Orbital-differentiated coherence-incoherence crossover identified by photoemission spectroscopy in LiFeAs

In iron-based superconductors (FeSCs), orbital differentiation is an important phenomenon, whereby correlations stronger on the $d_{xy}$ orbital than on the $d_{xz}/d_{yz}$ orbital yield quasiparticles with a $d_{xy}$ orbital character having larger mass renormalization and an abnormal temperature evolution. However, the physical origin of this orbital differentiation is debated between the Hund's coupling-induced unbinding of spin and orbital degrees of freedom and the Hubbard interaction instigated orbital-selective Mott transition. Here we use angle-resolved photoemission spectroscopy to identify an orbital-dependent correlation-induced quasiparticle (QP) anomaly in LiFeAs. The excellent agreement between our photoemission measurements and first-principles many-body theory calculations shows that the orbital-differentiated QP lifetime anomalies in LiFeAs are controlled by the Hund's coupling.

Figure 1

(a) Experimental setup of the polarization measurements. The $\pi$ polarization ($\pi$-pol) is pure, whereas the $\sigma$ polarization ($\sigma$-pol) is mixed with $c$-axis polarization. (b) FS mapping at 74 eV with $\sigma$-pol. Dashed lines are the extracted FSs. The blue and red colors represent $d_{xy}$ and $d_{xz/yz}$ orbitals, respectively. (c) Curvature of ARPES intensity of cut 1. The red and blue dashed dispersions are extracted from the high-resolution polarization data shown in (d)–(g).

Figure 2

(a) ARPES intensity plot of the $\beta$ band at 20 K. EDCs/MDCs that are crossing the colored circles in (a) are shown in (b) and (c). Black dashed curves in (b) are single Lorentz function fittings. (d) Extracted energy-dependent scattering rates $\mathrm{Im}\mathrm{\Sigma }\left(\omega \right)$ are represented by blue circles. The red shaded area represents the nonquasiparticle regime. The area does not go to zero because of the temperature correction $\pi k_{B}T$ to the self-energy. (e) Extracted energy-dependent renormalization factor of the $\beta$ band. $E_0$ at 16 meV shown in (d) and (e) corresponds to the energy of the spectral function anomaly [22]. Error bars are determined by the standard deviation of the fitting parameters.

Figure 3

(a)–(e) ARPES intensity plots of the $\beta$ band at 30, 50, 100, 150, and 200 K, respectively. (f)–(j) ARPES intensity plots of the $\delta$ and $\gamma$ bands at 30, 50, 100, 150, and 200 K, respectively. All spectra are divided by the Fermi-Dirac function convoluted with the system resolution. (k)–(p) DFT+DMFT calculated momentum- and energy-resolved spectral function at 58, 116, and 232 K. (k)–(m) and (n)–(p) correspond to hole bands and electron bands, respectively.

Figure 4

(a) and (b) EDCs of the $\beta$ and $\delta$ bands from 20 to 200 K. All curves are fitted by the QP spectral function plus a polynomial background [22]. The horizontal dashed lines represent zero intensity. (c) Extracted temperature-dependent QP scattering rates. The gray shaded background represents the maximum region of the derivative of the resistivity curve. (d) and (e) Extracted QP peaks at $k_F^{\beta }$ and $k_F^{\delta }$, respectively. The insets of (d) and (e) are DFT+DMFT calculated spectral functions at $k_F^{\beta }$ and $k_F^{\delta }$, respectively. To get the total QP spectral weight, we integrate the extracted and calculated spectral functions shown in (d) and (e) and plot the results in (f). Error bars shown in (c) and (f) are determined by the standard deviation of the fitting parameters.