How a defence contractor in San Diego built and operates the most scientifically productive tokamak in the United States—and one of the most versatile in the world.
DIII-D, operated by General Atomics (GA) in San Diego, California, has been the flagship tokamak of the U.S. Department of Energy’s fusion programme since 1986. Over nearly four decades it has pioneered plasma-shaping techniques, advanced divertor physics, and high-performance operating scenarios that have directly influenced the design of ITER and numerous private-sector fusion devices. Its story begins not with fusion at all, but with a nuclear-fission contractor’s decision to diversify.1
General Atomics entered fusion research in the 1960s under the leadership of Tihiro Ohkawa, a Japanese-born physicist who proposed the “doublet” plasma cross-section—an elongated, indented shape that increased plasma current and stability limits beyond what circular cross-sections could achieve. GA built a series of devices to test this idea: Doublet I (1969), Doublet II (1971), and Doublet III (1978). Doublet III was a major tokamak with a 1.67 m major radius funded jointly by the DOE and the Japan Atomic Energy Research Institute (JAERI).2
By the mid-1980s, the fusion community recognised that D-shaped plasmas with a single magnetic null (an X-point divertor) offered better confinement and impurity control than doublet shapes. GA undertook a bold modification: Doublet III was re-equipped with new poloidal-field coils, an open divertor, and upgraded heating systems, emerging in 1986 as DIII-D. The “D” signified the new D-shaped cross-section.
The scientific output of DIII-D is extraordinary in both breadth and influence. Among its landmark achievements:
H-mode and ELM physics. DIII-D has been a primary testbed for understanding edge-localised modes (ELMs) and for developing techniques to suppress them, including resonant magnetic perturbation (RMP) coils that have been adopted by the ITER design.3
Advanced Tokamak (AT) scenarios. GA scientists demonstrated sustained high-bootstrap-fraction, high-beta plasmas in DIII-D, proving that a tokamak could in principle operate in a quasi-steady-state without massive external current drive—a cornerstone of reactor economics.4
Negative triangularity. Recent campaigns on DIII-D have explored negative-triangularity plasmas, which exhibit L-mode-like edge behaviour without ELMs while maintaining core confinement competitive with H-mode. These results have attracted worldwide interest and motivated negative-triangularity reactor designs.
Divertor innovation. DIII-D has tested closed and open divertor geometries, advanced materials, and the “small-angle-slot” divertor concept, generating data used to validate the scrape-off-layer models for ITER.
GA is unusual in the fusion landscape: a private company that operates a major national user facility under a cooperative agreement with the DOE. Hundreds of researchers from universities and national laboratories worldwide participate in DIII-D experimental campaigns each year, making it one of the most collaborative facilities in the programme.5
DIII-D continues to operate, with upgrades planned to extend its capabilities in steady-state scenarios and divertor research. Its four-decade run has produced thousands of peer-reviewed publications and trained a large fraction of the U.S. fusion workforce.