從研發合作與專利引用探討半導體企業知識擴散強弱連結管道之差異

Abstract

在知識經濟時代的複雜環境及快速變遷下，不斷地創新成為企業生存的重要法則，如何獲取企業外部的知識就成為企業經營的重要課題，而外部知識的獲取則可透企業間彼此的知識流動、累積與擴散。本研究針對半導體企業的知識擴散進行研究，以六吋（1976∼1991）、八吋（1989∼1999）及十二吋（1997∼2011）晶圓期間來區分技術發展的分水嶺，透過研發合作及專利引用作為企業間知識擴散強連結與弱連結管道的衡量對象，採用專利計量分析與社會網絡分析方法，首先探討整體知識擴散網絡結構特徵與擴散管道的主要參與企業；再區分企業不同技術定位（技術領導型、仲介型、追隨型、獨立型企業）及產業鏈角色（整合元件製造廠、IC設計廠、晶圓代工廠、封裝測試廠、IC設備廠、其他企業），針對各類型企業利用強、弱連結管道進行知識擴散的差異進行探討，分析企業參與程度的差異、擴散對象的差異、以及企業技術定位的變動對於知識擴散的影響進行探討；最後對於半導體晶圓製造主要廠商的知識擴散進行個案分析。 研究結果顯示：半導體企業間知識擴散廣泛利用強連結管道，但個別企業群內則偏重利用弱連結管道；技術領導型企業與整合元件製造廠為半導體知識擴散高度參與者；技術定位中技術領導型企業為半導體知識主要擴散對象；產業鏈角色中整合元件製造廠及晶圓代工廠為半導體知識主要擴散對象；半導體企業技術定位變動與其知識擴散參與程度差異不大；半導體企業知識擴散強、弱連結管道交互牽動不明顯；主要晶圓製造廠商半導體知識擴散表現接近。 基於上述研究結果，建議監控整合元件製造廠及晶圓代工廠等半導體技術領導者研發動態，擬定半導體技術研發策略、透過強連結網絡掌握半導體共通技術開發者，透過弱連結網絡掌握半導體特定技術開發者、透過共同知識擴散對象衡量半導體企業與競爭者的技術發展相似性、晶圓代工廠可與整合元件製造廠從垂直分工轉為水平合作的研發夥伴、臺灣半導體晶圓代工廠應持續發展差異化技術，以確保研發聯盟各自的技術優勢。

In the knowledge economy and the complex environment of rapid change, the constant innovation has become an important survival rule for companies. How to catch the external knowledge has become an important issue, and external knowledge can be acquired from the knowledge flow. In this study, the analyzed periods are distinguished by wafer size in six-inches (1976~1991), eight-inches (1989~1999) and twelve-inches (1997~2011). The methodologies of patentometrics and social network analysis (SNA) are adopted, and research and development (R&D) cooperation and patent citations are treated as the agents for strong ties and weak ties of knowledge spillover. Firstly, the author analyzed the structure characterisitics of knowledge spillover networks and the main companies that utilize the strong and weak ties to spill knowledge. Secondly, The author classified companies into four technological positions: leader, broker, follower, and isolated company. On the other hand, companies were also classified into six roles of the industry value chain: integrated device manufacturer (IDM), IC design, foundry, packaging & testing, equipment, and others. The author analyzed the difference of utilizing the strong and weak tie of knowledge spillover, the difference of targets with which companies spill knowledge, and the the impact of the change of the companies with the technological position on knowledge spillover. Finally, A case analysis of knowledge spillover is addressed for the top IC foundries. The results of this study show that: 1. Knowledge spillover among semiconductor companies occurs broadly through strong ties, while the weak ones of knowledge spillover are mostly used within the company cluster; 2. Technological leaders and IDM are the highly active participants for knowledge spillover; 3. Technological leaders, IDMs, and foundries are the main targets with which companies spill knowledge; 4. Whether semiconductor companies with the technological position change or not, the patent count and the degree of their participation remain similar; 5. The interaction between strong ties and weak ties of knowledge spillover is insignificant during the chip era evolution; 6. There are similar knowledge spillover actions among main IC foundries. Based on the findings, it is recommended that the mainstream of semiconductor technology development could be captured by monitoring the companies that are technological leaders with the role of IDM or foundry. Semiconductor companies could trace the developers of common technology by observing the strong tie network of knowledge spillover, and they could trace the developers of particular technology by observing the weak tie network of knowledge spillover. Semiconductor companies could also measure the similarity between themselves and competitors based on common targets of knowledge spillover. Foundries could develop partnership with IDMs from vertical division of work into horizontal cooperation. Foundries in Taiwan should continue to develop differentiated technologies in order to ensure technical superiority of their R&D alliances.