Line Reconstruction DXA

An experimental dislocation extraction that reconstructs dislocation lines from the facets of a Delaunay tessellation whose lattice vectors don't close.

Line Reconstruction DXA is an experimental alternative to OpenDXA. It works on the same input, the crystal clusters of a structure identification stage, but reconstructs dislocation lines directly from the tessellation instead of sweeping Burgers circuits along a defect surface. Use it to cross-check OpenDXA or when you only need the lines.

Key	Version	Builds
`line-reconstruction-dxa`	2.2.0	Linux, macOS, Windows, browser

How it works

The plugin builds a Delaunay tessellation of the atoms and keeps the tetrahedra that pass an alpha-shape test, which removes the large, empty tetrahedra outside surfaces and inside voids.
It gives each tessellation edge an ideal lattice vector by following a path of crystal bonds between its two atoms, at most Crystal Path Steps long, through the crystal clusters. Edges it can't resolve are completed from their neighbors.
A triangular facet whose three edge vectors don't sum to zero is threaded by a dislocation, and the sum is its Burgers vector. The plugin joins these facets into segments and stitches the segments into lines, open or closed.
In FCC it also sorts the lines into perfect, Shockley, Frank, stair-rod and Hirth dislocations (in the log), and the lines are smoothed and coarsened.

Inputs

An earlier stage of Polyhedral Template Matching or Adaptive Common Neighbor Analysis in the same pipeline, with the same crystal. Its shared clusters_table, clusters_transitions and neighbor_lattice fill in this plugin's hidden arguments.

Arguments

Argument	Type	Default	Allowed values	Description
Crystal Structure `crystalStructure`	select	`FCC`	`FCC`, `BCC`, `HCP`, `CUBIC_DIAMOND` (Cubic Diamond), `HEX_DIAMOND` (Hexagonal Diamond)	The reference crystal. Match the structure identification stage.
Crystal Path Steps `crystalPathSteps`	number	`4`	≥ 1, step 1	Longest path of crystal bonds used to give an edge its lattice vector.
Alpha Scale `alphaScale`	number	`3.5`	≥ 0, step 0.1	Alpha-shape threshold for keeping a tetrahedron, as a multiple of the neighbor distance.
Smoothing Iterations `smoothingIterations`	number	`0`	≥ 0, step 1	Taubin smoothing passes applied to the lines.
Line Point Interval `linePointInterval`	number	`1.2`	≥ 0, step 0.1	Spacing of the points kept along each line.

These arguments are filled in from the earlier stage and don't appear in the form:

Argument	Type	Shared by
Clusters Table `clusters-table`	string	Polyhedral Template Matching, Adaptive Common Neighbor Analysis (as `clusters_table`)
Clusters Transitions `clusters-transitions`	string	Same (as `clusters_transitions`)
Neighbor Lattice `neighbor_lattice`	string	Same

The workflow also lists Pattern Structure Matching as a source, but version 2.1.0 of that plugin shares only neighbor_lattice.

Outputs

Exposure	What you get
Dislocation Lines	A listing, and 3D lines colored by Burgers family.

Each line carries the properties the workflow declares: burgers_family, burgers_family_label, crystal_structure, length (Line length), magnitude (Burgers magnitude), cluster_id and is_closed (Closed loop).

Notes

This plugin is experimental: it has no defect mesh, charts or network statistics. Use OpenDXA for those.

Reference

The plugin's own method has no separate publication. For the dislocation extraction algorithm it builds on, see A. Stukowski, K. Albe, "Extracting dislocations and non-dislocation crystal defects from atomistic simulation data", Modelling Simul. Mater. Sci. Eng. 18, 085001 (2010).