Diffraction is an optical phenomenon that is commonly investigated for its applications in many optical systems, such as diffractive optical elements, microscopy, and coronagraphs. Current models for predicting diffraction typically suffer from either efficiency or accuracy. This paper addressed both issues by implementing techniques inspired by Braunbek method and Bluestein method. A modification to the Kirchhoffs boundary conditions is used to improve the theoretical model, and the Chirp-z transform is applied instead of the fast Fourier transform for more flexible calculations. A comparison between diffraction patterns for different models shows that the new method exceeds in accuracy. A comparison of time between numerical methods demonstrates that the chirp-z transform is faster in computation than the fast Fourier transform by about a minute. The method introduced provides many implications, such as the enhancement of dynamic optical systems and the improvement of flexibility in other realms of numerical Fourier transform.
Read full abstract