Pdn-CO-Stability

Dataset, machine learning models and Monte Carlo simulations in Python for evaluating stability of subnanometer CO-adsorbed Pdn clusters supported on Ceria

Computational Framework

Notes:

• Machine learning models are efficient structure-to-energy mappings
• Low energy indicates higher stability
• Monte Carlo simualtions are used to optimize the strutcure

Dataset

The dataset contains Pdn cluster structures in the size range from 1 to 21, descriptors and CO-CO interactions.

• Cluster structures from DFT calculations
• Descriptors for single CO adsorption
• CO-CO interactions for multiple CO adsorption

Pdn cluster energy model

The machine learning model to predict Pdn energy from a given structure

• Pdn-CE

CO adlayer energy model

The machine learning model to predict CO adlayer energy from a given structure

• CO-CE

Structure optimization algoirthms in Grand Cannoical ensembles

Grand Cannoical Monte Carlo (GCMC)

Automatic discovery of optimal (lowest free energy) adsorbate layer structures at a given temperature and CO pressure

Dependencies

• Python version 3.6+
• Numpy: Used for vector and matrix operations
• Matplotlib: Used for plotting
• Scipy: Used for linear algebra calculations
• Pandas: Used to import data from Excel files
• Sklearn: Used for training machine learning models
• Seaborn: Used for plotting
• Networkx: Used for graph opertations
• ase: Used for atomic structure representation
• sympy: Used for geometry calculations
• mpi4py: Used for paralleling GCMC simulations

Publication