Source capture
Authors Francesco Rosa, Hoshang Sahib, Giacomo Merzoni, Leonardo Martinelli, Riccardo Arpaia, Nicholas B. Brookes, Daniele Di Castro, Krzysztof Wohlfeld, Maryia Zinouyeva, Marco Salluzzo, Daniele Preziosi, Giacomo Ghiringhelli
Relevance score 5.327
Primary category cond-mat.supr-con
Published 2026-05-18
Research paradigm Experimental
Sample form Thin Film

Summary

This study systematically compares the spin and orbital excitation properties of undoped superconducting infinite-layer nickelate PrNiO2 and insulating cuprate CaCuO2 using momentum-resolved and polarization-resolved resonant inelastic X-ray scattering (RIXS) measurements. The results show that the in-plane magnetic exchange integral of PrNiO2 (approximately 46 meV) is significantly smaller than that of CaCuO2 (approximately 82 meV), while the out-of-plane exchange integrals are similar (approximately 6–7 meV), indicating that both materials support three-dimensional antiferromagnetic order with comparable three-dimensionality of spin-spin correlations. The orbital excitations (intra-3d transitions) are well described by a single-ion model, but the Ni-dxy peak energy is notably lower than that of Cu-dxy, with opposite dispersion directions—nickelate exhibits orbital excitation propagation driven by nearest-neighbor orbital superexchange coupling, whereas cuprate is dominated by next-nearest-neighbor coupling. Despite a significant difference in charge-transfer energy (larger in the nickelate), the spin and orbital excitation characteristics are generally highly similar, with key distinctions only in the energy and dispersion of the Ni-dxy peak, attributed to differing orbital superexchange coupling mechanisms. This work reveals the core commonalities in magnetism and orbital dynamics between infinite-layer nickelates and cuprates, while also indicating smaller spin fluctuation energies and stronger localization of doped charges on metal sites in the nickelates.

Materials

Methods

Keywords

Highlights

  • Momentum- and polarization-resolved RIXS measurements on nominally undoped, superconducting PrNiO2 are compared with the reference infinite layer cuprate CaCuO2.
  • The Ni-dxy peak lies at significantly lower energy and shows an opposite dispersion to that of Cu-dxy, attributed to different orbital superexchange couplings.
  • The in-plane exchange integrals are approximately half in PrNiO2 (46 meV) compared to CaCuO2 (82 meV), while out-of-plane values are comparable (6-7 meV).
  • In PrNiO2, the magnon peak broadening is nearly constant (≈30 meV) over the explored momentum range, less than half the broadening in superconducting Bi2201 at lowest doping.

Conclusions

  • In PrNiO2, the in-plane magnetic exchange integrals are smaller than in CaCuO2, whereas the out-of-plane values are similar, indicating that both materials support a three-dimensional antiferromagnetic order.
  • The orbital dispersion in the infinite-layer nickelate primarily involves nearest neighbor orbital superexchange interaction, which in cuprates is strongly hampered by coupling to magnons.
  • Infinite-layer nickelates are unconventional superconductors closely related to cuprates, but lacking some of the ingredients that enhance Tc in the latter.
  • Self-doping has a much milder impact on spin order than chemical doping, endowing infinite-layer nickelates with a non-disruptive way to achieve superconductivity which is absent in copper oxides.

Main claims

  • In-plane magnetic exchange integral in PrNiO2 (≈46 meV) is about half that in CaCuO2 (≈82 meV), but out-of-plane exchange is similar (≈6-7 meV), indicating comparable three-dimensional antiferromagnetic correlations.
    • Evidence: LSW fits to RIXS spin wave dispersions give J1 values; out-of-plane exchange Jc is comparable
  • Orbital excitations show distinct differences: Ni-dxy peak is lower in energy (1.29 eV) than Cu-dxy (1.65 eV) and exhibits opposite dispersion direction.
    • Evidence: RIXS maps show dispersion of dxy peak; PNO dispersion is maximal at Γ and decreases; CCO is minimal at Γ and increases
  • The opposite orbiton dispersion is due to dominant nearest-neighbor orbital superexchange in PNO versus next-nearest-neighbor in CCO, originating from different covalency and charge-transfer energy.
    • Evidence: Three-band model calculations give NN orbital exchange dominant in PNO; NNN dominant in CCO; NN orbiton hopping in PNO enabled by Hund's exchange

Workflow

  • sample_preparation — High-quality infinite-layer thin films of both systems.
    • Materials: PrNiO2 thin films; CaCuO2 thin films
    • Methods: pulsed laser deposition; topotactic reduction for PrNiO2
    • Observations: PrNiO2 is superconducting (T_c ≈10 K); CaCuO2 is insulating
  • rixs_measurements — Comparative spin and orbital excitation spectra.
    • Materials: Ni-L3 edge for PrNiO2; Cu-L3 edge for CaCuO2
    • Methods: momentum-resolved RIXS with polarization analysis; ERIXS spectrometer at ESRF ID32
    • Observations: spin excitations up to ≈230 meV (PNO) and ≈320 meV (CCO); orbital excitations at 1-3 eV
  • spin_wave_analysis — In-plane exchange smaller in nickelate; out-of-plane similar.
    • Materials: RIXS dispersion data
    • Methods: linear spin wave theory with SpinW
    • Observations: J1 ≈46 meV (PNO) vs 82 meV (CCO); Jc ≈6 meV (PNO) vs 7 meV (CCO)
  • orbital_excitation_analysis — Orbiton propagation driven by nearest-neighbor orbital superexchange in PNO vs next-nearest-neighbor in CCO.
    • Materials: RIXS orbital excitation data
    • Methods: single-ion cross-section calculations; orbiton dispersion fitting
    • Observations: dxy peak at 1.29 eV (PNO) vs 1.65 eV (CCO); opposite dispersion direction