Superconducting pairing mechanism in $\mathrm{CeCo}{\mathrm{In}}_5$ revisited

Spectroscopic Imaging Scanning Tunneling Microscopy (SI-STM) measurements have previously been applied to the study of the heavy-fermion system $\mathrm{CeCo}{\mathrm{In}}_5$ to examine the superconducting gap structure and band dispersions via quasiparticle intereference. Here we directly measure the dispersing electron bands with angle-resolved photoelectron spectroscopy (ARPES) and compare with first-principles electronic structure calculations. By autocorrelating the ARPES-resolved bands with themselves we can measure the potential $q$ vectors and discern exactly which bands the STM is measuring. We find that the STM results are dominated by scattering associated with a cloverleaf shaped band centered at the zone corners. This same band is also a viable candidate to host the superconducting gap. The electronic structure calculations indicate that this region of the Fermi surface involves significant contributions from the Co $d$ electrons, an indication that the superconductivity in these materials is more three dimensional than that found in the related unconventional superconductors, the cuprates and the pnictides.

Figure 1

(a) Crystal structure of $\mathrm{CeCo}{\mathrm{In}}_5$ where the individual atoms are identified and (b) the associated Brillouin zone. Two momentum planes in the BZ, which are discussed in the text (ΓXM and ZAR), defined by $z=0$ for Γ and $z=\pi$ for Z, are highlighted. (c) Experimental Fermi surface corresponding to the ΓXM plane and (d)–(f) experimental dispersions along the high-symmetry directions Γ-M, Γ-X, and M-X.

Figure 2

(a), (b) Calculated Fermi surfaces in the ΓXM and ZAR momentum planes shown in Fig. 1. (c) Calculated band structure along high-symmetry lines in the BZ where the bands near the Fermi level are highlighted with their respective orbital characters. (d) Orbital-dependent density of states for the Ce $4f$ (red), Co $3d$ (blue), and In $5p$ (green) bands.

Figure 3

Comparison of our first-principles results with the corresponding experimental Fermi surface in the ΓMX plane and the measured dispersions in the high-symmetry directions Γ-M, Γ-X, and M-X. The chemical potential in the calculations is adjusted by approximately 0.1 eV to give full alignment. Only the calculated contributions from the $4f$ levels at the Fermi level are shown along the ΓX direction.

Figure 4

Comparison of our experimental Fermi surface and measured dispersions with two TB parametrizations. (a)–(d) are based on the parametrization in Eq. (1). (e)–(h) are the same as (a)–(d) except that these panels refer to STM parametrizations in Ref. [16] (magenta lines) and Ref. [17] (green lines).

Figure 5

(a) Our experimentally measured Fermi surface is overlaid with our TB parametrization of Eq. (1). (b) The idealized Fermi surface obtained from our TB parametrization. (c) Complex $q$-space structure obtained by autocorrelating the Fermi surface in (b). (d)–(f) Three components that make up the full Fermi surface shown in (b). (g)–(i) Autocorrelations of Fermi surface in (d)–(f) showing the $q$ vectors associated with intraband scattering. (j)–(l) The same as (g)–(i) but now the correlations reflect interband scattering as indicated. In (g)–(i) and (j)–(l) the critical $q$ vectors determined in Ref. [16] are indicated by conduction band $q$ vector (circle) and the superconducting gap $q$ vector (×). The most likely scattering $q$ vectors are also shown by the arrows in panel (f), solid lines correspond to the conduction band from Refs. [16, 17], and the dashed lines to the superconducting gap from Ref. [18]. The yellow line indicates a third critical $q$ vector connecting the conduction bands from Ref. [17].