| 摘要: | 當前的擴增實境（AR）技術有望將人機互動轉變為一種更為自然的形式。然而，該技術本質上受到光學合成器（Optical Combiner）的限制，這嚴重地約束了可穿戴式、眼鏡形式設備的發展。其根本問題在於投影時的透視清晰度與反射效率之間固有的權衡取捨；傳統的光學元件難以同時滿足在執行光學投影任務時的高反射率，以及在觀看真實世界時的高穿透率。本論文旨在應對此挑戰，透過提出並系統性地優化一種全新的全介電質、多層結構導模共振（Guided-Mode Resonance, GMR）光柵，以實現高性能的光學合成器。
本研究呈現了計算設計與優化的完整流程，所有模擬均透過嚴謹耦合波分析（Rigorous Coupled-Wave Analysis, RCWA）方法進行。研究中採用了系統性且分階段的方法，第一步為開發基礎的單色GMR濾光片，以探討其基本物理定律。此模型隨後進一步發展為更複雜的三色元件，可在實際的斜向入射模式（45度）下運作。設計中亦包含關鍵的功能性薄膜，例如用於減少基板漏光損耗的低折射率氟化鎂（MgF₂）緩衝層，以及一層精心客製化的氟化鎂覆蓋層，其同時作為寬頻抗反射鍍膜與阻抗匹配層，用以提升共振的品質因子（Q-factors）與繞射效率（diffraction efficiency）。
經過完整優化的最終結構（光柵/波導/緩衝層/基板）展現了在解決「透視-效率」權衡問題方面最先進的性能。它在目標波長下提供了接近完美的反射效率—藍光為97%、綠光為100%及紅光為100%，同時在整個可見光譜範圍（400-700 nm）內保持了高達93%的平均穿透率。此外，該設計具有出色的光譜純度，其半高全寬（FWHM）在藍、綠、紅通道上分別僅為1奈米、2奈米和3奈米，對應高達447的品質因子。此成果實現了極高的色彩純度，色域覆蓋率達到BT.2020標準的95%，色域佔比亦達99%。
最後，詳盡的製造容差與靈敏度分析證明，此設計對於標準奈米微影及薄膜沉積製程中的常見變異具有穩定性，從而證實其為一個務實可行的工程藍圖。本論文為光學合成器領域做出了重大貢獻，針對以往困擾該領域的難題，提出了一個兼具高穿透率、高效率與高色彩保真度的解決方案。;The way AR technology is now going to make human interaction with computers more natural. The optical combiner, however, inherently limits it, making possible the use of only a small number of wearables, eyeglass form factor devices, to implement the technology. The nature of the optical combiner physically constrains it, though, severely limiting the types of wearables, eyeglass form factor devices into which the technology can be adapted. The inherent issue is that there is a trade-off between see-through transparency and reflectivity at projection, and the optical components that are traditionally available have difficulties of being both highly reflective at optical projection tasks and highly transparent at real-world operation. This thesis addresses this challenge by presenting and systematically optimising a new all-dielectric, multi-layer and Guided-Mode Resonance (GMR) grating design, for a high-performance optic combiner.
This work gives the overall process of computational design and optimization, which is done in the Rigorous Coupled-Wave Analysis (RCWA) method. A systematic and multi-stage approach is implemented in the research, with the first step being the development of the fundamental monochromatic GMR filter to investigate the fundamental laws of physics. This model is then further evolved into a more complex, three-colour device to work practically in an oblique-incidence mode (45 degrees). The design also includes crucial functional thin films like a low-index MgF2 buffer layer to reduce substrate leakage loss and a specially tailored MgF2 overlayer that serves as a broadband Anti-reflection Coating (ARC) and also an impedance-matching layer to, respectively, increase resonant Q-factors and diffraction efficiency.
The final structure of full optimization of a Grating/Waveguide/Buffer layer/Substrate has a performance profile that is state of the art in solving transparency-efficiency trade-off. It offers close to unity reflection efficiencies of 97% (Blue), 100% (Green), and 100% (Red) at the target wavelength, with a very high average Transmittance (400-700 nm) of 93% across the visible range. In addition, the design would have great spectral purity with superior Full Width Half Maximum (FWHM), such as 1 nm, 2 nm, and 3 nm on blue, green, and red channels, respectively, corresponding to high Q-factors up to 447. This leads to high color purity, and even the color gamut reaches an impressive 95% of the BT.2020 standard with 99% gamut ratio as well.
Finally, fabrication tolerance and sensitivity study are performed in detail, which shows the design is robust to the typical variations of nanolithography and thin-film deposition processes and therefore a practical engineering blueprint that can be fabricated. The thesis is a significant contribution to the field of optical combiners, as it offers a highly transparent, highly efficient, and color-accurate solution to the exact same problem that afflicted this field of research in the past. |
