Extended Data Fig. 1: Positron raytracing with vacuum coil magnetic fields without reconnection.

a, The number of positrons near the laser spot are plotted as a function of OU-ESM channel, coil configuration, and coil current. In all coil configurations, positrons are deflected near the laser spot only for small finite coil currents 0 < I ≲ 3kA, as larger coil currents deflect the positrons below the bottom of the back plate. Further, across all channels, left and right coil configurations consistently exhibit larger positron impacts than the double coil configuration. This exercise implies that LPI-generated electrons are unlikely to be preferentially accelerated to the OU-ESM by double coil magnetic fields, as compared to single coil magnetic fields. b, A simulated five-channel energy spectrum is generated by transforming the experimental no-coil spectrum via positron raytracing at energies of 20 − 100keV, in 10keV increments. The applied magnetic field is the two-coil configuration, with a constant coil current of I = 1kA. Across all channels, simulated spectral dips are observed at 30 − 50keV, in contrast with the experimental spectral bumps (Fig. 2a,b,c) at 40 − 70keV. c, Application of the same raytracing technique to the left-coil configuration can demonstrate the selection mechanism at low energies in Channels 4 and 5, as described in the Fig. 2e caption. In the raytracing spectrum, dips are observed at low energy in Channels 4 and 5, as the lower-energy positrons are preferentially deflected to large θ − z angle by the coil magnetic field.