2D Semiconductors Are Advancing, But Silicon Is Improving Too

Researchers in the semiconductor industry have long expected the need for better materials to replace silicon in transistors. However, silicon devices keep improving, delaying that change. 

Silicon offers a unique mix of performance, ease of manufacturing, and cost-effectiveness. However, in recent years, the idea of moving beyond silicon has become more likely. Transistors require thinner channels for better control, but as the thickness drops below three nanometers, surface scattering significantly increases channel resistance.

Two-dimensional semiconductors seem to be the best alternative. They have no dangling bonds sticking out, which helps reduce surface scattering.

CFET architectures combine PMOS and NMOS transistors in a single vertical structure. This design could allow for silicon scaling to continue for another two decades.

An alternative channel material needs to do more than just match silicon’s performance. It must also match silicon’s manufacturability and cost. Two-dimensional materials still have a long way to go to be as affordable as silicon.

Fabs require consistent quality across a 300mm wafer. For a long time, the best 2D semiconductor devices were made from exfoliated flakes of bulk material.

PMOS and NMOS devices use different channel materials. It might be possible to stack CFET-like structures with the two devices vertically, but placing PMOS and NMOS devices side-by-side is very challenging.

Given the challenges of using 2D semiconductor CMOS channels, it's good that the future of silicon looks promising. This gives the industry time to work on 2D transistor development for less demanding applications.