Iuliana P. Radu, imec
How New and old Materials Research and Know-How Extend the Increase in Computation Power
For the past 20 years, doomsayers have been talking about the end of performance increase in compute because transistor scaling was supposed to end. We all know this was far from true and continue physical and performance scaling of the transistors has enabled continues increase in computing power. Maintaining physical or performance scaling of the switches themselves is becoming more difficult but other options exist and the field of very exciting research. Old concepts with new materials and new concepts with old materials are in the spotlight and materials research is paramount to enabling new growth in computing power and of the society in general.
In this talk, we will cover an old concept but with new materials: transistors with 2D semiconducting channels as these materials are expected to enable further scaling of transistor area and performance. At the other side of the spectrum, we will cover how better-understood materials such as Al, Nb, TiN contribute to better qubit for quantum computing. Uniting the two topics are not only the application of computing, but also the mindset and fundamental need for extreme control of materials and processes involved in fabricating these devices.
Transistors with 2D materials hold the promise for extreme gate length scaling compared to Si as they can be deposited with monolayer precision, and mobility is expected to be almost independent of material thickness. In this presentation, we will discuss imec’s work on 2D materials emphasizing channel and dielectrics deposition and how process steps could change device performance. We will cover how device and circuit modelling guides choices in integration schemes and ultimately defines specifications for the materials involved. We will describe how variability analysis can help differentiate between one-off observations and consistent process control and understanding.
Quantum computing is fundamentally different that classical computing but many of the technology challenges are similar to those for building classical systems. In this talk, we will outline how learning and practices from standard semiconductor development can be used to enable quantum computing. We will discuss how materials and process development can improve device performance.
In both topics, we will outline what we believe main materials research directions to be investigated and where fundamental discovery and innovation is key.