The counter-intuitive empirical finding that adding more heating power to a magnetically confined plasma makes it leak energy faster.
One of the most consequential empirical facts in magnetic fusion is that energy confinement time decreases as external heating power increases. In the standard H-mode scaling IPB98(y,2), confinement time scales as P−0.69, meaning that doubling the heating power cuts τE by roughly 38%.1 This phenomenon is called power degradation.
The root cause lies in turbulent transport. The dominant micro-instabilities—ion temperature gradient (ITG) modes, trapped electron modes (TEM), and electron temperature gradient (ETG) modes—are driven by gradients, not absolute values. As more power raises the temperature, gradients steepen until turbulent transport surges to clamp them near a critical threshold. This behaviour is called profile stiffness.2
Above the critical gradient, the turbulent heat diffusivity χ increases roughly in proportion to the excess gradient, so additional power is rapidly conducted outward. The pedestal region in H-mode is partially exempt from this stiffness, which is one reason H-mode confinement is superior—the pedestal provides an insulating layer whose height is set by separate, ELM-regulated physics.3
Power degradation means that brute-force heating cannot compensate for insufficient confinement. Reactor designers must instead increase confinement by raising plasma current, magnetic field, or device size—exactly the dependencies the scaling law prescribes.
Conversely, power degradation provides a natural negative feedback that can help stabilize the burn: a thermal excursion raises power, which degrades confinement, limiting further temperature rise. This partial self-regulation, while insufficient on its own, eases burn control requirements compared to a confinement time that is power-independent.