Quality Improvement of Memory Modules in Depanel Process and CMP Parameters Analysis Using Grey Theory and Taguchi Method

• Main Authors: Chih-yuan Ho, 何智遠
• Other Authors: Zone - ching Lin
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/mhbxev

博士 === 國立臺灣科技大學 === 機械工程系 === 94 === The major objective of this thesis is to apply theories such as Grey Theory, Taguchi Method and Statistical Process Control (SPC) etc. to the quality improvement of the depanel process of the memory modules, as well as the analysis of the CMP (Chemical Mechanical Polishing) parameters. This thesis firstly uses ANOVA (Analysis of Variance) and grey relation analysis to investigate into CMP experiment data of Taguchi Method such as: polishing down pressure, polishing platen speed, carrier speed and arm oscillation etc. The result shows that the order of importance is the same. At the same time, an analytical result is proposed regarding the sensitivity of the data regularization and data of the grey relation theory towards the order of importance. Factors affecting the precision of the products during the memory modules depanel process include: error between the tables and the fixtures, location of the tool, tool wear, machine vibration and cutting temperature etc. By combining the model of grey relation analysis and SPC, this thesis finds that the most important parameter affecting the accuracy of depanel process is the variation of tool, in which the preliminary tool wear can be regarded as the most important factor. This research proposes to use grey prediction compensation method, pre-treated method and shift center deviation method to solve the problems of dimensional variation that often occurs in the memory modules. The result shows that these three methods can all improve the quality of the products. Besides, this thesis uses process capability indices to calculate the result of these three methods before and after improvement, as well as the advantages and the disadvantages of them. In order to research into the tool life and the quality of the memory modules of the tungsten carbide router, this thesis uses two types of commercial available tungsten carbide router, which are TiAICN and Zr respectively, to undergo the experimental comparison towards the milling depanel process. Experimental result shows that the coated tools indeed carry a better effectiveness in their tool life and the quality of the memory module products. This thesis uses Pulsed-DC Reactive Magnetron Sputtering (ZrN) on the tungsten carbide router, and by combining Taguchi Method and grey relation analysis, it analyzes into the optimal parameters for sputtering process. Experiment result shows that, magnetron sputtering is indicative in direction. Therefore, this thesis designs a rotational device for the sputtering tool and finds the optimal sputtering speed. These research results can be provided as valuable references in solving the dimensional variation problems in the depanel process of memory modules.