AIAA 2017-5134

AIAA SPACE Forum

12 - 14 Sep 2017, Orlando, FL
AIAA SPACE and Astronautics Forum and Exposition

Design and Manufacture of Composite Liquid Oxygen

Propellent Tank for University Rocket

Neil Benkelman,∗ Russel Burger∗, Alex Farias∗

Francesca Frattaroli∗, Weldon Peterson∗ and Chris Wilson∗
Portland State University, Portland, Oregon, 97207, U.S.A.

When designing a rocket to overcome Earths gravitational force, the dry mass of the
propellant tank is the most limiting factor for amateur and university-level rocket teams.
Downloaded by 80.82.77.83 on September 26, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-5134

Most rocket groups are dependent upon heavy aluminum tanks to hold their liquid propel-
lant. Single-piece, carbon-fiber composite tanks are strong and light, but require expensive
automated fiber placement techniques that are not a viable option for amateur teams, who
operate with a tight budget and limited resources in comparison to their corporate or gov-
ernment counterparts. Portland State University’s composite-tank capstone and research
project is an attempt by senior undergraduate engineering students to design, build, and
test an affordable, lightweight, composite liquid oxygen propellant tank for eventual inte-
gration into Portland State Aerospace Society’s 100 kilometer rocket.

Nomenclature
LOX Liquid Oxygen
LN2 Liquid Nitrogen
P T F E Polytetrafluoroethylene

I. Introduction
PSAS is building a rocket with the intention of reaching an altitude of 100 km (the Von Karman line).
For this to be possible, the rocket’s mass ratio must be optimized. The total mass of a rocket is typically 85
percent propellant and 15 percent vehicle and payload. This ratio makes dry mass reduction a major point
in any rocket design. As such, PSAS desired to pursue the development of a composite material tank to
hold Liquid Oxygen (LOX), the elected propellant. PSAS is also developing an electric feed system that will
lower the internal pressure required to 45 psi. The primary assumption in pursuing this option is that a tank
fabricated using composite materials will significantly reduce the mass ratio when compared to a conventional
aluminum tank. Recognizing there are significant challenges associated with designing a composite material
tank, the team was tasked with providing a proof-of-concept design without being constrained by mass.
The focus on this project was to design, fabricate, and test a proof-of-concept, liquid oxygen compatible
tank made out of predominantly composite material for use with the Portland State Aerospace Society’s
liquid-fueled composite rocket.
∗ BSME Recent Graduate, Maseeh College of Engineering and Computer Science, AIAA Student Member

1 of 10

American
Copyright © 2017 by the American Institute of Aeronautics and Astronautics, Inc. Institute of Aeronautics and Astronautics
All rights reserved.
 II. Design and Manufacture
A. Mating Rings
Downloaded by 80.82.77.83 on September 26, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-5134

Figure 1: Main Ring Design V1.1

6061-T6 aluminum was chosen as the material for the mating rings. It is a common aircraft material due
to its high strength-to-weight ratio. Aluminum also has no ductile to brittle transition temperature, giving
it superior impact resistance at cryogenic temperature. The mating rings on the other airframe modules are
also made of this material. This eliminates any concerns about mismatched thermal expansion properties
and corrosion as a result of galvanic coupling at the interface between the tank and other modules.
The mating rings were designed to be the backbone that holds the whole design together. All other
subsystem components connect to these parts. The carbon fiber and Nomex layers have no strength before
curing, and therefore necessitated the use of a mandrel to wrap these materials into shape during compression
molding. The mandrel must be removed after curing, which means the tank at this stage must have open
ends. The rings have 12 threaded holes in a circular pattern that match a hole pattern on the end cap so that
the two components can be fastened together with 3-48 screws. With the knowledge that hole alignment
during the shrink fitting process would be important, a keyway was added to the mating rings as a method
of straightening the end cap during the shrink fit application so that all fastener holes are aligned once
completed. An 1/8” cylinder extends from the ring into the layers, providing a surface for the composite
sandwich structure to adhere to, and an offset on the interior to seat the PTFE liner (see Figure 1). A lip
extending away from the tank was designed as a placeholder on the ring which will gain mating features to
integrate the tank to the rockets airframe.

2 of 10

American Institute of Aeronautics and Astronautics

 B. End Caps
Downloaded by 80.82.77.83 on September 26, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-5134

Figure 2: End Cap Design V1.1

The material chosen for the end caps was 6061-T6 aluminum, which incorporated a 1/16” thickness plate
with a rounded central portion to account for the addition of fitting valves. Twelve through holes were
placed along the outer rim of each end cap, allowing fasteners to pass through and secure into the associated
mating rings. In between these locations, a round cut was added in order to align with the keyway mentioned
above, to aid in the shrink fitting process during tank fabrication. On the bottom portion of each cap, a
longitudinal surface was recessed down into the tank, in order to provide a seal by compressing the PTFE
liner in between this portion of the end cap against the mating ring.

3 of 10

American Institute of Aeronautics and Astronautics

 C. Layering Configuration (Adhesive, Nomex, Carbon Fiber)
An exploded view of the tank assembly is shown below in Figure 3.
Downloaded by 80.82.77.83 on September 26, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-5134

Figure 3: Final Tank Design

Carbon fiber material is advantageous in aerospace design due to its outstanding strength-to-weight
ratio, which is far superior to that of metals.1 The layering scheme used in our design was adapted from
the airframe design developed for PSAS by the LV3.0 Capstone Team.2 Three materials were donated to
the team by PSAS. These were woven carbon fiber sheets, a 1/4”-thick honeycomb core material called
Nomex, and a structural adhesive film designated Metlbond. Additional Nomex sheet in 1/8” thickness was
purchased to accomodate the geometry of our mating ring design. The carbon fiber layers provide the tank
with strength in the axial and tangential directions, and the honeycomb core layers provide strength in the
radial direction. When combined, the composite provides structural support from external loads including
the rockets weight and thrust, and internal loads from the pressure of the oxidizer.
A 1/8” Nomex layer is placed just outside the PTFE liner, followed carbon fiber, then 1/4” Nomex, and
another layer of carbon fiber. Structural adhesive is placed between each of the layers, and is the outermost
layer to provide a smoother skin than carbon fiber (see Figure 4).

Figure 4: Material Layers (left to right): Metlbond Adhesive, Nomex Honeycomb, Carbon Fiber

4 of 10

American Institute of Aeronautics and Astronautics

 D. Liner Materials
Fluorocarbon polymers are particularly well suited to low temperature applications because they are the
only known materials that retain a measurable ductility at temperatures very close to absolute zero (-269C).
This makes them ideal for use as insulators and as static seals at these temperatures.3
The liner would have to survive the carbon fiber curing process or be integrated after laying up the
structural shell of the tank. It would also need to be held in place by either an adhesive or some mechanical
joining method. The team was provided with a structural adhesive that could be cured in tandem with the
carbon fiber, so it made sense to place the liner and cure simultaneously. This decision to co-cure required
that the liner material maintain integrity over the full range of temperatures between the boiling point of
liquid oxygen (-183C) and the curing temperature of the composite materials (176C). Only two materials were
found with these capabilities. They were ethylene tetrafluoroethylene (ETFE) and polytetrafluoroethylene
(PTFE). Ultimately, PTFE was chosen on the basis of availability and cost.

E. Shrink Fit Seal

The sealing of the tank was accomplished with a shrink fit, which led to the overall success of the design.
This was performed by submerging the aforementioned end caps into a LN2 bath to sufficiently cool, and
Downloaded by 80.82.77.83 on September 26, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-5134

therefore shrink the diameter of the cap. Simultaneously the layered module was heated in an oven to provide
expansion to the mating rings. The end caps were then gently inserted, one end at a time, into the mating
rings, and pressed into place with a dead blow hammer. The achievement of this process was that the PTFE
liner was compressed between the two aluminum surfaces, effectively seating it so that it functioned as a
gasket and formed a complete interior seal to the tank (see Figure 5). This sealing method does not require
any additional sealant or caulking, as it purely mechanical with no seams. It also eliminated the need to
adhere the PTFE liner into place.
Axially oriented stainless steel fasteners help hold the end cap onto the mating rings. This was done as a
redundant safety precaution, in case the friction force between the PTFE and aluminum proved insufficient.

Figure 5: Section view of shrink-fit seal between end cap and mating ring.

5 of 10

American Institute of Aeronautics and Astronautics

 III. Testing & Results
Due to safety concerns and resource access issues, initial performance testing of the tank did not include a
pressurized LOX test. Instead, two proxy performance tests were performed. Multiple intermediate designs
were fabricated, however, due to high manufacturing costs, only a single final prototype was constructed and
tested. Presented results document the performance of this final prototype.

A. Hydrostatic Burst Test

The first test performed was a hydrostatic burst test, intended to evaluate the effectiveness of the mechanical
seal produced by the shrinkfit process. In application, the aluminum-PTFE seal must function at ambient
temperatures during the initial fill process and be able to withstand the 45 psi (3 atm) of internal pressure
during standard operation.
During the test, the tank was filled and pressurized with water at ambient temperatures via a manual
pump. The internal pressure was gradually increased until observable leakage occurred at a pressure of 100
psi. At that point, the tank failed to maintain constant internal pressure and fluid was observed leaking from
the threaded and plugged holes placed in the end-caps to attach future plumbing fixture, attributable to
manufacturing error during the tapping process. Therefore, this test yielded a factor of safety of 2.2 without
Downloaded by 80.82.77.83 on September 26, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-5134

observing failure in the region of interest (the liner seal).

Unable to exceed a hydrostatic pressure of 100 psi, the internal pressure was increased to past 200 psi
(maximum gauge reading) through continuous pumping, with no observed structural failure or leakage at
the aluminum-PTFE seal. This result does not confirm if hydrostatic pressure could be maintained at 200psi
over an extended period of time or if these pressures could be maintained at cryogenic temperatures.

Figure 6: Hydrostatic Pressure Test Setup

6 of 10

American Institute of Aeronautics and Astronautics

 B. Cryogenic Compression Testing
The second test, cryogenic compression, was intended to evaluate the structural integrity of the tank as an
airframe module. At cryogenic temperatures, the tank must be able to withstand the same compressive
forces as ambient temperature modules which have already been approved for use by PSAS.
The LV4 airframe modules are expected to experience approximately 900 lbs of axial compressive force
during launch. To simulate these conditions, the tank was filled and cooled with liquid nitrogen (-195 C)
before being compression tested. The tank withstood 9.6 kip of force before structural failure, defined by a
sudden deformation and decrease in experienced loading (see Figure 7).
Downloaded by 80.82.77.83 on September 26, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-5134

Figure 7: Cryo-Compression Test Results

Failure was visually identified by deformation of the top mating ring and outward buckling of the external
carbon fiber layer near the base of the tank (see Figure 8).

7 of 10

American Institute of Aeronautics and Astronautics

Downloaded by 80.82.77.83 on September 26, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-5134

Figure 8: Shrink-Fit Tank Design V1.0 after Cryo-Compression Failure

The full scale tank will have a greater height relative to wall thickness than this scaled-down prototype,
making it more susceptible to buckling, but the total loading will remain 900 lbs and be distributed across
a larger diameter ring. Testing of a full-scale tank will be required to confirm the structural soundness of
this tank design, but the current factor of safety of 10 suggests that a significant decrease in compressive
strength can likely be absorbed without rendering the tank unsafe.

IV. Further Research

We will continue to test successful designs using a cryogenic fluid such as LN2 and LOX at future stages
of the development process. We have shown that we exceed the design factor of safety for pressure at
ambient temperature, and compressive load at cryogenic temperature. We believe our current design has the
capability to meet the client requirements for all combined flight loadings. Because LOX compatibility has
not been verified, additional testing must be performed. Due to the dangers concerning testing with liquid
oxygen, we are working with PSAS and the Air National Guard to design and perform tests under proper
safety procedures. For these tests to be meaningful they will need to be performed at working pressurized
conditions.
Before any pressurized LOX testing is done a shrink fit tank should be pressurized with LN2 . There are
various methods that have been discussed. One would be to use the hand pump we used to hydro test the
tank. Questions arise about whether the hand pump will work and the accuracy of the pressure gauge at
cryogenic temperatures. Another avenue is to fill the tank with LN2 and seal it at both ends. As the LN2
boils the pressure will increase inside the tank, and pressure as a function of time can be estimated and a
burst pressure can be calculated.
It may be possible to achieve a stronger tank if the outside wall of the PTFE tube was chemically etched
similar to the sheet liner, especially when a thinner tube is used. Stress calculations were done under the
assumption of perfectly bonded surfaces within the layering. Due to the mismatched coefficients of thermal
expansion between the various materials, being able to adhere the PTFE to the layering scheme would

8 of 10

American Institute of Aeronautics and Astronautics

 mitigate any potential deformation or shifting of the liner.
When working with PTFE at cryogenic temperatures we observed that when thermally cycled the ma-
chined PTFE tubing does not recover to its original dimensions when brought back to ambient temperatures.
This leads to the recommendation that the PTFE be thermal cycled before final machining is performed.
This will allow for a better fit and less variability in manufacturing of the tank.

V. Conclusion
The Final Tank design is shown below in Figure 9.
Downloaded by 80.82.77.83 on September 26, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-5134

Figure 9: Final Tank Design, Shrink Fit Tube Liner V1.1

A composite pressure vessel capable of withstanding an internal pressure of 100 psi and an axial force
of 9 Kip was designed and manufactured. To hold the liquid fuel in the vessel a liner was required. A thin
walled PTFE tube was chosen as the liner material. PTFE was chosen due to it’s chemical inertness and high
operating temperature range. The pressure vessel has a modular design for integration with the intended
rocket. The aluminum mating rings that serve as an attachment point are also structural components of
the pressure vessel. A shrink fit was achieved between the mating rings and end caps that mitigated many
of the concerns associated with the project requirements. This design lengthened leak paths and removed
the necessity of gaskets or other sealing mechanisms. It also reduced the required bolt load subjected to the
fasteners.
The Pressure vessel was subjected to two different tests, a hydro static test and a cryogenic crush test.
The results from both tests showed the tank would be able to withstand operating conditions.
Although current data suggests that future iterations of this design will be LOX compatible further
testing must be done, such as pressurizing the tank with LN2 and subsequently with LOX.

9 of 10

American Institute of Aeronautics and Astronautics

 Acknowledgments
This project would not have been possible without the support of the mechanical engineering department
faculty, members the Portland Aerospace Society, members of the amateur rocket community, and the NASA
Oregon Space Grant Consortium.
We would like to thank Dr. Jun Jiao and Dr. Gerald Recktenwald for providing guidance, support, and
supervision throughout our capstone project. We would like to thank Mr. Tom Bennett for providing us
with access to equipment to conduct compression testing on our tank prototype. We would like to thank
Mr. Michael Chunning and Mr. Kristen Auclair for providing advice and assistance in the machine shop.
We would like to thank Mr. Joseph Shields, Mr. Erin Schmidt, and Mr. Robert Watzlavick for providing
design, test, and manufacturing advice for our composite tank. We would also like to thank our capstone
sponsor, The Portland State Aerospace Society, for providing composite molding equipment and materials.
This project was sponsored in part through Oregon NASA Space Grant Consortium, grant NNX15AJ14H.

References
1 Callister,
W. D., Materials science and engineering : an introduction, John Wiley & Sons, 2007.
Downloaded by 80.82.77.83 on September 26, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-5134

2 PSAS, “PSAS LV3 Airframe Github, https://github.com/psas/lv3.0-airframe,” .

3 Zesus, “Low Temperature Properties of Polymers,” Key Engineering Materials, Vol. 59-60, 1991, pp. 35–46.

10 of 10

American Institute of Aeronautics and Astronautics