Dt
In Kuo, Lu, and Tzengs original RDX analysis,186 three global reactions proposed by
Brill187 were used to account for reactions taking place inside the bubble. It was found
that there was very little reaction within the bubble; rather, mass flux from vaporization
dominated the bubble growth process. Bubble growth rate was much faster than RDX
solid surface regression rate. Hence, the bubble size was limited by foam layer thickness.
The biggest change necessary to adapt the analysis to nitromethane was to replace
locations through the thermal wave, instead of at a phase interface, and tracking their
development. The thermal wave dimension and profile was based on conduction heat
transfer calculations and measured temperature profiles. Bubble sizes introduced ranged
Liquid and gas properties and reactions were also replaced with ones appropriate for
nitromethane. Thermodynamic properties for the gases in bubbles (i.e., CH3NO2, CH3,
and NO2) were obtained from NASA coefficients in the September 2004 release of
Alexander Burcats database188 (this source will be discussed in greater detail in a later
115
third-body reaction) of nitromethane to CH3 and NO2 was taken from Zhang and Bauer.90
58,500
CH 3 NO 2
k1
CH 3 + NO 2 , k1 = 1.78x1016 exp (5.5)
RT
42, 000
CH 3 NO 2 + M
k2
CH 3 + NO 2 + M, k2 = 1.26x1017 exp (5.6)
RT
with time; in fact, bubbles shrunk quickly as the contents condensed. The lack of
vaporization was expected since conditions below surface were by definition below the
boiling point. The relatively high activation energy of the decomposition reactions and
relatively low temperature of the material also prevented any significant amount of
(either through evaporation or decomposition), the conclusion was that the subsurface
condensed-phase region could be treated uniformly as a single-phase liquid, and gas just
component liquid with an evaporative process at the surface. A control volume could be
chosen that that encompassed all processes of interest, starting below the thermal wave
and ending just above the surface. This approach could be applied to either feeding tests
or combustion into quiescent fluid in tubes by anchoring the control volume to the
surface location. Then, the burning surface could be treated as stationary, with a flow of
unreacted and unheated propellant entering the bottom of the control volume at a velocity
116
equivalent to the burning rate. Although developed independently, the model shown here
For the situation examined here, the area is constant. Since CHEMKIN uses the mass
flux (mass per time per area), continuity can be also expressed in terms of flow rate:
or in terms of burning rate (which is equivalent to the velocity at which fluid is flowing
FLRT
= rb (5.9)
init
As only one species is considered for the condensed phase with no reaction taking
place in this region, species mass conservation does not provide any additional
The energy balance across the control volume can be expressed as follows:
Q cond is the heat conducted into the control volume at the surface (dT/dx = 0 at the inlet
since the boundary was chosen to lie below the thermal wave), Q rad is the net radiative
flux at the surface, and Q reac is the energy released by chemical reactions within the
control volume. Q reac is assumed to be negligible based on the analysis in the previous
117
section, and will be discussed in detail later in this section. Q rad from the non-sooting
flame48 is also assumed to be negligible compared conduction to the surface. Latent heat
associated with phase change is captured in the enthalpy terms. With these assumptions,
Practically, this relation can be implemented and solved in two ways. The first is
to find the sensible enthalpy change in the condensed phase, then add the latent heat of
Q cond = m
( Ts
Tinit )
c p dT + Lv
(5.12)
An alternative technique suggested by Miller and Anderson189 is to find the total enthalpy
values directly for liquid at the initial temperature and at a point just above the surface
(where the gas temperature is still the same as the surface temperature) for the gas phase.
These values can then be used directly in Eq. (5.11). Either of these methods should
temperatures. The gas-phase total enthalpy in Eq. (5.11) was obtained from NASA
Enthalpy at initial conditions was found by adding the heat of formation to the (relatively
small) sensible enthalpy change from standard temperature to the initial temperature.
Tinit
hinit = h of + c p dT (5.13)
298
A temperature-dependent liquid specific heat correlation given later in Eq. (5.40) and
118
Table 5.2 was used for both techniques. Figure 5.4 shows the significant differences
between the two methods of finding the enthalpy change for typical conditions. The
baseline method displays virtually no change until a sudden drop-off with increasing
temperature, while the technique of Miller and Anderson shows the expected
correlation used to find liquid specific heat was only specified to be valid up to 473.15 K,
so the function beyond that point was based on an unreliable extrapolation. Because of
this, the method of Miller and Anderson was selected for use in this study.
6.0E+07
5.0E+07
Enthalpy Change (J/kmol)
4.0E+07
3.0E+07
2.0E+07
Baseline
1.0E+07
Miller-Anderson
0.0E+00
480 500 520 540 560 580 600
Surface Temperature (K)
Figure 5.4 Comparison of two different methods of calculating enthalpy change for
initial temperature of 273 K to gas phase at specified surface temperature.
or reaction of nitromethane below the surface, either in the condensed phase or in gas-
119
phase bubbles, if they existed. The only method of gasification is vaporization; therefore
the gas generated at the surface is pure nitromethane. This was justified in three ways.
First, previous studies had shown little to no reaction over long time scales for
nitromethane at temperatures near those postulated for the interface. For example, Shaw,
et al. found liquid nitromethane to be thermally stable at temperatures below 627 K for
pressures up to one kilobar (100 MPa).122 Similar results were also found in the high-
pressure diamond-anvil cell work of Piermarini, Block, and Miller.18 Second, the single-
bubble simulation of a bubble in the thermal wave (described in the previous section)
constant-pressure condition.191 To find the ignition time for a particular pressure, the
initial conditions used were pure nitromethane gas at the desired pressure and calculated
surface temperature (i.e. saturated vapor). This was considered a very conservative
approximation of conditions through the thermal wave in the subsurface region. Ignition
time was defined as when the temperature rose 400 K above the initial value. In Figure
5.5, a series of calculated ignition times at different pressures are plotted with an
estimated residence time in the thermal wave, found from the empirical burning rate. It
can be seen that over the range of pressures of interest, the ignition time was many orders
of magnitude greater than the residence time at elevated temperatures in the near surface
region, justifying the chemically-inert assumption for the subsurface region. These
5
10
t
ignition
4
10 t
residence
3
10
Time (sec)
2
10
1
10
0
10
-1
10
1000 2000 3000 4000 5000 6000
Pressure (kPa)
Initial temperature of the unreacted propellant was an input to the problem. For
conditions below the critical pressure, the surface temperature was found as the point
where the vapor pressure equaled the specified chamber pressure (boiling point). With
their explicit treatment of two-phase flow, Li, Williams, and Margolis181 found
evaporative equilibrium at the condition of high liquid volume fraction, which is met in
this case. In his studies of the combustion of melted energetic materials,192 Zarko also
used this assumption of phase equilibrium at the liquid surface, applying the Clausius-
121
the liquid and gas phases are not distinct. An Arrhenius-form pyrolysis law similar to
E
rb = As exp s (5.14)
RTs
can be used to extrapolate, with the parameters obtained by fitting the equation to
calculated surface temperatures over the range where a liquid-vapor transition exists.
This functional dependence has been shown to adequately capture a wide variety of
surface behavior. A unified gasification law of this form was applied by Zenin to
condensed-phase reaction for nitramines.194 For HMX, Ward, Son, and Brewster used it
to model a zero-order reaction for condensed phase to initial gas-phase species just above
the surface.195
(with increasing pressure and expected increasing surface temperature into the
supercritical region), both activation energy and pre-exponential factors could vary,
perhaps significantly. In fact, the definition of the surface becomes much less obvious.
However, this approach can still be useful, and has been implemented into the code using
values obtained from the vaporization curve and empirical burning rate data: A = 527.51
g/cm2-s with Ea = 9,071 cal/mol. For comparison, the value for the activation energy
122
processes in the region above the liquid/gas interface could be approximated by one-
above the interface is very flat and uniform. Its stand-off distance and thickness are very
small compared to the inner diameter of the feeding tube. In this one-dimensional
The conservation equations for the gas-phase region are listed below.
Conservation of Mass
d
( u ) = 0 (5.15)
dx
Conservation of Energy
dT d dT N dT N
c pu = i i pi
YV c i hi (5.17)
dx dx dx i =1 dx i =1
where the specific enthalpy for a particular species, hi, is a combination of the heat of
T
hi = hDfi + c
Tref
pi dT (5.18)
RuT
p= (5.19)
W
1
W= N
(5.20)
Yi
i =1 Wi
The diffusion velocity Vi can be found either with mixture-averaged properties or using a
techniques showed differences on the close order of one percent or less, while calculation
times using the multicomponent properties were up to ten times greater. Therefore, the
mixture-averaged properties were used. With mixture averaging, the diffusion velocity in
the energy and species equations can be written as the sum of three parts:
Vi = VD ,i + VT ,i + Vc (5.21)
The first term is ordinary mass diffusion caused by a concentration gradient (Ficks Law).
diffusion from a thermal gradient (Soret effect), also with a mixture-averaged diffusion
coefficient, and is only utilized for the low molecular weight species H and H2. The final
YV
i =1
i i =0 (5.22)
Species production rates in the above conservation equations come from reactions
N k fj N
ik M k ik M k for k = 1, 2, , M
i =1
krj
i =1
(5.23)
where M is the total number of elementary reactions. For the kth elementary reaction, the
Ea
kk = Ak T k exp k (5.24)
RuT
vj ,k
M
E N X p
i = Wi ( vi,k vi,k ) Ak T exp ak j
k
(5.25)
k =1 RuT j =1 RuT
The above equations have been implemented into the PREMIX driver code for
CHEMKIN.191
To predict the burning rates of nitromethane, the subsurface model was coupled to
the gas-phase model. This was accomplished by matching the mass, species, and energy
fluxes at the interface. Also, the temperature was required to be continuous at the
interface. With the formulations used, the interface between the condensed-phase and
gas-phase models was actually in the gas phase, just above the surface. For the energy
condition, the conduction heat flow into the condensed phase was obtained from the
conductivity and temperature gradient just above the surface. These requirements can be
expressed as follows, with the condensed-phase condition on the left and gas-phase
Ts ,0+ = Tg (5.26)
125
m s ,0+ = m g (5.27)
i ,0+ = mY
mY i,g (5.28)
dT
Q cond = g (5.29)
dx 0+
Applying these matching conditions resulted in the vaporization rate, surface
temperature, and species concentrations calculated from the condensed phase being used
In essence, independent solutions to each phase were found, which were then
coupled by applying and testing the interface boundary conditions, and modifying the
nested fashion, where the inner level was the gas-phase solver found in the PREMIX
driver code, and the outer level was the condensed-phase solver and coupling code.
implemented through the PREMIX driver code and CHEMKIN libraries and utilities.
126
Database files containing the species list and reaction mechanism, thermodynamic
properties, and transport properties are pre-processed into binary linking files for use by
the CHEMKIN library subroutines. The conservation equations are implemented in the
PREMIX driver code that calls the CHEMKIN subroutines as necessary to find properties
and reaction rates. When executed, the PREMIX program reads initial and boundary
conditions, as well as other problem parameters (e.g. tolerances and grid limits) from an
input file. The level of output during execution can be set by the user; at successful
convergence on a solution, the species and temperature profiles are written out.
In the PREMIX code, the governing gas-phase equations (Eq. 5.15 to 5.25) were
The convective terms used windward differencing, while derivatives in the energy and
species equations were approximated with central differences. At the cold boundary inlet
(x = 0), the initial temperature, species composition, and mass flux were specified. At the
hot boundary (outlet), the temperature and species gradients were required to be zero.
To find the solution, a steady-state damped Newton technique was applied until
user-specified relative and absolute tolerance criteria were met. To condition the steady-
state solution during iteration, a pseudo-timestepping technique was used where transient
terms were added to energy and species equations and a certain number of time steps
were taken. Steady-state techniques were then re-applied to the evolved solution. The
maximum number and size of the time steps were set by the user in the problem input
file, although the PREMIX code could adapt the usage (both size and number of steps)
within the user bounds as required by the specific situation. After a solution was found,
127
the first and second derivatives between grid points of both temperature and species
profiles (gradient and curvature) were checked against specified tolerances. If either was
higher than desired, new mesh points were introduced and a new solution was found on
the adapted finer mesh.196 Gas-phase inputs and properties are discussed in a later
section.
Figure 5.6 Flow chart for PREMIX gas-phase driver code and CHEMKIN system (from
Ref. 197).
128
both in absolute value (CPU time on a 3-GHz PC ranged up to hours per solution) and
relative to the rest of the problem. Since solutions were of a non-linear nature, many
projected points to speed convergence could not be applied directly. The use of empirical
The gas-phase and condensed-phase solutions were coupled using the flux and
an iterative process. A flow diagram of the FORTRAN code implementing the coupling
and iteration processes is shown in Figure 5.7. The code itself is listed in Appendix B.
An initial guess was assumed for the burning rate and an appropriate mass flux
(flowrate) was generated based on density found from the specified initial temperature.
This assumed flowrate value and the surface temperature (found either from the
assumption of phase equilibrium at the surface or a pyrolysis law) were used as input
Once a gas-phase solution was found, heat flux to the surface could be calculated
from the product of gas-phase temperature gradient and thermal conductivity just above
the surface, as expressed in Eq. (5.29). From this calculated heat flux, a new burning rate
value could be determined from the condensed phase equations. If this new burning rate
129
differed from the one previously assumed for the gas phase, a modified value based on
the difference was generated which was then fed back to the gas-phase solver as updated
input conditions and a new gas-phase solution found. This process was repeated in an
iterative manner until the burning rate calculated from heat transfer into the condensed
phase matched within a specified relative tolerance the assumed value used for the gas-
To find updated flowrate values for the gas-phase input, the bi-section method
was chosen for simplicity, but with a modified scaling factor parameter to tune it for this
particular problem. The true bisection method applies a factor of 0.5 to the difference
between the assumed and calculated solution values to find the correction for the next
guess. Too high a value could overshoot the true solution and lead to oscillation, while
too low a value would increase iterations, slows convergence, and possibly lead to greater
numerical error. To speed convergence for this problem, a factor of 0.6 was found to be a
optimal value. The solution tolerance selected was 5E-4, which gave a solution with
accuracy better than three significant figures, while not requiring an excessive number of
iterations. By using a well-chosen initial flowrate value, the program would generally
START
Declare variables
Initialize shell
environment for
CHEMKIN
Check if
Terminate with
iteration limit
error message
exceeded
Write updated
input file Run PREMIX
Check for
Terminate with
successful
error message
execution
Calculate rb based on
gas-phase heat feedback
Is rb,calc
rb,assumed within
tolerance?
Terminate with
success message
Figure 5.7 Flow chart of program implementing gas- and condensed-phase coupling.
131
are needed as inputs to the general model. These include the chemical kinetics
mechanism, thermal and transport properties for unreacted, intermediate, and final
species, initial guesses to aid the solution, and other properties used in calculations.
Many of these values have been found in the literature. Where necessary, estimated
values were determined either through analogy or calculation. In some cases (e.g. initial
values for burning rate and gas-phase temperature profile), empirical correlations and
measurements obtained in this study have been used. Details on the values and sources
mechanisms have been proposed for nitromethane. However, most were based on low-
pressure, low-concentration data, and developed to model specific shock tube or static
vessel conditions. Therefore, only the decomposition or at most, ignition process was
captured. For example, the mechanism proposed by Zhang and Bauer90 appears to be
involving N2, one of the major equilibrium products of combustion, making it obviously
Melius, and Yetter and Rabitz, were specifically developed for high-concentration
propellant ignition. However, they were used mostly as developmental models, and
Bendtsen, and Miller.94 Yetter. et al.s mechanism was used as a base to take advantage
of the refinement of gas-phase chemistry when the mechanism was validated against
RDX. Large modified nitramine species and fragments (e.g. RDXR, HCNNO2, etc.) and
associated reactions were removed from the reaction set, since nitromethane is a much
simpler molecule than RDX. Rates for the CH3 reactions were verified against the more
recent work of Warnatz, Maas, and Dibble199 and updated where appropriate. The
resulting overall mechanism was made up of 47 species and 250 reactions, and is listed in
Appendix C. Note that two of the elementary reactions (R111 & R114) have rates
modified from more recent literature and to match the measured burning rates of
nitromethane from the present study (this will be discussed in the following chapter).
molecule into CH3 and NO2, has been found to have a pressure-dependent rate constant
over the pressure range examined in this study. For pressure-varying reactions in the gas
phase, CHEMKIN provides multiple ways to express this dependency. Lindemann, Troe,
and SRI formulations are recognized to capture behavior in the fall-off region.200 All
To accurately represent reaction rates in the falloff region, Gilbert, Luther, and
Troe developed a set of equations with adjustable parameters that could be used to blend
reaction rates between low- and high-pressure limits.92 The low-pressure limit can be
133
written as
Ea
k0 = A0T 0 exp 0 (5.30)
RT
Ea
k = AT exp (5.31)
RT
Then, the actual reaction rate for a given pressure can be expressed as a function of the
P
k = k r F (5.32)
1 + Pr
The reduced pressure is a function of the mixture concentration and the rate limits:
k0 [ M ]
Pr = (5.33)
k
When F is 1, the expression reduces to the Lindemann form. For the Troe form, F is
given by
1
log Pr + c
2
T T T **
Fcent = (1 ) exp *** + exp * + exp (5.36)
T T T
Miller94 chose to use a centering parameter in the Troe form to accurately represent the
reaction rate in the fall-off region. Figure 5.8 shows measured reaction rates of Glnzer
and Troe69 (cited as [9] in the figure) and Zaslonko, et al.88 (cited as [26] in the figure) as
a function of concentration for three different temperatures. It can be seen that their use
of a center-broadening term Fcent with the Troe expression fits the data much better than
Figure 5.8 Fall-off curves for CH3NO2 dissociation (from Glarborg, Bendtsen, and
Miller94).
135
modeling, the concentration at a typical condition of 5 MPa can be located on Figure 5.8.
PV = nRT (5.37)
n P
= (5.38)
V RT
Given an intermediate temperature of 1000 K in the flame above the surface, the
concentration is 0.601 mol/L or 6.01x10-4 mol/cc. Although the conditions under which
nitromethane decomposition occurs vary, both through the flame and for different initial
conditions, it can be seen that the concentration is in the region where the center-
broadening curve deviates significantly from the either the asymptotic or Lindemann
CHEMKIN does not take a constant Fcent (as given by Glarborg, Bendtsen, and
Miller94) directly as input, however, but uses a function of temperature with either three
or four user-supplied parameters, , T*, T**, and T***, as shown in Eq. 5.36. T** is an
optional value, and its term will be omitted if the parameter is not supplied.196 In order to
use the constant Fcent as input to the gas-phase solution, suitable parameters had to be
found for at least , T*, and T***. These parameters were obtained by curve fitting Eq.
5.36 to the constant Fcent value over the temperature region of interest. By following the
technique used by Yetter, et al.198 of forcing T*** to be very small (i.e. 1E-90) and T* to
be very large (i.e. 1E90), the function showed essentially no variation with temperature
over a typical combustion range, and assumed the constant value of Fcent. The points
136
used and resulting fit to the three-parameter correlation used by CHEMKIN are shown in
Figure 5.9.
g
0.25
0.2
0.15 y = (1-M1)*exp(-M0/1E-90)+M1...
cent
Value Error
F
m1 1.83e-01 0e+00
0.1 Chisq 0e+00 NA
R 1e+00 NA
0.05
NM decomposition: CH NO (+M)<=>CH +NO (+M)
3 2 3 2
Glarborg, et al. F = 0.183
cent
0
0 500 1000 1500 2000 2500
Temperature (K)
Figure 5.9 Curve fit to center-broadening parameter (Fcent) of Glarborg, Bendtsen, and
Miller for Troe form of falloff reaction.
With the assumption that no reaction occurs in the condensed phase, the only
compound that needed to be considered was NM liquid. Density and specific heat as a
function of temperature were necessary to calculate mass flow and enthalpy change in the
subsurface region.
137
5.3.2.1 Density
measured experimentally by Berman and West14 between 130 and 200 C. They
(1+(1T C)D )
=A B (5.39)
where is in grams per cubic centimeter, and T in Kelvin. With this function and the
coefficients listed in Table 5.1, density can be found from the freezing to boiling points,
with less than 1% error expected over the entire range. It is assumed the compressibility
Heat capacity of the liquid was also determined by Berman and West14 over the
same range of conditions, and fit to a fourth-order polynomial. Again, properties over a
wider range of temperatures have been tabulated by the DIPPR project. A second-order
polynomial
Cp = A + BT + CT 2 (5.40)
was sufficient to give less than three percent error, providing Cp in J/kmol-K using the
Table 5.2 Coefficients for nitromethane liquid heat capacity correlation (from DIPPR6)
Temperature (K) A B C
244.60-473.15 1.1627E5 -1.3530E2 3.4500E-1
Change in enthalpy with temperature can be obtained from integrating the specific heat
polynomial over the temperature range of interest. If integrated from standard conditions
determined.
Other thermal and physical properties of the liquid were compiled, calculated, and
document are reproduced in Appendix D. Correlations for some of these properties are
Most of the species of interest in the kinetics mechanism are commonly studied in
combustion problems, and both thermal and transport parameters can be found in the
databases included with the CHEMKIN 4.0 distribution.191 However, there are a few that
are more rarely seen or specific to the nitromethane mechanism, and alternative methods
had to be found to develop values. Since CHEMKIN has a particular format for inputting
thermal and transport properties (not actually properties themselves, but parameters of
curve fits or equations that can be used to calculate them), a number of sources and
species transport and thermal data used in this study are listed in Appendix E.
Data for the thermal properties of most species included in this study were from
the comprehensive database compiled and maintained by Burcat,201 containing more than
1300 species. The April 2005 version202 was used for the calculations in this document.
Not all species in the reaction mechanism were present in the database, so other sources
were used for a few species. For those species that were not available in the the Burcat
database, thermodynamic property parameters in the CHEMKIN format were used from
It should be noted that Burcats database was undergoing a substantial update and
expansion at the time of writing, and at the request of the author will include in future
To evaluate the governing equations for the gas phase, a set of accurate transport
properties including viscosities, thermal conductivities, and mass and thermal diffusion
coefficients are needed. Instead of tabulating coefficients for each property, CHEMKIN
calculates these properties based on kinetic theory. Six molecular parameters are
database was used without modification or review for the species contained in it.
However, there were a number of species that did not have parameters in this database;
therefore, alternate sources and estimation methods were required to obtain the necessary
data. A study of laminar flame speeds in H2/air mixtures by Dong, et al.203 found that the
shown in a later section that the results of this study were relatively sensitive to transport
property values, so care was taken in finding values for the parameters of missing
species.
parameters are not as widespread. Where no existing data could be found, the parameters
based on rigorous kinetic theory was given by Hirschfelder, Curtiss, and Bird.204 From
the Lennard-Jones 6-12 intermolecular potential, Svehla later developed a series of rules
properties, or other techniques.205 These rules were further refined by Kee, Coltrin, and
141
expression obtained from kinetic theory, using the force constants and as adjustable
conductivity for obtaining values of and . If data of acceptable quality are not
available for measured transport properties, other ways to estimate force constants based
on physical properties, mixing rules of other species, or analogy with similar molecules
can be applied. The rules are listed in roughly declining order of quality and desirability.
If reliable data for neither of the above transport properties is available, the individual
force parameters can be estimated based on empirical relationships with other physical
properties.
If no reliable physical property data is available, the following mixing rules can be used
12. Estimate for monatomic species as that of diatomic species (or vice versa).
13. Find and for diatomic molecule AB based on values for species A2 and B2.
homologous compounds
of the technique. Parameters for other molecules missing from the CHEMKIN transport
database were found using similar methods following the rules given above.
function of temperature were readily available in the literature for nitromethane,6 Rule 1
26.693 MT
x106 = (5.41)
2 (2,2)*
where is viscosity in micropoise, M is molecular wt, T is absolute temperature in
Kelvin, the collision diameter in angstroms, and ij(2,2)* the reduced collision integral
be obtained from fitting), and the reduced collision integral can be approximated by the
following expression.206
( ) (
ij(2,2)* T * , ij* b1T *b2 + (T * + b3 ) f (2,2) T * , ij* )
b4
(5.42)
Here, the collision integral is shown as a function of reduced temperature,
143
Tk B
T* = (5.43)
where the temperature is non-dimensionalized by the ratio of the force constant in
Joules to the Boltzmann constant207 (kB = 1.3806505x10-23 J/K) and ij* is the reduced
dipole moment:
1 i j
ij* = (5.44)
2 ij ij3
This reduced dipole moment is a function of the dipole moment of the molecule(s) of
interest, (in Debyes), and the force constants and (to be obtained from fitting). In
( ) ( )
2
b5 T * T*
e e b6 *
( )
ij
f (2,2) T*
, *
= 1 + (5.45)
2 + 2.5ij*
ij
The coefficients bj in Eqs. 5.26 and 5.29 are given by Kee, Coltrin, and Glarborg206 and
are listed in Table 5.3 below. With the set of equations above, given viscosity as a
function of temperature, molecular weight, and dipole moment, the force constants and
Table 5.3 Coefficients for approximation of the reduced collision integral ij(2,2)* .
somewhere in the range of 3.1-3.57 Debye, with multiple measurements at 3.46.208 This
value was also the one recommended by DIPPR,6 and so was used here for the
calculation of transport properties. Viscosity data were also obtained from the same
144
source. In that database, the viscosity data are already correlated, although to a less
AT B
= (5.46)
1+ C
T
where A, B, and C are coefficients, T is the temperature in Kelvin, and the resulting
viscosity value is given in Pa-s. The coefficients for nitromethane used in this study are
Table 5.4 Coefficients for vapor viscosity correlation (from Ref. 6).
Coefficient Value
A 4.0700E-07
B 6.4850E-01
C 3.6750E+02
A spreadsheet was constructed with the input data and constants, Eqs. (5.41)
(5.45) were applied, and the Solver tool in Microsoft Excel209 was used to obtain the
best-fit values of = 4.271 and k B = 486.013 J. The quality of the fit can be seen
graphically in Figure 5.10, which shows the literature viscosity data with the curve fit of
Eq. (5.25) superimposed (marked as Svehla fit). It can be seen that the fit is very good;
the maximum error from the literature value was 1.89% at the lowest temperature. To
the more complex kinetic theory expression for viscosity with a simpler 3rd-order
N
ln k = an ,k ( ln T )
n 1
(5.47)
n =1
For the third-order fit, N = 4. This polynomial fit is shown as the third curve in Figure
145
5.10, and is also a very good fit to the literature values, with maximum variation of
1.31%. The four an coefficients of Eq. 5.47 for nitromethane are listed in Table 5.5.
2
3 10
2
2 10
2
2 10
2
1 10
1
5 10
300 400 500 600 700 800 900 1000
Temperature (K)
Figure 5.10 Comparison of literature values (DIPPR), curve fit, and CHEMKIN
polynomial approximation for nitromethane vapor viscosity.
146
Table 5.5 Coefficients for CHEMKIN nitromethane vapor viscosity 3rd-order polynomial
correlation.
Coefficient Value
a1 3.1350E-05
a2 1.1119E+00
a3 -9.1600E+01
a4 1.12800E+05
from the permanent dipole) due to an applied external electrical field. In essence, the
molecule deforms in the field, with the electrons shifting in one direction, while the
considers this parameter when calculating the binary interaction of a polar and a non-
polar molecule, with only the polarizability of non-polar molecules being used,196 as the
permanent dipole moment of a polar molecule would overwhelm any additional induced
dipole moment. As nitromethane has a dipole moment (discussed earlier in this section),
this parameter can be neglected for it. If it is necessary to find the polarizability of a
specifies the value at 298 K and a function for the temperature dependence is assumed.196
For nitromethane (and other species of interest), this parameter was not found in the
147
Zrot for most species are of O(1), with a few exceptions including Zrot,H2 = 280 ,
distributed about the center atom, possibly minimizing the rotational effects of collisions
with other molecules. Based on these observations and values given for structurally-
similar molecules in the database, it seems reasonable to assume a Zrot value of 2 for
nitromethane and other species with unknown values. To test this assumption, cases
were run with Zrot varying from 1 to 4 for nitromethane. The change in predicted burning
Since nitromethane is the major component just above the surface where energy
from the data in Table 5.8 were also compared to the literature value to ensure
consistency and accuracy. The function used by the DIPPR project for thermal
conductivity is
AT B
= (5.48)
C D
1+ + 2
T T
where A, B, C, and D are coefficients, T is the temperature in Kelvin, and the resulting
thermal conductivity value is given in W/m-K. Similar to the viscosity values, the
coefficients used in this study were obtained from the DIPPR database6 and are
Table 5.6 Coefficients for DIPPR nitromethane vapor thermal conductivity correlation
(from Ref. 6).
Coefficient Value
A 3.1350E-05
B 1.1119E+00
C -9.1600E+01
D 1.2800E+05
approximates the more complex expression with a simpler 3rd-order polynomial fit which
N
ln k = bn ,k ( ln T )
n 1
(5.49)
n =1
Coefficients of nitromethane for this polynomial fit are given in Table 5.7 and the results
graphically displayed in Figure 5.11. As with viscosity, it can be seen that the
correlations are very similar, with only slight variation between the two (maximum of
2.58%). Interestingly, the sign of difference varies along the curve, suggesting that much
of the variation is due to differences in the correlating function used between the two
Table 5.7 Coefficients for CHEMKIN nitromethane vapor thermal conductivity 3rd-order
polynomial correlation.
Coefficient Value
b1 3.1350E-05
b2 1.1119E+00
b3 -9.1600E+01
b4 1.12800E+05
149
-4
7 10
-4
6 10
-4
5 10
-4
4 10
-4
3 10
-4
2 10
Table 5.8 lists the supplemental transport properties for species present in this
study that did not have data in the tran.dat file distributed with CHEMKIN. Parameters
for a number of the minor species were taken from the RDX study of Liau.214 Although
it is obvious that some of them are estimates, since the mole fractions of those species are
relatively low at all points in the flame, the sensitivity to error in the properties should not
150
be significant. Properties for the last three species in the table were determined using the
Based on empirical data, an initial guess was made for the flowrate. Where the
vaporization curve was valid, surface temperature (i.e. gas-phase initial temperature) was
estimated flowrate and the pyrolysis law. The gas-phase solver was also initialized with
a temperature profile either from approximating the measured data, or derived from a
previous solution. The initial grid used was 100 points for the low pressures (6 MPa or
below) or 200 for the higher range; this grid was automatically refined as necessary.
Curvature and gradient limits were set to 0.3 and 0.15, respectively. As mentioned
previously, the initial gas composition was considered to be pure nitromethane, since no
At pressures higher than those considered here, non-ideal effects may need to be
considered. These are manifested in two ways: supercritical conditions for nitromethane
at the surface and real-gas effects in the gas phase. The critical pressure of
nitromethane is 6.31 MPa, within the range of normal rocket operation and a portion of
the experimental burning rate data. Above the critical pressure, there is no true phase
change (discontinuity in density) to form a distinct interface separating liquid and gas.
However, there is still a rapid change in the density due to intense heating from the flame
zone. Solubility of intermediate and product species at the surface can also be of
consequence. The supercritical droplet burning reviews by Kuo215 and Yang216 can
of state and method is used to define the surface, the energy balance given in Eq. 5.11
should still be valid, although the thermophysical properties exhibit great sensitivity to
pressure and temperature near the critical point, and can vary significantly from ideal gas
with correction terms that can be found from a real equation of state, such as the
gas-phase beyond the surface, as well. In an earlier nitromethane ignition study by Yetter
and Rabitz,117 introducing a real gas model for high pressure changed the characteristic
CHAPTER 6
phase models discussed in Chapter 5 was exercised over a range of conditions for
nitromethane combustion. Predicted burning rate values were obtained, and sensitivities
to various parameters, including kinetic rates and thermal and transport properties were
determined.
Predicted burning rates for nitromethane were found over the range of 3 to 6 MPa.
Comparison of two sets of calculated rates to measured tube rates is shown in Figure 6.1.
Agreement between the calculated rates and data is relatively good. As noted in the
figure legend, some of the reaction rates were modified slightly from the baseline values
given in the reaction list in Appendix C. The first set of calculations is based upon the
overall reaction mechanism with updated values of R111 and the basline value for R114,
and yields burning rates about 5% lower than the measured rates. The calculated values
showed a slightly higher pressure exponent, 1.27, compared to the empirical value of
1.17. In the second set, the rate for R114 was increased by 15% to better match the
measured burning rate behavior. This increase is justified because it falls within the
stated uncertainty for the kinetic rate of this reaction. These modifications, based on the
values reported in the literature, will be discussed in more detail in a later section when
10
Empirical correlation
Calculated (updated R111x0.5, M&A R114)
Calculated (updated R111x0.5, M&A R114x1.15)
Burning Rate (mm/s)
1.17
r = 1.73e-01 P
b
1.27
r = 1.40e-01 P R= 9.99983e-01
b
1.29
r = 1.44e-01 P R= 9.99955e-01
b
0.1
2 3 4 5 6 7
Pressure (MPa)
Figure 6.1 Comparison of calculated and measured burning rates for nitromethane at 3 to
6 MPa.
Burning rates at higher pressures were also calculated, and shown in Figure 6.2.
The predicted and measured values diverged where the empirical measurements showed
higher-slope regions (above 15 MPa). This is believed to be due to the pyrolysis law
assumed for gasification not fully capturing physical behavior over portions of the
pressure regime.
154
3 4 5
10 10 10
1000.0
Calculated (updated R111x0.5 M&A R114x1.15)
Empirical Correlations
Regression Rate (mm/s)
100.0
10.0
1.0
1.27
r = 0.148 P R= 9.999e-01
b
0.1
1 10 100 1000
Pressure (MPa)
understanding of the processes controlling the flame can be reached. At the most basic
level, equilibrium flame temperature and species concentrations vary very little with
pressure over the range of interest, so the final products and flame temperature of all
conditions should be similar. Equilibrium values calculated using NASAs CEA code176
2,460 over the range shown of 1 to 6 MPa. Because of the significantly negative oxygen
balance, nitromethane does not combust to fully-oxidized products. Water and carbon
monoxide are the most abundant species, with virtually identical mole fractions of about
0.272. Hydrogen and nitrogen follow at 0.228 and 0.167, respectively. A small amount
One may note that the products are highly under-oxidized (in fact, a significant
monopropellant based solely on the species. However, the figure of merit used for most
T flame
Isp (6.1)
W
allowing the lower molecular weight of species that are not fully oxidized (e.g. H2 or CO
vs. H2O or CO2, respectively) to partially offset the loss in energy from incomplete
often somewhat fuel-rich because of the very low molecular weight of unoxidized
3000
0.4000 T
f lame
2000
0.2000
1500
H2O N2
1000
0.1000 CO CO2
0.0900 H2
0.0800
500
0.0700
0.0600
0.0500 0
1.0 2.0 3.0 4.0 5.0 6.0
Pressure (kPa)
Figure 6.3 Variation of nitromethane equilibrium flame temperature and product species
with pressure.
Detailed species and temperature profiles for the nitromethane flame were
obtained as part of the gas-phase solution when burning rate was determined. Calculated
results from 3 MPa are shown in Figure 6.4, and profiles from 6 MPa are in Figure 6.5.
The trends seen in both sets of plots are similar, with the reactions occurring over a
For the 3 MPa case, the temperature increased monotonically from the initial
157
value to the final flame temperature of 2439 K, very close to the equilibrium value found
using CEA. No obvious stages or plateaus were apparent, but some small changes in the
slope could be discerned. The flame thickness (distance to final temperature) was
approximately 700 m. Major species included the reactant (CH3NO2) which was
consumed rapidly, and H2, H2O, CO, CO2, and N2, all of which are equilibrium product
species. Calculated final values were all within a less than half of a percent of the CEA
values, although slightly less CO and slightly more CO2 were present. Although
intermediate species, in addition to lower levels of CH2O, HNO, HONO, and N2O.
At 6 MPa, temperature profile shape was similar, but the flame thickness was
about half of the 3 MPa value, at 350 m. The final flame temperature of 2,451 K was
again very close to the equilibrium value of 2,460 K. Final species fractions of H2, H2O,
CO, CO2, and N2 were nearly identical to the 3 MPa case, as were the presence and level
of the species profiles showed three regions could be distinguished based upon the
chemical composition trends. The description given here is based on the 3 MPa profiles
and CH4, with smaller amounts of CH2O, N2O, HONO, and HNO. HONO and much of
the HNO disappeared by the end of this region. At the relatively low gas-phase
temperature at the burning surface, it may be expected that there would be a significant
induction time. However, due to the low gas-phase velocity near surface (less than 4
cm/s), diffusion of both energy and species is significant, initiating decomposition almost
immediately. For 3 MPa, this region could be considered to span from the surface up to
about 100 m, with temperature increase from surface temperature to about 1300 K.
The next 100 m was an intermediate zone, marked by the partial consumption of
NO and CH4, and total consumption of the other intermediate species (CH2O, N2O, and
Consumption of NO and CH4, with associated rise of N2, H2, and CO to final
values accelerated in the final zone after N2O was depleted. After a small overshoot of
The heat release distribution in the gas-phase zone is shown in Fig. 6.5. The peak
energy release location corresponds the end of the initial zone, when the nitromethane is
completely consumed. Smaller peaks occur subsequently in the intermediate and final
zones. The second peak corresponds to the consumption of CH2O and N2O, with the
162
final peak corresponding to the consumption of CH4 and NO, with only equilibrium
products remaining.
scrutinizing the species production rates and evolution of concentration profiles through
the flame.
Both the carbon and nitrogen pathways started off with the decomposition of the
nitromethane molecule. By far, the majority of CH3NO2 was decomposed through C-N
bond fission (R237) into CH3 and NO2, with production rates ten times higher than
hydrogen abstraction reactions with either CH3 (R243) or H (R239). Other nitromethane
reactions were even less active. The highly-reactive CH3 radical rapidly converted to
163
CH3O or CH4. While the CH4 was relatively stable in the initial stage, the CH3O rapidly
followed through CH2O with significant portion of the formaldehyde going immediately
to CO through HCO (which was very short-lived as an intermediate species with very
low mole fraction). In the intermediate zone, all of the CH2O and some of the CH4 were
consumed, with the CH2O following the same path to CO as in the initial stage (through
HCO). The CH4 that was consumed had a hydrogen removed by either OH or O (R224
and R225), then went through the CH3O path to CO. With some CO being oxidized to
CO2, the concentration of CO actually stayed approximately constant through this region.
In the final zone, CH4 was the only remaining carbon-containing intermediate species,
and was slowly oxidized to CO through similar process seen in the intermediate zone (to
Most nitrogen came from the formation of NO2 from the initial decomposition,
although a small amount (less than ten percent) went through CH2NO2 or HONO. All of
these intermediates then went to NO, the most abundant intermediate species in
nitromethane combustion. Some of this NO was converted to HNO through reaction with
HCO and CH3O (R58 and R248). A small amount of N2O was also formed, either
remaining HNO and HONO intermediates were quickly converted to NO, while N2O was
Although the absolute level of NH was very low in this zone, it acted as an active
to yield N2). After the N2O was completely consumed, NO was the only remaining
164
nitrogen-containing intermediate at the start of the final zone. There, it was slowly
The staged behavior seen here with nitromethane in one-dimensional flow was
with more explicit effects on the temperature profile (distinct stages) with time. Similar
species trends to the current study were noted, with a relatively large amount of NO being
decomposition to CH3 and NO2, which recombined to methyl nitrite (CH3ONO), and then
Bendtsen, and Miller94) included a reaction (R246) to directly convert CH3 and NO2 to
CH3ONO path is still available via R249); most of the production went through the R246
direct path.
propellants.160 There is great commonality between the gas phase mechanism of RDX
and nitromethane; in fact, they are identical after the initial decomposition of
reactive species, and the burning rate of nitromethane was about one-tenth that of RDX at
similar conditions, leading to a flame zone that was much more spatially compressed.
165
over the range where calculated burning rates closely match measured values. Similar
shape and behavior was seen in all cases, with the main difference being the flame width,
15
12
9
6
3 MPa
constant, though there was a small amount of overshoot before settling to the equilibrium
value at the highest pressure condition. This can be seen more clearly in Figure 6.8 for
just the 150 MPa case. Although this case is unrealistic in terms of surface conditions, in
can still be useful to examine trends in the gas phase behavior. Looking at the species
profiles, the higher temperature was caused by rapid reactions over-oxidizing to form
166
excess water and carbon dioxide. The remaining NO and CH4 were oxidized more
slowly, with the H2O and CO2 levels decreasing slowly to equilibrium balances with H2
and CO, respectively. This overshoot was also observed in nitramine simulations.182
(a) temperature
167
Figure 6.8 Temperature and species profiles calculated for 150 MPa.
168
For the high pressure situation, CH2O and N2O were present as intermediate
reaction was preferred at higher temperatures, while the one producing CH2O was more
Sensitivity analysis can give insight to the problem and relative importance of
Sensitivity can be stated in terms of the percent change of the output per percent change
using by differentiating the natural log of the output of interest with respect to the natural
ln X k
(6.2)
ln i
Temperature
ln T
(6.3)
ln i
Burning rate
ln rb
(6.4)
ln i
Here, i is the parameter being varied; it can be the Arrhenius pre-exponential A-factors
expressions can be written for any variable of interest. Because of the nature of the
solutions found in this study, temperature and species sensitivities can be determined for
each discrete time or space point, while a single burning rate sensitivity value
heat feedback from the gas phase, the slope of the temperature profile just above the
surface controls the burning rate. To determine which reactions were most important to
the pre-exponential factors. This sensitivity analysis was conducted at the 3 and 6 MPa
conditions. For 3 MPa, the reactions with the highest temperature sensitivity are plotted
N + CO 2 = NO + CO (R45)
NH + OH = HNO + H (R74)
HNO + H = H 2 + NO (R111)
HNO + NO = N 2 O + OH (R114)
CH 3 + NO 2 = CH 3O + NO (R246)
Near the surface, R111 and R246 had negative temperature sensitivity (increase in rate
resulted in lower temperature), while R45, R74, and R114 are positive. As species and
temperature evolved through the flame, the relative importance and even sign of the
sensitivity of these reactions changed. In general, sensitivities were highest near the
burning surface, and fell to nearly zero by the end of the reaction zone (approximately
170
0.07 cm for the 3 MPa case shown). This is because these rates affect the bulk
temperature the most where the highest chemical reaction rates and intermediate species
occur, just above the surface. Species concentrations were nearly at equilibrium values
by the end of the reaction zone, and changes in these reaction rates have relatively little
were very small, making these particular reactions unimportant in the equilibrium
situation. R74 was also found among the most sensitive reactions for RDX burning rate
by Yetter, et al.198 and R114 is very important in dark zone structure, especially in nitrate
ester propellants.194
slightly higher initial sensitivities, and damping occurring at a shorter distance due to the
more compact flame structure. At 150 MPa, NO became relatively less important, and
R45 and R246 were replaced by reactions converting N2O to N2 (R140 and R142) in the
proportional to the temperature gradient. Therefore, it was expected that reactions that
most strongly affected the initial temperature gradient should result in the greatest effect
on calculated burning rate. To test this, the reaction rates were modified (typically the
pre-exponential parameter was doubled) one at a time, new burning rate solutions found,
and the sensitivities determined. From looking at the slope near the surface in Figure 6.9,
R114 should have a relatively strong positive burning rate sensitivity, while the value for
R111 is expected to be slightly lower and negative. The other reactions should yield
much less effect. The calculated effect of the five most temperature-sensitive reactions
on burning rate is displayed in Figure 6.10. As expected, this correlates closely to the
initial slope of the sensitivity seen in Figure 6.9 since heat feedback at the surface
R45
R74
R111
R114
R246
After identifying the reactions that had the greatest effect on burning rate, rates
were examined in more detail to see if studies that occurred since the mechanism was
first formulated could contribute. It was found that the original expressions for R111 and
For R111, the source cited for the parameters found in the mechanism was Soto
and Page.218 However, the values of both pre-exponential factor and activation energy in
the mechanism were not consistent with those in the cited source. Accordingly, the
mechanism was corrected to match the published values. No more recent studies were
found.
R114 was among the reactions specifically examined by Diau, et al.219 in their
173
study of HNO + NO reactions. Compared to the original parameters from the mechanism
of Miller and Bowman,109 their pre-exponential factor was over four times higher, and the
determined that the value of the HNO heat of formation commonly cited was incorrect.
Based on this updated value, a very recent study Miller and Anderson194 determined that
the true rate of R114 should be twice that originally determined by Diau, et al. using the
incorrect heat of formation. The updated Miller and Anderson value was considered as
Based on the sensitivity results and literature evaluations for R111 and R114,
reaction rates were modified slightly to get burning rates closer to the measured ones.
Using the baseline literature values, calculated burning rate at 3 MPa was about 5%
below the empirical value. In the study of Soto and Page, the parameters cited for R111
are calculated ones; however, the rate differed by up to an order of magnitude from older
experimental values also shown. Although they attributed much of the difference to the
possibility of missing pathways in the previous experimental data analyses, it does not
seem unreasonable to assume a rate falling between the experimental and computational
values. Therefore, a pre-exponential factor of 0.5 times the calculated rate of Soto and
Page was used, with the same activation energy retained. For R114, it was found that
increasing the baseline value for the pre-exponential factor (from Miller and Anderson194)
by 15% produced a better fit to measured data. This small increase is well within the
When the A-factors of reactions 111 and 114 were modified as described above,
174
calculated burning rates (shown by open square symbols in Figure 6.2) much closer to the
measured values were obtained. The modifications introduced were small, and there is
justification for the selection of the modified values, including uncertainty in the
published rates.
accuracy, those of some short-lived highly-reactive radicals are not as well established.
Available databases are constantly being revised to include new and updated values based
Recently, Dong, et al.203 found that laminar flame speeds could be as sensitive to
diffusivities as to kinetics. It has also been pointed out that the methodology and
database used to obtain transport properties in the CHEMKIN package have been
These parameters control the transport of species, momentum, and energy. They will not
have any direct effect on the reactions, but can influence the temperature and
concentration fields, both of which control reaction rates. Effects should be most
significant for species with highest concentrations, although small changes in highly-
reactive species could also influence flame structure. In light of this, a sensitivity
analysis was conducted for selected thermal and transport properties to help quantify the
Thermal conductivity is used not only in the gas-phase code, but also directly in
Fouriers Law to find the gas-phase heat feedback to the condensed phase. Normally, it
175
is calculated from the species mole fraction and temperature found just above the surface
in gas phase solution. Although thermal conductivity can be calculated from the
molecular parameters, a more direct way to isolate the effect on burning rate is by setting
it to different constant values and comparing the results. Through this method, sensitivity
of burning rate to thermal conductivity was found to vary from 1.24 at 3 MPa to 1.03 at 6
MPa. These values are much higher than the sensitivity to any particular chemical
In general, the transport properties could not be varied directly since their
computation and use is embedded in the CHEMKIN 4 PREMIX program binary code.
However, sensitivity to the molecular parameters in the transport database from which
transport parameters were not directly available in any database, and were calculated or
estimated for this study. To find the effect of inaccuracy in these parameters, the burning
rate sensitivities were found by manually varying each parameter a small percentage,
then finding net effect on calculated burning rate. These sensitivities, found in the form
/k
B
Z
rot
for a non-polar molecule when finding the interaction between that molecule and a polar
one. Since nitromethane has a permanent dipole moment and its polarizability is not used
in any case, the sensitivity should be zero. Using nitromethane as an indicator, the
of the data used to obtain these molecular parameters and the importance of the species
affected to the flame structure, the effect on results could still be significant.
177
ln rb
p (6.5)
Ti P
Using this relation, p indicates the fractional variation in burning rate with respect to the
change in initial temperature, in units of inverse Kelvin. Figure 6.12 shows the logarithm
of the burning rate plotted versus propellant initial temperature for a number of chamber
pressure conditions. The p for each pressure can then be deduced from the slope of a
linear curve fit. Figure 6.13 shows the calculated values of p in relation to measured
values over the same range. Although the magnitudes are similar, the measured values
show a stronger function of pressure, declining by about a factor of two over the range
change the sensible enthalpy required to raise the liquid temperature from the initial value
to the surface temperature. As pressure increases, the heat feedback to the surface
higher mass flow rate and lessening of the effect of initial temperature change.
178
2.0
3 MPa ln rb 5 MPa ln rb
4 MPa ln rb 6 MPa ln rb
0.5
0.0
-0.5
240 260 280 300 320 340
T (K)
initial
Figure 6.12 Calculated burning rate as a function of initial temperature and pressure.
179
0.005
Empirical
Calculated
0.004
0.003
(K )
-1
p
0.002
0.001
0.000
2 3 4 5 6 7
Pressure (MPa)
Figure 6.13 Comparison of measured and calculated values for burning rate temperature
sensitivity.
rather, values were selected such that the solution was repeatable and did not change
within the tolerance. The initial uniform grid size (NPTS) in the gas phase was selected
to be 100 points (200 points at high pressures). Using a lower number of initial points
caused convergence problems in some cases; a larger number led to increased calculation
180
time with no change in final solution. The allowable gradients (GRAD) and curvature
(CURV) used to determine where mesh refinement was necessary were chosen to be 0.15
and 0.3, respectively (compared to default values of 0.1 and 0.5). Again, these were
selected to aid convergence of the gas phase solution. Specifying tighter bounds led to a
slightly finer mesh with much longer calculation time to solve, but no observed change in
burning rate solution. Gas phase tolerances were set at 0.002, which provided
CHAPTER 7
7.1. Conclusions
The push for higher performance and reduced toxicity monopropellant has led to
the search for a hydrazine replacement. One promising candidate, nitromethane, was
1. Measurement of basic propulsion parameters such as burning rate were made over
a wide range of conditions for both static (tube) and feeding tests. Three pressure
showed a very thin reaction zone and low-luminosity flame over a stable, smooth
liquid surface. The flame was too thin to be resolved by 25-m thermocouples.
wrapping the CHEMKIN PREMIX gas-phase code coupled with the condensed-
phase model. Predicted burning rates using the model showed good agreement
with measured rates over the subcritical range of 3 to 6 MPa, based on review of
recent literature, two rates (R111 and R114) were adjusted slightly to improve the
match.
4. Calculated species and temperature profiles showed three regions that could be
intermediate species, with smaller percentages of CH2O, N2O, HNO, and HONO
second by consumption of all intermediate species except CH4 and NO, and final
identified the importance of HNO (R74, R111, and R114) reactions to the
temperature profile, and therefore burning rate. Although the absolute levels of
concentration species.
Although significant findings have been made in this research, some additional
questions and refinements have been identified. In general, the next step is to address the
higher pressure behavior (including slope break mechanisms) and investigate why the
183
model does not fully capture the decrease of p (burning rate temperature sensitivity)
As for all but the simplest combustion modeling, the quest for an improved
few items can be singled out for further investigation. More recent reviews such as
Kuo222 and Warnatz199, or other studies of individual reactions can be used to update
higher pressures. Recent mechanisms for more complex propellants or mixtures often
significantly exceed 300 reactions; there may be pathways that have not been captured in
the mechanism utilized here. This particular problem could also benefit from a surface
either in as a unified expression or piecewise in the pressure range. Along with this, a
real-gas treatment for the gas-phase solution at higher pressures would address non-
common techniques and instruments compared to the small spatial dimension of the
temperature and species profiles could be used to verify the selected mechanism and
quantified ways at lower pressure would permit more complete profiles to be generated
184
APPENDIX A
The following output file was generated by NASAs CEA2000 chemical equilibrium
code. A rocket problem for pure nitromethane was run with typical conditions for an
upper stage or space thruster with a chamber pressure of 1000 psia and a nozzle
reac
name NM wt% 100 C 1 H 3 N 1 O 2 h,kJ/mol=-113.09 t,k=298.15
end
OPTIONS: TP=F HP=F SP=F TV=F UV=F SV=F DETN=F SHOCK=F REFL=F INCD=F
RKT=T FROZ=F EQL=F IONS=F SIUNIT=F DEBUGF=F SHKDBG=F DETDBG=F TRNSPT=F
Pc,BAR = 68.947304
Pc/P =
O/F = .000000
POINT ITN T C H N O
PERFORMANCE PARAMETERS
MOLE FRACTIONS
NOTE. WEIGHT FRACTION OF FUEL IN TOTAL FUELS AND OF OXIDANT IN TOTAL OXIDANTS
189
APPENDIX B
This code, implementing the integrated model discussed in the text to solve for
burning rate, was written in FORTRAN 90 and compiled for use on Microsoft Windows
using COMPAQ Visual Fortran 6.6C. Libraries from CHEMKIN 4.0.2 were used. The
PNNL Fortran Library (FLIB) Version 2.02R provided useful file and character string
manipulation functions.223
! nm_flame.f90
!
! FUNCTIONS:
! nm_flame - Entry point of console application.
!
!****************************************************************************
!
! PROGRAM: nm_flame
!
! PURPOSE: Entry point for the console application.
!
!****************************************************************************
!
!
!Additional libraries used:
! libstr
! (...add description of function used)
!
! flib
! (...add description of functions used)
PROGRAM nm_flame
! Variables
IMPLICIT none
CHARACTER(256) error_message
!Shouldn't need to change most of these unless trying out weird stuff
!not all implemented in shell calls, so be careful and check if not default...
therm_data_file = 'therm.dat' !therm filename
therm_output_file = '' !therm output filename
trans_data_file = '' !trans data file
trans_input_file = 'RDX+NM supplemental tran.dat' !supplemental trans data
file
trans_output_file = '' !trans output file
chem_input_file = 'Yetter NM+RDX.inp' !chem input file
chem_output_file = '' !chem output file
premix_input_file = 'pre-mixed_burner__burner_stabilized.inp' !premix input
file
premix_output_file = '' !premix output file
CHEMKIN_app_name = 'CKReactorBurnerStabilizedFlame' !CHEMKIN application/driver
binary
CALL Set_CHEMKIN_env()
!* Initialize variables
loopcount = 0
error_message=''
file name
IF (exists) THEN
success = SYSTEMQQ('%REACTION_DIR%\chemkin40_pc\bin\tran -i tran.dat -o
tran.out &
&-d %REACTION_DIR%\chemkin40_pc\data\tran.dat')
PRINT *, "Preprocessed transport data files...",success
ELSE
success = SYSTEMQQ('%REACTION_DIR%\chemkin40_pc\bin\tran -o tran.out &
&-d %REACTION_DIR%\chemkin40_pc\data\tran.dat')
PRINT *, "Preprocessed transport data file...",success
END IF
!* Main loop - checks rb match for convergence between assumed value and
! and calculated value from gas phase heat transfer
PRINT *, "rb_tol:", rb_tol, "rb_rtol:", rb_rtol
PRINT *, "delta_rb:", delta_rb, "rdelta_rb:", rdelta_rb
!Check for existence of executable output logfile and look for signs of
success
INQUIRE (FILE = 'pre-mixed_burner__burner_stabilized.out', EXIST = exists)
IF (.NOT. exists) THEN
PRINT *, "No pre-mixed_burner__burner_stabilized.out file
exists...please"
PRINT *, "check input files and output logs for possible errors, then
restart"
error_message = 'Missing output logfile '//error_message
EXIT
END IF
!Get the initial (surface) temperature from the gas phase solution in the
!PREMIX output file
Ts = Temp(1)
!* Find heat flux to liquid from gas phase (for interface energy balance)
rb_assumed = flowrate/rho_liquid
!then find difference from assumed value
delta_rb = rb - rb_assumed
195
!Now that I.C. guesses are updated, write out new config file
!
!input file modify/write routines...
!?use the restart from previous run setting by adding RSTR keyword
!to input file
!rewrite old input file with loopcount suffix
!
!This routine assumes input file linelength <= 128 characters
!Use CASE statement to check for keywords and modify lines appropriately.
196
!check for USEV keyword so updated rb (flowrate) is used for new solution
!CALL Check_keyword('USEV')
OPEN(UNIT=30, FILE='newconfig.inp')
TPROcount=0
REWIND (20)
DO WHILE (.NOT. EOF(20))
READ (20,'(A128)') configline
configlinetemp = configline
!PRINT *, TRIM(configline)
CALL pars(configlinetemp,word,nword)
!PRINT *, TRIM(configlinetemp) !pars() returns just 1st word in
configlinetemp?
!DO i=1,nword
! PRINT *, TRIM(word(i))
!END DO
IF (word(1) .EQ. 'FLRT') THEN !update flowrate value for FLRT line
!take new flowrate, turn into string, concatenate, and write
!entire FLRT line to new config file, replacing old one...
!write to 5 digits after decimal (oscillation in convergence at 4)
configline = '' !blank out configline so fresh start for
concatenation
CALL SP_CATNUM(new_flowrate, 'F5', 'FLRT ', configline)
WRITE (30,*) TRIM(ADJUSTL(configline))
PRINT *, " Writing new flowrate: FLRT = ", new_flowrate
ELSE IF (word(1) .EQ. 'TPRO') THEN !temperature profile entry
TPROcount=TPROcount+1
IF (TPROcount .EQ. 1) THEN !update temp to Ts for first TPRO entry
only (x=0)
!NOTE: this assumes first TPRO entry always at x=0!
!take new temp, turn into string, concatenate, and write
!entire TPRO line to new config file, replacing old one...
!write to 5 digits after decimal (saw oscillation in
!convergence at 4)
configline = '' !blank out configline so fresh start for
concatenation
CALL SP_CATNUM(new_Ts, 'F1', 'TPRO 0.0 ', configline)
configline = ADJUSTL(configline)
n=LEN_TRIM(configline)
configline = configline(1:n)//' ! Temperature (K)'
197
!CASE DEFAULT
!write line unchanged to new config file
!WRITE (30,*) TRIM(configline)
!END SELECT
END DO
CALL UNION(shell_cmd_tmp,'.inp',shell_cmd)
CALL Run_shell(shell_cmd, shell_err)
END DO
APPENDIX C
--------------------
ELEMENTS ATOMIC
CONSIDERED WEIGHT
--------------------
1. C 12.0112
2. H 1.00797
3. N 14.0067
4. O 15.9994
--------------------
-------------------------------------------------------------------------------
C
P H
H A
A R
SPECIES S G MOLECULAR TEMPERATURE ELEMENT COUNT
CONSIDERED E E WEIGHT LOW HIGH C H N O
------------------------------------------------------------------------------
-
1. H2 G 0 2.0159E+00 200 6000 0 2 0 0
2. O2 G 0 3.1999E+01 200 6000 0 0 0 2
3. H2O G 0 1.8015E+01 200 6000 0 2 0 1
4. O G 0 1.5999E+01 200 6000 0 0 0 1
5. H G 0 1.0080E+00 200 6000 0 1 0 0
6. OH G 0 1.7007E+01 200 6000 0 1 0 1
7. HO2 G 0 3.3007E+01 200 6000 0 1 0 2
8. H2O2 G 0 3.4015E+01 200 6000 0 2 0 2
9. CH2O G 0 3.0026E+01 200 6000 1 2 0 1
10. HCO G 0 2.9019E+01 300 5000 1 1 0 1
11. CO G 0 2.8011E+01 200 6000 1 0 0 1
12. CO2 G 0 4.4010E+01 200 6000 1 0 0 2
13. N G 0 1.4007E+01 200 6000 0 0 1 0
14. N2 G 0 2.8013E+01 200 6000 0 0 2 0
15. NO G 0 3.0006E+01 200 6000 0 0 1 1
16. NO2 G 0 4.6005E+01 200 6000 0 0 1 2
17. NH G 0 1.5015E+01 200 6000 0 1 1 0
18. NH2 G 0 1.6023E+01 200 3000 0 2 1 0
19. NH3 G 0 1.7031E+01 200 6000 0 3 1 0
20. NNH G 0 2.9021E+01 200 6000 0 1 2 0
21. HNO G 0 3.1014E+01 200 3500 0 1 1 1
22. HONO G 0 4.7013E+01 300 5000 0 1 1 2
23. HCN G 0 2.7026E+01 200 6000 1 1 1 0
24. HNC G 0 2.7026E+01 200 6000 1 1 1 0
25. N2O G 0 4.4013E+01 200 6000 0 0 2 1
26. CN G 0 2.6018E+01 200 6000 1 0 1 0
27. C2N2 G 0 5.2036E+01 200 6000 2 0 2 0
28. NCN G 0 4.0025E+01 200 6000 1 0 2 0
29. NCO G 0 4.2017E+01 200 6000 1 0 1 1
30. CNO G 0 4.2017E+01 200 6000 1 0 1 1
31. HNCO G 0 4.3025E+01 200 6000 1 1 1 1
200
(k = A T**b exp(-E/RT))
REACTIONS CONSIDERED A b E
APPENDIX D
APPENDIX E
Following are the additions made to the thermodynamics and transport data files used by
data are taken from the extensive database maintained by Prof. Alexander Burcat at
Technion - Israel Institute of Technology. The most recent version can be found in the
As the complete database contains over 1,100 entries, only the species of interest for
nitromethane combustion were extracted. These are taken from the April 2005 release.
A few species that were included in the combustion mechanism were not in Burcats
database.
THERMO
!CH3ONO values taken from old CH3NO2.dat file (origin unknown--Yetter NM?)
CH3ONO 103190C 1H 3N 1O 2G 0300.00 4000.00 1500.00 1
0.11361289E+02 0.04159349E-01-0.04145670E-05-0.01695140E-08 0.03028732E-12 2
-0.12814815E+05-0.03545434E+03 0.14903451E+01 0.02645433E+00-0.02112331E-03 3
0.09414399E-07-0.01811204E-10-0.09125782E+05 0.01813766E+03 4
!Following species missing from Burcat database copied from orig Yetter RDX inp
HCO 121286H 1C 1O 1 G 0300.00 5000.00 1000.00 1
0.03557271E+02 0.03345573E-01-0.01335006E-04 0.02470573E-08-0.01713851E-12 2
0.03916324E+05 0.05552299E+02 0.02898330E+02 0.06199147E-01-0.09623084E-04 3
0.01089825E-06-0.04574885E-10 0.04159922E+05 0.08983614E+02 4
NNH T07/93N 2H 1 0 0G 200.000 6000.000 1
0.37667544E+01 0.28915082E-02-0.10416620E-05 0.16842594E-09-0.10091896E-13 2
0.28650697E+05 0.44705067E+01 0.43446927E+01-0.48497072E-02 0.20059459E-04 3
-0.21726464E-07 0.79469539E-11 0.28791973E+05 0.29779410E+01 0.30009828E+05 4
HONO 31787H 1N 1O 2 G 0300.00 5000.00 1000.00 1
0.05486893E+02 0.04218065E-01-0.01649143E-04 0.02971877E-08-0.02021148E-12 2
-0.01126865E+06-0.02997002E+02 0.02290413E+02 0.01409922E+00-0.01367872E-03 3
0.07498780E-07-0.01876905E-10-0.01043195E+06 0.01328077E+03 4
NCO EA 93 N 1C 1O 1 0G 200.000 6000.000 1
0.51521845E+01 0.23051761E-02-0.88033153E-06 0.14789098E-09-0.90977996E-14 2
0.14004123E+05-0.25442660E+01 0.28269308E+01 0.88051688E-02-0.83866134E-05 3
0.48016964E-08-0.13313595E-11 0.14682477E+05 0.95504646E+01 0.21347373E+05 4
HOCN BDEA94H 1N 1C 1O 1G 300.000 5000.000 1368.000 1
5.89784885E+00 3.16789393E-03-1.11801064E-06 1.77243144E-10-1.04339177E-14 2
-3.70653331E+03-6.18167825E+00 3.78604952E+00 6.88667922E-03-3.21487864E-06 3
5.17195767E-10 1.19360788E-14-2.82698400E+03 5.63292162E+00 4
210
HCNO BDEA94H 1N 1C 1O 1G 300.000 5000.000 1382.000 1
6.59860456E+00 3.02778626E-03-1.07704346E-06 1.71666528E-10-1.01439391E-14 2
1.79661339E+04-1.03306599E+01 2.64727989E+00 1.27505342E-02-1.04794236E-05 3
4.41432836E-09-7.57521466E-13 1.92990252E+04 1.07332972E+01 4
H2CNH 41687C 1H 3N 1 G 0300.00 4000.00 1000.00 1
0.05221589E+02 0.04748526E-01-0.04179159E-05-0.02606612E-08 0.04703140E-12 2
0.08657219E+05-0.04500776E+02 0.02365879E+02 0.06780570E-01 0.02423000E-04 3
-0.06157824E-08-0.01615097E-10 0.09971141E+05 0.01230718E+03 4
END
Data for most of the species of interest were included with the CHEMKIN distribution;
these were used without modification or review. For the species not found in this listing,
found in the literature. Species of interest not included in the standard distribution were
REFERENCES
1
Sutton, G. P., Rocket Propulsion Elements: An Introduction to the Engineering of
Rockets, 6th Edition, John Wiley and Sons, Inc., New York, 1992.
2
Clark, J. D., Ignition! : An Informal History of Liquid Rocket Propellants, Quinn and
Boden Company, Inc., Rahway, NJ, 1972.
3
Wucherer, E. J., Christofferson, S. S., and Reed, B., Assessment of High
Performance HAN-Monopropellants, AIAA Paper 2000-3872, 36th
AIAA/ASME/SAE/ASEE Joint Propulsion Conference, Huntsville, AL, July 16-19,
2000
4
Anon., Nitromethane, Chapter 5 in Hazards of Chemical Rockets and Propellants.
Volume III, Liquid Propellants, Hannum, J. A. E., ed., CPIA Publication 394,
Chemical Propulsion Information Agency, Laurel, Maryland, September 1984.
5
Vogelsanger, B., Chemical Stability, Compatibility, and Shelf Life of Explosives,
Chimia, Vol. 58, No. 6, pp. 401-408, 2004.
6
Anon., CH3NO2 in Physical and Thermodynamic Properties of Pure Chemicals:
Data Compilation, ed. T. E. Daubert, et al., Taylor & Francis, Washington, DC, 1997.
7
David R. Lide, ed., CRC Handbook of Chemistry and Physics, Internet Version 2005,
<http://www.hbcpnetbase.com>, CRC Press, Boca Raton, FL, 2005.
8
Jones, W. M. and Giauque, W. F., The Entropy of Nitromethane. Heat Capacity of
Solid and Liquid. Vapor Pressure, Heats of Fusion and Vaporizaion, Journal of the
American Chemical Society, Vol. 69, pp. 983-987, 1947.
9
Wagman, D. D., et al., The NBS Tables of Chemical Thermodynamic Properties,
Journal of Physical and Chemical Reference Data, Vol. 11, Suppl. 2, 1982.
10
Cox, J. D. and Pilcher, G., Thermochemistry of Organic and Organometallic
Compounds, Academic Press, New York, NY, 1970.
11
"Critical Constants," in CRC Handbook of Chemistry and Physics, Internet Version
2005, David R. Lide, ed., <http://www.hbcpnetbase.com>, CRC Press, Boca Raton,
FL, 2005.
12
P. J. Linstrom and W. G. Mallard, eds., NIST Chemistry WebBook, NIST Standard
Reference Database Number 69, June 2005, National Institute of Standards and
Technology, Gaithersburg MD, 20899 (http://webbook.nist.gov).
13
McCullough, J. P., Scott, D. W., Pennington, R. E., Hossenlopp, I. A., and
Waddington, G., Nitromethane: The Vapor Heat Capacity, Heat of Vaporization,
Vapor Pressure and Gas Imperfection; the Chemical Thermodynamic Properties from
0 to 1500 K, Journal of the American Chemical Society, Vol. 76, pp. 4791-4796,
1954.
14
Berman, H. A. and West, E. D., Density and Vapor Pressure of Nitromethane 26 to
200 C, Journal of Chemical and Engineering Data, Vol. 12, No. 2, April 1967, pp.
197-199.
15
Kharbanda, O. P., Latent Heats of Vaporization, Industrial Chemist, Vol. 31, No. 3,
pp. 124-127, 1955.
16
Gallant, R. W., Physical Properties of Hydrocarbons. Part 37Nitrogen Containing
Compounds, Hydrocarbon Processing, Vol. 48, No. 8, pp. 117-126, August 1969.
212
17
Brasch, J. W., Irreversible Reaction of Nitromethane at Elevated Pressure and
Temperature, Journal of Physical Chemistry, Vol. 84, pp. 2084-2085, 1980.
18
Piermarini, G. J., Block, S., and Miller, P. J, Effects of Pressure on the Thermal
Decomposition Kinetics and Chemical Reactivity of Nitromethane, Journal of
Physical Chemistry, Vol. 93, pp. 457-462, 1989.
19
Courtecuisse, S., Cansell, F., Fabre, D., and Petitet, J.-P., Phase Transitions and
Chemical Transformations of Nitromethane Up to 350 C and 35 GPa., Journal of
Physical Chemistry, Vol. 102, No. 2, pp. 968-974, 1995.
20
Agrawal, P. M., Rice, B. M., and Thompson, D. L., Molecular Dynamics Study of
the Melting of Nitromethane, Journal of Chemical Physics, Vol. 119, No. 18, pp.
9617-9627, 2003.
21
Simon, F. and Glatzel, G., Bemurkungen zur Schmelzdruckkurve, Zeitschrift fr
anorganische und allgemeine Chemie, Vol. 178, pp. 309-316, 1929.
22
Babb, Jr., S. E., Parameters in the Simon Equation Relating Pressure and Melting
Temperature, Reviews of Modern Physics, Vol. 35, No. 2, pp. 400-413, 1963.
23
Kechin, V. V., Thermodynamically Based Melting-Curve Equation, Journal of
Physics: Condensed Matter, Vol. 7, No. 3, pp. 531-535, 1995.
24
Alavi, S. And Thompson, D. L., A Molecular-Dynamics Study of Structural and
Physical Properties of Nitromethane Nanoparticles, Journal of Chemical Physics,
Vol. 120, No. 21, pp. 10231-10239, 2004.
25
Corelli, R. M., Note Sulla Decomposizione Esotermica del Nitrometano e Sua
Utilizzazaione, La Chimica e LIndustria, Vol. 31, No. 12, pp. 436-439, 1949.
26
Cass, W. G., Exothermic Decomposition of Nitromethane, Aircraft Engineering,
Vol. 22, pp. 238-239, August 1950.
27
Zwicky, F. et al., Operation of Jet Propulsion Motors with Nitroparaffin, US Patent
No. 2,433,943, Jan. 6, 1948.
28
Zwicky, F. and Ewing, F. J., Decomposition of Nitro-Paraffins in Jet Propulsion
Motor Operation, US Patent No. 3,048,007, Aug. 7, 1962.
29
Weissbluth, M., A Study of the Nitromethane-Oxygen Combination as a Rocket
Propellant, Jet Propulsion Laboratory, California Institute of Technology, Pasadena,
Progress Report 1-22, 1944.
30
Klein, E. L., Fuels for Jets, SAE Journal, pp. 22-28, December 1947.
31
Zwicky, F. and Ross, C. C., Nitromethane Excels as Rocket Propellant, SAE
Journal, pp. 22-24, June 1949.
32
Anon., Evaluate Nitromethane for Rockets, Aviation Week, pp. 27-33 and 37, May
16, 1949.
33
Foster, C. R., Experimental Study of Nitromethane Rocket Motors with Varying
Fineness Ratio, Characteristic Length, and Chamber Pressure, Jet Propulsion
Laboratory, California Institute of Technology, Pasadena, Report 4-33, 27 May 1947.
34
Bellinger, F., et al., Chemical Propellants: Stability of Mononitromethane,
Industrial and Engineering Chemistry, pp. 1320-1323, July 1948.
35
Bellinger, F., et al., Chemical Propellants: Mononitromethane, Industrial and
Engineering Chemistry, pp. 1324-1331, July 1948.
36
Kindsvater, H. M., Kendall, K. K., Mueller, K. H., and Datner, P. P., Research on
213
1951.
52
Frjacques, M. C., Sur la cintique de dcomposition thermique en phase vapeur des
nitroalcanes, Comptes rendus hebdomadaires des seances de lAcademie des
sciences, Vol. 231, pp. 1061-1062, 1950.
53
Hillenbrand, Jr., L. J. and Kilpatrick, M. L., The Thermal Decomposition of
Nitromethane, Journal of Chemical Physics, Vol. 19, No. 3, p. 381, 1951.
54
Hillenbrand, Jr., L. J. and Kilpatrick, M. L., The Thermal Decomposition of
Nitromethane, Journal of Chemical Physics, Vol. 21, No. 3, pp. 525-535, 1953.
55
Gray, P., Yoffe, A. D., and Roselaar, L., Thermal Decomposition of the
Nitroalkanes, Transactions of the Faraday Society, Vol. 51, pp. 1489-1497, 1955.
56
Mueller, K. H., The Thermal Decomposition of Nitromethane at High Pressures,
Journal of the American Chemical Society, Vol. 77, No. 13, p. 3459-3462, 1955.
57
Makovky, A. and Gruenwald, T. B., The Thermal Decomposition of Nitromethane
Under High Pressure, Transactions of the Faraday Society, Vol. 55, Part 6, pp. 952-
958, 1959.
58
Bradley, J. N., Shock-Wave Decomposition of Nitroparaffins. Part 1.Mass-
Spectrometric Study of Nitromethane Decomposition, Transactions of the Faraday
Society, Vol. 57, Part 10, pp. 1750-1756, 1961.
59
Hiraoka, H. and Hardwick, R., Emission and Dissociation of NO2 in Shock Waves,
Journal of Chemical Physics, Vol. 39, No. 1, pp. 2361-2365, 1963.
60
Mader, C. L., A Study of the One-dimensional, Time-dependent Reaction Zone of
Nitromethane and Liquid TNT, Los Alamos Scientific Laboratory Report LA-3297,
Los Alamos, New Mexico, August 1965.
61
Mader, C. L., The Time-dependent Reaction Zones of Ideal Gases, Nitromethane and
Liquid TNT, Los Alamos Scientific Laboratory Report LA-3764, Los Alamos, New
Mexico, October 1967.
62
Borisov, A. A., Kogarko, S. M., and Skachkov, G. I., Thermal Decomposition of
Nitromethane, translated from Kinetika i Kataliz, Vol. 7, No. 4, pp. 589-596, 1966.
63
Borisov, A. A., Zaslonko, I. S., and Kogarko, S. M., Spectroscopic Investigation of
the Decomposition of Nitromethane in Shock Waves, Combustion, Explosion, and
Shock Waves (Engl. trans.), Vol. 4, No. 3, pp. 222-227, 1968.
64
Crawforth, C. G. and Waddington, D. J., Reactions of Nitroalkanes in the Gas Phase:
Part 1.Pyrolysis of Nitromethane, Transactions of the Faraday Society, Vol. 65,
Part 5, pp. 1334-1349, 1969.
65
Benson, S. W. and ONeal, II., E., Kinetic Data on Gas-Phase Unimolecular
Reactions, NSRDS-NBS 21, U.S. Government Printing Office, Washington, D. C.,
1970.
66
Phillips, L. and Shaw, R., Tenth International Symposium on Combustion, p. 453,
1964.
67
Shtern, V. Ya., Mechanism of Gas-Phase Nitrogenation of Alkanes with Carbon
Dioxide in Khimicheskaya Kinetika i Tsepnye Reaktsii (Chemical Kinetics and
Chain Reactions), pp. 286-322, Nauka, Moscow, 1966.
68
Dubikhin, V. V., Nazin, G. M., and Manelis, G. B., Thermal Decomposition of
Nitromethane, Russian Chemical Bulletin (Engl. trans.), Vol. 20, No. 6, pp. 1247-
215
1249, 1971.
69
Glnzer, K. and Troe, J., Thermische Zerfallsreaktionen von Nitroverbindungen I:
Dissoziation von Nitromethan, Helvetica Chimica Acta, Vol. 55, p. 2884-2893,
1972.
70
Zaslonko, I. S, Kogarko, S. M., Mozzhukhin, E. B., and Petrov, Yu. P., Thermal
Decomposition of Nitromethane in Shock Waves, translated from Kinetika i Kataliz,
Vol. 13, No. 5, pp. 1113-1118, 1972.
71
Alster, J., Slag, N., Dewar, M. J. S., Ritchie, J. P., and Wells, C., Chemical Paths in
the Shock Initiation of Detonation. A Theoretical Study, Internationale
Jahrestagung 1979; Combustion and Detonation Processes, Fraunhufer Institut fr
Treff und Explosivstoffe, Karlsruhe, West Germany, pp. 591-609, 1979.
72
Dewar, M. J. S., Ritchie, J. P., and Alster, J., Thermolysis of Molecules Containing
NO2 Groups, Journal of Organic Chemistry, Vol. 50, No. 7, pp. 1031-1036, 1985.
73
Perche, A., Tricot, J. C., and Lucquin, M., The Pyrolysis of Nitromethane. Part 1.
Experimental Study: Nitromethane Alone and in the Presence of Additives, Journal
of Chemical Research (S), pp. 116-117, 1979.
74
Perche, A., Tricot, J. C., and Lucquin, M., The Pyrolysis of Nitromethane. Part 2.
Determination of a Simplified Mechanism; Bibliographic Review of Rate Constants,
Journal of Chemical Research (S), pp. 304-305, 1979.
75
Perche, A., Tricot, J. C., and Lucquin, M., The Pyrolysis of Nitromethane. Part 3.
Simulation of a Simplified Mechanism, Journal of Chemical Research (S), pp. 306-
307, 1979.
76
McKee, M. L., Ab Initio Study of Rearrangements on the CH3NO2 Potential Energy
Surface, Journal of the American Chemical Society, Vol. 108, pp. 5784-5792, 1986.
77
Krause, H. H., Eisenreich, N., Hansen, R., Koppenhfer, W., and Volk, F.,
Untersuchung neuer Monergol-Varianten, ICT-Industrieauftrag 14/87.
78
Cohn, R. H., Thermal Decomposition of Nitromethane by Adiabatic Gas
Compression, Ph. D. Thesis, New Mexico Institute of Mining and Technology,
Socorro, NM, December, 1990.
79
Rice, B. M. and Thompson, D. L., Classical Dynamics Studies of the Unimolecular
Decomposition of Nitromethane, Journal of Chemical Physics, Vol. 93, No. 11, pp.
7986-8000, 1990.
80
Roszak, S. and Kaufman, J. J., Ab Initio Multiple Reference Double-Excitation
Configuration-Interaction Ground and Excited State Potential Curves for
Nitromethane Decomposition, Journal of Chemical Physics, Vol. 94, No. 9, pp.
6030-6035, 1991.
81
Melius, C. F., Thermochemical Modeling: I. Application to Decomposition of
Energetic Materials, in Chemistry and Physics of Energetic Materials, S. N. Bulusu,
ed., Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 21-49, 1990.
82
Zeman, S., New Application of Kinetic Data of the Low-Temperature Thermolysis
of Nitroparaffins, Thermochimica Acta, Vol. 261, pp. 195-207, 1995.
83
Khrapkovskii, G. M., Shamov, A. G., Shamov, G. A., and Shlyapochnikov, V. A.,
Some Peculiarities of the Influence of Molecular Structure on the Activation Energy
of Radical Gas-Phase Decomposition of Aliphatic Nitro Compounds, Russian
216
Chemical Bulletin (Eng. trans.), Vol. 45, No. 10, pp. 2309-2315, 1996.
84
Cant, N. W., et al., The Reactions of Nitromethane in the Gas Phase and on a Co-
ZSM5 Catalyst Under the Conditions of the Methane/NOx SCR Reaction, Catalysis
Letters, Vol. 46, pp. 207-212, 1997.
85
Bromly, et al., Kinetic and Thermodynamic Sensitivity Analysis of the NO-
Sensitized Oxidation of Methane, Combustion Science and Technology, Vol. 115,
No. 4-6, pp. 259-296, 1996.
86
Djebaili-Chaumeix, N., Abid, S., and Paillard, C. E., Shock Tube Study of the
Nitromethane Decomposition and Oxidation, Paper 1810 in Proceedings of the 21st
International Symposium on Shock Waves, ed. A. F. P. Houwing, Great Keppel
Island, Australia, July 20-25, St. Lucia, Queensland: University of Queensland, Dept.
of Mechanical Engineering, 1998.
87
Hsu, D. S. Y. and Lin, M. D., Laser Probing and Kinetic Modeling of NO and CO
Production in Shock-Wave Decomposition of Nitromethane Under Highly Diluted
Conditions, Journal of Energetic Materials, Vol. 3, pp. 95-127, 1985.
88
Zaslonko, I. S., Petrov, Yu. P., and Smirnov, V. N., Thermal Decomposition of
Nitromethane in Shock Waves: The Effect of Pressure and Collision Partners,
Kinetics and Catalysis (Eng. trans.), Vol. 38, No. 3, pp. 321-324, 1997.
89
Troe, J., Fall-off Curves of Unimolecular Reactions, Berichte der Bunsen-
Gesellschaft fr Physikalische Chemie, Vol. 78, No. 5, pp. 478-488, 1974.
90
Zhang, Y.-X. and Bauer, S. H., Modeling the Decomposition of Nitromethane,
Induced by Shock Heating, Journal of Physical Chemistry B, Vol. 101, pp. 8717-
8726, 1997.
91
Zhang, Y.-X. and Bauer, S. H., Gas-Phase Decomposition Mechanisms of C-NO2,
N-NO2 Energetic Materials: Reevaluations, International Journal of Chemical
Kinetics, Vol. 31, p. 655-673, 1999.
92
Gilbert, R. G., Luther, K., and Troe, J., Theory of Thermal Unimolecular Reactions
in the Fall-off Range. II. Weak Collision Rate Constants, Berichte der Bunsen-
Gesellschaft fr Physikalische Chemie, Vol. 87, pp. 169-177 ,1983.
93
Manaa, M. R. and Fried, L. E., DFT and ab Initio Study of the Unimolecular
Decomposition of the Lowest Singlet and Triplet States of Nitromethane, Journal of
Physical Chemistry A, Vol. 102, pp. 9884-9889, 1998.
94
Glarborg, P., Bendtsen, A. B., and Miller, J. A., Nitromethane Dissociation:
Implications for the CH3 + NO2 Reaction, International Journal of Chemical
Kinetics, Vol. 31, No. 9, pp. 591-602, 1999.
95
Khrapkovskii, G. M., Chachkov, D. V., and Shamov, A. G., Effect of Molecular
Structure on the C-N Bond Strength in a Nitroalkane Series: I. Nitromethane,
Fluoronitromethanes, Chloronitromethanes, and Fluorochloronitromethanes, Russian
Journal of General Chemistry (Eng. trans.), Vol. 71, No. 9, pp. 1449-1456, 2001.
96
Nguyen, M. T., et al., Nitromethane-Methyl Nitrate Rearrangement: A Persistant
Discrepancy Between Theory and Experiment, Journal of Physical Chemistry A,
Vol. 107, pp. 4286-4291, 2003.
97
Wodtke, A. M., Hintsa, E. J., and Lee, Y. T., The Observation of CH3O in the
Collision Free Multiphoton Dissociation of CH3NO2, Journal of Chemical Physics,
217
1987.
112
Miller, J. A., et al., "On the Conversion of HCN to NO and N2 in H2-O2-HCN-Ar
Flames at Low Pressure," Proceedings of the Twentieth Symposium (International) on
Combustion, The Combustion Institute, Pittsburgh, PA, p. 673, 1985.
113
Glarborg, P., Miller, J. A., and Kee, R. J., Kinetic Modeling and Sensitivity Analysis
of Nitrogen Oxide Formation in Well-Stirred Reactors, Combustion and Flame, Vol.
65, pp. 177-202, 1986.
114
Atkinson, R., Lloyd, A. C., and Winges, L., An Updated Chemical Mechanism for
Hydrocarbon/NOx/SO2 Photooxidations Suitable for Inclusion in Atmospheric
Simulation Models, Atmospheric Environment, Vol. 16, No. 6, pp. 1341-1355,
1982.
115
Kee, R. J., Miller, J. A., and Jefferson, T. H., CHEMKIN: A General-Purpose,
Problem-Independent Transportable, Fortran Chemical Kinetics Code Package,
Sandia Laboratories Technical Report SAND 80-8003, 1980.
116
Wodtke, A. M., Hintsa, E. J., and Lee, Y. T., ONR Technical Report NR 659-
819/10/85, 1985.
117
Yetter, R.A. and Rabitz, H., "Modeling the Decomposition of Nitromethane at High
Pressures," unpublished manuscript, 1990.
118
Melius, C. F., Thermochemical Modeling: II. Applications to Ignition and
Combustion of Energetic Materials, in Chemistry and Physics of Energetic
Materials, S. N. Bulusu, ed., Kluwer Academic Publishers, Dordrecht, The
Netherlands, pp. 51-78, 1990.
119
Melius, C. F., The Thermochemistry and Reaction Pathways of Energetic Material
Decomposition and Combustion, Philosophical Transactions of the Royal Society of
London. Series A, Vol. 339, pp. 365-376, 1992.
120
Melius, C. F., The Thermochemistry and Reaction Mechanisms in the Decomposition
of Energetic Materials, Sandia Report SAN093-1224C, presented at the 1994 Joint
USA-Russia Energetic Material Technology Symposium, Livermore, CA, May 18-
25, 1994.
121
Lutz, A. E., Kee, R. J., and Miller, J. A., SENKIN: A Fortran Program for Predicting
Homogeneous Gas Phase Chemical Kinetics with Sensitivity Analysis, Technical
Report SAND87-8248, Sandia National Laboratories, 1987.
122
Shaw, R., DeCarli, P. S., Ross, D. S., Lee, E. L., and Stromberg, H. D., "Thermal
Explosion Times of Nitromethane, Perdeuteronitromethane, and Six Dinitroalkanes as
a Function of Temperature at Static High Pressures of 1-50 kbar," Combustion and
Flame, Vol. 35, pp. 237-247, 1979.
123
Afeefy, H.Y., Liebman, J.F., and Stein, S.E., "Gas Phase Thermochemistry Data,"
NIST Standard Reference Database Number 69, Eds. P.J. Linstrom and W.G.
Mallard, March 2003, National Institute of Standards and Technology, Gaithersburg
MD, 20899.
124
Glarborg, P., Alzueta, M. U., Dam-Johansen, K., and Miller, J. A., Kinetic Modeling
of Hydrocarbon/Nitric Oxide Interactions in a Flow Reactor, Combustion And
Flame, Vol. 115, pp. 127, 1998.
125
Blais, N. C., Engelke, R., and Sheffield, S. A., Mass Spectroscopic Study of the
219
1956.
141
Starkman, E. S., Nitromethane as a Piston Engine Fuel, SAE Technical Paper
540186, 1954.
142
Bush, K. C., Germane, G. J., and Hess, G. L, Improved Utilization of Nitromethane
as an Internal Combustion Engine Fuel, SAE Technical Paper 852130, International
Fuels and Lubricants Meeting and Exhibition, Tulsa, OK, October 21-24, 1985.
143
Rahm, H., Low-Polluting Fuel, German Patent No. DE 10314275 A1, July 28, 2005.
144
Milewska, A. and Szydlowski, J., Deuterium Isotope Effects on Miscibility Curves
of Nitromethane + Isobutanol, Journal of Chemical Engineering Data, Vol. 44, pp.
505-508, 1999.
145
Tu, C-H., Lee, S-L., and Peng, I-H., Excess Volumes and Viscosities of Binary
Mixtures of Aliphatic Alcohols (C1-C4) with Nitromethane, Journal of Chemical
Engineering Data, Vol. 46, pp. 151-155, 2001.
146
Tronscoso, J., et al., Temperature Dependence of Densities and Speeds of Sound of
Nitromethane + Butanol Isomers in the Range (288.15-308.15) K, Journal of
Chemical Engineering Data, Vol. 46, pp. 312-316, 2001.
147
Schad, R. E., Nitrohydrocarbon Gels, United States Patent No. 2,891,852, June 23,
1959.
148
Novak, L. J., Thickened or Gelled Normally Liquid Nitroparaffin Propellants, United
States Patent No. 2,921,846, Jan 19, 1960.
149
Jones, L. R., Thickened Nitroparaffin Containing Stabilizer, United States Patent No.
3,318,741, May 9, 1967.
150
Griffith, G. L., Carroll, Jr., W. J, and Schwoyer, W. L., Process of Gelling TMETN
Nitrocellulose Explosives Using Nitroparaffin Solvents and TMETN Nitrocellulose
Explosive Gels, United States Patent No. 3.489,623, Jan. 13, 1970.
151
Purcell, R. F., Nitroparaffin Gels, United States Patent No. 2,954,350, Sept. 27, 1960.
152
Idar, D. I. and Asay, B. W., Improved Characterization of Nitromethane,
Nitromethane Mixtures, and Shaped-Charge Jet Properties, Propellants, Explosives,
and Pyrotechnics, Vol. 24, pp. 1-6, 1999.
153
Teipel, U. and Forter-Barth, U., Mechanical Properties of Gel Propellants with
Nanoparticles, Journal of Energetic Materials, Vol. 22, pp. 69-82, 2004.
154
Teipel, U. and Forter-Barth, U., Rheological Behavior of Nitromethane Gelled with
Nanoparticles, Journal of Propulsion and Power, Vol. 21, No. 1, pp. 40-43, 2005.
155
Boyer, E., Risha, G., Kuo, K. K., and Devendorf, T., Combustion Characteristics of
Non-Toxic Non-Hypergolic Bi-Propellants, AIAA Paper 2002-4297, 38th
AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit, Indianapolis,
Indiana, 7 10 July 2002.
156
Chang, Y. P., Boyer, E., and Kuo, K. K., Combustion Behavior and Flame Structure
of XM46 Liquid Propellant, Journal of Propulsion and Power, Vol. 17, No. 4, pp.
800-808, 2001.
157
Zenin, A., HMX and RDX: Combustion Mechanism and Influence on Modern
Double-Base Propellant Combustion, Journal of Propulsion and Power, Vol. 11, No.
4, pp. 752-758, 1995.
158
Chang, Y. P. and Kuo, K. K., Assessment Of Combustion Characteristics and
221
Moscow, 1977.
174
Kondrikov, B. N., Raikova, V. M., Sviridov, E. M., and Sidorova, T. T., Data of
Combustion Laboratory of Mendeleev University of Chemical Technology.
175
Kopicz, et al., Combustion Behavior and Thermochemical Properties of JA2
Propellant, in Challenges in Propellants and Combustion: 100 Years After Nobel
(Kuo ed.), Begell House, Inc., New York, NY, pp. 559-573, 1997.
176
Gordon, S. and McBride, B. J., Computer Program for Calculation of Complex
Chemical Equilibrium Compositions and Applications I. Analysis, NASA Reference
Publication 1311, 1994.
177
Parr, T. and Hanson-Parr, D., Nonintrusive Diagnostic Techniques for Research on
Nonsteady Burning of Solid Propellants, in Nonsteady Burning and Combustion
Stability of Solid Propellants (De Luca, et al. eds.), American Institute of Aeronautics
and Astronautics, Washington, D.C., 1992.
178
Vanderhoff, J. A. and Kotlar, A. J., Simultaneous Determination of Temperature and
OH Concentrations in a Solid Propellant Flame, Proceedings of the 23rd Symposium
(International) on Combustion, pp. 1339-1344, 1990.
179
Vanderhoff, J. A., Teague, M. W., and Kotlar, A. J., Determination of Temperature
and NO Concentrations through the Dark Zone of Solid Propellant Flames,
Proceedings of the 24th Symposium (International) on Combustion, pp. 1915-1922,
1992.
180
Margolis, S. B., Williams, F. A, and Armstrong, R. C., Influences of Two-Phase
Flow in the Deflagration of Homogeneous Solids, Combustion and Flame, Vol. 67,
pp. 249-258, 1987.
181
Li, S. C., Williams, F. A., and Margolis, S. B., Effects of Two-Phase Flow in a
Model for Nitramine Deflagration, Combustion and Flame, Vol. 80, pp. 329-349,
1990.
182
Liau, Y-C. and Yang, V., Analysis of RDX Monopropellant Combustion with Two-
Phase Subsurface Reactions, Journal of Propulsion and Power, Vol. 11, No. 4, pp.
729-739, 1995.
183
Liau, Y.-C., Kim, E. S., and Yang, V., A Comprehensive Analysis of Laser-Induced
Ignition of RDX Monopropellant, Combustion and Flame, Vol. 126, pp. 1680-1698,
2001.
184
Kim, E. S., Yang, V., and Liau, Y.-C., Modeling of HMX/GAP Pseudo-Propellant
Combustion, Combustion and Flame, Vol. 131, pp. 227-245, 2002.
185
Kuo, K. K. and Lu, Y. C., Modeling of Physicochemical Processes of Burning RDX
Monopropellants, Proceedings of the Twentieth International Pyrotechnics Seminar,
Colorado Springs, CO, pp. 583-600, July 25-29, 1994.
186
Kuo, K. K., Lu, Y. C., and Tseng, Y. S., Modeling and Numerical Simulation of
Physicochemical Processes Occurring in the Two-Phase Foam Layer of Burning
RDX, in Challenges in Propellants and Combustion: 100 Years After Nobel (Kuo
ed.), Begell House, Inc., New York, NY, pp.1133-1145, 1997.
187
Brill, T. B., Surface Pyrolysis Products and Flame Diagnostics by FTIR
Spectroscopy, in Non-Intrusive Combustion Diagnostics (ed. K. K. Kuo and T.
Parr), Begell House, 1994.
223
188
Alexander Burcat's Ideal Gas Thermochemical Database
<ftp://ftp.technion.ac.il/pub/supported/aetdd/thermodynamics>; September 2004.
189
Miller, M. S. and Anderson, W. R., Energetic-Material Combustion Modeling with
Elementary Gas-Phase Reactions: A Practical Approach, Chap. 2.12 in Solid
Propellant Combustion Chemistry, Combustion, and Motor Interior Ballistics, ed. V.
Yang, T. B. Brill, and W. Z. Ren, Vol. 185, Progress in Astronautics and Aeronautics,
AIAA, Reston, VA, pp. 501531, 2000.
190
Alexander Burcat's Ideal Gas Thermochemical Database
<ftp://ftp.technion.ac.il/pub/supported/aetdd/thermodynamics>; September 2005.
191
R. J. Kee, et al., CHEMKIN Release 4.0.2, Reaction Design, San Diego, CA, 2005.
192
Zarko, V. E., Gusachenko, L. K., and Rychkov, A. D., The Effect of Phase
Transitions on Combustion Stability of Melted Energetic Materials, 28th
International Annual Conference of ICT, Fraunhofer Institut Chemische Technologie,
Karlsruhe, Germany, pp. 16-1 16-13, 1997.
193
Zenin, A., HMX and RDX: Combustion Mechanism and Influence on Modern
Double-Base Propellant Combustion, Journal of Propulsion and Power, Vol. 11, No.
4, pp. 752758, 1995.
194
Miller, M. S. and Anderson, W. R., Burning-Rate Predictor for Multi-Ingredient
Propellants: NitrateEster Propellants, Journal of Propulsion and Power, Vol. 20,
No. 3, pp. 440-454, 2004.
195
Ward, M. J., Son, S. F., and Brewster, M. Q., Steady Deflagration of HMX with
Simple Kinetics: A Gas Phase Chain Reaction Model, Combustion and Flame, Vol.
114, pp. 556568. 1998.
196
R. J. Kee, et al., Theory Manual, CHEMKIN Release 4.0.2, Reaction Design, San
Diego, CA, 2005.
197
Kee, R. J., Grcar, J. F., Smooke, M. D., Miller, J. A., and Meeks, E., PREMIX: A
FORTRAN Program for Modeling Steady Laminar One-Dimensional Premixed
Flames, Reaction Design, San Diego, CA, April 1998.
198
Yetter, R. A., Dryer, F. L., Allen, M. T., and Gatto, J. L., Development of Gas-Phase
Reaction Mechanisms for Nitramine Combustion, Journal of Propulsion and Power,
Vol. 11, No. 4, pp. 683-697, 1995.
199
Warnatz, J., Maas, U., and Dibble, R. W., Combustion: Physical and Chemical
Fundamentals, Modeling and Simulation, Experiments, Pollutant Formulation, 2nd
Edition, Springer-Verlag, New York, NY, 1999.
200
R. J. Kee, et al., Input Manual, CHEMKIN Release 4.0.2, Reaction Design, San
Diego, CA, 2005.
201
Burcat, A., "Third Millennium Ideal Gas and Condensed Phase Thermochemical
Database for Combustion," Technion Aerospace Engineering (TAE) Report # 867
January 2001.
202
Alexander Burcat's Ideal Gas Thermochemical Database
<ftp://ftp.technion.ac.il/pub/supported/aetdd/thermodynamics>; April 2005.
203
Dong, Y., et al., Extinction of Premixed H2/Air Flames: Chemical Kinetics and
Molecular Diffusion Effects, Combustion and Flame, Vol. 142, No. 4 , pp. 374-387,
2005.
224
204
Hirschfelder, J. O, Curtiss, C. F., and Bird, R. B., Molecular Theory of Gases and
Liquids, John Wiley and Sons, Inc., New York, NY, 1964.
205
Svehla, R. A., Estimated Viscosities and Thermal Conductivities of Gases at High
Temperatures, NASA Technical Report R-132, Accession No. N63-22862, National
Aeronautics and Space Administration, Washington, D. C., 1962.
206
Kee, R. J., Coltrin, M. E., and Glarborg, P., Chemically Reacting Flow: Theory and
Practice, John Wiley and Sons, Inc., Hoboken, NJ, 2003.
207
Mohr, P. J. and Taylor, B. N., Fundamental Physical Constants, in CRC Handbook
of Chemistry and Physics, Internet Version 2005, David R. Lide, ed.,
<http://www.hbcpnetbase.com>, CRC Press, Boca Raton, FL, 2005.
208
McClellan. A. L., Tables of Experimental Dipole Moments, Vol. 1, W. H. Freeman
and Co., San Francisco, CA, 1963.
209
Microsoft Excel 2002 (10.6501.6735) SP3, Microsoft Corporation, Redmond, WA,
2004.
210
Bttcher, C. J. F., Theory of Electric Polarization, Volume I: Dielectrics in Static
Fields, 2nd ed., Elsevier Scientific Publishing Co., New York, NY, 1973.
211
Rhatigan, J. L, Effects of Gas-Phase Radiation and Detailed Kinetics on the Burning
and Extinction of Solid Fuel, NASA/TP-2001-210770, National Aeronautics and
Space Administration, June 2001.
212
Parker, J. G., Rotational and Vibrational Relaxation in Diatomic Gases, Physics of
Fluids, Vol. 2, No. 4, pp. 449-462, 1959.
213
Warnatz, J., Influence of Transport Models and Boundary Conditions on Flame
Structure in Numerical Methods in Flame Propagation, N. Peters and J. Warnatz,
ed., Vieweg-Verlag, Wiesbaden, pp. 87-111, 1982.
214
Liau, Y-C, A Comprehensive Analysis of RDX Propellant Combustion and Ignition
with Two-Phase Subsurface Reactions, Ph. D. Thesis, The Pennsylvania State
University, 1997.
215
Kuo, K. K., Principles of Combustion, 2nd ed., John Wiley and Sons, Inc., Hoboken,
NJ, 2005.
216
Yang, V., Modeling of Supercritical Vaporization, Mixing, and Combustion
Processes in Liquid-Fueled Propulsion Systems, Proceedings of the Combustion
Institute, Vol. 28, Part I, pp. 925-942, 2000.
217
Brill, T. B., Multiphase Chemistry Considerations at the Surface of Burning
Nitramine Monopropellants, Journal of Propulsion and Power, Vol. 11, No. 4, pp.
740-751, 1995.
218
Soto, M. R. and Page, M., Ab initio Variational Transition-State-Theory Reaction-
Rate Calculations for the Gas-Phase Reaction H+HNOH2+NO, Journal of
Chemical Physics, Vol.. 97, No. 10, pp. 7287-7296, 1992.
219
Diau, E. W., Halbgewachs, M. J., Smith, A. R., and Lin, M. C., Thermal Reduction
of NO by H2: Kinetic Measurement and Computer Modeling of the HNO + NO
Reaction, International Journal of Chemical Kinetics, Vol. 27, pp. 867-881, 1995.
220
Anderson, W. R., Heats of Formation of HNO and Some Related Species,
Combustion and Flame, Vol. 117, Nos. 12, pp. 394403, 1999.
221
Paul, P. and Warnatz, J., A Re-evaluation of the Means Used to Calculate Transport
225
Eric Boyer was born on November 28, 1971 and grew up amidst the rolling hills
from the Bermudian Springs High School, he attended Princeton University and received
Pennsylvania State University, he has worked on many different projects at the High
Pressure Combustion Laboratory (HPCL), including liquid, solid, gel, and hybrid rocket
systems as well as gun and mortar studies. He has also served as HPCL Safety Officer