Semi-Implicit Lax-Wendroff Schemes

• Semi-implicit Lax-Wendroff schemes are time-stepping methods that combine explicit treatment of advection with implicit discretization of stiff source or coupling terms.
• They achieve second-order accuracy and asymptotic-preserving properties by decoupling the time-step and cell size from microscopic or relaxation scales.
• These schemes offer robust stability and efficient simulation of multi-scale phenomena in kinetic, hydrodynamic, and reaction-transport applications.

A semi-implicit Lax–Wendroff scheme refers to a class of time-stepping methods for hyperbolic partial differential equations, particularly conservation laws, in which the Lax–Wendroff procedure is blended with implicit or semi-implicit treatment of stiff source, collision, or coupling terms. By leveraging an implicit or semi-implicit discretization of certain contributions—most commonly relaxation, scattering, or reaction terms—these schemes allow the time-step and cell size to be chosen independently of the underlying microscopic or relaxation scales, removing the severe stiffness constraints that affect fully explicit second-order (Lax–Wendroff-type) methods. Semi-implicit Lax–Wendroff schemes preserve the characteristic second-order (or higher) accuracy of Lax–Wendroff in the advective limit, while maintaining robust stability and accuracy from the ballistic to diffusive regime in multi-scale kinetic, hydrodynamic, and reaction-transport applications.

1. Theoretical Foundation and Derivation

The core of the Lax–Wendroff methodology is a second-order accurate predictor–corrector or Taylor/Cauchy–Kovalevskaya expansion strategy, in which temporal derivatives are systematically replaced using the governing equations. For a typical hyperbolic PDE or kinetic equation, a semi-implicit Lax–Wendroff scheme arises when the temporal discretization distinguishes advection ("clean transport") from stiff source/collision or coupling dynamics:

• The advective term is discretized by central or high-order explicit rules (e.g., midpoint).
• The stiff (e.g., relaxation, scattering, or source) term is discretized (semi-)implicitly, typically via a trapezoidal, backward Euler, or similar rule.
• For interfacial flux evaluation, the kinetic equation is solved again at faces using semi-implicit predictors, coupling advection and relaxation in a consistent fashion.

This results in schemes in which the update for the conservative or kinetic quantity $w^{n+1}$ takes the schematic form: $$\text{New value} = \text{Weighted previous value} + \text{Flux terms evaluated at half-step} + \text{Implicit source/coupling terms}$$ with explicit expressions for the coupling weights, determined by discrete analogs of the relaxation parameters and time-step. For example, the cell-centered update in the gray phonon BTE (Peng et al., 2024) reads: $$f_{i,j,k}^{n+1} = \frac{\tau}{\tau+h}f_{i,j,k}^n + \frac{h}{\tau+h}f^{\text{eq}, n+1}_{i,j,k} + \frac{h}{\tau+h}(f^{\text{eq}, n}_{i,j,k} - f^n_{i,j,k}) - \frac{2 h \tau}{\tau+h}\bigl[ u_k \delta_x f^{n+1/2}_{i,j,k} + v_k \delta_y f^{n+1/2}_{i,j,k} \bigr]$$ with $h = \Delta t/2$, and fluxes evaluated at the half-step.

2. Semi-Implicit Lax–Wendroff for Kinetic and Multiphysics Problems

The semi-implicit Lax–Wendroff scheme has been systematically applied to a variety of kinetic equations, multi-scale transport, and multi-physics contexts:

• Phonon hydrodynamics: Using the Callaway double-relaxation time BTE, a semi-implicit Lax–Wendroff kinetic scheme couples phonon migration (convection, treated by midpoint rule) and normal/resistive scattering (trapezoidal rule), resulting in an update not limited by collision or mean free path (Li et al., 29 Jan 2026). The interfacial fluxes are reconstructed by explicitly resolving the kinetic equation at faces, employing upwind or central stencil for the distribution function and its gradient.
• Electron-phonon coupling: For two-temperature kinetic models, the scheme simultaneously integrates electron and phonon migration, BGK relaxation, and electron–phonon coupling terms. All non-advective (collision and coupling) terms are handled via semi-implicitization within the Lax–Wendroff predictor-corrector framework (Li et al., 1 Feb 2026).
• Classic relaxation-time BGK models: The core strategy also applies to relaxation-time BGK kinetic transport, underlining the asymptotic-preserving property: as the relaxation time becomes vanishingly small, the scheme transitions to a second-order accurate discretization of the corresponding macroscopic (diffusion) equation (Peng et al., 2024).

Numerical tests confirm the ability of these approaches to stably resolve heat conduction, ballistic-to-diffusive transitions, and coupled processes—without the time-step or spatial restrictions that challenge explicit methods.

3. Stability, Accuracy, and Asymptotic-Preserving Properties

The stability and accuracy characteristics of semi-implicit Lax–Wendroff schemes depend on the treatment of the various terms:

• Second-order accuracy: The combination of a midpoint or Lax–Wendroff-like treatment for transport and trapezoidal (or higher-order) implicit rules for sources ensures formal second-order accuracy in both time and space (Peng et al., 2024, Li et al., 29 Jan 2026, Li et al., 1 Feb 2026).
• Stability constraint: The only remaining time-step restriction is the CFL condition for convection,

$$\Delta t < \mathrm{CFL} \cdot \min(\Delta x, \Delta y) / \max(|v_g|),$$

with no requirement that $\Delta t \ll \tau$ or $\Delta x \ll \lambda$ as in explicit schemes; all collision and coupling stiffness is absorbed into the algebraic update (Peng et al., 2024, Li et al., 29 Jan 2026, Li et al., 1 Feb 2026). In the finite-element setting, an additional weighted-norm stability bound may arise, determined by the largest eigenvalue of the implicitization matrix (Vabishchevich, 2017).

• Asymptotic-preserving property: In the diffusive regime ($\mathrm{Kn} \ll 1$), these schemes recover the correct macroscopic limit without resolving stiff kinetic sub-scales; they are thus asymptotic-preserving (Peng et al., 2024, Li et al., 29 Jan 2026, Li et al., 1 Feb 2026).

4. Spatial Discretization and Interface Flux Reconstruction

The spatial discretization in semi-implicit Lax–Wendroff schemes hinges on the following principles:

• Cell-center/finite volume update: The kinetic equation is integrated in conservative form over the grid cell. Advective derivatives are approximated at midpoints, and source terms are discretized semi-implicitly.
• Interfacial fluxes: To close the scheme, the interface distribution is computed by a semi-implicit predictor, resolving the kinetic equation at the interface over the appropriate time interval. Spatial derivatives at faces are reconstructed using second-order upwind for directional streaming or central stencils when appropriate.
• Macroscopic updates: The cell-averaged conserved variables (mass, energy, etc.) are updated by velocity-moment integrals of the interfacial flux, ensuring global conservation without additional closure approximations.

The same approach extends from finite-difference to Galerkin and Path-Conservative variational settings, enabling generalizations to multi-dimensional and multi-physics systems (Gatti et al., 11 Sep 2025).

5. Variants and Extensions

Several generalizations and refinements of the semi-implicit Lax–Wendroff approach have been developed:

• SDC (Spectral Deferred Correction): High-order time integration with semi-implicit Lax–Wendroff-inspired correctors can be systematically constructed. The deferred correction loop iteratively drives the solution toward collocation conditions, and the implicit Lax–Wendroff term is included in the corrector stage to enhance stability and flexibly increase order. Empirical results show $L$-stability up to order 11 with right-Radau quadrature (Stiller, 2024).
• κ-schemes and dimension-by-dimension extensions: Partial Lax–Wendroff expansions and parameter-dependent stencils, including Corner Transport Upwind corrections, enable unconditional stability and third-order accuracy in special velocity regimes (Frolkovič et al., 2018).
• IMEX–Lax–Wendroff–Galerkin methods: In stiff multiphysics PDEs, such as shallow water or lava flow equations, IMEX Runge–Kutta schemes are combined with Lax–Wendroff-inspired Taylor expansions in time to maximize space–time stability. Pseudo-staggered Galerkin and Path-Conservative spatial discretizations are used for achieving robust, well-balanced, second-order accurate, and optimally $L$-stable updates (Gatti et al., 11 Sep 2025).

6. Numerical Performance and Applications

Comprehensive numerical studies demonstrate that semi-implicit Lax–Wendroff schemes:

• Accurately capture non-equilibrium effects (e.g., ballistic heat transport in thin films) and recover diffusive, hydrodynamic, or coupled behavior at coarse grid spacings and large time steps (Peng et al., 2024, Li et al., 29 Jan 2026, Li et al., 1 Feb 2026).
• Achieve second-order convergence in time and space even when $\Delta t \gg \tau$ and $\Delta x \gg \lambda$.
• Permit efficient simulation of multi-scale phenomena—transient grating relaxation, steady/unsteady conduction, electron-phonon coupling, multi-component flows—without tuning of parameters or degradation of accuracy.
• Exhibit robust stability for large numbers of degrees of freedom in high-dimensional settings, with conservative, non-dissipative evolution in the appropriate norms (Vabishchevich, 2017).

7. Comparison to Related Methods

A direct comparison with explicit Lax–Wendroff, Crank–Nicolson, and fully implicit or high-order Padé schemes highlights tradeoffs:

Method	Accuracy	Stability constraint	Conservation
Explicit Lax–Wendroff	2nd order	CFL bound (transport & source)	L²-norm (if lumped)
Semi-implicit Lax–Wendroff	2nd order	CFL bound (transport only)	Weighted norm [E]
Crank–Nicolson (CN)	2nd order	Unconditional	Standard L²-norm
Fourth-order implicit (Padé)	4th order	Unconditional	Higher-weighted norm

Semi-implicit Lax–Wendroff schemes optimize the balance of accuracy, stability, and computational complexity, providing a practical compromise for multi-scale transport problems and stiff coupled systems (Vabishchevich, 2017, Peng et al., 2024, Li et al., 1 Feb 2026, Gatti et al., 11 Sep 2025).