Anomalous Transverse Thomson Effect

Abstract

We formulate the anomalous transverse Thomson effect (ATTE), the long-missing Thomson counterpart of the anomalous Hall and anomalous Nernst effects. Its coefficient is not unique but depends on the electrical boundary condition: ${\tau }_{\mathrm{ATTE}}^{(j)}$, defined for a spatially uniform imposed current density, and ${\tau }_{\mathrm{ATTE}}^{(E)}$, defined for equipotential electrodes. Both are fully determined by the total anomalous Nernst coefficient (ANC) and its temperature dependence, while ${\tau }_{\mathrm{ATTE}}^{(E)}$ acquires an additional contribution from the temperature dependence of the longitudinal resistivity. This relation is model independent within local linear response and applies to the total ANC regardless of whether its microscopic origin is intrinsic or extrinsic. Using a massive Dirac model for Fe${}_3$Sn${}_2$, we show that Berry curvature near the gapped Dirac cones enhances the ANC-related contributions to ${\tau }_{\mathrm{ATTE}}^{(E)}$. In the low-temperature limit, both ${\tau }_{\mathrm{ATTE}}^{(j)}/S_{xy}^{\mathrm{A}}$ and ${\tau }_{\mathrm{ATTE}}^{(E)}/S_{xy}^{\mathrm{A}}$ approach a value of three. We further extract ${\tau }_{\mathrm{ATTE}}^{(E)}$ from reported experimental data for CoS\(_2\), Co\(_3\)Sn\(_2\)S\(_2\) and CeCrGe\(_3\); for CeCrGe${}_3$ it reaches approximately -20 $\mu V/K$, and we estimate the maximum achievable cold-side heat current density to assess its cooling potential. Because the ATTE can be evaluated directly from existing ANC and longitudinal resistivity data, it is readily testable without additional equipment or measurements.

arXiv comment7 pages, 3 figures