An edit to my post: the line. "1 single-point calculation (0 CG steps) of the adsorbed molecule in the relaxed geometry of system A (use the final .XV file or .xyz file or whatever you want but remove the adsorbed molecule). Find its total energy. Call this E-noghost-AD"
should actually read 1 single-point calculation (0 CG steps) of the adsorbed molecule in the relaxed geometry of system A (use the final .XV file or .xyz file or whatever you want but remove the NANOTUBE). Find its total energy. Call this E-noghost-AD By the way, BSSE corrections should ALWAYS be performed in SIESTA because it uses an incomplete basis set (atomic orbitals), not just when your adsorption energy calculations are different than experiment or you want to lower your adsorption energy to make your calculations look nice for a paper. If you are still seeing vast differences between your adsorption energy calculations and experiments, it may be a good idea to look into different functionals like the van-der-Waals functionals implemented in the SIESTA-trunk version. Or, additionally, you can try to use diffuse basis functions (s-like) to help complete the basis set in the region corresponding to adsorption. On Wed, Mar 30, 2011 at 12:50 AM, Abraham Hmiel <abehm...@gmail.com> wrote: > "Should I set a ghost atom between the molecule and the nanotube ?" > > No, this is what you must do: > > You've already got the relaxed structure of the nanowire + adsorbed > molecule system, right? Let's call this system "A" and its total energy is > E-A > > You need 6 more calculations to complete the puzzle this is pretty much > what you should do: > > 1 calculation with the same k-point grid, mesh grid, and cell size as > system A, but only for the nanotube, fully relaxed to a similar tolerance as > system A. Find its total energy. Call this E-NT. > > 1 calculation with the same k-point grid, mesh grid, and cell size as > system A, but only for the adsorbed molecule, fully relaxed to a similar > tolerance as system A. Find its total energy. Call this E-AD > > 1 single-point calculation (0 CG steps) of the nanowire in the relaxed > geometry of system A (use the final .XV file or .xyz file or whatever you > want but remove the adsorbed molecule). Find its total energy. Call this > E-noghost-NT > > 1 single-point calculation (0 CG steps) of the adsorbed molecule in the > relaxed geometry of system A (use the final .XV file or .xyz file or > whatever you want but remove the adsorbed molecule). Find its total energy. > Call this E-noghost-AD > > 1 single-point calculation (0 CG steps) of the nanowire in the relaxed > geometry of system A, except replace any adsorbate chemical species with > ghost atoms. If you have species in the adsorbate that are present in the > nanowire, for example, a simulation of H2O on a hydrogen-passivated SiNW (or > C in methane on a CNT), then copy the H.psf file to a new file like > H_ghost.psf and then create a new chemical species H_ghost with atomic > number -1 and a different atomic species index. Replace any index of the H > in the adsorbate with the new index, and introduce a new basis set for > H_ghost that is identical to the one you used for H (except for the label > H_ghost). Find its total energy. Call this E-ghost-NT > > 1 single-point calculation (0 CG steps) of the adsorbed molecule in the > relaxed geometry of system A, except replace any nanowire chemical species > with ghost atoms and follow the procedure above if you have any species in > the adsorbate that are also present in the nanowire. Find its total energy. > Call this E-ghost-AD > > The counterpoise correction is: (E-ghost-AD - E-noghost-AD + E-ghost-NT - > E-noghost-NT) call this E-CC. It should be a fraction of an eV, have a > negative sign and very sensitive to the adsorption site geometry. The BSSE > should _reduce_ the adsorption energy... > > and the energy of adsorption is: E-NT + E-AD - E-A + E-CC > > And that is how you do the counterpoise correction with SIESTA. > > Best, > Abraham Hmiel > > Katherine Belz Groves Graduate Fellow in Nanoscience, Xue Group > The College of Nanoscale Science and Engineering at SUNY Albany > "Clouds are not spheres, mountains are not cones, coastlines are not > circles, > and bark is not smooth, nor does lightning travel in a straight line." - > Benoit Mandelbrot > > > On Wed, Mar 30, 2011 at 12:09 AM, yf liu <liuyf1...@gmail.com> wrote: > >> Dear Herbert Fruchtl: >> thank you very much for your reply. I have read the manual >> carefully, but find little message about the ghost atoms. the new question >> is: How can I use the ghost atom to correct the binding energy etc. ? For >> example, I want correct the calculation about the molecule adsorption on the >> nanotube. Should I set a ghost atom between the molecule and the nanotube >> ? >> >> looking forward your reply. >> >> 2011/3/20 Wei Hu <gyrw4...@mail.ustc.edu.cn> >> >>> I am sorry, I am a freshman about the siesta. The input is listed in >>> the following. >>> My calculations about the binding energy do not meet the experimental >>> results,so,I have to check the ghost atoms calculations.Now,The impact is >>> acceptable. >>> Another problem is how to get the chemical potential or total energy >>> of free C or N atom. Does it need to calculate the the ghost atoms affected >>> by the supercell? >>> >>> SystemName C62N_ghost >>> SystemLabel C62N_ghost >>> NumberOfSpecies 3 >>> >>> %block ChemicalSpeciesLabel >>> 1 6 C >>> 2 7 N >>> 3 -6 Cg >>> %endblock ChemicalSpeciesLabel >>> >>> %block PS.lmax >>> C 1 >>> %endblock PS.lmax >>> >>> %include coord.fdf >>> >>> PAO.BasisSize DZP >>> >>> #SolutionMethod dm_on >>> >>> >>> SolutionMethod diagon >>> >>> MeshCutoff 200.0000000 Ry >>> >>> #MD.TypeOfRun Broyden >>> #MD.TypeOfRun CG >>> #MD.NumCGsteps 500 >>> >>> WriteForces >>> >>> MD.MaxForceTol 0.04 eV/Ang >>> #DM.UseSaveDM T >>> >>> MaxSCFIterations 100 >>> DM.MixingWeight 0.1 >>> DM.NumberPulay 6 >>> #DM.MixingWeight 0.25 >>> #DM.NumberPulay 0 >>> >>> >>> SpinPolarized .true. >>> #FixSpin .true. >>> #TotalSpin 2.0 >>> >>> WriteMullikenPop 1 >>> >>> NetCharge -1.0 >>> >>> %block kgrid_Monkhorst_Pack >>> 2 0 0 0.0 >>> 0 2 0 0.0 >>> 0 0 2 0.0 >>> %endblock kgrid_Monkhorst_Pack >>> >>> >>> >>> > -----Original E-mail----- >>> > From: "Herbert Fruchtl" <herbert.fruc...@st-andrews.ac.uk> >>> > Sent Time: 2011-3-18 20:06:46 >>> > To: siesta-l@uam.es >>> > Cc: >>> > Subject: Re: [SIESTA-L] about basis-set superposition error (BSSE) >>> > >>> > From the energy part of the output alone we can't tell if the input >>> was >>> > correct. You see a smallish difference and lower energy with the ghost, >>> which >>> > one would expect. >>> > >>> > Total energies are meaningless, and of course adding a few basis >>> functions won't >>> > change the total by a lot. You need to compare differences (like a >>> binding >>> > energy, as you indicated in your original post). >>> > >>> > In the example outputs below, the difference (in absolute energies) is >>> 0.09 eV. >>> > That's 2 kcal/mol or 9 kJ/mol. Depending on the interaction you are >>> looking at, >>> > this may or may not be negligible. >>> > >>> > Herbert >>> > >>> > On 03/18/2011 02:00 AM, Wei Hu wrote: >>> > > Dear,the ghost atoms calculations seem to little effects on the total >>> energy from my results. Is there any wrong? >>> > > >>> > > My results about a nitrogen-vacancy center in bulk diamond C62N >>> (2*2*2): >>> > > >>> > > The results with ghost atoms: >>> > > >>> > > siesta: Program's energy decomposition (eV): >>> > > siesta: Eions = 16680.576136 >>> > > siesta: Ena = 3653.202863 >>> > > siesta: Ekin = 7163.588024 >>> > > siesta: Enl = -820.400928 >>> > > siesta: DEna = -182.458272 >>> > > siesta: DUscf = 15.575306 >>> > > siesta: DUext = 0.000000 >>> > > siesta: Exc = -3054.613637 >>> > > siesta: eta*DQ = 0.000000 >>> > > siesta: Emadel = 0.000000 >>> > > siesta: Ekinion = 0.000000 >>> > > siesta: Eharris = -9905.679557 >>> > > siesta: Etot = -9905.682781 >>> > > siesta: FreeEng = -9905.682783 >>> > > >>> > > siesta: Final energy (eV): >>> > > siesta: Kinetic = 7163.588024 >>> > > siesta: Hartree = 894.608032 >>> > > siesta: Ext. field = 0.000000 >>> > > siesta: Exch.-corr. = -3054.613637 >>> > > siesta: Ion-electron = -8393.758885 >>> > > siesta: Ion-ion = -6515.506314 >>> > > siesta: Ekinion = 0.000000 >>> > > siesta: Total = -9905.682781 >>> > > >>> > > The results without ghost atoms: >>> > > >>> > > siesta: Program's energy decomposition (eV): >>> > > siesta: Eions = 16680.576136 >>> > > siesta: Ena = 3653.202863 >>> > > siesta: Ekin = 7164.385562 >>> > > siesta: Enl = -820.554764 >>> > > siesta: DEna = -182.943454 >>> > > siesta: DUscf = 15.589130 >>> > > siesta: DUext = 0.000000 >>> > > siesta: Exc = -3054.694636 >>> > > siesta: eta*DQ = 0.000000 >>> > > siesta: Emadel = 0.000000 >>> > > siesta: Ekinion = 0.000000 >>> > > siesta: Eharris = -9905.587400 >>> > > siesta: Etot = -9905.591435 >>> > > siesta: FreeEng = -9905.591437 >>> > > >>> > > siesta: Final energy (eV): >>> > > siesta: Kinetic = 7164.385562 >>> > > siesta: Hartree = 894.861106 >>> > > siesta: Ext. field = 0.000000 >>> > > siesta: Exch.-corr. = -3054.694636 >>> > > siesta: Ion-electron = -8394.637153 >>> > > siesta: Ion-ion = -6515.506314 >>> > > siesta: Ekinion = 0.000000 >>> > > siesta: Total = -9905.591435 >>> > > >>> > > >>> > > >>> > > >>> > >> -----Original E-mail----- >>> > >> From: "Herbert Fruchtl"<herbert.fruc...@st-andrews.ac.uk> >>> > >> Sent Time: 2011-3-17 22:36:53 >>> > >> To: siesta-l@uam.es >>> > >> Cc: >>> > >> Subject: Re: [SIESTA-L] about basis-set superposition error (BSSE) >>> > >> >>> > >> You can calculate counterpoise-corrected energies by specifying >>> ghost atoms. >>> > >> There is no way of doing counterpoise-corrected geometry >>> optimisations in SIESTA. >>> > >> >>> > >> Herbert >>> > >> >>> > >> On 03/17/2011 01:57 PM, yf liu wrote: >>> > >>> Hello everybody, >>> > >>> >>> > >>> whether the basis-set superposition error (BSSE) is included in >>> siesta >>> > >>> calculation. when i calculate the molecule adsorption on carbon >>> nanotube, i >>> > >>> found the adsorption energy is some higher than the well know >>> value. how can i >>> > >>> add the BSSE in siesta calculation? >>> > >>> >>> > >>> yours >>> > >>> yufeng liu >>> > >>> >>> > >> >>> > >> -- >>> > >> Herbert Fruchtl >>> > >> Senior Scientific Computing Officer >>> > >> School of Chemistry, School of Mathematics and Statistics >>> > >> University of St Andrews >>> > >> -- >>> > >> The University of St Andrews is a charity registered in Scotland: >>> > >> No SC013532 >>> > > >>> > >>> > -- >>> > Herbert Fruchtl >>> > Senior Scientific Computing Officer >>> > School of Chemistry, School of Mathematics and Statistics >>> > University of St Andrews >>> > -- >>> > The University of St Andrews is a charity registered in Scotland: >>> > No SC013532 >>> >>> >> >
