Compute Reorganization Energy with QuantumPatch
Similar to Energy Disorder, Reorganization Energies are essential features to almost any ab-initio simulation of charge carrier and exciton dynamics in organic semiconductors. In this tutorial we will show you two different approaches to compute reorganization energies (short: Lambdas due to their denotation with the Greek letter in equations) in QuantumPatch.
In-vacuo Lambda computation
The easiest (and currently recommended) option is to compute reorganization energies on single molecules in vacuum.
Check the "Include in-vacuo Lambda/EA/IP computation" checkbox in the "General" tab. This will display several options in the (otherwise empty) LambdaEAIP tab. When this option is checked, QuantumPatch computes the reorganization energy for one molecule of each species in your morphology in addition to whatever QuantumPatch run you specified otherwise (MatixEAIP to compute EA and IP levels, or Polarized/Uncharged to compute energy disorder and Js). We recommend to combine Lambda computation with the disorder computation. In this case, set up QP according to this tutorial. If you are only interested in Lambda, disable the "Run QuantumPatch" checkbox in the General tab.
In either case, for the lambda computation, proceed as follows:
In addition to the engines needed for your QP run, define an engine for the lambda computation
- Engine: Turbomole
- Functional: b3-lyp
- Basis: def2-SVP Further, disable the disperion correction and leave convergence at "normal". Check "Fallback" and use the following fallback engine:
- Fallback engine: similar to the previous engine but with convergence set to extreme.
Remark: To speed up the lambda computation use multithreading in the engine by setting Threads to whatever your resources allow, max. 16. Remember to allocate sufficient CPUs in the "Resources" tab above and that the maximal number of threads is therefore limited by how many cores per node your architecture has.
This tab does not affect the in-vacuo lambda computation. Set it according to the performed QP run, or disregard in case of "Run QuantumPatch" is disabled in the "General" tab.
- Check the "Calculate Lambda" checkbox
- Lambda Type: chose whether to compute lambda only for holes, electrons or both types of charge carriers (we usually do both).
- Calculate EA/IP: Check to compute EA and IP for single molecules in vacuum. Please keep in mind that in-vacuo EA/IP should be considered mere estimates. Follow the EA IP tutorial to compute reliable EA and IP e.g. to use in LightForge KMC simulations.
- Vertical EA/IP: EA and IP values are computed by subtracting total energy of charged and neutral molecule. The geometry of the neutral molecule is relaxed. For the energy of the charged molecule, you can either use the relaxed geometry of the charged state or the relaxed geometry of the neutral state. For the latter, check this option. Important: Keep this option disabled to compute lambdas, as the reorganization energy is zero, if the geometry of charged and neutral state are the same.
- Set the same engine for both geometry optimization and energy evaluation (labeled "Single points") as defined in the Engines tab.
Lambda computation in environment
A more accurate, but still experimental approach (still in development) is to compute lambda for molecules that are confined by their surrounding molecules in the thin film. In general, these reorganization energies are smaller than the ones computed in vacuum and differ from molecule to molecule. Here we use DFTB3 with a Force-Field Environment to relax monomers within a rigid morphology matrix, i.e. restrictions to elements parametrized in DFTB3 applies. If you would like to follow this approach, you need to set up a Matrix EA IP run following this tutorial, disable the "in-vacuo Lambda/EA/IP Computation" checkbox, but check the "Include in-matrix Lambda Calculation". To get Lambda for both holes and electrons, make sure to define both a positive and a negative molstate in the General-tab. Then follow the steps provided in the Matrix EA IP tutorial.
Remark: This approach works only for molecules that are compatible with DFTB+ (see list of supported elements).
The results of the search are