[QE-users] Getting convergence on a cell obtained from a cif using vc-relax
Giuseppe Mattioli
giuseppe.mattioli at ism.cnr.it
Fri Jun 1 16:42:33 CEST 2018
Dear Reinaldo
I'm afraid there is no "strategy to systematically get convergence"
for all kinds of system. Your system is a metal-organic molecular
crystal, mainly kept together by strong dispersion interactions. Such
interactions are totally neglected by the PBE functional you are
using. If you want an accurate estimate of lattice parameters you can:
a) use one of the "native" vdwdf functionals (there is plenty, but you
should test them). Use some kind of empirical or partially empirical
dispersion correction such as DFT-D2 or the recently implemented
DFT-D3. Moreover, I have obtained very nice structural results in the
case of organic or metal-organic crystals by using the XDM correction,
but it works only together with PAW pseudopotentials.
HTH
Giuseppe
Quoting Reinaldo Pis Diez <reinaldo.pisdiez at gmail.com>:
> Dear folks,
>
> I usually help some experimentalist fellows from the Center I belong
> by calculating geometries, cell parameters, vibrational frequencies,
> electronic transitions, etc, of transition metal complexes. As the
> cif file is often available, I use it to create the input file for
> PWScf using Virtual Nano Lab from Quantum Wise.
>
> Although convergence of the vc-relax job is usually achieved, some
> cases are more problematic. Job ends with a high pressure and with
> atomic forces far from thresholds when 'bfgs' is used both for &ions
> and &cell namelists. Using 'damp' does not help to get geometry and
> cell parameters optimized. Thus my (very general) question is: does
> anybody know an strategy to systematically get convergence when
> doing vc-relax? A link to a tutorial or a document is welcome.
>
> For completeness, the input of one of the problematic cases is
> appended below. By the way, I'm using "Program PWSCF v.6.1 (gpu tag
> v1.0)" compiled with PGI compilers and MKL libraries on a CentOS box
> with a Titan X (Pascal) gpu.
>
> Thanks in advance
>
> Reinaldo
> Center of Inorganic Chemistry
> Natl Univ of La Plata, Argentina
>
> Input follows....
>
> &CONTROL
> title = 'Cu-ovatpnh2'
> calculation = 'vc-relax'
> restart_mode = 'from_scratch'
> outdir = './'
> pseudo_dir = './'
> prefix = 'Cu-ovatpnh2'
> etot_conv_thr = 0.0001
> forc_conv_thr = 0.0004
> tstress = .true.
> tprnfor = .true.
> nstep = 200
> /
>
> &SYSTEM
> ibrav = 14
> celldm(1) = 9.9320792573d0
> celldm(2) = 2.0161515879d0
> celldm(3) = 2.1156030541d0
> celldm(4) = -0.1705259449d0
> celldm(5) = -0.0219091103d0
> celldm(6) = -0.2238523262d0
> nat = 59
> ntyp = 6
> ecutwfc = 40
> ecutrho = 200
> input_dft = 'pbe'
> occupations = 'smearing'
> smearing = 'mv'
> degauss = 0.005
> nspin = 2
> starting_magnetization(1) = 0.7
> /
>
> &ELECTRONS
> conv_thr = 1d-06
> mixing_beta = 0.6
> mixing_mode = 'plain'
> mixing_ndim = 8
> diagonalization = 'david'
> /
>
> &IONS
> ion_dynamics = 'bfgs'
> /
>
> &CELL
> cell_dynamics = 'bfgs'
> press = 0.5
> /
>
> ATOMIC_SPECIES
> Cu 63.546000d0 cu_pbe_v1.2.uspp.F.UPF
> C 12.010700d0 c_pbe_v1.2.uspp.F.UPF
> H 1.007940d0 h_pbe_v1.4.uspp.F.UPF
> N 14.006700d0 n_pbe_v1.2.uspp.F.UPF
> O 15.999400d0 o_pbe_v1.2.uspp.F.UPF
> S 32.065000d0 s_pbe_v1.4.uspp.F.UPF
>
> ATOMIC_POSITIONS (crystal)
> C 0.082231085 0.204570350 0.616545553
> C 0.917768915 0.795429650 0.383454447
> C 0.172973917 0.203448044 0.733466608
> C 0.827026083 0.796551956 0.266533392
> H 0.367673209 0.258122966 0.770942028
> H 0.632326791 0.741877034 0.229057972
> C 0.987334077 0.125087827 0.797766960
> C 0.012665923 0.874912173 0.202233040
> H 1.019591574 0.112838755 0.892010655
> H -0.019591574 0.887161245 0.107989345
> C 0.756454381 0.067986708 0.729401520
> C 0.243545619 0.932013292 0.270598480
> H 0.580696185 0.004985342 0.755622580
> H 0.419303815 0.995014658 0.244377420
> C 0.224564542 0.266351629 0.518206870
> C 0.775435458 0.733648371 0.481793130
> H 0.388829444 0.346391746 0.560989945
> H 0.611170556 0.653608254 0.439010055
> H 0.304022379 0.192372172 0.457292153
> H 0.695977621 0.807627828 0.542707847
> C 0.956901991 0.245893003 0.340167231
> C 0.043098009 0.754106997 0.659832769
> H 0.022416710 0.153595990 0.313033710
> H 0.977583290 0.846404010 0.686966290
> C 0.771131988 0.269306261 0.255617666
> C 0.228868012 0.730693739 0.744382334
> C 0.652850645 0.378765293 0.282708985
> C 0.347149355 0.621234707 0.717291015
> C 0.462994923 0.392175655 0.192940332
> C 0.537005077 0.607824345 0.807059668
> C 0.403148927 0.302768506 0.084001267
> C 0.596851073 0.697231494 0.915998733
> H 0.259814627 0.314740185 0.016757517
> H 0.740185373 0.685259815 0.983242483
> C 0.525750373 0.196269824 0.058490144
> C 0.474249627 0.803730176 0.941509856
> H 0.478830729 0.128376775 0.971151210
> H 0.521169271 0.871623225 0.028848790
> C 0.705969389 0.179966480 0.142934709
> C 0.294030611 0.820033520 0.857065291
> H 0.804698818 0.098652977 0.124221693
> H 0.195301182 0.901347023 0.875778307
> C 0.184454672 0.523754793 0.135830920
> C 0.815545328 0.476245207 0.864169080
> H 0.287576875 0.544363666 0.053314825
> H 0.712423125 0.455636334 0.946685175
> H 0.118308380 0.611071476 0.178999457
> H 0.881691620 0.388928524 0.821000543
> H 0.013394359 0.440437861 0.110093147
> H 0.986605641 0.559562139 0.889906853
> N 0.053284394 0.319764373 0.444382321
> N 0.946715606 0.680235627 0.555617679
> O 0.705759527 0.466685324 0.381694456
> O 0.294240473 0.533314676 0.618305544
> O 0.356827705 0.498376849 0.225196110
> O 0.643172295 0.501623151 0.774803890
> S 0.766003561 0.110202377 0.587178900
> S 0.233996439 0.889797623 0.412821100
> Cu -0.000000000 0.500000000 0.500000000
>
> K_POINTS {automatic}
> 3 3 3 0 0 0
> _______________________________________________
> users mailing list
> users at lists.quantum-espresso.org
> https://lists.quantum-espresso.org/mailman/listinfo/users
GIUSEPPE MATTIOLI
CNR - ISTITUTO DI STRUTTURA DELLA MATERIA
Via Salaria Km 29,300 - C.P. 10
I-00015 - Monterotondo Scalo (RM)
Mob (*preferred*) +39 373 7305625
Tel + 39 06 90672342 - Fax +39 06 90672316
E-mail: <giuseppe.mattioli at ism.cnr.it>
More information about the users
mailing list