You can subscribe to this list here.
2010 
_{Jan}

_{Feb}
(2) 
_{Mar}
(22) 
_{Apr}
(11) 
_{May}
(19) 
_{Jun}
(25) 
_{Jul}
(13) 
_{Aug}
(28) 
_{Sep}
(17) 
_{Oct}
(12) 
_{Nov}
(17) 
_{Dec}
(14) 

2011 
_{Jan}
(1) 
_{Feb}
(11) 
_{Mar}
(20) 
_{Apr}
(7) 
_{May}
(44) 
_{Jun}
(33) 
_{Jul}
(6) 
_{Aug}
(10) 
_{Sep}
(38) 
_{Oct}
(12) 
_{Nov}
(21) 
_{Dec}
(9) 
2012 
_{Jan}
(2) 
_{Feb}
(15) 
_{Mar}
(37) 
_{Apr}
(29) 
_{May}
(36) 
_{Jun}
(24) 
_{Jul}
(16) 
_{Aug}
(14) 
_{Sep}
(27) 
_{Oct}
(63) 
_{Nov}
(34) 
_{Dec}
(10) 
2013 
_{Jan}
(6) 
_{Feb}
(25) 
_{Mar}
(34) 
_{Apr}
(23) 
_{May}
(36) 
_{Jun}
(54) 
_{Jul}
(9) 
_{Aug}
(19) 
_{Sep}
(22) 
_{Oct}
(11) 
_{Nov}
(35) 
_{Dec}
(33) 
2014 
_{Jan}
(20) 
_{Feb}
(21) 
_{Mar}
(24) 
_{Apr}
(31) 
_{May}
(29) 
_{Jun}
(31) 
_{Jul}
(49) 
_{Aug}
(51) 
_{Sep}
(15) 
_{Oct}
(28) 
_{Nov}
(28) 
_{Dec}
(45) 
2015 
_{Jan}
(23) 
_{Feb}
(24) 
_{Mar}
(51) 
_{Apr}
(54) 
_{May}
(8) 
_{Jun}
(26) 
_{Jul}
(36) 
_{Aug}
(61) 
_{Sep}
(62) 
_{Oct}
(77) 
_{Nov}
(55) 
_{Dec}
(29) 
2016 
_{Jan}
(118) 
_{Feb}
(50) 
_{Mar}
(68) 
_{Apr}
(97) 
_{May}
(44) 
_{Jun}
(26) 
_{Jul}
(34) 
_{Aug}
(33) 
_{Sep}
(52) 
_{Oct}
(58) 
_{Nov}
(57) 
_{Dec}
(57) 
2017 
_{Jan}
(34) 
_{Feb}
(93) 
_{Mar}
(56) 
_{Apr}

_{May}

_{Jun}

_{Jul}

_{Aug}

_{Sep}

_{Oct}

_{Nov}

_{Dec}

S  M  T  W  T  F  S 

1
(3) 
2

3
(3) 
4

5

6
(2) 
7

8

9

10

11

12

13

14

15

16
(3) 
17
(5) 
18

19
(1) 
20
(1) 
21

22

23

24

25

26

27

28
(3) 
29

30
(1) 





From: Yundi Quan <quanyundi@gm...>  20130930 18:35:02

Hi, Take the lastest version of phonopy as an example. The function get_independent_atoms_in_dot_scf in wien2k.py gets the positions of atoms from case.scf. In WIEN2k_11, the format of the line in case.scf starting with ':POS' is FORMAT(////,':POS',i3.3,':',1x,'ATOM ',I4,1X,'POSITION =', 3F8.5,2X,'MULTIPLICITY =',I2,' ZZ=',f7.3,2x,a10) see line 2094 of /SRC_lapw2/l2main.F In WIEN2k_13, the format becomes FORMAT(////,':POS',i3.3,':',1x,'ATOM ',I4,1X,'X,Y,Z =', 3F8.5,2X,'MULT=',I2,' ZZ=',f7.3,2x,a10) see line 2247 of /SRC_lapw2/l2main.F The function get_independent_atoms_in_dot_scf should be updated to something like this: def get_independent_atoms_in_dot_scf( filename ): positions = [] for line in open( filename ): if line[:4] == ":POS": x = float( line[27:34] ) y = float( line[35:42] ) z = float( line[43:50] ) num_atom = int(line[4:7]) positions.append([x,y,z]) return np.array( positions )[num_atom:] On Sat, Sep 28, 2013 at 4:35 AM, Atsushi Togo <atz.togo@...> wrote: > If you wish the update, could you describe it in detail? Where is it > written in the document? Is there no backward compatibility? What is > the string format in Fortran. etc... > > Togo > > On Sat, Sep 28, 2013 at 3:23 PM, Yundi Quan <quanyundi@...> wrote: > > Hi, > > In WIEN2k_13, the format of the atomic positions in case.scf file have > > changed. It now looks like this: > > :POS001: ATOM 1 X,Y,Z = 0.00000 0.00000 0.00000 MULT= 1 > > > > Could you update phonopy so that it can work with WIEN2k_13? Thanks a > lot. > > > > Yundi > > > > >  > > October Webinars: Code for Performance > > Free Intel webinars can help you accelerate application performance. > > Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most > > from > > the latest Intel processors and coprocessors. See abstracts and register > > > > > http://pubads.g.doubleclick.net/gampad/clk?id=60133471&iu=/4140/ostg.clktrk > > _______________________________________________ > > Phonopyusers mailing list > > Phonopyusers@... > > https://lists.sourceforge.net/lists/listinfo/phonopyusers > > > > > >  > Atsushi Togo > http://atztogo.github.com/ > atz.togo@... > 
From: torsten weissbach <torsten.weissbach@ph...>  20130928 13:27:59

Hi Yundi, a quick fix in phonopy1.6.4/lib/python/interface/wien2k.py: find if line[:4] == ":POS": change the numbers in line like this: x = float( line[27:34] ) y = float( line[35:42] ) z = float( line[43:50] ) num_atom = int(line[4:7]) should work but maybe some extra fixes are needed. It would be better to have a regex fit here when working with different wien versions. Unfortunately I dont know python. After you changed this, run setup.py like on install. Regards, Torsten 
From: Atsushi Togo <atz.togo@gm...>  20130928 11:35:11

If you wish the update, could you describe it in detail? Where is it written in the document? Is there no backward compatibility? What is the string format in Fortran. etc... Togo On Sat, Sep 28, 2013 at 3:23 PM, Yundi Quan <quanyundi@...> wrote: > Hi, > In WIEN2k_13, the format of the atomic positions in case.scf file have > changed. It now looks like this: > :POS001: ATOM 1 X,Y,Z = 0.00000 0.00000 0.00000 MULT= 1 > > Could you update phonopy so that it can work with WIEN2k_13? Thanks a lot. > > Yundi > >  > October Webinars: Code for Performance > Free Intel webinars can help you accelerate application performance. > Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most > from > the latest Intel processors and coprocessors. See abstracts and register > > http://pubads.g.doubleclick.net/gampad/clk?id=60133471&iu=/4140/ostg.clktrk > _______________________________________________ > Phonopyusers mailing list > Phonopyusers@... > https://lists.sourceforge.net/lists/listinfo/phonopyusers >  Atsushi Togo http://atztogo.github.com/ atz.togo@... 
From: Yundi Quan <quanyundi@gm...>  20130928 06:23:56

Hi, In WIEN2k_13, the format of the atomic positions in case.scf file have changed. It now looks like this: :POS001: ATOM 1 X,Y,Z = 0.00000 0.00000 0.00000 MULT= 1 Could you update phonopy so that it can work with WIEN2k_13? Thanks a lot. Yundi 
From: Atsushi Togo <atz.togo@gm...>  20130920 06:14:21

Hi Sung, Usually phonon calculation doesn't need a lot of computational demand. So you can just test it by increasing mesh numbers from small numbers until property that you want converges. I think numerical error is negligible. Togo On Fri, Sep 20, 2013 at 5:42 AM, Sung Gu Kang <sk2583@...> wrote: > > Dear Dr.Togo, > > When we set the mesh point (mesh sampling) in mesh.conf, what can be the standard (or criteria) to set the number of MP? I understand the higher number will be the better, but how high it should be? It could cause numerical error if it is too high. > > Best wishes, > Sung  Atsushi Togo http://atztogo.github.com/ atz.togo@... 
From: Sung Gu Kang <sk2583@co...>  20130919 20:43:42

Dear Dr.Togo, When we set the mesh point (mesh sampling) in mesh.conf, what can be the standard (or criteria) to set the number of MP? I understand the higher number will be the better, but how high it should be? It could cause numerical error if it is too high. Best wishes, Sung 
From: Christian Carbogno <christian.carbogno@fh...>  20130917 16:10:55

Dear Carlo, The second strategy (molecule in huge supercell) will work with phonopy / phonopyFHIaims. However, you can also use FHIaims' native interface for the calculations of vibrations in nonperiodic systems (cf. see the FHIaims manual for details). In that case, you would not even need the "molecule in huge supercell" trick. Cheers, Chris   Dr. Christian Carbogno Phone: +493084134816 FritzHaberInstitut der MaxPlanckGesellschaft Fax: +493084134701 Faradayweg 46 D14195 BerlinDahlem / Germany email: carbogno@...  WWW: http://www.fhiberlin.mpg.de/th/th.html key:10.07351160023.05:fhi On 17.09.2013, at 17:36, Carlo Motta <MOTTAC@...> wrote: > Hi, > > I have a question. is it possible to use phonopy to calculate the vibrations for an isolated (i.e. non periodic) system, like a molecule in gas phase? > I am using AIMS and the relative phonopy interface. > If the answer is negative, do you think that I might use a trick, like putting the molecule in a big supercell and calculate the vibrations at Gamma for the "periodic" fictitious system? > > Thank you in advance, > Carlo > >  > LIMITED TIME SALE  Full Year of Microsoft Training For Just $49.99! > 1,500+ hours of tutorials including VisualStudio 2012, Windows 8, SharePoint > 2013, SQL 2012, MVC 4, more. BEST VALUE: New MultiLibrary Power Pack includes > Mobile, Cloud, Java, and UX Design. Lowest price ever! Ends 9/20/13. > http://pubads.g.doubleclick.net/gampad/clk?id=58041151&iu=/4140/ostg.clktrk > _______________________________________________ > Phonopyusers mailing list > Phonopyusers@... > https://lists.sourceforge.net/lists/listinfo/phonopyusers 
From: Carlo Motta <MOTTAC@tc...>  20130917 15:36:28

Hi, I have a question. is it possible to use phonopy to calculate the vibrations for an isolated (i.e. non periodic) system, like a molecule in gas phase? I am using AIMS and the relative phonopy interface. If the answer is negative, do you think that I might use a trick, like putting the molecule in a big supercell and calculate the vibrations at Gamma for the "periodic" fictitious system? Thank you in advance, Carlo 
From: Tian Lan <tianlan@ca...>  20130917 05:00:26

Forget about my concern about the space group issue, I think 227 is correct. Sorry. Tian On Mon, Sep 16, 2013 at 9:42 PM, Tian Lan <tianlan@...> wrote: > Dear Dr. Togo, > > Thank you. Here I found an obvious problem, I did not notice that the > space group that phonopy showed is 227, which is not right, I guess I made > some mistakes. Because I am working on a diamond cubic silicon. I think it > is a fcc structure with a basis, how can its reciprocal space belong to > 227? I compare with the reference, this is what Si should look like > https://wiki.alcf.anl.gov/wparker/index.php/Silicon > > I am not familiar with the crystal symmetry, but what I did, as listed in > my first email, is I used a 8 atoms unit cell to generate a 2 by 2 by 2 > supercell, and as you emphasized, when it comes to the kpoint, like DOS or > band structure, I tried to convert it to a primitive cell first, my > Primitive axis: > [ 0. 0.5 0.5] > [ 0.5 0. 0.5] > [ 0.5 0.5 0. ] > I don't see any obvious problem. I generally followed your example input > here (but I don't have your example installed in my cluster): > https://github.com/atztogo/phonopy/tree/master/example/Siphonopyqha > I feel that I may make a silly mistake somewhere... > > Best, > Tian > > > On Mon, Sep 16, 2013 at 8:06 PM, Atsushi Togo <atz.togo@...> wrote: > >> Hi, >> >> Phonopy has its own symmetry handler and usually the topology of band >> structure is less sensitive to the VASP setting. >> But as Jan recommended, the force calculation should be done as high >> accuracy as possible. For example, I recommend always LREAL = .FALSE. >> >> In phonopy, qpoint has to be set in primitive basis of reciprocal >> space, i.e., I guess X point you mean would be (0.5, 0, 0.5). >> No. 227 is searched at http://www.cryst.ehu.es/cryst/get_kvec.html . >> In Bilbao crystallographic server convention, no name of R exists for >> this space group. >> >> The oddity of mesh sampling arises by three fold rotation of symmetry. >> If the center of rotation is not sampled, the sampling becomes >> notuniform. In the example, the odd numbers give to sample >> Gammapoint, but the even numbers mesh doesn't. >> >> Togo >> >> On Tue, Sep 17, 2013 at 7:29 AM, Tian Lan <tianlan@...> wrote: >> > >> > >> >  Forwarded message  >> > From: Tian Lan <tianlan@...> >> > Date: Mon, Sep 16, 2013 at 3:22 PM >> > Subject: Re: [Phonopyusers] Phonon dispersion error in Si >> > To: Jan Jaeken <Jan.Jaeken@...> >> > >> > >> > Hi Jan, >> > >> > Thank you. You are right, I think it is not stringent enough. But I >> think it >> > is basically okay because it has been converged. 4 by 4 by 4 kmesh >> for a 2 >> > by 2 by 2 supercell, is just a 8 by 8 by 8 mesh for a unit cell. I >> think for >> > a phonon calculation, this is usually sufficient. My experience is, >> this may >> > influence the frequencies a little bit, but would not destroy any >> symmetry >> > relation, which is my major concern here. >> > >> > As I said, for example, at X point, phonon branches should be converged >> to 3 >> > frequency points, i.e., there is certain degeneracy there. But my >> > calculation converged to 4 frequency points, which means some branches >> broke >> > the symmetry. And when I did the DOS calculation, as I mentioned, >> phonopy's >> > analysis gives 2000+ irreducible k points for a 30by30by30 meshes, but >> for >> > the same mesh, phonopy's example gives 800 k points, which means that my >> > setup decreased the order of symmetry, this could not happen for a >> > calculation of any accuracy, if I understand right. >> > >> > But I could not figure out my problem, since it looks a simple case for >> > phonopy,and I was following the procedure from phonopy manual. >> > >> > Best, >> > Tian >> > >> > >> > On Mon, Sep 16, 2013 at 3:03 PM, Jan Jaeken <Jan.Jaeken@...> >> wrote: >> >> >> >> Hello, >> >> >> >> Those settings you have been using aren't stringent enough. Especially >> >> the number of kpoints you have been using is much too low. First, try >> >> using a larger kpoint grid. You should compare the forces of your >> >> current calculation against those of the calculation at a higher >> >> setting. They shouldn't deviate by more than 1 meV/A. >> >> >> >> Furthermore, a 2x2x2 supercell might be too small. The forces farthest >> >> from the displaced atom (check periodic boundary conditions) should be >> >> a thousand times smaller than those on the displaced atom itself. >> >> Personally, I wouldn't trust anything smaller than a 3x3x3 supercell, >> >> as periodic images will almost certainly interfere with each other in >> >> a 2x2x2. >> >> >> >> Also, make sure to set ENCUT to at least 130% of the POTCAR's ENMAX, >> >> and LREAL = .False. always. Check that you are using a proper ISMEAR >> >> and SIGMA as well (see the VASP manual). >> >> >> >> These force calculations are highly sensitive, and it is important to >> >> run strict convergence tests. Problems in the phonon band structure >> >> often stem from poorlyconverged forces. >> >> >> >> Regards, >> >> >> >> Jan Jaeken >> >> >> >> >> >> Quoting Tian Lan <tianlan@...>: >> >> >> >> > Dear all, >> >> > >> >> > I plan to run a MD simulation for diamond cubic Si. Before this, I >> did >> >> > selfconsistent calculation and used phonopy to get phonon dispersion >> >> > and >> >> > DOS to check the correctness of my VASP setup. I thought it would be >> an >> >> > easy task, however, I found some strange problems especially for >> >> > dispersion. There are no big deviation(like frequency)from its >> >> > expectation >> >> > values, but some obvious errors. >> >> > >> >> > Specifically, at X (0.5,0,0) point, the modes should converge to 3 >> >> > points, >> >> > as I found from any reference. But my calculation gave 4 branches, it >> >> > looks >> >> > like some symmetry was broken. And gamma>R(0.5, 0.5,0), two TA >> branches >> >> > should not be degenerate, but my calculation gave a degenerate >> curves. >> >> > >> >> > The small displacement method is used. I used a 2 by 2 by 2 >> supercell, >> >> > with the unit cell 8 atoms; 4 by 4 by 4 kpoint mesh; the unit cell >> >> > POSCAR >> >> > like this >> >> >  >> >> > Diamond cubic Si >> >> > 5.465 >> >> > 1.0 0.0 0.0 >> >> > 0.0 1.0 0.0 >> >> > 0.0 0.0 1.0 >> >> > Si >> >> > 8 >> >> > Direct >> >> > 0.8750000000000000 0.8750000000000000 0.8750000000000000 >> >> > 0.8750000000000000 0.3750000000000000 0.3750000000000000 >> >> > 0.3750000000000000 0.8750000000000000 0.3750000000000000 >> >> > 0.3750000000000000 0.3750000000000000 0.8750000000000000 >> >> > 0.1250000000000000 0.1250000000000000 0.1250000000000000 >> >> > 0.1250000000000000 0.6250000000000000 0.6250000000000000 >> >> > 0.6250000000000000 0.1250000000000000 0.6250000000000000 >> >> > 0.6250000000000000 0.6250000000000000 0.1250000000000000 >> >> > >> >> >  >> >> > >> >> > My INCAR is like this, I don't think there is any obvious problem >> >> > compared >> >> > with, for example, phonopy example for Si >> >> >  >> >> > SYSTEM = Si Relax >> >> > ISTART = 0 >> >> > ICHARG = 2 >> >> > ENCUT = 245.345 >> >> > PREC = Accurate >> >> > NELMIN = 4 # do a minimum of four electronic steps >> >> > NELM = 40 # maximum of 40 electronic steps >> >> > EDIFF = 1E8 >> >> > ADDGRID = .TRUE. >> >> > IBRION = 1 >> >> > ISMEAR = 0 >> >> > SIGMA = 0.05 >> >> > LREAL = Auto >> >> > LWAVE = .False. >> >> > LPLANE = .TRUE. #recommended for Linux clusters >> >> > NPAR = 8 # twice the number of nodes >> >> >  >> >> > >> >> > Then I followed the standard steps for phonopy. >> >> > >> >> > (1)Get the Force sets >> >> > phonopy f disp001/vasprun.xml >> >> > (2) calculate DOS or band, in here, convert to primitive axes, the >> >> > band.conf is as follows: >> >> >  >> >> > ATOM_NAME = Si >> >> > DIM = 2 2 2 >> >> > PRIMITIVE_AXIS = 0 1/2 1/2 1/2 0 1/2 1/2 1/2 0 >> >> > BAND = 0 0 0 0 0 1/2 >> >> > #BAND = 0 0 0 1/2 0 0, 0 1/2 1/2 0 0 0, 0 0 0 1/2 1/2 1/2 >> >> > BAND_POINTS = 101 >> >> >  >> >> > And a problem really concerned me is, the phonopy analysis showed >> that >> >> > the >> >> > irreducible k points after DOS calculation(mesh.conf) are 2000 more >> >> > points, >> >> > while in phonopy example, there is only about 800, the symmetry >> analysis >> >> > is >> >> > like this >> >> >  >> >> > Mesh sampling mode >> >> > Settings: >> >> > Sampling mesh: [30 30 30] >> >> > Supercell: [2 2 2] >> >> > Primitive axis: >> >> > [ 0. 0.5 0.5] >> >> > [ 0.5 0. 0.5] >> >> > [ 0.5 0.5 0. ] >> >> > Spacegroup: Fd3m (227) >> >> > Calculating force constants... >> >> > Number of irreducible qpoints: 2480 >> >> >  >> >> > >> >> > I don't know why there is such a strange problem, I can not find any >> >> > evident issue related to my calculation. I hope I provide enough >> >> > information to evaluate it. Thanks. >> >> > >> >> > >> >> >  >> >> > Lan, Tian >> >> > Ph.D. Candidate, Department of Applied Physics and Materials Science >> >> > California Institute of Technology, >> >> > Caltech M/C 13878, Pasadena, CA, 91125 >> >> >> >> >> >> >> >> >> >> >> >> >>  >> >> LIMITED TIME SALE  Full Year of Microsoft Training For Just $49.99! >> >> 1,500+ hours of tutorials including VisualStudio 2012, Windows 8, >> >> SharePoint >> >> 2013, SQL 2012, MVC 4, more. BEST VALUE: New MultiLibrary Power Pack >> >> includes >> >> Mobile, Cloud, Java, and UX Design. Lowest price ever! Ends 9/20/13. >> >> >> >> >> http://pubads.g.doubleclick.net/gampad/clk?id=58041151&iu=/4140/ostg.clktrk >> >> _______________________________________________ >> >> Phonopyusers mailing list >> >> Phonopyusers@... >> >> https://lists.sourceforge.net/lists/listinfo/phonopyusers >> > >> > >> > >> > >> >  >> > Lan, Tian >> > Ph.D. Candidate, Department of Applied Physics and Materials Science >> > California Institute of Technology, >> > Caltech M/C 13878, Pasadena, CA, 91125 >> > >> > >> > >> >  >> > Lan, Tian >> > Ph.D. Candidate, Department of Applied Physics and Materials Science >> > California Institute of Technology, >> > Caltech M/C 13878, Pasadena, CA, 91125 >> > >> > >>  >> > LIMITED TIME SALE  Full Year of Microsoft Training For Just $49.99! >> > 1,500+ hours of tutorials including VisualStudio 2012, Windows 8, >> SharePoint >> > 2013, SQL 2012, MVC 4, more. BEST VALUE: New MultiLibrary Power Pack >> > includes >> > Mobile, Cloud, Java, and UX Design. Lowest price ever! Ends 9/20/13. >> > >> http://pubads.g.doubleclick.net/gampad/clk?id=58041151&iu=/4140/ostg.clktrk >> > _______________________________________________ >> > Phonopyusers mailing list >> > Phonopyusers@... >> > https://lists.sourceforge.net/lists/listinfo/phonopyusers >> > >> >> >> >>  >> Atsushi Togo >> http://atztogo.github.com/ >> atz.togo@... >> > > > >  > Lan, Tian > Ph.D. Candidate, Department of Applied Physics and Materials Science > California Institute of Technology, > Caltech M/C 13878, Pasadena, CA, 91125 >  Lan, Tian Ph.D. Candidate, Department of Applied Physics and Materials Science California Institute of Technology, Caltech M/C 13878, Pasadena, CA, 91125 
From: Tian Lan <tianlan@ca...>  20130917 04:42:28

Dear Dr. Togo, Thank you. Here I found an obvious problem, I did not notice that the space group that phonopy showed is 227, which is not right, I guess I made some mistakes. Because I am working on a diamond cubic silicon. I think it is a fcc structure with a basis, how can its reciprocal space belong to 227? I compare with the reference, this is what Si should look like https://wiki.alcf.anl.gov/wparker/index.php/Silicon I am not familiar with the crystal symmetry, but what I did, as listed in my first email, is I used a 8 atoms unit cell to generate a 2 by 2 by 2 supercell, and as you emphasized, when it comes to the kpoint, like DOS or band structure, I tried to convert it to a primitive cell first, my Primitive axis: [ 0. 0.5 0.5] [ 0.5 0. 0.5] [ 0.5 0.5 0. ] I don't see any obvious problem. I generally followed your example input here (but I don't have your example installed in my cluster): https://github.com/atztogo/phonopy/tree/master/example/Siphonopyqha I feel that I may make a silly mistake somewhere... Best, Tian On Mon, Sep 16, 2013 at 8:06 PM, Atsushi Togo <atz.togo@...> wrote: > Hi, > > Phonopy has its own symmetry handler and usually the topology of band > structure is less sensitive to the VASP setting. > But as Jan recommended, the force calculation should be done as high > accuracy as possible. For example, I recommend always LREAL = .FALSE. > > In phonopy, qpoint has to be set in primitive basis of reciprocal > space, i.e., I guess X point you mean would be (0.5, 0, 0.5). > No. 227 is searched at http://www.cryst.ehu.es/cryst/get_kvec.html . > In Bilbao crystallographic server convention, no name of R exists for > this space group. > > The oddity of mesh sampling arises by three fold rotation of symmetry. > If the center of rotation is not sampled, the sampling becomes > notuniform. In the example, the odd numbers give to sample > Gammapoint, but the even numbers mesh doesn't. > > Togo > > On Tue, Sep 17, 2013 at 7:29 AM, Tian Lan <tianlan@...> wrote: > > > > > >  Forwarded message  > > From: Tian Lan <tianlan@...> > > Date: Mon, Sep 16, 2013 at 3:22 PM > > Subject: Re: [Phonopyusers] Phonon dispersion error in Si > > To: Jan Jaeken <Jan.Jaeken@...> > > > > > > Hi Jan, > > > > Thank you. You are right, I think it is not stringent enough. But I > think it > > is basically okay because it has been converged. 4 by 4 by 4 kmesh for > a 2 > > by 2 by 2 supercell, is just a 8 by 8 by 8 mesh for a unit cell. I think > for > > a phonon calculation, this is usually sufficient. My experience is, this > may > > influence the frequencies a little bit, but would not destroy any > symmetry > > relation, which is my major concern here. > > > > As I said, for example, at X point, phonon branches should be converged > to 3 > > frequency points, i.e., there is certain degeneracy there. But my > > calculation converged to 4 frequency points, which means some branches > broke > > the symmetry. And when I did the DOS calculation, as I mentioned, > phonopy's > > analysis gives 2000+ irreducible k points for a 30by30by30 meshes, but > for > > the same mesh, phonopy's example gives 800 k points, which means that my > > setup decreased the order of symmetry, this could not happen for a > > calculation of any accuracy, if I understand right. > > > > But I could not figure out my problem, since it looks a simple case for > > phonopy,and I was following the procedure from phonopy manual. > > > > Best, > > Tian > > > > > > On Mon, Sep 16, 2013 at 3:03 PM, Jan Jaeken <Jan.Jaeken@...> wrote: > >> > >> Hello, > >> > >> Those settings you have been using aren't stringent enough. Especially > >> the number of kpoints you have been using is much too low. First, try > >> using a larger kpoint grid. You should compare the forces of your > >> current calculation against those of the calculation at a higher > >> setting. They shouldn't deviate by more than 1 meV/A. > >> > >> Furthermore, a 2x2x2 supercell might be too small. The forces farthest > >> from the displaced atom (check periodic boundary conditions) should be > >> a thousand times smaller than those on the displaced atom itself. > >> Personally, I wouldn't trust anything smaller than a 3x3x3 supercell, > >> as periodic images will almost certainly interfere with each other in > >> a 2x2x2. > >> > >> Also, make sure to set ENCUT to at least 130% of the POTCAR's ENMAX, > >> and LREAL = .False. always. Check that you are using a proper ISMEAR > >> and SIGMA as well (see the VASP manual). > >> > >> These force calculations are highly sensitive, and it is important to > >> run strict convergence tests. Problems in the phonon band structure > >> often stem from poorlyconverged forces. > >> > >> Regards, > >> > >> Jan Jaeken > >> > >> > >> Quoting Tian Lan <tianlan@...>: > >> > >> > Dear all, > >> > > >> > I plan to run a MD simulation for diamond cubic Si. Before this, I did > >> > selfconsistent calculation and used phonopy to get phonon dispersion > >> > and > >> > DOS to check the correctness of my VASP setup. I thought it would be > an > >> > easy task, however, I found some strange problems especially for > >> > dispersion. There are no big deviation(like frequency)from its > >> > expectation > >> > values, but some obvious errors. > >> > > >> > Specifically, at X (0.5,0,0) point, the modes should converge to 3 > >> > points, > >> > as I found from any reference. But my calculation gave 4 branches, it > >> > looks > >> > like some symmetry was broken. And gamma>R(0.5, 0.5,0), two TA > branches > >> > should not be degenerate, but my calculation gave a degenerate curves. > >> > > >> > The small displacement method is used. I used a 2 by 2 by 2 supercell, > >> > with the unit cell 8 atoms; 4 by 4 by 4 kpoint mesh; the unit cell > >> > POSCAR > >> > like this > >> >  > >> > Diamond cubic Si > >> > 5.465 > >> > 1.0 0.0 0.0 > >> > 0.0 1.0 0.0 > >> > 0.0 0.0 1.0 > >> > Si > >> > 8 > >> > Direct > >> > 0.8750000000000000 0.8750000000000000 0.8750000000000000 > >> > 0.8750000000000000 0.3750000000000000 0.3750000000000000 > >> > 0.3750000000000000 0.8750000000000000 0.3750000000000000 > >> > 0.3750000000000000 0.3750000000000000 0.8750000000000000 > >> > 0.1250000000000000 0.1250000000000000 0.1250000000000000 > >> > 0.1250000000000000 0.6250000000000000 0.6250000000000000 > >> > 0.6250000000000000 0.1250000000000000 0.6250000000000000 > >> > 0.6250000000000000 0.6250000000000000 0.1250000000000000 > >> > > >> >  > >> > > >> > My INCAR is like this, I don't think there is any obvious problem > >> > compared > >> > with, for example, phonopy example for Si > >> >  > >> > SYSTEM = Si Relax > >> > ISTART = 0 > >> > ICHARG = 2 > >> > ENCUT = 245.345 > >> > PREC = Accurate > >> > NELMIN = 4 # do a minimum of four electronic steps > >> > NELM = 40 # maximum of 40 electronic steps > >> > EDIFF = 1E8 > >> > ADDGRID = .TRUE. > >> > IBRION = 1 > >> > ISMEAR = 0 > >> > SIGMA = 0.05 > >> > LREAL = Auto > >> > LWAVE = .False. > >> > LPLANE = .TRUE. #recommended for Linux clusters > >> > NPAR = 8 # twice the number of nodes > >> >  > >> > > >> > Then I followed the standard steps for phonopy. > >> > > >> > (1)Get the Force sets > >> > phonopy f disp001/vasprun.xml > >> > (2) calculate DOS or band, in here, convert to primitive axes, the > >> > band.conf is as follows: > >> >  > >> > ATOM_NAME = Si > >> > DIM = 2 2 2 > >> > PRIMITIVE_AXIS = 0 1/2 1/2 1/2 0 1/2 1/2 1/2 0 > >> > BAND = 0 0 0 0 0 1/2 > >> > #BAND = 0 0 0 1/2 0 0, 0 1/2 1/2 0 0 0, 0 0 0 1/2 1/2 1/2 > >> > BAND_POINTS = 101 > >> >  > >> > And a problem really concerned me is, the phonopy analysis showed that > >> > the > >> > irreducible k points after DOS calculation(mesh.conf) are 2000 more > >> > points, > >> > while in phonopy example, there is only about 800, the symmetry > analysis > >> > is > >> > like this > >> >  > >> > Mesh sampling mode > >> > Settings: > >> > Sampling mesh: [30 30 30] > >> > Supercell: [2 2 2] > >> > Primitive axis: > >> > [ 0. 0.5 0.5] > >> > [ 0.5 0. 0.5] > >> > [ 0.5 0.5 0. ] > >> > Spacegroup: Fd3m (227) > >> > Calculating force constants... > >> > Number of irreducible qpoints: 2480 > >> >  > >> > > >> > I don't know why there is such a strange problem, I can not find any > >> > evident issue related to my calculation. I hope I provide enough > >> > information to evaluate it. Thanks. > >> > > >> > > >> >  > >> > Lan, Tian > >> > Ph.D. Candidate, Department of Applied Physics and Materials Science > >> > California Institute of Technology, > >> > Caltech M/C 13878, Pasadena, CA, 91125 > >> > >> > >> > >> > >> > >> >  > >> LIMITED TIME SALE  Full Year of Microsoft Training For Just $49.99! > >> 1,500+ hours of tutorials including VisualStudio 2012, Windows 8, > >> SharePoint > >> 2013, SQL 2012, MVC 4, more. BEST VALUE: New MultiLibrary Power Pack > >> includes > >> Mobile, Cloud, Java, and UX Design. Lowest price ever! Ends 9/20/13. > >> > >> > http://pubads.g.doubleclick.net/gampad/clk?id=58041151&iu=/4140/ostg.clktrk > >> _______________________________________________ > >> Phonopyusers mailing list > >> Phonopyusers@... > >> https://lists.sourceforge.net/lists/listinfo/phonopyusers > > > > > > > > > >  > > Lan, Tian > > Ph.D. Candidate, Department of Applied Physics and Materials Science > > California Institute of Technology, > > Caltech M/C 13878, Pasadena, CA, 91125 > > > > > > > >  > > Lan, Tian > > Ph.D. Candidate, Department of Applied Physics and Materials Science > > California Institute of Technology, > > Caltech M/C 13878, Pasadena, CA, 91125 > > > > >  > > LIMITED TIME SALE  Full Year of Microsoft Training For Just $49.99! > > 1,500+ hours of tutorials including VisualStudio 2012, Windows 8, > SharePoint > > 2013, SQL 2012, MVC 4, more. BEST VALUE: New MultiLibrary Power Pack > > includes > > Mobile, Cloud, Java, and UX Design. Lowest price ever! Ends 9/20/13. > > > http://pubads.g.doubleclick.net/gampad/clk?id=58041151&iu=/4140/ostg.clktrk > > _______________________________________________ > > Phonopyusers mailing list > > Phonopyusers@... > > https://lists.sourceforge.net/lists/listinfo/phonopyusers > > > > > >  > Atsushi Togo > http://atztogo.github.com/ > atz.togo@... >  Lan, Tian Ph.D. Candidate, Department of Applied Physics and Materials Science California Institute of Technology, Caltech M/C 13878, Pasadena, CA, 91125 
From: Atsushi Togo <atz.togo@gm...>  20130917 03:06:55

Hi, Phonopy has its own symmetry handler and usually the topology of band structure is less sensitive to the VASP setting. But as Jan recommended, the force calculation should be done as high accuracy as possible. For example, I recommend always LREAL = .FALSE. In phonopy, qpoint has to be set in primitive basis of reciprocal space, i.e., I guess X point you mean would be (0.5, 0, 0.5). No. 227 is searched at http://www.cryst.ehu.es/cryst/get_kvec.html . In Bilbao crystallographic server convention, no name of R exists for this space group. The oddity of mesh sampling arises by three fold rotation of symmetry. If the center of rotation is not sampled, the sampling becomes notuniform. In the example, the odd numbers give to sample Gammapoint, but the even numbers mesh doesn't. Togo On Tue, Sep 17, 2013 at 7:29 AM, Tian Lan <tianlan@...> wrote: > > >  Forwarded message  > From: Tian Lan <tianlan@...> > Date: Mon, Sep 16, 2013 at 3:22 PM > Subject: Re: [Phonopyusers] Phonon dispersion error in Si > To: Jan Jaeken <Jan.Jaeken@...> > > > Hi Jan, > > Thank you. You are right, I think it is not stringent enough. But I think it > is basically okay because it has been converged. 4 by 4 by 4 kmesh for a 2 > by 2 by 2 supercell, is just a 8 by 8 by 8 mesh for a unit cell. I think for > a phonon calculation, this is usually sufficient. My experience is, this may > influence the frequencies a little bit, but would not destroy any symmetry > relation, which is my major concern here. > > As I said, for example, at X point, phonon branches should be converged to 3 > frequency points, i.e., there is certain degeneracy there. But my > calculation converged to 4 frequency points, which means some branches broke > the symmetry. And when I did the DOS calculation, as I mentioned, phonopy's > analysis gives 2000+ irreducible k points for a 30by30by30 meshes, but for > the same mesh, phonopy's example gives 800 k points, which means that my > setup decreased the order of symmetry, this could not happen for a > calculation of any accuracy, if I understand right. > > But I could not figure out my problem, since it looks a simple case for > phonopy,and I was following the procedure from phonopy manual. > > Best, > Tian > > > On Mon, Sep 16, 2013 at 3:03 PM, Jan Jaeken <Jan.Jaeken@...> wrote: >> >> Hello, >> >> Those settings you have been using aren't stringent enough. Especially >> the number of kpoints you have been using is much too low. First, try >> using a larger kpoint grid. You should compare the forces of your >> current calculation against those of the calculation at a higher >> setting. They shouldn't deviate by more than 1 meV/A. >> >> Furthermore, a 2x2x2 supercell might be too small. The forces farthest >> from the displaced atom (check periodic boundary conditions) should be >> a thousand times smaller than those on the displaced atom itself. >> Personally, I wouldn't trust anything smaller than a 3x3x3 supercell, >> as periodic images will almost certainly interfere with each other in >> a 2x2x2. >> >> Also, make sure to set ENCUT to at least 130% of the POTCAR's ENMAX, >> and LREAL = .False. always. Check that you are using a proper ISMEAR >> and SIGMA as well (see the VASP manual). >> >> These force calculations are highly sensitive, and it is important to >> run strict convergence tests. Problems in the phonon band structure >> often stem from poorlyconverged forces. >> >> Regards, >> >> Jan Jaeken >> >> >> Quoting Tian Lan <tianlan@...>: >> >> > Dear all, >> > >> > I plan to run a MD simulation for diamond cubic Si. Before this, I did >> > selfconsistent calculation and used phonopy to get phonon dispersion >> > and >> > DOS to check the correctness of my VASP setup. I thought it would be an >> > easy task, however, I found some strange problems especially for >> > dispersion. There are no big deviation(like frequency)from its >> > expectation >> > values, but some obvious errors. >> > >> > Specifically, at X (0.5,0,0) point, the modes should converge to 3 >> > points, >> > as I found from any reference. But my calculation gave 4 branches, it >> > looks >> > like some symmetry was broken. And gamma>R(0.5, 0.5,0), two TA branches >> > should not be degenerate, but my calculation gave a degenerate curves. >> > >> > The small displacement method is used. I used a 2 by 2 by 2 supercell, >> > with the unit cell 8 atoms; 4 by 4 by 4 kpoint mesh; the unit cell >> > POSCAR >> > like this >> >  >> > Diamond cubic Si >> > 5.465 >> > 1.0 0.0 0.0 >> > 0.0 1.0 0.0 >> > 0.0 0.0 1.0 >> > Si >> > 8 >> > Direct >> > 0.8750000000000000 0.8750000000000000 0.8750000000000000 >> > 0.8750000000000000 0.3750000000000000 0.3750000000000000 >> > 0.3750000000000000 0.8750000000000000 0.3750000000000000 >> > 0.3750000000000000 0.3750000000000000 0.8750000000000000 >> > 0.1250000000000000 0.1250000000000000 0.1250000000000000 >> > 0.1250000000000000 0.6250000000000000 0.6250000000000000 >> > 0.6250000000000000 0.1250000000000000 0.6250000000000000 >> > 0.6250000000000000 0.6250000000000000 0.1250000000000000 >> > >> >  >> > >> > My INCAR is like this, I don't think there is any obvious problem >> > compared >> > with, for example, phonopy example for Si >> >  >> > SYSTEM = Si Relax >> > ISTART = 0 >> > ICHARG = 2 >> > ENCUT = 245.345 >> > PREC = Accurate >> > NELMIN = 4 # do a minimum of four electronic steps >> > NELM = 40 # maximum of 40 electronic steps >> > EDIFF = 1E8 >> > ADDGRID = .TRUE. >> > IBRION = 1 >> > ISMEAR = 0 >> > SIGMA = 0.05 >> > LREAL = Auto >> > LWAVE = .False. >> > LPLANE = .TRUE. #recommended for Linux clusters >> > NPAR = 8 # twice the number of nodes >> >  >> > >> > Then I followed the standard steps for phonopy. >> > >> > (1)Get the Force sets >> > phonopy f disp001/vasprun.xml >> > (2) calculate DOS or band, in here, convert to primitive axes, the >> > band.conf is as follows: >> >  >> > ATOM_NAME = Si >> > DIM = 2 2 2 >> > PRIMITIVE_AXIS = 0 1/2 1/2 1/2 0 1/2 1/2 1/2 0 >> > BAND = 0 0 0 0 0 1/2 >> > #BAND = 0 0 0 1/2 0 0, 0 1/2 1/2 0 0 0, 0 0 0 1/2 1/2 1/2 >> > BAND_POINTS = 101 >> >  >> > And a problem really concerned me is, the phonopy analysis showed that >> > the >> > irreducible k points after DOS calculation(mesh.conf) are 2000 more >> > points, >> > while in phonopy example, there is only about 800, the symmetry analysis >> > is >> > like this >> >  >> > Mesh sampling mode >> > Settings: >> > Sampling mesh: [30 30 30] >> > Supercell: [2 2 2] >> > Primitive axis: >> > [ 0. 0.5 0.5] >> > [ 0.5 0. 0.5] >> > [ 0.5 0.5 0. ] >> > Spacegroup: Fd3m (227) >> > Calculating force constants... >> > Number of irreducible qpoints: 2480 >> >  >> > >> > I don't know why there is such a strange problem, I can not find any >> > evident issue related to my calculation. I hope I provide enough >> > information to evaluate it. Thanks. >> > >> > >> >  >> > Lan, Tian >> > Ph.D. Candidate, Department of Applied Physics and Materials Science >> > California Institute of Technology, >> > Caltech M/C 13878, Pasadena, CA, 91125 >> >> >> >> >> >>  >> LIMITED TIME SALE  Full Year of Microsoft Training For Just $49.99! >> 1,500+ hours of tutorials including VisualStudio 2012, Windows 8, >> SharePoint >> 2013, SQL 2012, MVC 4, more. BEST VALUE: New MultiLibrary Power Pack >> includes >> Mobile, Cloud, Java, and UX Design. Lowest price ever! Ends 9/20/13. >> >> http://pubads.g.doubleclick.net/gampad/clk?id=58041151&iu=/4140/ostg.clktrk >> _______________________________________________ >> Phonopyusers mailing list >> Phonopyusers@... >> https://lists.sourceforge.net/lists/listinfo/phonopyusers > > > > >  > Lan, Tian > Ph.D. Candidate, Department of Applied Physics and Materials Science > California Institute of Technology, > Caltech M/C 13878, Pasadena, CA, 91125 > > > >  > Lan, Tian > Ph.D. Candidate, Department of Applied Physics and Materials Science > California Institute of Technology, > Caltech M/C 13878, Pasadena, CA, 91125 > >  > LIMITED TIME SALE  Full Year of Microsoft Training For Just $49.99! > 1,500+ hours of tutorials including VisualStudio 2012, Windows 8, SharePoint > 2013, SQL 2012, MVC 4, more. BEST VALUE: New MultiLibrary Power Pack > includes > Mobile, Cloud, Java, and UX Design. Lowest price ever! Ends 9/20/13. > http://pubads.g.doubleclick.net/gampad/clk?id=58041151&iu=/4140/ostg.clktrk > _______________________________________________ > Phonopyusers mailing list > Phonopyusers@... > https://lists.sourceforge.net/lists/listinfo/phonopyusers >  Atsushi Togo http://atztogo.github.com/ atz.togo@... 
From: Tian Lan <tianlan@ca...>  20130916 22:29:15

 Forwarded message  From: Tian Lan <tianlan@...> Date: Mon, Sep 16, 2013 at 3:22 PM Subject: Re: [Phonopyusers] Phonon dispersion error in Si To: Jan Jaeken <Jan.Jaeken@...> Hi Jan, Thank you. You are right, I think it is not stringent enough. But I think it is basically okay because it has been converged. 4 by 4 by 4 kmesh for a 2 by 2 by 2 supercell, is just a 8 by 8 by 8 mesh for a unit cell. I think for a phonon calculation, this is usually sufficient. My experience is, this may influence the frequencies a little bit, but would not destroy any symmetry relation, which is my major concern here. As I said, for example, at X point, phonon branches should be converged to 3 frequency points, i.e., there is certain degeneracy there. But my calculation converged to 4 frequency points, which means some branches broke the symmetry. And when I did the DOS calculation, as I mentioned, phonopy's analysis gives 2000+ irreducible k points for a 30by30by30 meshes, but for the same mesh, phonopy's example gives 800 k points, which means that my setup decreased the order of symmetry, this could not happen for a calculation of any accuracy, if I understand right. But I could not figure out my problem, since it looks a simple case for phonopy,and I was following the procedure from phonopy manual. Best, Tian On Mon, Sep 16, 2013 at 3:03 PM, Jan Jaeken <Jan.Jaeken@...> wrote: > Hello, > > Those settings you have been using aren't stringent enough. Especially > the number of kpoints you have been using is much too low. First, try > using a larger kpoint grid. You should compare the forces of your > current calculation against those of the calculation at a higher > setting. They shouldn't deviate by more than 1 meV/A. > > Furthermore, a 2x2x2 supercell might be too small. The forces farthest > from the displaced atom (check periodic boundary conditions) should be > a thousand times smaller than those on the displaced atom itself. > Personally, I wouldn't trust anything smaller than a 3x3x3 supercell, > as periodic images will almost certainly interfere with each other in > a 2x2x2. > > Also, make sure to set ENCUT to at least 130% of the POTCAR's ENMAX, > and LREAL = .False. always. Check that you are using a proper ISMEAR > and SIGMA as well (see the VASP manual). > > These force calculations are highly sensitive, and it is important to > run strict convergence tests. Problems in the phonon band structure > often stem from poorlyconverged forces. > > Regards, > > Jan Jaeken > > > Quoting Tian Lan <tianlan@...>: > > > Dear all, > > > > I plan to run a MD simulation for diamond cubic Si. Before this, I did > > selfconsistent calculation and used phonopy to get phonon dispersion and > > DOS to check the correctness of my VASP setup. I thought it would be an > > easy task, however, I found some strange problems especially for > > dispersion. There are no big deviation(like frequency)from its > expectation > > values, but some obvious errors. > > > > Specifically, at X (0.5,0,0) point, the modes should converge to 3 > points, > > as I found from any reference. But my calculation gave 4 branches, it > looks > > like some symmetry was broken. And gamma>R(0.5, 0.5,0), two TA branches > > should not be degenerate, but my calculation gave a degenerate curves. > > > > The small displacement method is used. I used a 2 by 2 by 2 supercell, > > with the unit cell 8 atoms; 4 by 4 by 4 kpoint mesh; the unit cell > POSCAR > > like this > >  > > Diamond cubic Si > > 5.465 > > 1.0 0.0 0.0 > > 0.0 1.0 0.0 > > 0.0 0.0 1.0 > > Si > > 8 > > Direct > > 0.8750000000000000 0.8750000000000000 0.8750000000000000 > > 0.8750000000000000 0.3750000000000000 0.3750000000000000 > > 0.3750000000000000 0.8750000000000000 0.3750000000000000 > > 0.3750000000000000 0.3750000000000000 0.8750000000000000 > > 0.1250000000000000 0.1250000000000000 0.1250000000000000 > > 0.1250000000000000 0.6250000000000000 0.6250000000000000 > > 0.6250000000000000 0.1250000000000000 0.6250000000000000 > > 0.6250000000000000 0.6250000000000000 0.1250000000000000 > > > >  > > > > My INCAR is like this, I don't think there is any obvious problem > compared > > with, for example, phonopy example for Si > >  > > SYSTEM = Si Relax > > ISTART = 0 > > ICHARG = 2 > > ENCUT = 245.345 > > PREC = Accurate > > NELMIN = 4 # do a minimum of four electronic steps > > NELM = 40 # maximum of 40 electronic steps > > EDIFF = 1E8 > > ADDGRID = .TRUE. > > IBRION = 1 > > ISMEAR = 0 > > SIGMA = 0.05 > > LREAL = Auto > > LWAVE = .False. > > LPLANE = .TRUE. #recommended for Linux clusters > > NPAR = 8 # twice the number of nodes > >  > > > > Then I followed the standard steps for phonopy. > > > > (1)Get the Force sets > > phonopy f disp001/vasprun.xml > > (2) calculate DOS or band, in here, convert to primitive axes, the > > band.conf is as follows: > >  > > ATOM_NAME = Si > > DIM = 2 2 2 > > PRIMITIVE_AXIS = 0 1/2 1/2 1/2 0 1/2 1/2 1/2 0 > > BAND = 0 0 0 0 0 1/2 > > #BAND = 0 0 0 1/2 0 0, 0 1/2 1/2 0 0 0, 0 0 0 1/2 1/2 1/2 > > BAND_POINTS = 101 > >  > > And a problem really concerned me is, the phonopy analysis showed that > the > > irreducible k points after DOS calculation(mesh.conf) are 2000 more > points, > > while in phonopy example, there is only about 800, the symmetry analysis > is > > like this > >  > > Mesh sampling mode > > Settings: > > Sampling mesh: [30 30 30] > > Supercell: [2 2 2] > > Primitive axis: > > [ 0. 0.5 0.5] > > [ 0.5 0. 0.5] > > [ 0.5 0.5 0. ] > > Spacegroup: Fd3m (227) > > Calculating force constants... > > Number of irreducible qpoints: 2480 > >  > > > > I don't know why there is such a strange problem, I can not find any > > evident issue related to my calculation. I hope I provide enough > > information to evaluate it. Thanks. > > > > > >  > > Lan, Tian > > Ph.D. Candidate, Department of Applied Physics and Materials Science > > California Institute of Technology, > > Caltech M/C 13878, Pasadena, CA, 91125 > > > > > >  > LIMITED TIME SALE  Full Year of Microsoft Training For Just $49.99! > 1,500+ hours of tutorials including VisualStudio 2012, Windows 8, > SharePoint > 2013, SQL 2012, MVC 4, more. BEST VALUE: New MultiLibrary Power Pack > includes > Mobile, Cloud, Java, and UX Design. Lowest price ever! Ends 9/20/13. > http://pubads.g.doubleclick.net/gampad/clk?id=58041151&iu=/4140/ostg.clktrk > _______________________________________________ > Phonopyusers mailing list > Phonopyusers@... > https://lists.sourceforge.net/lists/listinfo/phonopyusers >  Lan, Tian Ph.D. Candidate, Department of Applied Physics and Materials Science California Institute of Technology, Caltech M/C 13878, Pasadena, CA, 91125  Lan, Tian Ph.D. Candidate, Department of Applied Physics and Materials Science California Institute of Technology, Caltech M/C 13878, Pasadena, CA, 91125 
From: Jan Jaeken <Jan.Jaeken@UGent.be>  20130916 22:03:49

Hello, Those settings you have been using aren't stringent enough. Especially the number of kpoints you have been using is much too low. First, try using a larger kpoint grid. You should compare the forces of your current calculation against those of the calculation at a higher setting. They shouldn't deviate by more than 1 meV/A. Furthermore, a 2x2x2 supercell might be too small. The forces farthest from the displaced atom (check periodic boundary conditions) should be a thousand times smaller than those on the displaced atom itself. Personally, I wouldn't trust anything smaller than a 3x3x3 supercell, as periodic images will almost certainly interfere with each other in a 2x2x2. Also, make sure to set ENCUT to at least 130% of the POTCAR's ENMAX, and LREAL = .False. always. Check that you are using a proper ISMEAR and SIGMA as well (see the VASP manual). These force calculations are highly sensitive, and it is important to run strict convergence tests. Problems in the phonon band structure often stem from poorlyconverged forces. Regards, Jan Jaeken Quoting Tian Lan <tianlan@...>: > Dear all, > > I plan to run a MD simulation for diamond cubic Si. Before this, I did > selfconsistent calculation and used phonopy to get phonon dispersion and > DOS to check the correctness of my VASP setup. I thought it would be an > easy task, however, I found some strange problems especially for > dispersion. There are no big deviation(like frequency)from its expectation > values, but some obvious errors. > > Specifically, at X (0.5,0,0) point, the modes should converge to 3 points, > as I found from any reference. But my calculation gave 4 branches, it looks > like some symmetry was broken. And gamma>R(0.5, 0.5,0), two TA branches > should not be degenerate, but my calculation gave a degenerate curves. > > The small displacement method is used. I used a 2 by 2 by 2 supercell, > with the unit cell 8 atoms; 4 by 4 by 4 kpoint mesh; the unit cell POSCAR > like this >  > Diamond cubic Si > 5.465 > 1.0 0.0 0.0 > 0.0 1.0 0.0 > 0.0 0.0 1.0 > Si > 8 > Direct > 0.8750000000000000 0.8750000000000000 0.8750000000000000 > 0.8750000000000000 0.3750000000000000 0.3750000000000000 > 0.3750000000000000 0.8750000000000000 0.3750000000000000 > 0.3750000000000000 0.3750000000000000 0.8750000000000000 > 0.1250000000000000 0.1250000000000000 0.1250000000000000 > 0.1250000000000000 0.6250000000000000 0.6250000000000000 > 0.6250000000000000 0.1250000000000000 0.6250000000000000 > 0.6250000000000000 0.6250000000000000 0.1250000000000000 > >  > > My INCAR is like this, I don't think there is any obvious problem compared > with, for example, phonopy example for Si >  > SYSTEM = Si Relax > ISTART = 0 > ICHARG = 2 > ENCUT = 245.345 > PREC = Accurate > NELMIN = 4 # do a minimum of four electronic steps > NELM = 40 # maximum of 40 electronic steps > EDIFF = 1E8 > ADDGRID = .TRUE. > IBRION = 1 > ISMEAR = 0 > SIGMA = 0.05 > LREAL = Auto > LWAVE = .False. > LPLANE = .TRUE. #recommended for Linux clusters > NPAR = 8 # twice the number of nodes >  > > Then I followed the standard steps for phonopy. > > (1)Get the Force sets > phonopy f disp001/vasprun.xml > (2) calculate DOS or band, in here, convert to primitive axes, the > band.conf is as follows: >  > ATOM_NAME = Si > DIM = 2 2 2 > PRIMITIVE_AXIS = 0 1/2 1/2 1/2 0 1/2 1/2 1/2 0 > BAND = 0 0 0 0 0 1/2 > #BAND = 0 0 0 1/2 0 0, 0 1/2 1/2 0 0 0, 0 0 0 1/2 1/2 1/2 > BAND_POINTS = 101 >  > And a problem really concerned me is, the phonopy analysis showed that the > irreducible k points after DOS calculation(mesh.conf) are 2000 more points, > while in phonopy example, there is only about 800, the symmetry analysis is > like this >  > Mesh sampling mode > Settings: > Sampling mesh: [30 30 30] > Supercell: [2 2 2] > Primitive axis: > [ 0. 0.5 0.5] > [ 0.5 0. 0.5] > [ 0.5 0.5 0. ] > Spacegroup: Fd3m (227) > Calculating force constants... > Number of irreducible qpoints: 2480 >  > > I don't know why there is such a strange problem, I can not find any > evident issue related to my calculation. I hope I provide enough > information to evaluate it. Thanks. > > >  > Lan, Tian > Ph.D. Candidate, Department of Applied Physics and Materials Science > California Institute of Technology, > Caltech M/C 13878, Pasadena, CA, 91125 
From: Tian Lan <tianlan@ca...>  20130916 19:38:10

Dear all, I plan to run a MD simulation for diamond cubic Si. Before this, I did selfconsistent calculation and used phonopy to get phonon dispersion and DOS to check the correctness of my VASP setup. I thought it would be an easy task, however, I found some strange problems especially for dispersion. There are no big deviation(like frequency)from its expectation values, but some obvious errors. Specifically, at X (0.5,0,0) point, the modes should converge to 3 points, as I found from any reference. But my calculation gave 4 branches, it looks like some symmetry was broken. And gamma>R(0.5, 0.5,0), two TA branches should not be degenerate, but my calculation gave a degenerate curves. The small displacement method is used. I used a 2 by 2 by 2 supercell, with the unit cell 8 atoms; 4 by 4 by 4 kpoint mesh; the unit cell POSCAR like this  Diamond cubic Si 5.465 1.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 1.0 Si 8 Direct 0.8750000000000000 0.8750000000000000 0.8750000000000000 0.8750000000000000 0.3750000000000000 0.3750000000000000 0.3750000000000000 0.8750000000000000 0.3750000000000000 0.3750000000000000 0.3750000000000000 0.8750000000000000 0.1250000000000000 0.1250000000000000 0.1250000000000000 0.1250000000000000 0.6250000000000000 0.6250000000000000 0.6250000000000000 0.1250000000000000 0.6250000000000000 0.6250000000000000 0.6250000000000000 0.1250000000000000  My INCAR is like this, I don't think there is any obvious problem compared with, for example, phonopy example for Si  SYSTEM = Si Relax ISTART = 0 ICHARG = 2 ENCUT = 245.345 PREC = Accurate NELMIN = 4 # do a minimum of four electronic steps NELM = 40 # maximum of 40 electronic steps EDIFF = 1E8 ADDGRID = .TRUE. IBRION = 1 ISMEAR = 0 SIGMA = 0.05 LREAL = Auto LWAVE = .False. LPLANE = .TRUE. #recommended for Linux clusters NPAR = 8 # twice the number of nodes  Then I followed the standard steps for phonopy. (1)Get the Force sets phonopy f disp001/vasprun.xml (2) calculate DOS or band, in here, convert to primitive axes, the band.conf is as follows:  ATOM_NAME = Si DIM = 2 2 2 PRIMITIVE_AXIS = 0 1/2 1/2 1/2 0 1/2 1/2 1/2 0 BAND = 0 0 0 0 0 1/2 #BAND = 0 0 0 1/2 0 0, 0 1/2 1/2 0 0 0, 0 0 0 1/2 1/2 1/2 BAND_POINTS = 101  And a problem really concerned me is, the phonopy analysis showed that the irreducible k points after DOS calculation(mesh.conf) are 2000 more points, while in phonopy example, there is only about 800, the symmetry analysis is like this  Mesh sampling mode Settings: Sampling mesh: [30 30 30] Supercell: [2 2 2] Primitive axis: [ 0. 0.5 0.5] [ 0.5 0. 0.5] [ 0.5 0.5 0. ] Spacegroup: Fd3m (227) Calculating force constants... Number of irreducible qpoints: 2480  I don't know why there is such a strange problem, I can not find any evident issue related to my calculation. I hope I provide enough information to evaluate it. Thanks.  Lan, Tian Ph.D. Candidate, Department of Applied Physics and Materials Science California Institute of Technology, Caltech M/C 13878, Pasadena, CA, 91125 
From: Atsushi Togo <atz.togo@gm...>  20130906 01:52:00

Hi, VaspToTHz = 15.633302. THzToEv = 0.00413566733. So the factor you set for factor option is, I think, VaspToTHz * THzToEv * 1000. If you just want to convert phonopy result given in THz, it's THzToEv * 1000 as you wrote. Togo On Fri, Sep 6, 2013 at 10:32 AM, Aldo Humberto Romero <aromero@...> wrote: > I am a bit confused on how to change the output units in Phonopy. For Vasp > calculations > the default is THz, does it means that if I want the output in meV, should > I use a factor > that converts from THz to meV?) (=27.21138*1000/6579.68)? or does it > corresponds > to the atomic units to meV?... it should be simple, I can do a calculation > for a known spectra > but if possible, I would like to avoid that. > > Thanks in advance > > >  > > Prof. Aldo Humberto Romero > CINVESTAVUnidad Queretaro > Libramiento Norponiente 2000 > CP 76230, Queretaro, QRO, Mexico > tel: 442 211 9909 > fax: 442 211 9938 > > email: aromero@... > aldorome@... > www: qro.cinvestav.mx/~aromero > >  > Learn the latestVisual Studio 2012, SharePoint 2013, SQL 2012, more! > Discover the easy way to master current and previous Microsoft technologies > and advance your career. Get an incredible 1,500+ hours of stepbystep > tutorial videos with LearnDevNow. Subscribe today and save! > http://pubads.g.doubleclick.net/gampad/clk?id=58041391&iu=/4140/ostg.clktrk > _______________________________________________ > Phonopyusers mailing list > Phonopyusers@... > https://lists.sourceforge.net/lists/listinfo/phonopyusers  Atsushi Togo http://atztogo.github.com/ atz.togo@... 
From: Aldo Humberto Romero <aromero@qr...>  20130906 01:36:09

I am a bit confused on how to change the output units in Phonopy. For Vasp calculations the default is THz, does it means that if I want the output in meV, should I use a factor that converts from THz to meV?) (=27.21138*1000/6579.68)? or does it corresponds to the atomic units to meV?... it should be simple, I can do a calculation for a known spectra but if possible, I would like to avoid that. Thanks in advance  Prof. Aldo Humberto Romero CINVESTAVUnidad Queretaro Libramiento Norponiente 2000 CP 76230, Queretaro, QRO, Mexico tel: 442 211 9909 fax: 442 211 9938 email: aromero@... aldorome@... www: qro.cinvestav.mx/~aromero 
From: Atsushi Togo <atz.togo@gm...>  20130903 07:26:24

And just as an example, I attach a graphene band structure that I calculated just now. As we can see, the acoustic modes are very soft, in this case the structure optimization is better to be done by the supercell though I didn't in this calculation. Then, I don't (can't) describe all about my VASP settings (please don't ask me about VASP), but some are shown as follows: For the structure optimization of theunit cell, INCAR: PREC = Accurate IBRION = 2 NSW = 20 NELMIN = 5 ISIF = 3 ENCUT = 500 EDIFF = 1.000000e08 EDIFFG = 1.000000e08 ISMEAR = 1 SIGMA = 2.000000e01 IALGO = 38 LREAL = .FALSE. ADDGRID = .TRUE. LWAVE = .FALSE. LCHARG = .FALSE. KPOINTS: Automatic mesh 0 Monkhorstpack 12 12 1 0.500 0.500 0.000 For force calculations of supercells, INCAR: PREC = Accurate IBRION = 1 NELMIN = 5 ENCUT = 500 EDIFF = 1.000000e08 ISMEAR = 1 SIGMA = 2.000000e01 IALGO = 38 LREAL = .FALSE. ADDGRID = .TRUE. LWAVE = .FALSE. LCHARG = .FALSE. NPAR = 4 KPOINTS: Automatic mesh 0 Monkhorstpack 2 2 1 0.500 0.500 0.000 phonopy: % phonopy dim="6 6 1" band="0 1/2 0 0 0 0 1/3 2/3 0 0 1/2 0" p c POSCARunitcell Togo On Tue, Sep 3, 2013 at 2:04 PM, Atsushi Togo <atz.togo@...> wrote: > Hi, > > Now I think your problem is not in phonopy but VASP. From vasp web site, > > For the calculation of the total energy in bulk materials we recommend > the tetrahedron method with Blöchl corrections (ISMEAR=5). This > method also gives a smooth nice electronic density of states (DOS). > The only drawback is that the methods is not variational with respect > to the partial occupancies, therefore the calculated forces might be > wrong by a few percent for metals. For the calculation of phonon > frequencies based on forces we recommend the method of > MethfesselPaxton (ISMEAR>0). For semiconductors and insulators the > forces are always correct, because partial occupancies do not vary! > > So you can read VASP manual carefully before asking VASP questions to > the phonopy mailing list or you can ask them at the VASP forum. > > Togo > > On Tue, Sep 3, 2013 at 12:02 PM, Yang Ming <phyym@...> wrote: >> Dear Prof. Togo, >> >> Thank you so much for the kind reply. >> >> I just checked the KPOINTS and the INCAR setting, which are correct for the electronic structure (see the band structure attached). >> >> Actually, I also increased denser kpoint meshes (18×18×1), and also changed the INCAR setting (ENCUT =500, ISMEAR=5, EDIFF=1.0E8, and ADDGRID = .True) , but still got negative frequencies. >> >> Let me show you my calculation processes again below, and please help me to check the problems and let me know if the settings are incorrect: >> >> 1. Generate a 6×6×1 supercell from below bulk by using the commad: phonopy d dim="6 6 1". >> >> 2. Calculate the force using the generated structure POSCAR001 and standard VASP setting. >> >> 3. Create the FORCE_SETS by using: phonopy f disp001/vasprun.xml >> >> 4 Phonon band structure was calculated by using the band.conf like below: >> >> ATOM_NAME = C >> DIM = 6 6 1 >> BAND = 0.0 0.5 0.0 0.0 0.0 0.0 0.3333 0.6667 0.0 0.0 0.5 0.0 >> >> 5. Plot the band structure by using: phonopy p band.conf >> >> >> Thanks a lot. >> Best regards, >> Yang Ming >> ****** >> Computational CondensedMatter Physics Lab >> Department of Physics >> National University of Singapore >> Singapore 117542 >> Email: phyym@...; ming_young@... >> Tel: +65 6516 4335 (office) >> >> >> Original Message >> From: Yang Ming >> Sent: Sunday, 1 September, 2013 9:22 PM >> To: 'Atsushi Togo' >> Subject: RE: [Phonopyusers] Phonon dispersion of graphene >> >> Dear Dr. Togo, >> >> Thanks a lot for the prompt reply. >> >> Please find the attached data and calculated eigenvalue. >> >> Thanks. >> >> Best regards, >> Yang Ming >> ****** >> Computational CondensedMatter Physics Lab Department of Physics National University of Singapore Singapore 117542 >> Email: phyym@...; ming_young@... >> Tel: +65 6516 4335 (office) >> >> Original Message >> From: Atsushi Togo [mailto:atz.togo@...] >> Sent: Sunday, 1 September, 2013 8:30 PM >> To: Yang Ming >> Cc: phonopyusers@... >> Subject: Re: [Phonopyusers] Phonon dispersion of graphene >> >> Hi, >> >> Without watching your band structure, it is difficult to guess what your problem is. >> >> Togo >> >> On Sun, Sep 1, 2013 at 6:32 PM, Yang Ming <phyym@...> wrote: >>> Dear All, >>> >>> >>> >>> I am new in Phonopy. I was just using Phonopy to calculate phonon >>> dispersion of graphene for testing, but the results are incorrect >>> (negative frequencies were found). Could anyone there help to check my inputs below? >>> >>> >>> >>> >>> >>> Basically, I followed the Phonopy tutorial for VASP step by step: >>> >>> >>> >>> 1. Generate a 6×6×1 supercell from below bulk by using the commad: >>> phonopy d dim="6 6 1". >>> >>> >>> >>> C >>> >>> 1.00000000000000 >>> >>> 2.1304215583187101 1.2299994601826900 0.0000000000000000 >>> >>> 0.0000000000000000 2.4599989203653800 0.0000000000000000 >>> >>> 0.0000000000000000 0.0000000000000000 12.0000000000000000 >>> >>> 2 >>> >>> Direct >>> >>> 0.0000000000000000 0.0000000000000000 0.5000000000000000 >>> >>> 0.6666666699999979 0.3333333300000021 0.5000000000000000 >>> >>> >>> >>> 2. Calculate the force using the generated structure POSCAR001 and >>> standard VASP setting: >>> >>> >>> >>> ENCUT = 400 >>> >>> ISMEAR = 0; SIGMA = 0.01 >>> >>> PREC = A >>> >>> >>> >>> EDIFF = 1.0E8 >>> >>> NELM = 400 >>> >>> >>> >>> IDIPOL = 3 >>> >>> >>> >>> LWAVE = .F. >>> >>> LCHARG = .F. >>> >>> LREAL = .F. >>> >>> NPAR = 2 >>> >>> NELMDL = 5 >>> >>> LPLANE = .T. >>> >>> >>> >>> The KPOINTS for VASP calculation were set to 3×3×1. >>> >>> >>> >>> 3. Create the FORCE_SETS by using: phonopy f disp001/vasprun.xml >>> >>> 4. >>> >>> 5. Phonon band structure was calculated by using the band.conf like >>> below: >>> >>> >>> >>> ATOM_NAME = C >>> >>> DIM = 6 6 1 >>> >>> BAND = 0.0 0.5 0.0 0.0 0.0 0.0 0.3333 0.6667 0.0 0.0 0.5 0.0 >>> >>> 6. Plot the band structure by using: phonopy p band.conf >>> >>> >>> >>> After all these steps, I found some negative frequencies and the >>> values of the related frequencies are not consistent with those of >>> previous results also. I doulechecked the calculation processes, but don't know the problem. >>> >>> >>> >>> Thank you a lot for your time, and any comments are appreciated. >>> >>> >>> >>> Best regards, >>> >>> Yang Ming >>> >>> ****** >>> >>> Computational CondensedMatter Physics Lab >>> >>> Department of Physics >>> >>> National University of Singapore >>> >>> Singapore 117542 >>> >>> Email: phyym@...; ming_young@... >>> >>> Tel: +65 6516 4335 (office) >>> >>> >>> >>> >>>  >>>  Learn the latestVisual Studio 2012, SharePoint 2013, SQL >>> 2012, more! >>> Discover the easy way to master current and previous Microsoft >>> technologies and advance your career. Get an incredible 1,500+ hours >>> of stepbystep tutorial videos with LearnDevNow. Subscribe today and save! >>> http://pubads.g.doubleclick.net/gampad/clk?id=58040911&iu=/4140/ostg.c >>> lktrk _______________________________________________ >>> Phonopyusers mailing list >>> Phonopyusers@... >>> https://lists.sourceforge.net/lists/listinfo/phonopyusers >>> >> >> >> >>  >> Atsushi Togo >> http://atztogo.github.com/ >> atz.togo@... > > > >  > Atsushi Togo > http://atztogo.github.com/ > atz.togo@...  Atsushi Togo http://atztogo.github.com/ atz.togo@... 
From: Atsushi Togo <atz.togo@gm...>  20130903 05:04:57

Hi, Now I think your problem is not in phonopy but VASP. From vasp web site, For the calculation of the total energy in bulk materials we recommend the tetrahedron method with Blöchl corrections (ISMEAR=5). This method also gives a smooth nice electronic density of states (DOS). The only drawback is that the methods is not variational with respect to the partial occupancies, therefore the calculated forces might be wrong by a few percent for metals. For the calculation of phonon frequencies based on forces we recommend the method of MethfesselPaxton (ISMEAR>0). For semiconductors and insulators the forces are always correct, because partial occupancies do not vary! So you can read VASP manual carefully before asking VASP questions to the phonopy mailing list or you can ask them at the VASP forum. Togo On Tue, Sep 3, 2013 at 12:02 PM, Yang Ming <phyym@...> wrote: > Dear Prof. Togo, > > Thank you so much for the kind reply. > > I just checked the KPOINTS and the INCAR setting, which are correct for the electronic structure (see the band structure attached). > > Actually, I also increased denser kpoint meshes (18×18×1), and also changed the INCAR setting (ENCUT =500, ISMEAR=5, EDIFF=1.0E8, and ADDGRID = .True) , but still got negative frequencies. > > Let me show you my calculation processes again below, and please help me to check the problems and let me know if the settings are incorrect: > > 1. Generate a 6×6×1 supercell from below bulk by using the commad: phonopy d dim="6 6 1". > > 2. Calculate the force using the generated structure POSCAR001 and standard VASP setting. > > 3. Create the FORCE_SETS by using: phonopy f disp001/vasprun.xml > > 4 Phonon band structure was calculated by using the band.conf like below: > > ATOM_NAME = C > DIM = 6 6 1 > BAND = 0.0 0.5 0.0 0.0 0.0 0.0 0.3333 0.6667 0.0 0.0 0.5 0.0 > > 5. Plot the band structure by using: phonopy p band.conf > > > Thanks a lot. > Best regards, > Yang Ming > ****** > Computational CondensedMatter Physics Lab > Department of Physics > National University of Singapore > Singapore 117542 > Email: phyym@...; ming_young@... > Tel: +65 6516 4335 (office) > > > Original Message > From: Yang Ming > Sent: Sunday, 1 September, 2013 9:22 PM > To: 'Atsushi Togo' > Subject: RE: [Phonopyusers] Phonon dispersion of graphene > > Dear Dr. Togo, > > Thanks a lot for the prompt reply. > > Please find the attached data and calculated eigenvalue. > > Thanks. > > Best regards, > Yang Ming > ****** > Computational CondensedMatter Physics Lab Department of Physics National University of Singapore Singapore 117542 > Email: phyym@...; ming_young@... > Tel: +65 6516 4335 (office) > > Original Message > From: Atsushi Togo [mailto:atz.togo@...] > Sent: Sunday, 1 September, 2013 8:30 PM > To: Yang Ming > Cc: phonopyusers@... > Subject: Re: [Phonopyusers] Phonon dispersion of graphene > > Hi, > > Without watching your band structure, it is difficult to guess what your problem is. > > Togo > > On Sun, Sep 1, 2013 at 6:32 PM, Yang Ming <phyym@...> wrote: >> Dear All, >> >> >> >> I am new in Phonopy. I was just using Phonopy to calculate phonon >> dispersion of graphene for testing, but the results are incorrect >> (negative frequencies were found). Could anyone there help to check my inputs below? >> >> >> >> >> >> Basically, I followed the Phonopy tutorial for VASP step by step: >> >> >> >> 1. Generate a 6×6×1 supercell from below bulk by using the commad: >> phonopy d dim="6 6 1". >> >> >> >> C >> >> 1.00000000000000 >> >> 2.1304215583187101 1.2299994601826900 0.0000000000000000 >> >> 0.0000000000000000 2.4599989203653800 0.0000000000000000 >> >> 0.0000000000000000 0.0000000000000000 12.0000000000000000 >> >> 2 >> >> Direct >> >> 0.0000000000000000 0.0000000000000000 0.5000000000000000 >> >> 0.6666666699999979 0.3333333300000021 0.5000000000000000 >> >> >> >> 2. Calculate the force using the generated structure POSCAR001 and >> standard VASP setting: >> >> >> >> ENCUT = 400 >> >> ISMEAR = 0; SIGMA = 0.01 >> >> PREC = A >> >> >> >> EDIFF = 1.0E8 >> >> NELM = 400 >> >> >> >> IDIPOL = 3 >> >> >> >> LWAVE = .F. >> >> LCHARG = .F. >> >> LREAL = .F. >> >> NPAR = 2 >> >> NELMDL = 5 >> >> LPLANE = .T. >> >> >> >> The KPOINTS for VASP calculation were set to 3×3×1. >> >> >> >> 3. Create the FORCE_SETS by using: phonopy f disp001/vasprun.xml >> >> 4. >> >> 5. Phonon band structure was calculated by using the band.conf like >> below: >> >> >> >> ATOM_NAME = C >> >> DIM = 6 6 1 >> >> BAND = 0.0 0.5 0.0 0.0 0.0 0.0 0.3333 0.6667 0.0 0.0 0.5 0.0 >> >> 6. Plot the band structure by using: phonopy p band.conf >> >> >> >> After all these steps, I found some negative frequencies and the >> values of the related frequencies are not consistent with those of >> previous results also. I doulechecked the calculation processes, but don't know the problem. >> >> >> >> Thank you a lot for your time, and any comments are appreciated. >> >> >> >> Best regards, >> >> Yang Ming >> >> ****** >> >> Computational CondensedMatter Physics Lab >> >> Department of Physics >> >> National University of Singapore >> >> Singapore 117542 >> >> Email: phyym@...; ming_young@... >> >> Tel: +65 6516 4335 (office) >> >> >> >> >>  >>  Learn the latestVisual Studio 2012, SharePoint 2013, SQL >> 2012, more! >> Discover the easy way to master current and previous Microsoft >> technologies and advance your career. Get an incredible 1,500+ hours >> of stepbystep tutorial videos with LearnDevNow. Subscribe today and save! >> http://pubads.g.doubleclick.net/gampad/clk?id=58040911&iu=/4140/ostg.c >> lktrk _______________________________________________ >> Phonopyusers mailing list >> Phonopyusers@... >> https://lists.sourceforge.net/lists/listinfo/phonopyusers >> > > > >  > Atsushi Togo > http://atztogo.github.com/ > atz.togo@...  Atsushi Togo http://atztogo.github.com/ atz.togo@... 
From: Yang Ming <phyym@nu...>  20130903 03:03:06

Dear Prof. Togo, Thank you so much for the kind reply. I just checked the KPOINTS and the INCAR setting, which are correct for the electronic structure (see the band structure attached). Actually, I also increased denser kpoint meshes (18×18×1), and also changed the INCAR setting (ENCUT =500, ISMEAR=5, EDIFF=1.0E8, and ADDGRID = .True) , but still got negative frequencies. Let me show you my calculation processes again below, and please help me to check the problems and let me know if the settings are incorrect: 1. Generate a 6×6×1 supercell from below bulk by using the commad: phonopy d dim="6 6 1". 2. Calculate the force using the generated structure POSCAR001 and standard VASP setting. 3. Create the FORCE_SETS by using: phonopy f disp001/vasprun.xml 4 Phonon band structure was calculated by using the band.conf like below: ATOM_NAME = C DIM = 6 6 1 BAND = 0.0 0.5 0.0 0.0 0.0 0.0 0.3333 0.6667 0.0 0.0 0.5 0.0 5. Plot the band structure by using: phonopy p band.conf Thanks a lot. Best regards, Yang Ming ****** Computational CondensedMatter Physics Lab Department of Physics National University of Singapore Singapore 117542 Email: phyym@...; ming_young@... Tel: +65 6516 4335 (office) Original Message From: Yang Ming Sent: Sunday, 1 September, 2013 9:22 PM To: 'Atsushi Togo' Subject: RE: [Phonopyusers] Phonon dispersion of graphene Dear Dr. Togo, Thanks a lot for the prompt reply. Please find the attached data and calculated eigenvalue. Thanks. Best regards, Yang Ming ****** Computational CondensedMatter Physics Lab Department of Physics National University of Singapore Singapore 117542 Email: phyym@...; ming_young@... Tel: +65 6516 4335 (office) Original Message From: Atsushi Togo [mailto:atz.togo@...] Sent: Sunday, 1 September, 2013 8:30 PM To: Yang Ming Cc: phonopyusers@... Subject: Re: [Phonopyusers] Phonon dispersion of graphene Hi, Without watching your band structure, it is difficult to guess what your problem is. Togo On Sun, Sep 1, 2013 at 6:32 PM, Yang Ming <phyym@...> wrote: > Dear All, > > > > I am new in Phonopy. I was just using Phonopy to calculate phonon > dispersion of graphene for testing, but the results are incorrect > (negative frequencies were found). Could anyone there help to check my inputs below? > > > > > > Basically, I followed the Phonopy tutorial for VASP step by step: > > > > 1. Generate a 6×6×1 supercell from below bulk by using the commad: > phonopy d dim="6 6 1". > > > > C > > 1.00000000000000 > > 2.1304215583187101 1.2299994601826900 0.0000000000000000 > > 0.0000000000000000 2.4599989203653800 0.0000000000000000 > > 0.0000000000000000 0.0000000000000000 12.0000000000000000 > > 2 > > Direct > > 0.0000000000000000 0.0000000000000000 0.5000000000000000 > > 0.6666666699999979 0.3333333300000021 0.5000000000000000 > > > > 2. Calculate the force using the generated structure POSCAR001 and > standard VASP setting: > > > > ENCUT = 400 > > ISMEAR = 0; SIGMA = 0.01 > > PREC = A > > > > EDIFF = 1.0E8 > > NELM = 400 > > > > IDIPOL = 3 > > > > LWAVE = .F. > > LCHARG = .F. > > LREAL = .F. > > NPAR = 2 > > NELMDL = 5 > > LPLANE = .T. > > > > The KPOINTS for VASP calculation were set to 3×3×1. > > > > 3. Create the FORCE_SETS by using: phonopy f disp001/vasprun.xml > > 4. > > 5. Phonon band structure was calculated by using the band.conf like > below: > > > > ATOM_NAME = C > > DIM = 6 6 1 > > BAND = 0.0 0.5 0.0 0.0 0.0 0.0 0.3333 0.6667 0.0 0.0 0.5 0.0 > > 6. Plot the band structure by using: phonopy p band.conf > > > > After all these steps, I found some negative frequencies and the > values of the related frequencies are not consistent with those of > previous results also. I doulechecked the calculation processes, but don't know the problem. > > > > Thank you a lot for your time, and any comments are appreciated. > > > > Best regards, > > Yang Ming > > ****** > > Computational CondensedMatter Physics Lab > > Department of Physics > > National University of Singapore > > Singapore 117542 > > Email: phyym@...; ming_young@... > > Tel: +65 6516 4335 (office) > > > > >  >  Learn the latestVisual Studio 2012, SharePoint 2013, SQL > 2012, more! > Discover the easy way to master current and previous Microsoft > technologies and advance your career. Get an incredible 1,500+ hours > of stepbystep tutorial videos with LearnDevNow. Subscribe today and save! > http://pubads.g.doubleclick.net/gampad/clk?id=58040911&iu=/4140/ostg.c > lktrk _______________________________________________ > Phonopyusers mailing list > Phonopyusers@... > https://lists.sourceforge.net/lists/listinfo/phonopyusers >  Atsushi Togo http://atztogo.github.com/ atz.togo@... 
From: Atsushi Togo <atz.togo@gm...>  20130901 13:43:49

Hi, Your band structure is really broken. I doubt your setting of ISMEAR = 0; SIGMA = 0.01, and few kpoints though I don't know how much it affect to phonons. So I recommend you to check carefully about VASP setting, e.g. by drawing electronic band structure. If you don't know well about electronic structure calculations, you may ask it to your colleagues. Togo On Sun, Sep 1, 2013 at 10:21 PM, Yang Ming <phyym@...> wrote: > Dear Dr. Togo, > > Thanks a lot for the prompt reply. > > Please find the attached data and calculated eigenvalue. > > Thanks. > > Best regards, > Yang Ming > ****** > Computational CondensedMatter Physics Lab > Department of Physics > National University of Singapore > Singapore 117542 > Email: phyym@...; ming_young@... > Tel: +65 6516 4335 (office) > > Original Message > From: Atsushi Togo [mailto:atz.togo@...] > Sent: Sunday, 1 September, 2013 8:30 PM > To: Yang Ming > Cc: phonopyusers@... > Subject: Re: [Phonopyusers] Phonon dispersion of graphene > > Hi, > > Without watching your band structure, it is difficult to guess what your problem is. > > Togo > > On Sun, Sep 1, 2013 at 6:32 PM, Yang Ming <phyym@...> wrote: >> Dear All, >> >> >> >> I am new in Phonopy. I was just using Phonopy to calculate phonon >> dispersion of graphene for testing, but the results are incorrect >> (negative frequencies were found). Could anyone there help to check my inputs below? >> >> >> >> >> >> Basically, I followed the Phonopy tutorial for VASP step by step: >> >> >> >> 1. Generate a 6×6×1 supercell from below bulk by using the commad: >> phonopy d dim="6 6 1". >> >> >> >> C >> >> 1.00000000000000 >> >> 2.1304215583187101 1.2299994601826900 0.0000000000000000 >> >> 0.0000000000000000 2.4599989203653800 0.0000000000000000 >> >> 0.0000000000000000 0.0000000000000000 12.0000000000000000 >> >> 2 >> >> Direct >> >> 0.0000000000000000 0.0000000000000000 0.5000000000000000 >> >> 0.6666666699999979 0.3333333300000021 0.5000000000000000 >> >> >> >> 2. Calculate the force using the generated structure POSCAR001 and >> standard VASP setting: >> >> >> >> ENCUT = 400 >> >> ISMEAR = 0; SIGMA = 0.01 >> >> PREC = A >> >> >> >> EDIFF = 1.0E8 >> >> NELM = 400 >> >> >> >> IDIPOL = 3 >> >> >> >> LWAVE = .F. >> >> LCHARG = .F. >> >> LREAL = .F. >> >> NPAR = 2 >> >> NELMDL = 5 >> >> LPLANE = .T. >> >> >> >> The KPOINTS for VASP calculation were set to 3×3×1. >> >> >> >> 3. Create the FORCE_SETS by using: phonopy f disp001/vasprun.xml >> >> 4. >> >> 5. Phonon band structure was calculated by using the band.conf like >> below: >> >> >> >> ATOM_NAME = C >> >> DIM = 6 6 1 >> >> BAND = 0.0 0.5 0.0 0.0 0.0 0.0 0.3333 0.6667 0.0 0.0 0.5 0.0 >> >> 6. Plot the band structure by using: phonopy p band.conf >> >> >> >> After all these steps, I found some negative frequencies and the >> values of the related frequencies are not consistent with those of >> previous results also. I doulechecked the calculation processes, but don't know the problem. >> >> >> >> Thank you a lot for your time, and any comments are appreciated. >> >> >> >> Best regards, >> >> Yang Ming >> >> ****** >> >> Computational CondensedMatter Physics Lab >> >> Department of Physics >> >> National University of Singapore >> >> Singapore 117542 >> >> Email: phyym@...; ming_young@... >> >> Tel: +65 6516 4335 (office) >> >> >> >> >>  >>  Learn the latestVisual Studio 2012, SharePoint 2013, SQL >> 2012, more! >> Discover the easy way to master current and previous Microsoft >> technologies and advance your career. Get an incredible 1,500+ hours >> of stepbystep tutorial videos with LearnDevNow. Subscribe today and save! >> http://pubads.g.doubleclick.net/gampad/clk?id=58040911&iu=/4140/ostg.c >> lktrk _______________________________________________ >> Phonopyusers mailing list >> Phonopyusers@... >> https://lists.sourceforge.net/lists/listinfo/phonopyusers >> > > > >  > Atsushi Togo > http://atztogo.github.com/ > atz.togo@...  Atsushi Togo http://atztogo.github.com/ atz.togo@... 
From: Atsushi Togo <atz.togo@gm...>  20130901 12:30:33

Hi, Without watching your band structure, it is difficult to guess what your problem is. Togo On Sun, Sep 1, 2013 at 6:32 PM, Yang Ming <phyym@...> wrote: > Dear All, > > > > I am new in Phonopy. I was just using Phonopy to calculate phonon dispersion > of graphene for testing, but the results are incorrect (negative frequencies > were found). Could anyone there help to check my inputs below? > > > > > > Basically, I followed the Phonopy tutorial for VASP step by step: > > > > 1. Generate a 6×6×1 supercell from below bulk by using the commad: > phonopy –d –dim="6 6 1". > > > > C > > 1.00000000000000 > > 2.1304215583187101 1.2299994601826900 0.0000000000000000 > > 0.0000000000000000 2.4599989203653800 0.0000000000000000 > > 0.0000000000000000 0.0000000000000000 12.0000000000000000 > > 2 > > Direct > > 0.0000000000000000 0.0000000000000000 0.5000000000000000 > > 0.6666666699999979 0.3333333300000021 0.5000000000000000 > > > > 2. Calculate the force using the generated structure POSCAR001 and > standard VASP setting: > > > > ENCUT = 400 > > ISMEAR = 0; SIGMA = 0.01 > > PREC = A > > > > EDIFF = 1.0E8 > > NELM = 400 > > > > IDIPOL = 3 > > > > LWAVE = .F. > > LCHARG = .F. > > LREAL = .F. > > NPAR = 2 > > NELMDL = 5 > > LPLANE = .T. > > > > The KPOINTS for VASP calculation were set to 3×3×1. > > > > 3. Create the FORCE_SETS by using: phonopy –f disp001/vasprun.xml > > 4. > > 5. Phonon band structure was calculated by using the band.conf like > below: > > > > ATOM_NAME = C > > DIM = 6 6 1 > > BAND = 0.0 0.5 0.0 0.0 0.0 0.0 0.3333 0.6667 0.0 0.0 0.5 0.0 > > 6. Plot the band structure by using: phonopy –p band.conf > > > > After all these steps, I found some negative frequencies and the values of > the related frequencies are not consistent with those of previous results > also. I doulechecked the calculation processes, but don’t know the problem. > > > > Thank you a lot for your time, and any comments are appreciated. > > > > Best regards, > > Yang Ming > > ****** > > Computational CondensedMatter Physics Lab > > Department of Physics > > National University of Singapore > > Singapore 117542 > > Email: phyym@...; ming_young@... > > Tel: +65 6516 4335 (office) > > > > >  > Learn the latestVisual Studio 2012, SharePoint 2013, SQL 2012, more! > Discover the easy way to master current and previous Microsoft technologies > and advance your career. Get an incredible 1,500+ hours of stepbystep > tutorial videos with LearnDevNow. Subscribe today and save! > http://pubads.g.doubleclick.net/gampad/clk?id=58040911&iu=/4140/ostg.clktrk > _______________________________________________ > Phonopyusers mailing list > Phonopyusers@... > https://lists.sourceforge.net/lists/listinfo/phonopyusers >  Atsushi Togo http://atztogo.github.com/ atz.togo@... 
From: Yang Ming <phyym@nu...>  20130901 09:32:21
