Get Iterative Hirshfeld Charges

Jacob's Notebook

For a detailed explanation see hirschfeld-i-analysis

A simpler method is outlined below:

• Set up a directory in your NERSC home for the analysis of one compound
• Copy the density file outputted from Abinit (e.g. mz122070_DEN) into this directory
• Also copy into this directory all contents of the "hirshfeld" folder within "Norquist Lab Files" on the Norquist lab computer
• You will notice a bunch of ae files, a separate atom types file, a meta_run.pl, a run_hi.pl file, a interp_charges.pl file, and a hirshfeld.sub file.
• To generate .basenames and .sep files use the following command: ./separate_atomtypes.r.abinit_paw.12apr12 mz12207o_DEN
• To generate all neutral .ae files use perl interp_charges.pl mz12207o_DEN.basenames -init
• You will then need to edit the hirshfeld.sub file so that it uses the correctly named .basenames and .sep files.
• Then just qsub hirshfeld.sub to submit the meta_run with 100 iterations.
• To make sure the charges converged, check the charge_delta_iterations.dat file. At some iteration, all changes in charge should converge to zero for each atom.
• You can then copy the charge_iterations.dat file onto a computer with Excell and origin, and plot the charge as a function of iteration.