Dear Colleagues, I want to calculate molecules adsorbing on metallic surfaces, and since the adsorption energy of them is know to be small, I would like to apply the vdW-DF/vdW-DF2 for the calculations. When I calculate the lattice constant with those functionals, I seem to get very large lattice constants, 5-10 % larger than the experimental one. An example for bulk, fcc gold is attached. I perform the calculation looping over different values for the lattice constant; the result seems to be independent of the cut-off energies (in this case I have large values for 'ecutwfc'/'ecutrho' because this was the last run for the convergence tests).
Thus my question, does any one see something wrong in my input? Since the exchage part of the exchange and correlation is treated with a GGA functional (independently whether I use vdW-DF2 and PW86 or vdW-DF and revPBE) and correlation with LDA, plus an attractive(?) long-range contribution, I would expect lattice constants much closer to the experiments. Greetings from Wroclaw, apsi PS Apologies if the topic of my mail is an FAQ/discussed already! -=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=- Ari P Seitsonen / Ari.P.Seitsonen at iki.fi / http://www.iki.fi/~apsi/ Physikalisch-Chemisches Institut der Universitaet Zuerich Tel: +41 44 63 54 497 / Mobile: +41 79 71 90 935 -------------- next part -------------- &control calculation = 'scf' restart_mode = 'from_scratch' prefix = 'bulk-Au' pseudo_dir = '/users/aps/usr/espresso/PP_LIBRARY/' outdir = './scratch/' tstress = .true. tprnfor = .true. verbosity = 'high' etot_conv_thr = 1e-6 forc_conv_thr = 1e-5 / &system ibrav = 2 a = 4.60 nat = 1 ntyp = 1 nbnd = 9 occupations = 'smearing' smearing = 'fermi-dirac' degauss = 0.0036749326 ecutwfc = 80 ecutrho = 800 input_dft = 'vdW-DF2' / &electrons conv_thr = 1e-10 mixing_beta = 0.7 / ATOMIC_SPECIES Au 1.00 Au.bp-van_ak.UPF ATOMIC_POSITIONS crystal Au 0 0 0 K_POINTS AUTOMATIC 3*32 0 0 0