Electronic and magnetic properties of light rare-earth cubic Laves compounds derived from XMCD experiments

TL;DR

This paper develops a comprehensive XMCD/XAS framework for light rare-earth cubic Laves-phase intermetallics Nd_1-xPr_xCoNi and Ce_0.25Pr_0.75CoNi, integrating experiment with DFT and crystal-field multiplet theory. It demonstrates that Ni and Co carry finite 3d moments while Nd and Pr 4f moments are crystal-field suppressed, and reveals a tunable mixed-valence Ce ground state dependent on B-site electronegativity, with substantial Ce 4f^0/4f^1 character. The results underscore the importance of accurately determining the number of 3d/4f holes $n_h$ and the limitations of spin sum rules for LREs, providing a robust protocol for interpreting XMCD in light rare-earth intermetallics. Collectively, the work offers a pathway to tune Ce magnetism via composition and establishes quantitative benchmarks for XMCD analyses in these materials.

Abstract

This work presents electronic and magnetic properties of selected members in the cubic Laves phase series Nd1-xPrxCoNi and Ce0.25Pr0.75CoNi, together with the corresponding binary compositions (NdCo2, NdNi2, PrCo2, PrNi2, CeCo2, CeNi2), using soft x-ray absorption spectroscopy, x-ray magnetic circular dichroism (XMCD), density-functional theory, and crystal field multiplet calculations. All transition-metal moments saturate below 1 T, while the rare-earth moments do not saturate even at 5 T, consistent with van Vleck paramagnetic contributions and crystal field suppression. While the sum rules are widely used to extract element-specific magnetic moments from XMCD, we show that for 3d transition metals, their application requires accurate estimates of the number of unoccupied 3d states. We observe a finite magnetic moment on Ni, challenging the common assumption of its nonmagnetic character in Laves phases. The orbital magnetic moments were determined using the spin rules, while the spin moments were estimated from single-ion values from multiplet calculations, due to the invalidity of the spin sum rule for light rare-earth elements. The magnetic moments of Nd and Pr are found to be suppressed relative to their free-ion values, with multiplet theory indicating that this is due to crystal field effects. Our results confirm that Nd and Pr maintain localized 4f3 and 4f2 configurations, respectively, and that their element-specific magnetic moments are robust to rare-earth substitution. Ce, on the other hand, exhibits a tunable mixed-valent ground state with both magnetic 4f1 and nonmagnetic 4f0 components. The relative fraction of these states varies with the electronegativity of the surrounding 3d transition metals, revealing a pathway to tune Ce magnetism via composition. This work establishes a framework for accurately interpreting XMCD in light rare-earth-based intermetallics.

Figures (6)

• Figure 1: (a) Co and (b) Ni experimental $L_{2,3}$ and (c) Nd and (d) Pr simulated and experimental $M_{4,5}$ edges measured at 4.2 K and 5 T. The XAS spectra represents the average of the left and right polarized x-ray absorption, while XMCD shows the difference.
• Figure 2: (a) The orbital (hatched) and spin (filled) moment for Nd, Pr, Co, and Ni. Ni was only measured for NdCoNi and PrCoNi. $\mu_\text{S}$ for Pr and Nd was calculated using multiplet theory, while all other values are found by applying the sum rules to the experimental XMCD data. (b) The total magnetic moment of Nd and Pr per atom, together with the total LRE moment per formula unit.
• Figure 3: The total Co and Ni moment as a function of the $n_h$ chosen in the sum rules. The $n_h$ determined using DFT calculations in this work are shown with a vertical line for each of the edges.
• Figure 4: Field-dependent XMCD measurements performed at 4.2 K, at the energy of the strongest peak for each edge. The curves are scaled using the results from the sum rules at 5 T. Data points close to zero field have been removed.
• Figure 6: Ce $M_{4,5}$ edges measured at 4.2 K and 5 T. XAS shows the average of the right and left polarized x-ray absorption, while XMCD shows the difference.