Aerogel insulation applications for liquid hydrogen launch vehicle tanks

J.E. Fesmire
*
, J.P. Sass
Cryogenics Test Laboratory, NASA Kennedy Space Center, Mail Code KT-E, Kennedy Space Center, FL 32899, USA
a r t i c l e i n f o
Article history:
Received 13 July 2007
Received in revised form 4 January 2008
Accepted 6 March 2008
Keywords:
A. Adsorbents
B. Hydrogen
B. Liquid helium
C. Thermal conductivity
D. Temperature sensors
a b s t r a c t
Solutions to thermal insulation problems using aerogel beads were demonstrated for space launch vehi-
cles using a model of the space shuttle external tanks liquid hydrogen (LH
2
) intertank. Test results using
liquid helium show that with aerogel, the nitrogen mass inside the intertank is greatly reduced and free
liquid nitrogen is eliminated. Physisorption within the aerogel was also investigated, showing that the
sorption ratio (liquid nitrogen to aerogel beads) is about 62%. The insulating effectiveness of the aerogel
shows that cryopumping is driven by thermal communication between warm and cold surfaces. This
technology can solve heat transfer problems and augment existing thermal protection systems on launch
vehicles.
Published by Elsevier Ltd.
1. Introduction
Space launch vehicles experience a recurring problem with the
uninsulated areas on the cryogenic propellant tanks and feedlines
that allow air to condense or ice to form. These areas, including
ange joints, bracket supports, expansion bellows, and other cavi-
ties, are uninsulated by design. Conventional thermal insulation
materials, if applied to these critical system components, could
cause damage to the launch vehicle during ight because the vehi-
cles mechanical articulations have been restricted or its cavities
have been imperfectly sealed. New aerogel-based thermal insula-
tion systems have been developed for use in critical locations on
launch vehicles. Aerogel materials and a liquid oxygen (LO
2
) feed-
line bellows have been previously described [1]. Recent work has
focused on ambient-pressure insulation systems for liquid hydro-
gen tank dome applications. This research was motivated by the
need to solve a longstanding problem on the liquid hydrogen
(LH
2
) tank dome area of the space shuttle external tank.
Postight photographs have shown a concentration of foamdeb-
ris around the intertank-to-LH
2
-tank ange joint of the external
tank (ET). In the cavity formed between the intertank wall, ange,
and the LH
2
tank dome (termed the crevice as shown in Fig. 1),
there exists a periphery of bare metal at 20 K. The nitrogen purge
gas within the intertank volume condenses on this cold surface. Li-
quid nitrogen (LN
2
) accumulates in this crevice in the hours before
launch when propellant loading and conditioning are underway. A
small amount of solid nitrogen (SN
2
) also forms on the colder side of
the crevice. The LN
2
can then seep through the ange joint to the
underside of the foam and into voids in the foam. Upon launch,
the LN
2
can vaporize due to the ullage pressurization (hot gas),
aerodynamic heating, and reduction in ambient pressure. As the
nitrogen expands from liquid to gas, the voids or discontinuities
within the foam can pressurize, creating foam debris that can dam-
age the vehicle during ight. In addition, the lift-off mass penalty is
signicant, up to 230 kg for both the liquid and solid nitrogen.
A thermal insulation system using bulk-ll aerogel material has
been successfully developed and demonstrated in a 1/10th scale
model of the LH
2
intertank ange area. This aerogel insulation sys-
tem and method of application were developed over several years
of materials research and demonstration testing at the Cryogenics
Test Laboratory at NASAs Kennedy Space Center.
2. Experimental investigations
Experimental investigations included performance testing of
cryogenic thermal conductivity, mass transport (cryopumping),
and physical adsorption. Various types and forms of aerogel mate-
rials were studied. Flexible blanket and bulk-ll aerogel materials
are now commercially available on a large scale for low-tempera-
ture applications [2,3].
Aerogel beads or granules in bulk-ll form could provide the
thermal insulation for the LH
2
intertank ange area. The small sur-
face area of the LH
2
tank dome that is not covered by spray-on
foam insulation (SOFI) material would be insulated with this mate-
rial, which has very low thermal conductivity. The aerogel is an
open-cell material with an average pore size of 20 nanometers
and a surface area of approximately 800 m
2
/g. The average size
0011-2275/$ - see front matter Published by Elsevier Ltd.
doi:10.1016/j.cryogenics.2008.03.014
* Corresponding author. Tel.: +1 321 867 7557; fax: +1 321 867 9197.
E-mail address: james.e.fesmire@nasa.gov (J.E. Fesmire).
Cryogenics 48 (2008) 223231
Contents lists available at ScienceDirect
Cryogenics
j our nal homepage: www. el sevi er . com/ l ocat e/ cr yogeni cs
of the spherical beads is 1 mm. With a particle density of 140 kg/
m
3
and a bulk density of about 80 kg/m
3
, an interstitial space of
roughly 60% is typical. The aerogel beads are lightweight, easy to
handle and convey, and present no safety hazard. The bead parti-
cles pack well, are capable of a high level of elastic compression,
and are generally reusable. The material is commercially available
from Cabot Corporation under the trademark Nanogel [4].
2.1. Thermal conductivity
Thermal characterization of the aerogel materials in various
congurations was done using insulation test cryostats [2,3,5].
The temperature prole for a 25-mm thickness of aerogel beads
for an ambient pressure test is given in Fig. 2. The 2.5-mmlayer po-
sition gives a temperature difference of approximately 57 K. Based
on this information, an inference is made that the thickness re-
quired to develop an isotherm in the range of 63 K (SN
2
)77 K
(LN
2
) from a cold boundary of 20 K (LH
2
) is less than 2.5 mm be-
cause of the increased effectiveness of thermal insulation at lower
boundary temperatures [6].
2.2. Mass transport (cryopumping)
Cryopumping experiments for a cold boundary temperature of
77 K were performed using both a closed-chamber apparatus (cryo-
stat) and an open-chamber apparatus (cold column). Results show
that stabilization occurs in approximately 2 h after cooldown and
isotherms form in approximately 3 h, as previously reported [1].
Cryopumping effects between the boundary temperatures of
300 K and 20 K were observed in the demonstration testing using
liquid helium (LHe) and were found to follow a similar pattern of
stabilization in the 23 h after cooldown. Thermal equilibrium oc-
curred in approximately 810 h, as evidenced by on-off oscillation
in the periodic ow rate of nitrogen purge gas supplied to the
apparatus.
2.3. Physical adsorption (physisorption)
In general, aerogel materials have a very high capacity for sorp-
tion of nitrogen molecules. All of the nitrogen mass that is added
during the material cooldown process is sequestered inside the
Fig. 2. Experimentally measured temperature distribution through a 25-mm thickness of aerogel beads with boundary temperatures of 77 K and 293 K at one atmosphere
nitrogen pressure.
Fig. 1. The LH
2
intertank ange area of the space shuttle external tank is shown on the side of the tank that attaches to the orbiter (left). A cutaway view of the intertank
ange area shows the crevice (right).
224 J.E. Fesmire, J.P. Sass / Cryogenics 48 (2008) 223231
nanoporous structure of the aerogel. The nitrogen is adsorbed to
the surfaces in monolayers of molecules. Differences in surface
energies do not permit condensation and formation of liquid mol-
ecules within the pores [7].
The capacity of bulk-ll aerogel for physical adsorption, or
physisorption, of LN
2
was measured using various methods. In
the mass uptake method (cup test) shown in Fig. 3, the aerogel
material is placed in a bottom-perforated cup, fully saturated using
LN
2
, and then placed on a precision scale to measure the weight
gained. Nitrogen is taken up inside the nano-sized pores of the
aerogel as the material surfaces are cooled. This process is known
as physisorption, meaning that layers of nitrogen molecules are
stacked inside the porous structure. Results show that the sorption
ratio (nitrogen to aerogel) is about 62% by volume. The sorption ra-
tio on a mass basis is about seven to one. For the aerogel material
in the ET intertank case, the available sorption capacity for nitro-
gen is many times more than the minimum necessary for the 8
10 h propellant loading timeline. This feature means that any extra
aerogel insulation material in the intertank crevice would simply
provide additional insulating benet without intoducing any ad-
verse effect.
2.4. Free liquid nitrogen experiment (cold nger apparatus)
The capability of the aerogel to insulate a cold metal surface be-
low 20 K and eliminate formation of free LN
2
was tested using a
novel liquid nitrogen production method. This cold nger appara-
tus uses gravity to visually determine if any collection and move-
ment of nitrogen in the liquid state occurs when the cold surface
is insulated with aerogel material. The ET intertank crevice is also
simulated in that it does not positively contain LN
2
that may col-
lect. The cold nger apparatus has slits in the bottom analogous
to the shim gaps in the intertank ange joint. The apparatus con-
sists of a stainless steel cooling coil inside a clear plastic tube with
slits cut just above a bottomfoam plug. The tube is lled with aero-
gel beads (or left empty) and purged with gaseous nitrogen. The
cooling coil is supplied from a vacuum-jacketed transfer hose con-
nected to an LHe dewar, as shown in Fig. 4.
Three experimental congurations were made using the cold
nger apparatus. The rst conguration, with no aerogel insula-
tion, is shown in Fig. 5. In this case the free LN
2
began owing from
the tip 6 minutes into the test and owed continuously for the
whole -h test. The rate of nitrogen liquefaction was approxi-
mately 0.42 ml per second as measured by weighing the LN
2
that
Fig. 3. The nitrogen sorption capacity of the aerogel material is measured by the
mass uptake method using liquid nitrogen (cup test). Individual nitrogen molecules
are sequestered within the nanoporous internal structure.
Fig. 4. Cold nger experiment with no aerogel insulation inside tube (conguration 1) determined the rate of nitrogen liquefaction.
J.E. Fesmire, J.P. Sass / Cryogenics 48 (2008) 223231 225
was caught in an insulated cup after it owed through the bottom
drain slits. At that rate, LN
2
could ll the internal volume of the
apparatus in about 2 h. Conguration 2, in which the cylinder
is partially lled with aerogel insulation, is shown in Fig. 6. Even
in this atypical case, none of the LN
2
produced in the exposed cold
coil region above the aerogel was able to move to the bottom of the
apparatus during the 1-h test. The full aerogel insulation case,
conguration 3, is shown in Fig. 7. No LN
2
owed from the appara-
tus during the entire 4-h test. At the end of this test, the entire con-
tents of the tube were spilled into a pan, and the aerogel material
was found to be completely loose and friable with no condensed
nitrogen observed.
3. LH
2
intertank demonstration test unit
The demonstration test unit depicted in Fig. 7 was designed to
determine the operational limits of the new aerogel insulation sys-
tem. The test unit, at approximately 1/10th scale, is constructed
from thin-wall stainless steel, with a G-10 feedthrough for the
LHe supply and an aluminum lid. The crevice prole is approxi-
mately full scale (with a 12.7-mm bottom surface and a 50-mm ex-
posed cold metal tank surface).
Temperature sensors were located from the bottom of the joint
crevice to above the aerogel beads, as well as on the tank, intertank
barrel section, and foam surfaces, as shown in Fig. 8. Pressure
Fig. 5. Cold nger experiment with partially lled aerogel insulation (conguration 2) showed thermal stability even with oversaturated aerogel.
Fig. 6. Cold nger experiment with full aerogel insulation inside tube (conguration 3). No liquid nitrogen was produced within the tube even after 4 h of cooling with LHe.
226 J.E. Fesmire, J.P. Sass / Cryogenics 48 (2008) 223231
sensors on the base of the crevice and on the lid measured any
pressure difference from top to bottom. The intertank barrel sec-
tion is designed for full vacuum to allow cryopumping tests for a
closed-mass system. SOFI was applied to the sides of the unit
and the dome of the tank in the thicknesses specied for ight con-
guration. Additional thicknesses of SOFI were applied to the cryo-
gen feedthrough and the bottom of the unit to provide a stable
thermal prole for the tank dome and intertank ange.
The test setup included an LHe supply system from a 42,000-l
tanker, a GN
2
supply system, temperature sensors, pressure trans-
ducers, a GN
2
ow control and metering system, and a full data
acquisition system (see Fig. 9). Viewports on the lid (two, each
with a 50-mm diameter) allowed visual examination of the crevice
area. Cold helium gas was distributed through a spray header on
the underside of the tank dome to maintain a cold boundary
temperature (CBT) of 20 5 K. Two electric tape heaters were af-
xed to the lid to maintain a warm boundary temperature (WBT)
of 300 2 K.
4. Test results
The test results show that the aerogel insulation material elim-
inates production and accumulation of LN
2
in the ange area of the
intertank. Using this material also avoids any gas entrapment that
Fig. 7. Isometric cutaway view shows the main features of the demonstration test
unit of the external tanks LH
2
intertank.
Fig. 8. Section view of 1/10th scale demonstration test unit shows where temperature sensors are located. Closeup of the full-scale intertank crevice area is shown on the
right.
Fig. 9. Demonstration testing using an LHe tanker (left). 1/10th scale demonstration unit (right).
J.E. Fesmire, J.P. Sass / Cryogenics 48 (2008) 223231 227
could cause collateral problems during ascent to orbit. Tests were
performed in two series and are summarized as follows.
Part A of the test series included two tests performed with a
bare crevice and six tests with aerogel insulation installed. For
these tests, the nitrogen gas was sealed in the intertank. The LHe
tanker provided a continuous liquid helium supply for long tests.
The WBT was controlled by using lid heaters and a shield, or hat,
to minimize the thermal inuence of the cold helium boiloff va-
pors. Visual observations were made at regular intervals by looking
through the two viewports on the lid of the test unit.
Observations from Part A of the testing showed that the aerogel
insulation dramatically reduced nitrogen cryopumping and gener-
ally increased temperatures throughout the intertank. The appear-
ance of the aerogel material remained unchanged throughout all
testing. In the tests with the bare (uninsulated) crevice, solid nitro-
gen did not readily form due to the heat input from the intertank
wall and the exothermic process of freezing. The LN
2
in the bare
crevice was observed to boil at the outer edge near the intertank
wall, as was expected.
Fig. 10 compares the total mass ow of the nitrogen purge gas
due to cryopumping for tests run with and without aerogel insula-
tion. Measurements are taken for the entire system, including the
crevice region (both with or without aerogel), the center feed-
through, and the dome of the tank. Fig. 11 presents photographs
of tests in which insulation is not used, which show LN
2
accumu-
lations, and photographs of tests in which the aerogel insulation
is used, which show that the aerogel material is unchanged.
Fig. 12 summarizes the steady-state temperatures and shows the
effect of insulating the cold metal surface of the intertank cavity.
Following Part A of the testing, which had a quantied GN
2
ow,
Part B of the test series was performed to obtain better visual and
physical evidence of the condition of the material in the crevice re-
gion. The GN
2
purge ow was maintained as before to keep a dry
nitrogen environment at ambient pressure inside the intertank,
but the ow rate was not measured. Viewports were routinely
opened during the tests and used for access to the crevice area to
take measurements, obtain samples, and probe the aerogel mate-
rial. The following items were examined: LN
2
level versus time,
location of solid N
2
formation, friability of the aerogel material,
and whether LN
2
was present in the aerogel material. Photos and
video were taken at regular intervals during all tests.
Observations made in the Part B tests revealed that only a min-
imal depth of aerogel insulation is required to prevent LN
2
forma-
tion and that the aerogel material remains completely dry and
friable (no free LN
2
). With just enough aerogel insulation to cover
the bare metal portion of the tank dome, LN
2
was prevented from
accumulating. Also, solid nitrogen forms more readily with smaller
amounts of LN
2
due to the heat capacity of the liquid. The aerogel
insulation system is tolerant of an imperfect installation as shown
by tests in which a portion of the crevice was left bare. A test using
Fig. 10. Totalized GN
2
ow for the Part A (closed system) test series shows the effect on the crevice from having aerogel insulation versus the effect without insulation during
cooldown and stabilization phases.
Fig. 11. Photographs looking directly down at the crevice: tan area with black depth indicator lines is the dome of the tank; red and green items are 14-mm diameter corks to
indicate the depth; wires are thermocouple leads. A test without the aerogel insulation, showing LN
2
accumulation inside the crevice region, is shown in the left and center
images. A test with aerogel insulation, showing an unchanged appearance, is shown on the right.
228 J.E. Fesmire, J.P. Sass / Cryogenics 48 (2008) 223231
an insulation system with a minimal aerogel spot on one side and
bare metal spot on the other side is shown in Figs. 13 and 14. The
crevice temperature is maintained at approximately 20 K during
all tests. Fig. 15, from another test of an insulation system with a
bare metal portion, shows that liquid nitrogen produced in the
exposed crevice progresses into the adjacent aerogel material by
physisorption.
5. Analysis and discussion
A summary of the thermal effect of aerogel beads on the inter-
tank is given in Fig. 16. The aerogel keeps the warm side warm and
the cold side cold. A minimal depth of aerogel was found to be en-
ough to prevent nitrogen from liquefying. In fact, the primary
requirement is to simply cover the exposed bare metal portion of
the tank dome. In practical use of the new insulation system on
the ight tank, the aerogel will be lled to a level that is convenient
to verify.
Beyond the obvious advantages in eliminating LN
2
in the crevice
region on the External Tank, there are ancillary advantages and dis-
advantages to using the proposed aerogel insulation system.
Advantages are as follows:
Fig. 12. Comparison of tests with and without aerogel, showing the effectiveness of the insulation system in keeping the warm side (intertank wall) thermally isolated from
the cold side (tank dome). Temperature ranges at multiple locations are given for Tests 6 and 10 (with aerogel insulation) and Tests 7 and 8 (without aerogel insulation).
Fig. 13. Test of insulation system shows minimal aerogel spot at beginning (left) and end of test (right). A small amount of aerogel near bare metal portion of tank dome is
saturated with nitrogen at the end of the test.
Fig. 14. Test with aerogel insulation shows bare metal spot at beginning (top) and
end of test (bottom). Liquid nitrogen is produced in the exposed crevice and is then
adsorbed by the adjacent aerogel material.
Fig. 15. Another test with exposed tank dome at 20 K. The aerogel material darkens
over time as it is saturated with nitrogen by physisorption.
J.E. Fesmire, J.P. Sass / Cryogenics 48 (2008) 223231 229
Aerogel is high thermal performance material.
s Has the lowest thermal conductivity of bulk-ll materials at
ambient nitrogen pressure at 14 mW/m-K (293 K and 77 K
boundary temperatures).
s Insulates all bare metal surfaces of the cavity.
Material is cavity-lling, free-owing, and hydrophobic.
s Withstands thermal cycling due to its excellent compression
strength.
s Lightweight at 80 kg/m
3
.
High physisorption capacity of nitrogen.
s About 62% by volume (LN
2
:aerogel).
s During cooldown, some nitrogen molecules are adsorbed
within the pores of the aerogel.
s After cooldown, cryopumping processes are essentially
stopped.
Total nitrogen purge uptake is dramatically reduced.
s Requires approximately 66% less compared to uninsulated
case.
Free LN
2
in intertank ange and crevice region is eliminated.
s For all phases of loading and launch operations.
s Under both transient and steady-state conditions.
The only disadvantages identied are associated with installa-
tion. The present aerogel insulation system would need to be in-
stalled in the vertical position. In the demonstration unit, the
aerogel stayed in place even with a forceful impingement of purge
gas directed straight downward from the lid. In the ight tank the
aerogel would be protected further by: (1) the presence of a stiff-
ening ring immediately above the crevice area, (2) the favorable
location of purge entrances to the large intertank volume, and (3)
the very steep angle and depth of the crevice. An enhanced system
using a breathable cap material to hold the aerogel in place could
be developed if it were necessary to install the aerogel in a hori-
zontal position.
Important ndings also resulted from baseline demonstration
tests of the uninsulated intertank. Results showed a delicate
thermodynamic balance that determined whether the nitrogen
remained in the liquid state or solidied in the crevice. In one
long-duration test without aerogel, solid nitrogen was observed
to grow across the part of the crevice and subsequently melt before
the end of the test. On other occasions, a probe was used to feel for
the bottom of the LN
2
-lled crevice. This technique indicated that
solid nitrogen formed to a thickness of about 19 mm at the bottom
of the crevice near the tank wall and tapered down toward the
ange gap. The overall amount of LN
2
present appeared to affect
the likelihood of solidication, with frozen nitrogen occurring more
readily in smaller amounts of liquid. Understanding this solidica-
tion process is important in determining the severity of the problem
of liquid nitrogen seeping through the intertank ange.
A key advantage of the aerogel system is that it will remain
completely loose and friable for the entire loading and launch pro-
cess. Conventional attempts to solve the problem of LN
2
in the
intertank have relied on creating a perfect seal to isolate the nitro-
gen from the cold surfaces. These approaches were not successful
due to the difculty of the intertank remaining sealed when ex-
posed to a sudden thermal and pressure gradient during operation.
A failed seal is potentially worse than no seal at all because the LN
2
has a long time in which to accumulate and only a short time to
escape through the failure point as an evaporated gas during vehi-
cle ascent. Because the aerogel insulation system is not intended to
seal out nitrogen, it avoids the problem of nitrogen entrapment
during ascent.
6. Conclusion
A new thermal insulation system using a bulk-ll aerogel mate-
rial was developed to keep liquid nitrogen from accumulating
within the intertank of the Space Shuttle. The proposed solution
is to insulate the exposed cold surface of similar cavities on the
LH
2
tank dome with aerogel beads. Typically, these cold surface
areas are difcult or impossible to insulate with conventional
materials such as polyurethane foam. The aerogel insulation sys-
tem eliminates all free liquid nitrogen within the crevice and thus
removes a cause of foam debris being created in the intertank
ange area. And unlike conventional insulation materials, aerogel
beads do not restrict mechanical movements and do not rely on
the cavity being perfectly sealed. As an added benet, the Shuttles
ight weight can be reduced by up to 230 kg. The installation pro-
cess is straightforward, and no adverse effects have been identied.
The operability of the aerogel-based thermal insulation system
has been demonstrated in testing down to 20 K. The experimental
design, research, development, and demonstration testing were
performed by the Cryogenics Test Laboratory at NASAs Kennedy
Space Center. Sixteen tests were conducted using liquid helium
to provide the necessary cold boundary temperature for LH
2
simu-
lation. Laboratory testing of the liquid nitrogen sorption character-
istics and other physical properties of the aerogel material was also
performed.
The new technology can improve thermal performance of fu-
ture space launch vehicles. Using aerogel bulk-ll materials and
other aerogel-based insulation systems will provide new options
for the design and safe operation of cryogenic systems.
Acknowledgements
The authors are glad to acknowledge the expert assistance of
many engineers and scientists dedicated to the Space Shuttle Pro-
gram. We especially appreciate the insight and assistance of Char-
lie Stevenson, Trent Smith, Stephen Huff, and Wes Johnson for the
demonstration testing. This work was supported by the NASA Engi-
neering and Safety Center (NESC) through the leadership of Mike
Kirsch and Andreas Dibbern. We thank Adam Dokos and the NASA
Prototype Laboratory team for constructing the test unit.
Fig. 16. Effect of the aerogel insulation system on temperature distribution within
the intertank. The average steady-state temperature at each location was calculated
for tests with aerogel (Tests 6, 9, and 10) and tests without aerogel (Tests 7 and 8).
The reported temperature differences (tests with aerogel minus tests without ae-
rogel) show that insulating the crevice resulted in warmer temperatures through-
out the intertank.
230 J.E. Fesmire, J.P. Sass / Cryogenics 48 (2008) 223231
References
[1] Fesmire JE. Aerogel insulation systems for space launch applications. Cryogenics
2006;46(23):1117.
[2] Cabot Corporation, <http://www.nanogel.com>.
[3] Fesmire JE, Augustynowicz SD, Rouanet S. Aerogel beads as cryogenic thermal
insulation system. Adv Cryogenic Eng Am Inst Phys 2002;47:15418.
[4] Begag R, Fesmire JE. Nonammable, hydrophobic aerogel composites for
cryogenic applications. In: Presented at international thermal conductivity
conference 29, Birmingham, AL, USA; June 2007.
[5] Fesmire JE, Morris DL, Augustynowicz SD, Nagy ZF, Sojourner SJ. Vibration and
thermal cycling effects on bulk-ll insulation materials for cryogenic tanks. Adv
Cryogenic Eng Am Inst Phys 2006;51B:135966.
[6] Choi YS, Barrios MN, Change HN, Van Sciver SW. Thermal conductivity of
powder insulations for cryogenic storage vessels. Adv Cryogenic Eng Am Inst
Phys 2006:51B.
[7] De Boer JH. The dynamical character of adsorption. 2nd ed. London: Oxford
Press; 1968.
J.E. Fesmire, J.P. Sass / Cryogenics 48 (2008) 223231 231