Deiar Abraham Hmiel
Thank you very much for you detailed reply. I will try it.
2011/3/30 Abraham Hmiel <abehm...@gmail.com>
> 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
>>>>
>>>>
>>>
>>
>
