Covalent bonding and magnetism in cuprates

TL;DR

This paper demonstrates that covalent bonding significantly influences magnetism in cuprates, challenging the traditional ionic model, and shows that strong hybridization affects magnetic excitations, which is crucial for understanding high-temperature superconductivity.

Contribution

It provides a detailed neutron scattering study revealing the inadequacy of the ionic model and highlights the importance of covalency in cuprate magnetism.

Findings

Magnetic intensity is strongly affected by Cu 3d and O p hybridization.

The ionic Heisenberg model is insufficient to describe cuprate magnetism.

Strong covalency explains the missing magnetic intensity in INS experiments.

Abstract

The importance of covalent bonding for the magnetism of 3d metal complexes was first noted by Pauling in 1931. His point became moot, however, with the success of the ionic picture of Van Vleck, where ligands influence magnetic electrons of 3d ions mainly through electrostatic fields. Anderson's theory of spin superexchange later established that covalency is at the heart of cooperative magnetism in insulators, but its energy scale was believed to be small compared to other inter-ionic interactions and therefore it was considered a small perturbation of the ionic picture. This assertion fails dramatically in copper oxides, which came to prominence following the discovery of high critical temperature superconductors (HTSC). Magnetic interactions in cuprates are remarkably strong and are often considered the origin of the unusually high superconducting transition temperature, Tc. Here we report a detailed survey of magnetic excitations in the one-dimensional cuprate Sr2CuO3 using inelastic neutron scattering (INS). We show that although the experimental dynamical spin structure factor is well described by the model S=1/2 nearest-neighbour Heisenberg Hamiltonian typically used for cuprates, the magnetic intensity is modified dramatically by strong hybridization of Cu 3d states with O p states, showing that the ionic picture of localized 3d Heisenberg spin magnetism is grossly inadequate. Our findings provide a natural explanation for the puzzle of the missing INS magnetic intensity in cuprates and have profound implications for understanding current and future experimental data on these materials.