.png)
4 months ago
4
The U.S. Department of Energy is inviting companies to propose experiments using its Microreactor Application Research Validation and Evaluation (MARVEL) reactor, a testbed scheduled to begin operations in 2028 at Idaho National Laboratory’s TREAT facility.
Only U.S. private-sector entities can lead experiments, though national labs or universities may serve as partners. The program aims to validate a range of microreactor end uses, including electricity generation, industrial process heat for chemical plants, and advanced controls and safeguard systems for reactor designs. The DOE is looking, in particular, for “novel” concepts that have never been connected to a nuclear reactor or previously demonstrated concepts for which data is limited.
MARVEL will produce 85 kilowatts of thermal power and 10 kW of electricity, but it isn’t intended to feed into the grid. The system, cooled with a liquid sodium potassium alloy, will be roughly the height of a railroad car. The short-term demonstration platform will operate between 2028 and 2030.
John Jackson, the national technical director of the DOE’s microreactor program, says MARVEL was intentionally defined as a limited R&D project in its funding determination, which avoids schedule delays otherwise associated with capital projects. “Part of the goal is to share, as much as possible, lessons learned from our engineering challenges with industry, so that as we trail-blaze a path forward for advanced reactors, we can lift all boats and help commercial developers mature their concepts faster,” Jackson says.
The guard vessel—which surrounds the reactor core and primary coolant loop—and other key components are already complete or in storage. The DOE plans to assemble and install MARVEL in 2026. Initial dry criticality is expected to follow in 2027, in which MARVEL will achieve a self-sustaining nuclear chain reaction before the coolant is loaded. Full-power operations are slated for 2028.
That timeline aligns with many companies’ goals to enter the market by 2030, including designs supported by the DOE, like Westinghouse’s heat-pipe-cooled eVinci and Radiant Industries’ gas-cooled Kaleidos.
2030 is a reasonable goal for microreactors to come online, says Mary Lou Dunzik-Gougar, associate dean of the Idaho State University College of Science and Engineering and former president of the American Nuclear Society. “It’ll happen, so the real question is who’s going to go first?”
Dunzik-Gougar foresees high demand for MARVEL. “I think they’re probably going to be flooded with ideas, and then it’s just a matter of determining who can be served and what ideas mesh with what capabilities,” she says. “If they get a number of proposals, they should have a good assortment of tests that could be of value.” (Dunzik-Gougar serves on the advisory board of the Idaho National Laboratory’s Materials and Fuels Complex, where MARVEL is being built; however, she says she has no direct role in reviewing the expressions of interest from groups looking to use the testbed.)
Given that MARVEL will be in operation for only two years, Jackson wants to conduct as many experiments as possible, likely focusing on demonstrations that are scalable in size and duration. He confirmed preliminary discussions with at least a dozen companies so far, targeting a variety of potential experiments utilizing electricity or process heat.
“I can’t share specific details, but some examples would include powering a containerized high-performance computer, demonstrating supercritical CO2 power conversion, high-temperature gasification, and thermal management of a data center,” Jackson says.
Coolant and Reactor Design
MARVEL is a liquid-metal reactor running on uranium-zirconium-hydride TRIGA fuel, a safety-demonstrated and certified option mainly used in research reactors. The system operates between 500 and 550 °C, but process-heat experiments can receive inlet fluid temperatures of 300 °C.
The choice of sodium potassium as a cooling mechanism has two key advantages, according to Jackson: Because it’s a liquid metal at relatively low temperatures, MARVEL benefits from superior heat transfer and a low-pressure system. This allows for a more compact reactor design and simpler safety and containment systems.
“Sodium potassium is a very efficient coolant,” says Dunzik-Gougar. “The design is such that you don’t have to pump or force it. Natural convection in the metal will carry the heat away naturally, whereas when you use a gas or water, you have to pump it. So it’s a good choice of heat sink, and it means the system can be physically smaller.”
Many components had to be designed from scratch to make MARVEL a reality, such as the four-loop primary coolant system. But the best example, Jackson says, is a custom-designed safety feature: a four-drum reactivity control system with a “central insurance absorber“ that inserts down into the core for active reactivity control. “This design is unique to MARVEL and has garnered interest from multiple other projects,” Jackson adds.
Fuel and component fabrication on the US $80 million project began last year. France-based TRIGA International was selected to supply 37 fuel elements, while U.S.-based Carolina Fabricators is handling the primary coolant system and guard vessel for the core.
Remaining parts include the reactor support frame, top shield, instrumentation and control, and the heat extraction, power conversion, and primary coolant systems. Jackson adds, “We also need to complete the safety basis and have it approved by DOE, assemble the reactor, get authorization to load fuel, perform initial dry criticality, load NaK [sodium potassium alloy], do zero-power physics testing, and then achieve full criticality.”
Barriers to Launch
Of all the potential use cases microreactors can serve, Dunzik-Gougar sees remote military applications as an early proving ground. “The next logical deployment will be on the civilian side in rural communities like those in Alaska,” she adds, pointing to diesel-reliant grids.
Despite their promised flexibility in markets that conventional reactors cannot address, microreactors have been slow to launch, partly because there’s no way to test them. MARVEL plans to fix that, but even its startup schedule was pushed from a 2023–2024 window, to early 2025, and then 2027. Jackson says those delays were primarily due to “the realization of unknown unknowns,” adding, “There’s a reason this is an R&D project. It hasn’t been done before, so we’re learning as we go.”
The U.S. supply chain for nuclear-grade components has atrophied in recent decades, which further complicates build timelines for new reactor projects. In a 2022 presentation, MARVEL leaders cited fuel prioritization and supply-chain gaps as key risks. Some have been addressed now, with fabrication underway and TRIGA fuel contracted for 2026 delivery. The team has also procured much of the stainless steel structural material but still faces long lead times on beryllium.
Slow regulatory guidance is a broader underlying challenge. “The regulatory regime is really set up for larger capacity plants, not smaller reactor designs,” says Dunzik-Gougar.
Last month, U.S. President Donald Trump issued executive orders directing substantial reforms to the Nuclear Regulatory Commission’s operations, aiming to kick-start high-volume licensing of microreactors and modular reactors. Before that, the NRC released an integrated plan to provide clarity for diverse microreactor designs and deployment models.
MARVEL’s “first-mover” role could help realize the goals of reducing regulatory barriers and costs. “The process for establishing partnerships with fabricators has been a years-long effort,” Jackson says. “With DOE authorization, MARVEL staff have trained engineers to perform the critical work to address technical questions and developed methods to reasonably assure safety.”
