I’m really sorry for how long it has taken me to get to the second part of a response to Peter Zeihan’s brief video about small modular reactors. The video was released on November 28, 2023, and I threw up a blog response the same day, but spooled up a bunch of thoughts as a background without reaching the actual conclusion.

So here’s my attempt to reach that conclusion, and if you haven’t read the first part of this post, please do so here. But the brief recap is that small modular reactors are probably still be a great idea, despite the announcement on Friday, November 8, 2023, that the Utah Associated Municipal Power Systems (UAMPS) was going to cancel its Carbon Free Power Project with NuScale.

I would go further than that. I don’t even think this announcement is going to make any difference in the prospects for small modular reactors, and I’ll tell you why: NuScale’s approach to small modular reactors wasn’t destined to make any significant market impact.

The big picture is this: the type of SMR represented by NuScale, a small modular reactor-pressurized water reactor (SMR-PWR) was never going to do what its proponents claimed. It was never going to be a drop-in replacement for coal plants. It just never had (and still doesn’t have) what it takes to do something like that, for several reasons. The biggest one is that it operates at high pressures. The second biggest one is that it is fuel-inefficient. And the third one is that its cooling system is too thirsty. There’s more reasons than these but these are the big ones.

If you want to have a potentially drop-in replacement for a coal plant, you’re going to need a small reactor. A reactor that is physically small and is physically light, so you can move it around without too much trouble. That just takes anything cooled by pressurized fluids like water or gas off the table. They weigh too much and they’re too hard to move around. Now some of you are probably saying, hey, they *could* do this, and I don’t disagree. You could do it. But it’s a lot harder to do than an unpressurized reactor. And so in the larger sense, you won’t do very much.

If you confine yourself to unpressurized coolants then you’re talking about a reactor cooled by liquid metals like sodium or lead, or a reactor cooled by molten salt. These are your choices. Molten metals are better understood but molten-salt carries a lot more thermal energy per unit volume, and that is your basic yardstick for the size of the reactor. So a molten-salt-cooled reactor will just be smaller for the same job. Molten metals are usually chemically reactive. Sodium is viciously reactive. Lead is less so. But fluoride salts are chemically stable and unreactive. They are so chemically stable, in fact, that they have the opposite problem, where you worry about corrosion. But if I had to choose between chemical reactivity on one pole or chemical corrosion on the other pole, I know which one I would choose. One process is violent and rapid and explosive and the other is slow and boring and addressable.

So you’re far better off with a molten-salt small modular reactor (SMR-MSR) than with an SMR-PWR or an SMR-LMR. It’s just going to be more compact and there will be no danger of an explosive interaction with the oxidizing materials of this planet. This is even more important if you’re wanting to have a reactor with minimal personnel attendance.

Fuel efficiency has been an Achilles’ heel of PWRs since they were thought up, and SMR-PWRs didn’t do a thing to solve the problem. If anything they made it worse, because many of their designs were highly integrated and that made it very difficult to access the fuel assemblies for replacement. NuScale’s design had the entire integrate steam generator sitting right on top of the reactor core. In large PWRs you keep that area open so you can refuel the reactor but NuScale’s modular design did not permit that. So you had a bunch of tall, skinny reactors that are very difficult to take apart to get to the fuel assemblies, and you still had to change the fuel assemblies.

Many of the SMR-LMR designs were much better than this because they were breeder reactor designs. The Toshiba 4S and others took the approach of being a “nuclear battery” and that would permit the fuel assemblies to last much longer. Some designs postulated many years of operation before the entire unit would be replaced with another nuclear battery. That still left the problem of trying to transport a relatively thin-skinned reactor vessel (because it uses a low-pressure coolant) along with its load of liquid metal (probably explosive sodium) and a core full of high-fissile fuel (mainly plutonium at this point) to some putative disassembly and reprocessing center. Ooof, I wouldn’t want the job of designing that.

But at least the SMR-LMR designs had a strategy around better fuel utilization, I will grant them that advantage over the SMR-PWRs like NuScale. But an SMR-MSR can be so much better. Integrated chemical processing will keep the reactivity constant in the reactor. If it’s a breeder reactor design like LFTR, it will make its own fuel from thorium. If it’s a burner design then it is still straightforward to add external fuel to the reactor during operation. At the end of operation, the fuel salt can be drained from the core and transferred to another core. These concepts have been proposed for many years by several MSR developers, including Flibe Energy.