Reliability Allocation And Prediction For Dual-stage System Of Lithography

Lithography is the core equipment of large-scale integrated circuit manufacturingequipment. The high-end lithography technology directly reflects the country’smanufacturing level. The dual-stage system is the critical subsystem of lithography,which has a great impact on the reliability level of lithography directly. The fact showsthat the lack of reliable precision stage has become important factor to restrict thedevelopment level of lithography in our country. So, it is urgent to improve reliability ofprecision stage. This thesis devotes to develop a suite of theoretical methods andpractical tools to allocate and predict the reliability of the dual-stage system at its designstage. The major contributions of this paper cover the following aspects:(1) According to the physical structure of the studied system, the whole dual-stagesystem are divided into several subsystems, and the structure and composition of eachsubsystem are clarified, thus laying foundation for carrying out these works likereliability allocation and prediction.(2) According to analysis of the existing problems in reliability allocation ofdual-stage system and its characteristics, establishing its reliability allocation process.Aiming at the problem that the lack of reliability data of system and the highersubjective uncertainty in reliability allocation process, a comprehensive considerationmultiple weights reliability allocation method is proposed in this paper and use it toallocate the reliability index of dual-stage system. On the basis of the AHP, this methoduses triangle fuzzy numbers to describe the fuzziness of expert judgments and considersthe experts’ weight. This method can get a more reasonable results than traditionalmethods.(3) Based on the structure of dual-stage system, the reliability prediction model ofwafer stage subsystem is simplified. Due to the lack of sufficient reliability data indesign phase, an improved analogous argumentation method of similar products isproposed in this paper to predict the subsystems reliability. On the basis of analogousargumentation method of similar products, the weight of expert is considered in the newmethod and a diversity factor of expertise is proposed to amend the results of reliability prediction. As demonstrated in the case studies, the proposed method can get moreaccurate results than traditional methods.