Nonreflecting Boundary Condition for the free Schrödinger equation for hyperrectangular computational domains

Samardhi Yadav,Vishal Vaibhav
2024-08-25
Abstract:In this article, we discuss the efficient ways of implementing the transparent boundary condition (TBC) and its various approximations for the free Schrödinger equation on a hyperrectangular computational domain in $\field{R}^d$ with periodic boundary conditions along the $(d-1)$ unbounded directions. In particular, we consider Padé approximant based rational approximation of the exact TBC and a spatially local form of the exact TBC obtained under its high-frequency approximation. For the spatial discretization, we use a Legendre-Galerkin spectral method with a boundary-adapted basis to ensure the bandedness of the resulting linear system. Temporal discretization is then addressed with two one-step methods, namely, the backward-differentiation formula of order 1 (BDF1) and the trapezoidal rule (TR). Finally, several numerical tests are presented to demonstrate the effectiveness of the methods where we study the stability and convergence behaviour empirically.
Numerical Analysis,Computational Physics
What problem does this paper attempt to address?
The problem that this paper attempts to solve is: in the numerical solution of the free Schrödinger equation, how to effectively implement the transparent boundary condition (TBC), especially on high - dimensional (such as two - dimensional and three - dimensional) hyper - rectangular computational domains. Specifically, the article focuses on how to deal with the non - locality in time and space under the condition of periodic boundary conditions in \(d - 1\) unbounded directions. ### Main research contents 1. **Efficient implementation of transparent boundary conditions (TBC)**: - The article proposes a rational approximation method based on Padé approximation for the efficient implementation of transparent boundary conditions. - This method can effectively reduce the non - locality in time and space, thereby improving the efficiency and stability of numerical calculations. 2. **Extension to high - dimensional problems**: - Extend from rectangular domains to hyper - rectangular domains (such as two - dimensional and three - dimensional), and handle periodic boundary conditions in multi - dimensional cases. - For two - dimensional problems, elliptic and non - elliptic Schrödinger equations are considered, and the form of TBC under high - frequency approximation is discussed. - For three - dimensional problems, these methods are further generalized, and the relevant mathematical derivations and numerical implementations are discussed in detail. 3. **Numerical discretization methods**: - In terms of spatial discretization, the Legendre - Galerkin spectral method is used, and boundary - adapted basis functions are constructed to ensure the bandedness of the linear system. - In terms of time discretization, two single - step methods are adopted: the first - order backward difference formula (BDF1) and the trapezoidal rule (TR), and it is ensured that the discretization of the boundary conditions is consistent with the time step of the internal problem. 4. **Numerical tests and verification**: - Through a series of numerical experiments, the effectiveness and stability of the proposed numerical methods are verified. - The performance of different TBC approximation methods (such as the convolution quadrature method, the traditional Padé method, the new Padé method, and the high - frequency approximation) are compared. ### Formula summary - **Transparent boundary operator**: \[ \sqrt{\partial_t - i\Delta_\Gamma} \] where \(\Delta_\Gamma\) is the Laplace - Beltrami operator. - **Padé approximation**: \[ R_M(z) = b_0 - \sum_{k = 1}^M \frac{b_k}{z + \eta_k^2} \] where \[ b_0 = 2M+ 1, \quad b_k = \frac{2\eta_k^2(1 + \eta_k^2)}{2M + 1}, \quad \eta_k = \tan\left(\frac{k\pi}{2M + 1}\right) \] - **DtN mapping**: \[ \partial_{x_1}u(x, t)+e^{-i\pi/4}(\partial_t - i\beta\partial_{x_2}^2)^{1/2}u(x, t)=0 \] Through these methods, the paper aims to provide an efficient and stable numerical solution framework for high - dimensional Schrödinger equations, especially suitable for hyper - rectangular computational domains with periodic boundary conditions.