以多準則決策法、技術探勘、主路徑分析、模糊背包問題及模糊能力集合擴展佈局半導體奈米線場效電晶體專利

Abstract

自二十世紀中葉以來，半導體科技持續進步，領導廠商持續發展新製 程與元件，以製造更微小、更有效率的晶片。然而面臨物理極限，半導體 元件的微縮技術，已面臨瓶頸，在不久的將來，摩爾定律即將發展至盡頭。 為了延展莫爾定律，奈米線成為最有潛力之元件。奈米線為一維結構的材 料，其特性為長寬比大、電阻低。應用奈米線技術於場效電晶體中，可以 減少功率損失、提升元件效能、微縮尺寸，為延長摩爾定律的重要元件之 一。因此，半導體廠商紛紛研發技術，並且申請專利，以鞏固其競爭優勢。 顯然，奈米線場效電晶體(Nanowire Field-Effect Transistor, NWFET)技術與 專利為半導體廠商之重要無形資產，然而較少研究著墨於奈米線場效電晶 體的專利佈局。因此，本研究擬探討奈米線場效電晶體之專利，跨越此研究缺口。首先，本研究擬確認專利範圍，並透過探勘美國專利商標局(United States Patent and Trademark Office, USPTO)之專利資料庫，擷取相關專利， 其次，利用主路徑分析法(main path analysis)探索奈米線場效電晶體之專利， 建構引用網路中之主要路徑及專利。接著透過決策實驗室分析法(Decision Making Trial and Evaluation Laboratory, DEMATEL)及基於決策實驗室分析 法之網路流程(DEMATEL based Analytic Network Process, DANP)，找出個 別專利間的影響關係和權重值。最後應用模糊背包問題(Fuzzy Knapsack Problem, FKP)演算法推演 25 項關鍵技術後，藉由能力集合擴展(Fuzzy Competence Set Expansion)法，規劃技術開發路徑圖。藉由以上方法，可了 解奈米線技術專利發展過程中的重要技術，並可作為後進廠商佈局專利， 發展研發策略之依據，本研究之實證成果，可作為後進晶圓代工廠商佈局 未來專利之依據，所發展之研究方法，也可為其他產業所用。Since the middle of the twentieth century, semiconductor technology has continued to advance, and leading manufacturers have continued to develop new processes and devices to fabricate smaller and more efficient chips. However, facing the physical limit, the miniaturization technology of semiconductor technology has already faced bottlenecks. In the near future, the Moore's Law is about to reach its end. In order to extend the Moore's law, nanowires (NWs) have become the most promising device. NWs are materials with a one-dimensional structure, which are characterized by a large aspect ratio and low electrical resistance. The application of NWs technology in field-effect transistors (FETs) can reduce power loss, improve component efficiency, and shrink size. It is one of the important components that extend Moore's Law. Therefore, semiconductor manufacturers have developed technologies and applied for patents to foster their competitive advantages. Obviously, NWFET (Nanowire Field-Effect Transistor) technology and patents are important intangible assets of semiconductor manufacturers, but there is little research on the patent landscape of NWFET. Therefore,this research intends to mine the patents of NWFETs and bridge the gap in this research. First, this research intends to confirm the scope of patents, and extract related patents by mining the patent database of the United States Patent and Trademark Office (USPTO). The patents of the NWFETs were retrieved; then, the main path and patent citation network can be constructed accordingly. Then, through the Decision-Making Trial and Evaluation Laboratory (DEMATEL) and the DEMATEL based Analytic Network Process (DANP), the influence relationships and weight values versus individual patents are derived. Finally, after applying the algorithm of fuzzy knapsack problem (FKP) to derive 25 key technologies, the Fuzzy Competence Set Expansion method is adopted to derive the technology roadmap. Through the above methods, we can derive the most important technologies associated with the important patents of NW technology. These technologies can be used as the basis for defining the landscape of patents and research and development strategies of latecomers in the semiconductor manufacturing industry. The empirical results of this research can be used in landscaping patents of latecomers. The research methods developed can also be used by other industries.