Infinite-layer nickelates as Ni-eg Hund's metals

Unconventional superconductivity, while still not fully understood, has been observed in various materials. Cuprate superconductors, for example, exhibit superconductivity emerging from bad metallic states formed by doping a charge-transfer insulator, a phenomenon linked to 'Mottness'. In contrast, multi-orbital Fe-based superconductors demonstrate bad metallic behavior influenced by on-site Hund's coupling, a concept termed 'Hundness', leading to the identification of Hund's metals. The recent discovery of superconductivity in Ni-based materials, specifically NdNiO₂ and its isostructural similarity to cuprates (CaCuO₂), provides a compelling opportunity to further explore unconventional superconductivity. While NdNiO₂ shares structural similarities with cuprates, key differences exist, including the absence of a charge-transfer insulator parent state and long-range magnetic order, along with resistivity upturns and a sign change in the Hall coefficient, hinting at a multi-orbital nature and potentially Hund's metal physics. This study aims to investigate the multi-orbital physics in infinite-layer nickelates using first-principles calculations to uncover the origin of electron correlation in their normal phases.

The study extensively reviews previous works on unconventional superconductivity in cuprates and Fe-based superconductors, highlighting the roles of Mottness and Hundness, respectively. It cites numerous publications detailing the theoretical frameworks and experimental findings supporting these concepts. The literature review also discusses existing research on infinite-layer nickelates, noting both similarities and discrepancies with cuprates and Fe-based materials. This includes studies on the electronic structure, magnetic properties, and the debate surrounding the nature of doped holes and their potential implications for Hund's metal physics. The existing theoretical models and experimental results are evaluated to set the stage for the present research and its potential contributions to the field.

The researchers employed an ab initio linearized quasiparticle self-consistent GW (LQSGW) and dynamical mean-field theory (DMFT) method (LQSGW+DMFT) to explore the electronic structure of infinite-layer nickelates. This approach combines ab initio linearized quasiparticle self-consistent GW calculations with a one-shot DMFT correction to the local electron self-energy. The correlated orbitals are chosen to span a large energy window, encompassing strongly hybridized bands and Hubbard bands. Parameters for the DMFT calculations are determined from the local GW approximation and constrained random phase approximation (cRPA) calculations. The methodology has been previously validated against various correlated electron systems, including Mott insulators, Hund's metals, and narrow-gap semiconductors. The researchers used La₁₋ₓBaₓNiO₂ instead of Nd₁₋ₓSrₓNiO₂ to simplify calculations, justified by existing literature demonstrating similarities in their electronic structures near the Fermi level. The effect of Ba doping was handled using the virtual crystal approximation. The calculations were performed using ComDMFT and FlapwMBPT codes. Specifically, the authors calculated the on-site Coulomb interactions, orbital occupations, temperature and doping dependence of local spectra, spin and orbital susceptibilities, to identify characteristic signatures of Hundness versus Mottness. They also investigated the Ni-3d multiplet states to clarify the microscopic origin of the observed behaviors. Finally, they propose an additional experimental test: examining the doping dependence of the Ni-dx²-y² band effective mass.

The study reveals several key findings. Firstly, the calculated orbital occupations show a significant deviation from predictions based on oxidation state rules, particularly for the Ni-$d_{z^2}$ orbital, which is attributed to strong hybridization with La-$d$ orbitals. Secondly, the analysis of Coulomb interactions using cRPA indicates that Hundness is more dominant than Mottness in La₁₋ₓBaₓNiO₂. The static U is smaller than in NiO and even smaller than in FeSe, while the J is larger than in FeSe. Thirdly, the temperature and doping dependence of Ni-$d_{xy}$ local spectra and susceptibilities demonstrate multiple signatures of Hundness, primarily in the active Ni-$e_g$ orbitals and not the inactive Ni-$t_{2g}$ orbitals. This includes spin-orbital separation, absence of a pseudo-gap, and the inactivity of Ni-$t_{2g}$ orbitals in spin and orbital fluctuations. The study further demonstrates the dominance of the spin-triplet configuration over spin-singlet configuration in the Ni-$e_g²$ subspace, consistent with some previous experimental observations and another necessary condition for Hund metal physics. The analysis of the Ni-$d_{x^2-y^2}$ spectral function shows a single incoherent peak without a pseudo-gap formation at high temperatures. The temperature evolution of this spectral function exhibits characteristic features of Hund's metallicity, like a shoulder-like structure in the electron self-energy and an inverted slope in the real part. Lastly, the doping dependence of the effective mass of Ni-$d_{x^2-y^2}$ bands demonstrates a monotonic increase from the electron-doped to the hole-doped side, further supporting the two-orbital Hund's metal picture. The investigation of the Hubbard U's effect on Hundness concludes that U does not play a dominant role in the observed electronic properties.

The findings of this study strongly support the classification of infinite-layer nickelates as two-orbital Hund's metals, driven primarily by on-site Hund's coupling. The observed differences from both cuprates (dominated by Mottness) and Fe-based superconductors (five-orbital Hund's metal) highlight the unique role of Hund's coupling in shaping the electronic properties of these materials. The detailed analysis of various spectroscopic quantities, including orbital occupations, Coulomb interactions, spectral functions, and susceptibilities, provides a comprehensive picture of the correlated electronic structure. The proposed additional experiments, such as measuring the doping dependence of the effective mass of the Ni-dx²-y² band, offer further avenues to validate the two-orbital Hund's metal scenario. The results emphasize the importance of considering multi-orbital effects and Hund's physics in understanding unconventional superconductivity in correlated materials.

This research demonstrates that infinite-layer nickelates exhibit a two-orbital Hund's metal character, fundamentally driven by on-site Hund's coupling. This contrasts with the Mottness-dominated cuprates and the five-orbital Hund's metals found in Fe-based superconductors. The study's comprehensive analysis of various electronic properties provides a solid foundation for future investigations into the intriguing superconducting behavior of these materials. Further experimental validation of the proposed two-orbital Hund's metal picture is crucial and the doping dependence of the Ni-dx²-y² band effective mass provides a readily accessible experimental avenue for such validation.

The study uses La₁₋ₓBaₓNiO₂ as a model system instead of Nd₁₋ₓSrₓNiO₂, although justifications are provided based on existing literature. The virtual crystal approximation used to treat Ba doping might not fully capture the complexity of dopant distribution and its effect on the electronic structure. The choice of Wannier orbitals, while thoroughly described, can still impact the results. While the study strongly suggests a two-orbital Hund's metal scenario, further investigations are needed to conclusively rule out other possibilities. The exact values of U, as obtained from different methods, may also introduce some variability.