Fig. 1

The scheme of polarization-based layer-resolved all-optical toggling of magnetization. a Physical scheme and sample design. The p-polarized/s-polarized optical pulse (and the associated targeted layer for magnetization reversal) is shown by red/blue. b, c The distribution of |e|² (dashed-dotted line) and the partial δA and total A absorption (solid line) inside the multi-layered heterostructure for p-polarized and s-polarized optical pulses respectively, with fixed wavelength 800 nm and angle of incidence 59°. d Calculated angular dependence of the absorbed optical energy by the top (dashed line)/bottom (dotted line) layer with incident p-polarized (red)/s-polarized (blue) light. e Calculated (thin solid lines) and experimentally measured (thick solid lines) reflectance angular spectra for incident p-polarized (red)/s-polarized (blue) light. Also shown is the calculated strength of |e|² at the bottom surface of the bottom Gd₂₇(FeCo)₇₃ layer with incident p-polarized (red dashed-dotted line)/s-polarized (blue dashed-dotted line) light. In panels d-e, the green background colour indicates the predicted region where the absorption of p-polarized light in the bottom Gd₂₇(FeCo)₇₃ layers is greater than in the top Gd₂₆(FeCo)₇₄ layer. This is the region with expected polarization-controlled layer selective switching of magnetization. The black and grey arrows indicate the angles where the polarization-based all-optical switching experimentally succeeded and failed respectively