NEB Oxygen Migration Toolkit

Automated DFT workflow for computing oxygen vacancy migration barriers in perovskite oxides using VASP, pymatgen, ASE, and the climbing-image Nudged Elastic Band (CI-NEB) method on SLURM-based HPC clusters.

The toolkit takes a relaxed bulk structure and, end to end, predicts favorable vacancy sites, enumerates symmetry-distinct O→vacancy hop pathways, generates and interpolates NEB images, submits and monitors multi-stage VASP calculations with automatic restarts, and post-processes the results into energy barriers, transition-state analyses, and publication-quality figures.

Workflow overview

 Bulk structure (POSCAR)
        │
        ▼
 1. Vacancy prediction ──────────  Oxygen_pathways/O_vacancy_predictor.py
    (optional ShakeNBreak            oxygen_workflow_improved.py
     distortion sampling)            oxygen_workflow_shakenbreak.py
        │
        ▼
 2. Pathway enumeration ─────────  oxygen_hopping_vacancy_pathway_generation.py
    initial/final structures         pathway_analyzer.py, report_generator.py
     for each distinct hop
        │
        ▼
 3. Endpoint relaxation ─────────  relax_NEB/multistage_relaxation.py
    multi-stage (coarse→fine),       parallel initial+final relaxations
        │
        ▼
 4. NEB setup + run ─────────────  relax_NEB/neb_workflow_from_migration_analysis.py
    ASE interpolation with           neb_path_generation_with_constraint_control.py
     constraint control,             neb_multistage.py (2/3/5-stage CI-NEB)
     SLURM auto-restart              slurm_manager.py, vasp_inputs.py
        │
        ▼
 5. Analysis + visualization ────  energy_analyzer.py, neb_path_visualizer.py
    barriers, TS detection,          improved_transition_state_detection.py
     TS frequencies,                 ts_frequency_calc.py
     local-environment analysis      oxygen_migration_analyzer_complete.py

Key features

• Multi-stage NEB (2, 3, or 5 stages): rough optimization → climbing image → tight convergence, with automatic CONTCAR→POSCAR handoff between stages and per-stage SLURM resource sizing.
• Multi-stage endpoint relaxation with progressive convergence criteria (EDIFFG −0.1 → −0.01), run in parallel for initial and final structures.
• Constraint control during ASE path interpolation — selective freezing/unfreezing of the host lattice around the migrating oxygen.
• HPC automation: SLURM script generation, job submission, live monitoring, failure detection, and automatic restart from the last checkpoint.
• Vacancy-site prediction with optional ShakeNBreak symmetry-breaking distortions to escape metastable defect configurations.
• Transition-state characterization: TS detection from the converged path, vibrational frequency calculations, and imaginary-mode verification.
• Analysis pipeline: energy profiles from OUTCAR/NEBEF.dat, migration-path geometry tracking, oxygen coordination/local-environment evolution (CrystalNN), and automated plots and CSV/JSON reports.

Repository layout

Path	Contents
relax_NEB/	Core NEB workflow: relaxation, path generation, multi-stage NEB, SLURM + VASP I/O, analysis, visualization
Oxygen_pathways/	Vacancy prediction and O-hop pathway enumeration (pre-NEB stage)
examples/data/	Sample output: a converged LaNiO₃ pathway (energy profile, migrating-O trajectory, bond analysis, per-image POSCARs)
docs/	Component reference and workflow notes

Installation

git clone https://github.com/<your-username>/neb-oxygen-migration.git
cd neb-oxygen-migration
pip install -r requirements.txt

Runtime requirements on the cluster side: VASP (with VTST tools for CI-NEB) and a SLURM scheduler. Set your POTCAR library path and SLURM account/partition in the workflow configuration block of the driver scripts.

Quick start

Generate migration pathways from a relaxed bulk structure:

from Oxygen_pathways.oxygen_workflow_improved import ImprovedEnergyDrivenOxygenWorkflow
# see Oxygen_pathways/start_oxygen_workflow.py for a complete driver script

Run a full NEB calculation from a generated pathway (initial/final structures + moving O index):

# relax_NEB/neb_workflow_from_migration_analysis.py
# - multi-stage relaxation of both endpoints (parallel)
# - ASE image interpolation with constraint control
# - multi-stage CI-NEB with SLURM auto-restart
# - energy profile + barrier extraction

Visualize a finished calculation:

python relax_NEB/neb_visualizing.py   # energy profile, path geometry, TS analysis

Example output (LaNiO₃, O→V_O hop, 7 images): see examples/data/LaNiO3_pathway_001_7images/ — per-image energies, forces, band gap, and magnetization in *_energy_profile.csv, the migrating-oxygen trajectory in *_moving_O_path.csv, and the full image structures in structures/.

Physics background

Oxygen-ion transport governs the performance of solid-oxide fuel cell cathodes, oxygen-permeation membranes, and resistive-switching memristors. The rate-limiting step is the hop of a lattice O²⁻ into a neighboring vacancy; its activation energy is the NEB barrier computed here.

Requirements

Python ≥ 3.9 with pymatgen, ASE, NumPy, SciPy, pandas, matplotlib, seaborn, networkx (see requirements.txt). ShakeNBreak is optional (vacancy distortion sampling).

License

MIT — see LICENSE.