SpaceX says Crew Dragon capsule exploded due to exotic titanium fire
teslarati.com
teslarati.com
"The following metals have been found to be incompatible with nitrogen tetroxide and must not be used: Aluminum 2024, Zinc, Aluminum 7075, Silver, K-Monel, Titanium, Brass Cadmium, Bronze Hastelloy, Copper, EZ Flow 45 Braze."[0]
[0]: Materials Compatibility with Liquid Rocket Propellants, https://apps.dtic.mil/dtic/tr/fulltext/u2/866010.pdf
>SpaceX: "It is worth noting that the reaction between titanium and NTO at high pressure was not expected."
SpaceX did not expect a reaction known to Boeing almost 50 years ago. My guess would be that this is actually known by a lot of people in rocketry, since UDMH/NTO is a standard propellant - the Proton is entirely fueled with it.
I also guess that the titanium component of the valve is not normally exposed to the oxidizer, and only came in contact with it when the valve broke. It would seem that the simple issue of shattered valve fragments flying down the oxidizer line would have been sufficient to destroy the thruster irrespective of a titanium fire...
But the same report also notes:
"Nitrogen tetroxide with a water content in excess of 2.5 percent can accelerate titanium intergranular corrosion to the point of violent reaction."
I think the assumption is that by the time a RUD starts you no longer care about whether the various parts of your fuel system are going to explode, since the fuel is a far larger concern and it's going to go up anyways.
There’s really no point in designing additional failsafes into a launch vehicle when it disintegrates which can happen for a variety of reasons beyond anyone’s control — RUD is really just an euphemism for an explosion.
It doesn't mean there are no failsafes against RUDs or to protect the crew and equipment in case a RUD occurs.
While as a ceo and "quick drafting solutions" that works. But when it comes to stuff the like this. You get this. The reaction is known for 59 years. And they still did it..
Titanium is practically the only metal that can store NTO for long periods of time over a reasonably large temperature range. This does come at the expense of shock sensitivity (particularly after a soak/adsorb), but the industry has to use it anyway, while working to mitigate the shock problem.
[0] https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/197200...
Expecting them to see the 'water hammer' issue is a different story
The implication being that there's just no way SpaceX could have suspected this problem because structural titanium use in a spacecraft would inform decisions regarding its use in contact with corrosive propellant.
Obviously, this was written by a PR person, not engineering management, but it plays to an image of SpaceX being a company to handwave past the hard-earned lessons of decades of "Old Aerospace" companies. They should be more careful in releasing these kinds of statements.
And sometimes you think you're mixing "milk and cereal" and it catches fire and you don't know why.
This paragraph really drove the point home for me. I was able to easily visualize the event.
Bismuth(III) oxide, boron(III) oxide, silicon(IV) oxide, chromium(III) oxide, manganese(IV) oxide, iron(III) oxide, iron(II,III) oxide, copper(II) oxide, and lead(II,IV) oxide.
"The following metals have been found to be incompatible with nitrogen tetroxide and must not be used: Aluminum 2024, Zinc, Aluminum 7075, Silver, K-Monel, Titanium, Brass Cadmium, Bronze Hastelloy, Copper, EZ Flow 45 Braze."[0]
"Titanium is practically the only metal that can store NTO for long periods of time over a reasonably large temperature range." [1]
[0]: Materials Compatibility with Liquid Rocket Propellants, https://apps.dtic.mil/dtic/tr/fulltext/u2/866010.pdf
[1]: https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/197200...
For example, from [0] (which is declassified from 1974!) :
"Titanium can react violently with strong oxidizers if the oxi-dizer penetrates the oxidized surface layer."
"Free chloride, ions in nitrogen tetroxide can cause an accelera-tion of intergranular corrosion in titanium."
The ancient Boeing report is quoted at length as gospel but the reasoning is not explained. Here's some weird facts:
1) Most metals are not happy with NTO because any moisture turns NTO into concentrated nitric acid and nitric eats lots of metals.
2) Some industrial nitric acid plant components are made of very expensive titanium because its strong and essentially nitric acid corrosion proof. This makes Ti sound fantastic for NTO... however...
3) It is true that titanium is one of the few metals you can dunk in seawater where stress-corrosion does NOT occur. Which makes people even more hyper optimistic about Ti for NTO tanks as per above... however...
4) Unfortunately titanium DOES stress corrode in NTO. Stress corrosion is where essentially microcracks that would normally heal in other metals instead get a higher volume "rust" oxidation in the crack slowly breaking the metal apart after thousands or millions of cycles even if an unstressed part would never wear. Think of a perfectly painted garage that would never have wood rot due the the paint protecting the wood, but if you bend the wood back and forth tiny cracks will form in the paint and the water will soak thru the cracks into the wood and the garage wood will rot eventually even if the paint was theoretically perfect under unstressed conditions. Its actually a little more complicated than this.
So from an engineering standpoint if you build something under millions of cycles of repetitive stress like the shaft of a circulation pump in an industrial production NTO tank out of titanium, you're doomed to have it shear in a "sooner than economically viable" amount of time like a couple months of continuous use maybe. On the other hand if you build a valve that will only exist for a little while and only be under a couple stress cycles in its lifetime and only possibly operate once for a few seconds, thats very likely OK.
There is also a systems issue where if a life-critical failure occurs upstream resulting in everyone being dead already, it doesn't matter if downstream a system that should never be contaminated fails and burns. Or rephrased if enough stuff upstream broke such that the helium system literally caught fire, the crew are already dead men and the ships already lost so it doesn't really matter much anymore if the helium system burns or does not burn at that point. I mean its technically interesting but if you're gonna lose the vehicle and pax it doesn't matter really in the long run. Kinda like the old line about how we could build jet liners to the same standards as black boxes, but the external G forces of the crash would kill the passengers inside the giant black box so realistically why bother. If something upstream fails badly enough that the helium system is compromised, they're already dead anyway so engineering a theoretically indestructible helium system is a waste of time if the mission, ship, and passengers are already lost, so may as well save the weight and labor.
A burst disk comes from a different concept than a check valve. With a check valve, one presumably desires a fluid flowing in one direction—in this case the pressurant for the oxidizer vessel—and no back flow. In the rocket engine industry this setup tends to be used to maintain cleanliness. A burst disk, however, embodies the same concept as a relief valve—it prevents the pressurization of a system above its rated capacity. I fail to see how replacing a check valve with a burst disk fixes the problem and maintains the same system capability.
0: https://www.spacex.com/news/2019/07/15/update-flight-abort-s...
[0] https://www.amazon.com/Some-birds-dont-Gary-Paulsen/dp/B0006...
0: https://www.amazon.com/Ignition-informal-history-liquid-prop...
EDIT: archive.org appears to have it: https://archive.org/details/somebirdsdontfly00paul
On December 29, 1953, a technician at Edwards Air Force Base was examining a set of titanium samples immersed in RFNA, when, absolutely without warning, one or more of them detonated, smashing him up, spraying him with acid and flying glass, and filling the room with NO₂. The technician, probably fortunately for him, died of asphyxiation without regaining consciousness.
There was a terrific brouhaha, as might be expected, and JPL undertook to find out what had happened. J. B. Rittenhouse and his associates tracked the facts down, and by 1956 they were fairly clear. Initial intergranular corrosion produced a fine black powder of (mainly) metallic titanium. And this, when wet with nitric acid, was as sensitive as nitroglycerine or mercury fulminate. (The driving reaction, of course, was the formation of TiO₂.) Not all titanium alloys behaved this way, but enough did to keep the metal in the doghouse for years, as far as the propellant people were concerned.
(RFNA is a mixture of NTO and anhydrous nitric acid).
Yeah, it didn't click with me immediately. I saw "titanium fire" but my brain substituted "magnesium fire".
The idea that titanium can catch on fire in a similar way, is…concerning.
… and of course there's a small-scale YouTube video: https://www.youtube.com/watch?v=cH76ObzYIxc
Ditto.
It's rather surprising that any metals at all can withstand the remarkably unhappy oxidizers used in modern rocket fuels! Probably the metal part in contact is an oxide layer that's easily abraded by big chunks of crap.
Designing exotic thermites is a literal direct application of the elementary reaction thermodynamics that you learn in basic chemistry classes. Bootstrapping ignition is a bit more involved and often multi-stage.
How can chemical interactions be "exotic"? Or did they mean "exothermic" or "exciting" or "esoteric"? Is "exotic" a technical term, or did the chemicals simply come from a distant foreign country, and how does that affect their physical properties?
https://www.quora.com/What-are-exotic-compounds-in-chemistry
Is there such a thing as non-baryonic titanium?
https://en.wikipedia.org/wiki/Exotic_matter
>Hypothetical particles and states of matter that have "exotic" physical properties that would violate known laws of physics, such as a particle having a negative mass.
I can understand why you'd want to fill a rocket with exotic matter of negative mass: to make it lighter! ;)
Webster's 1913 for "exotic" includes: extraneous. Same source for "extraneous" includes: without or beyond a thing; not essential or intrinsic; foreign.
First, it can mean observed chemistry that is seen neither in nature nor the lab under ordinary conditions. In other words, it is theoretically possible but the required conditions are so atypical that the expected probability of it occurring unintentionally are very small. I'm guessing this is the case that applies here.
Second, "exotic" can mean engineered chemistry that is far outside the standard industrial spectrum, typically because there is a unique application with very special requirements. Industrial chemistry is all about price performance of the chemistry not absolute performance, but some use cases are more performance sensitive than price sensitive so people will design bespoke and unusual ("exotic") chemistry for the purpose. It is a curve. Everyone uses the cheapest chemistry that still does the job adequately.
More simply, its politely described as "exotic" if you go to a conference and talk about it and people are jealous of your gear / skills / budget / risk taking.
An electronics analogy is a dude who can't solder at all thinks thru-hole soldering is exotic skills and equipment. A dude good at thru-hole thinks SMD is exotic. A dude with good SMD ability thinks owning your own bare die wire-bonder is exotic. A dude with great wire-bond ability thinks working at a foundry and designing custom monolithic microwave chips is exotic. I have no idea what a custom microwave chip designer would think is exotic, probably some Area 51 stuff involving UFOs, realistically maybe working as a physical chemist or as an experimental semiconductor physicist. Those guys probably think super colliders and warp drives and fusion reactors are exotic, I donno.
In fairness, way back when I was playing thermites I just bought the required materials. Metal oxides in particular tend to be boring chemicals and arguably the more malleable half of that equation. Of course, this was in the 1990s when you could legally buy military explosives with a driver's license and cash...
Thermite made from aluminum and iron can start with a child's firework sparkler, so it is pretty easy to make and use.
Titanium, while not a noble metal, is more resistant to heat and corrosion than aluminum and is just generally less reactive. You might call it "less un-noble" than aluminum. I expect that you'd have a much harder time getting thermite started using titanium. Just look at how much it took to get the titanium fire going on the Crew Dragon- a chunk of NTO had to slam into a valve at really high speed.
Edit: Since you piqued my curiosity, I decided to see if it had been done before. It has: http://developing-your-web-presence.blogspot.com/2008/10/on-... You replace iron oxide with titanium oxide (since titanium pieces oxidize only on a thin outer layer I expect you would need a very fine mesh), and keep the aluminum. This seems to need a little work but it has been done.
Sparklers (i.e. the fireworks) sometimes use titanium to make the sparks, but the primary thing you are igniting is a conventional oxidizer salt (nitrates) and an organic binder. Basically weak rocket fuel with some particles of flammable metal in it to throw off sparks. This is not a thermite reaction. The surface areas of the metals have a huge impact on practical flammability, hence why fine powders are easy to ignite. Iron is also quite flammable as a fine powder but I don't expect my skillet to spontaneously combust on my stove.
Typical thermite bootstraps are three stage for the simple reason that chemistry that is easy to ignite usually does not generate enough thermal power to bootstrap a thermite reaction. A sparkler has a composition that is very similar to a primary stage, a mixture of oxidizer and organics plus a bit of metal for better thermal power (or sparkles, in the case of a sparkler). This is used to ignite the booster stage, which does deliver sufficient thermal power to bootstrap a thermite reaction but is difficult to ignite directly because it is essentially inert anywhere close to room temperature. Boosters are typically direct metal oxidation reactions e.g. aluminum and sulfur.
Useful thermite also tends to be fairly coarse mesh size, which makes it more difficult to ignite. You can trigger a thermite reaction with a fine mesh thermite mixture but that will mostly just give you sunburn and blind you for a day -- typically you want a controlled burn.
I suppose they might retain the ability in case of a chute failure, but that would be it.
I'd still like to see them try it.
https://twitter.com/iamkostmos/status/1104510464627625984
I hadn't heard any thing from official channels since then.
During a propulsive landing, the Dragon capsule would test the SuperDracos while still at a reasonably high altitude; this way, in the event of a failed SuperDraco, the parachute backup can be used. (The landing burn for Crew Dragon would start only a few seconds before landing, far too late to deploy the parachutes if the engines don't perform properly.)
That said, this is old information, and many things could have changed since then.
read: NASA says it's not happening.
After landing there would be a process to make the craft safe by releasing all pressurised gasses, and either sealing or burning off remaining propellants. With no pressuring remaining there is no way to repeat this anomaly.
There will of course still be issues with the residual hypergolics since they are highly toxic and the propellant tanks will need to be decanted: whether than happens before or after crew recovery is another discussion.
For the time being, NASA has published a tentative target of mid-November 2019 for Crew Dragon’s first crewed launch to the International Space Station
I'd rather fly 1,000,000 miles in a Boeing 737 MAX than risk one ride in this contraption. It surely needs quite a bit more unmanned testing before it can honestly be considered to be "man rated".
Of course, NASA has a long and sordid history of pushing spaceflight while playing down many potential catastrophic problems. The Shuttle problems prove that.
And there's no need to belabor Musk's shortcomings. As someone on HN recently quipped, in response to Musk's claim of full driving autonomy by 2020: Tesla has an advantage here in that they don't feel the need for their autonomy to be particularly safe.
Space travel is one of the most dangerous forms of transport ever devised. Even the most insane, wildly optimistic estimates for the safety of the Shuttle were orders of magnitude more dangerous than an airliner.
Comparing the two does not make sense. If airliners are your standard of safety for space travel, you might as well take a shortcut and just say “don’t launch humans into space.”
It most certainly makes sense. A lot of sense.
The Dragon capsule spectacularly blew itself to smithereens. Now is not the time to bring out some duct tape and baling wire to slap the thing back together so that it can attempt to fly astronauts into space before the end of the year.
Now is the time for careful and considered testing. Not rushing. This thing needs to be safe.
I can guarantee you one thing. If American astronauts die this year because of inadequate testing, that will be the last time you see American astronauts being launched into space on a NASA sponsored platform.
If Americans die because this was rushed, you can kiss NASA's manned space efforts goodbye. We'll pay the Russians what it takes to keep going to ISS. And when ISS ends, NASA will be out of the manned space flight business. The operative word going forward will be "robotics".
You may still be right about your prediction, though. But that just says something about people having unrealistic expectations.
And that's why the US shut down the manned space program after Challenger.
> If Americans die because this was rushed, you can kiss NASA's manned space efforts goodbye. We'll pay the Russians what it takes to keep going to ISS.
The lack of recent fatalities with the Soyuz does not mean the Soyuz is "safe". 141 safe launches just implies that the probability of success is p^141=0.5 (since I have no idea how likely this outcome was, I default to 50/50), so 99.5%. (edit: 99.3% with Laplace's rule.) That's very good, but it's also very very far from the safety record of airplanes.
In the US (important qualifier), commercial air travel sees a fatality rate of 0.07 per billion passenger miles. Also (important qualifier), that number excludes acts of suicide and terrorism.
Buses see a fatality rate of 0.11 per billion passenger miles. Trains see a rate of 0.15 per billion passenger miles (not including fatalities of non-passengers).
Both of those modes of transportation compare favorably with air travel. If you consider a worldwide perspective (in which pilots are on average less well trained), and include the effects of hijacked planes, it's possible one of the two might even beat it.
(I'm not disagreeing with you that on a per-mile basis, air travel is very safe. I just think the specific numbers here are interesting. Also worth pointing out that for trips less than a few hundred miles, other modes are probably safer. If rail was used more for long distance travel, it would probably improve its safety record by this metric. And air travel is pretty terrible for the environment as well.)
Source: https://journalistsresource.org/studies/environment/transpor...
Doing the math using the data from https://www.bts.gov/content/transportation-fatalities-mode and https://www.bts.gov/content/us-passenger-miles, I come up with 6,059 billion airline passenger-miles for the decade ending in 2017 (the last year provided there) and 66 fatalities, for a rate of about 0.01 per billion passenger miles.
In the same period, railroads did about 176 billion passenger miles with 94 fatalities, for a rate of about 0.53. Buses did 3,234 billion passenger miles with 486 fatalities, for a rate of 0.15.
The numbers vary in other countries, but I suspect they tend to vary in the same direction and in a similar amount.
There’s an argument that instead of passenger-miles, most people’s intuitive sense of transportation danger is actually based on passenger-hours, not passenger-miles, which makes buses and trains look better.
But I completely agree with your fundamental point that air travel is incredibly safe!
The argument is that humans, in the aggregate, don’t choose trips exclusively based on passenger miles; they (often, though not always) think about how many hours they’re willing to devote to travel.
If cheap air travel availability as a mode induces people to travel 10x more passenger miles, in a world where airlines were only as safe per passenger mile as cars, there would be a net reduction in public health.
But we don’t live in that world; planes are better in safety when measuring by passenger hour than cars, so in fact we’ve improved public health.
The train data is less clear. It's described as "Train accidents" so it could potentially be more general, but grade crossings and trespassers are separated out, which seems to imply that "Train accidents" is occupants.
People often repeat this claim online without clarifying that airlines measure by number of departures.
I haven't done a fair comparison where the same metric was used (say, fatalities per mile traveled in both transport types) but it would be interesting to see one. I'm sure airliners would still come out on top against cars, but the ratio would be different. Then if we extend this fair comparison to space, the amount of miles traveled gets rather large, so that would be interesting to see as well.
The standard metric is fatalities per passenger mile. In the US, cars account for roughly 10x the number of passenger miles as airlines, so you’d expect 10x the fatalities if they were equally safe. In recent years, cars average around 40,000 fatalities per year, and airliners average around zero.
I wonder if there have been any numbers published for space. Looks like the ISS does the equivalent distance of (only) 134 crossings of North America in one day. Or, stated another way, 134 people crossing it once. Clearly airlines carry way more people than that, and the fatalities are lost in the noise. So it would seem that yeah you're right, airlines are still beating space for safety by a big margin. To make space win you'd have to cheat and count something absurd like the miles our solar system travels as it orbits the galactic center... but then to be fair you'd have to credit the airlines with the same distance... so nevermind.
Losing a crew on the first manned SpaceX mission would end the company. Musk is not stupid; moral or not, he knows that these people need to make it back for the military dollars to keep coming in so he can make it to Mars.
Yep, as one of Russian cosmonauts said when asked if he was afraid: ones, who are afraid, do not fly to space.