Characteristics Of Plasma-Activated Dielectric Film Surfaces For Direct Wafer Bonding
2020 IEEE 70TH ELECTRONIC COMPONENTS AND TECHNOLOGY CONFERENCE (ECTC 2020)(2020)
摘要
The low-temperature wafer bonding has been studied on two types of dielectric material (SiO, SiCN) as final bonding layers. It is important for the wafer bonding technology to obtain the higher interfacial energy between two bonding wafers, and oxygen and nitrogen (O-2, N-2) plasma treatments have been studied to properly activate the dielectric film surfaces prior to a bonding process that includes chemical-mechanical polishing, hydration with DI water and heat treatment. The surface activation by the plasma treatments with DI hydration formed similar to 10nm thick SiOx layer on the SiCN films. It is found that a newly formed surface SiOx layer played a role as a bonding medium. The dielectric film surfaces were treated by plasma treatment, then characterized by analyzing chemical binding states in the surface SiOx layer. The characteristics of the new SiOx layers were found to be dependent on the plasma species and bulk dielectric films. The obtained properties of the surface layer have been co-related to the initial bonding energy of the bonded wafers as well as the final bonding energy with heat treatment. As a result, the N-2 plasma treatment to the dielectric films enhanced the initial bonding energy and SiCN-SiCN bonding wafers treated by the O-2 plasma have the better initial bonding energy rather than SiO-SiO bonding due to high hydrogen contents in the surface oxide films. In addition, the chemical analysis (XPS) has revealed the surface activity of the films from the results of the chemical binding states of Si. Basically, we focus on the initial bonding energy and it is crucial to ensure that the Cu pads facing each other comes into contact prior to the heat treatment that causes Cu diffusion across the opposite Cu pads.
更多查看译文
关键词
Direct wafer bonding, initial bonding energy, plasma-modified dielectric surface, SiCN-SiCN bonding, SiO-SiO bonding
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络