A study on optimization process for wafer cleaning through supercritical carbon dioxide fluid method

• Main Authors: Chi-jean Chain, 錢啟禎
• Other Authors: Jir-ming Char
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/66582846389884487193

碩士 === 輔英科技大學 === 環境工程與科學系碩士班 === 99 === Abstract In recent years, with the development of semiconductor integrated circuit industry from mm to nm process, the traditional cleaning methods became inadequate in cleaning the wafer. A wafer manufacturing process using traditional wet cleaning technology requires a large amount of acid-base solvents and chemicals. The average daily cleaning process of a wafer fabrication produces 15,000~22,000 tons of waste water and a large amount of waste solvents, which result in serious environmental problems. The application of Supercritical Carbon-dioxide (SCCO2) Cleaning Technology is a new attempt to solve the wafer cleaning problem. With the characteristics of low surface tension, low viscosity, high diffusivity and high density, the SCCO2 can be spread into the nanopore of wafer and clean away contaminants. As a result, the supercritical carbon-dioxide cleaning technology produces relatively fewer wastewater problems and requires lower operating cost. These advantages give the supercritical carbon-dioxide cleaning technology a great potential of developing a “green clean process” in wafer cleaning. The purpose of this study was to evaluate the ability of SCCO2 in removing the photo resister from the wafer surface. Otherwise, soak and spray cleaning methods are assessed based on Box-Behnken’s design method and compared to obtain a more efficiency method in photo resister removing. The experimental results are analyses through applying the Design-Expert software to construct second-order regression model and search for the optimal combination of variables. After he experimental data sets were substituted into the governing equations to obtain the optimized solution, an evaluation of feasibility are executed to verify the result.