SU8 photoresists are used widely in the development and fabrication of the high-aspect-ratio microstructure. However, SU8 process is difficult to repeat when the film thickness and aspect ratio are increasing. In this paper, the soft bake is investigated. The retained solvent level in SU8 film is studied by using the theory of free volume and diffusion process. The result shows that the proper and uniform solvent level in the photoresist film is crucial to achieve good SU8 microstructures in the ultra-thick film process. The microstructure with 5 lm linewidth and aspect ratios larger than 50 was obtained. IntroductionHigh aspect ratio microstructures are needed in many advance applications. A variety of thick photoresists were applied and optimized for microfabricating of high aspect ratio MEMS (Micro Electro-Mechanical System). The SU8 has been proven the suitable material for microfabrication combining the processing of ultra-thick photoresist layers and an outstanding UV-based lithographic performance. However, the quality of microstructures is sensitive to the variation of process parameters in the thick film process [1-3]. Typical aspect ratio would be 20, and 25 for good process. Although the microstructure with aspect ratio larger than 50 has been obtained, the repeatability may be a problem.Processing of SU8 is a challenging task, especially in the ultra-thick photoresist film. Every process step influences strongly the results obtained and has to be adapted individually to the desired application. Therefore, the parameter values described in the literature vary in a wide range according to various structural features and it is difficult to decide which parameters are most crucial to achieve good SU8 microstructures. Although the previous research [4] showed that the soft bake plays the key role in the SU8 process, the details about this process step haven't given. In this paper, the soft bake is investigated. The retained solvent level in SU8 film is studied by using the theory of free volume and diffusion process. The result shows that the proper and uniform solvent level is crucial to achieve good SU-8 microstructures in the thick film process. ExperimentalThe SU8-2075 photoresist was chosen in this paper because it allows a wide coverage of thickness. The SU8 was statically dispensed onto a 2 inch cleaned silicon wafer and spread at 500 rpm for 5 s. The different rotation speed was applied and held for 30 s to attain the target film thickness (100, 200, 300, 400 micron). The weight of silicon wafer and SU8 film were measured by using DTG160 balance (Shanghai balance factory). The wafer was heated on the hotplate at 95°C for predetermined time. The weight was measured after the wafer has been cooled to room temperature. Then, the SU8 film was exposed in contact mode using a narrowband mercury source, followed by a PEB (post exposure bake) at 95°C. The mask with different linewidth structures was applied to fabricate the microstructures with different aspect ratio. Finally, the microstructures were deve...

The deep lithography of thick resist layer is the primary step of LIGA technology. UV Proximity lithography, which is used to fabricate high aspect ratio MEMS normally, is investigated in this paper. The light intensity distribution in the thick photoresist layer mainly depends on the diffraction produced by the gap between mask and wafer in proximity lithography. Fresnel diffraction model is used to simulate lithography process normally, which the thin photoresist layer is used in the lithography process. But it is not accurate in deep lithography process. A correction to the Fresnel diffraction theory is used to simulate the lithography process depending on the scalar diffraction theory of light propagation in this paper. The difference of this two models is given in this paper. The simulation results show that the corrected model can obtain more accurate simulation results than the Fresnel diffraction model. The experiment results and the theory analysis both indicate that: the structure contrast decreases with the increase of the film thickness. The smaller the structure linewidth is, the faster the structure contrast decreases. The linewidth of microstructure is not equivalent from the top to the bottom, broaden in the middle part normally. The theory simulation gives the quantitative analysis.

Soft X-ray microscopy has excellent characteristics for imaging cells and subcellular structures. In this paper, the yeast strain, Candida utilis, was imaged by soft X-ray microscopy and three-dimensional volumes were reconstructed with the SART-TV method. We performed segmentation on the reconstruction in three dimensions and identified several types of subcellular architecture within the specimen cells based on their linear absorption coefficient (LAC) values. Organelles can be identified by the correlation between the soft X-ray LAC values and the subcellular architectures. Quantitative analyses of the volume ratio of organelles to whole cell in different phases were also carried out according to the three-dimensional datasets. With such excellent features, soft X-ray imaging has a great influence in the field of biological cellular and subcellular research.