Surface Passivation in Empirical Tight Binding

TL;DR

This paper introduces a versatile implicit passivation method for empirical tight binding simulations, applicable to various passivation scenarios, and validates it against ab-initio results and experiments, highlighting its advantages over traditional methods.

Contribution

The work presents a novel implicit passivation approach for TB methods that is broadly applicable and improves accuracy in simulating passivated surfaces.

Findings

The new method accurately matches ab-initio and experimental results.

Certain oxidation configurations significantly degrade transistor performance.

Implicit H passivation overestimates transistor performance.

Abstract

Empirical Tight Binding (TB) methods are widely used in atomistic device simulations. Existing TB methods to passivate dangling bonds fall into two categories: 1) Method that explicitly includes passivation atoms is limited to passivation with atoms and small molecules only. 2) Method that implicitly incorporates passivation does not distinguish passivation atom types. This work introduces an implicit passivation method that is applicable to any passivation scenario with appropriate parameters. This method is applied to a Si quantum well and a Si ultra-thin body transistor oxidized with SiO2 in several oxidation configurations. Comparison with ab-initio results and experiments verifies the presented method. Oxidation configurations that severely hamper the transistor performance are identified. It is also shown that the commonly used implicit H atom passivation overestimates the transistor performance.