Graphite, Graphene and Graphene Oxide Builder

MakeGraphitics is a library to ceate various atomistic graphitic structures for molecular dynamics.

Available structures:

• Hexagonal graphene flake
• Rectangular graphene flake
• Rectangular perodic graphene sheet (no edges)
• Periodic graphite
• Graphene and graphite oxide

Output:

• .xyz
• lammps data file

Automatically parameterise by forcefields:

• OPLS
• GraFF
• ReaxFF

Install

Clone this repository. Install using Python2.7. Run the tests to check the installation has worked.

git clone https://github.com/velocirobbie/make-graphitics
cd make-graphitics
python setup.py install
pytest

A conda environment is provided if you do not have the right packages. If you have conda set up, execute these commands to create a working python environment before the install setp.

conda env create --file graphene-env.yml
conda activate graphene

Running pytest will create a bunch of unwanted output files. Sorry about this, I will try and tidy up the outputs soon. In the mean time you can remove with rm *xyz *data.

Examples

See the scripts in the examples/ directory for a number of sample structures.

1. Make a rectangular graphene sheet that extends through periodic boundaries. Parameterised with OPLS and outputs to .xyz for easy veiwing with VMD and a LAMMPS data file.

python2.7 graphene_sheet.py

Size of the sheet can be specified in graphene_sheet.py.

2. Make a hexagonal flake of graphene oxide. Parameterised with OPLS and outputs to .xyz for easy veiwing with VMD and a LAMMPS data file.

python2.7 GO_flake.py

There are several tunable parameters in GO_flake.py that you may be interested in. Including:

• flake radius
• C/O target ratio, ratio
• Rate at which new nodes are added, new_island_freq
• output snapshots of the oxidation process every N steps with video_xyz=N. Viewed in VMD with topo readvarxyz out.xyz

Notes on the Oxidiser

The Oxidiser takes a graphitic structure and attempts to oxidise it by the process described in (Yang, Angewandte Chemie, 2014; Sinclair, 2019). The algorithm proceeds as follows:

1. If hydrogens exist (i.e. edge of a flake), 1/4 are changed to alcohol groups and 1/4 to carboxyl groups (Lerf and Klinowski model). These values can be changed by passing the Oxidiser object the optional arguments: edge_OHratio = 0.25, edge_carboxyl_ratio = 0.5.

2. The reactivity of every possible site is calculated. This is done by using a ranodom forest approach to extend the data set of GO reactivites given by Yang et al. We do not take into account the reactivity of the edges.

3. A site is oxidised at random weighted by each site's reactivity. The chance of an oxidisation producing an alcohol or epoxy group on the surface is by default 50:50, but can be specified by passing Oxidiser the optional argument: surface_OHratio = 0.5

4. The time elapsed between oxidations is estimated from the reactivity of the site that has been oxidised.

5. New nodes are added proportionally to time_elapsed * new_island_freq. Note this can be 0. The reasons for doing this are outlined in (Sinclair 2019).

6. Steps 2-5 are repeated until the target C/O ratio is reached or no new sites are available, usually C/O ~ 1.7 . We recommend setting the target, ratio, to over 2 as this is what is seen experimentally.

Notes on Parameterisation

Not all the bonded interactions that can occur in graphene oxide are included in the OPLS parameterisation. We make some neccesary like for like atom-type substitutions to get around this problem. It is not ideal but common practice in molecular dynamics. The substitutions used are outputed after a parameterisation step. Each substitution line outputs the origional atom types, the atom types used to parameterise them, and a summary string that you can use to find in the script makegraphitics/params.py. Substitutions keep atom types as close to the origional as possible e.g. replaces an aromatic C with an alkene C, whcih are both sp2 carbon atoms.

More structure examples

More examples of building structures with this script are in the examples directory.

Note that differenct structures can be combined into one simulation object with Combine. Also coordinates can be manipulated before writing to a lammps file. An examploe of this is shown in peel_sim.py.

Citing

The work contained here has been published in some of my own papers e.g.

• Graphene–Graphene Interactions: Friction, Superlubricity, and Exfoliation https://doi.org/10.1002/adma.201705791

• Modeling Nanostructure in Graphene Oxide: Inhomogeneity and the Percolation Threshold https://doi.org/10.1021/acs.jcim.9b00114

• The Role of Graphene in Enhancing the Material Properties of Thermosetting Polymers https://doi.org/10.1002/adts.201800168

I would appreciate a citation if you any of the code in any published work :) You could cite the graphene oxide structure paper, this github page (if the journal allows), or the latest release on the zenodo repository

@article{sinclair2019modelling,
  title={Modelling nanostructure in graphene oxide: inhomogeneity and the percolation threshold},
  author={Sinclair, Robert Callum and Coveney, Peter Vivian},
  journal = {Journal of Chemical Information and Modeling},
  volume = {59},
  number = {6},
  pages = {2741-2745},
  year = {2019},
  doi = {10.1021/acs.jcim.9b00114},
}
@misc{make-graphitics-github,
    url = {https://github.com/velocirobbie/make-graphitics},
    howpublished = {\url{https://github.com/velocirobbie/make-graphitics}},
    note = {Accessed: \today},
    author = {Sinclair, Robert C.},
    year = {2019}
}
@misc{make-graphitics_zenodo,
  author    = {Sinclair, Robert. C. },
  title     = {make-graphitics},
  version   = {0.1.0},
  publisher = {Zenodo},
  year      = {2019},
  doi       = {10.5281/zenodo.2548538}
}