Litcius/Paper detail

Sustained Benefits of NCFETs Under Extreme Scaling to the End of the IRDS

Thomas Cam, Ji Kai Wang, Michael Wong, Kyle D. Holland, Prasad S. Gudem, Diego Kienle, Mani Vaidyanathan

2020IEEE Transactions on Electron Devices12 citationsDOI

Abstract

We use full quantum-transport simulations by coupling the Landau-Khalatnikov (LK) and Poisson equations self-consistently with the nonequilibrium Green's function (NEGF) formalism, and calibrated to experimental results, to investigate extremely scaled negative-capacitance, field-effect transistors (NCFETs) having dimensions toward the end of the international roadmap for devices and systems (IRDS), that is, to sub-10-nm gate lengths, where channel transport can be expected to be governed by quantum-mechanical effects. We identify how the ferroelectric affects both thermionic emission and quantum-mechanical tunneling of electrons, both of which are relevant transport mechanisms for these ultrascaled devices. Our detailed results show that while NCFETs are not immune to the increase in the tunneling as they undergo extreme channel-length scaling, the metal-ferroelectric-insulator-semiconductor (MFIS) structure will continue to offer benefits to a subthreshold slope, ON- and OFF-currents, drain-induced barrier lowering, and output conductance until the end of the roadmap. These improvements allow MFIS NCFETs of any given node to achieve similar performance to nonferroelectric devices of the immediately preceding (higher-dimension) node. The fundamental reason for the improvements is identified to be the presence of voltage amplification at the top of the barrier (TOB) and suppression of TOB movement with drain voltage.

Topics & Concepts

Quantum tunnellingPhysicsScalingTransistorSubthreshold slopeThermionic emissionQuantumField-effect transistorQuantization (signal processing)Ballistic conductionCondensed matter physicsElectronVoltageQuantum mechanicsComputer scienceComputer visionGeometryMathematicsFerroelectric and Negative Capacitance DevicesSemiconductor materials and devicesAdvanced Memory and Neural Computing