Opening of the superconducting gap in the hole pockets of Ba(Fe1−xCox)2

Abstract

Despite many ARPES investigations of iron pnictides, the structure of the electron pockets is still poorly understood. By combining ARPES measurements in different experimental configurations, we clearly resolve their elliptic shape. Comparison with band calculation identify a deep electron band with the dxy orbital and a shallow electron band along the perpendicular ellipse axis with the dxz/dyz orbitals. We find that, for both electron and hole bands, the lifetimes associated with dxy are longer than for dxz/dyz. This suggests that the two types of orbitals play different roles in the electronic properties and that their relative weight is a key parameter to determine the ground state.PACS numbers: 79.60.-i, 71.18.-y, 71.30.-h There is a consensus that the multiband nature of iron pnictides is essential for defining their electronic properties. Many models for the superconducting and antiferromagnetic orders heavily rely on the interaction between different electron and hole Fermi Surface (FS) sheets [1,2]. On the experimental side, transport [3,4], Raman [5,6] and quantum oscillations experiments [7,8] seem to detect predominantly electrons, suggesting they have longer lifetimes than holes. However, ARPES experiments that can image independently hole and electron bands has not evidenced so far a clear difference between them that could explain this behavior. In fact, it has been more difficult to clearly image the electron pockets than the hole pockets with ARPES [9][10][11]. Recently, an ARPES study resolved them in more details [12], but proposed a structure so different from theory, that it clearly calls for more investigation.In this paper, we report ARPES measurements in Ba(Fe 0.92 Co 0.08 ) 2 As 2 , which corresponds to optimal Co doping for superconductivity (T c =23K). Similar features are found for other members of the BaFe 2 As 2 family [13]. We first clarify that, in appropriate experimental conditions, only the electron pockets of the 1Fe-Brillouin Zone (BZ) [14] are observed, which greatly simplifies the spectra. The electron pocket is an ellipse made out of two different orbitals with opposite parities and markedly different dispersions. Along one axis, the band is odd and very shallow (50meV), while it is even and much deeper (more than 100meV) along the perpendicular axis. These dispersions correspond very well to band structures renormalized by a factor of 3, suggesting to associate the shallow band with the d xz /d yz orbitals and the deep band with the d xy orbital. The parity measured by ARPES for the deep electron band is not that expected for d xy , which is due to interferences between the 2 Fe of the unit cell [15]. We find that, both for hole and electron pockets, the lifetimes are about twice longer on the d xy parts of the FS than on the d xz /d yz parts. These anisotropies are in good agreement with a recent study by Kemper et al. [16]. As the d xy parts of the electron pockets are also associated with higher Fermi velocities on the electron pockets, this probably explains why ...

Comprehensive Eliashberg analysis of microwave conductivity and penetration depth of K-, Co-, and P-substituted BaFe2As2

“…As already stated, this study represents an attempt to coherently interpret, within the same theoretical model, multiple properties measured experimentally on samples of the same system but with different types of doping. It was shown above that the measured critical temperature and the temperature dependence of the penetration depth and quasiparticle conductivity are, on the whole, very satisfactorily reproduced within a three band s ± Eliashberg model, also yielding gap values in good agreement with ARPES measurements 12,[48][49][50][51] . Indeed, it has to be stressed that the constraints assumed for the calculations and the requirement to reproduce so different experimental outcomes reduced a lot the number of free parameters and their range of variability.…”

“…T c values are in accordance with literature 48,53 , for the optimally doped systems, and were considered as a starting point for the calculations: in solving the Eliashberg equations, parameters were adjusted to give the experimental T c values. Concerning the energy gaps, in the calculation we fixed the input parameters so that the calculated gaps would reproduce as initial low-temperature values those reported in literature for ARPES measurements, when available 12,[48][49][50][51] , but then we let them change to better reproduce our experimental data. At the end of the procedure, a good agreement with ARPES measurements still holds.…”

“…The parameters values were chosen to give the best agreement with the experimental curves, within the very limited range of variability allowed by all the constraints assumed in the model. Among them, we imposed that the low-temperature values of the gaps must be comparable with those reported in literature for ARPES measurements 12,13,[48][49][50][51] (see real-axis values in Table I).…”

Orbital-dependent electron dynamics in Fe-pnictide superconductors

“…However, for Fe-pnictides, different experiments and theory show that the differentiation of the correlation strength among orbitals is low for the parent compound and decreases further with electron doping [11,19]. Since superconductivity emerges even in electron-doped compounds where all the electrons appear to be itinerant, this raises the question whether strong electron correlation is really the key ingredient required to explain the unconventional properties of the Fe-pnictide family.…”