DFT + GWBSE Optimization Using CO¶

Introduction¶

This tutorial explains how to perform a molecular geometric optimization using the GWBSE method. See the GW Compendium: A Practical Guide to Theoretical Photoemission Spectroscopy, for an excellent introduction to the method.

Requirements¶

You will need to install VOTCA using the instructions described here
Once the installation is completed you need to activate the VOTCA enviroment by running the VOTCARC.bash script that has been installed at the bin subfolder for the path that you have provided for the installation step above

Preparing the input¶

To run a DFT-GWBSE calculation we will use the xtp_tools calculator.

we can now change the calculator mode from single point energy calculation to optimization as follows by adding -c geometry_optimization.maxiter=1 which enables optimization with one iteration.

Because geometry optimsation is quite expensive numerically we chose a smaller basisset and G0W0 Finally, to run the calculation we just need the following command (This still takes a while)

[1]:

!xtp_tools -c job_name=CO geometry_optimization.maxiter=1 dftpackage.basisset=3-21G dftpackage.auxbasisset=aux-def2-svp gwbse.gw.mode=G0W0 -e dftgwbse -t 4

==================================================
========   VOTCA (http://www.votca.org)   ========
==================================================

please submit bugs to https://github.com/votca/votca/issues

xtp_tools, version 2023-rc.2 gitid: 891004f (compiled Jun  3 2023, 22:08:53)

Initializing tool
... dftgwbse Evaluating tool
... dftgwbse  Using 4 threads
... ... Reading structure from CO.xyz
... ... Requested geometry optimization of excited state s1
... ... Initial state: s1
... ... WARNING: No tracker is used
... ... Convergence of total energy: 0.000001 Hartree
... ... Convergence of RMS Force:    0.000030 Hartree/Bohr
... ... Convergence of Max Force:    0.000100 Hartree/Bohr
... ... Convergence of RMS Step:     0.000600 Bohr
... ... Convergence of Max Step:     0.001000 Bohr
... ... Initial trust radius:        0.018897 Bohr
... ...
... ... =========== OPTIMIZATION SUMMARY =================================
... ... At iteration  0
... ...  ---- POSITIONS (Angstrom)
... ...  Atom    x        y       z
... ...    0    +0.0000  +0.0000  +0.0000
... ...    1    +1.2000  +0.0000  +0.0000
... ...    Total energy:     -112.26831841 Hartree
... ...    Trust radius:       0.01889726 Bohr
... ...  ---- FORCES (Hartree/Bohr)
... ...       central differences
... ...       displacement 0.0010 Angstrom
... ...  Atom    x        y       z
... ...    0    -0.1314  -0.0000  -0.0000
... ...    1    +0.1314  +0.0000  +0.0000
... ...
... ... =========== OPTIMIZATION SUMMARY =================================
... ... At iteration  1
... ...  ---- POSITIONS (Angstrom)
... ...  Atom    x        y       z
... ...    0    -0.0071  -0.0000  -0.0000
... ...    1    +1.2071  +0.0000  +0.0000
... ...    Total energy:     -112.27152481 Hartree
... ...    Trust radius:       0.03779452 Bohr
... ...  ---- FORCES (Hartree/Bohr)
... ...       central differences
... ...       displacement 0.0010 Angstrom
... ...  Atom    x        y       z
... ...    0    -0.1081  +0.0000  +0.0000
... ...    1    +0.1081  -0.0000  -0.0000
... ...    energy change:     -0.00320640 Hartree      Not converged (1e-06)
... ...    RMS force:          0.02547456 Hartree/Bohr Not converged (3e-05)
... ...    Max force:          0.10807939 Hartree/Bohr Not converged (0.0001)
... ...    RMS step:           0.00314954 Bohr         Not converged (0.0006)
... ...    Max step:           0.01336238 Bohr         Not converged (0.001)
... ...  ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
... ...
... ... Saving data to CO.orb
... ... Writing output to CO_summary.xml

Notice that we are using 4 threads. The results will be stored in a file named CO_summary.xml in the current work directory, together with the optimization step in optimisation.trj and the orbitals in hdf5 format saved on CO.orb.