Heat Sink Heat Shields

(Henry Spencer) wrote in message ...
In article ,
John Carmack wrote:
If you are willing to trade mass for it, copper heat shields have been
shown to work, and they would be cheap and easy to fabricate...

A good point. Heat-sink heatshields have been out of fashion for a long
time because of their mass, but they're the only flight-proven TPS that's
both durable and fully reusable. (It's hard to imagine anything much more
durable than a thick slab of solid metal...)

Several questions...

*What re-entry vehicles demonstrated copper heat shields?

*How thick is "a thick slab of solid metal"? Aerospace definitions of
"thick slab" probably differ from, say, shipbuilding definitions of
"thick slab."

*For water-cooled heat shields, what percentage of the re-entry
vehicle's mass is typically needed as cooling water?

Mike Miller, Materials Engineer

In article ,
Mike Miller wrote:
A good point. Heat-sink heatshields have been out of fashion for a long
time because of their mass, but they're the only flight-proven TPS that's
both durable and fully reusable. (It's hard to imagine anything much more
durable than a thick slab of solid metal...)

*What re-entry vehicles demonstrated copper heat shields?

Early ICBM and IRBM warheads. The suborbital Mercury capsules also had
heat-sink heatshields, but using beryllium instead of copper.

*How thick is "a thick slab of solid metal"?

I don't have numbers handy, but think 10-20cm. Slab, not sheet.

*For water-cooled heat shields, what percentage of the re-entry
vehicle's mass is typically needed as cooling water?

Only a couple of percent, I think, but here my memory is quite vague.
(Again, references aren't handy.)
--
MOST launched 1015 EDT 30 June, separated 1046, | Henry Spencer
first ground-station pass 1651, all nominal! |

In the table below I multiply the specific heat by the
melting point to get a figure of merit I call "HeatSink Joules/Kg"
(should really have subtracted some starting temp like 20 C).
Note that a heatsink also has to have a high conductivity,
which rules out titanium. Beryllium looks far better than
the others. Copper is very conductive, but it stores less heat
than anything else on this list.

Is there any chance that the the US ICBMs with "copper heatsinks"
could have really been copper coated beryllium? Maybe they
coated it to reduce the danger of handling beryllium?
I have never seen details of ICBMs with heatsinks or transpiration.
I don't know of any non-military transpiration flights and suspect
there are none.

Beryllium has flown on at least John Glen's suborbital Mercury
flight. So for sure it can be done.

In my simulations a Beryllium heatsink looks like a fine reusable
heat shield for reentry from a 5 km/sec rotovator/space-tether.
I am simulating a 4 meter diameter capsule that weighs 4,000 Kg.
For this a heatsink of around 5% of the mass would be enough for
suborbital and around 15% for orbital. You can calculated the
thickness from the density below and the 4 meter diameter.
I would use a suborbital sized heatsink and water/transpiration
to handle the extra heat in the case of missing the LEO tether
on the way down and having to do a full orbital speed reentry.

Material Conductivity Density Specific Melting HeatSink
Heat Point
W/m-C kg/m3 J/kg-C C
Joules/Kg
Beryllium 175 1,859 1885 1278 2,409,030
Titanium 16 4,507 544 1668 907,392
Iron 80 7,874 449 1538 690,562
Lithium 85 535 3582 181 648,342
Aluminum 220 2,707 896 660 591,360
Tungsten 180 19,350 134 3422 458,548
Copper 386 8,954 380 1085 412,300

Ice Melting 333,000
Heating Water 100 C * 4184 J/Kg-C 418,400
100 C Water to Steam 2,500,000
Ice to Steam 3,251,400
If steam used in transpiration x4 13,005,600

Charing Ablative
Char radiation / vaporization / Transpiration very good

The only bad thing about a charing ablative is that it is
not testable/reusable.

Beryllium
Strong, very light, resistant to oxidization like aluminum
high melting point, very high specific heat
Used in aerospace
One of the lightest metals
Stronger than steel pound for pound
Brittle
Something like $160/lb or $350/Kg. About this all through 1990s.
So could afford for reusable vehicle.
Berylliosis
Breathing fumes or dust, or getting them on open cut.
DOE has worker standards. Machining can expose worker to risk.
Solid it is not a health hazard.
In 1998 US consumed 240 tons and exported 60 tons.
Brush Wellman Inc is only US ore processor. Has 60 years reserve.
Primary processor for world.

Some sources for some of the above info:
http://www.arkthermal.com/metals2.doc.

Conductivity, Density, Melting Point:
http://www.webelements.com/

Specific heats:
http://www.allmeasures.com/Formulae/...cific_heat_cap acity_300K/

-- Vince

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~
Vincent Cate Space Tether Enthusiast
http://spacetethers.com/
Anguilla, East Caribbean http://offshore.ai/vince
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~

You have to take life as it happens, but you should try to make it
happen the way you want to take it. - German Proverb

In article ,
Vincent Cate wrote:
In the table below I multiply the specific heat by the
melting point to get a figure of merit I call "HeatSink Joules/Kg"...
Note that a heatsink also has to have a high conductivity...

Conductivity is very important, because the heatsink surface must not
melt. Copper wins big there, and a bunch of otherwise-attractive metals
flunk completely. High-temperature oxidation resistance is also
significant. According to the old books, copper and beryllium are the
only heatsink materials that looked useful.

Is there any chance that the the US ICBMs with "copper heatsinks"
could have really been copper coated beryllium?

Nope, straight copper. Heavy, yes, but cheap, easily fabricated, and
mechanically durable. (Whereas beryllium, although light, is costly,
very difficult to work with, and brittle.)

Remember that those warhead designs were done in desperate haste, to get
*something* operational ASAP. My reading of the history (from limited
information, mind you) is that they might well have gone to beryllium for
a second-generation design, except that they went to ablators instead.

Beryllium has flown on at least John Glen's suborbital Mercury
flight. So for sure it can be done.

Glenn never made a suborbital Mercury flight. Shepard and Grissom flew on
Mercury's original beryllium heatsink heatshield. The ablative design was
ready in time for the orbital flights, and was deemed superior. (I think
they did qualify the heatsink design for orbital flight, at least on paper.)

Beryllium ...
Used in aerospace

Even aerospace use is declining, due to practical hassles and competition
from carbon composites.

Stronger than steel pound for pound
Brittle

The brittleness is not only a problem for the final structure, but greatly
complicates machining etc. It's inherent in the crystal structure and is
not fixable (this was studied in great depth), although with considerable
difficulty you can make beryllium that is ductile in two dimensions and
only brittle in the third. The brittleness makes the practical strength
much less than theoretical values in most applications, because you must
design very conservatively to avoid local stress concentrations that would
be of no importance with a more ductile metal.

Its one big advantage is something that actually isn't in your list:
stiffness. Not how much load it will take before breaking, but how much
it will resist flexing under lesser loads. In particular, specific
stiffness -- stiffness per kilogram -- does not vary a lot between metals,
except that beryllium is way out in front of everything else. Until
carbon composites came along, that is.

Berylliosis
Breathing fumes or dust, or getting them on open cut.
DOE has worker standards. Machining can expose worker to risk.

And that too complicates working with it.

Wild idea of the week: I wonder if you could take a leaf from Apollo's
book, and make a heatsink heatshield out of hexagonal beryllium rods in a
copper or stainless-steel honeycomb? The honeycomb would take mechanical
loads and hold the beryllium together, eliminating brittleness issues,
while the beryllium handled most of the heat.

One book, interestingly enough, mentions the idea of adding expendable
(perhaps liquid) coolant behind a heatsink heatshield, but says the idea
was not pursued, because straight heatsinks seemed adequate for satellite
applications, while nothing short of ablators would do for the most
demanding warhead flight profiles.
--
MOST launched 1015 EDT 30 June, separated 1046, | Henry Spencer
first ground-station pass 1651, all nominal! |

Henry Spencer wrote

In article ,
Vincent Cate wrote:
In the table below I multiply the specific heat by the
melting point to get a figure of merit I call "HeatSink Joules/Kg"...
Note that a heatsink also has to have a high conductivity...

Conductivity is very important, because the heatsink surface must not
melt. Copper wins big there, and a bunch of otherwise-attractive metals
flunk completely. High-temperature oxidation resistance is also
significant. According to the old books, copper and beryllium are the
only heatsink materials that looked useful.


Diamond oxidises, like copper and beryllium, but it has 6 times the thermal
conductivity of copper. And a much higher melting point.

Expensive though...


--
Peter Fairbrother

(Henry Spencer) wrote in message ...

Its one big advantage is something that actually isn't in your list:
stiffness. Not how much load it will take before breaking, but how much
it will resist flexing under lesser loads. In particular, specific
stiffness -- stiffness per kilogram -- does not vary a lot between metals,
except that beryllium is way out in front of everything else. Until
carbon composites came along, that is.

Some heavy metals exceed beryllium for stiffness. Beryllium has
a 44MSi stiffness (vs 29-30 for steel, 16ish for titanium, 10 for
aluminum), but tungsten is up to 58Msi, and I think rhenium and
osmium are stiffer.

Of course, there's that whole density angle that makes beryllium
so special for its stiffness.

One book, interestingly enough, mentions the idea of adding expendable
(perhaps liquid) coolant behind a heatsink heatshield,

Mr. Spencer, if you ever do get to a place where you have reference
books handy, I'd love to see more precise estimates for the mass
fraction of water in transpiration cooling.

If I understood the exchange between you and Mr. Carmack, were you
saying that if you were willing to accept the higher mass and made
a Cu or Al transpiration heat shield thicker, you could use larger,
more manageable pore sizes? Would the larger pores reduce the
insulation efficiency of the vented steam outside the shield?

Mike Miller, Materials Engineer

In article ,
Peter Fairbrother wrote:
Conductivity is very important, because the heatsink surface must not
melt. Copper wins big there, and a bunch of otherwise-attractive metals
flunk completely. High-temperature oxidation resistance is also
significant...

Diamond oxidises, like copper and beryllium, but it has 6 times the thermal
conductivity of copper. And a much higher melting point.

Diamond doesn't really have a melting point. If you get it hot enough, it
reverts to graphite -- diamond is only metastable. The process starts as
low as 1000degC.

And in an oxygen-containing atmosphere, the oxidation rate becomes
significant even before that. Copper and beryllium oxidize, yes, but the
result is a durable surface layer of solid oxide. But diamond oxidizes
to CO2...
--
MOST launched 1015 EDT 30 June, separated 1046, | Henry Spencer
first ground-station pass 1651, all nominal! |

Peter Fairbrother wrote:
Henry Spencer wrote


In article ,
Vincent Cate wrote:

In the table below I multiply the specific heat by the
melting point to get a figure of merit I call "HeatSink Joules/Kg"...
Note that a heatsink also has to have a high conductivity...

Conductivity is very important, because the heatsink surface must not
melt. Copper wins big there, and a bunch of otherwise-attractive metals
flunk completely. High-temperature oxidation resistance is also
significant. According to the old books, copper and beryllium are the
only heatsink materials that looked useful.



Diamond oxidises, like copper and beryllium, but it has 6 times the thermal
conductivity of copper. And a much higher melting point.

Expensive though...





Diamond burns, is a somewhat more accurate statement. A lot of
jewelers have found that out the hard way when trying to cast rings
around a diamond inset. There isn't a large separation between solid
diamond and carbon dioxide in certain operational environments.

(Mike Miller) writes:

But brittle, brittle, brittle. Worse than refractory metals,
and much worse than some alloys like W-27Re. Suddenly, I
like Mr. Spencer's idea for a beryllium-copper composite.

Haven't Beryllium-copper and beryllium-aluminum alloys also been proposed ???


-- Gordon D. Pusch

perl -e '$_ = \n"; s/NO\.//; s/SPAM\.//; print;'

(Mike Miller) wrote in message
. com...
This link discusses early use of copper heat shields for ICBMs:

http://www.centennialofflight.gov/es...try/Tech19.htm

Thanks!

For this a heatsink of around 5% of the mass would be enough for
suborbital and around 15% for orbital.

I tried to post a correction but the post does not seem to have
worked. Im my simulation the mass for orbital heatsink is
about 25%, not 15%. See sample inputs 51 and 51 in my simulator at:

http://spacetethers.com/spacetethers.html

This simulator has not been tested against any experimental
heatsink data because I have none (I will be reading that link
next). So there is the very real chance that it has bugs.
If anyone has any real numbers please let me know what they are.

In particular, if Beryllium has to be 25% of the mass, then
copper would have to be more than 100%. Part of this is
that in my simulations I had a L/D of 0.4 and the ICBMs
had 0. I have not re-run the simulation with 0 because my
home computer is a text only Linux box. Plan to tomorrow.
But this could explain the ICBMs. Also, I don't know how
close to orbital speed the copper heatsink ICBMs got.
It is far easier at 5 km/sec than at 7.7 km/sec.

For a capsule with humans you really do want some lift
though. It reduces the peak G load. You can see this
in my simulator.

But brittle, brittle, brittle. Worse than refractory metals,
and much worse than some alloys like W-27Re. Suddenly, I
like Mr. Spencer's idea for a beryllium-copper composite.

It does seem like there should be some way to reenforce it.

So, a beryllium heat sink would represent 15% of the mass
of an orbital capsule. Questions:

25% for my simulation (sorry).

For the same re-entry velocity, would the mass of the heat sink
heat shield vary for different capsule shapes (biconic vs raked
cone vs Soyuz) and/or different different re-entry paths
(capsule-type lifting vs. ballistic)?

The more lift you have the more total heat. You can look
at this as the longer trajectory means more total heat.
Also, if you are getting lift then you are angling your
heatshield to the flow air and the shockwave is not going
to be as far away from the capsule, so more heat.

The capsule has a certain amount of energy given by 1/2 MV^2
that is going to turn into heat. The big question is how much
of that heat goes into the air and how much goes into the
capsule. The "stanton number" tells you what portion goes
into the capsule. For blunt bodies it is really very small,
numbers like 0.1%. There is a formula on page 256 of
Hypersonic Aerotherodynamics to estimate how much heat the
capsule will get and I use it in my simulator (if you don't
specify a stanton number in the input). Here is that part
of the code from mass.java:

blackBodyRadius = Math.sqrt(blackBodyArea/k.pi);
heatRatePerCC = 18300.0 * Math.pow(ourAir.density, 0.5)
* Math.pow(airRelativeVelocity.magnitude()/10000.0, 3.05)
/ Math.sqrt(blackBodyRadius);
heatFromAir = k.timePerDisplay * heatRatePerCC
* k.SqCMinSqMeter * blackBodyArea;

All of the code is available at http://spacetethers.com/source/

If I read this correctly, could you replace 5.4kg of Be with
1kg of water with transpiration cooling (neglecting the mass
of the metallic portion of a transpiration heat shield)?

If so, would that mean a water transpiration heat shield would
be about 2-4% of an orbital capsule's mass?

Yes except that since I should have said 25%, so it is more like 5%,
neglecting the metallic portion.

Another note. The specific heat of materials changes over
temperature, so just taking the value at 300 K and
multiplying it by the melting point like I did is only
an approximation. I am not sure how bad it is.

-- Vince