Evan Davis, Massachusetts Institute of Technology

Photo of Evan Davis

Net energy production via controlled nuclear fusion requires confinement of plasma for sufficiently long times to allow appreciable fusion reactions to occur. Currently, the most promising approach to magnetic fusion is the tokamak. Tokamaks are toroidal vacuum vessels that use large magnetic fields (2-3 T) to suspend a hot plasma (about 100 million K) in space, preventing the hot core from touching any material walls. Energy confinement in tokamaks is strongly controlled by plasma stability and turbulence in the edge region, where sharp pressure gradients and large current densities can drive a host of instabilities.

Injection of sufficient heating power into the plasma can trigger spontaneous phase transitions to a high-performance operational regime known as H-mode. H-mode is characterized by drastically reduced edge turbulence and markedly improved confinement. While ideal magnetohydrodynamics (MHD) constrains H-mode operational limits on many tokamaks, it fails to explain the Enhanced D-Alpha (EDA) H-mode on MIT’s Alcator C-Mod tokamak. Stable to ideal MHD, the EDA H-mode is always accompanied by a beneficial edge fluctuation known as the quasi-coherent mode (QCM) that exhausts impurities, allowing for steady-state operation with excellent energy confinement. Despite comprehensive investigation, the physical mechanism that constrains the EDA regime and triggers the QCM has been an outstanding question for many years.

However, recent computational and experimental advances are rapidly constraining the physical mechanism(s) responsible for the EDA’s performance. BOUT++ simulations with realistic values of plasma resistivity predict that resistive ballooning modes drive the QCM. In contrast, recent measurements with a novel mirror Langmuir probe suggest that the QCM is predominantly a drift wave — an altogether different instability. Dedicated experiments to resolve this ambiguity are planned for the near future. Meanwhile, efforts to excite a drift wave-like response in BOUT++ are underway.

Abstract Author(s): E.M. Davis, M. Porkolab, J.W. Hughes, B. LaBombard, P.B. Snyder, and X.Q. Xu