CMP Diamond Conditioner Dressing Characteristics of Polyurethane Pad

• Main Authors: Ming-Yi Tsai, 蔡明義
• Other Authors: Yunn-Shiuan Liao
• Format: Others
• Language: zh-TW
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/37242272770187493119

博士 === 國立臺灣大學 === 機械工程學研究所 === 95 === The wafer performance is often determined by the texture of the pad surface which is controlled by a diamond pad conditioner in the CMP process. In the thesis, the fundamental characteristics of dressing action on the polyurethane pad are investigated first via dressing by single diamond of different orientations, under various dressing parameters and dressing path. Experimental results show that a groove with pile-up on both side walls forms as the diamond moves over the pad. Plowing is found to be the major mechanism responsible for this surface topology if dressing is conducted by the face of a diamond. On the contrary, cutting action dominates when the point of a diamond is responsible for dressing. The depth of the groove is smaller than the dressing depth due to the spring back of the pad. There is less variation of ridge with the dressing depth and dressing path for point dressing. These results imply that the use of point dressing is recommended. The second part of the thesis study the effects of ridges on the polishing rates of silicon dioxide are investigated. Experimental results reveal that there exists an optimal dressing depth. Below this value, the polishing rate decreases with the reduction of the dressing depth. Above this value, the polishing rate saturates and then decreases gradually. The use of point dressing is beneficial since it results in a higher but less variation polishing rate. As a result, the point dressing is recommended to the design of diamond pad conditioner. Finally dressed by a novel advanced diamond disk (ADD) containing the same height and identically shaped tip diamond are studied. They are compared those dressed by with conventional diamond disk (DG). Experimental results demonstrate that the ADD can dress asperities of the pad more uniformly than the DG. The dressing rate of the ADD is reduced by about 30% and the pad is less worn during the polishing process, and hence the pad life is extended. As a result, reduction of the cost CMP is expected. The polishing rate of the ADD is higher than that of the DG. There is less variation of polishing rate with the polishing time for ADD as compared with that of DG.