strike-point sweep is the deliberate movement of the divertor strike point back and forth across the target during a pulse, driven by the shaping coils, so that the time-averaged heat load is spread over far more tile surface than the stationary footprint would touch.
The strike point is where the 📝separatrix intersects the 📝divertor target — the line along which the exhaust channel lands. Sweeping it exploits the fact that a tile's survival is governed by two different quantities on two different timescales. The instantaneous heat flux is unchanged by sweeping: at any moment the same power arrives in the same narrow strip, and the surface temperature spike it produces is set by the local flux and the material's thermal response. What changes is the bulk heating. Move the strip fast enough and no point on the target sees the flux long enough to accumulate steady-state temperature, so the time-averaged load — the quantity that sets bulk tile temperature and cooling requirements — falls roughly in proportion to the swept width. This makes a sweep a complement to 📝divertor detachment and 📝flux expansion, not a substitute: detachment removes power, flux expansion spreads it geometrically, a sweep redistributes it in time.
The technique belongs to short-pulse machines. Its value depends on the pulse ending before the target reaches thermal equilibrium, so a steady-state power plant gains little from it and must dissipate the power instead. 📝SPARC sits on the useful side of that line, and its baseline design turns the constraint into a tool. 📝Adam Kuang and colleagues (2020, DOI 10.1017/S0022377820001117) projected SPARC's unmitigated peak parallel heat flux above 10 gigawatts per square metre — close to an order of magnitude beyond anything demonstrated — and adopted a roughly 1 hertz strike-point sweep across the target as the baseline mitigation, spreading the load enough to keep tile surface temperatures tolerable without active divertor cooling over a pulse of about 25 seconds with a 10-second 📝flattop. Both the heat-flux projection and the sweep's effectiveness are empirical-scaling predictions for a machine that has not yet operated; the paper is explicit that SPARC will test the limits of those scalings rather than confirm them beforehand.