Finding out whether a new nuclear reactor design actually works is a slow, expensive business. Before anyone lets you anywhere near real fission, you’ll need licensing, shielding, alloys that can survive molten salt, and a budget with an uncomfortable number of zeros. So a team at the University of Idaho asked a pretty reasonable question: how much of the physics can you check with a clear column, table salt, water and an electric current?<\/p>\n
Quite a lot, it turns out.<\/p>\n
The experiment itself isn’t new, and I’m not going to pretend it is. The run I’m about to describe happened on April 23, 2020, and the full writeup landed in a University of Idaho doctoral thesis in May 2022. The reason it’s back on my desk is that on September 17, the INL Foundation and MicroNuclear announced an agreement<\/a> to push the reactor that column was standing in for toward detailed engineering, a manufactured prototype and, eventually, a first-fission test. Which makes this a good moment to look at the bench test holding the whole idea up.<\/p>\n MicroNuclear, a Tennessee-based startup, calls its design the Molten Salt Nuclear Battery, or MsNB. It’s a compact reactor, roughly 10 feet across and 10 feet tall, and instead of solid fuel rods, the nuclear fuel is dissolved directly into molten salt. The company describes it as a 20 megawatt thermal unit, meaning 20 megawatts of raw heat before any of it gets turned into electricity, and it’s meant to run at full power for about 10 years before the whole unit goes back for fuel and salt recovery.<\/p>\n The unusual part is that there’s no pump. The salt heats up in a central cylinder, rises because hot liquid is lighter than cold liquid, sheds its heat at an exchanger up top, and sinks back down the gap between the inner and outer cylinders. Then it does the lap again, driven by nothing but its own temperature difference. It’s basically the physics of a lava lamp, applied to radioactive salt running north of 1,000 degrees Fahrenheit.<\/p>\n Why would you want that? Because pumps and valves are the parts that break. Even America’s newest full-size reactors<\/a> move their coolant with pumps in normal operation, and a machine that can’t lose its coolant pump because it doesn’t have one is, frankly, the entire sales pitch for something meant to sit at a military base or on a remote grid for a decade with minimal babysitting.<\/p>\n Testing that loop with actual molten salt is miserable. You’d need high-temperature alloys, serious safety controls and a pile of money. But the Idaho researchers leaned on a convenient piece of physics: the force driving natural circulation depends on the temperature difference between the hot and cold legs of the loop, not on how hot the fluid is in absolute terms. Which means you can study the flow at temperatures your kitchen faucet could deliver, as long as the stand-in fluid behaves the right way.<\/p>\n Their stand-in was water carrying 25% salt by mass. Ordinary NaCl, basically very committed seawater. And instead of bolting a heater to a wall, they heated it the way fission would: everywhere at once, by pushing an electric current straight through the liquid itself. Engineers call it ohmic heating. According to the team’s own account in Nuclear News, the very first proof of concept involved a welder and a container of saturated salt water, which hit a boil and turned into a fountain within 20 minutes. I’d pay good money for the lab notes from that afternoon.<\/p>\n From there they built a flow loop, and then two full surrogates of the reactor: a steel full-height unit constructed by Premier Technology, a nuclear manufacturer in Blackfoot, Idaho, and a see-through version in clear PVC, so you could literally watch the water move. Testing ran at Premier’s facility.<\/p>\n The bench behind the published numbers is that transparent column, an 8.08-foot vessel with a 2.47-inch central channel, and the 2022 thesis documents that April 2020 run in detail. The team stepped the electric input up through three power levels, holding each one for half an hour to 45 minutes, and let the loop settle at every stage, with a 3D-printed nylon insert inside the vessel guiding the water along the reactor’s intended path.<\/p>\n At the top setting of 5.2 kilowatts, the water leaving the heated channel ran 19.4 degrees Fahrenheit hotter than the water coming back down, and working that split through the heat balance gives a flow of about 0.18 pounds of water per second, or roughly an inch per second, with a full lap of the loop taking 339 seconds.<\/p>\n An inch per second is a crawl. You could stroll past this cooling system without noticing it’s on. But it doesn’t need to be fast, it needs to be reliable, and a flow driven by gravity and a temperature difference has pleasantly few ways to fail.<\/p>\n One caveat, and the writeup is upfront about it: nobody had a flow meter on the rig. Those flow and speed figures are calculated from the temperature measurements, not read off an instrument. So I can’t tell you the water moved at exactly an inch per second. I can tell you the heat balance says it must have, and that the circulation itself was right there in a clear tube for anyone to watch.<\/p>\n You might be wondering how a few kilowatts in a plastic column tells you anything about a 20 megawatt reactor. Because the bench wasn’t trying to match the power. It was trying to match the behavior. If your models predict how salt water circulates on the bench and the column agrees with them, you’ve earned some trust for what those same models say about molten salt at full scale. MicroNuclear’s own testing page<\/a> frames it exactly that way: cheap physical validation of the thermal hydraulics models, for pennies next to what reactor testing costs.<\/p>\n So, back to September 17. The agreement teams MicroNuclear up with the INL Foundation, engineering firm EXP, Premier Technology and Porter House, and it’s built to carry the design through engineering and licensing work, prototype manufacturing and a first-fission test, with the Foundation handling early fundraising and access to Idaho National Laboratory. Premier, for its part, was already picked to manufacture the MsNB in May 2023<\/a>, and the design’s paper trail runs back to a 2017 Department of Energy voucher project with Idaho National Laboratory. So this thing’s been a slow burn, and the salt water column sits near the start of it.<\/p>\n None of this proves the reactor works. A column of warm salt water can’t tell you how the design copes with actual fission, with fuel salt chemistry, or with a decade of corrosion at four-digit temperatures, and as far as I can tell nobody involved claims otherwise. What the bench settled is narrower and arguably more useful: the pump-free loop at the center of the whole concept behaves the way the models said it would, in a vessel you could stand next to in a t-shirt. Small reactors keep turning up in odd places lately, from a borehole reactor program in Kansas<\/a> to a canister lowered down a Florida well<\/a>, and those projects are asking where a reactor can go. This one has spent years on a narrower question: whether a particular reactor can cool itself.<\/p>\n The MsNB still has to earn everything past the salt water, and the fission part is precisely what the new collaboration exists to fund. The INL Foundation and MicroNuclear announced the agreement on September 17, 2026, and a first-fission test is on its published to-do list.<\/p>\n","protected":false},"excerpt":{"rendered":" Finding out whether a new nuclear reactor design actually works is a slow, expensive business. Before anyone lets you anywhere … <\/p>\nSo what is a molten salt nuclear battery?<\/h2>\n
Salt water can stand in for molten salt<\/h2>\n
What the clear column actually measured<\/h2>\n
From salt water to first fission<\/h2>\n