I was recently an invited speaker at a recent American
Carbon Society meeting held at North Carolina State University in March
2024. The meeting topic was thermal
management, and it mostly (but not
entirely) dealt with carbon-based materials.
Dr. Cheryl Xu of the school’s Department of Mechanical and Aerospace
Engineering ran across me and invited me to attend. The meeting was managed by Dr. Xu and by Dr.
Weiming Lu of the American Carbon Society,
a senior employee of Collins Aerospace.
I submitted 3 presentations in both slide set and
submittable paper format, which were all
accepted. Only one could be presented
live, the other two would be poster
presentations. In point of fact, all three were made into posters. Being
about the oldest person present, I felt
quite obsolete among specialized experts used to dealing with complex
software. But, apparently because of my having turned an
awful lot of heat protection materials into char during my defense career, they seemed glad to have me there.
The “live” one was “Early Ballpark Analysis: Entry”,
which took entry as the specific example for a generalizable concept
screening process that makes true brainstorming feasible in terms of time and
budget. Plus, being able to get ballpark estimates without
software makes possible the recognition of garbage-in, garbage-out problems when using
software. Those were my messages. They seemed well-received.
My other two presentations were “Old Ceramic
Composite”, an update on a paper I gave
previously at a Mars Society convention a few years ago, and “Ramjet Ablative Liners”, about some experimental combustor insulations
I tested years ago as part of a very productive ramjet direct-connect test
series.
I have already posted a sort of trip report on the
conference, the entry screening
paper, and the ramjet ablatives
paper, on this site. They are just below as “Presenter at
Workshop” 23 April 2024, “Entry Concept
Screening” 3 May 2024, and “Ramjet
Ablative Liners” 1 June 2024.
What follows is the text of the submittable paper for “Old
Ceramic Composite”. A related article
here on this site is “The ‘Warm Brick’ Ramjet Device”, posted 2 November 2021.
-----------------------
Old Ceramic Composite GW Johnson 12-31-2023
Abstract
At one job many decades ago,
the author had need of an insulating liner for an airbreathing
combustor, which could be fired many
times without needing refurbishment.
Inspired by NASA’s Shuttle tile,
he tried low-density ceramics,
and was successful on his second attempt with a fabric-reinforced
ceramic composite.
That effort long ago was done on company R&D
funding, with a low budget. There was no money for real materials
development and characterization activities,
it had to be made from commercially-available products, and not the most expensive ones at that. What worked was a commercial ceramic pipe
insulation paste, reinforced with a
commercially-available ceramic fire curtain cloth used in aircraft engine
nacelles.
Many years after that effort, the author realized that this material had
potential as a refractory entry heat shield material. He identified some ways and means to use it
for that purpose.
Background
The author was doing feasibility testing for a towed
aircraft infrared decoy concept back in 1984.
This thing used a fuel-fired, ram-fed airbreathing combustor as a hot
gas source for creating the infrared signature.
It was intended mostly for high subsonic to transonic flight, at low altitudes. That corresponds to the ground attack and
ground support scenarios.
The author needed a cheap combustor insulator that could
support multiple burns without any refurbishment, and the then-available ablatives were not
“it”. Inspired by NASA’s low-density and
consequently low thermal conductivity shuttle tile, he decided to try low-density ceramics himself. This was low-budget company R&D, though.
He could only use ordinary commercial materials, and certainly not the most expensive ones.
Experimental Material
The author selected a moldable pipe insulation paste from
Cotronics Corporation called 360M, which
was a water-based slurry of alumino-silicate flakes and fibers, cured by drying out. Handbook data indicated far lower density and
hot thermal conductivity than most dense ceramics.
It was quite porous and he did not want hot combustion gas
entering those pores, so he needed a
surface sealant. He used as a “paint” a
ceramic adhesive from Cotronics, designated 901, for that sealing function. It,
too, was water-based, cured by drying. Both were recommended as max 100 C oven
cures, although the author cured them at
103 C “just to make sure”. Both of these
materials are still available from Cotronics today, although you generally have to ask them about
the 360M paste, it is considered
obsolete.
The author’s liners were retained inside the combustor shell, but not bonded to it. See Figure 1. The first example worked fine, until the fuels testing reached the point of determining rich and lean combustor blow-out limits. Rich blowout is marked by very strong instabilities, showing up as harsh pressure oscillations. Those cracked the fragile insulator all the way through, and the combustor promptly spit the whole unit out in pieces, starting a grass fire that had to be stomped out.
Figure 1 – Basic Data About the Two Experimental Liner
Materials
Taking a cue from his experiences building fiberglass canoes
and kayaks with his father, the author
decided to try fabric reinforcement for his low density ceramic liner as the
means to withstand the nasty oscillating pressure loads. As it turned out, 3M Corporation offered as a commercial
product a ceramic fire curtain cloth for aircraft engine nacelles. This cloth was going to be used anyway, as the heated source of infrared signature
from the decoy.
A that time, 3M
offered Nextel 312 alumino-silicate fibers as yarns woven into cloth. If memory serves, the cloth designation, specifically chosen for its high air
permeability, was AF-14. Since then,
more ceramic materials have become available from 3M for a variety of
purposes. But Nextel 312 is still
available in the AF-14 cloth from 3M today.
The author molded the ceramic composite liner as alternating
layers of 360M paste and wraps of the cloth,
again cured at 103 C, just like
the plain liner. This one was very
successful, resisting many rich blow-out
instability events, several dozen test
burns of varying lengths, and
accumulating several hours of burn time,
all without any refurbishment at all.
When it proved successful, he
just used it; he never characterized it
for any of its properties. There was no
money for that.
About The Old Decoy Testing
See Figure 2.
It shows the test combustor rig hardware, Figure 3 below shows the test facility used
for this work. The combustor shell has
the dimensions of 2-inch iron pipe,
about 1.5 inch diameter as insulated,
and about 3 inches long, inside. Ahead of it is a duct piece where fuel-air
mixing takes place, and ahead of that is
an inlet piece where air capture and diffusion,
and fuel injection, take
place.
The author and his team used an automotive spark plug and an
aftermarket automotive high-energy ignition set to achieve fuel-air
ignition, mounted in the recirculation
zone just downstream of the step area change that was our flameholder. They successfully tested with hydrogen, propane,
aviation gasoline, commercial jet
fuel, and technical grade ethanol as the
fuels.
Figure 2 – The Decoy Test Article
Figure 3 – Improvised Free Jet Test Facility
The test facility comprised a PVC pipe stilling chamber with
a wooden convergent-only nozzle block that created a jet of high-speed
air, up to about Mach 0.9. It was powered by what in 1984 was the
largest-capacity rental air compressor trailer, anywhere in Texas. The temperature of the air stream was
measured as 190 F in the exiting jet.
The test article was mounted on a heavy steel stand made of
plate and pipe. This was positioned
about a stilling chamber diameter downstream of the air nozzle block’s exit, immersed in the exiting jet. That’s basically a “poor man’s free jet test
facility” for ramjets. Fuel control was
by the sight, sound, and smell,
as sensed by the author, standing
a few feet away in the hot jet blast behind the test article.
As a minimum, it took
the author and a technician to operate the facility and run tests. Sometimes the team included a second
technician, and up to two junior
engineers as assistants.
Simple Molding Tools,
Near-Pristine Used Liner
See Figure 4 for an image of the molding tool set
that the author used constructing the combustor liner and nozzle pieces as
ceramic composites. The molding plug
mandrels are wooden items he made using his drill press to spin them, in lieu of a lathe. (As previously
mentioned, this was low-budget company R&D stuff.) The finish on the mandrel plugs is just a
coat of varnish. The combustor shell was
its own liner outer mold shell. He had a
separate outer shell for molding the nozzle blocks.
Figure 4 – Very Simple Molding Tools
In Figure 5 is a view of the combustor liner after
all the testing was done. Note the
near-pristine appearance of the ceramic composite. The nozzle was similarly near-pristine. The surface temperatures approached the 3250
F melt point of the material, as
evidenced by some very localized surface-flowing downstream, most of that in the nozzle block throat, where heat transfer film coefficients are
highest.
This material normally has an ultimate max service
temperature of about 2350 F, limited by
shrinkage cracks upon cooldown. There
are some such cracks visible, but they
do not go all the way through thanks to the fabric layers, and there was very little in the way of fluid
shear forces to “pick” at these cracks,
since the internal flow velocities were so low.
Figure 5 – Liner Condition After Testing Concluded
Attempted Material Characterizations, Decades Later
Many years later, the
author analyzed the combustor performance for typical test conditions, and then used these to help set up a
cylindrical steady-state heat balance model of the insulated combustor. This hardware was immersed in 190 F air as
its external coolant, turbulently flowing
by at high subsonic speeds. On the
inside, full-rich flame temperatures
dropped across the internal boundary layer to very near the meltpoint of the
ceramic, at its internal surface.
After an hour-long burn at the indicated conditions, one can presume that heat transfer is
steady-state. Back in 1984, the author licked his thumb and touched it to
the shell, to find that it would just
barely boil spit, meaning it was near
215-220 F in thermal equilibrium. The
shell itself was 300-series stainless steel,
with a high thermal conductivity,
and thus very nearly isothermal,
as the numbers indicated.
Not knowing what the actual hot thermal conductivity of the
ceramic composite was, the author ran his
thermal analysis repeatedly in a spreadsheet,
across a large range of possible thermal conductivity values. The one that “fit” the observations of
“near-melt” and “boil-spit” the best (and this is crude, of course), was 0.02 BTU/hr-ft-F, which is in metric 0.035 W/m-C. See Figure 6 below for all the data, and Figure 7 below, for the temperature distribution plot obtained
at the “best” thermal conductivity value.
That hot thermal conductivity is actually lower than the hot
thermal conductivity reported in the Cotronics handbook for the 360M paste
alone! Because low thermal conductivity
correlates with low density (or high void volume fraction), this implies the density of the author’s
composite was lower than the handbook density,
as well. He honestly does not
know what that finished density really was,
but it subjectively felt like it was somewhere near that of
industrial-grade styrofoam.
Figure 6 – Simple Cylindrical Steady State Heat
Balance, Parametric On Liner Thermal
Conductivity
Figure 7 – Temperature Distribution for “Best-Fit” Thermal
Conductivity
While the heat balance analysis was actually done in US
customary units, I converted the
temperature distribution to metric for Figure 7. The thermal gradient indicated for the
evidently low thermal conductivity liner is quite remarkable! It’s in the vicinity of 270-275 C per mm of thickness, or in US units 12,000-13,000 F per inch! And that effect is a very real result! The hot gas temperature and outer shell
temperature are known well enough to support that assertion.
Possible Explanation for Low Thermal Conductivity and
Density
The apparent low thermal conductivity and density, estimated years afterward for this ceramic
composite, are well below the Cotronics
handbook values for their cured 360M material.
See Figure 8 below. There
are only two differences here between the normal use of that paste as trowel-on
pipe insulation, and what the author did
with it: (1) he cured his liner at a
higher-than-recommended temperature by about 3 C, and (2) there is a cloth of some inherent
porosity embedded within it.
Cotronics recommends at most 100 C for the cure temperature
that drives the water out by evaporation.
The paste is aluminosilicate solids slurried in water, with something in the water that enables the
solids to adhere together, once the
water is gone.
The elevation where the author did this work is about 500
feet above mean sea level, so the normal
boiling point of water there is very slightly below the sea level value of 100
C. He cured at 103 C, so the slurry water in his layup actually
flashed into steam, at least
partly. He suspects that the steam
“wormholing” its way out of the layup, acted to increase its void fraction, thereby lowering both finished density and the
resulting thermal conductivity.
Only the water in the slurry could penetrate into the pores
of the fabric, and into the pores
between the fibers in its yarns. The
solid flakes and fibers could not do that,
and yet being similar materials,
the cure chemistry did allow them to bond with that fabric. But the large inherent porosity in that
fabric would necessarily add to the void fraction in the finished article. That would also act to lower density and
thermal conductivity of the finished product.
At least, that is the
author’s hypothesis. There is no proof, other than “it worked”.
Figure 8 – Rough Estimates of Material Properties Made
Decades Afterward
Adaptations for Use as an Entry Heat Shield
To use this stuff as an entry heat shield material, three things must be addressed: (1) achieving high thermal emissivity to make
re-radiation cooling more effective, (2)
the risk of shrinkage cracks in a high fluid shear environment, and (3) mounting on a exterior surface.
The two plots in Figures 9 and 10 show calculated
results for a black surface of high emissivity in Figure 9, and in Figure 10 for the stock white surface
of low emissivity. In making the
plots, “high” thermal emissivity was
assumed to be 0.80, and “low” as 0.20. The plot ordinates are equilibrium surface
temperature such that re-radiation to Earth temperatures exactly balances
stagnation heating, versus how much
stagnation heat flux as the abscissa.
Also shown are the max surface temperature limits imposed by the onset
of cracking and melting.
Figure 9 – Calculated Steady-State Heat Balance For High
Emissivity
Figure 10 – Calculated Steady-State Heat Balance For Low
Emissivity
The switch to a high emissivity black surface
would seem to be easy enough to address.
The author called Cotronics and asked if significant carbon black could
be added to their 901 adhesive. Their
answer was that a fair amount actually could be added, and they could do it. So,
you just seal the surface pores of your layup with a black “paint”
instead of a white “paint”. Nothing
exotic.
To address the risk of high fluid shear forces tearing
at shrinkage crack lips, one
simply operates at conditions below the shrinkage crack limit, in turn well below melting. That makes stagnation zone service for Earth
entries at around 25 W/sq.cm very marginal,
but it would have application to Mars entries either from low orbit or directly
off the interplanetary trajectory there. This does presume very blunt shapes
with very large effective “nose radii”. Stagnation convective heating flux is modeled
per H. Julian Allen’s 1953-vintage warhead entry analysis: Qconv/A =
K(density/nose radius)0.5(velocity)3.
NASA‘s shuttle tile was bonded directly to the structural
substrate (metal airframe surface) with an RTV adhesive. That bondline is a single point failure
mode, however! The same could be done with my ceramic
composite, but there is another
possibility arising from the fabric reinforcement that is embedded within it!
If made in panels upon a metal substrate panel as depicted
in Figure 11, it can be bonded
with RTV to that substrate, but the free
edges of the reinforcing fabric can also be folded around the edges of that
substrate, and clamped in place on its
backside. That provides the
redundant retention that NASA’s tiles never had!
Figure 11 – Achieving Redundant Retention
All this requires in the way of vehicle design change is
putting the airframe’s metal skin panels onto the airframe framing as
already-insulated substrate panels,
instead of installed bare-metal panels to be insulated afterwards. That does require fasteners instead of welding
or rivets, to facilitate removing
individual panels for repairs.
Development Status
The author’s ceramic composite from 1984 is basically decades
obsolete, and never received any
materials development effort, rendering
it technologically immature. Others have
since gone far beyond where the author ever got. But the concept of a ceramic cloth
reinforcement embedded in a ceramic matrix still has very good promise today! The greatest of these promises is the
possibility of redundant retention. But
his notion of a ceramic matrix composite has already been used.
Figure 12 shows what NASA developed from its Shuttle
tile technology, that is currently
flying on the USAF X-37B. This serves
even in its stagnation zones, where the
older shuttle tile was inadequate. NASA
calls this “Tufroc”. It features a
ceramic composite cap of higher density,
strength, and temperature
resistance, atop a low-density ceramic substrate, in turn applied as a tile.
Figure 12 – NASA “Tufroc” 2-Piece Ceramic Tile as Used On
USAF X-37B
The cap piece is a reinforced composite, similar to the author’s concept: a carbon fiber preform reinforcing a ceramic
matrix. The substrate piece is a
low-density ceramic that is a strength improvement upon the older shuttle tile
material. It is not a composite, although it could be. The two pieces are mechanically linked, but the substrate is still just bonded with
an RTV adhesive to the metal airframe substrate.
This material will handle Earth low orbit entries in
stagnation zones, it is much more
resistant to impact and erosion damage,
and it does not suffer from surface shrinkage cracking. However,
the bond to the substrate is still a possible single point failure!
If the substrate piece were at least partly a fabric-reinforced
composite, then it could possibly
incorporate the author’s clamped-fabric idea for redundant retention.
None of these NASA materials are ordinary commercial
products. They are special-made, and available only in smaller sizes. The
author’s stuff is made of fairly-inexpensive ordinary commercial
materials, and it could be made in
rather large panels. Nobody is doing
that, just yet.
Conclusions
The author’s material is very promising, but totally undeveloped. It “just worked” as a combustor liner., and it was made of ordinary commercially-available
materials.
There are other more recent materials of good promise, that are mature and flying.
Adding the author’s clamped reinforcing fabric concept to
those newer materials might well solve the redundant retention problem.
About the author:
The author had a 20 year career in aerospace defense doing
new product development design,
analysis, test, and evaluation, entering the workforce in the
slide rule days with a master’s degree in aerospace engineering. Transition to the then-expensive pocket
calculators was underway, but desktop
computers were still years in the future.
That career was mostly (but not entirely) in rocket and ramjet missile
propulsion. It ended with a plant
shutdown and layoff in 1994, just when
the industry was shrinking drastically.
The author then had a second 20 year career that was mostly in teaching
(at all levels from high school to university),
plus some civil engineering and aviation work. He earned a doctorate in general engineering
late in life, to support that second
career. He is now retired.
