Purpose of ppacf.x:
ACF analysis and print files to track signatures of binding
(PRB 97, 085115 (2018)).
For an illustration of how to use this code to set hybrid mixing
value, please refer to JCP 148, 194115 (2018) doi: 10.1063/1.5012870.
The code reads the output produced by pw.x, extracts and calculates
$E_{c}^{nl}$, $T_{c}^{nl}$, $E_{c,\lambda}^{LDA}$, $E_{c,\lambda}^{nl}$,
$E_{xc,\lambda}$, $T_c^{LDA}$.
If lfock is set to .True., the code also computes the total Fock
exchange value.
With flag code_num = 2, the codes can read output produced by VASP.
With flag lplot, the codes also out puts files containing spatial
variation in most of these quantities.
The input data of this program is read from standard input or from file
and has the following format:
Structure of the input data:============================&PPACF.../
Intermediate results can be saved to disk (see variable lplot in &PPACF)
and later read by pp.x.
Since the file with intermediate results is formatted, it can be safely
transferred to a different machine. This also allows plotting of a
linear combination (for instance, energy density differences) by saving
two intermediate files and combining them (see variables in &PLOT
from pp.x .)
All output quantities are in ATOMIC (RYDBERG) UNITS unless otherwise
explicitly specified.
prefix of files saved by program pw.x
prepended to input/output filenames:
prefix.ecnl, prefix.tcnl, etc.
directory containing the output data from pw.x, i.e. the same as in pw.x
value of the ESPRESSO_TMPDIR environment variable if set;
current directory ('./') otherwise
Number of fragments in coupling-constant scaling curve.
In the default case, only $\lambda=0$ and $\lambda=1$ ends are calculated.
1
If .True. print out the spatial distribution of energy density.
prefix.tclda the LDA component of kinetic-correlation energy density.
prefix.tcnl(prefix.tcgc) the non-local (gradient corrected) component of kinetic-correlation energy density.
prefix.exlda the LDA component of exchange energy density.
prefix.eclda the LDA component of correlation energy density.
prefix.exgc the gradient-corrected component of exchange energy density.
prefix.ecnl(prefix.ecgc) the non-local(gradient-corrected) component of correlation energy density.
prefix.vcnl If vdW-DF: the non-local correlation-potential variation (at nspin=1).
prefix.vcnl1,2 If spin-vdW-DF: spin-reolved non-local correlation-potential variations.
.False.
.True.
If .True. also print out
prefix.tks the Kohn-Sham kinetic energy density.
In case of spin-polarized calculations, prefix.tks1 and prefix.tks2
save the spin-up and spin-down components.
If .True. calculate the Fock exchange based on input Kohn-Sham orbitals.
.False.
If .True. use Lin Lin's ACE (J. Chem. Theory Comput. 12(5), 2242-2249 (2016),
doi: 10.1021/acs.jctc.6b00092).
.True.
Select from which code to read output files.
1 = Quantum ESPRESSO
2 = VASP
The codes will read vasprun.xml and CHGCAR from VASP
calculations.
Please note that in VASP-based analysis:
- Core charge is ignored.
- The ppacf-from-VASP-read-in only works for VASP
calculations done in PBE, revPBE, vdW-DF, vdW-DF2, or vdW-DF-cx
- The ppacf-from-VASP-read-in only always uses the full Ecnl kernel
for coupling-constant scaling analysis of vdW-DF versions.
- Wavefunction based analysis (Fock exchange energy and
Kohn-Sham kinetic energy) are not available from VASP
- When lplot = .True., the code will also print out
charge density in prefix.chg (prefix.chg1 and prefix.chg2
save the spin-up and spin-down components in case of
spin-polarized calculations), which can be processed by pp.x.
1
Select type of vdw kernel table used in ppacf coupling-constant scaling
analysis of nonlocal-correlations in vdW-DF versions:
- vdW_analysis = 0: Full Ecnl kenel of vdW-DF method
- vdW_analysis = 1: The cumulant- or susceptibility-Ecnl kernel component
- vdW_analysis = 2: The pure-vdW-Ecnl kernel component
See IOP JCPM (2020) for presentation of the latter two (non-default) options
o