Computing electron affinity of organic materials
Correct EA levels (sometimes called LUMO) are essential for device simulations. Here we show you how to compute a acceptably accurate EA on single molecules in vacuum.
Setup EA computation
- Convert the molecule.pdb from Parametrizer to cml, e.g. with babel.
- Disable "Run QuantumPatch"
- Enable "Include in-vacuo Lambda/EA/IP Calculation"
- Disregard molecular states
Define the following engines:
- b2plyp: Turbomole, B2PLYP, def2-SVP, set SCF and Geometry Convergence to normal.
- m062x: Turbomole, m062x, def2-SVP, set SCF and Geometry Convergence to normal.
Resources: Parallelization of up to 32 cores works well. Remember to leave one core for the QP master thread, e.g. if you have 32 cores and 64GB RAM available, specify 31 threads and 64000MB Memory in the engines definition and then allocate 32 cores and 64000MB memory in the Resources tab of SimStack.
Disregard the shells tab
- Disable "Calculate Lambda" and "Calculate IP"
- Enable "Calculate EA" and "Vertical EA/IP". With the "Vertical" option, EA (or IP) are computed by subtracting total energy of the charged and neutral molecule on the relaxed ground state geometry. Keeping this unchecked would also optimize the geometry in the charged state.
- As engine for geometry optimization, chose the m062x engine
- Engine for Single Points: b2lyp
Command line submission
A settings template for command line usage is available here. To execute EA computation, rename this file to LamdaEAIP_settings.yml and execute the following command, after sourcing quantumpatch.config as usual:
export NANOMATCH=/path/to/your/nanomatch export NANOVER=V4 source $NANOMATCH/$NANOVER/configs/quantumpatch.config $NANOMATCH/$NANOVER/QuantumPatch/MolecularTools/LambdaEAIP.py
The results of the search are