Coexistence of 3D and quasi-2D Fermi surfaces driven by orbital selective Kondo scattering in UTe$_2$

TL;DR

This study reveals the coexistence of 3D and quasi-2D Fermi surfaces in UTe$_2$, driven by orbital-selective Kondo effects, clarifying conflicting experimental observations and implications for topological superconductivity.

Contribution

The paper introduces a combined ab initio GW and dynamical mean-field theory approach to uncover orbital-dependent Fermi surfaces and their temperature evolution in UTe$_2$, highlighting the role of Kondo coherence.

Findings

At high temperature, both Fermi surfaces are 3D with weak spectral weight.

Cooling to 25 K induces one Fermi surface to become quasi-2D due to Kondo coherence.

The 3D Fermi surface persists at low temperature, supporting topological superconductivity.

Abstract

The 3D Fermi surface, along with a chiral in-gap state and a Majorana zero energy state, is suggested to play a crucial role in the topologically nontrivial superconductivity in UTe$_2$. However, conflicting experimental observations of the 2D Fermi surface raise questions about topological superconductivity. By combining ab initio many-body perturbation GW theory and dynamical mean-field theory based on Feynman diagrams, we discovered the coexistence of two orbital dependent Fermi surfaces, both centered at the $\Gamma$ point in the Brillouin zone, which are heavily influenced by the orbital-selective Kondo effect. At high temperature, both Fermi surfaces exhibit 3D nature with weak spectral weight due to incoherent Kondo hybridization. Upon cooling down to 25 K, due to the pronounced Kondo coherence, while one Fermi surface remains a robust 3D Fermi surface, the other transforms surprisingly into a quasi-2D Fermi surface, which should be responsible for the experimental observation of 2D character. Our results suggest that the 3D Fermi surface should exist at low temperature for the topological superconductivity. Our findings call for further investigation of the interplay between the two orbital-dependent $\Gamma$-centered Fermi surfaces.