Python library for ultra-fast structural optimisation and property calculations of organic co-polymers. This code was developed while doing my post-doc in the Zwijnenburg group, https://www.zwijnenburg-group.org/.
polyhts takes base functionality from:
rdkit: transforms SMILES strings into monomer building blocksstk: constructs linear co-polymer structures from these building blocksxtb: optimises geometries, calculates properties
Combining the above, polyhts can be used for high-speed, accurate screening of
organic co-polymer compositions, both for property screening and quick, exploratory
calculations. Currently supported properties are:
- Ionisation potentials (IP)
- Electron affinities (EA)
- Excitation energies & oscillator strengths
- Solvation free energies
Though, in principle, any property that can be calculated via xtb or stda
can be obtained.
polyhts calculations start by defining a Session, in which information like
co-polymer repeat unit length, number of repeat units that will be used to
construct a polymer chain, number of conformers to be explored and solvent type
are specified.
For example, we can start a Session in which we will construct polymers with 4
repeat units (each of which are 2 monomers long) and explore 100 conformers while
applying an implicit solvent model for benzene:
session = polyhts.Session('my_session', 2, 4, 100, solvent='benzene')This way, we can explore co-polymer compositions with arbitrary repeat unit complexity and polymer chain length.
Within this session, we can screen all combinations of pre-supplied monomer unit SMILES from a text file:
session.screen('smiles-list.txt', nprocs=20)where smiles-list.txt has the format:
0001 smiles1
0002 smiles2
0003 smiles3
. .
. .
. .
Note that, not only will all compositions of monomers be screened, but all permutations of a given compositions as well (e.g. AABB as well as ABAB).
This created a directory called 'my_session', containing one sub-directory for all the co-polymers screened. Here you can find the starting and optimised structures. Lastly, within 'my_session', there will be an output file containing all results.
We can also calculate properties for a single co-polymer, where we supply a list of smiles explicitly, where the length of this list is equivalent to the repeat unit length:
session.calc_polymer_properties(['c1c(Br)cc(Br)cc1', 'c1c(Br)cc(Br)cc1'], 'polymer-name')Following the stk documentation, Br atoms are places within SMILES strings
where monomer units are to be connected to one another.
clone the repo and then:
$ cd polyhts
$ pip install -r requirements.txt
$ pip install -e .
You will also need to install RDKit, which can be installed via conda:
$ conda install -c rdkit rdkit
- [1] J. Chem. Inf. Model. 2015, 121, 2562-2574
- [2] J. Chem. Theory Comput. 2017, 13, 1989-2009
- [3] Comput. Theor. Chem. 2014, 1040, 45-53
- [4] J. Chem. Phys. 2016, 145, 054103