在這感測器需求量大增的世代中， CMOS感測器目前一直追求像素的小型化，而像素的小型化將帶來效率低下以及衍生出來的低靈敏度問題。而傳統的CMOS感測器中的色彩濾波器方案受限於先天結構的影響，將使問題變得更加嚴重，而近年來受到廣泛關注的超穎表面色彩路由器可以將特定波長的光引導到對應的像素中進而使效率提高。這樣一來，正好可以克服效率與靈敏度的問題。但在微小像素中塞入許多奈米結構並不易於製作。 本篇論文中我們提出一個基於全介電質奈米結構設計出拜耳式排列的色彩路由器，我們在1.6 µm*1.6 µm的單位晶胞下僅放入4個對應像素的奈米圓柱，並且運用基因演算法對結構參數優化。在我們所設計的兩個方案中，在模擬上RGB效率分別可以達到57.16%、46.88%、49.94%與58.47%、52.98%、32.77%，並且能量大部份分佈在各顏色像素的中心，來避免光線落入低感光區域的問題。在製程方面，最終結構尺寸是有機會使用Krf掃描式曝光機進行低成本量產製造。 ;In this era of increased demand for sensors, CMOS sensors have been continuously striving for pixel miniaturization. However, the downsizing of pixels brings inefficiency and the associated low sensitivity issues. Traditional color filter solutions are limited by their inherent structures, exacerbating these problems. In recent years, metasurface color routers have gained significant attention as they can efficiently route specific wavelength light to corresponding pixels, thereby improving efficiency. This approach effectively overcomes the challenges of efficiency and sensitivity. However, incorporating numerous nanoscale structures into tiny pixels poses fabrication challenges. In this paper, we propose a metasurface color router based on an all-dielectric nanostructure arranged in a periodic pattern. We only place four nanocylinders corresponding to pixels in a 1.6 µm * 1.6 µm unit cell. We employ a genetic algorithm to optimize the structural parameters. In our two designed schemes, the simulated RGB efficiencies achieve 57.16%, 46.88%, 49.94% and 58.47%, 52.98%, 32.77%, respectively. Furthermore, the majority of the energy is concentrated at the centers of each color pixel, mitigating the issue of light falling into low photosensitive areas. In terms of the manufacturing process, there is a possibility to utilize a Krf scanner exposure machine to achieve low-cost mass production of the final structure.
