Imeall is a database framework for the calculation of the atomistic properties of grain boundaries. The grain boundaries are generated and labeled according to a uniform process indicated by the following notation:

  Theta[XXX](YYY)

sigma = (number of coincident sites in an elementary cell)/ (total number of all lattice sites in an elementary cell)

Theta is the angle of misorientation about the vector [XXX] and (YYY) determines the plane of the interface between the misoriented Grain A and Grain B. In this way it is possible to uniquely characterize all grain boundaries.

An additional set of three degrees of freedom is then used to extend the above definition. These degrees of freedom correspond to a rigid body translation in the two dimensional plain of the grain boundary one relative to the other, and a volume relaxation orthogonal to the grainboundary plane. These atomic scale degrees of freedom compose subgrains of the canonical parent grain. Finally, we assign the ideallized crystalline bulk structure the notation:

  0000000000

Computationally each grain boundary so identified can be considered an object with all its physical properties as attributes (to be described below). The structures are stored locally in the database as Extended XYZ files.

Dependencies

• Flask
• ASE
• QUIP
• OVITO

Documentation

Full documentation can be found here.

Using Imeall

0. To get the development branch of Imeall:

  git clone https://github.com/Montmorency/imeall.git

Set the environment variable (preferably in your .bashrc or .bash_profile) for your ovito executable and the directory containing any .xml files required for QUIP potential calculators.

  export POTDIR=~/imeall/imeall/potentials
  export OVITO=/Applications/ovito

1. Initialize the imeall app:

  python runserver.py

This will launch the web app hosted locally and can be accessed through your browser at http://127.0.0.1:5000/.

The home page contains a list of links for each material in the database. Each link leads to a view of different orientation axes vectors for which grain boundary properties have been computed.

For instance /alphaFe/110 contains all computed 110 grain boundaries for iron. The 000 orientation axes implies a perfect crystal, different cleavage planes and singular defects and vacancies in a perfect crystal.

Within each orientation axes is a list of full grain boundaries. Each of these contains a snap shot of the unrelaxed grain boundary structure (the canonical boundary), the coincident site lattice, and a list of sub-directories containing the type of calculations performed on the grain boundary. The subdirectories (calc_types) store atomic scale properties according to calculation type. Examples of calculation types include: tight binding models, EAMs, DFT. If multiple "flavours" of each calculation type are to be used the directories can take more specific names: EAM_Mishin, DFT_PW, DFT_VASP etc. Within each calc_type of directories is the pattern gbid_'suffix'. The suffix denotes the modification to the canonical grain boundary structure. This could constitute a pattern like _d2.0 which denotes: delete one of a pair of atoms from the original grain boundary structure with nearest neighbour distance less than 2 A. Another suffix pattern might be displace the canonical grain along the in-plane lattice vectors by a distance 0.25*lattice vector: _tv0.25bxv0.25. Suffices can be combined _tv0.25bxv0.25_d2.0. A key for some common suffix patterns is given here:

  _v2bxv6        : supercell 2 times along v 6 times along bxv.
  _tv0.13bxv0.25 : displace one grain along the 0.13*v along v and 0.13*bxv along bxv.
  _d2.0z         : delete one of pair of atoms within distance 2.0 A of each other.
  _h2.0n1        : 1 additional hydrogen atom in system 2.0 A from a grainboundary.

Systematic use of these naming conventions ensures that at the analysis stage grain boundary properties can be retrieved, compared, and combined easily.

The canonical grain contains a file:

  gb.json

subsequent directories each contain a file:

  subgb.json

This json file contains information about relevant grain boundary properties in the usual key:value format. Some common keys are:

  E_gb_init : Unrelaxed Grain Boundary Energy.
  E_gb      : Relaxed Grain Boundary Energy.
  n_at      : Number of atoms in the grain boundary.
  A         : Grain Boundary area in the unit cell.

Upon subsequent analysis new key:values pairs can be added to the json dictionaries as required.