Fundamental Characterization of Low Dimensional Carbon Nanomaterials for 3D Electronics Packaging

Abstract: Transistor miniaturization has over the last half century paved the way for higher value electronics every year along an exponential pace known as 'Moore's law'. Now, as the industry is reaching transistor features that no longer makes economic sense, this way of developing integrated circuits (ICs) is coming to its definitive end. As a solution to this problem, the industry is moving toward higher hanging fruits that can enable larger sets of functionalities and ensuring a sustained performance increase to continue delivering more cost-effective ICs every product cycle. These design strategies beyond Moore's law put emphasis on 3D stacking and heterogeneous integration, which if implemented correctly, will deliver a continued development of ICs for a foreseeable future. However, this way of building semiconductor systems does bring new issues to the table as this generation of devices will place additional demands on materials to be successful. The international roadmap of devices and systems (IRDS) highlights the need for improved materials to remove bottlenecks in contemporary as well as future systems in terms of thermal dissipation and interconnect performance. For this very purpose, low dimensional carbon nanomaterials such as graphene and carbon nanotubes (CNTs) are suggested as potential candidates due to their superior thermal, electrical and mechanical properties. Therefore, a successful implementation of these materials will ensure a continued performance to cost development of IC devices. This thesis presents a research study on some fundamental materials growth and reliability aspects of low dimensional carbon based thermal interface materials (TIMs) and interconnects for electronics packaging applications. Novel TIMs and interconnects based on CNT arrays and graphene are fabricated and investigated for their thermal resistance contributions as well electrical performance. The materials are studied and optimized with the support of chemical and structural characterization. Furthermore, a reliability study was performed which found delamination issues in CNT array TIMs due to high strains from thermal expansion mismatches. This study concludes that CNT length is an important factor when designing CNT based systems and the results show that by further interface engineering, reliability can be substantially improved with maintained thermal dissipation and electrical performance. Additionally, a heat treatment study was made that enables improvement of the bulk crystallinity of the materials which will enable even better performance in future applications.