Screenshot instructions:
Windows
Mac
Red Hat Linux
Ubuntu
Click URL instructions:
Rightclick on ad, choose "Copy Link", then paste here →
(This may not be possible with some types of ads)
From: Huan Tran <huantd@gm...>  20121018 14:08:52
Attachments:
Message as HTML
image.png

From: Atsushi Togo <atz.togo@gm...>  20121018 15:33:35

Hi, What you think is correct within harmonic approximation level. Did you draw the curve of energy as a function of amplitude? How is the shape of the curve? Best regards, Togo On Thu, Oct 18, 2012 at 11:08 PM, Huan Tran <huantd@...> wrote: > Dear all > > I am using phonopy (version 1.5) to calculate the phonon band of a crystal > structure. The result is attached, showing that there is a phonon mode with > imaginary frequencies. The qpath used in this figure is > > [0.500 0.000 0.500] > [0.000 0.000 0.000] > [0.000 0.500 0.500] > > [0.500 0.500 0.000] > [0.000 0.000 0.000] > > I have chosen 2 qpoints to do modulation with various amplitudes: [0.500 > 0.000 0.500] and [0.500 0.500 0.000] (of course, I modulated the > structure with one q point each time). For the second q point, I did try to > modulate with 1 and/or 2 lowest phonon modes. In all of the attempts, the > single point calculation for any obtained MPOSCAR returns an energy which is > ALWAYS HIGHER than the singlepoint calculation with MPOSCARorig. It is > strange for me, because I expect that along the unstable phonon mode, energy > has to go down, even with a singlepoint calculation (without relaxation). > Is this true? Please correct me if what I expect is wrong. And please > request me any more information if you need to help me to solve this > problem. > > Thank you > Huan Tran > >  > Dr. Huan Tran > Assistent > Departement Physik > Universitaet Basel > Klingelbergstrasse 82 > CH4056 Basel > > Tel.: +41 61 267 39 10 > Fax: +41 61 267 37 84 > Email: huan.tran@... > >  > Everyone hates slow websites. So do we. > Make your web apps faster with AppDynamics > Download AppDynamics Lite for free today: > http://p.sf.net/sfu/appdyn_sfd2d_oct > _______________________________________________ > Phonopyusers mailing list > Phonopyusers@... > https://lists.sourceforge.net/lists/listinfo/phonopyusers >  Atsushi Togo http://atztogo.users.sourceforge.net/ atz.togo@... 
From: Atsushi Togo <atz.togo@gm...>  20121018 16:22:19

Hi, I have forgotten to mention one more important point, especially for nongamma point calculation. The phase factor can be, roughly speaking, determined randomly, if the eigenvector of x element of the first atom is very small. The following procedure should be carefully considered: 1. Watch modulation.yaml first. There the complex atomic modulations are written. The real values are used for MPOSCAR. 2. Draw energy vs phase factor (e.g. 0 to 180) with fixed amplitude. And again watch modulation.yaml with respect to phase factor. 3. If the eigenmodes are degenerated, all the linear combinations can be also possible modulations. Best regards, Togo On Fri, Oct 19, 2012 at 12:33 AM, Atsushi Togo <atz.togo@...> wrote: > Hi, > > What you think is correct within harmonic approximation level. Did you > draw the curve of energy as a function of amplitude? How is the shape > of the curve? > > Best regards, > > Togo > > > On Thu, Oct 18, 2012 at 11:08 PM, Huan Tran <huantd@...> wrote: >> Dear all >> >> I am using phonopy (version 1.5) to calculate the phonon band of a crystal >> structure. The result is attached, showing that there is a phonon mode with >> imaginary frequencies. The qpath used in this figure is >> >> [0.500 0.000 0.500] > [0.000 0.000 0.000] > [0.000 0.500 0.500] >> > [0.500 0.500 0.000] > [0.000 0.000 0.000] >> >> I have chosen 2 qpoints to do modulation with various amplitudes: [0.500 >> 0.000 0.500] and [0.500 0.500 0.000] (of course, I modulated the >> structure with one q point each time). For the second q point, I did try to >> modulate with 1 and/or 2 lowest phonon modes. In all of the attempts, the >> single point calculation for any obtained MPOSCAR returns an energy which is >> ALWAYS HIGHER than the singlepoint calculation with MPOSCARorig. It is >> strange for me, because I expect that along the unstable phonon mode, energy >> has to go down, even with a singlepoint calculation (without relaxation). >> Is this true? Please correct me if what I expect is wrong. And please >> request me any more information if you need to help me to solve this >> problem. >> >> Thank you >> Huan Tran >> >>  >> Dr. Huan Tran >> Assistent >> Departement Physik >> Universitaet Basel >> Klingelbergstrasse 82 >> CH4056 Basel >> >> Tel.: +41 61 267 39 10 >> Fax: +41 61 267 37 84 >> Email: huan.tran@... >> >>  >> Everyone hates slow websites. So do we. >> Make your web apps faster with AppDynamics >> Download AppDynamics Lite for free today: >> http://p.sf.net/sfu/appdyn_sfd2d_oct >> _______________________________________________ >> Phonopyusers mailing list >> Phonopyusers@... >> https://lists.sourceforge.net/lists/listinfo/phonopyusers >> > > > >  > Atsushi Togo > http://atztogo.users.sourceforge.net/ > atz.togo@...  Atsushi Togo http://atztogo.users.sourceforge.net/ atz.togo@... 
From: Huan Tran <huantd@gm...>  20121019 10:41:15
Attachments:
Message as HTML
modulation.eps

From: Atsushi Togo <atz.togo@gm...>  20121019 13:19:37

Hi, > I have tried several amplitudes within 0.01, the default displacement used > in phonopy, and the energy curve in included: it still goes up. The slope > for these lines are 0.002 eV/A while I used EDIFFG = 1E04, which is quite > smaller. Hopefully EDIFFG here is small enough. Probably EDIFFG = 1E04 is OK, I think. The curve should be parabolic, but yours is not. Why? > As the default displacement is 0.01 \AA, do you have any estimation on how > big a supper cell should be, like 20 x 20 x 20 \AA, for example. I guess > that it is not easy, and depends on the system. If you want the modulation at [0.500 0.000 0.500], then you need only 2x1x2 supercell. It is not good if you didn't calculate with like 2x1x2 or 2x2x2 supercells. > My particular case is not easy for me as I worked with it for a month > without solution. Can you suggest me any material which describes in more > detailed the treatment for nonGamma point. It's dependent on what you want to do. So I have no idea what I can show. Togo > > Regards > Huan > > On Fri, Oct 19, 2012 at 3:23 AM, Atsushi Togo <atz.togo@...> wrote: >> >> Can I reply this to the mailing list? I want to share questions and >> answers because I don't want to do same Q&A repeatedly. And also >> perhaps somebody else can answer questions. >> >> > Thank you for your reply. The energy goes up nearly parabolic with the >> > amplitude ranged from 0.1 to 2  I tried both sign, plus and minus >> > (sorry >> > i can not access my plot at the moment). I think that an amplitude of >> > 0.1 >> > may be OK for the harmonic approximation. How do you think abou it? >> >> The default displacement used for the force constant calculation is >> (in the case of finite displacement methods) 0.01. So you should >> calculate to see how it behaves 0.01 and around. >> >> > For the next steps, I am going to follow your instructions tomorrow, but >> > let >> > me make sure to inderstand you correctly. I understand that I should >> > play >> > with the phase factor and find the best value that makes the energy go >> > down >> > the most, is this correct? >> >> It depends on the system. What is difficult is that the wave vector is >> nonGamma. If Gamma, the eigenvectors can be chosen to be real values. >> You can calculate the Brillouin zone boundary as a Gammapoint to make >> your unit cell as supercell. In this case, the wave vectors on the >> boundary folded to the Gamma point. But in this case, those >> eigenvalues will be degenerate. How to handle the degeneracy is quite >> the matter physically and practically. It is beyond how to use >> phonopy. We may need discussion. >> >> > There is no degenerated phonon modes in my problem. As shown in the band >> > figure, the is one imaginary mode at [0.500 0.000 0.500] so only this >> > mode should be modulation, is this right? >> >> I hope so. But I'm not sure unless I see your calculation (or some >> example the situation can be explained). I don't mean I want to see >> your research. >> >> Togo >> >> > >> > Thank you very much. >> > >> > Regards >> > Huan >> > >> > >> > On Thursday, October 18, 2012, Atsushi Togo <atz.togo@...> wrote: >> >> Hi, >> >> >> >> I have forgotten to mention one more important point, especially for >> >> nongamma point calculation. >> >> The phase factor can be, roughly speaking, determined randomly, if the >> >> eigenvector of x element of the first atom is very small. The >> >> following procedure should be carefully considered: >> >> >> >> 1. Watch modulation.yaml first. There the complex atomic modulations >> >> are written. The real values are used for MPOSCAR. >> >> 2. Draw energy vs phase factor (e.g. 0 to 180) with fixed amplitude. >> >> And again watch modulation.yaml with respect to phase factor. >> >> 3. If the eigenmodes are degenerated, all the linear combinations can >> >> be also possible modulations. >> >> >> >> Best regards, >> >> >> >> Togo >> >> >> >> >> >> On Fri, Oct 19, 2012 at 12:33 AM, Atsushi Togo <atz.togo@...> >> >> wrote: >> >>> Hi, >> >>> >> >>> What you think is correct within harmonic approximation level. Did you >> >>> draw the curve of energy as a function of amplitude? How is the shape >> >>> of the curve? >> >>> >> >>> Best regards, >> >>> >> >>> Togo >> >>> >> >>> >> >>> On Thu, Oct 18, 2012 at 11:08 PM, Huan Tran <huantd@...> wrote: >> >>>> Dear all >> >>>> >> >>>> I am using phonopy (version 1.5) to calculate the phonon band of a >> >>>> crystal >> >>>> structure. The result is attached, showing that there is a phonon >> >>>> mode >> >>>> with >> >>>> imaginary frequencies. The qpath used in this figure is >> >>>> >> >>>> [0.500 0.000 0.500] > [0.000 0.000 0.000] > [0.000 0.500 >> >>>> 0.500] >> >>>> > [0.500 0.500 0.000] > [0.000 0.000 0.000] >> >>>> >> >>>> I have chosen 2 qpoints to do modulation with various amplitudes: >> >>>> [0.500 >> >>>> 0.000 0.500] and [0.500 0.500 0.000] (of course, I modulated the >> >>>> structure with one q point each time). For the second q point, I did >> >>>> try >> >>>> to >> >>>> modulate with 1 and/or 2 lowest phonon modes. In all of the attempts, >> >>>> the >> >>>> single point calculation for any obtained MPOSCAR returns an energy >> >>>> which is >> >>>> ALWAYS HIGHER than the singlepoint calculation with MPOSCARorig. It >> >>>> is >> >>>> strange for me, because I expect that along the unstable phonon mode, >> >>>> energy >> >>>> has to go down, even with a singlepoint calculation (without >> >>>> relaxation). >> >>>> Is this true? Please correct me if what I expect is wrong. And please >> >>>> request me any more information if you need to help me to solve this >> >>>> problem. >> >>>> >> >>>> Thank you >> >>>> Huan Tran >> >>>> >> >>>>  >> >>>> Dr. Huan Tran >> >>>> Assistent >> >>>> Departement Physik >> >>>> Universitaet Basel >> >>>> Klingelbergstrasse 82 >> >>>> CH4056 Basel >> >>>> >> >>>> Tel.: +41 61 267 39 10 >> >>>> Fax: +41 61 267 37 84 >> >>>> Email: huan.tran@... >> >>>> >> >>>> >> >>>> >> >>>>  >> >>>> Everyone hates slow websites. So do we. >> >>>> Make your web apps faster with AppDynamics >> >>>> Download AppDynamics Lite for free today: >> >>>> http://p.sf.net/sfu/appdyn_sfd2d_oct >> >>>> _______________________________________________ >> >>>> Phonopyusers mailing list >> >>>> Phonopyusers@... >> >>>> https://lists.sourceforge.net/lists/listinfo/phonopyusers >> >>>> >> >>> >> >>> >> >>> >> >>>  >> >>> Atsushi Togo >> >>> http://atztogo.users.sourceforge.net/ >> >>> atz.togo@... >> >> >> >> >> >> >> >>  >> >> Atsushi Togo >> >> http://atztogo.users.sourceforge.net/ >> >> atz.togo@... >> >> >> > >> >  >> > Dr. Huan Tran >> > Assistent >> > Departement Physik >> > Universitaet Basel >> > Klingelbergstrasse 82 >> > CH4056 Basel >> > >> > Tel.: +41 61 267 39 10 >> > Fax: +41 61 267 37 84 >> > Email: huan.tran@... >> >> >> >>  >> Atsushi Togo >> http://atztogo.users.sourceforge.net/ >> atz.togo@... > > > > >  > Dr. Huan Tran > Assistent > Departement Physik > Universitaet Basel > Klingelbergstrasse 82 > CH4056 Basel > > Tel.: +41 61 267 39 10 > Fax: +41 61 267 37 84 > Email: huan.tran@...  Atsushi Togo http://atztogo.users.sourceforge.net/ atz.togo@... 
Sign up for the SourceForge newsletter:
No, thanks