This tutorial is for lightforge version 2. If you are using an older version, please contact firstname.lastname@example.org for upgrades.
The LightForge GUI
You can find an introduction to the GUI of LightForge here. In the following, we will skip the basics of the GUI and provide a step-by-step guide for this use case, going through all tabs of the GUI.
Resources and general and IO tabs
| resources tab | general tab | IO tab|
We allocate 31 CPU's as we will later set up the simulation to calculate 30 replicas of the system. On a smaller amount of CPU's the simulation will take longer. On 16 it would mean every CPU has to do 2 simulations, on 30 CPU's that one CPU has to do 2 simulations. In both cases the overall runtime would be similar.
We require all three particles types and we will not attach electrodes. If electrodes were attached new particles could be created and the PLQ measurement would never terminate.
We don't need any input files.
Materials and device tabs
|materials tab||exciton presets||device tab|
For convenience we use exciton presets to determine exciton properties of the material. We define one material with only parametric input for the transport properties. By checking (temporarily) the "show exciton_preset" box, we can see the exciton properties of the used "fluorescent" preset.
Exciton presets are used if microscopic input is missing or some information is not availbe from the microscopic input. Since we will only generate singlets by light illumination and the intersystem crossing rate (ISC) is low the most relevant setting here is max_SSA rate, which determines the pairwise rate [1/s] for singlet singlet annihilation at 1nm distance.
We use a single layer if 50 nm width and thickness. The luminescence simulations yield better quenching statistics in larger systems than smaller. As morphology we choose a simple cubic lattice.
Topology and and physics and operation tabs
|topology tab||physics presets||operation tab|
We choose an amount of 6 neighbours to save memory as Foerster transport will be dominant and Foerster mediated hops are indepenent of the nearest neighbour topology in lightforge. We use parametric transfer integrals for the nearest neighbour hops. The high inverse wf_decay_length for Dexter transfer integrals leads to a suppression of Dexter in this simulation. This is only done for convenience to fastly obtain smooth simulation results.
We only change two settings from the defaults:
- show_advanced: True
- td_scaling: 0.01 We require the first so that the latter will appear. Using the exciton presets the magnitude of the transition dipole of the exciton will be calculated from the lifetime and the singlet energy of the exciton. The transition dipole then is used to calculate Foerster transport rates for the excitons. In order to reduce the rates for Foerster transport (for convenience) we scale all transition dipoles by 0.01.
We do 30 simulations to obtain a smooth luminesence plot. We choose the measurement mode "luminescence" to obtain transients for excitons and photons. We choose a very high rate of exciton generation to get good statistics on the exciton quenching with a moderatly sized system (50nm x 50nm x 50nm). The light irradiation is pulsed, it will be turned on for 10 ns, then turned off for 0.01s . If the simulation would still run after that time, the light would be turned on again for 10 ns. The simulations however terminates, when all excitons have decayed. Take care to set the number of holes and electrons to 0 to avoid SPQ. The computional settings don't influence this measurement mode.
Starting the simulation
Save the workflow, connect to your workflow server and hit run
Once the calculation is finished download the folder lf_output.zip to inspect the results.
The results of the search are