Low-Density Non-Ablative Ceramic Heat Shields

In the article dated 3-2-13 and titled “A Unique Folding-Wing Spaceplane Concept” on this site, I investigated the concept of a folding-wing spaceplane to side-step the spaceplane designers’ dilemma. That investigation took the form of a simplified bounding analysis, indicating only basic feasibility-or-not.

------------------------------------------------------------

Update 10-11-13:

This material was incorporated into a paper on ceramic entry heat shields that GW presented at the 16^th Annual Mars Society Convention in Boulder,  Colorado,  in August 2013.  That paper was very well-received.  That presentation and many others were videotaped for publication on YouTube.  Those videos are now available.  Be aware they take a while to load,  being 20-to-30 minutes long. 

GW's presentation at the 2013 Mars Society convention on a lightweight thermal protection ceramic material is available on Youtube:
Reusable Ceramic Heat Shields - GW Johnson - 16th Mars Society Convention.
http://www.youtube.com/watch?v=3MXYY3jnNr0

(Control+click to follow link)

------------------------------------------------------------
Update 1-4-15:

Using yttria-stabilized zirconia instead of alumino-silicates could increase the operational surface temperature limit to something on the order of 4000 F (2204 C),  making LEO entry shields far more feasible at attractive ballistic coefficients.  It remains to be seen whether available fibrous zirconia products could be successfully processed into a low-density re-radiating heat shield with redundant retention and enhanced structural strength.  However,  I am now looking at these materials for another combustor liner application.  No answers yet.

-------------------------------------------------------------

The spaceplane designer’s dilemma is that stagnation point heating is dramatically reduced by flying dead-broadside during entry, but, that same broadside attitude causes extreme air loads that rip off wings. Plus, highly-swept wings cause bad landing characteristics.

This dilemma is not a solvable problem in fixed geometry, no matter how highly-swept the wings might be. The real solution lies in some kind of variable geometry.

Of many possibilities, the one not involving jettisoning major hardware might simply be folding the wings into the wake zone behind a fuselage. That fuselage then enters dead broadside, and thus as blunt as possible, like a space capsule. This is not a new idea, having been proposed unsuccessfully by Lockheed in the 1950’s for what eventually became the X-20 Dyna-Soar project.

The results of my bounding analysis study implied a folding wing spaceplane, with straight wings of subsonic airfoil section. It unfolds long after entry, while hanging from a parachute, at entirely subsonic speeds. The wings are unfolded during ascent, for purposes of abort.

In order to obtain characteristics more like those of an ordinary aircraft instead of a traditional spacecraft, a robust non-ablative heat shield is needed, so that periodic refurbishment or replacement is unnecessary. The same simplified entry bounding calculations indicate that this might be possible with a non-ablating ceramic heat shield for dead-broadside entry attitudes, if the ballistic coefficient can be kept low enough, and the bluntness high enough.

A low-enough stagnation point convective peak heating makes a low-density re-radiative ceramic heat shield feasible, because the equilibrium skin temperature for re-radiation of the heat load is low enough to stay within material limitations. One example of such a material is the ceramic tiles used on the Space Shuttle. The black ones on the windward side were useful at higher heating rates than the white ones on the leeward side. That is the influence of surface color upon effective or average emissivity.

The entry environment at Mars is generally less demanding than that for entry from Earth orbit, primarily because the vehicle velocities are far lower there. The low-density re-radiative ceramic heat shield would be even more feasible there. The main advantage is the very low weight of such a heat shield, something very important to any vehicle designs intended to be used at Mars, and especially to those intended to fly more than once at Mars.

One such material already available is the low-density ceramic tile used on the Space Shuttle. Another version of this same low-density ceramic material that is less susceptible to damage, and less expensive to install and maintain, than Shuttle tile, would be very highly desirable. Such a material may actually be available, although it is still quite experimental.

In 1984, I constructed a low-density, fabric-reinforced ceramic composite material, and used it as an experimental low-conductivity, non-ablating monolithic liner insert in a very small ramjet combustor. This was, in turn, part of a device called “Warm Brick”, whose purpose and characteristics are not germane here.

Entry Conditions Comparison

Entry from low Earth orbit (LEO) was modeled as a surface-grazing elliptical transfer from a circular orbit at 600 km altitude. From low Mars orbit (LMO), the same kind of surface-grazing transfer was assumed from a circular orbit at 200 km. These provide very shallow trajectory angles at the entry interface point, as demonstrated in the article dated 8-10-12 and titled “Big Mars Lander Entry Sensitivity Study” on this site.

Additionally, a direct entry at Mars from interplanetary transfer was also investigated (see the article dated 8-12-12 and titled “Direct-Entry Addition to Mars Entry Sensitivity Study” on this site. This direct entry was assumed to be a “typical” higher velocity, but at the same very shallow entry angle, mostly for analytical convenience. Control of entry angle from direct transfer trajectories can be difficult, just as it was returning from the moon. These three scenarios are compared in Figure 1 (all figures are at end of this article).

The simplified entry analysis is very old and very simple: an exponential model based on a density scale height approach. This was originally developed for the early warhead entry work of the 1950’s. It derives ultimately from the work of H. Julian Allen at NACA, more recently referenced in the work of Justus and Braun at NASA. I published an item dealing (in part) with this model as the article dated 6-30-12 and titled “Atmosphere Models for Earth, Mars, and Titan” on this site. I corrected the convective heating model given in the Justus and Braun report, in the article dated 7-14-12 and titled “’Back of the Envelope’ Entry Model” on this site.

For entry velocities under about 10 or 11 km/sec, convective heating dominates the entry heating, and is usually empirically taken to be proportional to the square root of the density/nose radius ratio, and to the cube of the velocity. At hypersonic speeds, vehicle drag coefficient is essentially constant, allowing the use of a very simple constant ballistic coefficient model.

Vehicle Models

The Earth orbital spaceplane model is depicted in Figure 2. Unfortunately, that illustration includes mixed unit systems, but the key parameters ballistic coefficient and bluntness (belly or heat shield radius of curvature) are given in metric. That design concept required only a modest increase in belly radius (from the original Spaceplane1 configuration to the final Spaceplane2) to achieve a peak heating low enough for a high-emissivity ceramic shield to be feasible.

In this type of bounding analysis, the inherent uncertainty is large, while for the LEO scenario the surface temperature margin is nil. Therefore, this Earth return application of low density ceramics may, or may not, actually be as feasible as it seems. That would require far more sophisticated design analyses to resolve. Margins were definitely larger for both Mars entry scenarios, so the use of these materials is very definitely a feasible option there.

The Mars entry vehicle application is completely different from the spaceplane, being essentially a ballistic entry capsule shape of very large size (60 metric tons), to support manned landings and major cargo shipments there. This was originally envisioned as a one-use chemical vehicle, with a simple rocket ascent stage, as depicted in Figure 3, and described in the article dated 8-28-12 and titled “Manned Chemical Lander Revisit” on this site. However, the same basic shape and concept might also apply to any refueled, reusable “landing boat” or “Mars ferry” in that size class.

Key to the success of such a design is a substantial reduction in heat shield weight, and the elimination of heat shield refurbishment and replacement costs. This particular Mars lander vehicle model was used simply for analytical convenience, not necessarily strict realism.

Re-Radiation Equilibrium Analyses

For a non-ablative, re-radiating low-density ceramic heatshield, there is convective input from the entry flow, re-radiated output due to a high skin temperature, and essentially zero conduction through the material. There is essentially no radiative heat input from the shock layer at entry speeds under about 10 or 11 km/sec. The Stefan-Boltzmann law applies, and with no ablation products in the shock layer, re-radiation is essentially unobstructed. Whether you consider radiation to space at 3 K, or to the Earth at around 300 K, makes no real difference. It might as well be 0 K at this level of analysis.

The controlling factor in the balance is the spectrally-averaged emissivity of the surface. Both Shuttle tile and my experimental composite are alumino-silicate materials. These are normally white in color, and have high reflectance at visible and near infrared wavelengths, and low reflectance at long infrared wavelengths. These are most definitely not “gray emissivity” materials.

In essence, the emissivity of these plain white alumino-silicate materials is roughly 0.2 from visible wavelengths up to around 2 microns, rising to about 0.8 at around 3 microns, and then remaining high at the longer wavelengths. Both my old composite, and the white leeside Shuttle tile, fall in this category. A conservative average emissivity value would be 0.2, it could be a little higher.

All of these alumino-silicate materials (regardless of form) melt near 3200-3250 degrees F, and experience a solid phase change that very seriously degrades structural properties, near 2300-2350 degrees F. Thus the material limit is only about 2300 degrees F. The mechanism of degradation is irreversible shrinkage and the resulting cracking, plus a very serious embrittlement of the affected materials.

Black-coated windward-side Shuttle tiles have a much higher average infrared emissivity, in the vicinity of 0.8 or perhaps higher. For black Shuttle tile, this was done with a glass-like surface layer. I never developed a black surface coating for my old “Warm Brick” composite material, although it could be done. Carbon black (by itself a very nearly spectrally “black” material, at average emissivity above 0.90), used as a pigment in the surface cement coating, offers one promising and easy-to-investigate avenue. So, I assumed that an average emissivity of 0.8 was possible for it, as well.

This energy balance showed that the critical value of peak stagnation heat flux (during entry from LEO) was 25 W/sq.cm, for the allowable skin temperature of 2300 degrees F, and an average emissivity of 0.8. The low emissivity white surfaces were not feasible for that mission. This value set the belly radius finally settled-upon for the Spaceplane2 configuration.

See Figure 4 for the black surface (high-emissivity) data, and Figure 5 for the white surface (low emissivity) data. Meltpoints and phase change limits are shown on these figures. If one exceeds the phase change limit at the surface, the material is definitely damaged, but will likely still protect you through that one entry, but then also likely require replacement after landing. The LEO scenario corresponds to the 25 W/sq.cm heating in these figures.

For Mars direct entry (shallow angle) at a generic 5.6 km/sec, the entry model showed a peak heat flux of about 11 W/sq.cm. From LMO, the peak heat flux was only 2.6 W/sq.cm. These were generated with the “Mars Lander Revisited” model, which uses a ballistic coefficient of 400 kg/sq.m, and a heat shield radius of curvature of 12.4 meters. These might be “typical” of any ballistic capsule shape, in the vicinity of 60 metric tons mass. These are the other two heating levels shown in the same two figures. Skin surface temperatures were well within the feasible range for the low density ceramics (of either type).

Entry Scenario Conclusions

A black surface (high emissivity) low-density ceramic heat shield might possibly be made to work for smaller (lower ballistic coefficient) vehicles coming back from LEO. In that scenario, one must be very careful balancing achievable surface emissivity against all possible reductions to peak heating.

The white surface (low emissivity) version cannot serve in the LEO scenario.

A black surface (high emissivity) low-density ceramic heat shield could definitely serve for vehicles landing directly upon Mars from an interplanetary transfer orbit. There is considerable margin below phase change limits to support a wide variety of such designs.

The white surface (low emissivity) form of the low-density ceramic heat shield is not feasible for direct entry at Mars.

Both the white surface (low emissivity) and black surface (high emissivity) forms of the low-density ceramic heat shield are feasible for a wide variety of vehicle designs entering from LMO at Mars.

In the appropriate roles, both white and black surface shuttle tile could serve in these heat shield roles, as just described. They could save considerable weight, but are subject to the same vulnerabilities, and the same high maintenance costs, as were demonstrated on the Space Shuttle. The real remaining question is whether the old “Warm Brick” ceramic composite material could also be feasible.

The Experimental Composite as an Alternative Material

The “Warm Brick” feasibility test device included as a part of its assembly a subminiature ramjet combustor, depicted in Photo A. In that photo, the inlet is to the right, and the combustor to the left, connected by a narrower inlet mixing tube. The combustor shell piece has the same exterior and interior dimensions as a piece of schedule-40 two-inch pipe, about 4 inches long.

The combustor shell itself was used as the insulator molding shell, as depicted in Photo B, along with the slightly-tapered wooden plug that formed the inner insulation surface. There was also a shell-and-plug combination to form the combustor nozzle from the same low-density ceramic material. These are also shown in the same photo.

The first article (not shown here), was a simple low-density molding without any reinforcement. It did not survive rich blowout instability in the combustor. For the second (successful) article, alternating layers of the low-density molding compound and a standard fire curtain cloth (both alumino-silicate) were emplaced upon the molding plugs, then forced into the molding shells, and cured all-at-once. The nozzle was made first, then used as part of the mold tooling setup for the combustor liner.

Once these articles were “cured” (by driving the water off by several hours’ exposure in a low-temperature oven at 215 F), then the surface was painted with a ceramic cement. This material is normally used to bond two molded parts together, but for this purpose, it acted as a gas-impermeable hard surface layer, same as Shuttle tile. The same low-temperature oven “cure” hardens this cement.

Photo C is a view of the second article, removed from the combustor shell, after many hours of accumulated burn time at full-strength fuel-air mixtures. The view is aft looking forward, toward the sudden-dump flameholder lip. Photo D is a view into the corresponding nozzle, looking aft.

The near-pristine condition of the ceramic is quite evident, in spite of exposures resulting in internal skin temperatures approaching the actual meltpoint of the alumino-silicate materials. After some (but not all) of these tests, some localized surface melting was noted. Strength properties were never evaluated; however, the rich blowout instability it survived was near 0.8 atm amplitude, at audio frequencies (several hundred Hertz).

A recent re-analysis of one of those tests provided a crude (but realistic) estimate of the as-built thermal conductivity of this material: 0.02 BTU/hr-ft-F = 0.035 W/m-C = 0.00035 W/cm-C. The density was never characterized, but falls in the range of commercial Styrofoam products. In both respects, this material strongly resembles the properties of Shuttle tile material.

The differences with Shuttle tile are threefold: (1) this material is substantially tougher and stronger, (2) this material was not bonded to the combustor shell at all (being a free-standing internal insert), and (3) this material can be made (and used) in very large-dimension monolithic parts.

Figure 6 shows a concept for an externally-mounted heat shield panel. It is not bonded to the baseplate metal panel, being retained instead by the ends of the fabric reinforcing material layers, wrapped around the edges of the base plate, and clamped in place. Panels like this can be simply bolted to the framing members of the vehicle. (Any vehicle capable of entering dead broadside will have substantial internal framing.) The “trick” is to lay-up and cure this item as one monolithic piece, then add the surface coating to the cured item.

Update 12-19-2013:  Actually,  the ceramic composite can be bonded to the baseplate as well as mechanically retained by holding onto the reinforcing fabric.  That make redundant retention possible,  something the Space Shuttle never had.  

The actual materials used back then are no longer available, but I have recently contacted those manufacturers, and determined that modern equivalents are currently available. The molding compound provides an open, porous matrix made of flakes and fibers of alumino-silicate mineral crystals. The fire curtain cloth is woven from yarns of alumino-silicate fibers, that are each about 0.020 inch (0.5 mm) diameter.

Update 12-19-2013:  Actually,  the original materials are still available,  but no longer show as catalog items.  You have to ask for them by name and product ID number.  And,  the ceramic cement manufacturer agrees that,  up to a point,  carbon black can be used to create a black instead of white hard surface coat.  

Other than actual testing, there are two things that need to be done to support application as an entry heat shield. One is a black (high emissivity) surface coating. The other is the detailed process development for constructing the illustrated panels. For the black coating, it might be possible to mix significant amounts of carbon black into the ceramic cement. How that would affect cure and processing is unknown, but it is the most obvious idea to try. The process development work for a shield panel assembly is seemingly pretty straightforward (it just needs to be done).

Note:  if this kind of engineering,  or this kind of novel ceramic composite material,  interests you,  do please contact me about consulting help.  Be aware that some of this (beyond this level of detail) is subject to the ITAR restrictions,  though. 

Figure 1 -- Comparison of Entry Conditions

Figure 2 -- "Spaceplane2" Concept and Characteristics

Figure 3 -- "Chemical Mars Lander Revisit" Concept and Characteristics

Figure 4 -- Re-Radiation Balance for Black Surface Ceramic Heatshield

Figure 5 -- Re-Radiation Balance for White Surface Ceramic Heatshield

Photo A -- The "Warm Brick" Feasibility Test Device Combustor Subsystem from 1984

Photo B -- Molding Tools for the "Warm Brick" Combustor Insulator (1984)

Photo C -- View Looking Forward Into the Combustor Insulator After Hours of Accumulated Burn Time (1984)

Photo D -- View Looking Aft Into the Combustor Nozzle Block After Hours of Accumulated Burn Time (1984)

Figure 6 -- Concept for Large Bolt-On Ceramic Composite Heat Shield Panel

A Unique Folding-Wing Spaceplane Concept

The spaceplane-designer’s dilemma is that peak entry heating is reduced by flying at higher angle of attack (AOA),  but also that flying at high AOA is more likely to tear the wings off.  This is because physics clearly says that stagnation point heating is reduced the larger the effective “nose” radius (really,  just the far higher effective bluntness of the broadside attitude).  Physics also says that the air loads on extended wing and tail surfaces broadside to entry flow are truly enormous. 

--------------------------------------------------------

Update 10-11-13:

This material was incorporated into a paper on ceramic entry heat shields that GW presented at the 16^th Annual Mars Society Convention in Boulder,  Colorado,  in August 2013.  That paper was very well-received.  That presentation and many others were videotaped for publication on YouTube.  Those videos are now available.  Be aware they take a while to load,  being 20-to-30 minutes long. 

GW's presentation at the 2013 Mars Society convention on a lightweight thermal protection ceramic material is available on Youtube:

Reusable Ceramic Heat Shields - GW Johnson - 16th Mars Society Convention.
http://www.youtube.com/watch?v=3MXYY3jnNr0

(Control+click to follow link)

---------------------------------------------------------

This dilemma is why all early US spacecraft were compact capsules entering blunt heat shield first,  not rocket planes with wings.  It is also why all US warhead designs from that time featured a big rounded nose on an otherwise conical shape.  All of these were heat-protected by ablative heat shields,  which are intrinsically “reusable” only a finite number of times,  at best.  The silica-phenolic ablatives on those early capsules were also quite heavy materials. 

The first flying exception was the US Space Shuttle,  essentially a “rocket plane” that entered at high AOA (around 30 degrees),  yet not dead-broadside (which would be AOA about 90 degrees).   It had to remain “streamline-enough” to not rip its wings off,  yet still enter the atmosphere at high-enough AOA to get some significant reduction of heating over much of the vehicle. 

The Shuttle nose cap,  and its wing and tail leading edges,  required combined ablative-refractory protection (in the form of carbon-carbon composite),  while the rest of the surface was cool enough to need only a reusable low-density refractory heat shield in the form of rather-fragile tiles bonded to the airframe.  It landed very fast,   with a very high sink rate on approach to landing.  It had no go-around capability.  These are all very unfavorable landing characteristics.  The X-37B is quite similar,  just a lot smaller. 

Another exception (that never actually flew) was the X-20 Dyna-Soar project,  cancelled as the first articles were coming off the production line at Boeing in the early 1960’s.  This was a highly-swept delta-wing hypersonic glider that entered pretty much “streamline”,  with its nose and leading edges protected by an early analog of the carbon composite ablative-refractory used on the Space Shuttle.  The rest of the structure was exotic alloy metal,  cooled radiatively at high skin temperatures.   Problem:  very highly-swept delta-wing hypersonic glider shapes invariably have very poor subsonic landing characteristics,  as did the Shuttle. 

Avoiding the Dilemma

The fundamental need here is to enter dead-broadside,  belly-first,  to very dramatically-reduce peak entry stagnation heating so that other shielding options become feasible,  while at the same time avoiding the dead-broadside air loads that rip off wings.  This is simply impossible in fixed geometries,  no matter how highly-swept the wings are.  And,  further,  highly-swept wings always make the landing characteristics extremely poor.  That’s also part of the dilemma.

One way to avoid this dilemma is to fold the wings (and probably the tails) into the wake zone behind a fuselage flown dead-broadside,  belly-first during entry.  This is effectively a blunt capsule shape with some “feathers” folded stream wise out behind.  This obtains the full benefit of space capsule-like bluntness to reduce peak stagnation heating.  One unfolds these surfaces long after entry is over,  when the broadside air load threat is reduced.  See Figure A.  This idea is actually quite old:  I found mention of it as a Lockheed proposal in the early 1950’s, in documents describing the history of the Dyna-Soar effort. 

Figure A – Folding-Wing Spaceplane Concept

Depending upon ballistic coefficient during entry,  and upon the amount of bluntness in the fuselage belly shape,  stagnation heating is reduced to a point where multiple well-demonstrated choices become feasible for the heat shield material on the vehicle belly surface:  ablatives,  refractories,  sweat cooling,  heat sinking,  and combinations of any of these approaches.   That selection depends upon how the numbers work out for heating,  versus any applicable material limits. 

The lighter,  the better,  in terms of reusability,  though.  That is why low-density ceramic refractories that do not ablate (such as Shuttle tiles) are attractive.  The maximum skin temperature is set by heat to be re-radiated,  and it is limited by a solid phase-change that shrinks,  embrittles,  and cracks the alumino-silicate material,  degrading it structurally.

One Step Further

Nearly every single spaceplane concept or proposal I have ever seen assumes that the vehicle flies as a lifting craft from hypersonic speeds all the way to landing,  because of the fixed-wing geometry.  No one design can accommodate such a range of speeds with good characteristics at landing.  So,  with a folding-wing design,  what if you do not fly it as a lifting vehicle hypersonically?  Or even supersonically?  What if you decelerate all the way to subsonic speeds before you pitch-over stream-wise and fly as a lifting vehicle?  The structural air loads problem is greatly reduced,  and so is the practical problem of providing the means to fold and unfold the wings (usually by hydraulics). 

That choice,  a radical departure from other designs,  would allow the designer to use straight (unswept) wings and fully-subsonic airfoil sections,  for excellent landing characteristics.  That kind of entry sequence might look like Figure B.  And that is where this design feasibility study began.

If this sort of vehicle design innovation is attractive to your project,  then I might be able to help as a consultant.  Do please contact me.  Depending upon the application and the details,  the International Trade in Arms Restrictions (ITAR) might apply.   

Figure B – Descent Sequence for a Fully-Subsonic Unfold

“Spaceplane1”

A fully-reusable spaceplane intended to fly hundreds,  thousands,  or even tens-of-thousands of missions would have operating and physical characteristics not unlike a typical utility-category aircraft.  Those are usually rather robust,  tough airframes,  capable of enduring abuse without requiring much maintenance. 

Because of the straight folding wings (and perhaps folding tails),  it would also have to have operating and physical characteristics not unlike a straight-wing aircraft carrier fighter plane.  Those are also very tough,  robust airframes designed to absorb the incredible punishments of carrier landings and combat,  without a lot of maintenance. 

The common thread here is “tough,  robust,  low-maintenance”.   One would expect this to show up in the inert weight fraction representing the structure:  essentially the aircraft empty weight divided by its max gross weight.  The wing loading and aspect ratio both impact flight handling and landing characteristics,  so those are also important parameters. 

Because I have an old book in my library with some properties of many and various aircraft in it,  I picked (more-or-less arbitrarily) a Cessna 195 utility aircraft,  and the World War 2-vintage Grumman F6F-5 “Hellcat” carrier fighter.  I used the averages of the mass properties and aerodynamic characteristics of these planes,  specifically empty weight ratioed to max gross weight,  aspect ratio,  and gross weight wing loading.  See Table 1.  Surprisingly,  both aircraft shared almost exactly the same empty-to-gross weight ratio at 60%.  (In “rocket equation terms” for the weight statement,  that is the inert weight fraction.) 

Table 1 – Aircraft Properties Comparison Table

This is to be a simple rocket airplane,  propelled into space by an existing commercial rocket.  It would leave the sensible atmosphere while being boosted,  at something like Mach 2.  Even with straight subsonic wings,  this is feasible,  and provides an avenue for abort.  The on-board propulsion that I “modeled” is quite generic:  I used 305 sec specific impulse (Isp) as “typical” of practical storable,  or ordinary cryogenic,  rocket combinations,  such as MMH-NTO and kerosene-LOX.  This would also correspond to the shorter expansion bells that would be useful near sea level.  Many design options are realistically “covered” by that value. 

I tried to mildly-overestimate the propellant fraction necessary to cover orbit circularization,  de-orbit burn,  and (upon landing) climb to 3000 feet for go-around capability.  The details are not given here,  but that ended up being a total of about 35% propellant fraction,  which then leaves 5% for payload,  at 60% inerts for a “super robust” airframe structure.  This represents a very conservative design. 

The target orbit was circular at 600 km altitude.  I assumed any circularization burn would require the same delta-V as the de-orbit burn.  This was 13% propellant fraction,  the other 22% being the “go-around kitty” for landing.  For a surface-grazing “transfer” return orbit,  I computed entry interface conditions at 140 km altitude as 7.742 km/sec and 2.35 degrees depression below local horizontal.    

I used 180 pounds per person,  and 180 pounds per pressure suit,  plus 40 pounds of oxygen and drinking-water consumables per person for a short flight of a few hours.  That’s 400 pounds per person.  Assuming 4 people,  or fewer if trading people for cargo,  that’s 1600 pounds payload.  Thus I obtained 32,000 pounds ignition weight for the plane,  as it is to be launched by commercial rocket booster such as Atlas-V or Delta-IV.  It’s a bit heavy for Falcon-9,  as currently sized,  but this is very preliminary. 

I roughed-out the fuselage,  wing,  and tail dimensions from these numbers,  worked out a full weight statement as a rocket vehicle,  the wing loadings,  and the ballistic coefficient.  These numbers correspond to “Spaceplane1”,  for which I simply guessed a radius of curvature in the cross section plane of 20 feet (6.1 meters).  See Figure C,  (which also covers “Spaceplane2”,  which design differs only in the belly radius). 

Figure C – Basic Configuration of “Spaceplane1” and “Spaceplane2”

I then ran my back-of-the-envelope (BOE) entry and heating spreadsheet,  to estimate the entry dynamics and stagnation heating loads,  using methods that trace ultimately to the early warhead work of H. Julian Allen,  circa 1953.  I also ran a comparison to Apollo entering from the same orbit.  Apollo has a much higher ballistic coefficient and a somewhat-smaller heat shield radius. 

Apollo showed peak stagnation heating 55.8 W/sq.cm at 55 km altitude,  peak deceleration 6.2 gees at 45 km altitude,  and end-of-hypersonics (at Mach 3) at 37.2 km altitude.  Its ballistic coefficient is about 313.5 kg/sq.m,  and the radius was 4.59 m.  Its ablative heat shield was suitable for the far more demanding entry returning from the moon.  The final altitude at end-of-hypersonics was more than sufficient to slow the capsule by a sequence of drogue and main chutes to a very survivable ocean impact speed in the vicinity of 20 mph. 

“Spaceplane1” has a ballistic coefficient of 196.4 kg/sq.m and a belly radius of 6.1 m.  It showed peak stagnation heating 38.1 W/sq.cm at 60 km altitude (substantially lower heating than Apollo),  peak deceleration 6.1 gees at 50 km altitude,  and came out of hypersonics (Mach 3) at 40.6 km altitude.  An ablative heat shield should be quite feasible,  and the new lower-density PICA-X is very attractive for this application.  Based on the Apollo experience,  there is plenty of altitude for the chute sequence to bring the craft well subsonic,  easily enabling unfolding of the wings and tails. 

One should note that running the BOE entry analysis varying only the heat shield radius produces exactly the same dynamics results.  Only the stagnation peak heating rates (and integrated heat absorbed) vary. 

Reusable Heat Protection?

There are two questions I wanted to resolve:  (1) Could a low-density ceramic refractory (such as Shuttle tile) be made to work as the belly heat shield?  (2) Could a structurally-stronger ceramic-composite version of this material (what I call the “Warm Brick” combustor insulator) be used instead?  In either version of the material,  the mechanism of function is re-radiation of the entry heating as a heat balance per unit area of applied convection and re-radiation at the skin temperature which balances. See also the 3-18-13 article on ceramic heatshields for more details. 

Surface emissivity becomes quite important.  The belly and under-wing tiles on the Space Shuttle were black for that very reason.  Elsewhere on the airframe where the heating was far lower,  the tiles were white.  With these alumino-silicate materials,  especially in the natural white-in-color form,  this surface emissivity is not “gray”,  it is strongly spectrally-variable,  but the details of that are beyond scope in this feasibility study.  Conservative gross averages are used herein. 

The true material limit is maximum allowable skin temperature.  It is the same 2300-to-2350 degrees F for Shuttle tile as it is for the “Warm Brick” composite:  both are the same alumino-silicates.  This is far less than the typical meltpoint of 3200-3250 degrees F.  It is set by the structural degradation induced by a solid phase change,  not by actual melting. 

I created and ran a spreadsheet-calculation that balances entry convective heat flux to re-radiation by the emissivity-corrected Stefan-Boltzmann law (absolute temperature to the 4^th power).  At the “Spaceplane1” peak heat flux and a “black surface” average emissivity of about 0.80,  skin temperatures were too high,  above the phase-change limit.  That ruled out even black-surfaced Shuttle tile. 

But,  I increased the belly radius of the spaceplane to a minimum of about 14.2 meters,  making it effectively flatter (more blunt),  and achieved success at lowering skin temperatures to acceptable values at the stagnation point (or line).   The maximum tolerable peak stagnation heating rate was 25 W/sq.cm.  In the radiation balance spreadsheet at 0.80 emissivity,  skin temperature was reduced to just the max allowable at 2300 F.  Black Shuttle tiles would work.  The “Warm Brick” material could work if it had a blackened,  highly-emissive surface. 

This result makes a fully reusable form of the folding-wing spaceplane concept technologically feasible with a non-ablative heat shield.  That raises the possibility of many repeated flights without periodic maintenance,  even of the heat shield (barring damage).   Dynamically speaking from the BOE analysis,  the peak deceleration experienced by the crew is 6.1 gees,  “eyeballs-down” unless the seats recline,  for no more than about 60 seconds above 5 gees.   All of this is quite feasible,  even for civilian passengers.

“Spaceplane2”

For “Spaceplane2” there are no changes to any of the shapes and dimensions in Figure C,  except the belly radius,  which is not shown.  (“Spaceplane1” had 6.1 meters,  that was increased to 14.2 m in “Spaceplane2” to get the peak heating down to 25 W/sq.cm.)    There are no changes to any of the mass,  ballistic,  and aerodynamic properties in Table 1.  The results of the final BOE entry analysis are given in Table 2 and Figures 1 through 5.  The chute descent sequence is still just as depicted in Figure B,  leading to a chute-retarded subsonic unfold.  Table 3 shows the heat balance and skin temperature for the stagnation line along the belly.   Weight statement is in Figure 6.  

Table 2 – BOE Entry Results for “Spaceplane2”

Figure 1 – Entry Decelerations-Down-Trajectory for “Spaceplane2”

Figure 2 – Velocity-Down-Trajectory for “Spaceplane2”

Figure 3 – Range and Slant Range Versus Altitude for “Spaceplane2”

Figure 4 – Stagnation Heating Quantities-Down-Trajectory for “Spaceplane2”

Figure 5 – Velocity and Deceleration-Versus-Time Traces for “Spaceplane2”

Table 3 – Heat Shield Heat Balance Data at Entry Peak Heating Rate

Figure 6 – Weight Statement and Related Data for “Spaceplane1” and “Spaceplane2”

The vehicle then flies as a simple straight-wing,  entirely-subsonic airplane glider to its landing field.  It is capable of a very slow landing speed,  in the vicinity of  70 knots at sea level.  The go-around scenario is as depicted in Figure D. 

The orbit and entry conditions are depicted in Figure E.  These were merely assumed, not optimized.  For reducing peak heating,  the keys items are very large (blunt) “nose” radius,  low ballistic coefficient,  shallow entry angle,  and low velocity at entry interface.  This study only looked at belly radius,  with fixed ballistic coefficient,  and fixed entry speed,   and fixed entry angle.  As the design matures,  ballistic coefficient will probably adjust slightly.  There is also opportunity to look at a different entry trajectory that might be less demanding.  Somewhat larger vehicles might be possible that way. 

Figure D – Go-Around Scenario for “Spaceplane1” and “Spaceplane2”

Figure E – Orbit and Descent to Entry Data (not drawn to scale!)

Ascent abort conditions would be Mach 2 or less,  with the higher speed near 80,000 feet altitude,  and air loads dropping rapidly due to low density.  The boosted ascent is made with wings and tails open in the flying position, so that the on-board rocket provides ascent abort capability from the pad to beyond the sensible edge of the atmosphere.  Even with straight wings and subsonic airfoils,  this kind of flight is feasible (as in the X-1). 

The straight wings with fat blunt subsonic airfoils are merely high-drag items,  providing much-needed inherent deceleration for the abort.  It is likely that abort air loads will structurally size the wings,  very likely at the max-q point near Mach 1 on ascent,  somewhere near 30 or 40,000 feet.   Abort will also likely size installed engine thrust at something near twice the weight.  Thrust to weight 33% (as in most airliners) is all that is needed for go-around,  leading to an estimated engine turn-down ratio of 6:1.  All of this should be easily achievable with existing technologies.

Other Applications

There is an inherent design limitation to smaller vehicle sizes with this ceramic heat shield approach,  because of the way ballistic coefficient increases as vehicle size increases,  all else equal.  Higher ballistic coefficients increase peak stagnation heating,  and there are practical limits to how flat the belly can really be.  I doubt that this folding-wing approach can be used with a ceramic heat shield at Shuttle-like sizes.  Big vehicles will require ablatives like PICA-X,  in turn requiring periodic heat shield replacement. 

                Two-Stage Horizontal-Takeoff Aircraft-to-Orbit

This kind of folding straight-wing design is not practical for the two-stage airplane-to-orbit launch method.  That orbiter will see mildly-hypersonic flow while attached to the booster aircraft during ascent,  and would therefore need swept wings,  especially during an abort.  However,  the very same folding-wing approach might still be used for entry dead-broadside belly-first. 

The same ceramic heat shield might prove feasible in the same way as described for this study.  The same chute sequence could be used to unfold subsonically at limited air loads,  and the same glide-to-landing could be done,  with the same go-around capability.  With swept wings,  approach handling is more difficult,  and the landing speed is somewhat higher.  But this should still be quite feasible.

                Other Boosters Such As Falcon-9

The gross weight in this study of 32,000 pounds (16 US tons,  14.5 metric tons) is within the capability of Atlas-V and Delta-IV,  as I have rough-sized this concept.  It is a little too heavy for Falcon-9 at 10-to-13 metric tons published capability to LEO out of Canaveral.  But,  design refinement might well shrink that gross weight somewhat,  for the same payload (or more).  Making the low-density ceramic heat shield feasible for this application offers a very much lighter heat shield,  for a significant inert weight savings. 

The actual detailed structural sizeouts (not done here) could reduce that very conservative 60% inert weight fraction that I used.  Even a small change gets you well below 13 metric tons gross weight,  making Falcon-9  a viable candidate launcher.  I think this inert weight fraction reduction is a very likely outcome,  since the inert weight fraction of the X-15 was 40%,  and it was robust enough to fly many times.  I would venture a guess of 50% inerts in the final version of a design,  with a very long service life (tens of thousands of flights) and very little pre-launch maintenance (rather like an ordinary aircraft). 

The main expense of vertical launch would then be the purchase of the booster rocket launch services.

Using the Warm Brick Material?

Black-surface Shuttle tile is a well-proven way to incorporate a low-density ceramic refractory heat shield.  Because of the many differently-shaped individual tiles,  this was expensive to install and repair.  Because of the fragility of the low-density ceramic,  repair was often needed,  especially in a side-mounted ascent configuration where debris-shedding is a serious risk.  The material is simply a very porous,  open structure of alumino-silicate crystals,  hard-surfaced on the exposed face.  For the windward-side tiles on the Space Shuttle,  that hard surface ceramic coat had to be black for high emissivity and efficient heat re-radiation. 

The “Warm Brick” material is different in construction and properties,  and is very experimental.  I made the successful version of this stuff only once,  long ago in 1984,  for an application completely different from entry heat protection.  I used it for a low-conductivity insulator inside a subminiature,  subsonic-flight ramjet combustor.  That combustor was not propulsive,  but instead served as a hot gas source in a device known as “Warm Brick”,  the purpose and details of which are not germane here. 

                “Warm Brick” Device Combustor Application

I built two of these,  inspired by the low density ceramic tiles on the Space Shuttle.  The first was essentially a copy of the Shuttle material,  being merely a very porous (as cured) alumino-silicate potting compound purchased commercially.  I sealed its surface against gas permeability with a coat of ceramic cement from the same supplier.  Those specific materials are no longer available,  but that same supplier has equivalent materials available today (I checked recently). 

I did not have to develop any bonding,  since the insulator was simply (and very effectively) trapped within the combustor shell.  It was shattered to pieces by one instance of the severe pressure oscillations of rich blow-out instability in this combustor (amplitude near a dozen psi at audio frequencies of a few hundred Hertz,  shaking the device visibly upon its heavy steel test sting). 

For the second insulator,  I decided to build a fabric-reinforced ceramic composite material,  analogous in concept to an organic resin-fiberglass composite.  I used an alumino-silicate woven fire-curtain fabric common in the aviation industry.  That specific cloth material is no longer available,  but there is a modern equivalent today from the same supplier (I checked that,  too). 

That ceramic composite combustor liner was quite successful:  it withstood extreme instability oscillations,  and it protected the shell during continuous burns measured in hours.  The combustor shell never got any hotter than “barely able to boil spit”,  even immersed in a (measured) 190 F airstream. 

                Material Characterization of the Ceramic Composite

I never characterized the density of this tough ceramic composite,  but it should be no more than twice that of Shuttle tile,  and is very much more likely to be only around 30% higher than Shuttle tile density.  But those are just guesses.  My physical sense handling the pieces was “similar to Styrofoam”. The published density of Shuttle tiles is effective specific gravity 0.0144;  for a typical alumino-silicate mineral specific gravity of 3.2,  the mineral volume fraction is 0.0045,  and the void space volume fraction is 0.9955.  My guesses just above would indicate a "Warm Brick" material effective specific gravity between about 0.019 to 0.029,  corresponding to void fractions from 0.994 to 0.991.  But,  these are just guesses.  Styrofoam products vary in density, from 0.02 to 0.05 effective specific gravity.   -----  3-4-13 update

Back then,  I never determined the as-built thermal conductivity of the composite.  Recently,  I did find enough dimensional data on the old “Warm Brick” device,  so that I could analyze the heat protection after-the-fact,  for the longest-burn test that I remember.  To accomplish this,  I had to re-analyze the combustor,  inlet,  and nozzle as a “low-speed range,  pitot-inlet ramjet” to determine the internal hot gas properties and flow state.  That required a proper ramjet cycle analysis suited to dump-combustor flame stabilization.  I worked out a spreadsheet I could converge manually for this,  although it is quite a laborious process.    Figure I below is the spreadsheet setup,  Figure II is the converged analysis. 

For folks interested in practical engine cycle analysis of such ramjets,  I have not published any details,  but I am writing a book.  I did such work professionally for about 2 decades,  long ago,  as one of the industry leaders.  Please contact me if you need consulting help with ramjets.   I can handle both low-speed range and high-speed range designs,  a variety of types and species of fuels,  and a variety of inlet geometries.  Some of this is subject to the ITAR restrictions,  though. 

I used that internal flow state and the external slipstream conditions to set up a cylindrical-geometry heat transfer model.  It had film coefficients using well-accepted correlations inside and outside.  There were two conductive layers:  the steel shell and the low-density ceramic-composite liner.  I knew the thermal conductivity of the steel,  but not the ceramic.  I knew the inside and outside gas temperatures,  and a rough estimate of the shell temperature (215 F or so).  I also knew the ceramic inside was running near its meltpoint (meaning somewhere near 3000 F),  because there was a little surface melting evident in some tests,  just not this one.  In this particular application,  the phase-change-induced shrinkage,  embrittlement,  and surface cracking was not a problem.  So,  I could run a lot hotter.

I set the spreadsheet up as repeated calculations for a variety of ceramic conductivities from very low to rather high.  Then I picked the one calculation that best matched my shell-temperature and liner-near-melting observations. That spreadsheet is given as Figure III below.  The corresponding as-built thermal conductivity is 0.02 BTU/hr-ft-F (0.035 W/m-C or 0.00035 W/cm-C). 

Those figures are crude,  probably plus or minus at least 50%.  But it’s a good ballpark figure,  useful for protection applications by low thermal conduction.  That property does not apply to radiative entry protection,  but it does confirm the essential cut-off of all heat flow through the material during entry. 

That has implications for the integrated absorbed entry heat,  where heat capacity can be important.  Those phenomena were not even considered in the “Spaceplane1” / “Spaceplane2” study.  I never characterized the heat capacity of this material,  but it has to be very low,  because the density is low.  Volumetrically,  this stuff is mostly void space. 

From a manufacturing standpoint,  this material is amenable to construction in large bolt-on panels instead of small tiles.  Because of the layers of reinforcing fabric,  bonding can be avoided if desired,  even for external insulation applications.  A typical panel might comprise a backplate onto which the potting compound is troweled.  Over this,  a layer of the cloth is stretched,  bent around the edges,  and clamped on the backside.  This is heat-cured by water drive-off at about 215 F,  compatible even with aluminum backplates.  One builds and cures multiple layers to the desired thickness this way,  then surface coats the assembly with the hard ceramic cement layer,  and cures that in the same way. 

The one thing I never did,  because it was not needed in the “Warm Brick” device,  was to develop a black,  highly-emissive surface coating.  The normal color of these alumino-silicate potting and cement materials is a bright white.  From a nearly-closed combustor,  there is no escape of surface radiation,  and thus no measurable radiative cooling,  regardless of the spectral-averaged emissivity.  The situation is essentially a “reverberatory furnace” with strong gas flow throughput. 

I do think that mixing carbon black into the ceramic cement compound might be a practical way to achieve easily (and reasonably cheaply) the hard black emissive surface coat.  I have no idea how that might affect the cure or properties,  but I suspect it might work acceptably well. 

This stuff has real potential for both low-conductivity and emissive-re-radiation applications.  Lots of investigators and projects might find this technology useful.  There are no others in the world who have experience with materials quite like these.  If you need some consulting help in this area,  please contact me.  Again, in some cases the ITAR restrictions might apply. 

Figure I – Setup Data for Ramjet Cycle Analysis of the “Warm Brick” Device Combustor

Figure II – Cycle Analysis Results for the “Warm Brick” Device Combustor

Figure III – Thermal Model for the “Warm Brick” Combustor Insulation Sleeve