| 查看: 3459 | 回复: 9 | ||||
[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
346求调剂[0856]
已经有7人回复
-
一志愿山大07化学 332分 四六级已过 本科山东双非 求调剂！
已经有3人回复
-
310求调剂
已经有3人回复
-
机械专硕299求调剂至材料
已经有4人回复
-
070300化学319求调剂
已经有7人回复
-
08工科 320总分 求调剂
已经有6人回复
-
一志愿天津大学化学工艺专业（081702）315分求调剂
已经有12人回复
-
307求调剂
已经有3人回复
-
265求调剂
已经有3人回复
-
301求调剂
已经有10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
ifmc1234
木虫 (著名写手)
- 应助: 18 (小学生)
- 金币: 2797.9
- 散金: 6001
- 红花: 14
- 帖子: 1761
- 在线: 572.6小时
- 虫号: 825500
- 注册: 2009-08-10
- 专业: 理论和计算化学
★ ★ ★
tailor(金币+2):谢谢参与
qasd(金币+2):谢谢 2010-09-11 18:57:22
tailor(金币+2):谢谢参与
qasd(金币+2):谢谢 2010-09-11 18:57:22
|
不知道你算的是分子还是固体，建议不要用它来讨论。你可在userguide里搜一下，我没看有什么人用siesta讨论过BSSE,在userguide里的27页提到了BSSE。不过具体怎么做还不知道。说明书中只提到PAO.Basis里设置时，用Ghost原子可以处理BSSE。不过如果根据你上面的定义，我想可以这样E_interaction = E_AB - E_A - E_B + Sum_E_BSSE,其中Sum_E_BSSE为A,B单个的BSSE之和。也就是说用鬼原子来处理BSSE，然后根据上面求的 |
2楼2010-09-06 21:29:23
zhangguangping
木虫 (著名写手)
- 1ST强帖: 23
- 应助: 71 (初中生)
- 贵宾: 0.031
- 金币: 1973.1
- 散金: 9059
- 红花: 79
- 帖子: 2681
- 在线: 2128.8小时
- 虫号: 529624
- 注册: 2008-03-20
- 性别: GG
- 专业: 原子和分子物理
3楼2010-09-06 21:30:41
★
tailor(金币+2):谢谢参与
tailor(金币+2):谢谢参与
 |
4楼2010-09-06 21:41:19
请教老外，得到的回复
★
zxzj05(金币+1):谢谢回帖交流 2010-09-25 22:34:07
zxzj05(金币+1):谢谢回帖交流 2010-09-25 22:34:07
|
What we calculated is the Basis Set Superposition Error (BSSE) that is an error occurring when you use a finite basis set to study, for example, a chemical reaction. In a tipical reaction you calculate the electronic structure of the reactants separately, let us say A and B, and then the electronic structure of the product, A+B. What happens in this case is that you are using different basis sets to represent the reactants in each step of the calculation. When you calculate the isolated reactant A, you use the basis set of A, and the electrons of A will be arranged spatially on these provided orbitals. However, when you put both systems together in the A+B, the electrons of A will have now a greater spatial freedom, since they can now occupy orbitals coming from B. The same happens with B. You have different basis sets to distribute the electrons of A in the two situations. To correct this, you need to recalculate the energy of the isolated reactants (A and B) by considering the basis set of both reactants, i.e., the calculations of the isolated reactant A will now include the basis sets of B (without the system B itself). It needs to be done for each configuration that you take into account in the reaction of A+B. Now you will have the same basis set for the isolated reactants and the product. You can look for more information searching for "counterpoise correction" |
5楼2010-09-10 19:36:33
zhangguangping
木虫 (著名写手)
- 1ST强帖: 23
- 应助: 71 (初中生)
- 贵宾: 0.031
- 金币: 1973.1
- 散金: 9059
- 红花: 79
- 帖子: 2681
- 在线: 2128.8小时
- 虫号: 529624
- 注册: 2008-03-20
- 性别: GG
- 专业: 原子和分子物理
6楼2010-09-11 10:02:01
夕阳西下
金虫 (著名写手)
- 应助: 138 (高中生)
- 贵宾: 0.041
- 金币: 9247.1
- 散金: 73
- 红花: 28
- 沙发: 1
- 帖子: 1268
- 在线: 1164.4小时
- 虫号: 220750
- 注册: 2006-03-20
- 专业: 凝聚态物性 II ：电子结构
- 管辖: 物理
7楼2010-09-11 12:08:57
cenwanglai
荣誉版主 (知名作家)
老和山猥琐派九段
- 1ST强帖: 5
- 应助: 46 (小学生)
- 贵宾: 8.842
- 金币: 7440.4
- 散金: 9654
- 红花: 118
- 沙发: 23
- 帖子: 5306
- 在线: 1961.4小时
- 虫号: 537452
- 注册: 2008-04-01
- 性别: GG
- 专业: 理论和计算化学
- 管辖: 仿真模拟
8楼2010-09-25 16:50:46
skyscorpioa
新虫 (初入文坛)
- 应助: 0 (幼儿园)
- 金币: 529.9
- 散金: 12
- 红花: 1
- 帖子: 38
- 在线: 76.6小时
- 虫号: 1089120
- 注册: 2010-09-04
- 专业: 凝聚态物性 II ：电子结构
9楼2010-11-20 02:43:11
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
