Extended Data Fig. 2: Absence of any detectable photo-Nernst effect in WTe₂ at room temperature.

(a) NV center phase image Δφ(r) in WTe₂ Device A for an external dc magnetic field, B_ext = + 28.2 mT. (b) Δφ(r) image for the opposite field direction, B_ext = − 28.2 mT. By flipping the direction of B_ext, we probe the projection of the photocurrent’s magnetic field, B_PC, along opposite NV axes. Afterwards, the NV center’s acquired phase simply changes sign everywhere, indicating that the current flow is independent of the magnetic field direction. (c) Reconstructed J(r) for B_ext = + 28.2 mT. (d) Reconstructed J(r) for B_ext = − 28.2 mT. (e) The difference between the experimental photocurrent patterns at positive and negative external field: J(r, + B_ext) − J(r, − B_ext). The difference highlights the field-antisymmetric component of the photocurrent flow, which isolates the photo-Nernst effect^1,29. The overlaid false colormap denotes the magnitude of the difference, which is experimentally consistent with noise. (f,g) Simulated photocurrent pattern J(r) when including both the APTE term and a possible Nernst term (\(\propto {B}_{\rm{ext,z}}\hat{z}\times \nabla T\)) for f) positive and g) negative B_ext (see Supplementary Section 7). (h) The difference J(r, + B_ext) − J(r, − B_ext) between the simulated patterns clearly shows a chiral photocurrent vortex due to the photo-Nernst effect, which is absent in the experiment. The colormaps in (f-h) denote the magnitude ∣J(r)∣.