Three-Dimensional (3-D) Integrated Systems
Vasileios F. Pavlidis, PhD, EPFL-IC-LSI
Keywords: Vertical integration, 3-D ICs, Interconnect design, Circuit design
Technology scaling and CMOS technologies have steadily supported an increase in the performance of integrated circuits (ICs) over the past several decades. These driving forces are expected, however, to lose momentum as the fabrication of nanoscale devices at gigascale densities will not be economically viable. Three-dimensional (3-D) integration is a novel design paradigm with great potential to fundamentally advance the computational power and functionality of modern integrated systems. The introduction of three-dimensional integration will result in highly complex integrated circuits since an unprecedented number (and possibly types) of devices and functions will be integrated as a single system.
Schematic of a heterogeneous 3-D integrated system.
Considering this nonpareil complexity, the usual objectives of contemporary circuit design are significantly challenged. The same situation also applies for the CAD tools that will need to handle these complex systems. Therefore, several research efforts have focused on developing these design methodologies and design aids that can sufficiently support the emerging 3-D integrated systems. We emphasize a variety of problems related to the design, modeling, and the development of design tools for 3-D ICs. The physical and interconnect design of these systems is of particular interest, since 3-D vertical integration is often considered to be an interconnect centric design paradigm. We also put considerable efforts to develop techniques and related design aids for envisioned heterogeneous 3-D systems that can potentially include mixed-signal circuits and other components, such as sensors and biomedical circuits. The overall objective is to provide methodologies and design tools that can facilitate the development of high-performance and highly heterogeneous 3-D systems.
Our research related to 3-D integration is funded by the Swiss National Foundation and Intel Corporation (Intel Germany Labs). The main foci of these projects are on the power and signal integrity of 3-D integrated systems. The harmonious integration of vastly different circuits and technologies, which can be included in a heterogeneous 3-D system, requires the careful investigation of signal behavior as signals within a 3-D IC can change in form (i.e., analog, digital, RF) and traverse multiple planes. The correct design of the power network supplying current and ground to all these planes can also affect the performance and operation of the system components. Since power and signal integrity are intertwined, both of these issues are currently investigated. The following doctoral projects are pursued within the goals and objectives of the funded programs: