Thermodynamic properties of the Yb2Ti2O7 pyrochlore as a function of temperature and magnetic field: validation of a quantum spin ice exchange Hamiltonian

TL;DR

This study validates a quantum spin ice Hamiltonian for Yb2Ti2O7 by comparing calculated thermodynamic properties with experimental data, confirming the effectiveness of high-field neutron scattering for determining exchange constants.

Contribution

The paper demonstrates that an effective anisotropic-exchange spin-1/2 Hamiltonian accurately describes Yb2Ti2O7's thermodynamics, validated through numerical calculations and experimental comparison.

Findings

Excellent agreement between calculated and experimental magnetization data.

High-field neutron scattering fitting effectively determines exchange parameters.

RPA analysis does not accurately describe low-temperature correlations.

Abstract

The thermodynamic properties of the pyrochlore Yb2Ti2O7 material are calculated using the numericallinked-cluster (NLC) calculation method for an effective anisotropic-exchange spin-1/2 Hamiltonian with parameters recently determined by fitting the neutron scattering spin wave data obtained at high magnetic field h. Magnetization, M(T,h), as a function of temperature T and for different magnetic fields h applied along the three high symmetry directions [100], [110] and [111], are compared with experimental measurements on the material for temperature T>1.8K. The excellent agreement between experimentally measured and calculated M(T,h) over the entire temperature and magnetic field range considered provides strong quantitative validation of the effective Hamiltonian. It also confirms that fitting the high-field neutron spin wave spectra in the polarized paramagnetic state is an excellent method for determining the microscopic exchange constants of rare-earth insulating magnets that are described by an effective spin-1/2 Hamiltonian. Finally, we present results which demonstrate that a recent analysis of the polarized neutron scattering intensity of Yb2Ti2O7 using a random phase approximation (RPA) method [Chang et al., Nature Communications {3}, 992 (2012)] does not provide a good description of M(T,h) for $T\lesssim 10$ K, that is in the entire temperature regime where correlations become non-negligible.