Extreme-Scale Code Models Extremely Hot Plasma to Explain Spontaneous Transition

For the first time, scientists modeled the spontaneous bifurcation of turbulence to high-confinement mode, solving a 35-year-old mystery.

Click to enlarge photo. Enlarge Photo

(Left) A typical plasma-density-fluctuation pattern from edge turbulence in a realistic tokamak geometry. The black line shows the magnetic separatrix surface with the magnetic X-point at the bottom. (Right) The suppression of high amplitude turbulence (red) starts around 0.175 milliseconds (ms), leftmost vertical  line, and completes around 0.2 ms, at the vertical solid line, followed by the suppression of lower amplitude turbulence (shades of green and blue). The large eddies transform into smaller ones over time.

The Science

Running the world’s largest magnetic fusion experiment known as ITER, currently under construction in France, which aims to prove the feasibility of fusion as an energy source, depends on achieving high-confinement mode. This mode suppresses turbulence at the plasma’s edge. After 35 years, scientists simulated the fundamental physics of the bifurcation of turbulence into the high-confinement mode (known as the “H-mode”). This simulation became possible due to the rapid development of computational hardware and software.

The Impact

Resolving the longstanding question of how and why the plasma bifurcates into high-confinement mode enables scientists to predict the heating power required to achieve the goal of a 10-fold return on energy production in ITER (500 MW of fusion power from 50 MW of input power). Because ITER conditions will be very different from those in today’s tokamaks, any projections of the required heating power based on empirical extrapolations from present data or from simulations with approximate models could well have large uncertainty.


To achieve its goal, ITER’s core ion temperature must be around 15 keV, or over 10 times hotter than in the core of the sun. At the same time, the ions in contact with the tokamak wall must remain cold. However, the cold edge will not allow the core to reach 15 keV unless the distance between the core and the edge is unreasonably large, given the invariability of the steep radial slope of the ion temperature. For more than 35 years, scientists knew that the edge temperature can spontaneously bifurcate into a high pedestal, or transport barrier, in a layer whose width is only a few percent of the total plasma size. This happens as the heating power is increased above a critical level and enables the core ion temperature to rise above 15 keV following the same invariable radial slope.

Experiments have verified that the spontaneous buildup of the edge pedestal results from a spontaneous suppression of its turbulence. Numerous simple, speculative theories attempted to explain this bifurcation.

The reason that only these simple, speculative theories exist has been the multiscale, multiphysics, kinetic nature of the edge plasma physics and the lack of computing power. The multiscale edge gyrokinetic code XGC, utilizing 90 percent of the 27 petaflop Titan supercomputer for three days, achieved the edge turbulence bifurcation for the first time at a first-principles level. The simulation revealed that the turbulence bifurcation is achieved through synergistic effects between the turbulent Reynolds-stress generated sheared flow and the non-turbulent generated sheared flow, also called the “X-point orbit loss-driven” sheared flow. The discovery represents a significant implication for the operation of ITER. If the X-point orbit loss-driven sheared flow is not as strong in ITER as in the present tokamaks, the bifurcation to H-mode may require higher plasma heating power than what is planned. On the other hand, the heating power need may be lower if the X-point orbit loss-driven sheared flow proves stronger in ITER.


C.S. Chang
Princeton Plasma Physics Laboratory, Princeton University


U.S. Department of Energy, Office of Science, Fusion Energy Sciences and Advanced Scientific Computing Research


C.S. Chang, S. Ku, G.R. Tynan, R. Hager, R.M. Churchill, I. Cziegler, M. Greenwald, A.E. Hubbard, and J.W. Hughes, “A fast low-to-high mode bifurcation dynamics in a tokamak edge plasma gyrokinetic simulationExternal link.” Physical Review Letters 118, 175001 (2017). [DOI: 10.1103/PhysRevLett.118.175001]

Highlight Categories

Program: ASCR, FES

Performer/Facility: University, DOE Laboratory, SC User Facilities, ASCR User Facilities, OLCF

Last modified: 8/18/2017 3:54:10 PM