45th AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit AIAA 2009-4890

2 - 5 August 2009, Denver, Colorado

Feasibility Study and Demonstration

of an Aluminum and Ice Solid Propellant
Tyler D. Wood1, Mark A. Pfeil2, Timothee L. Pourpoint3, John Tsohas4, and Steven F. Son5
Purdue University, West Lafayette, IN, 47907, USA

T.L. Connell, Jr6, Grant A. Risha7 and Richard A. Yetter8

The Pennsylvania State University, University Park, PA, 16802, USA

Aluminum-water reactions have been proposed and studied for several decades for
underwater propulsion systems, and other applications such as hydrogen generation.
Aluminum and water has also been proposed as a propellant and there have been proposals
for other refrigerated propellants that could be mixed, frozen in place and used as solid
propellants. However, little work has been done to determine the feasibility of these
concepts. With the recent availability of nano-scale aluminum (nAl), a simple binary
formulation of nAl and water is now plausible. Nano-sized aluminum has a lower ignition
temperature than micron-sized aluminum particles, partly due to its high surface area, and
burning times are much faster than micron Al. We have previously reported that frozen nAl
and ice mixtures considered are stable, as well as insensitive to electrostatic discharge,
impact and shock. Here we report a study of the feasibility of a nAl-ice propellant in small-
scale rocket experiments. The focus here is not to develop an optimized propellant, however
improved formulations are possible and could be explored in future work. Several static
motor experiments have been conducted, including using a flight-weight casing. In this
flight-weight test the grain configuration was 6.75” long, 3” outside diameter, with a 1”
center perforation. It produced a peak 500lbf of thrust at 1650 psi. The flight weight casing
will be used in the first sounding rocket test of an aluminum-ice propellant soon.

Nomenclature
a,n = propellant burning rate coefficients
Ab = burning area
At = throat area
c* = characteristic velocity
Dt = throat diameter
dt = time increment
ISP = specific impulse
g = gravity
m = mass
m˙ = mass flow rate
pc = chamber pressure
rb = burning rate
W = web thickness
€
1
Graduate Student, Mechanical Engineering, 500 Allison Road, Chaffee Hall, West Lafayette, IN 47907,
Email: tdwood@purdue.edu. AIAA Member.
2
Graduate Student, Aeronautics & Astronautics Engineering, Purdue University,
3
Research Assistant Professor, Aeronautics & Astronautics, Purdue University, Senior AIAA Member
4
Graduate Student, Aeronautics & Astronautics Engineering, Purdue University, AIAA Member
5
Associate Professor, Mechanical Engineering and Aeronautics & Astronautics (courtesy), Purdue University,
AIAA Member
6
Graduate Student, Mechanical Engineering, The Pennsylvania State University, AIAA Member
7
Assistant Professor, Division of Business and Engineering, The Pennsylvania State University, Altoona College,
AIAA Member
8
Professor, Mechanical Engineering, The Pennsylvania State University, AIAA Member
1
American Institute of Aeronautics and Astronautics

Copyright © 2009 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved.
ε = thickness of alumina deposit
ρp = propellant density
φ = mixture ratio

Subscript
in = in
out = out
p = propellant

I. Introduction

A LUMINUM powder is a common ingredient in conventional rocket propellant to increase specific impulse, ISP,
as well as stability. The properties and recent availability of nano-scale aluminum (nAl) has motivated research
of material. For example, Kuo et al.1 discussed the potential use of nano-sized powders for rocket propulsion in a
recent paper. Many of the advantages listed for these particles are shorter ignition delay, faster burn times, and the
possibility to act like a gelling agent, thus replacing the inert low-energy gellants. Using nAl has been shown to
produce a significant increase in performance of propellants2,3. Researchers showed that replacing 50µm particles
with the same amount of nominally 100nm particles in AP-based propellants could result in a burning rate increase
of up to 100%4. Most of these characteristics can be attributed to the high specific surface area1, 5. The possible
disadvantages of nAl are the reduction in active aluminum content, electrostatic discharge (ESD) sensitivity when
dry, and rheology difficulties. Other research has been conducted pairing this increased reactivity of nAl with less
reactive oxidizers such as water in addition to conventional oxidizers6, 7. Aluminum and water propellants may
prove to be suited for deep space exploration, in that potentially propellants could be made in situ from available
water and aluminum. Also, the products of this propellant, mainly H2 and Al2O3, are relatively non-toxic, making it
a “green” propellant8, 9.
The objectives of this paper are to present recent results of nAl/ice (ALICE) small-scale static experiments.
Another objective was to develop larger scale (kilogram scale) mixing procedures that produce a consistent material.
A classical mixer and a newly available Resodyn mixer were considered. The burning rate has been characterized
for these propellants in a strand burner. Recent results of the static experiments are also compared against internal
ballistic predictions. Trajectory simulations have also been made for the flight of the sounding rocket to be flown
this summer.

II. Background
While widely available bulk commercial nAl has only recently been developed, the water-aluminum reaction has
been of considerable interest since at least the 1940’s. In 1942, Rasor10 filed a patent, which proposed to use
seawater and aluminum to provide the propulsion for a submarine. While thermodynamically this reaction would be
viable, the kinetics of bulk aluminum would not yield complete reaction. This was evident by work done by Elgert
et al.11 that used U235 to melt the aluminum. However only 0.2% of the aluminum reacted, even though
temperatures reached 2200°F. Even work done by Leibowitz et al.12, who tried igniting the aluminum with a laser,
found that if the melting temperature of the aluminum oxide was not reached, an ignition would not occur.
There have also been several studies investigating the use of micron sized Al powders with water for purpose of
underwater propulsion13, 14. In 2004, Ingenito et al.8 produced a paper discussing the potential uses for an aluminum-
water mixture for space propulsion. Using the NASA CEA equilibrium code, they calculated specific impulse, ISP,
over various expansion ratios, O/F ratios, and pressures. The vacuum ISP was calculated over 300s, greater than that
of most solid propellants, at an O/F typically around 1.2 (Expansion ratio of 100). A typical AP-based solid rocket
propellant has vacuum ISP values ranging from 260 s to 300 s15. Small motors are expected to have much lower
measured ISP values. Ingenito et al. also proposed the idea of adding hydrogen peroxide (H2O2) to increase
performance. Indeed, many other propellant formulations are possible.
Nano-scale aluminum can dramatically increase the reaction rates of aluminum and water. Ivanov et al.16
reported the earliest combustion work with nanoaluminum and water. Mixtures of stoichiometric mixtures of
aluminum and water were considered. They reported that they needed 3% polyacrylamide to thicken (or gel) the
water or the nAl-water reaction would not occur without the gelling agent16. In 2006, Risha et al.17 reported
combustion of nAl and water for the first time without the use of a gelling agent. The nanoaluminum used likely
had a higher surface area than previously used, which may explain the different observations. Risha et al. found that
stoichiometric mixtures of nAl-water propellant have a pressure dependence of around 0.47 and have densities of
around 1.5 g/cc. The mass and linear burning rates are much faster than those of advanced propellants17. While the
2
American Institute of Aeronautics and Astronautics
burning rate for a fuel-lean mixture was lower than a stoichiometric mixture, the pressure exponents were similar.
This suggests that the propellant has the same dependence on pressure, regardless the amount of excess water.
Sippel et al.18 performed several experiments to characterize a nAl-water propellant. Non-pyrophoric aluminum
powders have a passivation layer of alumina. This is used by the nAl manufacturers, such as Novacentrix Inc. and
Argonide Corp., to prevent pyrophoric ignition with air, and to extend shelf life. Even with this passivation, nAl-
water can have a short shelf life, on the order of weeks, when exposed to moist air due to its high affinity for
oxidizers18. This inert alumina shell inversely affects the performance of the mixture5, 17. Due to the smaller size of
the aluminum particles (from micron to nano), the alumina layer accounts for more of the mass. The oxide layer
does not add to the energy produced by the particle, and the active aluminum content decreases with the size of the
particles and the increase in oxide passivation layer thickness. Dokhan et al.19 estimates that active aluminum
content of micron aluminum is 99.5% or better, while oxide passivated nano-aluminum typically ranges from 50%
to 95%18, 20, 21.
Franson et al.22, at SNPE, performed perhaps the first work on the implementation of ALICE in a rocket motor
configuration. The outer diameter of the grain was 86mm, with an inner perforation diameter of 60mm, and a length
of 157mm. The total mass of the grain using a combination of micron and nanoscale aluminum was measured at
550g. Pressure readings were estimated and recorded for the test. Post inspection of the motor revealed large
amounts of alumina residue in the chambers. Analysis of the slag showed that an estimated 17% of the initial
aluminum did not participate in the reaction. This helped explain the 1.6MPa pressure observed in contrast to the
3MPa expected pressure, and a subsequent lack in performance using a mixture of nanoscale and micron aluminum.
In previous work by our group, we examined aging issues of aluminum and water. Nanoscale aluminum in
liquid water will oxidize. One method to increase the shelf life is to freeze the aluminum water mixture to form
ALICE. Sippel et al.18 showed that nAl and water stored at -25C, retained the original active aluminum content after
40 days, modeling the procedure by Cliff et al.21. This was a significant increase from the previous findings that had
a value less then 10% after the same time period in liquid water18. In addition, Sippel et al. found that over the
course of six months, the active aluminum content was unchanged within the uncertainty of the measurement.
Safety testing was also performed on the experimental propellant18. Impact sensitivity testing showed that a
mixture of frozen nAl using nominally 80nm powders and water (ALICE) had a drop height greater then the
capacity of the experiment apparatus (>2.2meters), while dry 200µm AP had a drop height of 38.5cm. ESD testing
showed that stoichiometric ALICE (80nm nAl) had an energy threshold greater than 1.5J, over one thousand times
the amount of energy typically released in a human ESD event. Shock sensitivity was performed to determine
whether the propellant would propagate a detonation wave. The results displayed the stability of the frozen
propellant using 80nm nAl, with no indication of damage to the witness plate18. All of these tests show that nAl/ice
is a very insensitive propellant.
Complementary testing has also been conducted at Penn State, with a center-perforated configuration that shows
promising results (also reported in these proceedings). Motors with outer diameters 0.75” and 1.5”, and an inner
diameter of 0.25” and 0.5” respectively, have given thrust and chamber pressure readings that are repeatable23.
Recent testing has also been performed with a 3” motor using a hand-mixed grain that compares fairly well with
results presented below.

III. Mixing Techniques

Early mixtures in this work were obtained using a Ross (Charles Ross & Son
Company. Hauppauge, New York) DPM-1Q dual planetary mixer or by hand.
However, inconstancies in mixing and packing densities motivated other
approaches. A Resodyn (Resodyn Acoustic Mixer Inc, Butte, Montanta) LabRAM
resonating mixer (Fig. 1) is currently being used to mix the ALICE propellant. The
LabRAM mixer operates by applying a force to the system being mixed and
varying the frequency at which that force is applied until the resonant frequency of
the system being mixed is found. Once found, the system continues to apply a
force at that frequency at the user specified intensity (ranging from 0 to 100)24.
The density and viscosity of the system will change as it mixes causing the
resonant frequency of the system and the energy put into the mixture (measured by
the acceleration level) to change. The mixer is designed to track the resonant
frequency of the system and make changes to the mixing frequency while mixing
to match the resonant frequency.
Figure 1. ResoDyn LabRAM
resonating mixer
3
American Institute of Aeronautics and Astronautics
 The changes of acceleration and frequency provide important information of the stage of how mixed the
propellant is. Typically, the frequency of the mixer increases for a while during the mix and then drops, while the
acceleration exhibits a general continual increase. These changes occur due to variance of the propellant properties
throughout mixing. These changes can be seen in Fig. 2. ALICE initially starts out as de-ionized water and
aluminum powder (80nm from Novacentrix) and begins to form clumps until it becomes a uniform paste like
(gelled) substance. The properties of the propellant reach a uniform state—in other words the propellant is fully
mixed—when the acceleration and the frequency finally level off for a period of time. This can be seen in Fig. 3(b).
The unmixed propellant was originally placed in taped plastic bags as mixing with Resodyn mixer began. These
bags were inserted into an ice bath within a solid container. The ice bath was used to keep the propellant at low
temperatures to inhibit undesired reactions. After mixing at one minute intervals, the bags were removed and
inspected visually to determine if they were uniformly mixed. A sample of the acceleration and frequency output
from a bag mix can be seen in Fig. 3(a). The problem with the bags is that they can sometimes leak or break. The
contents within the bag were always inspected for consistency and eliminated if apparent variations in propellant
characteristics were found. Six or more bags were needed to mix all of the propellant for an ALICE rocket grain
depending on the length of the desired grain. These bags were then combined in the grain case while being vibrated
and packed after which the grain was placed in the freezer. Using this method, well-formed and consistent grains
can be obtained.
Mixing procedures are still being refined at this time in hopes of making mixing quicker and easier while
eliminating any variance in propellant consistency. A method was developed recently in which a smaller container
was placed within an ice bath in a larger container. This method has eliminated leaks and made it possible to
observe the acceleration and frequency more consistently to determine when the mixture is mixed. Those are the
results shown on Figs. 2 and 3(b). This new mixing method and other improvements will be explored more fully in
the future and incorporated into the mixing of the ALICE grain.

(a) (b) (c)

Figure 2. Images of various stages of mixing: a) Mixing consistency after first cycle; b) Mixing consistency
after second cycle; c) Mixing consistency after final cycle.

(a) (b)
Figure 3. Traces from the Resodyn mixer: a) Acceleration and frequency of consecutive multiple mixing
cycles; b) Acceleration and frequency of single mixing cycle of near-constant intensity.

4
American Institute of Aeronautics and Astronautics
 IV. Burning Rate Measurements Technique and Results
In previous work with the Ross dual planetary mixer, stoichiometric mixtures proved to be too viscous for the
size of nAl used. The propellant became too thick to mix effectively with the Ross mixer. These challenges could
be overcome by using a powder with less surface area. This viscous behavior prompted the current ALICE mixtures
to be fuel-lean with an target equivalence ratio, φ, of 0.75. Fuel-lean mixtures had an overall decrease in burning
rate when compared to stoichiometric mixtures, but Risha et al.17 did show similar pressure exponents for both the
fuel-lean and stoichiometric mixtures.
Mixing procedures used with the Resodyn mixer have evolved and improved throughout this project. Initial
procedures were developed based on the equivalence ratio of 0.75. However, safety concerns related to the reactive
nature of the nano-aluminum powder led to the decision of passivating the powder in air for two days prior to
mixing. This passivation step lowered the active aluminum content by about 4% leading to an equivalence ratio
closer to 0.71 and providing for a less reactive propellant. Again, the formulation studied here is far from optimum.
Future research with the ResoDyn resonating mixer will focus on mixing a stoichiometric mixture successfully.
It is important to quantify changes in the burning rate, since several modifications have been made to the mixing
procedure since results found by Sippel et al.18 Strand burn experiments were performed using material from each
mixing batch used to produce static fire grains. Propellant is loaded into 8 mm ID tubes and the samples are
immediately frozen. The samples are tested in a high-pressure combustion bomb. Prior to combustion, the bomb is
pressurized with argon to a determined pressure. During combustion the chamber pressure is recorded by a data
acquisition system. A video of the event is analyzed to determine the linear burning rate of the propellant.
Measurements are repeated for a series of pressures and a power law is used to fit a burning rate as a function of
pressure. Shown on Fig. 4 is the burning rate of the propellant mixed in the resonating mixer. Over 25 tests were
performed and averaged in the results shown. The pressure exponent for this mixing procedure is 0.57, which is
somewhat larger pressure dependency then the Al-water propellant tested by Risha et al.17

Figure 4. Burning rate data of ALICE propellant mixed with the ResoDyn mixer is shown here.

V. Motor Performance Prediction

An internal ballistics analysis of the combusting ALICE motor grain was developed using a lumped-parameter
model. The control volume considered in this model takes into account the geometry of the grains tested in the
battleship and flightweight configurations. These configurations are summarized in Table 1.

5
American Institute of Aeronautics and Astronautics
 Table 1. Aluminum and Ice Grain Geometries Tested to Date at Purdue University
Grain Dimensions Casing Dimensions
Test Chamber Nozzle
Outer Diameter Inner Bore Length
Length Throat
[in] Diameter [in] [in] [in] Diameter [in]
1 to 3 3 1 ~3.5 5 0.36
4 to 5 3 1 ~5 5 0.42
6 3 1 ~6.75 7 0.52

While a simple approach, the assumptions inherent to a lumped-parameter model are quite appropriate in the
present application as the grains tested had low aspect ratios L/D ranging from 1.2 for the 3.5” long grains to 2.3 for
the 7” long grains and, therefore, the pressure variations along the chamber length can be neglected15. Furthermore,
while propellant and motor parameters are adjusted in the model, detail accounting of potentially important two-
phase flow losses or nozzle flow losses is not within the scope of the present study. Instead, the model is used to
predict the peak chamber pressure and thrust developed by the ALICE grains and to indicate the history of both
parameters based on the measured burning rate and the geometry of the grain.
The ALICE propellant formulation assumed in the model has an equivalence ratio of 0.71 and a characteristic
velocity of 1360 m/s. Further, based on previous experimental results reported in the literature25 and theoretical
performance calculations, specific impulse of 210 s is assumed for the thurst calculations.
The results presented below include that of two variants of the model. In the first variant, the aforementioned
propellant characteristics and nozzle geometries are assumed as nominal. It is used to predict the maximum thrust
and chamber pressure prior to experimental testing of a new grain or chamber geometry. In the second variant,
combustion and flow losses in the combustion chamber and through the nozzle are evaluated with model. These
losses are taken into account in two ways: first, since post-test examination of the experimental hardware reveals
alumina deposition on the throat and the expansion section of the nozzle, a simple deposition model is included in
the analysis. The thickness of the alumina deposit is assumed to increase linearly with time up to the deposit
thickness measured upon examination of the hardware. Second, performance losses are included by reducing the
nominal propellant characteristic velocity and specific impulse values until a resonable agreement with the
experimental data is obtained.
A final simplifying assumption included in both variants of the model is that the total impulse and mass flow rate
produced by the igniter are negligeable compared to that of the ALICE grain. The validity of this assumption is
discussed below. At any given instant in a lumped-parameter model, all exposed surfaces in the control volume are
assumed to contribute to amount of mass produced by the combustion of the propellant,


 .
 (1)



Conversely, the mass flow exiting the nozzle is given by,


 .
 (2)



Combining Eqs. 1 and 2 with the conservation of mass equation under steady state conditions leads to,

. (3)

Equation 3 can then be solved for the chamber pressure using St-Robert’s burning rate law, rb = aPcn,

. (4)

Since neither end of the grain is inhibited, the ALICE grain burning surface area is a function of the grain outer and
inner bore diameter and the grain length as given by,
6
American Institute of Aeronautics and Astronautics
 , (5)

with both Ri and L functions of the web thickness consumed normal to the local burn surface. The web thickness,
W, can therefore be defined as the integral of the burning rate history as given by

(6)

The theoretical mass flow rate and thrust can then be calculated based on Eqs. 1 or 2 and


 .
 (7)


Both variants of the lumped parameter model incorporate Eqs. 1 to 7 using a method of Euler numerical integration
with an adequately small time step (typically 1 ms). The second model variant reflects the previously mentioned
performance losses and the alumina deposition on the nozzle according to Eq. 8,


 ,
 (8)



where, ε is the thickness of the alumina deposit measured around the circumference of the nozzle throat. The
chamber pressure and thrust profiles calculated with both variants of the model are shown in Fig. 5 for a 5” long
ALICE grains. Modeling results for the 6.75” long ALICE grain are provided below along with the experimental
data.

Figure 5. Calculated chamber pressure and thrust for 5” long ALICE grain

As shown on Fig. 5, a peak chamber pressure of ~2100 psi is calculated with both variants of the model.
However, the peak pressure obtained with the second variant follows a longer chamber pressurization period and
7
American Institute of Aeronautics and Astronautics
occurs a quarter of a second later than with the first variant. This delay is a result of the reduced characteristic
velocity, assumed to be 85% of nominal in the second model variant. The peak pressures calculated with both
variants are almost identical as a result of the assumed alumina deposition model.
Also shown on Fig. 5 is the reduction in peak thrust from ~450 lbf to ~400 lbf from the first to the second variant
of the model. This reduction is the result of the lower specific impulse value assumed in the second variant of the
model. The main modeling results obtained for a 5” long grain are summarized in Table 2. Similar observations
resulting from calculations with longer grains are presented below. As illustrated by the computional results
outlined above, the lumped parameter model provides the information necessary to prepare static test fires of ALICE
grains of increasing length.
Table 2. Modeling summary for 5” long grain
Model Assumptions Total Impulse [N-s] Peak Pc [psi] Peak Thrust [lbf]
No losses for 5" long grain 1484 2130 445
With deposit & losses for 5" long grain 1336 2150 400

VI. ALICE Battleship Static Thrust Stand Experiments

Several static rocket tests have been conducted in the Purdue University Propulsion laboratory. The test cell for
the static tests has a remote control room, where experiments are monitored and initiated. Pressure and thrust are
recorded using LabView, and a 16-bit National Instruments, 32 channel data acquisition system. At least two
video cameras are used to observe and record the experiments. One camera monitors the outside where the plume is
expelled, and another high-speed camera, recording at 300 fps, monitors the side profile of the exhaust plume.
Based on the strand burn tests, the ALICE propellant combustion does not perform well at pressures less than
1000 psi; therefore a thick steel “battleship” motor casing is used (see Fig. 6). This casing has been sized to
withstand internal pressures exceeding 5,000 psi, to ensure a sufficient factor of safety. However, constraints in the
design of the flight-weight casing influenced the operating pressure of ALICE to be below 3000 psi. To eliminate an
additional variable between the battleship tests and the flight-weight tests, the same bolts were selected to secure the
ALICE motor assembly together. These bolts are designed to fail around 3200 psi so over-pressurization does not
result in the failure of the casing. With the anticipation that fluctuation in mixing and casting will cause variations in
performance, nozzle throat diameter is varied to provide a nominal goal pressure of 1500-2000 psi peak pressure.
The battleship casing is attached horizontally to the metal stand frame. The metal framing is attached to a pair of
flexures, which transfers the thrust produced by the engine to a 1000 lbf load cell (Interface, Scottsdale, Arizona).
Chamber pressure is measured using two PMP 1260 diaphragm pressure transducers, (Druck (a division of GE
Electrics), Billerica, MA) with a 0.25% full-scale accuracy.

Figure 6. Image of the battleship motor casing

Following a few experimental tests with various igniter motor sizes, the igniter of choice in all test
configurations was a commercially available Aerotech H180 motor26. A summary of the motor specifications of
interest in the present study is provided in Table 3. The reported ISP of 178s is not unexpected for small motors such
as these. As listed in Table 3, the Aerotech H180 motor has a total impulse of 218 N-s or about 15% of the total

8
American Institute of Aeronautics and Astronautics
impulse predicted with the first variant of the lumped parameter model for a 5” long ALICE grain (Table 2). While
a smaller igniter would be highly desirable, the selected igniter size is necessary for reliable and fast ignition of the
ALICE formulation evaluated in the present study. The required igniter could be minimized and formulations could
be developed that dramatically decrease the ignition energy required, but that is beyond the scope of this study.

Table 3. Aerotech H180 Motor Specifications26

Parameter Value Unit

Outer Diameter 29.0 mm
Total Length 23.8 cm
Total Weight 252 g
Propellant Weight 124 g
Average thrust 180.0 N
Maximum Thrust 228.5 N
Total Impulse 217.7 N-s
Burn Time 1.3 s
Isp 178 s

Several tests have been performed with the battleship motor. Initial testing started with 3.5” long grains. The
results of these tests are not presented herein for conciseness. Following three successful tests with the 3.5” long
grains, the length of the grain was increased to 5” to provide more thrust and better approximate the scale required
for the future sounding rocket. The experimental results of the two tests performed with 5” long grains (Test 4 & 5)
are presented and compared with the modeling results in Fig. 7 and Table 4 below.

Figure 7. Comparison of 5” long ALICE motor tests with lumped parameter models
Although the two tests are not precisely replicated, there are several key points to note. First, the length and
packing densities of both 5” long grains varied by 2.3% and 4.8% respectively with the first 5”-class grain about
0.25 inches longer than the second one. Second, the pressure rise on both tests are extremely similar. Aluminum

9
American Institute of Aeronautics and Astronautics
agglomeration on the nozzle or variations in casting could explain the differences in peak pressure, but it is
reassuring to see that the rise in pressure is similar. In addition, the experimental peak pressures and peak thursts
compare well with the modeling results thus providing a increased level of confidence for performance prediction of
longer grains. The experimental and modeling results obtained for the 5”-class grains are provided in Table 4
including the calculated total impulse values which are of particular interest for the souding rocket trajectory
predictions.

Table 4. Comparison of Modeling and Experimental Results for 5” long grains

Total Impulse [N-s] Peak Pc [psi] Peak Thrust [lbf]
Model assuming no losses for 5" long grain 1484 2130 445
Model with deposit & losses for 5" long grain 1336 2150 400
5" long grain experimental results of Test 4 970 2160 480
5" long grain experimental results of Test 5 890 1700 350

An analysis of both videos shows a non-oscilatory plume (Fig. 8), which indicated a steady burn. The first
picture is the start of the igniter flame, and initial chamber pressureization. As the H-180 motor burns, gases expand
in the ALICE casing and exit the nozzle as a dark smoke. Based on the recorded data (pressure & thrust), it is
believed that ALICE begins to burn in the second picture. This is evident from the sudden oscillatory change in
thrust from the load cell, that has been consistent thoughout the battleship tests. As the pressure increases further, the
flame continues to increase in size until the peak pressure is reached. The pressure and thrust decay rapidly
following the consumption of the ALICE grain.

Figure 8. Plume image as recorded by the high-speed camera.

VII. Flight-weight Rocket Design

A goal of the project is to demonstrate an ALICE powered sounding rocket. While characterization and static
tests have been proof of the concept, this will be the first flight test implementing an ALICE propellant. Flight tests
are necessary to study conditions that can not be duplicated in a test cell such as the effect of rocket acceleration on
alumina agglomeration, and propellant grain structural integrity.

10
American Institute of Aeronautics and Astronautics
 The ALICE flight-vehicle consists of an all-carbon-fiber, minimum diameter, 98mm high power rocketry kit
which is composed of two fuselage sections, connected together by a carbon-fiber interstage coupler and an avionics
bay which contain two redundant R-DAS (Rocket Data-Acquisition System) Tiny units (AED Electronics; The
Netherlands) as shown in Fig. 9(a). The R-DAS units are pre-programmed to eject a drogue parachute at apogee and
a main parachute at a pre-determined altitude. An ogive nose cone is placed on the forward end of the vehicle and
three carbon-fiber fins are attached to the aft end in order to provide aerodynamic stability. The fins are attached
with carbon-fiber plain weave cloth by using a wet hand layup technique to apply the cloth from fin-tip to fin-tip.
Following the layup process, vacuum bagging is used to provide pressure on the composite layer assembly in order
to remove any excess resin and improve bond strength. An exploded view of the flight vehicle is shown in Fig. 9
(b).
A test launch of the flight-vehicle was performed at a remote site located approximately 12 miles west of West
Lafayette, IN. Appropriate approvals were obtained for this test and planned tests. The purpose of the launch was
to verify that the avionics, recovery, and structural sub-systems were working properly, and to ensure aerodynamic
stability throughout all phases of the flight. The all-carbon-fiber flight-vehicle was powered by an Aerotech K780R
commercial 75mm solid rocket motor, which produces an average total impulse of 2,400 N-sec for a burn time of 3
seconds. The motor ignited as planned, produced approximately 200 lbf of initial thrust, and lifted the 23 lb vehicle
with an initial acceleration of 6.5Gs as recorded by both on-board R-DAS flight computers. The vehicle
successfully cleared the Mobile Launch Platform, and after a 3.8 second burn achieved a maximum velocity of 395
mph (Mach 0.52), at an altitude of 850 ft as shown in Fig. 11(b). The vehicle coasted to an apogee of 4,480 ft,
which took place in 18 seconds. As soon as the R-DAS units detected apogee, the drogue parachute ejection charges
fired and the drogue parachute deployed as planned. The vehicle began descending at a rate of 57 ft/sec under the
drogue parachute, and at 85 seconds into the flight the main parachute ejection charge fired at a pre-programmed
altitude of 700 ft. The main parachute deployed successfully, bringing the entire vehicle to a soft landing at 25
ft/sec as shown in Fig. 11(c).

(a) (b)
Figure 9. Images of the sounding rocket: a) Altimeter Bay with RDAS units; b) Exploded view of Mongoose
98 Rocket. The entire length of the assembled rocket is about 8 feet 6 inches and the outside diameter is 4
inches.
Overall, launch and recovery of the flight-vehicle went as planned. The carbon-fiber aerostructures (fuselage
and fins) were able to withstand the propulsive and aerodynamic forces during the entire flight, and the aerodynamic
stability was good throughout the boost and coasting phases of the flight. The avionics systems worked as planned,
successfully ejecting charges and logging flight data. The drogue and main parachute recovery systems worked
adequately by deploying at pre-determined altitudes and landing the vehicle intact, approximately 1000 ft
downrange of the launch area. A plot of the R-DAS flight data (acceleration, velocity, and altitude) is presented in
Fig. 10.

11
American Institute of Aeronautics and Astronautics
 Figure 10. This figure shows the R-DAS flight-data from test launch of ALICE flight-vehicle powered by
K780R motor.
The test flight also provided valuable data for validation and calibration of the RockSim27 Model Rocket Design
& Simulation Software that is being used for performing simulations of the ALICE flight-vehicle. At the moment,
RockSim v.827 and PRO versions are used to predict the vehicle acceleration, velocity, altitude, angle-of-attack and
flight path for various flight-weight ALICE propulsion systems. Simulations are being run to estimate the landing
location assuming successful parachute deployment, as well as ballistic flight in case the parachutes fail to open.
Using Monte Carlo analysis, the maximum expected radius of impact is being calculated for variable conditions
(wind speed, variable thrust profiles, launch tower angles, etc) that are used for range safety planning purposes.
Flights are only attempted for wind speeds less than 10 mph.
Static and dynamic stability analysis is also being performed by using the RockSim code. The code requires the
input of mass and geometry parameters for each component (motor, propellant, external carbon-fiber aerostructure,
interstage couplers, etc). The motor thrust profile, tower launch angle, wind speeds, temperature and other launch
factors are inputted in the code as well. Extensive simulations are being performed to ensure vehicle stability at all
Mach numbers and at varying atmospheric conditions.
Based on the thrust profile from the hot-fire test performed with the 7” long ALICE grain, as well as the new
flight-weight motor design, the simulations predict that the 30 lb mass flight-vehicle will depart the launch rail in 0.9
seconds, achieving an exit velocity of 67 ft/sec. The simulations also predict a maximum acceleration of 16 G’s,
velocity of 187 mph (Mach 0.24), and a nominal altitude of 1,200 ft as shown in Fig. 12.

12
American Institute of Aeronautics and Astronautics
 (a) (b) (c)
Figure 11. Images above are from the flight test: a) Rocket on launch platform, b) Liftoff of rocket, c) Descent
of rocket under main parachute

Figure 12. Shown here is a trajectory simulation with thrust profile from 7” ALICE grain hot-fire tests (Test
6).

Following the successful launch of the sounding rocket with a K780 commercial motor, attention has shifted to
the design, test, and flight of an ALICE flight weight motor. The flight weight motor shown schematically on Fig.
13 is built out of a solid piece of 7075-T6 aluminum. This method is preferred over welding on flanges to the end of
the casing, which could potentially cause changes in the mechanical properties of the aluminum. Bolts are threaded
into steel threaded inserts located in the aluminum flange. These steel inserts help to distribute the load evenly over
the length of the thread. The bolts are the same ones used on the battleship motor, which were selected to fail at
3200 psi.

13
American Institute of Aeronautics and Astronautics
 Figure 13. Schematic of the flight-weight motor casing
A structural analysis of the flight-weight motor was performed. A solid mesh of 3512 elements and a 2000-psi
internal load was applied to the 3-D ProMechanica© model. The resulting analysis showed a failure index of 0.29
based on the tensile yield strength of aluminum 7075-T6 of 73,000 psi (Fig. 14). Upon completion of the casing, the
vessel was hydro-tested. Water supplied an internal pressure of 2000psi, which held for several minutes. Passing
this test, allowed for the first static test with the ALICE propellant.

Figure 14. VonMises stress calculation (top) and failure index along internal casing length (0”: left on top
view – nozzle end; 12”: right on top view – igniter end) based on 7075-T651 aluminum (bottom)

14
American Institute of Aeronautics and Astronautics
 The first configuration of the flight-weight motor was in the horizontal configuration, similar to that of the
battleship tests. A nozzle throat of 0.52” was selected to provide a chamber pressure of 1500-psi, based on
predictions. The first test used a grain length of 6.75”, due to limitations of the grain-casting tool. A modification
of the tool has been completed since the casting of the first grain. The experimental results obtained with the first
flight-weight grain are presented and compared with the modeling results in Fig. 15 and Table 5 below.

Figure 15. Comparison of 7” long ALICE motor test with lumped parameter models

As with the two 5” grains, the experimental results compare well with the predictions of the first variant of the
lumped parameter model. The test results show an average peak pressure around 1500-psi and a thrust of 500lbf.
The specific impulse for the 6.75" long grain (Test 6) is estimated based on the total impulse and the mass of
propellant expelled from the combustion chamber. The total impulse is obtained by integrating the measured thrust
over the estimated duration of the ALICE burn. For test 6, the start of the ALICE burn is set at time = 0.36 s while
the end of the burn is set at a time of 1.32 s. While somewhat arbitrary, both time points are selected based on events
occurring during the test as indicated by small thrust spikes observed on the experimental data. In turn, the mass of
propellant expelled for the combustion chamber is calculated based on the initial mass of the ALICE grain and the
mass of the slag leftover in the chamber following the test. Using this very conservative approach, the estimated
specific impulse of the 6.75" long ALICE grain is 160 s. While this value is much lower than predicted, the poor
combustion efficiency of the current propellant formulation greatly reduces the total impulse of the grain. Two-
phase flow losses and nozzle efficiency losses also contribute to the lower specific impulse value.
It is encouraging, however, that the model discussed in section V captures the experimental peak values of
pressure and thrust for both the 5” and 7” grains. Despite the simplifying assumptions in the model, capturing these
performance metrics provides an indication of an attainable ISP of 210. Improved propellant formulations with higher
equivalence ratio (closer to stoichiometric) and appropriate additives should increase the experimental specific
impulse to values well above 200s. This is the subject of on-going work at Purdue University and Penn-State
University.
Using the same simplifying assumptions for alumina deposition, characteristic velocity and specific impulse
losses, the second variant of the model reflects the progressive nature of the grain burning but over predicts the peak
chamber pressure.

15
American Institute of Aeronautics and Astronautics
 Table 5. Modeling summary for 6.75” long grain
Total Impulse [N-s] Peak Pc [psi] Peak Thrust [lbf]
Model assuming no losses for 6.75" long grain 2004 1670 548
Model with deposit & losses for 6.75" long grain 1805 2000 530
6.75" long grain experimental test results 1285 1650 550

While the original target altitude for the sounding rocket was set at 3000 ft, several constraints limit the
achievable altitude with the current ALICE powered rocket to approximately 1200 ft (as shown on Fig. 12). First,
the combustion and flow losses observed during the last six static test firings lead to total impulse values about 60%
that of the predicted values. These losses can be greatly reduced and are being addressed in on-going work with
improved ALICE propellant formulations including additives and alternative formulations to achieve higher specific
impulse and lower the alumina content of the products. Second, the flight-weight casing for the ALICE propellant
has to sustain pressures up to 2000 psi requiring thicker walls than that for commercial motors such as the Aerotech
K780. In addition, the energy required for igniting the current ALICE propellant formulation is significantly higher
than that required for a standard solid propellant. This leads to added weight for an igniter casing and an interface
with the ALICE casing capable of sustaining high pressures and designed in such a way the combustion gases do not
impact the aluminum walls. While designed for flight with safety factors around 1.5, the heavier casings reduce the
maximum altitude achievable with the rocket. Finally, the burning rate of the current ALICE formulation is on the
order of 1 inch per second at the nominal operating pressure of 1500 psi. Such high burning rate means that larger
grains are required to sustain the ALICE combustion over sufficiently long durations to adequately distribute the
thrust and therefore the rocket acceleration. In turn, larger grains require heavier casings. The current design is a
trade-off between the aforementioned constraints and has for sole purpose the demonstration of the ALICE
propellant in a flight capable environment. Further improvements of the propellant formulation should address these
constraints.

Recommended Future Work & Conclusion

We have shown that refrigerated solid propellants can be used for rocket motors and the ALICE propellant has
shown promise as a successful rocket propellant in static test firings. Six small-scale static experiments have shown
consistent results when compared to the prediction codes. Although this current propellant formulation is far from
optimized, improvements in the mixing procedure have produced a consistent and homogeneous propellant. An
internal ballistic model developed to support the experiments provides a simplified account of a complex series of
events within the igniter and the main combustion chamber. The model is based on measured burning rate
parameters and exact grain geometries tested at the Purdue Propulsion Laboratories. Perturbations to the model can
be introduced to reflect the reduction of the nozzle throat diameter due to alumina deposition and to take into
account losses in the combustion chamber and the nozzle.
While the model overpredicts the total impulse of the ALICE propellant grains, it is a very useful tool for peak
chamber pressure and thrust predictions and, based on consistency between model and experiment over several tests,
it is also a prediction tool for flightweight motor performance and, therefore, rocket trajectory predictions.
The consistent ratio between the prediction and actual measurements indicate a further understanding of the
propellant. A successful launch of the carbon fiber rocket using a commercially available motor and a successful
flight-weight casing static test furthers the goal of this project to a test flight with ALICE propellant.
Future work will include a test of the flight-weight motor in the vertical position. Once this test is complete, a
launch with ALICE propellant in the sounding rocket will be the next step. In addition to these launches, the
alumina slag left should be analyzed to determine the amount of initial aluminum that did not react. This will help to
explain some of the differences between the predication and the test values, as well as yield a better understanding
into the kinetics of the reaction, and the modeling of nAl particles. Another direction for the ALICE propellant is to
begin testing with a stoichiometric mixture, or using fuel additives. Previous work by other researchers has
indicated that the ISP of ALICE propellant can increase with the addition of hydrogen peroxide or other ingredients
such as alane in place of aluminum powders. The use of fuel additives is an attractive idea, as it could modify the
burning rate, decrease alumina produced, and improve performance.

16
American Institute of Aeronautics and Astronautics
 Acknowledgments
The authors would like to thank Dr. Mitat Birkan of the Air Force Office of Scientific Research and NASA
under contract numbers FA9550-09-1-0073 and FA9550-07-1-0582. The authors would also like to thank Mr. Cody
Dezelan for his research contributions.

References
1
Kuo, K. K., Risha, G. A., Evans, B. J., Boyer, E., "Potential Usage of Energetic Nano-sized Powders for Combustion and
Rocket Propulsion," Proceedings of the Materials Research Society Symposium, Vol. 800, 2004, pp. 3-14.
2
Dokhan, A., Price, E. W., Seitzman, J. M., Sigman, R. K., "The Effects of Bimodal Aluminum with Ultrafine Aluminum on
the Burnign Rates of Solid Propellants," Proceedings of the Combustion Insititute, Vol. 29, 2002, pp. 2939-2945.
3
Shalom, A., Aped, H., Kivity, M., Horowitz, D., "The Effect of Nanosized Aluminum on Composite Propellant Properties,"
41st Annual AIAA JPC Conference , Vol. 41, 2005, pp.
4
Galfetti, L., Severini, F., De Luca, L. T., Meda, L., "Nano-Propellants for Space Propulsion," Proceedings of the 4th
International Spacecraft Propulsion Conference, Vol. 4, 2004 pp.
5
Trunov, M. A., Schoenitz, M., Dreizin, E. L., "Ignition of Aluminum Powders Under Different Experimental Conditions,"
Propellants, Explosives, Pyrotechnics, Vol. 30, No. 1, 2005 pp. 36-43.
6
Risha, G. A., Huang, Y., Yetter, R. A., Yang, V., Son, S. F., Tappan, B. C., "Combustion of Aluminum Particles with Steam
and Liquid Water," 44th AIAA Aerospace Sciences Meeting and Exhibit, Vol. 44, 2006, pp.
7
Trunov, V. G., Safronov, M. N., Gavrilyuk, O. V., "Macrokinetics of Oxidation of Ultradisperse Aluminum by Water in the
Liquid Phase," Combustion, Explosion, and Shock Waves, Vol. 37, No. 2, 2001 pp. 173-177.
8
Ingenito, A., Bruno, C., "Using Aluminum for Space Propulsion”, Journal of Propulsion and Power, Vol. 20, No. 6, 2004
pp. 1056-1063.
9
Shafirovich, E., Bocanegra, P. E., Chauveau, C., Gokalp, I., "Nanoaluminum – Water Slurry: A Novel “Green” Propellant
for Space Applications," Proceedings of the 2nd International Conference on Green Propellants for Space Propulsion, Vol. 2,
2004 pp.
10
Rasor, O., Portland, OR, U.S. Patent Application for a “Power Plant,” filed 9 Oct. 1939.
11
Elgert, O. J., Brown, A. W., "In Pile Molten Metal-Water Reaction Experiment”, U.S. Atomic Energy Publication IDO
16257, 1956 pp.
12
Leibowitz, L., Mishler, L. W., "A Study of Aluminum-Water Reactions by Laser Heating”, Journal of Nuclear Materials,
Vol. 23, 1967 pp. 173-182.
13
Miller, T. F., Herr, J. D., "Green Rocket Propulsion by Reaction of Al and Mg Powders and Water," 40th Annual AIAA JPC
Conference, Vol. 40, 2004, pp.
14
Foote, J. P., Lineberry, J. T., Thompson, B. R., Winkelman, B. C., "Investigation of aluminum particle combustion for
underwater propulsion applications," 32nd AIAA JPC Conference, Vol. 32, 2006, pp.
15
Humble, R. W., Henry, G. N., Larson, W. J., Space Propulsion Analysis and Design, 1st ed., McGraw-Hill, New York,
1995, Chaps. 6.
16
Ivanov, V. G., Gavrilyuk, O. V., Glazkov, O. V., Safronov, M. N., "Specific Features of the Reaction between Ultrafine
Aluminum and Water in a Combustion Regime," Combustion, Explosion, and Shock Waves, Vol. 36, No. 2, 2000 pp. 213-219.
17
Risha, G. A., Son, S. F., Yetter, R. A., Yang, V., Tappan, B. C., "Combustion of nano-aluminum and liquid water,"
Proceedings of the Combustion Insititute, Vol. 31, 2007, pp. 2029-2036.
18
Sippel, T. R., Son, S. F., Risha, G. A., Yetter, R. A., “Combustion and Characterization of Nanoscael Aluminum and Ice
Propellants," 44th Annual AIAA JPC Conference, Vol. 44, 2008, pp.
19
Dokhan, A. “ The effects of Aluminized Particle Size on Aluminized Propellant Combustion,” Ph.D. Thesis, Aeronautics
and Astronautics Dept., Georgia Institute of Technology, Atlanta, GA, 2002.
20
Kwon, Y. S., Gromov, A.A., Strokova, J. I., “Passivation of the surface of aluminum nanopowders by protective coatings of
the different chemical origins," Applied Surface Science, Vol. 253, 2007 pp. 5558-5564.
21
Cliff, M., Tepper, F. & Lisetsky, V., "Ageing characteristics of alex nanosize aluminum," 37th Annual AIAA JPC
Conference Vol. 37, 2001, pp.
22
Franson, C., Orlandi, O., Perut, C., Fouin, G., Chauveau, C., Gokalp, I., Calabro, M., "Al/H20 and Al/H20/H202 frozen
mixtures as examples of new composite propellants for space application,"
23
Risha, G. A., Connell, T. L., Yetter, R. A., Yang, V., Wood, T. D., Pfeil, M. A., Pourpoint, T. L., Son, S. F., “Aluminum-
Ice (ALICE) Propellants for Hydrogen Generation and Propulsion," 45th Annual AIAA JPC Conference, Vol. 44, 2009, pp.
24
LabRam ResonantAcoustic Mixer Manual, ResoDyn, Butte, MT, www.resodynmixer.com.
25
Yetter, R. A., Risha, G.A., Connell, T, Yang, V., Son, S. F., Sippel, T. S., Pourpoint, T. L., "Novel Energetic Materials for
Space Propulsion," Presentation, AFOSR/NASA Office of Chief Engineer Joint Contractors / strategic Planning Meeting in
Chemical Propulsion, Vienna VA, 2008.
26
ThrustCurve, John Coker, www.thrustcurve.org, Online – July 2009.
27
RockSim, Ver. 8.0, Apogee Components, Inc., Colorado Springs, CO, 2009.

17
American Institute of Aeronautics and Astronautics