The largest magnetic-mirror fusion device ever built—completed at Lawrence Livermore National Laboratory in 1986 at a cost of $372 million, then mothballed without running a single plasma experiment.
The Mirror Fusion Test Facility–B (MFTF-B) stands as one of the most dramatic episodes in the history of fusion energy. Constructed at Lawrence Livermore National Laboratory (LLNL) in California, it was the culmination of more than two decades of mirror-machine research and, at the time of its completion in February 1986, the largest and most expensive single device the U.S. magnetic-fusion programme had ever built. It was also the device that never fired: within weeks of its dedication ceremony, the Department of Energy ordered it mothballed, a casualty of shifting budget priorities and the ascendancy of the tokamak.1
Mirror confinement relies on magnetic-field gradients rather than a closed torus. Ions bouncing between regions of high field strength at each end of a linear solenoid can, in principle, be confined long enough to reach fusion temperatures. Livermore pursued this idea from the late 1950s through a series of machines—Alice, 2XIIB, Baseball II, and the Tandem Mirror Experiment (TMX)—that progressively attacked the chief weakness of mirrors: end losses. By the late 1970s, the tandem-mirror concept, in which electrostatic plugs at each end suppress particle escape, had restored faith in the approach.2
The original MFTF was a single-cell mirror authorised in 1977. By 1980, results from TMX convinced programme leaders that the tandem-mirror configuration was superior, and MFTF was redesigned as MFTF-B—a tandem mirror with thermal-barrier end cells. The machine stretched 58 metres end to end. Its central cell used a large superconducting solenoid to produce a 1 T field over a plasma roughly one metre in diameter. At each end, yin-yang superconducting coil pairs generated 3.5 T peak fields to form the mirror plugs. Neutral-beam injectors totalling 29 MW of deuterium beams were arrayed along the device, supplemented by electron-cyclotron and ion-cyclotron heating systems.3
Building the superconducting magnets was itself a landmark engineering achievement. The yin-yang coils, each weighing over 300 tonnes, were the largest superconducting magnets fabricated for fusion up to that date. Niobium-titanium cable-in-conduit conductor was wound, reacted, and tested in a purpose-built facility at Livermore.
By the mid-1980s, the U.S. fusion budget was under severe pressure. The Magnetic Fusion Energy Engineering Act of 1980 had promised a demonstration reactor by 2000, but congressional appropriations fell far short. Tokamaks—buoyed by results from PLT, TFTR, and JET—consumed an ever-larger share of funds. In March 1986, MFTF-B was declared operational but immediately placed in standby. It was never reactivated. The hardware was eventually scrapped in the mid-1990s.4
MFTF-B’s cancellation effectively ended large-scale mirror research in the United States. The scientific questions it was designed to answer—whether a tandem mirror could achieve energy-relevant confinement—remain open. Its superconducting-magnet technology, however, contributed to subsequent programmes including ITER. More broadly, MFTF-B became a cautionary symbol of the political and budgetary hazards that shadow long-horizon science projects.