Generalized binary spiral zone plates with a single focus obtained by feedforward neural network.

Area of Science:

• Optics and Photonics
• Artificial Intelligence in Optics

Background:

• Traditional spiral zone plates (SZPs) generate optical vortices but suffer from multiple foci, limiting applications.
• Existing methods to eliminate high-order foci involve binary structures with sinusoidal transmittance, which are not always optimal.

Purpose of the Study:

• To propose and design generalized binary spiral zone plates (GBSZPs) capable of producing a single focus.
• To utilize feedforward neural networks (FNNs) for designing GBSZPs with suppressed high-order foci.

Main Methods:

• Concept development of generalized binary spiral zone plates (GBSZPs).
• Training a feedforward neural network (FNN) to map structural parameters to focus intensities.
• Predicting GBSZP structural parameters for single-focus generation using the trained FNN.

Main Results:

• Simulations and experiments demonstrated that GBSZPs can reduce high-order foci intensities to as low as 0.2% of the first-order focus.
• The optimized GBSZP structure exhibits a radial transmittance function different from the traditional sinusoidal form.
• This indicates that non-sinusoidal radial profiles can also effectively eliminate high-order foci.

Conclusions:

• The proposed GBSZPs, designed via FNNs, offer a novel approach to generating single-focus optical vortices.
• The findings provide a new direction for designing binary zone plates with improved performance.
• This research has potential implications for advancing optical image processing and quantum computation.