Hacker News Reader

> The standard way to test material's resistance to molten salt is to put a coupon in a crucible full of salt for a few hundred hours.

Q: What is a coupon in this context?

Great Q: maybe? https://www.researchgate.net/figure/a-MAR-M247Nb-coupon-b-Al...

A coupon is an informal term for a small sample of material, most often metal. It’s perhaps most commonly used in welding where practice welds for people first training to weld are done on pieces of metal often referred to as coupons. :)

It is usually an artifact of materials like metal which are hard to cut, and can’t efficiently just be cut as needed during a task, but are prepared and pre-cut into small pieces you can grab easily. Each of those pieces is called a coupon.

I used to store mine in a Chinese takeout container box.

In the mid 1970s, in my welding class, the school was located down the street from a small manufacturer that formed various metal alloys.

They would provide the school with bins of scrap from their manufacturing free of charge. Those were our practice coupons in class. We had to be careful to return the welded scraps back to the proper bin so the manufacturer could recycle them.

Our class was just a very small part of their recycling flow. It was a great relationship, we needed practice materials, and they hired welders.

It's used 2 ways in metallurgy. Mainly it is a term for test pieces cut from a batch of metal and used for proof testing. Tensile, elongation, fatigue, weldability, and many other tests require destructive testing coupons. When you take welding certification tests, you weld two coupons together, and then submit your coupon for testing.

It is also a term for strips of metal that you mount in a corrosive environment as a way to monitor corrosion of other susceptible components. So if you were worried about your bronze bushings, you would buy some matching bronze coupons, put them in an area of similar flow, and measure their thickness periodically.

https://www.alspi.com/coupons.htm https://en.wikipedia.org/wiki/Fatigue_testing#Coupon_tests

>test pieces cut from a batch of metal

>strips of metal

I mean in essence you just have to look at the word. Coupon is from French couper, which means to cut. So it's a cut-out. It has the same meaning when you look at extreme couponing and metal testing.

Yeah, I didn't think of it as an overly obscure usage, but neither Merriam-Webster or Wikipedia list the definition. Dictionary.com does. Maybe its time to edit Wikipedia over lunch =)

Coupon is used in metalworking to mean a sample you can submit for destructive testing that is representative of the whole piece under test.

A weld coupon, for example, is a piece of test material of the same thickness and weld preparation as the main part being welded but is not integral to the part itself. That coupon is then sawed off and submitted for destructive tests like impact testing (to determine minimum operating temperature), tensile strength testing (specimen is pulled apart to find how strong and stretchy it is), and crystallographic etching (looking at the crystal structure of the metal to see if it has defects).

Does it even need to be a metal? (Since the pressure is so low strength requirements are lower)... How about ceramic or glass (or quartz), or something else non-metal?

non-metals have their own problems and glass tends to have some really weird properties.

I would think a major one would be their failure mode. Metals flex and expand before they eventually fail. Glass/ceramic is fine until suddenly it isn't and has a total failure.

Think of a window being hit. If it were metal it would probably deform but if it is glass it shatters.

Next would be joining them on-site. If needed, metal piping can be bent and welded in-place. what do you do with a glass pipe that needs a join? what do you do if there is a small variation in the plans and the pipe needs an adjustment?

I think there are a host of reasons why glass is not used for pipes.

what about glass pipes encased in said metals :-)

Glass-lined vessels and pipes are already used in the chemical industry so it's a somewhat proven technology.

Not sure if it's suited to the chemistry and temperatures (and radiation) of a molten salt reactor, but it seems like an interesting technology.

Composites? Such as the stuff used in Aviation?

Most of those contain plastics which are usually not good for high temperatures but also have long chain molecules which get broken by neutrons and other particles and cannot heal defects the way metals can.

Glass is already basically welded, or maybe it's more like brazing. But an oxyacetylene torch can be used to work with glass just as well as to cut or weld metal.

Metals also have weird properties. Like tempering and hardening based on temperature. In an industrial setting you need expert welders with deep knowledge of the materials or a weld is going to fail and ruin your day.

So it doesn't seem like a huge leap to me actually, assuming ceramics or glass actually have desirable properties.

Doesn't glass.... melt?

At 3100 F.

Salt melts at 1474 F.

1704C

801C

(or

1978K

1074K)

So does metal. Welding is just melting two parts together.

https://en.wikipedia.org/wiki/Cold_welding

https://en.m.wikipedia.org/wiki/Welding

What point are you trying to make here?

no heat, no melting. just welding

Perhaps it would have been better to say "melting two parts together is just welding".

We can do this with glass, however, lapping is incredibly expensive for something as simple as joining two pieces of glass together.

I noticed you focused almost exclusively on the welding while ignoring the brittle nature of glass?

I would think that putting liquid salt in a a glass pipe is somewhat asking for trouble. One of the many issues with glass is "thermal shock".

Let's say there is a fire and water based sprinklers are activated.. what will the 1,400F glass do once water touches it?

Or there is an accident and someone bangs into the glass. Metal can deform and not fail, glass cant.

“Glass” is a generic term for a wide variety of materials. So is “metal”. Metal can be brittle, sometimes that’s even desirable. That’s what tempering and hardening are about.

A molten salt reactor is a material science problem. Conventional “metal” doesn’t work because of the corrosion. If glass has some desirable property then we can overcome the “bumping in to it” problem. Maybe with a hand rail. Or staying away from the operational nuclear reactor.

I’m not suggesting glass actually be used. I’m saying if it was I wouldn’t be surprised.

It doesn't, but any material used needs to withstand 4 extremes - high temperatures (650-850 C), corrosion (dissolves uranium), very high neutron flux, and US govt. regulation.

Graphite is the current favorite non-metal option. It is already widely used in traditional reactors, so it is approved, and its interaction with radiation is well understood. On the other hand, its interaction with radiation and molten salt is not as well understood, but hopefully this test reactor and others like it will solve that.

One of my favorite solutions seals the reactor in with a crane and 8 graphite containment vessels. The best estimate of the lifespan of structural graphite in this environment is ~7 years, so the plan would be to monitor the system and move the reactor to a new vessel as needed. The goal is 50 year life-span for these reactors.

Anyway, there is research into alternate materials, but to really test them you need to expose them to both molten salts and radiation. And if they pass, you need to get government approval. The last material to get govt approval was Hastelloy N, and I heard the process was a slog.

Ah yes, the most extreme thing one withstands is U.S. regulation.

Nothing requires this research to to be carried out in the US specifically beyond funding being available here. It’s really the inherent difficulties which is holding back progress.

My understanding is that you are not allowed to build a research reactor that is bigger then a university reactor but isn't a full scale energy producing reactor. My understanding is that for such a reactor you would need the full operating license just as a grid connected PWR.

And since in the US its essentially impossible to get a license for anything but a PWR, that isn't rally possible.

Technology independent regulatory framework is one of the main reasons Canada has so many reactor startup, even those that started in other countries.

https://nuclearsafety.gc.ca/eng/reactors/power-plants/pre-li...

> Technology independent regulatory framework is one of the main reasons Canada has so many reactor startup

Canada's vast tracts of land wouldn't hurt either. If a meltdown or containment breach happens, and no population centers are within 400km, that's a much better bad scenario.

yeah im not exactly mad about it being hard / impossible for people to set up a lab-based experimental nuclear reactor near me due to regulations

Those reactors will be much safer then the street in front of your house.

Prove it.

And thus the reason for all this regulation.

We have prove how unsafe the streets are and yet ...

In fact, nuclear reactor to be built have a higher barrier of 'prove of safety' then almost anything else.

In fact, nobody in the US ever died because of civilian nuclear reactor research or at least not in the last 40+ years.

The street is too dangerous, you are right. An experimental nuclear reactor, however, does not make the street safer.

Additionally, the public is relatively unaware of how nuclear plants fail. If people at nuclear reactor research labs suddenly stop showing up to work - what happens? People kind of assume it will explode, or become a radiation hot-zone rendering the local area unusable for a hundred+ years.

And considering the experimental reactors being discussed need parts replacement due to corrosion, what happens if those replacements dont happen?

The dangers of human inaction seem much higher for nuclear than other things. With standard fuel sources, if people stop showing up to work then power simply stops being made - that's pretty much it. And while they are at work, the process is pretty simple and robust (compared to nuclear) with a lot of room for error.

I know I've been surprised (in a good way) to learn about some of the safety mechanisms existing within nuclear reactors, but I still only sort of understand what anything means that I read - and the safety mechanisms that gave me some sense of relief is based on a lot of assumptions I had to make about the way nuclear works as a lay person.

how about nuclear scientist stop saying "trust me bro" and more aggressively educate people on nuclear fail-safes? People have trouble voting for things they do not understand. And educating the public on something they are either uninterested in or incapable of understanding is an unfair burden to put on nuclear scientists, but they are the only ones qualified to do so.

the ball will move a lot faster once lay people can exchange stories about nuclear safety that go beyond "we barely use nuclear, and no one has died yet, it's actually really safe because ... reasons? scientists said so?"

I mean you don't build reactors somewhere out of in the middle of no-where, you need a labor force and infrastructure to build them.

Which of these do you think ACU's NEXT project is pursuing?

Well, one thing about liquid fuel vs solid fuel is that you can move the fuel to a new reactor and, uh... ?"recycle"? / ?"overhaul"? the old one. So you could use one for 7 years why the other is retrofitted. You'd need two reactors... or... you'd need 8 and seven are online while 1 is replaced per year.

So unless materials is solved, you need scalability and replacement. At least the OTHER problem with solid fuel, lots of waste, is generally not a problem with breeder reactors and liquids. You remove the fission products from the fluid and feed it back in. I'm no chemist so I don't know all the separation nastiness involved, but it's better than carting ten thousand year waste across the country to Yucca.

So the question is, what about the other reactor designs, don't they need replacement with respect to the vessel? And as I understand it, fusion reactors also have issues with high speed neutrons so their vessels would need periodic replacement, even if they get to sustained ignition and positive energy.

Your replacement containment layers seems like the "constant replacement" strategy. What if you could simply inject a new layer that hardens and pushes out the older layer?

Also, why not have solid uranium or some similar material as the inner container? Could simple saturation of the existing uranium in the salt prevent excessive wear?

I wonder how much of this approval is because the Chinese brought one online.

There are a lot of options, I'm no nuclear engineer but looking at all the different companies, they all have different approaches:

Terrestrial Energy: The whole reactor, including heat exchanges and so on is defined for 7 years of life. After that a second reactor is running. After a few years the original reactor is put into a storage silo and then a new reactor is placed their ready to be switch to.

From memory they don't seem to pump the fuel from 1 to the other.

Flibe Energy: They are doing the famous lifter. Simular concept, 2 reactor cores with graphite moderation, after 7 years the fuel is pumped to the second reactor, the first is getting its graphite core replaced. This will have longer lead time to deployment.

Moltex Energy: They have totally different approach. Instead of 7 year lift-time they are building a traditional reactor with much longer life. They are basically building a sodium cooled fast reactor, but replace sodium with salt solution. And then in the fuel assemblies are also like those in sodium reactor, but contain liquid salt with uranium in them. They want to produce the fuel salt from spent Canadian CANDU fuel.

Kairos energy: This is a molten salt cooled reactor that uses pellet fuel but little balls instead of the traditional pellets.

I've been thinking about the dumb idea of using solid uranium/thorium as the (closest) salt containment layer: you have a solid uranium (melting point 2000C) or thorium layer (3000C) around the salt, which may not degrade it that quickly if the salts (which are a solution basically, if it's a liquid?) are more or less saturated. Or even if it isn't, how long would a layer of uranium or thorium last? I guess that's the big question.

You don't care if the thorium or uranium captures neutrons I would think. Thorium neutron capture is a good thing.

So depending on how long the solid "breeder" inner shell hold up, when you "recondition" the inner shell, I assume you can just dissolve it into salt, feed it into the normal fission products processing that the salt fuel use, and put in a new solid uranium/thorium salt shell.

Or maybe thorium could be alloyed for more endurance properties as the containment. Of course I have no idea about the various cracking / strength / fatigue properties of thorium as a metal.

I wonder if pellets/spheres could use thorium as a surrounding material.

So how nuts is all of that?

Maybe you could do a weekly monthly re-coat of the inner layer with more thorium or uranium to replace that which gets dissolved/degraded.

Edit: ORNL on thorium properties in a nuclear environment

https://www.osti.gov/servlets/purl/4622065

I don't know enough to answer any of that.

> weekly monthly re-coat

Any operation inside the reactor is a bad idea.

Pure speculation here but what about using salt to contain the salt? Make a big block of salt (possibly foamed) and melt a puddle in the middle?

The corrosion issue is part of what drove Moltex to their rather interesting design (sterile coolant salt is a fluoride; fuel salt in tubes is chloride; this is a fast reactor.) The absence of uranium in the fluoride allows it to be operated at a redox potential where chromium does not dissolve.

Yes, one of the reason I like that reactor design. They basically put something in the fluoride that is basically designed to corrode so that the actual reactor vessel doesn't. If I understand correctly.

Their design doesn't require the 7 year swap cycle most MSR do.

Right! They've got the same thing going that ordinary LWRs do: the neutrons lose their energy in a surround liquid rather than in a solid moderator or solid structural materials other than the fuel rods (which are designed to be replaced often anyway). The design even adds some hafnium to the coolant salt (substituting for zirconium, which is the sacrificial metal you refer to there); hafnium serves to shield the reactor walls from thermalized neutrons.

One additional advantage of separating the fuel salt from the coolant salt is that the coolant salt volume can be increased as desired, making the thermal inertia of the reactor as large as one likes independent of reactor power or size of the fuel load.

> One additional advantage

The problem with that is that you would likely require a different design and go threw full licensing. So in practice this will likely not be used.

Presumably it would happen at design time and a large, high thermal inertia design would be the one licensed.

What about using molten lead or something that can act like a lubricant/insulation at that temperature away from the corrosion in the same way airflow is used over jet turbine blades to stop them melting? It there any high temperature liquid that is molten salt phobic like oil and water?