New Mars Forums

I’m unsure where to post this.  It and some of the preceding posts may well belong somewhere in the new SSTO topic. 

Composite materials are tricky,  and the advertising hype about them is very,  very misleading. 

The advertised high strengths are always ONLY in the fiber lay direction!  Strength in the other directions is very much lower,  and can be near-zero out of the fiber plane,  depending essentially only upon the matrix properties.  You have to “play God” and know all the necessary fiber lay directions to handle all the loads your structure might ever encounter.  The odds of actually doing that successfully,  first time up in a new design,  are actually quite low.

These composite materials have strength only in the direction of the fiber lay,  and even then ONLY so long as the matrix holds them in place!  If the matrix fails for any reason at all,  your material essentially turns into loose rags or threads or yarns.  You can get strength in two directions,  for some variable-with-angle level of strength in-plane,  by using real woven cloth as your fiber. 

The strength between two layers of fiber laid in different directions,  or two layers of woven cloth oriented differently,  is ONLY that of the matrix gluing them together,  which is usually down in the 5-10 ksi range,  and even then usually only by shear!  If your stresses are 3-D,  the material will fail at very low stress levels,  in precisely the direction where fibers do not lay.  It is inherent.

There is no practical way around that,  other than fully-3-D fiber preforms,  the manufacture of which is extremely difficult and horribly expensive.   And the infiltration of those 3-D preforms with a matrix is also extremely difficult and horribly expensive!  This is usually only done with metal matrix composites.

The tightly-woven tapes in particular are very strong in the long direction,  when used in composite work.  But if you try to use that by itself,  in multiple layers of spiral wrap in alternating directions to make a cylindrical vessel,  the only thing connecting one strip of tape to the next is matrix shear,  with a low shear-bearing area from one tape strip to the next tape strip!  Pressure test results of the vessel will be extremely disappointing!  You might have considerable hoop strength,  but you will inherently have almost zero longitudinal strength,  and you will have essentially no bending strength as a tubular bending-loads-resisting structure (as in a cylindrical tank that is also an airframe component).  These are best made with a few layers of woven fabrics,  or many,  many layers of yarn spin-wrapped in multiple different directions.

You cannot use composites,  especially woven composites,  in arbitrarily thin layers!  The thinnest thickness of a one-layer woven cloth composite is the thickness of that cloth layer.  Period.  Set by the threads or yarns from which it was woven,  but a bit larger due to the weave geometry.  You are extremely unlikely to get the strengths you need,  unless there are actually multiple layers with different fiber lay directions!  This stuff simply cannot be arbitrarily thin!  Formable metals can be arbitrarily thin,  except that alpha-phase titanium cannot,  only beta phase titanium can be formed into thin sheet (which ages at room temperature into uselessness,  with big,  weakly-bound grains).

The ceramic and carbon fibers that are now popular are indeed “good” for substantial exposure temperatures,  by themselves.  Generally speaking,  the organic polymer matrices you must use with them are not!  Most of the hydrocarbon-type polymers start charring at around 300-400 F,  and even the silicones start charring at about 600 F. 

Everything organic or silicone is fully carbonized by no more than 1000 F,  into something soft and crumbly,  resembling the charcoal in your BBQ grill,  except that it may still be reinforced by the fibers,  but only if those fibers were ceramic or carbon.  That is not a particularly reusable material for heat exposures,  but if the charcoal “hangs together” better with the fiber reinforcement,  the hype can truthfully claim their material “survived” the exposure,  as long as NOT A WORD was said about ever using it again! 

None of the polymers,  or other plastics that I have ever heard of,  have more than utterly-trivial elongation capability soaked out cryogenically cold (way under 1%).  Using such materials in composite cryogenic propellant tanks thus presents two extremely serious problems:  (1) very brittle behavior conferring extreme fragility in any kind of handling,  and (2) leakage of propellant through the inherent porosity of the composite,  unless you can successfully add some sort of internal sealing layer (not a trivial exercise in and of itself). 

Taken together,  those cryo-cold fragility troubles,  the weight of multiple fabric layers to get strength in more than one direction,  and the inability to sustain high temperature exposures,  plus the manufacturing costs,  are EXACTLY why SpaceX abandoned its original composites-construction concept for its Starship,  in favor of simply welding-up 300-series stainless steel sheets! 

Not having to put a heat shield on the lee side surfaces with stainless steel saves a great deal of heat shield weight,  and also saves some rather large manufacturing costs,  which made stainless steel the far better choice for a stage that must also qualify as a re-entry vehicle from orbit,  plus endure descent and landing loads.  You can take 304/304L SS to 1200 F as many times as you like,  although it will start corroding and scaling if you let it get any hotter than that.  Nothing re-radiates efficiently enough for self-cooling until it reaches temperatures in that 1000-1200 F class.  Organic-matrix composites never will survive that exposure,  and still be usable a second time.

There is also the issue of stiffness,  which is partly the moment of inertia of the geometry,  and partly the modulus of elasticity of the material.  Steels have a modulus of elasticity in the 30 million psi range.  Aluminum is about 10 million psi.  Carbon-epoxy gets close to aluminum,  and the other things like glass-vinyl ester (or glass-polyester) are far less stiff than that.  Even Kevlar-vinyl ester,  while quite tough against impacts and punctures,  is far less stiff.  Depending upon what the structure has to do,  the achievable stiffness may (or may not) be crucial.  But carbon-epoxy (and the others I mentioned) is limited to under about 300 F exposure,  if the epoxy matrix is to survive in a reusable condition!

And then there is detection of impact damage,  to which any composite is critically vulnerable.  Unless of catastrophic magnitude,  it’s often hidden damage,  you see nothing on the surface.  Ignoring the advertising hype otherwise,  the only way known to reliably detect hidden damage is by comparing before and after x-ray images,  looking for differences.  There has to be one taken before the impact,  and the other after,  for this to work.

Every square inch of the composite has to be in an as-built x-ray somewhere,  so that after a suspected impact,  a post-impact x-ray can be taken for direct comparison to the as-built x-ray.  The advertising hype is wrong,  there are NO general criteria for evaluating post-impact x-rays only!  Nothing is reliable enough to permit that!  These materials are simply too variable from one square inch to another,  to permit that sort of thing.  This requirement for as-built x-rays during manufacture acts to further drive-up costs.

There is also the question of joints.  Everyone is familiar with the way the fiberglass fenders on a Corvette eventually tear out around the machine screws holding them in place.  You simply cannot attach composite parts to other structures that way,  it is not reliable,  and it won’t last,  especially if highly-loaded.  Fiberglass or something else,  it does not matter!  Composites and conventional fasteners are simply incompatible.

What is required is a hybrid of alternating layers of the composite and shim-stock thin metal,  with enough layer bond area to carry the loads by only matrix shear,  from the metal layers holding the fasteners,  into the composite layers that extend into the rest of the composite structure.  There must be enough layers of the metal shim material to take the fastener loads as bearing loads without failing,  which in turn means there must also be many layers of composite,  in the region of the joint. 

That’s very labor intensive (and therefore expensive) to accomplish,  which is EXACTLY why I do not trust Boeing to have done that job correctly on the B-787.  Neither do I trust Airbus to have done this job correctly,  not after that Airbus lost its composite vertical fin and crashed in NYC many years ago.  The fin tore out around the bolts holding it to the aluminum airframe (it was not a hybrid joint,  surprise,  surprise). But this hybrid technique worked like a charm with composite solid rocket motor cases and metal end closures,  up to 4-5000 psi,  as tested and well-proven.

If you get the idea that designing and building with these materials is difficult,  and fraught with fatal pitfalls,  then you understand the main point here.  If you do this job wrong,  you will achieve better weight reductions and more cost savings (although still more expensive than metal construction).  However,  the product WILL INEVITABLY FAIL unexpectedly in service,  causing very serious and expensive legal issues of the “whose fault was this?” type.  Such can be quite catastrophic. 

On the other hand,  if you do the design correctly,  and avoid those unexpected failures in service,  your product will cost even more,  and it will save only a little weight,  compared to equivalent metal construction.  Top managers hate that.  Rightly or wrongly (wrongly in my opinion),  they tend to discount the legal troubles for unexpected failures in service,  and will push you very hard to do the job wrong.  (Which in turn is partly why the advertising hype lies so egregiously.) 

I’m probably obsolete and out-of-date,  regarding all the latest composite materials.  But none of these older ones (or the new ones I am not familiar with) will have any different extreme directionality of their properties,  or any different severe limitations on exposure temperatures for their organic matrices (including the silicones).  And the metal matrix composites are still horribly expensive,  and probably always will be.   As well as exceedingly difficult to manufacture.

There are no “magic” materials!

GW