[Q-e-developers] Huge esa when H2O background is added

Sun Mar 8 12:41:17 CET 2015

Dear Xu Huang,

I am one of the main developers of the continuum solvent model you are referring to and I guess I am also the one that provided it to your previous colleague, probably together with some detailed info on how to run it. I would be more than happy to help you with the set up of your calculations, you can contact me at my institution email address, as your previous colleague did, oliviero.andreussi 'at' epfl.ch

In general, the input you use seems reasonable, the reason of poor convergence may be numerical or physical. In the first case, my guess is that the tolrhopol keyword that you use is not tight enough, try using 1.d-12 or even smaller.

The meaning of the keywords and their use to help convergence is explained briefly in the documentation, which in the pre-release of the code is in a Doc/ directory inside Environ/, I guess you also have it, right? In recent versions of the code you can compile the documentation by typing 'make doc' inside the Environ directory.
Regards,

Oliviero Andreussi

________________________________
From: q-e-developers-bounces at qe-forge.org [q-e-developers-bounces at qe-forge.org] on behalf of Huang, Xu [xu-huang at uiowa.edu]
Sent: Friday, March 06, 2015 6:07 PM
To: q-e-developers at qe-forge.org
Subject: [Q-e-developers] Huge esa when H2O background is added

Dear QE-developers,

I'm using QE-5.1 to do LiCoO2 surface modeling in aqueous condition. The official implementation of continuum solvent model hasn't been released yet. I got it from a previous colleague who received the implementation from the model developer himself. I added H2O as background in relaxation pw.x calculation. However, the structure couldn't get convergence and esa (estimated scf accuracy) are huge as following:

......

estimated scf accuracy < 170.60730387 Ry

estimated scf accuracy < 1.99665060 Ry

estimated scf accuracy < 1.22910197 Ry

estimated scf accuracy < 6.96663706 Ry

estimated scf accuracy < 5.03277607 Ry

estimated scf accuracy < 1.15153871 Ry

estimated scf accuracy < 4.07745354 Ry

estimated scf accuracy < 0.78684128 Ry

estimated scf accuracy < 1067.72222994 Ry

estimated scf accuracy < 1026.29146455 Ry

estimated scf accuracy < 1037.18484566 Ry

estimated scf accuracy < 1056.50019233 Ry

estimated scf accuracy < 1058.57442299 Ry

estimated scf accuracy < 1059.96747709 Ry

estimated scf accuracy < 1125.30480321 Ry

estimated scf accuracy < 1175.02131330 Ry

estimated scf accuracy < 1015.64507162 Ry

estimated scf accuracy < 4362.77094323 Ry

estimated scf accuracy < 2422.16259874 Ry

estimated scf accuracy < 296.33753823 Ry

estimated scf accuracy < 80.46912372 Ry

estimated scf accuracy < 100.42046977 Ry

estimated scf accuracy < 108.34637365 Ry

estimated scf accuracy < 62.34104403 Ry

......

There is no such problem when I just apply vacuum as background. So I think the structure may not be the source of problem. There might be something wrong with the parameters controlling H2O background. Here I added the script that I'm using for relaxation pw.x calculation for you to look at. I'm not sure what solvent parameter gone wrong.

Thank you very much for your help!

Regards,

Xu Huang

(And this is my script:

#############################################

#!/bin/tcsh

#PBS -l nodes=1:ppn=16

#PBS -N licoo2

#PBS -q standard_16

#PBS -l walltime=48:00:00

#PBS -j oe

cd /import/c/w/uaxhuang1/LiCoO2_Co_0.5_Li_aq_51

source /etc/profile.d/modules.csh

module purge

module load pgi/13.4 openmpi-pgi-13.4/1.4.3

setenv PATH /usr/local/pkg/espresso/espresso-5.1/bin:$PATH

setenv ESPRESSO_PSEUDO /import/u/u1/uaf/uaxhuang1/pwf

setenv ESPRESSO_TMPDIR $PWD/temp

setenv ESPRESSO_ROOT /usr/local/pkg/espresso/espresso-5.1

setenv temp ./temp

### ELECTROSTATIC EMBEDDING PARAMETERS ###

set verbose="0"

set environ_thr="1.0"

set environ_type="input"

set eps_mode="full"

### PERIODIC BOUNDARY CONDITIONS ###

set assume_isolated="slabz"

set epsilon="80"

rm -r $temp/*

echo " running the relax calculation in aq"

echo " with $assume_isolated periodic boundary correction"

cat > LiCoO2.relax.in << EOF

&CONTROL

calculation = 'relax',

prefix = 'LiCoO2',

pseudo_dir = '/import/u/u1/uaf/uaxhuang1/pwf/',

outdir = './tmp/',

restart_mode = 'from_scratch',

verbosity = 'high',

/

&SYSTEM

ibrav = 4,

celldm(1) = 10.778758346,

celldm(3) = 6.000000000,

nat = 48,

ntyp = 3,

tot_charge = 0,

ecutwfc = 40,

ecutrho = 320,

occupations = 'smearing',

smearing = 'mv',

degauss = 0.02,

nspin = 1,

assume_isolated = '$assume_isolated',

/

&ELECTRONS

electron_maxstep = 100,

diagonalization = 'cg',

conv_thr = 1.d-6,

mixing_beta = 0.12,

/

&IONS

/

ATOMIC_SPECIES

Li 1.00 Li.pbe-s-rrkjus_psl.0.2.1.UPF

O 1.00 O.pbe-rrkjus.UPF

Co 1.00 Co.pbe-nd-rrkjus.UPF

ATOMIC_POSITIONS {crystal}

Li 0.500042167 -0.000042425 0.297620529

Li 0.000042425 0.499957833 0.297620529

O 0.334395935 0.165599716 0.335312931

O 0.832181938 0.167818078 0.334812285

O 0.332180448 0.667819568 0.334812569

O 0.834400300 0.665604081 0.335312931

Co 0.167464285 0.332535738 0.364086635

Co 0.665907921 0.334092095 0.364185341

Co 0.165908082 0.834091934 0.364185249

Co 0.667464278 0.832535731 0.364086635

O -0.000188488 0.000188488 0.393686106

O 0.500193879 -0.000198269 0.393757442

O 0.000198269 0.499806121 0.393757442

O 0.499810179 0.500189821 0.393686404

Li 0.333343150 0.166656724 0.432225225

Li 0.833324158 0.166675858 0.432221252

Li 0.333324107 0.666675909 0.432221246

Li 0.833343292 0.666656866 0.432225225

O 0.166669492 0.333330748 0.470795212

O 0.666664094 0.333335922 0.470795937

O 0.166664167 0.833335849 0.470795886

O 0.666669268 0.833330524 0.470795212

Co -0.000003501 0.000003501 0.499997899

Co 0.500002855 -0.000003354 0.499997642

Co 0.000003354 0.499997145 0.499997642

Co 0.499996334 0.500003666 0.499997957

O 0.333336976 0.166663244 0.529212723

O 0.833329769 0.166670247 0.529213198

O 0.333329846 0.666670170 0.529213182

O 0.833336772 0.666663040 0.529212723

Li 0.166675991 0.333324110 0.567777107

Li 0.666656854 0.333343162 0.567773146

Li 0.166656883 0.833343133 0.567773153

Li 0.666675906 0.833324025 0.567777107

O -0.000193620 0.000193620 0.606243508

O 0.500186548 -0.000187666 0.606314390

O 0.000187666 0.499813452 0.606314390

O 0.499805972 0.500194028 0.606243587

Co 0.334091926 0.165908118 0.635813965

Co 0.832535917 0.167464099 0.635914203

Co 0.332535903 0.667464113 0.635914199

Co 0.834091898 0.665908090 0.635813965

O 0.167820108 0.332181049 0.665187526

O 0.665601004 0.334399012 0.664686011

O 0.165601367 0.834398649 0.664685917

O 0.667818967 0.832179908 0.665187526

Li -0.000042383 0.000042383 0.702379471

Li 0.499957596 0.500042404 0.702379467

K_POINTS {automatic}

4 4 1 1 1 1

EOF

cat > environ.in << EOF

&ENVIRON

verbose = $verbose

environ_thr = $environ_thr

environ_type = '$environ_type',

eps_mode = '$eps_mode',

tolrhopol = 1.D-10

mixrhopol = 0.5

env_static_permittivity = $epsilon

env_surface_tension = 0.D0

env_pressure = 0.D0

/

EOF

mpirun -np 16 -x ESPRESSO_TMPDIR -x ESPRESSO_ROOT pw.x -environ environ.in < LiCoO2.relax.in > LiCoO2.relax.out

)

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/developers/attachments/20150308/c7ab03dc/attachment.html>