What are the risks of fusion and fission power compared to each other?

[maximum411]

I am curious about the relative dangers of fusion and fission power. Obviously fusion is much safer, but it is not without any risk. Radioactive materials are produced, and containment systems can fail. I am wondering if anyone can give a comparison of a fission meltdown vs. a worst-case scenario fusion plant accident. In each case, what would the extent and reach of the damage be, and how long would affected sites remain contaminated?

 

[tommyxu3]

For the danger of any application of physics field, I think, if we don't abuse them terribly, the damage could be controlled.
The effect of fusion may not be obvious now for it cannot be perfectly controlled without extremely high temperature.

 

[rootone]

You can't get a 'meltdown' in a fusion reactor.
If the containment of the plasma is lost everything just stops.
There could be contamination of the immediate area with short lived radioactive isotopes but not a Chernobyl kind of situation with uncontrolled reactions going on for days or weeks.

 

[DEvens]

The amount of working material present in a fusion reactor is dramatically smaller than in a fission reactor. For a typical commercial nuclear power plant there are 100s of tonnes of fuel in the core at any given time. For a fusion reactor there are likely to be a few 100 milligrams to possibly a few grams if it is a really big design. The worst conceivable problem in a fusion plant will thus involve a tiny amount of material.

The 100s of tonnes of fission fuel, after operating for some time, has decay heat. The fission products put out a small but significant amount of heat. Thus it must be cooled for at least months after it comes out of the core. Since it is one big lump, so to speak, this is a challenge.

That's the good news for a fusion plant.

Most designs for fusion produce copious amounts of tritium. They burn it also, since D-T is the usual target interaction. But they tend to produce much more than they burn. So a 1000 MW plant would produce many kg per year of tritium. This is a challenge to store because it is radioactive, and it is hydrogen. Hydrogen gets out of things pretty easily. And you have a very different sort of explosion risk.

Recently I was involved in advising a station on the location of their tritium extraction plant. Fission reactors produce some tritium, and extracting it means they will have lower activity in the coolant. They had thought to put the extraction and storage just outside the containment for the reactor. We pointed out that, if the quantity of tritium they would have on hand (from 20 years of operation) were to explode, it would quite possibly punch a hole in the containment, and certainly damage many of the support structures of the reactor. They were persuaded to change the design and put the facility some distance away with a protective Earth berm between.

A comparable power fusion plant would produce this much tritium in about 6 months.

 

[Drakkith]

Most designs for fusion produce copious amounts of tritium. They burn it also, since D-T is the usual target interaction. But they tend to produce much more than they burn. So a 1000 MW plant would produce many kg per year of tritium.

Can you elaborate on this? I wasn't aware that fusion plants would produce tritium.

 

[DEvens]

Can you elaborate on this? I wasn't aware that fusion plants would produce tritium.

The reaction people contemplate for commercial fusion is D-T. This is the reaction that people believe is the easiest to produce, and probably the easiest to keep going or make happen often enough to produce commercially useful power levels.

The biggest challenge with D-T fusion is that it produces a massively high energy neutron.

https://en.wikipedia.org/wiki/Fusion_power#Deuterium.2C_tritium

There are designs that want other reactions besides D-T, but they are harder to make work. D-T is hard enough.

So you get a 14 MeV neutron. And this is about 80% of the total energy of the fusion reaction. So you need a way to catch them and get their energy out.

Catching 14 MeV neutrons is a big challenge. At the same time, you want to get some more tritium to replace the T you burn. And the material you use to catch the neutrons has to be such that you can get the heat out and into the material. And you need to then be able to get that heat into some kind of electrical generator, most often some kind of steam generator driving a turbine. And the material has to be such that it can stand up to operating under conditions of this sort for at least some years. Otherwise it won't be commercial.

So the usual solution to these constraints that people come up with is a blanket of molten lead with lithium in it. The neutrons bounce around in the lead, giving up their energy. They also find lithium and produce tritium. Then the lead is circulated to heat a boiler.

You need a blanket of lead approximately 2 meters thick to adequately catch enough energy. Depending on what the structures around the reactor can deal with, you might need to increase that a little.

Now lead has a disagreeable aspect when you use it as shielding. It's great for gamma shielding. But for neutrons it has this feature where a high energy neutron will knock a neutron off a lead nucleus. It is sometimes referred to as "spalling." So it means you start with a 14MeV neutron. And suddenly you have two neutrons at, say, 4 MeV and 5 MeV, and the lead nucleus gets the rest. Or possibly it emits a gamma with the rest, and the gamma is quickly absorbed in the lead.

But now you begin to see the issue. One neutron at 14 MeV can become more than one neutron at lower energies. Sometimes 3 or 4. And sometimes those extra neutrons find lithium to convert to tritium.

A couple years ago I was on a contract to a fusion reactor research company. Their design does the stuff I described here. (That's me in front of their test device in my avatar.) I did some MCNP work to estimate things like the heat load to the structure at the outer surface of the lead. That load was disagreeable, but could probably be managed. But I also estimated the amount of tritium produced per tritium used up in the reaction. And it was much more than 1. Depending on the details of the core and the lead and how much lithium, it might be more than 2 tritiums produced per tritium burned.

There has been some work on being able to dial this number. I did some myself in my spare time. But it was not part of the contract, so I didn't spend that much time. And there are a lot of constraints on the lead. You have to get a very large percentage of the neutron flux absorbed in the lead. Neutrons do a disagreeable thing when they hit iron. Even very slow neutrons will get absorbed by the iron, and release a gamma. That means even very low energy neutrons will heat the iron parts of the reactor structure. So it's a problem. If you wind up with a 1000 MW fusion reactor, and the iron parts of the core get heated to by 1 percent of that, that's 10MW into the iron parts of the core. This is a challenge.

As well, most fusion reactor research is much more concerned with getting the plasma at all, never mind how to get the heat out. So maybe this number can easily be dialed to what is required to keep the reaction fed. But as of this time, it has not been.

 

[mheslep]

Can you elaborate on this? I wasn't aware that fusion plants would produce tritium.

Li-6 is placed in the blanket to breed the tritium fuel supply.

Li-6 + n --> He-4 + H-3 + 4.8 MeV