Ultrafast dynamics of moments in bulk ferromagnets

Basic Information

• Paper ID: 2502.07375
• Title: Ultrafast dynamics of moments in bulk ferromagnets
• Authors: Mouad Fattouhi, Pascal Thibaudeau, Liliana D. Buda-Prejbeanu
• Classification: cond-mat.mes-hall cond-mat.mtrl-sci physics.app-ph
• Publication Date: May 22, 2025 (arXiv preprint)
• Paper Link: https://arxiv.org/abs/2502.07375

Abstract

This paper develops a robust and efficient model for investigating the ultrafast dynamics of magnetic materials excited by laser pulses, integrating the dynamic Landau-Lifshitz-Bloch equation, quantum thermal bath, and two-temperature model. The model has been successfully applied to three representative materials from the literature: nickel, cobalt, and iron. Furthermore, analysis of the ultrafast dynamic susceptibility tensor reveals that off-diagonal elements display specific characteristics depending on the presence or absence of a continuous external magnetic field.

Research Background and Motivation

Core Problems Addressed

1. Precise control of thermal fluctuations: In miniaturized magnetic devices, precise control of thermal fluctuations in relevant physical quantities to achieve reliable operation remains a challenge
2. Ultrafast magnetization dynamics modeling: Due to complex processes involving different length and time scales, no single unified model for ultrafast magnetization dynamics currently exists
3. Computational efficiency issues: Existing stochastic Landau-Lifshitz-Gilbert (sLLG) equations require multiple stochastic realizations and averaging, resulting in low computational efficiency

Research Significance

1. Ultrafast spintronics development: Femtosecond laser excitation schemes can rapidly generate spin currents, entering the field of ultrafast spintronics
2. THz spintronic devices: Paving the way for development of THz spintronic devices, representing the first practical application of ultrafast spin physics
3. Urgent need for theoretical understanding: Optimization and reliable control of ultrafast magnetic devices can only be achieved through theoretical understanding

Limitations of Existing Methods

1. Classical LLB equation: Requires prior knowledge of precise thermodynamic functions, including temperature dependence of exchange stiffness, spontaneous magnetization, and damping parameters
2. Computational cost of stochastic methods: sLLG equations require numerous stochastic realizations to obtain reliable averages
3. Thermal coupling issues: The manner of coupling electron system temperature with stochastic or deterministic forms of atomic spin dynamics remains problematic

Core Contributions

1. Dynamic LLB (dLLB) model: Proposes a set of dLLB equations based on statistical averaging procedures, capable of describing first- and second-order moment dynamics of spin variables
2. Quantum thermal bath integration: Combines quantum fluctuation-dissipation theorem with the dLLB model, using Bose-Einstein statistics to replace classical Maxwell-Boltzmann statistics
3. Efficient deterministic method: Achieves over 100-fold computational speedup compared to sLLG equations
4. Ultrafast dynamic susceptibility analysis: First systematic analysis of diagonal and off-diagonal element behavior of the ultrafast dynamic susceptibility tensor

Methodology Details

Task Definition

Investigation of ultrafast demagnetization and remagnetization processes in ferromagnetic materials under laser pulse excitation, including:

• Input: Laser pulse parameters (power density, pulse width, wavelength), material parameters
• Output: Magnetization evolution over time, susceptibility tensor dynamics
• Constraints: Fluctuation-dissipation theorem, energy conservation

Model Architecture

1. dLLB Equation Set

For each atomic site i, the dLLB model reduces to the following equation set:

First-order moment dynamics equation:

(1 + α²) ds_i/dt = ω_i × s_i - α(Σ_i - Tr(Σ_i)1)ω_i - 2D/(1 + α²) s_i

Second-order moment dynamics equation:

(1 + α²) dΣ_i/dt = 2ΩΣ_i - 2α([M^A_i, Σ_i] + Tr(M_i)(Σ_i - 2Γ_i) - (Tr(Σ_i) - 2Tr(Γ_i))M^S_i) - 2D/(1 + α²)(3Σ_i - 1Tr(Σ_i))

Where:

• s_i ≡ ⟨S_i⟩: Average magnetic moment direction
• Σ_i ≡ ⟨S_i ⊗ S_i⟩: Second-order moment tensor
• α: Transverse Gilbert damping parameter
• D: Noise amplitude

2. Quantum Thermal Bath Model

Classical Fluctuation-Dissipation Relation (CFDR):

D = (α/ℏ)k_B T

Quantum Fluctuation-Dissipation Relation (QFDR):

D = (α/ℏ) ∫₀^∞ [ℏω/(exp(ℏω/k_B T) - 1)] g_m(ω, T) dω

Where g_m(ω, T) is the magnon density of states.

3. Two-Temperature Model (2TM)

C_e dT_e/dt = -G_{e-ph}(T_{ph} - T_e) + P(t)
C_{ph} dT_{ph}/dt = G_{e-ph}(T_{ph} - T_e)

Laser excitation term:

P(t) = (μF)/(ℓ_p τ) exp(-(t-t₀)²/(τ²/4log(2)))

Technical Innovations

1. Statistical averaging method: Rigorous statistical averaging procedure based on functional calculus, avoiding the need to provide thermodynamic functions a priori
2. Quantum corrections: Uses fourth-order magnon dispersion relation and magnon density of states with Van Hove singularity parameters
3. Self-consistent closure: Achieves self-consistent closure of hierarchical equations through second-order moment dynamics equations
4. Deterministic solution: Avoids the need for multiple realizations required by stochastic methods

Experimental Setup

Material Parameters

Three typical 3d ferromagnetic metals were investigated:

Simulation Setup

• Ni and Co: 5×5×5 unit cells, N=500 atomic spins
• Fe: 7×7×7 unit cells, N=686 atomic spins
• Boundary conditions: Periodic boundary conditions
• Numerical method: Fourth-order Runge-Kutta scheme

Evaluation Metrics

1. Curie temperature prediction accuracy: Comparison with experimental values
2. Demagnetization time scale: Characteristic time for magnetization recovery after laser pulse
3. Computational efficiency: Speed improvement relative to sLLG method

Experimental Results

Main Results

1. Curie Temperature Prediction

• CFDR method: Systematically overestimates Curie temperature
• QFDR method: Excellent agreement with experimental data
• Improvement mechanism: Quantum thermal bath captures magnon physics at low temperatures

2. Ultrafast Demagnetization

Testing three materials at different laser power densities:

• Ni: Good agreement with experimental data in the range of 22-50 J/m²
• Fe: Accurately reproduces demagnetization and remagnetization processes in the range of 15-60 J/m²
• Co: Captures main dynamical features in the range of 22-50 J/m²

3. Computational Efficiency Improvement

• Speed improvement: Over 100 times faster than sLLG method
• Reason: Deterministic method avoids the need for multiple stochastic realizations

Dynamic Susceptibility Analysis

Without External Magnetic Field

• Diagonal elements: χ_, χ_, χ_ approach 1/3 in the paramagnetic phase (isotropic)
• Off-diagonal elements: χ_, χ_, χ_ remain zero

With External Magnetic Field (B = 1.5T along z)

• Diagonal elements: Behavior similar to field-free case
• Off-diagonal elements: Display peaks and relaxation behavior, reflecting correlations between spatial orientations

Pulse Width Dependence

Under different laser pulse widths (50 fs, 200 fs, 800 fs):

• Diagonal elements: Insensitive to pulse width variations
• Off-diagonal elements: Peak amplitude decreases and width increases with increasing pulse width

Related Work

Classification of Ultrafast Magnetism Theory Models

1. Direct coherent interactions: Direct interaction between laser pulses and spin degrees of freedom
2. Local spin dynamics: Local spin dynamics triggered by laser heating or excitation
3. Non-local angular momentum transfer: Such as superdiffusive spin transport

Existing Thermal Spin Models

• Stochastic methods: Direct simulation of stochastic forms of atomic spins
• Deterministic methods: Such as thermal macrospin models and LLB equations
• Atomic-scale methods: Atomic spin dynamics based on known Hamiltonians

Advantages of This Work

1. Predictive capability: Unlike classical LLB equations, requires no prior knowledge
2. Efficiency: Significant computational speedup compared to sLLG methods
3. Accuracy: Quantum corrections provide more accurate low-temperature and critical behavior

Conclusions and Discussion

Main Conclusions

1. dLLB-QFDR model: The dLLB method successfully combined with quantum thermal bath can accurately predict thermal effects in ferromagnetic materials
2. Curie temperature prediction: Quantum corrections significantly improve predictions of Curie temperature and low-temperature magnetization curves
3. Ultrafast dynamics: The model accurately captures main features of ultrafast demagnetization in 3d ferromagnetic elements
4. Computational advantages: Deterministic method shows significant computational efficiency advantages over stochastic methods

Limitations

1. Mean-field approximation: Neglects spatial correlations, potentially affecting precise description of critical behavior
2. Simplified 2TM: Assumes electron and spin temperatures are identical, which may lack sufficient accuracy
3. Material scope: Currently validated only for simple 3d ferromagnetic metals

Future Directions

1. Extended applications: Method can be extended to ferrimagnets, antiferromagnets, and multilayer structures
2. Improved 2TM: Consider more complex temperature dependence and nonlinear effects
3. Spatial correlations: Include higher-order fluctuations to improve critical behavior description

In-Depth Evaluation

Strengths

1. Theoretical innovation: Statistical averaging derivation of dLLB equations is rigorous and systematic
2. Quantum corrections: Introduction of QFDR significantly improves physical accuracy of the model
3. Computational efficiency: Efficiency improvement from deterministic method has important practical value
4. Comprehensive validation: Sufficient verification across multiple materials and conditions

Weaknesses

1. Theoretical approximations: Mean-field approximation may limit model precision in critical regions
2. Parameter dependence: Model requires multiple material parameters, some of which are difficult to determine experimentally
3. Applicable scope: Currently primarily focused on simple ferromagnetic metals; applicability to complex magnetic materials remains to be verified

Impact

1. Academic contribution: Provides new theoretical tools and computational methods for ultrafast magnetism
2. Practical value: Efficient computational methods facilitate device design and optimization
3. Reproducibility: Detailed method and parameter descriptions facilitate reproduction and application by other researchers

Applicable Scenarios

1. Ultrafast spintronic device design: Theoretical design of THz-band magnetic devices
2. Laser-induced magnetization dynamics research: Fundamental research on femtosecond laser interaction with magnetic materials
3. Thermal effects analysis of magnetic materials: Theoretical prediction of magnetic material properties at finite temperatures

References

The paper cites 73 related references, covering important works in ultrafast magnetism, spin dynamics, statistical physics, and other fields, providing a solid theoretical foundation and experimental comparison data for this research.