Interferometric light microscopy for wafer defect inspection and three-dimensional object reconstruction

Interferometric optical microscopy;Defect inspection;Metrology;Quantitative phase imaging;Optical diffraction tomography;White-light diffraction tomography;Tomographic phase microscopy

Zhou, Renjie

关键词: Interferometric optical microscopy;    Defect inspection;    Metrology;    Quantitative phase imaging;    Optical diffraction tomography;    White-light diffraction tomography;    Tomographic phase microscopy;   

Others  :  https://www.ideals.illinois.edu/bitstream/handle/2142/72977/Renjie_Zhou.pdf?sequence=1&isAllowed=y

美国|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

The first topic of the dissertation is semiconductor wafer defect inspection. We developed a highly sensitive defect inspection system based on laser interferometric microcopy, called epi-illumination diffraction phase microscopy (epi-DPM), to measure both the phase and amplitude of the scattered field from the wafer. To detect deep sub-wavelength defects, we further reduced the noise in the images by developing an image post-processing method, called 2DISC (2nd-order image difference, image stitching, and convolution). With the 2DISC method, we examined a 22 nm node intentional defect array (IDA) wafer. The results showed that we can reliably detect defects down to areas as small as 20 nm by 100 nm. Moving forward, we adapted our epi-DPM system to inspect a densely patterned 9 nm node IDA wafer, a significantly more difficult task. To improve our system’s sensitivity, we replaced the old 532 nm solid-state laser with a new 405 nm diode laser which has 10x smaller noise. Using the 2DISC method, we were able to detect defects down to 15 nm by 90 nm.In order to further increase our system’s sensitivity, we proposed several approaches. The first approach utilizes precision z-scans to produce a three-dimensional (3D) wafer image, and looks at different cross-sectional planes in this 3D image. The second approach utilizes a spatial light modulator (SLM) to make a dark-field filter to selectively filter the image, such that we maintain the defect signal and suppress the wafer’s underlying structure. Therefore, after increasing the power of the incident light, the CCD (charge-coupled device) camera’s full dynamic range can be used to measure the signal coming from the defect. For a preliminary demonstration, we used an inverse filter in a bright-field imaging system to high-pass the imaging beam at the Fourier plane to remove the low frequency laser speckle noise. Our experimental results demonstrated detection sensitivity improvements with this inverse filter dark-field imaging. In order to find the best filter for the dark-field inspection system, we also worked on simulating dark-field and bright-field imaging for the 22 nm and 9 nm node IDA wafers. The third approach is using a white-light interferometric imaging system, which is expected to have better image contrast due to its speckle free nature. Recently, we finished building a white-light epi-illumination DPM (epi-wDPM) system. This system has the modalities of performing 3D scanning and dark-field filtering. A real-time inspection software and a table-top clean-room have also been built for this system. We have been working on implementing this system for 9 nm node wafer inspection, with the goal to break the previous defect detection limit.The second topic of the dissertation is solving the inverse scattering problem for 3D object reconstruction in quantitative phase imaging (QPI). Three-dimensional optical reconstruction typically uses a laser light source, suffering from speckle noise, and the measurements are usually done in the far zone by scanning the laser angles. Building on the ideas of laser diffraction tomography, we developed a new tomographic reconstruction method, called white-light diffraction tomography (WDT). WDT is speckle free and reconstructs the depth dimension by simply scanning through the focus of an object. The depth resolution of WDT is limited by the coherence property of the white-light source. With this approach we successfully reconstructed cells in 3D with resolution beyond the diffraction limit. The wavevector space method used in WDT can be applied to solve the inverse scattering problem in systems under the Fresnel approximation, such as optical coherence tomography (OCT) and angle-resolved low-coherence interferometry (aLCI). We have also proposed using our wavevector space method to calculate optical trapping force, lens focusing, and light diffraction by an aperture, to better understand optical resolution, and to solve the light diffusion and time-reversal problems in thick tissue.