Session Index

An AR display is integrated with a full holographic curved waveguide combiner. Ghosting phenomena can occur in curved waveguides due to the path of the incident light. To design efficiently and to accurately calculate the dimensions of Volume Holographic Optical Elements (VHOE) to avoid such ghosting, it is necessary to establish a mathematical model. This model uses the ABCD matrix to calculate the position of total internal reflection each time when the light travels through the waveguide.

  Preview abstract

As applying volume holographic optical elements (VHOEs) on the light guide in the see-through in the mixed-reality (MR) glasses, the color breaking in different view angles has somehow become a serious drawback in the practical applications and induces urgent need to be solved. In this work, we extensively explored with the color breaking effect according to the wavelength --dependent and angle--dependent diffraction efficiency spectrum of VHOEs for MR displays. As considering the color metamerism, a certain possible illumination light sources for reduction of the color breaking within FOVs can then be concluded. Some corresponding experiments are also performed for verification.

  Preview abstract

This study presents the development of a near-eye display with a field-of-view extension array system. It is composed of three volume holographic optical elements and light guides. Due to the volume holographic optical elements' ability to record interference fringes, we move the position of the microscopic lens to create interference fringes with varying focal lengths. This creates different focal lengths of the volume holographic optical elements, thus generating varying levels of field-of-view magnification.

  Preview abstract

Two-dimensional transition metal dichalcogenides (2D TMDCs) have emerged as promising semiconductor materials due to their remarkable properties, including an ultrathin atomic layer structure and unique valley-spin characteristics. These materials have the potential to revolutionize semiconductor technology beyond silicon. However, their production methods, such as chemical vapor deposition (CVD) and mechanical exfoliation, have limitations. To address these challenges, this study explores the application of deep learning, specifically a Transformer model, for the automated identification of monolayer 2D TMDCs. The novel approach leverages semantic representations to enhance accuracy and adaptability, providing a promising solution for the mass production of high-quality 2D TMDCs.

  Preview abstract

A metalens is an ultrathin, flat lens made from metasurface that manipulates light and offers compactness and aberration correction. Though there has not been any studies on the metalens disclosure for optical distortion optimization. Here, we propose that it can be addressed by co-optimizing the distortion and the optical performance (MTF/focusing efficiency) in the wide-angle metalens(field of view 160°). A single-element wide-angle metalens with correction of f-theta distortion will be particularly optimized. It assists in enlarging the pixels per degree (PPD) at the image corner.

  Preview abstract

In previous publication, we design a liquid crystal (LC) lens set consisting of three polarization switching LC lenses for varifocal images and vision corrections with fast response time. We demonstrate AR and VR systems by adopting the LC lens set to solve VAC as well as vision problem. However, there is some scattering at the edge of the lens due to the excessive thickness. It leads to a decrease in the effective aperture size and field of view. For increasing the aperture size, we conducted experiments to optimize the polarization-switching LC lens.

  Preview abstract

Recent advances in deep learning have opened up exciting possibilities in exploring 2D materials for photonic devices, known for their unique light emission and modulation properties. However, the manual identification process using optical microscopy is time-consuming. To address this, we leveraged the power of deep learning with Mask-RCNN, a state-of-the-art AI model. Training on a curated dataset of 100 labeled 2D material images, we obtained encouraging results. This automated approach offers an efficient solution, eliminating labor-intensive inspections and accelerating the development of 2D materials in photonics applications.

  Preview abstract