Materials Science Fragmentation

AutoFragment provides specialized support for materials science systems, including periodic solids, surfaces, and extended structures like MOFs and zeolites.

Periodic System Handling

Fundamental to materials science is the treatment of periodic boundary conditions (PBC). AutoFragment handles this through several core components:

Lattice Vectors

The Lattice class represents the crystallographic unit cell. It supports:

  • Conversion between Fractional and Cartesian coordinates.

  • Reciprocal Lattice calculations.

  • Cell volume and parameter extraction.

Minimum Image Convention

All distance-based calculations (bond detection, radial partitioning) in periodic systems automatically use the Minimum Image Convention, ensuring that the shortest distance through periodic boundaries is considered.

Supercells

AutoFragment can generate supercells (make_supercell) that maintain full bonding continuity. This is essential for studying larger domains or performing fragmentation on expanded systems.

Materials Partitioners

Specialized partitioners are available for materials, inheriting from MatSciPartitioner.

  • RadialPartitioner: Shell-based fragmentation around a specific center (e.g., an defect site).

  • SlabPartitioner: Layer-based fragmentation along a specified lattice axis. High useful for surface science.

  • UnitCellPartitioner: Divides a supercell into smaller units based on the underlying grid.

  • SurfacePartitioner: Automatically distinguishes between Top Surface, Bottom Surface, and Bulk regions based on vacuum distribution.

Material-Specific Rules

AutoFragment includes pre-defined rules for common material classes:

Zeolites

  • Siloxane bridges (Si-O-Si) are identified as valid fragmentation points.

  • Brønsted acid sites (Al-O(H)-Si) are specifically protected to maintain chemical integrity of the active site.

Metal-Organic Frameworks (MOFs)

  • Metal Clusters (Nodes) are preserved.

  • Organic Linkers are kept intact.

  • Metal-Carboxylate bonds can be marked as breakable to separate nodes from linkers.

Polymers

  • Backbone Detection: Identifies C-C chains and distinguishes between internal and terminal monomers.

  • Side-chain Preservation: Keeps functional side-groups attached to the backbone.

Perovskites

  • Octahedral preservation: Maintains the integrity of $MO_6$ octahedra while allowing for corner-sharing separation.

Usage Example

from autofragment.partitioners.geometric import SurfacePartitioner
from autofragment.io import read_cif

# Load a slab system
system = read_cif("slab_system.cif")

# Partition by surface vs bulk
partitioner = SurfacePartitioner(surface_axis=2, surface_depth=4.5)
tree = partitioner.partition(system)

# tree.fragments now contains "surface_top", "surface_bottom", and "bulk"