Uncovering the origin of interface stress enhancement and compressive-to-tensile stress transition in immiscible nanomultilayers

TL;DR

This study uses atomistic simulations with neural network potentials to uncover how interfacial intermixing and metastable phases influence interface stress and cause a transition from compressive to tensile stress in immiscible nanomultilayers.

Contribution

It introduces a chemically-accurate neural network potential to analyze the origin of interface stress and the stress transition in immiscible nanomultilayers.

Findings

Interfacial intermixing affects stress levels.

Metastable phases influence interface stress.

Explains the compressive-to-tensile transition.

Abstract

The intrinsic stress in nanomultilayers (NMLs) is typically dominated by interface stress, which is particularly high in immiscible Cu/W NMLs. Here, atomistic simulations with a chemically-accurate neural network potential reveal the role of interfacial intermixing and metastable phase formation on the interface stress levels. These results rationalize an experimentally-reported compressive- to-tensile transition as a function of NML deposition conditions and the extremely high interface stresses under some conditions.