| 查看: 3318 | 回复: 9 | ||||
| 当前只显示满足指定条件的回帖，点击这里查看本话题的所有回帖 | ||||
[交流]
【求助】siesta计算中，如何进行BSSE修正？
|
||||
|
即基组重叠误差，全称为Basis Set Superposition Error(BSSE)。对于双分子A和B的相互作用能应该表述为E_interaction = E_AB - E_A - E_B + E_BSSE。第三项即为修正项。对于弱相互作用，E_BSSE所占E_interaction的比例往往不小，甚至超过它，如果不进行校正，可能正负号都不对。 这些也是从网上查到的，但是网上主要介绍的都是高斯算法中的BSSE修正，由于本人用siesta算，想麻烦各位高手多多赐教！！ |
回复此楼» 收录本帖的淘帖专辑推荐
 atk&transiesta |
» 猜你喜欢
导师想让我从独立一作变成了共一第一
已经有8人回复
-
博士读完未来一定会好吗
已经有23人回复
-
到新单位后，换了新的研究方向，没有团队，持续积累2区以上论文，能申请到面上吗
已经有11人回复
-
读博
已经有4人回复
-
JMPT 期刊投稿流程
已经有4人回复
-
心脉受损
已经有5人回复
-
Springer期刊投稿求助
已经有4人回复
-
小论文投稿
已经有3人回复
-
申请2026年博士
已经有6人回复
» 本主题相关商家推荐: (我也要在这里推广)
» 本主题相关价值贴推荐，对您同样有帮助:
如何萃取4-氯丁醇
已经有6人回复- gaussian计算
已经有4人回复
- 急！ 关于Ms从Windows提交到Linux计算的问题？
已经有3人回复
- 请教：orca 中用CCSD(T)计算单点能出错--提示内存不足 该怎么办？
已经有10人回复
- 难道每次计算前都咬进行收敛性测试吗？
已经有14人回复
- siesta 计算电子占据数的问题
已经有3人回复
- siesta计算dipole moment（电偶极矩）
已经有7人回复
- 关于SIESTA中如何计算功函数的一种做法探讨
已经有8人回复
- siesta计算弱相互作用VDW
已经有17人回复
- 【求助】siesta进行ordern计算出错求助
已经有4人回复
- 【求助】SIESTA 中ATOM生成赝势测试后如何看赝势的好坏
已经有4人回复
- 【求助】如何在siesta中设置ghost state原子的计算
已经有16人回复
- 【求助】BSSE计算后如何进行零点能校正？
已经有8人回复
valenhou001
至尊木虫 (职业作家)
- 1ST强帖: 13
- 应助: 241 (大学生)
- 金币: 25701.7
- 散金: 602
- 红花: 166
- 帖子: 3782
- 在线: 873.8小时
- 虫号: 1007127
- 注册: 2010-04-27
- 专业: 凝聚态物性 II ：电子结构
★
tailor(金币+2): 谢谢参与
tailor(金币+2): 谢谢参与
|
这里有很详细的步骤： https://www.mail-archive.com/siesta-l@uam.es/msg02916.html 拷贝如下： Re: Re: [SIESTA-L] about basis-set superposition error (BSSE) Abraham Hmiel Tue, 29 Mar 2011 21:51:07 -0700 "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 >> >> > |
10楼2014-12-02 09:25:05