Why sawtooth crashes in tokamaks leave core reconnection incomplete
In plain words
The tokamak core heats up and then periodically collapses in a sawtooth crash, which the classic model of the Soviet theorist Boris Kadomtsev explains as complete reconnection of the central field lines. Measurements show the core field is only partly reconnected, and the crash is faster than the model allows.
Precise statement
Kadomtsev full reconnection predicts that the central safety factor $q_0$ rises to 1 after each crash on a time $\sim (\tau_A\tau_R)^{1/2}$ ($\tau_A$ Alfven time, $\tau_R$ resistive time), while motional-Stark-effect and Faraday-rotation measurements find $q_0$ remaining near 0.7 to 0.8 and crash times of order $1e-4\,\mathrm{s}$ in large hot tokamaks. Identify the mechanism (two-fluid fast reconnection, island stochastization, pressure-driven secondary instability, stabilization by fast ions or diamagnetic effects) that sets the crash time, the post-crash $q$ profile and the complete versus incomplete outcome. An answer is a model predicting these three quantities across devices.
What would settle it
Two-fluid or kinetic 3D simulations reproducing measured crash times and post-crash q profiles on at least two tokamaks.
Status in the literature
Incomplete reconnection with $q_0$ below 1 after the crash was documented in TFTR (Yamada et al. 1994, https://doi.org/10.1063/1.870479); no model predicts the outcome across devices.