pulling-gold-junctions

This dataset contains all files and scripts used in the study "Bridging the Gap: Using Machine Learning Force Fields to Simulate Gold Break Junctions at Pulling Speeds Closer to Experiments". Below, the folder structure and its contents are explained.

---
## Folder Structure

### Root Directory

- `10x10\_thinned.dat`: Thinned sample of gold junction with 10x10 electrodes.

### `holdout_set/`

Contains files and scripts related to testing models on holdout data:

- `ace_to_nep_small.txt`: Parameter file for NEP.
* `dft_forces/get_forces.py`: Script to extract DFT-calculated forces.
- `eam_data/au.lammps.eam`: EAM potential file for gold.
- `generate_data.py`: Script for sampling test structures from the reference trajectories.
- `get_eam_forces.py`: Script to calculate forces using EAM potential.
- `get_reaxff_forces.py`: Script to calculate forces using ReaxFF potential.
- `in.simple_*,`: LAMMPS input files for the reference simulations.
- `lammps_*.dat`: LAMMPS output data files.
- `log.lammps_*`: LAMMPS log files.
- `reaxff_data/`: Subdirectory for ReaxFF potential files.
- `control.reax_c.tatb`: ReaxFF control file.
- `ffield.reax.2013_ausch`: ReaxFF potential file for gold.
- `reference_structures/`: Subdirectory for reference structures. Also contains data from the DFT calculations.
- `lammps_*/`: Reference configurations for LAMMPS simulations.
- `1, 2, 3, .../`: Subdirectories for individual DFT calculations.
- `*.npy`: File containing forces from the DFT calculation.
- `*.txt`: GPAW output file.
- `.xyz`: Structure file for the DFT calculation.
- `stripped_8x8.dat`: Dataset with reduced configurations for testing.

#### Notes
A junction with 8x8 electrodes is used instead of 10x10 due to computational constraints. Specifically, too much memory was required to calculate the results for the 10x10 electrode with DFT.
### `production_data/`

Contains the main simulation data categorized by force field:

- `eam/`: Data from EAM-based simulations.
- `1e-2, 1e-3, 1e-4`: Subdirectories for simulations with varying pulling speeds.
- `au.lammps.eam`: EAM potential file for gold.
- `nep/`: Data from NEP-based simulations.
- `1e-2, 1e-3, 1e-4, 1e-5, 1e-6`: Subdirectories for simulations with varying pulling speeds.
- `ace_to_nep_small.txt`: Parameter file for NEP.
- `reaxff/`: Data from ReaxFF-based simulations.
- `1e-2, 1e-3`: Subdirectories for simulations with varying pulling speeds.
- `control.reax_c.tatb`: ReaxFF control file.
- `ffield.reax.2013_ausch`: ReaxFF potential file for gold.

### `test_equilibration/`

Contains data and scripts for testing the length of equilibration in the beginning of the simulation. Results are presented in the SI:

- `times_10/` and `times_100/`: Results for equilibration 10 times longer than the production data.
- `disjoint_submit.sh`: Script for submitting batch simulations.
- `early_stopped_run_10times_equil`: Executable for running the early-stopped runs with 10 times longer equilibration.
- `input_files.tar.gz`: Archived input files.
- `in.simple`: Template LAMMPS input file.
- `lammps_*.tar.gz`: Simulation results archived by run number.
- `logs.tar.gz`: Archived log files.

### `transmission_data/`

Contains simulation data for transmission data categorized by force field:

- `eam/`: Data from EAM-based simulations.
- `1e-2, 1e-3, 1e-4`: Subdirectories for simulations with varying pulling speeds.
- `nep/`: Data from NEP-based simulations.
- `1e-2, 1e-3, 1e-4, 1e-5, 1e-6`: Subdirectories for simulations with varying pulling speeds.
- `reaxff/`: Data from ReaxFF-based simulations.
- `1e-2, 1e-3`: Subdirectories for simulations with varying pulling speeds.