Fig. 9

Experimental geometry considerations. In panels (a, b), we plot the Fermi surface of PtCoO₂ over several neighbouring Brillouin zones, neglecting (a) and including (b) the rotation of the polarization vector associated with the sample orientation. The range of polarization vectors are illustrated by arrows in panel (b): for the horizontal (k_x) analyzer entrance slit, the polarization is fixed along lines of constant k_y. Moving from bottom to top, the sample rotates by ≈90°, and the polarization goes from entirely in-plane, projected almost along the \({\hat{\mathbf y}}\) direction, to almost entirely out of plane along the \({\hat{\mathbf z}}\) direction. The geometry is drawn schematically in (c): the sample is rotated about the red (\({\hat{\mathbf x}}\)) axis to access the full domain of k_y, with the polarization as drawn: \(\hat \epsilon = \sqrt {\frac{1}{2}} [0,1,1]\) in the laboratory frame. Finite \(\epsilon _x\) at all angles results from a rotation of 7.2° about the grey axis in (c) to select the desired k_x window. In (d–f) we compare the measured spin polarization in the first (d) and second Brillouin zones (e, f). Axis labels indicate distance from Γ_1,2. While a chiral Rashba spin texture is observed near normal emission, contamination between the different spin-channels is manifest as a substantial and artificial out-of plane spin projection in the second zone. P_z is zero near Γ₁ and not shown here. All related colourscales are represented on the same scale to facilitate direct comparison