The scaling of electronic devices has continued unabated for the past 5 decades, despite repeated predictions of the "end of scaling". The "more than Moore" slogan has been, and still is, viewed by industry as a need to explore realistic low-to-medium-risk avenues, some already in production -- such as strained Si/Ge or high-κ dielectrics, some yet to be translated to products in the near future or never -- such III-V semiconductors or tunnel-FETs. On the contrary, Academia has looked far ahead, even beyond carbon-based electronics (carbon nanotubes, graphene, nanoribbons). Now that the ULSI technology is looking at a feature size of 5 nm by 2026, even more “improbable” alternatives are considered, including transition metal dichalcogenides (TMDs), silicene/ane, germanene/ane (together with other *enes and *anes), topological insulators, Weyl semimetals, and other strongly correlated fermion systems. Here I will first discuss the fact that, if we really wish to scale to 5 nm, simple electrostatic scaling laws demand that we take two-dimensional materials very seriously, despite their thermodynamic instability (Mermin-Wagner theorem) and the related technological difficulties. I will first consider how electronic transport in non mirror-symmetric 2D crystals (e.g., silicene, germanene, stanene, TMDs of the 1T/T’ crystal structure) may be severely affected by coupling of the electrons with the acoustic flexural modes (ZA-phonons). I will then consider whether graphene has any chance of replacing Si by looking at how the extremely promising electronic properties it exhibits in its ideal form become much less interesting when it becomes a component of some non-ideal structure (such when supported and gated and/or in nanoribbon form). Finally, I will discuss a couple of very interesting admittedly improbable but fun ideas: 1. The Bose-Einstein condensation in bilayer systems (motivating UT-Austin’s BiSFETs) as an example of how issues of practical implementation may regrettably transform an excellent idea into a pure academic exercise; and 2. Monolayer tin (“stannanane”) as a 2D topological insulator with potential applications in spintronics and low-power high-performance devices (… assuming we can make it, of course…). For the entire discussion I will keep density functional theory (DFT), empirical pseudopotentials, and transport equations in the background. This work qas done in collaboration with Shela Aboud, Bo Fu, Jiseok Kim, Sudarshan Narayanan, Ana Suarez Negreira, Zhun-Yong Ong, Cathy Sachs, and William Vandenberghe.
Read full abstract
Oct 8, 2024
M R Edwards + 9
Open Access
