研究目的
To develop a strategy for optical microscopy of high-aspect-ratio (HAR) nanoelectromechanical systems (NEMS) that combine large feature spacing and large height with sub-wavelength width, and to demonstrate algorithms for separation of NEMS, MEMS and background in microscope images based on valley detection, thresholding and masking.
研究成果
The study demonstrates that widely spaced, high-aspect ratio silicon NEMS are visible under white-light illumination in a bright-field optical microscope as dark lines on a light background. This characteristic allows for the separation of nanoscale features and microscale edges from background using valley detection and non-linear processing. Further work is needed to improve the specificity of the separation algorithm, especially for closely spaced features and intersections of nanoscale features.
研究不足
The method's effectiveness is limited by the need for widely spaced features and high numerical aperture lenses for optimal resolution. Additionally, the presence of surface debris and non-uniform illumination can affect image quality and feature extraction accuracy.
1：Experimental Design and Method Selection:
A 2D model of incoherent imaging based on modal diffraction theory was used to simulate line images of HAR NEMS. The model accounts for the diffraction, reflection, and interference effects of light interacting with sub-wavelength features.
2：Sample Selection and Data Sources:
Silicon NEMS containing 100 nm wide features were imaged in a bright-field microscope to confirm the simulation results.
3：List of Experimental Equipment and Materials:
A Leica DMR white light microscope with ×5, ×20, and ×50 objectives (NA = 0.15, 0.40, and 0.55) and a Donpisha 3CCD camera module were used for imaging.
4：15, 40, and 55) and a Donpisha 3CCD camera module were used for imaging.
Experimental Procedures and Operational Workflow:
4. Experimental Procedures and Operational Workflow: Images were acquired and processed in MATLAB to separate NEMS, MEMS, and background based on valley detection, thresholding, and masking.
5：Data Analysis Methods:
The spatial variation of the major eigenvalue was analyzed to detect valleys in brightness, and the Otsu method was used for thresholding to separate feature classes.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
