Field-direction sensitivity of Kondo hybridization in UTe$_2$

Abstract

Neutron scattering experiments on the spin-triplet superconductor UTe$_2$ have established that the dominant low-energy magnetic response is along Brillouin zone boundaries, resembling the magnetic susceptibility of narrow-gap interband excitations. We report a study of the sensitivity of these excitations to magnetic field along the crystallographic $\hat{a}$-axis. Up to fields of $μ_0 H$=13 T, the maximal inelastic neutron spectral weight increases in energy transfer, with a pronounced increase in $d\hbarω_{peak}/dH$ near $μ_0 H$=7 T. This behavior parallels the field and temperature dependent features of the electrical resistivity that are associated with Kondo hybridization. Our measurements suggest that $\hat{a}$-axis fields near $μ_0 H$=7~T induce a change in the hybridization between heavy $f$-electrons and the bare conduction band.

Figures (12)

• Figure 1: Inelastic neutron scattering from UTe$_2$ at $T=$45 mK in the $(0kl)$ scattering plane integrated in $\hbar\omega \in\{\textcolor{black}{3.0},4.0\}~$meV. Each quadrant shows the symmetrized scattering in $\hat{a}$-axis magnetic fields of magnitude 0 T (a), 3 T (b), 7 T (c), and 11 T (d). The scattering, which is strongest at Brillouin zone edges which are denoted by white lines, is substantially reduced by magnetic field along the $\hat{a}$-axis.
• Figure 2: Symmetrized inelastic neutron scattering from UTe$_2$ along the $(0,\frac{1}{2}l)$ direction at zero field (a) and 11 T (b) from the CAMEA experiment. Here, the scattering has been integrated in $k\in\{0.3,0.7\}$ and $h\in\{-0.4,0.4\}$. In (c), the scattering intensity is shown as a function of $(0\frac{1}{2}l)$ in all four measured fields, showing that the excitations are peaked at $l=0$ and $l=4$. Scattering is integrated between $\hbar\omega\in\{1,4\}$ meV and $k\in\{1.25,1.75\}.$ The red line in (c) is a fit to two gaussian peaks centered at $l=0$ and $l\approx3.85$. Both peaks are of the same amplitude, but modulated by the U$^{3+/4+}$ magnetic form factor. All error bars represent one standard deviation.
• Figure 3: Inelastic neutron scattering from UTe$_2$ from second experiment on SEQUOIA at $T$=1.9(1) K, integrated in $\hbar\omega\in\{3,4\}$ meV in (a-d) and in $\hbar\omega\in\{8,9\}$ meV in (e-h). All slices have symmetrized and background subtracted as described in the appendix, and have been integrated in the out of plane scattering direction $h\in\{-0.5,0.5\}~$meV. Brillouin zone boundaries are shown as solid white lines, and the panels have fields applied along the crystalline $\hat{a}$-axis of (a,e) 0 T, (b,f) 7 T, (c,g) 11 T, and (d,h) 13 T
• Figure 4: Dispersion of magnetic excitations from UTe$_2$ along the $(0k0)$ direction integrated in $h\in\{-0.4,0.4\}$ and $l\in\{-1,1\}$. Scattering has been background subtracted and symmetrized as described in the appendix. The white points are fits to better visualize the dispersion of the excitations at zero-field (a) and 11 T (b). Error bars represent one standard deviation uncertainty in the fitted dispersion energies.
• Figure 5: Summary of the evolution of the low-energy magnetic excitations in UTe$_2$ as function of magnetic field. (a) Energy-dependent spectra centered at the ($0,\frac{3}{2},0$), integrated between $h\in\{-0.4,0.4\}$, $k\in\{-1.7,-1.3\}$, and $l\in\{-0.5,0.5\}$. Intensities are shifted by 10 arbitrary units for each field, the solid lines represent fits to a Lorentzian form. The vertical solid lines denote peak positions as extracted from the fits. (b) Black points show the integrated intensity of the cuts in (a) normalized to the zero-field integrated intensity. The red points show the shift in the peak in spectral weight as a function of field. In this figure, error bars are smaller than the points but are included. (c) Dispersion of magnetic excitations along the $(0k0)$ direction as a function of field, as extracted by fits to cuts along the energy direction. All error bars in (a) represent one standard deviation, and in (b,c) represent one standard deviation in fitted parameters.