Numerical scheme for a spatially inhomogeneous matrix-valued quantum Boltzmann equation

Jianfeng Lu, Christian B. MendlJournal of Computational Physics 2015 · DOI

Introduction

In the study of quantum materials, such as semiconductors and superconductors, understanding the transport and relaxation of particles with internal degrees of freedom is critical. While classical kinetic theory utilizes the standard Boltzmann equation to describe gas density evolution, quantum systems require a more sophisticated approach to account for spin. This is typically modeled using a "matrix-valued" quantum Boltzmann equation (QBE), where the distribution function is represented by a $2 \times 2$ matrix rather than a scalar value.

Despite its theoretical importance, the matrix-valued QBE presents significant computational hurdles. This paper introduces a high-performance numerical scheme designed to solve the spatially inhomogeneous QBE, enabling the simulation of spin-polarized systems across spatial gradients and under various boundary conditions.

The Computational Challenge

The primary difficulty in solving the QBE lies in the collision integral, which models how particles scatter and exchange momentum and spin. This integral is high-dimensional and computationally expensive to evaluate directly. Furthermore, while previous research often focused on "homogeneous" systems—where the state is uniform across space—real-world applications involve "inhomogeneous" environments. These require accounting for spatial variations, external magnetic fields, and boundary interactions, such as the edges of a microchip or the walls of a containment vessel.

Proposed Numerical Framework

The researchers developed a multi-step framework to address these complexities efficiently:

1. Pauli Matrix Representation

To make the physics of spin rotation more intuitive and computationally manageable, the $2 \times 2$ spin-density matrix is decomposed into a scalar component (representing density) and a 3-vector component (the Bloch vector). This decomposition allows for a clearer separation of mass transport and spin dynamics.

2. Fast Spectral Method and Carleman Representation

The core innovation lies in the treatment of the collision integral. The team employed a Fourier spectral method in the velocity variable, utilizing a Carleman representation. This mathematical transformation converts the complex scattering integral into a structure that can be evaluated using Fast Fourier Transforms (FFTs). This reduces the computational complexity from a prohibitive scale to approximately $O(N^3 \log N)$, where $N$ is the number of grid points.

3. Time Splitting and Boundary Conditions

The algorithm employs a time-splitting technique to decouple different physical phenomena—convection (spatial movement), collisions, and magnetic field rotations—into sequential steps within each time increment. To ensure the model's versatility, the researchers implemented three types of boundary conditions:

  • Periodic: For simulating bulk properties without edge effects.
  • Dirichlet: For systems with fixed states at the boundaries.
  • Maxwell: A sophisticated model combining specular reflection and diffusive scattering to simulate realistic surface interactions.

Key Insights and Results

Numerical Accuracy and Stability

The scheme demonstrates spectral accuracy, with relative errors in the collision operator dropping to $10^{-8}$ or lower using modest grid sizes (e.g., $N=32$). Crucially, the method perfectly respects the conservation of mass, momentum, and energy, which is essential for maintaining physical stability over long-duration simulations.

The Quantum H-Theorem

The simulations provide a numerical validation of the "Quantum H-Theorem." The system’s entropy was shown to increase monotonically until it reached a Fermi-Dirac equilibrium state, confirming that the numerical scheme accurately captures the underlying thermodynamics of the Hubbard model.

Relaxation Dynamics

A significant observation from the simulations is the distinction between local and global relaxation. The system tends to reach a local equilibrium (matching local density and temperature) quite rapidly. However, achieving global equilibrium across the entire spatial domain is a much slower process, governed by the transport properties of the system.

Magnetic Field Effects

The study also explored the impact of external magnetic fields, which act as a unitary rotation on the spin vector. The results show that while the magnetic field changes the direction of the spin, it does not alter the overall energy or density, consistent with theoretical expectations.

Conclusion and Future Directions

This work provides a robust foundation for simulating spin-dependent transport in quantum systems. By overcoming the computational bottlenecks of the collision integral, the researchers have enabled high-fidelity modeling of inhomogeneous quantum gases.

Future work will focus on extending the current 1D2V (one spatial, two velocity dimensions) implementation to a full 3D3V model for more realistic material simulations. Additionally, the authors suggest that this high-fidelity solver can be used to derive and validate simpler "hydrodynamic" equations, which would offer even faster computation for large-scale engineering applications in quantum technology.

Summary updated

BibTeX
@article{lu2015numerical,
  author = {Jianfeng Lu and Christian B. Mendl},
  title = {{Numerical scheme for a spatially inhomogeneous matrix-valued quantum Boltzmann equation}},
  journal = {Journal of Computational Physics},
  year = {2015},
  doi = {10.1016/j.jcp.2015.03.020},
}
Figure