Possibility of ferro-octupolar order in Ba$_2$CaOsO$_6$ assessed by X-ray magnetic dichroism measurements

Basic Information

• Paper ID: 2510.00448
• Title: Possibility of ferro-octupolar order in Ba$_2$CaOsO$_6$ assessed by X-ray magnetic dichroism measurements
• Authors: G. Shibata, N. Kawamura, J. Okamoto, A. Tanaka, H. Hayashi, K. Yamaura, H. Y. Huang, A. Singh, C. T. Chen, D. J. Huang, S. V. Streltsov, A. Fujimori
• Classification: cond-mat.str-el (Strongly Correlated Electron Systems), cond-mat.mtrl-sci (Materials Science)
• Publication Date: November 7, 2025
• Paper Link: https://arxiv.org/abs/2510.00448

Abstract

This study explores the possibility of ferro-octupolar order in the cubic double perovskite Ba$_2$CaOsO$_6$ through X-ray absorption spectroscopy (XAS) and X-ray magnetic circular dichroism (XMCD) measurements. The material contains Os$^{6+}$ (5d$^2$) ions and exhibits a "hidden order" phase transition below T$^*$ ~ 50 K. The findings include: (1) the t$_{2g}$-e$_g$ ligand field splitting of Os 5d is approximately 4 eV; (2) the temperature dependence of XMCD spectra is consistent with a residual cubic splitting of ~18 meV in the lowest J$_{\rm eff}$ = 2 multiplet, forming a non-Kramers E$_g$ doublet ground state and T$_{2g}$ triplet excited state; (3) ligand field multiplet calculations under virtual strong magnetic fields indicate that to stabilize ferro-octupolar order, the nearest-neighbor octupolar exchange interaction should be as strong as ~1.5 meV, consistent with theoretical predictions of ~1 meV.

Research Background and Motivation

Research Questions

This study aims to address the microscopic mechanism of "hidden order" in 5d$^2$ electron systems, particularly the nature of the phase transition in Ba$_2$CaOsO$_6$ below T$^*$ ~ 50 K.

Importance of the Problem

1. Frontier exploration of multipolar order physics: 5d electron systems with strong spin-orbit coupling can form higher-order multipolar orders beyond traditional magnetic dipolar order (such as electric quadrupolar and magnetic octupolar), representing an emerging research direction in condensed matter physics.
2. The "hidden order" mystery: Ba$_2$CaOsO$_6$ exhibits an anomalous peak in magnetic susceptibility at 50 K, but neutron diffraction reveals no magnetic Bragg peaks, μ-SR detects only tiny local magnetic moments (~0.2 μ$_B$) and staggered moments (~0.05 μ$_B$), and X-ray diffraction shows the crystal maintains cubic symmetry, ruling out Jahn-Teller distortion from quadrupolar order.
3. Experimental verification of theoretical predictions: Theoretical calculations predict that 5d$^2$ ions in cubic crystal fields may form magnetic octupolar order, but direct experimental evidence is lacking.

Limitations of Existing Methods

1. Blind spots in conventional magnetic probes: Neutron diffraction is insensitive to magnetic octupoles, and μ-SR can only detect local magnetic moments.
2. Uncertainty of indirect evidence: Macroscopic susceptibility and crystal structure alone cannot determine the microscopic nature of the order parameter.
3. Parameter dependence of theoretical calculations: Exchange interaction strengths from DFT and model calculations require experimental validation.

Research Motivation

By combining XMCD technique with ligand field multiplet theory calculations, probing the fine structure of Os 5d electronic states and their magnetic field response, we infer the possibility of octupolar order and its exchange interaction strength, providing experimental constraints on the microscopic mechanism of "hidden order."

Core Contributions

1. First systematic measurement of Os L$_{2,3}$-edge XMCD spectra in Ba$_2$CaOsO$_6$: High-quality XAS and XMCD data at 10 K and 60 K under 7 T magnetic field, revealing extremely strong L$_2$-L$_3$ asymmetry (L$_2$-edge XMCD signal ~1%, L$_3$-edge <0.1%).
2. Determination of electronic structure parameters for Os$^{6+}$ ions: Quantitative determination of t$_{2g}$-e$_g$ ligand field splitting Δ$_{LF}$ ~ 4 eV from XAS, and residual cubic splitting Δ$_c$ ~ 18 meV from XMCD temperature dependence.
3. Establishment of quantitative relationship between octupolar order and exchange interaction: Ligand field multiplet calculations show that to form ferro-octupolar order causing 4 meV non-Kramers doublet splitting (corresponding to T$^*$ ~ 50 K), an equivalent magnetic field ~300 T is required, corresponding to nearest-neighbor exchange interaction J ~ 1.5 meV, consistent with theoretical predictions of ~1 meV.
4. Provision of indirect experimental evidence for octupolar order: Through analysis of field-induced orbital and spin magnetic moments (M$_{\rm orb}$ = -0.016 μ$_B$/Os, M$_{\rm spin}^{\rm eff}$ = 0.058 μ$_B$/Os) and large magnetic dipolar term M$_T$ = 0.158, confirming anisotropic spin distribution caused by strong spin-orbit coupling, supporting the octupolar order picture.

Detailed Methodology

Experimental Techniques

X-ray Absorption Spectroscopy and Magnetic Circular Dichroism Measurements

Experimental Setup:

• Measurements performed at BL39XU beamline of SPring-8 synchrotron radiation facility in Japan
• Superconducting magnet providing up to 7 T magnetic field parallel to X-ray beam direction
• Diamond phase retarder switching circular polarization at ~30 Hz frequency
• Silicon drift detector (SDD) collecting signal in partial fluorescence yield (PFY) mode

Measurement Strategy:

• Monitoring Os L$_α$ (3d → 2p$_{3/2}$) and L$_β$ (3d → 3p$_{1/2}$) fluorescence intensities, corresponding to L$_3$ and L$_2$ edges respectively
• Averaging XAS and XMCD spectra at B = ±7 T to eliminate spurious signals from optical path differences
• Non-paralyzable model for SDD dead time correction
• Probing depth ~several micrometers, no surface treatment required

Data Processing:

1. XAS spectrum normalization: Setting L$_3$ and L$_2$ edge jump height ratio to 2:1
2. Background subtraction: Fitting two Lorentz functions and their integrals (arctangent functions), extracting Lorentz component
3. XMCD and sum rules analysis: Extracting orbital magnetic moment M$_{\rm orb}$ and effective spin magnetic moment M$_{\rm spin}^{\rm eff}$ using sum rules:

$M_{\rm orb} = -\frac{4\int_{L_3+L_2}(\mu^+ - \mu^-)d\nu}{3\int_{L_3+L_2}(\mu^+ + \mu^-)d\nu}(10-N_d)$

$M_{\rm spin}^{\rm eff} \equiv M_{\rm spin} + \frac{7}{2}M_T = -\frac{2\int_{L_3}(\mu^+ - \mu^-)d\nu - 4\int_{L_2}(\mu^+ - \mu^-)d\nu}{\int_{L_3+L_2}(\mu^+ + \mu^-)d\nu}(10-N_d)$

where M$_T$ ≡ -2∑$_i$⟨s$_i$ - 3(s$_i$·r$_i$)r$_i$/r$_i^2$⟩ is the magnetic dipolar term, reflecting anisotropy of spin distribution.

Theoretical Calculation Methods

Ligand Field Multiplet Theory

Calculation Framework: Using XTLS 8.5 software package for multiplet calculations of 5d$^2$ electrons in cubic crystal field.

Hamiltonian: $\hat{H} = \hat{H}_{\rm Coulomb} + \hat{H}_{\rm SOC} + \hat{H}_{\rm LF} + \hat{H}_{\rm Zeeman}$

• Coulomb interaction: Using Slater integrals F$^k$ and G$^k$ (k=0,2,4), Hund coupling J$_H$ = $\frac{3}{49}$F$^2$ + $\frac{20}{441}$F$^4$ = 0.27 eV
• Spin-orbit coupling: ζ = 0.33 eV (ζ' = -ζ for t$_{2g}$ shell)
• Ligand field: Cubic field splitting Δ$_{LF}$ = 4.3 eV (t$_{2g}$-e$_g$)
• Zeeman term: -μ$_B$B·(L + 2S), B adjustable to include molecular field effects

Parameter Reduction: Wave function extension in solids causes reduction of Slater integrals and SOC parameters, using reduction factors R$_{\rm Slater}$ = 0.6 (40% reduction) and R$_{\rm SOC}$.

Spectral Broadening: Only Lorentz broadening considered (HWHM = 2.7 eV, corresponding to L$_{2,3}$ core hole lifetime), no Gaussian broadening added.

Initial State Treatment: Lowest 5 J$_{\rm eff}$ = 2 states weighted according to Boltzmann distribution.

Technical Innovations

1. High-precision XMCD measurements: Through ±7 T magnetic field averaging and dead time correction, achieving magnetic moment data with systematic error <10%.
2. Virtual strong magnetic field calculation strategy: Introducing effective magnetic fields far exceeding experimental values (B ~ 12±2 T, including molecular field) in ligand field calculations to simulate electronic structure under octupolar order, establishing quantitative relationship between order parameter and exchange interaction.
3. Multiplet analysis framework: Combining XAS to determine crystal field parameters, XMCD temperature dependence to determine excitation gap, and virtual magnetic field calculations to infer exchange interaction, forming a complete chain from experiment to microscopic parameters.

Experimental Setup

Sample Preparation

High-quality Ba$_2$CaOsO$_6$ polycrystalline samples synthesized using methods described in reference 15, confirmed by X-ray diffraction to have cubic double perovskite structure (space group Fm$\bar{3}$m).

Measurement Conditions

• Temperature: 10 K and 60 K (using closed-cycle cryostat)
• Magnetic field: ±7 T, parallel to X-ray propagation direction
• Photon energy range: Os L$_3$ edge ~10.87 keV, L$_2$ edge ~12.39 keV
• Energy resolution: Determined by monochromator bandwidth and core hole lifetime broadening

Data Quality Control

• Averaging spectra at B = +7 T and -7 T to eliminate systematic errors
• Subtracting white line background and EXAFS oscillations
• Normalizing edge jump height ensuring L$_3$:L$_2$ = 2:1
• Removing magnetic EXAFS oscillations (near 12.40 and 12.41 keV)

Experimental Results

Main Results

XAS Spectrum Characteristics (Figures 2, 3)

1. Double-peak structure: Both L$_3$ and L$_2$ edges show clear double peaks, with low-energy peak corresponding to 2p → t$_{2g}$ transition and high-energy peak to 2p → e$_g$ transition, peak separation ~4 eV.
2. Ligand field splitting: Quantitative fitting yields Δ$_{LF}$ = 4.3 eV, consistent with O K-edge XAS measurement (~4 eV).
3. Spin-orbit coupling: Optimal fitting parameter ζ = 0.33 eV, approximately 60% of free ion value (accounting for solid-state effects).

XMCD Spectrum Characteristics (Figures 2, 3)

1. Extremely strong L$_2$-L$_3$ asymmetry:
   • L$_2$-edge XMCD intensity ~1% (relative to XAS)
   • L$_3$-edge XMCD intensity <0.1%
   • This asymmetry also observed in ferromagnetic Os-containing double perovskites
2. Magnetic moment values (T = 10 K, B = 7 T):
   • Orbital magnetic moment: M$_{\rm orb}$ = -0.016 μ$_B$/Os
   • Effective spin magnetic moment: M$_{\rm spin}^{\rm eff}$ = 0.058 μ$_B$/Os
   • |M$_{\rm orb}$/M$_{\rm spin}^{\rm eff}$| = 0.28 (experimental value)
   • Theoretical calculation (B = 7 T + 12 T molecular field): |M$_{\rm orb}$/M$_{\rm spin}^{\rm eff}$| = 0.297
3. Magnetic dipolar term:
   • M$_T$ = 0.158 (separated through multiplet calculations)
   • Large M$_T$/M$_{\rm spin}^{\rm eff}$ ratio reflects anisotropic spin distribution caused by strong SOC
4. Temperature dependence:
   • M$_{\rm spin}^{\rm eff}$ decreases by 0.004 μ$_B$ from 10 K to 60 K
   • Ligand field calculations attribute this to thermal excitation from E$_g$ ground state to T$_{2g}$ excited state

Energy Level Structure (Figure 1)

1. Splitting of J$_{\rm eff}$ = 2 multiplet:
   • Residual cubic splitting Δ$_c$ = 18 meV (theoretical calculation value)
   • Ground state: non-Kramers E$_g$ doublet
   • Excited state: T$_{2g}$ triplet
2. Magnetic field effects (B = 7 T + 12 T molecular field):
   • E$_g$ doublet splitting: Δ$E_{E_g}$ ~ 1 μeV (∝ B$^3$)
   • T$_{2g}$ triplet Zeeman splitting: Δ$E_{T_{2g}}$ = 0.74 meV (∝ B)

Key Findings

Indirect Evidence for Octupolar Order

1. Nature of non-Kramers doublet:
   • Absence of Jahn-Teller distortion rules out quadrupolar order
   • Magnetic field along ⟨111⟩ direction produces pure octupolar state |ψ$_{g,±}$⟩ ≡ $\frac{1}{\sqrt{2}}$(|ψ$_{g,↑}$⟩ ± i|ψ$_{g,↓}$⟩)
   • |ψ$_{g,↑}$⟩ ≡ |J$_z^{\rm eff}$ = 0⟩, |ψ$_{g,↓}$⟩ ≡ $\frac{1}{\sqrt{2}}$(|J$_z^{\rm eff}$ = 2⟩ + |J$_z^{\rm eff}$ = -2⟩)
2. Quantitative estimation of exchange interaction:
   • If T$^*$ ~ 50 K corresponds to E$_g$ doublet splitting ΔE$_{E_g}$ ~ 4 meV
   • From ΔE$_{E_g}$ ∝ B$^3$, equivalent magnetic field B ~ 300 T is required
   • 12 nearest-neighbor Os ions in fcc lattice (Figure 4)
   • Exchange interaction J ~ 300 T / 12 ~ 1.5 meV
   • Consistent with DFT calculation (J ~ 1 meV) and Schrieffer-Wolff transformation results
3. Origin of small magnetic moments:
   • Exchange field induces small moments through E$_g$-T$_{2g}$ hybridization
   • Consistent with ~0.05 μ$_B$ staggered moment detected by zero-field μ-SR

Theory-Experiment Comparison

Related Work

Multipolar Physics in 5d Electron Systems

1. J$_{\rm eff}$ = 1/2 Mott insulators: Ir$^{4+}$ (5d$^5$) oxides such as Sr$_2$IrO$_4$, with strong SOC leading to novel magnetic ground states.
2. Kitaev quantum spin liquid: Ru$^{3+}$ (4d$^5$) honeycomb compounds α-RuCl$_3$.
3. Theoretical predictions for 5d$^2$ systems: Chen & Balents (2011), Paramekanti et al. (2020) and others predicted magnetic octupolar order.

Previous Studies on Ba$_2$CaOsO$_6$

1. Macroscopic properties (Thompson et al., 2014):
   • Magnetic susceptibility shows sharp peak at T$^*$ ~ 50 K
   • Neutron diffraction shows no magnetic Bragg peaks
   • μ-SR detects ~0.2 μ$_B$ local moment and ~0.05 μ$_B$ staggered moment
2. Crystal structure (Hirai et al., 2020):
   • X-ray diffraction confirms cubic symmetry maintained at low temperature
   • Rules out Jahn-Teller distortion from quadrupolar order
3. Spectroscopic studies (Okamoto et al., 2025; Frontini et al., 2025):
   • O K-edge XAS and RIXS determine Δ$_{LF}$ ~ 4 eV
   • RIXS observes fine structure of t$_{2g}^2$ multiplet
4. Theoretical calculations (Voleti et al., 2023):
   • DFT + Schrieffer-Wolff transformation predicts J ~ 1 meV
   • Proposes detection of octupolar order through Janus impurities

Unique Contributions of This Work

• First XMCD experiment: Providing direct information on Os 5d electronic states under magnetic field
• Quantitative exchange interaction: Establishing connection between J and experimental observations through virtual field calculations
• Multi-technique fusion: XAS + XMCD + ligand field theory forming complete evidence chain

Conclusions and Discussion

Main Conclusions

1. Electronic structure confirmation: Os$^{6+}$ ions in Ba$_2$CaOsO$_6$ have 5d electrons in strongly correlated localized state, with ligand field splitting Δ$_{LF}$ ~ 4 eV and spin-orbit coupling ζ ~ 0.33 eV.
2. Energy level picture: Lowest J$_{\rm eff}$ = 2 multiplet splits into non-Kramers E$_g$ doublet ground state and T$_{2g}$ triplet excited state, with residual cubic splitting Δ$_c$ ~ 18 meV.
3. Possibility of octupolar order:
   • Cubic symmetry and absence of Jahn-Teller distortion support magnetic octupolar rather than electric quadrupolar order
   • If T$^*$ ~ 50 K corresponds to ferro-octupolar phase transition, nearest-neighbor exchange interaction J ~ 1.5 meV is required
   • This value is of the same order as theoretical prediction ~1 meV, providing quantitative support for octupolar order
4. Indirect detection capability of XMCD: Although XMCD does not directly probe octupolar moments, by analyzing electronic state response to magnetic field and temperature dependence, one can infer exchange interaction strength of octupolar order.

Limitations

1. Indirect nature of evidence:
   • XMCD measures field-induced magnetic moments, not octupolar moments themselves
   • Requires theoretical calculations (virtual strong field) to establish connection, depending on model assumptions
2. Limitations from polycrystalline samples:
   • Cannot perform angle-dependent measurements to directly verify octupolar state along ⟨111⟩ direction
   • Powder averaging may obscure certain anisotropic information
3. Uncertainty in molecular field:
   • Effective magnetic field B = 12±2 T determined by fitting XMCD intensity, with ±2 T error
   • Leading to ~15% uncertainty in exchange interaction J estimate
4. Absence of direct evidence for order parameter:
   • Cannot directly observe spatial distribution of octupolar order parameter
   • Distinction from quadrupolar order mainly based on elimination (absence of Jahn-Teller distortion)

Future Directions

1. Single crystal studies:
   • Synthesizing Ba$_2$CaOsO$_6$ single crystals for angle-resolved XMCD
   • Large-angle neutron scattering to directly observe octupoles (similar to Ce pyrochlore studies)
2. Impurity doping:
   • Inducing detectable magnetic dipoles through Janus impurities (breaking local symmetry)
   • Verifying theoretical predictions of magnetic moment distribution around impurities
3. Resonant inelastic X-ray scattering (RIXS):
   • Directly measuring E$_g$-T$_{2g}$ excitation spectrum to confirm Δ$_c$
   • Studying dynamics of excited states
4. Pressure and doping control:
   • Exploring stability of octupolar order and phase diagram
   • Studying variation of J with lattice parameter
5. Time-resolved spectroscopy:
   • Probing dynamical response of octupolar order
   • Distinguishing characteristic timescales of different multipolar orders

In-Depth Evaluation

Strengths

1. Methodological innovation:
   • First application of XMCD to octupolar order system, opening new experimental probe pathway
   • Virtual strong magnetic field calculation strategy cleverly establishes quantitative connection between octupolar order and observable quantities
   • Multi-technique fusion (XAS determining crystal field, XMCD determining magnetic response, theory determining exchange interaction) forms complete evidence chain
2. Excellent experimental technique:
   • High-precision XMCD measurements (systematic error <10%)
   • Rigorous data processing (dead time correction, background subtraction, magnetic EXAFS removal)
   • Temperature-dependent measurements providing crucial excitation gap information
3. Tight theory-experiment combination:
   • Ligand field multiplet calculations accurately reproduce XAS and XMCD spectral shapes
   • Exchange interaction J ~ 1.5 meV consistent with DFT prediction ~1 meV at same order of magnitude
   • Good theory-experiment consistency for key ratios like |M$_{\rm orb}$/M$_{\rm spin}^{\rm eff}$|
4. Clear physical picture:
   • Systematic exposition of energy level structure of 5d$^2$ ions in cubic field (Figure 1)
   • Clear distinction between electric quadrupolar and magnetic octupolar responses
   • Schematic diagram of ferro-octupolar order (Figure 4) intuitively shows spatial arrangement
5. Academic integrity:
   • Objectively points out indirect nature of XMCD and limitations of polycrystalline samples
   • Clearly labels sources of systematic errors
   • Proposes verifiable future experimental directions

Shortcomings

1. Absence of direct evidence:
   • Cannot directly observe octupolar order parameter (e.g., through neutron scattering)
   • Distinction from quadrupolar order mainly relies on elimination method
   • Validity of virtual field calculation needs verification by other experiments
2. Single sample form:
   • Only polycrystalline samples studied, cannot perform angle-dependent measurements
   • Cannot directly verify prediction of octupolar state along ⟨111⟩ direction
   • Powder averaging may lose important information
3. Limited temperature range:
   • Only two temperature points measured (10 K and 60 K)
   • Cannot finely track phase transition behavior near T$^*$ ~ 50 K
   • Lacking critical behavior analysis
4. Model dependence:
   • Exchange interaction estimate depends on accuracy of ligand field multiplet model
   • Slater integral reduction factor, molecular field magnitude and other parameters have some arbitrariness
   • Does not discuss other possible "hidden order" mechanisms (e.g., orbital currents)
5. Quantitative precision:
   • J ~ 1.5 meV estimate based on simplified equivalent magnetic field model
   • Does not consider longer-range interactions and quantum fluctuations
   • Molecular field ±2 T uncertainty leads to ~15% error in J

Impact

1. Contribution to the field:
   • Provides important experimental constraints for multipolar physics in 5d$^2$ systems
   • Promotes application of XMCD in multipolar order research
   • Provides quantitative parameters (J ~ 1.5 meV) for theoretical models
2. Practical value:
   • Virtual magnetic field calculation method can be generalized to other multipolar order systems
   • Provides motivation and parameter guidance for single crystal growth and neutron scattering experiments
   • Experimental foundation for impurity detection scheme (reference 19)
3. Reproducibility:
   • Detailed experimental condition description (Appendix A)
   • Clear theoretical calculation parameters (Appendix B)
   • Using commercial software XTLS 8.5, calculations can be reproduced
   • Sample preparation referenced in 15
4. Academic impact:
   • Published on high-impact preprint platform arXiv
   • International collaboration (Japan, Taiwan, Russia)
   • Cites latest progress in the field (2025 references)
   • May stimulate follow-up research on single crystal synthesis and neutron scattering

Applicable Scenarios

1. Applicable systems:
   • Double perovskites containing 5d$^2$ (Os$^{6+}$, Re$^{5+}$) ions
   • Strong SOC systems maintaining cubic symmetry
   • Correlated electron materials exhibiting "hidden order"
2. Experimental technique generalization:
   • Synchrotron XMCD studying multipolar order
   • Ligand field multiplet theory analyzing transition metal oxides
   • Virtual field method estimating exchange interactions
3. Limiting scenarios:
   • Not applicable to systems already undergoing Jahn-Teller distortion
   • Single crystal samples require angle-dependent measurements
   • May not apply to weak SOC systems (3d electrons)

Key References

7 G. Chen and L. Balents, Phys. Rev. B 84, 094420 (2011) - First theoretical prediction of octupolar order in 5d$^2$ systems

9 A. Paramekanti et al., Phys. Rev. B 101, 054439 (2020) - Theory of octupolar order in d-orbital Mott insulators

12 C. M. Thompson et al., J. Phys.: Condens. Matter 26, 306003 (2014) - Magnetic and structural characterization of Ba$_2$CaOsO$_6$

15 J. Okamoto et al., npj Quantum Mater. 10, 44 (2025) - O K-edge RIXS study

17 D. D. Maharaj et al., Phys. Rev. Lett. 124, 087206 (2020) - Neutron scattering determining Δ$_c$ ~ 20 meV

19 S. Voleti et al., npj Quantum Mater. 8, 42 (2023) - DFT calculation J ~ 1 meV and Janus impurity proposal

Overall Assessment: This is a high-quality experimental condensed matter physics paper that cleverly uses XMCD technique and ligand field theory to provide important indirect evidence and quantitative constraints for ferro-octupolar order in Ba$_2$CaOsO$_6$. Although limited by polycrystalline samples and indirect nature of XMCD, its methodological innovation (virtual strong magnetic field calculation) and rigorous theory-experiment combination open new pathways for multipolar order physics research. The exchange interaction estimate J ~ 1.5 meV is consistent with theoretical predictions, providing important guidance for subsequent single crystal and neutron scattering experiments. The paper's main limitation is the absence of direct observation of octupolar order, but the authors have honestly pointed this out and proposed feasible future directions.