48th AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit AIAA 2012-4201

30 July - 01 August 2012, Atlanta, Georgia

Nitrous Oxide Cooled, Reusable Hybrid Aerospike Rocket

Motor: Experimental Results
Patrick Lemieux1 and William R. Murray2
California Polytechnic State University, San Luis Obispo, 93047

We developed an innovative, simple, low-cost, reusable, oxidizer-cooled aerospike nozzle. Our innovation is
to use the refrigerant capabilities of nitrous oxide (N2O) to provide the cooling required for reusable
operation of an aerospike nozzle in conjunction with a N2O-PMMA hybrid rocket motor. Cooling is provided
by the latent heat of vaporization of liquid N2O as it flows in the hottest sections of the nozzle, maintaining
nozzle structural integrity and limiting nozzle erosion to acceptable levels, thereby enabling reusable
Downloaded by "Indian Institute of Technology, Bombay" on March 31, 2025 | http://arc.aiaa.org | DOI: 10.2514/6.2012-4201

operation, longer burn times, and improved performance. The N2O used for cooling here is subsequently
ejected into the exhaust plume and provides base bleed, reducing the thrust performance penalties that would
normally be associated with a severely truncated aerospike. We present some of the experimental results of a
test series that we conducted to verify this idea.

Nomenclature
BNi-2 = nickel/chromium-based high temperature brazing alloy
CNC = Computer Numerically Control
O/F = oxidizer to fuel ratio
HTPB = hydroxyl-terminated polybutadiene
N2O = nitrous oxide
NPR = nozzle pressure ratio; combustion chamber total pressure divided by ambient pressure
PMMA = Polymethyl methacrylate
T* = Critical temperature
γ = ratio of specific heats

I. Introduction

A. Overview

T he Mechanical Engineering Department at the California Polytechnic State University, San Luis Obispo (Cal
Poly), is developing a laboratory for the study of hybrid rocket motors, with an emphasis on annular aerospike
nozzles. One disadvantage of these nozzles, however, is that the main spike is fully engulfed in the hot gas flue of
the rocket exhaust, and is thus susceptible to rapid ablation, weakening and damage, making it unusable after a
single flight and even restricting the length of time it may operate during that flight. Also, the very large temperature
gradient to which an uncooled center spike is subjected is likely to cause geometric variations that will, at the very
least, affect the predictability of its performance.i This paper describes how nitrous oxide (the primary oxidizer used
in our facility), a substance that may also be used as a coolant for the purpose of mitigating the above problemii, has
been implemented in the integral design of a cooled aerospike and tested in our facility.

B. Technical Objectives
Because of its high vapor pressure, N2O can be self-pumping, thereby making it an ideal choice of oxidizer for
simple, low-cost applications such as this one. Even when used in conjunction with a higher pressure inert gas in a
pressure-fed design, the use of N2O still leads to simpler, less expensive designs than those requiring turbo-pumps.
As a practical nozzle, our innovation fits well with N2O-HTPB hybrid rocket designs, which tend toward simpler,
less expensive design alternatives. We tested it here on N2O-PMMA rocket motors because of the ease of
manufacture of PMMA compared to HTPB, although the concept applies independently of the fuel used.
Our innovation is original: despite the widespread use of N2O as an oxidizer for hybrid rocket motors, no one
appears to have taken advantage of the potential for N2O to be used as a coolant for rocket nozzles as it is flashed
from a liquid into a vapor. Compressed liquid fuel cooled rocket motors also typically do not use phase change for

1
Bently Endowed Professor, Mechanical Engineering, M/S 13-254, AIAA Senior Member.
2
Director, Donald E. Bently Center for Engineering Innovation, Mechanical Engineering, M/S 13-254.
1
American Institute of Aeronautics and Astronautics

Copyright © 2012 by Patrick Lemieux. Published by the American Institute of Aeronautics and Astronautics, Inc., with permission.
 cooling.iii In fact, following the demonstration of the concept in the present study, the principle is now pending
patent approval. iv
The main technical objective of this stage in the project is to demonstrate the feasibility of using N2O oxidizer
for cooling an annular aerospike nozzle, and to verify that such a nozzle is reusable. This work built on the results
that we presented recently on the advantages of the hybrid aerospike concept and on the detailed design and analysis
of a model aerospike nozzle for this purpose.v In summary, the goals of our study consist of:

1. Aerospike nozzle design: Starting from a baseline aerospike profile which we developed, include coolant
passages terminating at the aft end of the aerospike, truncated to allow for release of the coolant in the
exhaust plume.
2. Heat load assessment of nozzle: The throat region of rocket nozzles experiences the highest heat transfer
rate of any part of the motor. The resulting high heat load is of particular concern in the case of the
aerospike design, where the center spike itself is fully engulfed in the hot gas flue and thus has no means to
Downloaded by "Indian Institute of Technology, Bombay" on March 31, 2025 | http://arc.aiaa.org | DOI: 10.2514/6.2012-4201

mitigate the heating action of these gases. Develop models to predict the heat transfer rates along the
profile of the aerospike.
3. Material selection: Determine and select the best material (or materials) to build the nozzle from the
standpoint of ease of manufacturing, dimensional stability throughout operation, chemical compatibility
with the propellant and the oxidizer coolant, and structural requirements, as well as their ability to provide
suitable heat sink. Conduct structural analyses to estimate strain ranges for the desired fatigue life.
4. Optimum cooling strategy: Determine the cooling method that provides the most effective way to maintain
the throat temperature within design guidelines, while having the least impact on the overall motor
performance. For candidate cooling channel configurations, develop FEA models of the material assembly
of the nozzle to determine the stress levels in the assembly, and mitigate for the harshest conditions that the
nozzle is expected to experience.
5. Detailed nozzle design: Conduct the full design analysis of the nozzle configuration, and investigate the
nozzle design survivability through start-up and shut-down transients. This detailed design work includes a
finite element model of appropriate detail for the cooled chamber nozzle assembly, coupled with a CFD
model of the flow in the nozzle and analyses of the coolant flow.
6. Prototype build of nozzle: Manufacture the nozzle and cooling hardware for the nozzle, suitable for testing
on the Cal Poly test stand.
7. Nozzle testing: Perform a series of tests to verify the performance of the nozzle, and quantify the
effectiveness of the cooling strategy and materials selected in maintaining reusability of the nozzle over
multiple firing runs.

The present paper focuses and reports on the final two goals.

II. Aerospike Design: a Summary

A. Design Requirements
The cooled aerospike rocket nozzle designed at Cal Poly was built specifically for laboratory testing on the
labscale N2O-PMMA hybrid rocket already in place at Cal Poly. As such, many parameters in the design are driven
by the combustion conditions of the laboratory rocket and the facility itself. The rocket operating conditions,
including chamber pressure, oxidizer/fuel (O/F) ratio, and ratio of specific heats, γ, determine the spike contour and
heat loads on the aerospike. The aerospike nozzle design targets withstanding heat loading from a burn at
stoichiometric O/F ratio, at our baseline pressure condition (NPR=20). While the aerospike nozzle may not be
initially tested at these conditions, this design should withstand the heat and pressure loads that correspond with the
stoichiometric combustion condition, the highest loads that it might see in testing. The stoichiometric O/F ratio of
N2O- PMMA is 5.3. Our design chamber pressure is 294 psia (corresponding to a nozzle pressure ratio of 20 at sea
level). Using the NASA-Glenn Chemical Equilibrium Program (CEA), we determined that the products of
combustion have a ratio of specific heats of γ=1.14, and this formed the basis of our evaluation of the peak heat load
that must be sustained by the aerospike during any of our test runs.

B. Contour Design
The contour of an aerospike nozzle outside of the throat is determined based on the Prandtl-Meyer expansion fan
of the exhaust gases flowing through the nozzle. The contour of a minimum length aerospike can be calculated for a
2
American Institute of Aeronautics and Astronautics
 specific nozzle pressure ratio and ratio of specific heats at a given operating atmospheric pressure. The contour
developed for our aerospike came about as part of an MS thesis project at Cal Poly.vi Shown in Figure 1 is the
contour of the aerospike based on NPR=20, γ=1.14, and Patm=14.7 psia. We also spent significant efforts
streamlining the gas flow upstream of the aerospike, developing the post-chamber configuration shown in Figure 2.
Downloaded by "Indian Institute of Technology, Bombay" on March 31, 2025 | http://arc.aiaa.org | DOI: 10.2514/6.2012-4201

Figure 1. The aerospike contour calculated for NPR=20 and γ=1.14 at Patm=14.696 psia. The blue line represents the contour of
the spike, while the red line shows the position of the nozzle throat.

Figure 2. The flow geometry for the aerospike. The green region represents the net shape of the aerospike assembly. The yellow
regions represent the end of inner and outer fuel grains for an annular combustion chamber. Although the physical, cooled
aerospike is truncated, a non-truncated aerospike is shown in the above 2 figures for simplicity.

3
American Institute of Aeronautics and Astronautics
 C. Design Heat Loads
Based on the geometry of the aerospike and the design combustion conditions described above, we modeled the
flow of hot gas over the aerospike using FLUENT, a commercial computational fluid mechanics (CFD) program, in
order to quantify the heat load that must be absorbed by the coolant. This CFD model was used to predict the heat
transfer coefficients on the surface of the aerospike, assuming a constant spike surface temperature of 400°F. This
assumption, used to determine the mass flow rate of coolant required, is inherently conservative since the aerospike
is expected to reach a higher temperature at steady state, consequently reducing the heat flux to the coolant. Show in
Figure 3 is the heat load per unit length that will be transferred into the aerospike based on this analysis. This heat
load is calculated from the heat transfer coefficient and the exposed surface area at each axial position along the
length of the aerospike, assuming a constant temperature difference between the exhaust gas stagnation temperature
of T0=5267°F and the aerospike surface temperature of Ts=400°F. We integrated this heat load along the length of
the aerospike and determined that the total heat transfer rate to the aerospike assembly at design conditions is 26.7
BTU/s. These conditions are used to design the aerospike cooling passages and required coolant flow rate.
Downloaded by "Indian Institute of Technology, Bombay" on March 31, 2025 | http://arc.aiaa.org | DOI: 10.2514/6.2012-4201

Figure 3. The heat load being transferred into the aerospike per unit length, calculated based on a constant surface temperature of
400°F. This represents the rate of heat transferred to the coolant, at each aerospike cross-section.

4
American Institute of Aeronautics and Astronautics
 D. Final Design
The final design of the cooled aerospike nozzle is shown in Figure 4 and Figure 5. The manufacturing of the
aerospike nozzle and the resulting as-built hardware are presented in the following section.
Downloaded by "Indian Institute of Technology, Bombay" on March 31, 2025 | http://arc.aiaa.org | DOI: 10.2514/6.2012-4201

Figure 4. Upper left: The final concept of the aerospike assembly, consisting of a stainless steel core covered with a copper shell
and protected by a graphite insulator upstream. Upper right: The stainless steel core has 20 grooves cut in it with a 5/64" ball end
mill. Those grooves act as coolant channels. There are two different lengths of grooves: 10 long grooves and 10 short grooves.
Center: Two of the 10 long coolant channels can be seen along the edges of the aerospike assembly, just beneath the copper shell.
All coolant channels terminate in a common exit tube, which is sized to obtain the correct coolant mass flow rate through the
aerospike. Bottom: Two of the 10 short coolant channels can be seen along the edges of the aerospike assembly, just beneath the
copper shell. The short coolant channels are used to cool primarily the throat region, where the heat load is highest. Also shown
is the thermocouple that provides spike temperature during hot-flow testing. This thermocouple is routed up the main spike
support tube through which coolant will flow and mechanically staked in place in a slot on the stainless steel nozzle core, prior to
furnace brazing of the spike assembly.

5
American Institute of Aeronautics and Astronautics
Downloaded by "Indian Institute of Technology, Bombay" on March 31, 2025 | http://arc.aiaa.org | DOI: 10.2514/6.2012-4201

Figure 5. Left: Vertical section of the hybrid rocket motor showing coolant feed tube, the dual annular PMMA fuel grains, the
aerospike nozzle, and the nozzle body. Right: View of the top of the nozzle body. Visible are the copper throat plate, the nozzle
cap, the aerospike nozzle, and five thermocouples embedded in the copper throat plate. The innermost of these thermocouples is
0.25" from the nozzle throat, and each of the remaining thermocouples is 0.50" farther out in the radial direction than the
previous thermocouple. Note that the aerospike is hollow at its truncated face to allow for the exhaust of spent coolant.

III. Nozzle Manufacturing

The aerospike core is made of stainless steel. As seen in Figure 4, a copper shell, with a wall thickness of
0.040"3, covers the core from a point forward of the throat all the way to the tip, which is truncated. At the base of
the aerospike, the stainless steel core is covered with an insulator that protects the stainless steel from being exposed
to the hot flow and which retains the center fuel grain. This insulator is made of graphite. The N2O coolant is fed
into the stainless steel core from the base and run up the center of the aerospike through a passage. The N2O coolant
then is routed to the copper shell and flows between the copper shell and stainless steel core in several longitudinal
passages, as seen in Figure 4. A total of 20 coolant channels are specified: 10 long channels that run the length of
the spike, and 10 short channels that run along the 60% of the spike nearest the throat where the heat load is the
highest.
The principal processes in manufacturing our nozzles are machining of the intricate coolant channels in the
stainless steel core of the aerospike, furnace brazing of the copper shell to the stainless steel core, and careful
assembly of the nozzle into the rocket motor.

A. Furnace Brazing
For the purpose of furnace brazing the aerospike nozzle components in this project, we modified a 3kW electric
furnace to heat a retort made to provide an inert gas environment, as shown in Figure 6. After considerable
experimentation and process development, we chose 0.0015"-thick BNi-2 foil as the filler metal, 1910°F as the
brazing temperature, and ultra-high-purity (UHP) argon for the inert gas environment. We conducted two sets of
coupon tests to verify that the strength of the resulting brazed joints was higher than that of the copper parent
material used in the cooled aerospike nozzle.

3
We added a draft angle to the stainless steel core to aid in assembly, so the wall thickness of the copper shell was
0.040" at the aft end near the throat, but only 0.030" at the forward (upstream) end of the copper shell, varying
linearly.
6
American Institute of Aeronautics and Astronautics
Downloaded by "Indian Institute of Technology, Bombay" on March 31, 2025 | http://arc.aiaa.org | DOI: 10.2514/6.2012-4201

Figure 6. Left: Exploded view of furnace with retort (from left to right: retort plug, high-temperature gasket to provide pressure
seal, retort in redesigned clamp-on door, and furnace). Right: modified furnace for oven brazing cooled aerospike nozzles, a flow
of argon is necessary through the spike coolant channels as well as into the retort during the brazing cycle.

B. Nozzle Construction
Construction of these cooled aerospike nozzles is an involved process, requiring many steps. The blanks for the
stainless steel nozzle core must first be machined. Then the grooves that will form the cooling channels must be cut
into these blanks. The coolant paths are of complex shapes, and a five-axis CNC milling machine is necessary to cut
these groves. The aerospike nozzle shape must be machined into the copper rod that will be brazed onto the
stainless steel nozzle core. At braze time the thermocouple that will provide spike temperature during hot-fire
testing must be inserted into the spike and brazed along with the copper and stainless steel parts. Once that
assembly has been brazed, the assembly is turned in a CNC lathe to give the copper shell the specified shape and
wall thickness.
At this point, the brazed spike assemble is ready to mount to the spike support tube that passes through the center
fuel grain. Finally both fuel grains are loaded into the combustion chamber, which is connected to the base of the
rocket motor. This whole assemble is then chucked up on a lathe and centered to check the runout of the spike.
Adjustments are made if necessary. These steps are shown in Figure 7.
An example of the final hardware is shown in Figure 8.

7
American Institute of Aeronautics and Astronautics
Downloaded by "Indian Institute of Technology, Bombay" on March 31, 2025 | http://arc.aiaa.org | DOI: 10.2514/6.2012-4201

Figure 7. Upper left: coolant grooves being machined in the stainless steel spike core. Upper right: spike core, copper piece,
braze filler metal foil and thermocouple for a cooled spike. Center left: brazed spike assembly. Center right: finished aerospike,
the snout is used in proof testing and then parted off. Lower left: brazed spike along with the graphite insulator and end cap.
Lower right: entire rocket motor except for the nozzle body mounted in a lathe to check the runout on the aerospike.

8
American Institute of Aeronautics and Astronautics
Downloaded by "Indian Institute of Technology, Bombay" on March 31, 2025 | http://arc.aiaa.org | DOI: 10.2514/6.2012-4201

Figure 8. Cooled aerospike nozzle. Upper left: assembled cooled aerospike with the graphite insulator and the end cap. Upper
right: second view of assembled cooled aerospike. Note the thermocouple sheath extending from the nozzle. Lower left: cooled
nozzle on the rocket motor ready to be fired; and Lower right: close-up of cooled nozzle on the rocket motor ready to be fired.

IV. Experimental Testing

We fabricated three aerospike nozzles for our test stand to carry out the experimental part of this study:

1. A mass flow rate and coolant pressure tests aerospike nozzle. Its purpose is to provide a measurable estimate
of the mass flow rate achievable through the nozzle coolant passages geometry (i.e., accounts for all friction
losses including exhaust discharge coefficient). This nozzle is never hot-fired, and is referred here as the ‘cold-
flow aerospike’. Its flow rate verifications do not account for Rayleigh effect present during the hot runs;
2. An uncooled aerospike nozzle. Its purpose is to provide baseline heat transfer data that the aerospike
experiences without cooling (i.e., a ‘control aerospike’); and
3. A N2O-cooled aerospike nozzle. This is the ‘test aerospike’, benchmarked against Nozzle 2 above.

We test fired the uncooled aerospike nozzle once, the cooled nozzle twice, and the cold-flow aerospike seven
times, at different flow orifice configuration (both upstream and downstream) to tailor the specific flow
requirements of various hot test runs considered. We present results of some of these tests in the following sections.
All aerospike nozzles use the same nozzle body and hybrid rocket motor.

A. Coolant Flow Rate Analysis and Testing

We carried out a series of ‘cold flow’ runs (i.e., unfired test runs) to calibrate the coolant flow parameters that
correspond to our needs in an actual run. For a maximum design heat load to the aerospike of 26.7 BTU/sec, we
estimated that approximately 0.8 lbm/sec of cooling N2O flow is required so that the coolant exits the aerospike with
9
American Institute of Aeronautics and Astronautics
 a quality not exceeding 60%, a safety milestone that we arbitrarily chose to minimize the risk of runaway thermal
decomposition of N2O inside the cooling passages during hot fired tests. This is a flow rate that is significantly
higher than any previous experiments that we had conducted, and it was unclear whether our laboratory setup could
accommodate it. Releasing that much spent coolant into the hot plume of a running engine also represented a new
safety concern. Prior to conducting any hot-fire test, it was therefore imperative to verify that the cooling flow setup
that we developed could achieve this rate, and that we properly characterized the parameters that achieve this.
Without relying on analytical or numerical models, this requires a series of cold flow test experiments specifically
for this purpose.

These cold flow experiments build on our previous experience with nitrous oxide cooling the annulus of a
conventional converging-diverging nozzle, where we learned to expect a relatively negligible reduction in mass flow
rate between the cold flow and the hot-fire flow runs at the same coolant side condition, thus making non-fired
calibration tailoring runs a safe and valid process. In this case, we used four nitrous oxide bottles mounted onto a
Downloaded by "Indian Institute of Technology, Bombay" on March 31, 2025 | http://arc.aiaa.org | DOI: 10.2514/6.2012-4201

parallel manifold to help prevent choking at the bottle valves. There are no pressure regulators fitted to the bottles; a
single ball valve, downstream of the manifold, isolates them from the main cooling line. When the valve is opened,
the line is effectively charged at the bottle pressure. The smallest area in the system is expected to choke the flow, if
it is smaller than the exit hole at the end of the aerospike.

The cold-flow measurements consist of:

1. Mass flow rate. Our orifice flow meter is not designed for two-phase flow; we used a digital scale under the
bottle rack to measure the weight of the four bottles in real time to quantify flow rate.
2. Bottle pressure and temperature. Each bottle contains saturated N2O. Thus, the steady-state temperature of
the bottle determines their initial pressure. An analog gage is mounted on the manifold to verify the saturation
pressure at the beginning of the experiment (Pbot). During the experiment, the bottle liquid level drops as the fluid
flows out, and this causes the remaining liquid to boil off in the bottle, cooling it and dropping the bottle pressure
somewhat. The bottle state is monitored electronically during the test, at a location approximately six feet
downstream of the manifold (Pbot and Tbot). This station includes an orifice flow meter which, although inaccurate
for direct flow measurements, allows us to ‘throttle’ the flow at that location by varying the size of its orifice plate,
as well as measure the pressure just upstream of the plate.
3. Feed line pressure and temperature. The line pressure is monitored at two additional stations (P0 and P1),
and the temperature is monitored at one station (T1), along the feed line before entering the aerospike support stem.
The feed line size is constant from the manifold through the P0 measuring station, downstream of which it drops in
size.
4. Aerospike pressure. Pressure is measured at three points inside the aerospike in the cold flow test: at the top
of the stem just upstream of the aerospike (P2), at the end of the short cooling passages (P3), and at the end of the
long cooling passages (P4). At the aerospike nozzle outlet, the pressure is atmospheric during tests.
The first 3 measurement stations remain in use during the hot flow experiments; the three aerospike pressures
(P2, P3, P4) are available on the cold-flow nozzle only, however, and cannot be monitored during hot-flow runs.

We carried out a series of 7 cold test runs using this setup. The parameters that we vary between runs are the
orifice size at the flow meter station (‘upstream’ flow restriction; station ‘bot’ in the above description), and the
aerospike exit hole diameter (atmospheric station 5 in the above discussion). We can replace the orifice plate
relatively easily to change the upstream conditions. The aerospike outlet diameter, however, requires drilling
progressively larger exit holes in the aerospike. The cold-flow runs are summarized in Table 1. The first run used
the nominal orifice plate of the flow meter (11/64" diameter), and a small 1/10" exit hole for the aerospike. This
provides an inadequate flow rate through the system, and showed that the flow meter causes a large pressure drop in
the feed line. The flow meter was then completely removed for the following three tests (which prevented the direct
measurement of bottle pressure and temperature for those runs. Previous experiments showed however that these
values remain relatively constant from run to run) to minimize pressure drop upstream of the aerospike, and leaving
the aerospike nozzle outlet diameter as the only variable. An aerospike exit hole size of 0.180" gives us a close
estimate of the maximum outlet flow size required before the flow chokes at some other point upstream, as seen by
comparing the results of Run 3 and Run 4. Fortunately, the mass flow rate there is close to our design point of 0.8
lbm/sec so that either Run 3 or Run 4 represents a good design set point for the hot flow experiments.

10
American Institute of Aeronautics and Astronautics
 Table 1: Summary of cold-flow test results.

Stand Inlet Spike Inlet Supply Spike Inlet Cooling

Aerospike Quality
Orifice Plate Number Supply Coolant (Station 1) (Station 2) Bottle (Station 2) Capacity
COLD Outlet
Diameter of Supply Pressure Flow Heat of
FLOW Diameter Liquid Gas
(Station 'bot') Bottles (Station 'bot') Rate Pressure Temperature Pressure Enthalpy vaporiztion
TEST (Station 5) Enthalpy Enthalpy
remaining
in in psia lbm/s psia °F psia BTU/lbm BTU/lbm BTU/lbm % BTU/lbm
1 0.100 0.172 4 683.489 0.273 679.1 57.4 637.1 51.78 48.42 131.13 4.07 79.35
2 0.154 0.400 4 N/A 0.626 679.0 57.9 607.5 N/A N/A N/A N/A N/A
3* 0.180 0.400 4 N/A 0.744 649.8 54.2 553.8 N/A N/A N/A N/A N/A
4 0.196 0.400 4 N/A 0.742 635.6 52.7 533.1 N/A N/A N/A N/A N/A
5** 0.154 0.100 4 689.938 0.380 458.4 29.2 423.5 52.25 31.90 134.76 19.78 82.52
6 0.196 0.100 4 696.593 0.381 356.6 12.3 290.4 52.73 19.58 135.45 28.61 82.72
7*** 0.196 0.125 4 696.490 0.502 462.5 29.9 385.3 52.72 28.60 135.09 22.65 82.37
* Design Condition
** Cooled spike test 1 conditions
*** Cooled spike test 2 conditions
Downloaded by "Indian Institute of Technology, Bombay" on March 31, 2025 | http://arc.aiaa.org | DOI: 10.2514/6.2012-4201

Also summarized in Table 1 are our calculations of flow enthalpy and quality entering the spike cooling
passages. These calculations are based on the assumption that the flow is adiabatic from the bottle to the spike inlet,
and that kinetic energy changes are negligible throughout the system. This allows us to quantify the maximum
cooling capacity of the fluid, before it becomes superheated.

While Cold Flow Test Run 3 and 4 provide parameters where the spike cooling power is maximized (defined by
cooling capacity multiplied by flow rate), we became concerned with the possible safety effect of releasing so much
oxidizer in the combustion plume of the test cell during hot test run. We developed the conditions for Cold Flow
Test Run 5, at approximately half of our target coolant flow rate, so that in a 5-second hot test run we would not
release more N2O into the room than during our previous conventional nozzle test runs. Cold Flow Test Run 5
became the coolant flow set point for our first hot test run. Cold Flow Test Run 6 showed that the flow most likely
choked at the orifice plate when its diameter is 0.100"; Cold Flow Test Run 7 provided us with a suitable condition
for a more aggressive combustion experiment, once we observed that the effect of releasing N2O in the hot flue did
not have safety consequences significantly different from our previous experiments.

B. Uncooled Aerospike Nozzle

The physical uncooled (solid) aerospike nozzle is shown in Figure 9, mounted on the hybrid rocket motor in our
test facility. An axial section of the uncooled aerospike is shown in Figure 10. Details of the parts that make up the
uncooled aerospike, the nozzle body into which the spike fits, and the hybrid rocket motor onto which the nozzle
mounts are shown in Figure 11.
We conducted one hot-fire test of this rocket motor, using the new fuel rich test condition. The purpose of the
uncooled test was to verify the baseline surface temperature against which the cooled nozzle may be compared. The
uncooled aerospike core has the same internal passage as the cooled spike, but the passage is plugged and no coolant
flows. This passage does provide a route for a thermocouple used for measuring spike temperature.
A thermal finite element modeling and analysis performed on this nozzle predicted a life of 1.3 seconds for the
uncooled aerospike subjected to stoichiometric burn temperatures. Of course, this analysis, which is based on a
constant heat flux consistent with an isothermal nozzle surface, is notably conservative, as is the fuel-rich test
condition. Still, it illustrates that the timescale of concern is on the order of seconds, not shorter, so that we could
monitor the aerospike temperature in real time, and can respond to an event requiring test abort.
For the hot-fire test of the uncooled aerospike nozzle, plots of various test parameters are shown Figure 11. In
this 8.2 second burn, the aerospike temperature climbed to over 1100°F, at which point we manually terminated the
burn. Due to the time required for shut down, the spike temperature went on to reach a maximum of 1251°F. Due
to the non-aggressive level of the test, the relatively short duration of the burn, and the large amount of thermal
capacitance in the nozzle throat plate, the radial thermocouple located closest to the throat (TT1, 0.25" from the edge
of the throat) recorded a maximum temperature of less than 275°F. Finally, this test shows that an uncooled
aerospike core temperature rises by over 1200°F in less than 10 seconds, becoming our benchmark for the cooled
tests that followed.

11
American Institute of Aeronautics and Astronautics
Downloaded by "Indian Institute of Technology, Bombay" on March 31, 2025 | http://arc.aiaa.org | DOI: 10.2514/6.2012-4201

Figure 9. Uncooled aerospike nozzle. Upper left: copper shell, stainless steel shell, and thermocouple. Upper right: assembled
uncooled aerospike with thermocouple, the graphite insulator, and the end cap. Lower left: uncooled spike after brazing and
final assembly. Lower right: uncooled aerospike mounted on the center fuel grain.

Figure 10. The uncooled aerospike core has the same internal passage as the cooled spike, but the passage is plugged and no
coolant flows. This passage does provide a route for the thermocouple for measuring spike temperature.

12
American Institute of Aeronautics and Astronautics
Downloaded by "Indian Institute of Technology, Bombay" on March 31, 2025 | http://arc.aiaa.org | DOI: 10.2514/6.2012-4201

Figure 11. Upper left: uncooled aerospike mounted in combustion chamber. Upper right: lower portion of the nozzle body, the
copper nozzle throat ring, and the nozzle cap. Lower left: close-up of the five thermocouples embedded in the nozzle throat ring.
Lower right: uncooled aerospike nozzle fully assembled.

13
American Institute of Aeronautics and Astronautics
Downloaded by "Indian Institute of Technology, Bombay" on March 31, 2025 | http://arc.aiaa.org | DOI: 10.2514/6.2012-4201

Figure 11. Hot-fire test results for the uncooled aerospike nozzle [Run #89]. TT1 through TT5 are the temperatures recorded by
the thermocouples embedded in the nozzle throat plate.

C. Cooled Aerospike Nozzle

The rocket motor setup is modular, so that many of its parts are the same whether the aerospike nozzle is cooled
or not. The spikes look very similar from the outside, except for the coolant exhaust port in the truncated face of the
cooled aerospike nozzle. The switchover from uncooled spike to cooled spike is therefore straightforward, except for
the additional plumbing setup for the cooled tests. We drilled the exit port of the cooled spike to correspond to the
geometry established in Cold Flow Test Run 5, giving a cold flow rate of 0.36 lbm/sec. This first test of the cooled
spike was similar in burn intensity to the test of the uncooled spike, and is shown in Figure 13. In fact, Table 2,
which compares the burn conditions of the uncooled test with the two cooled ones, show that T* for the two
different runs is essentially the same. However, in the test with the cooled nozzle, the maximum nozzle temperature
was only 470°F over a 21.1-second burn, compared to 1251°F in 8.2 seconds for the uncooled nozzle. Clearly, for
this non-aggressive burn, the 0.36 lbm/sec coolant flow rate was more than enough to effectively cool the nozzle.
For the second hot-fire test of the cooled nozzle, we wanted a significantly more intense burn, so we chose an
oxidizer flow rate that was predicted to result in a chamber pressure of 280 psig, roughly the design pressure for an
NPR 20 nozzle operating at sea level. For this aggressive burn, we chose a restrictor plate and exit ID of the
aerospike coolant exhaust port to yield a coolant flow rate of approximately 0.5 lbm/sec, corresponding to Cold Flow
Test Run 7. The average chamber pressure for the actual run was 268 psig and the measured coolant flow rate was
0.49 lbm/s. The relative aggressiveness of this burn is evident in the significantly increased T* and by contrasting
the plumes shown in Figure 12 Figure 14. The key result, moreover, is that the maximum spike temperature was
held to a maximum of 1050°F during an18.4-second burn.
The results of all three hot-fire tests are shown in Table 2. Note that the mass flow rate predictions agree closely
to the measured values. Since the two cooled aerospike tests utilized the same aerospike nozzle, and that after the

14
American Institute of Aeronautics and Astronautics
 second test its dimension ns were essenttially unchangeed, we believee that ‘reusabillity’ has been demonstrated, to the
extent posssible in this sho
ort study. The spike was in all
a measurable ways intact annd reusable afteer the second teest.
Downloaded by "Indian Institute of Technology, Bombay" on March 31, 2025 | http://arc.aiaa.org | DOI: 10.2514/6.2012-4201

Figure 12. Rocket motor plume during th

he first test of thee cooled aerospikke nozzle [Run ##90].

Figure 13: Hot-fire tesst results for the first test of the ccooled aerospikee nozzle [Run #990]

15
Americcan Institute off Aeronautics aand Astronautiics
Downloaded by "Indian Institute of Technology, Bombay" on March 31, 2025 | http://arc.aiaa.org | DOI: 10.2514/6.2012-4201

Figure 14: Rocket motor plume during the second test of the cooled aerospike nozzle [Run #91]

Figure 15: Hot-fire test results for the second test of the cooled aerospike nozzle [Run #91]

16
American Institute of Aeronautics and Astronautics
 Table 2: Hot-fire test results for the three tests conducted.

O/F Burn Coolant Spike

Test Ratio Time Pc Thrust C* T∞ P* T* Flow Rate Temp
ave max ave (CEA) (CEA) (CEA) (CEA) peak
(sec) (psig) (psig) (lbf) (ft/sec) (°F) (psia) (°F) (lbm/s) (°F)
Un-Cooled Spike Test 1.67 8.2 132.0 148.2 29.3 4142 2743 80.0 2328 NONE 1251

Cooled Spike Test 1 1.68 21.1 142.2 217.1 44.6 4154 2766 85.7 2349 0.36 470

Cooled Spike Test 2 2.40 18.4 267.8 302.8 81.3 4723 4108 156.3 3601 0.49 1050

V. Conclusion
Downloaded by "Indian Institute of Technology, Bombay" on March 31, 2025 | http://arc.aiaa.org | DOI: 10.2514/6.2012-4201

We demonstrated experimentally that N2O provides sufficient cooling to extend the range of operation and make
an annular aerospike nozzle reusable. These proof-of-concept experiments followed extensive analysis and design,
including, sequentially-coupled and fully-coupled finite element models, for both uncooled and cooled aerospike
nozzles.
We designed and built a furnace brazing system to braze a copper spike shell onto a stainless steel spike core
into which coolant channels are machined, and developed a process for making successfully brazed assemblies. A
combination of cold-flow testing and CFD analysis was used to guide final adjustments to the cooled aerospike
nozzle.
A multiphase, cold-flow coolant analysis of the cooled aerospike geometry was developed and used to estimate
coolant flow during hot-fire tests, with good agreement with the actual test results. We carried out a total of 7 cold
flow test runs to calibrate our system, where we varied the size of two orifices in the system to control both the
coolant pressure and mass flow rate. The result of this series is a set of conditions that provide mass flow rates
varying from approximately 0.28 lbm/sec to 0.74 lbm/sec, and aerospike coolant pressure varying between 275 psig
and 620 psig.
We conducted three hot-fire tests on aerospike nozzles: one on an uncooled aerospike and two on the same
N2O-cooled aerospike nozzle. The aerospike copper surface temperature was reduced significantly, using our N2O-
cooling strategy, and the cooled nozzle maintained dimensional stability and integrity during the course of two hot-
fire tests, making it ‘reusable’.
We feel that the concept could be further improved by adding regenerative injection of the cooling N2O into the
combustion chamber. This change to the setup would also directly improve specific impulse, and thus the overall
rocket motor performance. The engineering required for this improvement was beyond the scope of the present
study.

Acknowledgments
The authors gratefully acknowledge the support of the Donald E. Bently Center for Engineering Innovation. We also
wish to acknowledge Dr. Michael Kerho of Rolling Hills Research Corporation for conducting the Fluent CFD
analysis discussed in this paper, Mr. Terry Cooke of Cal Poly for his help with instrumentation and data acquisition
in the experiments discussed herein, and Mr. James Gerhardt of Cal Poly for machining the nozzles presented in this
paper.
This work has been made possible in part due to financial support from the following three Small Business
Technology Transfer grants (STTR) from NASA: “A Reusable, Oxidizer-Cooled, Hybrid Aerospike Rocket Motor
for Flight Test,” Phase I Grant; “A Reusable, Oxidizer-Cooled, Hybrid Aerospike Rocket Motor for Flight Test,”
Phase II Grant; and “A Rayleigh Flow Experiment on Nitrous Oxide (N2O) to Assess the Limits of N2O Cooling,”
Phase III Grant.

References
i
Bui, T., et al., "Flight Research of an Aerospike Nozzle Using High Power Solid Rockets," AIAA Paper 2005-
3797.

17
American Institute of Aeronautics and Astronautics
 ii
Lemieux, P., “Nitrous Oxide Cooling in Hybrid Rocket Nozzles”, Progress in Aerospace Sciences, V46, 2010.
iii
Huzel, D. K., and Huang, D. H., Modern Engineering for Design of Liquid-Propellant Rocket Engines, AIAA
Progress in Astronautics and Aeronautics, Vol. 147, 1992.
iv
"System, Method and Apparatus for Cooling Rocket Motor Components using a Saturated Liquid Vapor Coolant
Mixture," US Patent Application No. 12/821,100, filed on 6/22/2010. Patrick Lemieux, William R. Murray,
Thomas W. Carpenter and James A. Gerhardt.
v
Grieb, D., Lemieux, P., Murray, W. R., Mello, J. D., Cooke, T. H., and Gerhardt, J. A., “Design and Analysis of a
Reusable N2O-Cooled Aerospike Nozzle for Labscale Hybrid Rocket Motor Testing”, AIAA Paper 2011-5539-133.
vi
Imbaratto, D, The Interaction between Throttling and Thrust Vectoring of an Annular Aerospike Nozzle, MS
Thesis, ME Department, California Polytechnic State University, Sept. 2009.
Downloaded by "Indian Institute of Technology, Bombay" on March 31, 2025 | http://arc.aiaa.org | DOI: 10.2514/6.2012-4201

18
American Institute of Aeronautics and Astronautics