# Shear Deformation Pathways of Cu-Ni Multilayers via Neural Network Potential
A trained neural net potential (NNP) was used to simulate shear in the [011] direction in Cu-Ni multilayers. We applied the NNP to a Monte Carlo scheme to generate parallel shear simulations, demonstrating the range of possible trajectories that can be obtained due to speedups provided by the NNP. Data collected in FY22.

## Naming Convention
* pristine: Cu-Ni multilayer with 250 Ni and 250 Cu atoms
* 1-Ni-1: Cu-Ni multilayer with 249 Ni and 250 Cu atoms; Ni vacancy located at the Ni-Cu interface
* 1-Ni-3: Cu-Ni multilayer with 249 Ni and 250 Cu atoms; Ni vacancy located two layers away from the Ni-Cu interface
* 1-Cu-1: Cu-Ni multilayer with 250 Ni and 249 Cu atoms; Cu vacancy located at the Ni-Cu interface
* 1-Cu-3: Cu-Ni multilayer with 250 Ni and 249 Cu atoms; Cu vacancy located two layers away from the Ni-Cu interface
* 2-Ni-1: Cu-Ni multilayer with 248 Ni and 250 Cu atoms; Ni divacancy located at the Ni-Cu interface
* 2-Cu-1: Cu-Ni multilayer with 250 Ni and 248 Cu atoms; Cu divacancy located at the Ni-Cu interface

## Contacts
* DFT: Qin Pang (qin.pang@pnnl.gov)
* NNP: Jenna Pope (jenna.pope@pnnl.gov)
* NNP: Henry Sprueill (henry.sprueill@pnnl.gov)
* PI: Peter Sushko (peter.sushko@pnnl.gov)

## Repositories 
* Original SchNetPack codebase: https://github.com/atomistic-machine-learning/schnetpack
* Active learning SchNet codebase: https://github.com/pnnl/Active-Sampling-for-Atomistic-Potentials

## Acknowledgements
This research was supported by the Laboratory Directed Research and Development Program under the Solid Phase Processing Science Initiative at Pacific Northwest National Laboratory, which is operated by Battelle for the U.S. Department of Energy under contract DE-AC06-76RLO.