Novel Self-passivation Rule and Structure of CdTe sigma3 (112) Grain Boundary

TL;DR

This study introduces a new self-passivation rule for modeling CdTe grain boundaries, revealing reconstructed configurations with improved electronic properties and implications for photovoltaic performance, challenging previous unreconstructed models.

Contribution

It proposes a general self-passivation rule for low-energy grain boundary reconstruction and demonstrates its effectiveness in revealing more accurate, self-passivated configurations with better electronic properties.

Findings

Reconstructed GBs have lower formation energies than previous models.

Reconstructed GBs exhibit self-passivated electronic properties without deep-level states.

CdCl2 treatment enhances photovoltaic properties by promoting self-passivation and n-type doping.

Abstract

The theoretical study of grain boundaries (GBs) in polycrystalline semiconductors is currently stalemated by their complicated nature, which is difficult to extract from any direct experimental characterization. Usually, coincidence-site-lattice (CSL) models are constructed simply by aligning two symmetric planes, ignoring various possible reconstructions. Here, we propose a general self-passivation rule to determine the low-energy GB reconstruction, and find new configurations for the CdTe sigma3 (112) GBs. First-principles calculations show that it has lower formation energies than the prototype GBs adopted widely in previous studies. Surprisingly, the reconstructed GBs show self-passivated electronic properties without deep-level states in the band gap. Based on the reconstructed configurations, we revisited the influence of CdCl2 post-treatment on the CdTe GBs, and found that the addition of both Cd and Cl atoms in the GB improves the photovoltaic properties by promoting self-passivation and inducing n-type levels, respectively. The present study provides a new route for further studies of GBs in covalent polycrystalline semiconductors and also highlights that previous studies on the GBs of multinary semiconductors which are based on the unreconstructed prototype GB models, should be revisited.