molten salt is an ionic compound held above its melting point so that it flows as a liquid, used in nuclear systems as a high-temperature heat-transfer fluid and, in 📝fusion blanket designs, as the 📝tritium breeder as well.
The salts of interest are fluorides and chlorides of light metals — lithium fluoride with beryllium fluoride, the lithium-sodium-potassium fluoride eutectic FLiNaK, various chloride mixtures — typically liquid between roughly 400 and 800 degrees Celsius. Their appeal is a set of properties water does not have at those temperatures. Vapor pressure stays low, so a salt loop can run near atmospheric pressure while a water loop at comparable temperature needs a thick pressure boundary and carries stored energy that wants to escape. Volumetric heat capacity is high, so a modest flow carries a large thermal load. And because the fluid is already an ionic melt, it does not decompose radiolytically the way water does under intense radiation. The costs are equally concrete: melting points several hundred degrees above ambient mean every pipe, valve and heat exchanger needs trace heating and freeze protection, and the salts attack structural alloys unless their chemistry is controlled.
In a fusion 📝blanket a molten salt can do three jobs with one fluid — absorb the energy of 📝14 MeV neutrons as heat, breed tritium on its 📝lithium content, and shield the structures behind it. That is what distinguishes a liquid 📝breeding blanket from a solid one built of ceramic breeder pebbles and a separate coolant. 📝FLiBe is the specific salt at issue for 📝ARC: it adds 📝beryllium as a 📝neutron multiplier, which is what lifts the 📝tritium breeding ratio above one. The heat the salt collects is then handed to a power cycle — a 📝Brayton cycle in the ARC concept.
The controlling chemistry problem is redox. Corrosion of a structural alloy in a fluoride salt is governed by the salt's redox potential, and a D-T fusion blanket drives that potential in the corrosive direction as tritium fluoride forms from bred tritium; the countermeasure is to hold the potential down deliberately with a dissolved or contacting reducing agent. Savannah River National Laboratory and 📝Commonwealth Fusion Systems pursued exactly this under a DOE Fusion Energy Sciences project, Active Redox Control of Molten Salts For Fusion Blankets (Olson, Garcia-Diaz and Barthel, report SRNL-STI-2024-00280), which frames FLiBe as a blanket material whose risks the company is working to retire. Redox control also bears directly on 📝tritium handling, since the chemical form the tritium takes determines how it is extracted from the loop.
Molten salt loops themselves are not hypothetical: 📝Oak Ridge National Laboratory ran the Molten Salt Reactor Experiment on a fluoride salt in the 1960s, and salt test loops have operated since. What has never operated is a fusion breeding blanket. ARC's salt loop is a design study — the 2015 conceptual paper and its successors — and its breeding ratio, corrosion rate and heat-transfer performance are calculated values awaiting a machine to check them against.