High fidelity laser beam shaping using liquid crystal on silicon spatial light modulators as diffractive neural...

Area of Science:

• Optics
• Photonics
• Machine Learning

Background:

• Liquid crystal on silicon spatial light modulators (LCoS-SLMs) enable dynamic laser beam shaping for arbitrary intensity distributions.
• Beam shaping with LCoS-SLMs is prone to artifacts like astigmatism and 0-th order diffraction.
• These artifacts arise from non-normal incidence, pixel crosstalk, and cover glass reflections.

Purpose of the Study:

• To introduce a novel method for compensating inherent LCoS-SLM properties during laser beam shaping.
• To address astigmatism and 0-th order diffraction artifacts.
• To leverage a diffractive neural network approach for improved beam shaping accuracy.

Main Methods:

• Modeling the LCoS-SLM as a diffractive neural network.
• Developing compensation strategies for astigmatism and 0-th order diffraction.
• Implementing and validating the proposed method for laser beam shaping.

Main Results:

• Demonstrated effective compensation for astigmatism and 0-th order diffraction artifacts.
• Achieved more accurate and artifact-free laser beam shaping.
• Validated the diffractive neural network approach for LCoS-SLM characterization and correction.

Conclusions:

• The diffractive neural network approach offers a robust method to overcome limitations in LCoS-SLM based beam shaping.
• Accurate compensation of device-specific properties is crucial for high-fidelity laser beam shaping.
• This work advances the application of LCoS-SLMs in fields requiring precise optical control.