Modelling of nanoscale multi-gate transistors affected by atomistic interface roughness.

Area of Science:

• Semiconductor device physics
• Quantum transport phenomena
• Advanced transistor technology

Background:

• Interface roughness scattering (IRS) is a key performance limiter in advanced non-planar transistors like FinFETs.
• Understanding IRS is crucial for optimizing carrier transport in nanoscale devices.

Purpose of the Study:

• To investigate and compare two physical models (Ando's and multi-sub-band) for electron scattering due to interface roughness.
• To predict the impact of IRS on FinFET performance considering realistic channel shapes and IRS parameters (correlation length Λ and RMS height).

Main Methods:

• Utilized an in-house developed 3D finite element ensemble Monte Carlo simulation toolbox.
• Incorporated parameter-free 2D Schrödinger equation quantum correction for non-equilibrium carrier transport.
• Simulated n-type SOI FinFETs with varying channel orientations and Fin widths.

Main Results:

• The multi-sub-band IRS model demonstrates a much stronger effect on electron transport than the Ando model.
• FinFETs are significantly affected by IRS in the ON-region, with effects intensifying as Fin width decreases.
• FinFETs with <110> channel orientation are more susceptible to IRS than those with <100> orientation.
• IRS correlation length (Λ) and RMS height (Δ) impact device performance similarly.

Conclusions:

• IRS is a critical factor limiting FinFET performance, particularly in scaled devices and during high-current operation.
• The choice of IRS model significantly influences simulation accuracy.
• Device design parameters like channel orientation and Fin width must be carefully considered to mitigate IRS effects.