1: \begin{abstract}

2: Computing the dispersion relation for two-dimensional photonic crystals is a notoriously

3: challenging task: It involves solving parameterized Helmholtz eigenvalue problems with

4: high-contrast coefficients. %The existing global polynomial interpolation method suffers

5: %from slow convergence due to the presence of unknown singular points \cite{wang2023dispersion}.

6: To resolve the challenge, we propose a novel $hp$-adaptive sampling scheme that can detect

7: singular points via adaptive mesh refinement in the parameter domain, and meanwhile, allow for

8: adaptively enriching the local polynomial spaces on the elements that do not contain singular

9: points. In this way, we obtain an element-wise interpolation on an adaptive mesh. We derive

10: an exponential convergence rate when the number of singular points is finite, and a

11: first-order convergence rate otherwise. Numerical tests are provided to illustrate its performance.\\

12: \textbf{Key words:} photonic crystals, band function reconstruction, adaptive mesh, element-wise interpolation

13: \end{abstract}

14: