A stochastic asymptotic-preserving scheme for the bipolar semiconductor Boltzmann-Poisson system with random inputs and diffusive scalings

Liu Liu

TL;DR

This paper introduces a novel numerical method, a stochastic asymptotic-preserving (s-AP) scheme, for simulating the complex behavior of bipolar semiconductors under uncertainty. This method efficiently captures the transition from kinetic to diffusive behavior while accurately handling random variations in material properties, paving the way for more reliable semiconductor device modeling.

What Problem Does This Paper Address?

Semiconductor devices are fundamental to modern technology, but their behavior is governed by complex kinetic equations (like the Boltzmann-Poisson system) that are challenging to solve, especially when accounting for uncertainties. These uncertainties can arise from variations in material properties, manufacturing processes, or environmental factors. Traditional numerical methods struggle to handle both the kinetic nature of charge carriers (electrons and holes) and the diffusive behavior that emerges at larger scales, particularly when these uncertainties are present. Furthermore, existing methods often fail to maintain accuracy across different physical scales (e.g., from microscopic collisions to macroscopic transport), a problem known as the "stiffness" issue. This paper addresses the need for a unified numerical approach that can accurately simulate bipolar semiconductor systems across various scales and under realistic uncertainty conditions.

Methods

The authors develop a stochastic asymptotic-preserving (s-AP) scheme based on the generalized polynomial chaos-stochastic Galerkin (gPC-SG) method. Here's a breakdown of the key techniques:

  • Boltzmann-Poisson System: The core of the problem is the bipolar semiconductor Boltzmann-Poisson system, which describes the evolution of electron and hole distribution functions and their self-consistent electric field.
  • Diffusive Scaling: The paper focuses on the "diffusive scaling" where the system's behavior transitions from kinetic to a simpler drift-diffusion system as a parameter (Knudsen number) approaches zero. Asymptotic-preserving (AP) schemes are designed to maintain accuracy across this transition.
  • Stochastic Galerkin (SG) Method: To handle uncertainties, the SG method is employed. This approach represents uncertain quantities as polynomial expansions in terms of random variables, allowing for a spectral-accuracy treatment of uncertainties.
  • Asymptotic-Preserving (AP) Property: The scheme is designed to be "asymptotic-preserving," meaning it remains accurate and stable even when the system's behavior shifts towards diffusion (small Knudsen number). This is crucial for capturing the macroscopic behavior of semiconductors.
  • Stochastic AP (s-AP) Property: This is an extension of the AP concept to systems with uncertainties. The authors formally prove that their scheme satisfies this property, ensuring that the numerical solution of the uncertain kinetic system converges to the solution of the uncertain drift-diffusion system as the kinetic effects become negligible.
  • Hypocoercivity Theory: This mathematical tool is used to analyze the convergence rate of the numerical scheme and the analytical solutions, demonstrating an exponential decay in time towards equilibrium and an exponential decay in numerical error in the random space.
  • Even-Odd Decomposition: A technique used to split the kinetic equations into separate equations for even and odd velocity components, which helps in deriving the AP schemes.

Datasets

Not discussed in the paper. The paper focuses on developing and validating a numerical method for a physical model, rather than training on or evaluating against specific datasets.

Key Insights and Ideas

  • Unified Treatment of Scales and Uncertainties: The most significant insight is the successful development of a single numerical framework that can handle both the kinetic-diffusive transition and the presence of random inputs in bipolar semiconductor systems. This is a major step towards more realistic simulations.
  • Formal Proof of s-AP Property: The paper provides a formal proof for the stochastic asymptotic-preserving property, which is a strong theoretical guarantee for the scheme's behavior in the diffusive limit under uncertainty.
  • Exponential Convergence: The use of hypocoercivity theory reveals that both the analytical solutions and the numerical errors exhibit exponential convergence towards equilibrium and in the random space, respectively. This indicates high accuracy and efficiency.
  • Sensitivity Analysis: The method allows for sensitivity analysis, helping to understand how different random input parameters affect the device's behavior, which is vital for design and optimization.
  • First of its Kind: The authors highlight that this is the first study to develop and analyze both deterministic and stochastic AP schemes for the bipolar semiconductor Boltzmann-Poisson system.

Key Results

  • Numerical Validation of s-AP Property: Numerical experiments demonstrate that for small kinetic parameters (ɛ), the solutions of the kinetic system closely match the solutions of the corresponding drift-diffusion system, confirming the s-AP property.
  • Accuracy and Efficiency: The gPC-SG scheme shows good accuracy and efficiency in capturing the behavior of the bipolar semiconductor system, both in deterministic and uncertain cases.
  • Spectral Convergence: The analysis confirms uniform spectral convergence in the random space, with errors decaying exponentially in time, for a simplified model without the electric field.
  • Robustness to Random Inputs: Numerical tests (Test 2 a, b, c) show that the gPC-SG method provides accurate results even when doping profiles, collision kernels, or initial data are random, with good agreement with reference solutions obtained by the stochastic collocation method.
  • Fast Convergence with gPC Order: For spectral convergence tests (Test 2 d), errors saturate quickly at modest gPC orders (K=4), indicating that the temporal and spatial discretization become the limiting factors for accuracy, rather than the gPC expansion itself.

Limitations and Future Work

  • Full Model Analysis: The paper acknowledges that a full convergence rate analysis for the complete bipolar Boltzmann-Poisson system (including the self-consistent electric field) is complex and remains a future research direction.
  • Higher-Order Schemes: The current scheme uses a second-order spatial discretization and a first-order time splitting. Future work could involve developing fully second-order schemes for improved temporal accuracy.
  • Disparate Mass Binary Gas/Plasma Systems: The authors suggest extending the AP schemes to describe the dynamics of more complex systems like disparate mass binary gases or plasma systems across various time scales.
  • Sensitivity Analysis for Full Boltzmann Equations: Conducting sensitivity analysis for multi-species full Boltzmann equations with random inputs is proposed as a challenging but important future direction.

Summary updated

Abstract

In this paper, we study the bipolar Boltzmann-Poisson model, both for the deterministic system and the system with uncertainties, with asymptotic behavior leading to the drift diffusion-Poisson system as the Knudsen number goes to zero. The random inputs can arise from collision kernels, doping profile and initial data. We adopt a generalized polynomial chaos approach based stochastic Galerkin (gPC-SG) method. Sensitivity analysis is conducted using hypocoercivity theory for both the analytical solution and the gPC solution for a simpler model that ignores the electric field, and it gives their convergence toward the global Maxwellian exponentially in time. A formal proof of the stochastic asymptotic-preserving (s-AP) property and a uniform spectral convergence with error exponentially decaying in time in the random space of the scheme is given. Numerical experiments are conducted to validate the accuracy, efficiency and asymptotic properties of the proposed method.
BibTeX
@article{liu2019stochastic,
  author = {Liu Liu},
  title = {{A stochastic asymptotic-preserving scheme for the bipolar semiconductor Boltzmann-Poisson system with random inputs and diffusive scalings}},
  journal = {Journal of Computational Physics},
  year = {2019},
  doi = {10.1016/j.jcp.2018.09.055},
}
Figure