Why Not Titanium Alloy Rocket Engine Nozzles?

[Ozen]

TL;DR Summary

INCONEL is typically used for the outer liner of the nozzle, but why not titanium alloys?

Industry standard right now is a copper alloy (usually an aluminum bronze) inner liner to promote heat transfer to the coolant channels and the walls of the channels and outer liner are usually made of an INCONEL alloy, usually either 625 or 718.

But INCONEL is VERY expensive and heavy. The outer liner of the nozzle actually is not exposed to high temperatures, the inner liner however can commonly reach 700K+ especially without film cooling. But the outer liner is completely different, it is exposed to the fuel (my case being CH4 which is cryogenic) and depending on how the channels are designed, the max temperatures it should see (outside of ambient temp) are 140K to 190K. So INCONEL's high temperature performance is not needed. Instead it comes down to cryogenic performance.

Unfortunately I couldn't find much data on titanium alloys like TI-6AL-4V, the only information I found was this: https://apps.dtic.mil/sti/pdfs/ADA398407.pdf which concluded titanium is very bad for usage in liquid oxygen tanks, often resulting in an explosion. Since the outer nozzle liner isn't exposed to LOX (only the injector manifold, plate, and inner liner which are made of different materials), it would be safe.

Why titanium? My reasoning for wanting to use titanium is it's performance is comparable to INCONEL. It has good high temperature performance, high melting point, can be used for low temperatures, corrosion resistant (oxidizes at high temperatures), and most importantly a very high strength to mass ratio, which is key to developing the type of rocket I want. The two major concerns I have are: 1) insufficient data on cryogenic performance and cycle life 2) INCONEL can briefly withstand temperatures as high as 1400F, which the rocket this would be used on will have the booster re-enter without a retro burn, so brief elevated temperatures are expected.What do you guys think? Just use the industry standard (INCONEL) or does titanium alloys actually seem like a good alternative?

 

[Benjies]

Hi there,

You've cited a good number of different materials in your post here, and also mentioned something I wasn't aware of: that cooling manifolds which make up the bell of a nozzle actually differ in material make up based on whether it is coolant/ambient facing, or exhaust facing. I understand the copper lining situation, but your post seems to indicate that the cooling manifolds themselves are hybrids of materials, which I am skeptical of. Do you have a source for this? Perhaps I'm misunderstanding your post?

The root of your question, however, I'd venture to say can be answered by the material availability of Inconel alloys WRT titaniums. I admittedly have little exposure to material sciences but I know that titanium is not highly available or cheap. These are factors that need to be considered when designing an engine that is scaleable, not just an engine that will earn you 2% more performance relative to modern engines.

 

[Ozen]

Hi there,

You've cited a good number of different materials in your post here, and also mentioned something I wasn't aware of: that cooling manifolds which make up the bell of a nozzle actually differ in material make up based on whether it is coolant/ambient facing, or exhaust facing. I understand the copper lining situation, but your post seems to indicate that the cooling manifolds themselves are hybrids of materials, which I am skeptical of. Do you have a source for this? Perhaps I'm misunderstanding your post?

The root of your question, however, I'd venture to say can be answered by the material availability of Inconel alloys WRT titaniums. I admittedly have little exposure to material sciences but I know that titanium is not highly available or cheap. These are factors that need to be considered when designing an engine that is scaleable, not just an engine that will earn you 2% more performance relative to modern engines.

Thanks for the reply!

Above are the different methods for the cooling channels. The most common are the tubular jacket and milled channels. My design uses the milled channels but with 3D printing. But you can see the outer liner is only exposed to the coolant channels containing the cryogenic fuel. The Merlin 1D engine as an example, uses a copper alloy for the inner liner and the outer liner is a nickel alloy (Elon himself stated this).

Titanium I have not heard of be in short supply, Nickel I was warned is increasing in price from demand. Cost wise, titanium is much cheaper than INCONEL, as INCONEL is a nickel based super-alloy. The cost is a primary demand of why I would like to switch, but also the weight savings would be tremendous (current stimate has the nozzle at 55kg alone with INCONEL). I hope for the rocket this engine will be mounted on to be able to re-enter without a retro burn for protection, which a big aspect on whether it can or not is the dry mass/surface area exposed on re-entry (the engines will be facing towards Earth, otherwise it's going to want to flip around until it is in that position and will likely destroy itself). So by eliminating a large portion of the dry mass, I not only get a boost from a higher dry mass to propellant ratio, but the atmosphere will help slow it down so it can be recovered.

 

[DaveC426913]

Titanium, while strong and light, is brittle is it not?

I wonder how even an alloy would stand up to the extreme vibrations and shock waves generated in the nozzle. (Isn't that why they hit the exhaust with huge gouts of water on liftoff? Because the shock waves would tear the launch pad to bits?

I know little in this area. Take my comments with a grain of salt.

 

[Ozen]

Titanium, while strong and light, is brittle is it not?

I wonder how even an alloy would stand up to the extreme vibrations and shock waves generated in the nozzle. (Isn't that why they hit the exhaust with huge gouts of water on liftoff? Because the shock waves would tear the launch pad to bits?

I know little in this area. Take my comments with a grain of salt.

Titanium does become more brittle at cryogenic temperatures, but that is a common trait even with steels and aluminums which been extensively used in rocketry (steels especially have been used on the nozzles while aluminum for propellant tanks).

From what I know, the waters primary purpose is to absorb the heat from the exhaust and protect the launch site, tower, etc. Water is used due to the high heat capacity it has and being easy to get. But I think you are right, it does have secondary effects like dampening vibrations and disrupting shock waves being formed.

 

[Benjies]

I hope for the rocket this engine will be mounted on to be able to re-enter without a retro burn for protection, which a big aspect on whether it can or not is the dry mass/surface area exposed on re-entry (the engines will be facing towards Earth, otherwise it's going to want to flip around until it is in that position and will likely destroy itself). So by eliminating a large portion of the dry mass, I not only get a boost from a higher dry mass to propellant ratio, but the atmosphere will help slow it down so it can be recovered.

Thank you for the info in your last post- informative read to see how the cooling channels are in fact made of multiple materials.

Another puzzling thing to hear... Am I understanding you correctly that this craft will be re-entering with the bells absorbing the reentry? From what I've learned I'm not aware of systems that can withstand reentry like this without a blunt-body and a bow shock in front of it. Again, please educate me if I'm wrong here, but it seems like your system would just disintegrate.

 

[DaveC426913]

...the engines will be facing towards Earth...

That sounds super unstable at re-entry velocities.

A convex shield has a negative feedback loop when it comes to deviation (small deviations are self-correcting).
A concave shield will have a strong positive feedback loop (small deviations lead to larger deviations).

I suppose, in this modern era of real-time millisecond computer-controlled trim, that's not a showstopper, if you can adjust quickly enough, but it'll be quite a challenge*.*Isn't that why Grumman abandoned the F-29 with its forward swept wings? It had to computer-correct, like 50 times a second, didn't it?

 

[Ozen]

Thank you for the info in your last post- informative read to see how the cooling channels are in fact made of multiple materials.

Another puzzling thing to hear... Am I understanding you correctly that this craft will be re-entering with the bells absorbing the reentry? From what I've learned I'm not aware of systems that can withstand reentry like this without a blunt-body and a bow shock in front of it. Again, please educate me if I'm wrong here, but it seems like your system would just disintegrate.

Yes you are correct, the engines are first taking the initial blunt of re entry. However they create a stagnation zone. It has been proven this method does work for recovery, Rocket Labs Electron booster has been successfully recovered multiple times now this way. They don't even use a landing burn but a parasail and a helicopter to catch it midair. They do have a heat shield on the bottom around the engines and they posted a picture of the returned engines, which had black "burn" marks on them. It is speculated they use INCONEL for their engines. The key to why they can do this is the dry mass/exposed surface area ratio makes it where it acts "fluffly", so it slows down much faster than you would expect.

That sounds super unstable at re-entry velocities.

A convex shield has a negative feedback loop when it comes to deviation (small deviations are self-correcting).
A concave shield will have a strong positive feedback loop (small deviations lead to larger deviations).

I suppose, in this modern era of real-time millisecond computer-controlled trim, that's not a showstopper, if you can adjust quickly enough, but it'll be quite a challenge*.*Isn't that why Grumman abandoned the F-29 with its forward swept wings? It had to computer-correct, like 50 times a second, didn't it?

It should be stable, the engine bells cause a stagnation in airflow around them. And the center of mass is ahead of the center of pressure (which for a cylinder is the center), which that dictates rocket stability. That's why older rockets used fins, to move the CoP behind the CoG. It would not be possible to re-enter with the top of the rocket first in this case, the mass being so rearward will cause it to flip while reentering and the forces it would experience would rip it apart before it can reorient itself.

I'm not too sure about the forward swept wings, only that it was very unstable but greatly enhanced performance.

 

[DaveC426913]

And the center of mass is ahead of the center of pressure

Why is this so? Because the engine volume is significantly denser than the remainder of the craft?

 

[anorlunda]

The Falcon 9 first stages don't do a full re-entry, but they do descend engines pointing down. Falcon 9s also deploy grid fins that appear to dominate the stability.

 

[Ozen]

Why is this so? Because the engine volume is significantly denser than the remainder of the craft?

Correct. The engines are VERY heavy, they are usually made from dense materials like copper, steels, INCONEL, etc. The propellant tanks making up most of the remainder are usually a aluminum alloy, carbon fiber matrix, a composite, or even now 304L SS. the walls are very thin as well, so the CoG should always be rearward.

The Falcon 9 first stages don't do a full re-entry, but they do descend engines pointing down. Falcon 9s also deploy grid fins that appear to dominate the stability.

View attachment 304354

The Falcon 9 does a burn which we call the re-entry burn to slow it down and guide it to its landing spot. They have two methods they use: 1) about a 700m/s burn to land on their barge + 600m/s for the landing burn 2) a boostabck burn for 1600m/s + 600m/s landing burn. Method 2 is to return and land at launch site. The grid fins deploy to assist with steering, not stability so much. That is why they are designed like they are, when parallel with the rocket, they do virtually nothing, but when they rotate them, they give a good control surface. That's why you see on the Starship Booster they don't even bother with retracting them, the drag and lift effects are so low they are negligible compared to the added mass and cost of retracting them.

 

[Gary Feierbach]

My understanding is Titanium is only available in significant quantities from Russia. That doesn't seem like a good choice just based on that. I recall that the spy plane SR-71 had a titanium skin. To get that much titanium they had to use a third party and some misdirection to get it from Russia. The Russians would certainly not have provided it otherwise. Rocket engines are very much in the same category.

 

[Astronuc]

My understanding is Titanium is only available in significant quantities from Russia. That doesn't seem like a good choice just based on that. I recall that the spy plane SR-71 had a titanium skin. To get that much titanium they had to use a third party and some misdirection to get it from Russia. The Russians would certainly not have provided it otherwise. Rocket engines are very much in the same category.

That's somewhat dated information, back in the day.

More recently - https://pubs.usgs.gov/periodicals/mcs2022/mcs2022-titanium.pdf

Historically - https://www.usgs.gov/centers/nation...on-center/titanium-statistics-and-information

I interact with Ti-alloy producers, as well as producers of other reactive and refractory metals and their alloys.

 

[Astronuc]

Summary: INCONEL is typically used for the outer liner of the nozzle, but why not titanium alloys?

does titanium alloys actually seem like a good alternative?

Titanium alloys are complicated (as are other reactive and refractory metals). Similarly, Zr and Hf are complicated. The group elements are hcp (α-phase) at room temperature to a transus temperature at which they undergo allotropic transformation to bcc (β-phase). Depending on the composition there maybe a region in which the alloy has both α- and β-phases, meaning over some temperature range, some grains are α-phase and some are β-phase. The phases are very sensitive to alloy composition and homogeneity (or heterogeneity) and impurities.

In case of Ti, "The beta phase in pure titanium has a body-centered cubic structure and is stable from approximately 882˚C (1620˚F) to the melting point of about 1688˚C (3040˚F)."

For information only, no endorsement expressed or implied.
https://www.spacematdb.com/spacemat/manudatasheets/TITANIUM ALLOY GUIDE.pdf

http://www.phase-trans.msm.cam.ac.uk/2004/titanium/titanium.html

Titanium is a reactive metal, and the reactivity increases with temperature. Metals like Ti, Zr, and Hf (and their alloys), must be processed (e.g., hot worked, annealed or welded) in a vacuum (or inert gas) to prevent uptake of oxygen and nitrogen. Fusion welding is usually performed by TIG (inert gas, e.g., He), or under vacuum with electron beam (EB) or laser. If possible, one would want to use 'cold welding' or magnetic force welding (not always practical).

Another concern would by hydrogen pickup from the propellant, especially if one uses LOX with H₂, NH₃ or CH₄, and runs fuel rich mixture. Ti, Zr, and Hf absorb hydrogen which forms brittle hydrides, especially in high stress locations, e.g., sharp notches, cracks and defects.

Another factor would be thermo-mechanical compatibility with other alloys used in the combustion chamber and throat. Aside from the allotropic transformation, differential thermal expansion and thermal creep may be issues.

And there are other design issues to consider.

 

[Ozen]

Titanium alloys are complicated (as are other reactive and refractory metals). Similarly, Zr and Hf are complicated. The group elements are hcp (α-phase) at room temperature to a transus temperature at which they undergo allotropic transformation to bcc (β-phase). Depending on the composition there maybe a region in which the alloy has both α- and β-phases, meaning over some temperature range, some grains are α-phase and some are β-phase. The phases are very sensitive to alloy composition and homogeneity (or heterogeneity) and impurities.

In case of Ti, "The beta phase in pure titanium has a body-centered cubic structure and is stable from approximately 882˚C (1620˚F) to the melting point of about 1688˚C (3040˚F)."

For information only, no endorsement expressed or implied.
https://www.spacematdb.com/spacemat/manudatasheets/TITANIUM ALLOY GUIDE.pdf

http://www.phase-trans.msm.cam.ac.uk/2004/titanium/titanium.html

Titanium is a reactive metal, and the reactivity increases with temperature. Metals like Ti, Zr, and Hf (and their alloys), must be processed (e.g., hot worked, annealed or welded) in a vacuum (or inert gas) to prevent uptake of oxygen and nitrogen. Fusion welding is usually performed by TIG (inert gas, e.g., He), or under vacuum with electron beam (EB) or laser. If possible, one would want to use 'cold welding' or magnetic force welding (not always practical).

Another concern would by hydrogen pickup from the propellant, especially if one uses LOX with H₂, NH₃ or CH₄, and runs fuel rich mixture. Ti, Zr, and Hf absorb hydrogen which forms brittle hydrides, especially in high stress locations, e.g., sharp notches, cracks and defects.

Another factor would be thermo-mechanical compatibility with other alloys used in the combustion chamber and throat. Aside from the allotropic transformation, differential thermal expansion and thermal creep may be issues.

And there are other design issues to consider.

Thanks for the detailed information and links! It'll take me some time to read through them though.

Manufacturing wise, it wouldn't be an issue. The plan is the nozzle would be made by 3D metal printing, which often are in a vacuum environment, the machines I'm eyeing even list Ti-6AL-4V as a suitable material.

You make a great point about the hydrogen embrittlement. That would indeed be a major concern, mostly for a landing burn where the exhaust can hit the external sides of the nozzle where the titanium would be. During forward flight it shouldn't be an issue, even with the exhaust flow wrapping back like shown below:

The current estimate of hydrogen and water at exit are (given in mass fractions): H = 0.00001496, H2 = 0.0088510, and H2O = 0.4967540. So Dihydrogen makes up not even 1% of the mass ejected.

Thermal-mechanical properties indeed may be an issue. I can begin analyzing that once I select the materials.

 

[Benjies]

The current estimate of hydrogen and water at exit are (given in mass fractions): H = 0.00001496, H2 = 0.0088510, and H2O = 0.4967540. So Dihydrogen makes up not even 1% of the mass ejected.

What fuel is your engine running? I'm running CEA to investigate if this sort of a reactant mix of 50% water occurs when. burning either Hydrogen or Methane and neither are occurring naturally in these simulations (equilibrium flow):

Methane/LOX:

LH2/LOX:

Just used SpaceX Raptor and RS-25 designs to see what the reactants came to. Was wondering what your system's unique situation was that gets you to that 50% H2O spot. Thanks

 

[Ozen]

What fuel is your engine running? I'm running CEA to investigate if this sort of a reactant mix of 50% water occurs when. burning either Hydrogen or Methane and neither are occurring naturally in these simulations (equilibrium flow):

Methane/LOX:
View attachment 304473

LH2/LOX:
View attachment 304474

Just used SpaceX Raptor and RS-25 designs to see what the reactants came to. Was wondering what your system's unique situation was that gets you to that 50% H2O spot. Thanks

Fuel is CH4 + LOX. Software I used to help automate design (after verifying the software calculates it correctly) was RPA. 3.2 mixture ratio with 700 psi chamber pressure. The fraction of elements on exit also depends on the nozzle geometry, but that's a secret!

One thing I am finding out is, if you can afford to get the pressure high, the benefits outweigh the cons. It shrinks the nozzle a bit, primarily with the thrust chamber. This is beneficial because it makes less surface area and you WILL need film cooling. The less surface area the film coolant has to cover, the less fuel is wasted on cooling purposes. The reason I chose a low pressure is because I intend to use an electric turbopump. I'm not sure how energy demanding it currently will be or how much a factor it compares to increasing chamber pressure, so I figured I'd play it safe with a lower pressure.

 

[Benjies]

Okay, well now I can't avoid the question. What are you designing? What is your mission/flight corridor?

 

[Nik_2213]

Not in same league, but a friend was involved with turbo-jet afterburners and 'NOT plenum-chamber burning'.

( I gathered latter was an arcane joke, akin to engines' oil & inter-cooler exhausts providing 'negative drag' ...)

Whatever: IIRC, part of problem with duct design was that 'burning' could exhibit a scary range of conditions, especially when throttling. Never mind 'back-fires' or 'standing waves' with modes, nodes and anti-nodes that could set up structural resonances, fatigue then rend the duct, but also local hot / cold / oxidising / reducing spots that 'effin' moved around'.

The 'best' duct material was not optimal for steady flight, but one that would tolerate such 'combustion excursions', be fit to fly again...

 

[Benjies]

The 'best' duct material was not optimal for steady flight, but one that would tolerate such 'combustion excursions', be fit to fly again...

Something we need to remember. It's easy to analyze design parameters based on steady combustion environments, but ignition, shutdown, thrust beyond test stand conditions, etc., are when you, as a designer, have to answer the tough question of "who is responsible?" when an engine bell bends.

Frankly I understood half of the paragraph you posted before this sentence, but what I've quoted is well received, haha.