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: Alejandro Rébola <arebol2@ui...>  20130823 05:56:02

Dear Santu, The way phonopy works is by creating the dynamical matrix for each q [let's call it D(q)] and then diagonalizing it. To create D(q) in theory you'd need all the force constants for each lattice vector R (let's assume no basis). They are all the possible pair of interactions in the lattice. However since the number of R's is very large, this would be impossible to do in practice, so the summation has to be cut off at some point. What I understand phonopy is doing is to consider only interactions within the range of the supercell you are using. The larger the supercell you use the more accurate your calculation, also because in this way you reduce interactions between periodic images. Hope this helps. Best, Alejandro On Thu, August 22, 2013 11:50 pm, Santu Baidya wrote: > Dear phonopy users, > > I did a phonon band dispersion calculation without making a supercell > using > phonopy. > > First I used VASP, IBRION=8 (DFPT) to calculate forces. Then I used > phonopy > to get band dispersion. > > What I know for single unit cell I should always get GAMMA point phonons > only. But phonopy is capable of giving frequencies at other kpoints also > ! > > Could anyone please please explain it how is it possible. > > Thanks. > > Santu Baidya > > > > > *"The happiest people do not always have the best of all,* * they simply > appreciate what they find on their way!!!" SANTU > * >  > Introducing Performance Central, a new site from SourceForge and > AppDynamics. Performance Central is your source for news, insights, > analysis and resources for efficient Application Performance Management. > Visit us today! > http://pubads.g.doubleclick.net/gampad/clk?id=48897511&iu=/4140/ostg.clktrk_______________________________________________ > Phonopyusers mailing list > Phonopyusers@... > https://lists.sourceforge.net/lists/listinfo/phonopyusers > 
From: Santu Baidya <santubaidya2009@gm...>  20130823 07:16:15
Attachments:
Message as HTML

 Forwarded message  From: Santu Baidya <santubaidya2009@...> Date: 23 August 2013 12:44 Subject: Re: [Phonopyusers] How from a single unit cell (not supercell) phonon dispersion appears? To: Alejandro Rébola <arebol2@...> Dear phonopy users, I have got an explanation for the single unit cell phonon calculation using phonopy and I need to know if that is correct or not. >From phonopy manual I found that A force and a secondorder force constant [image: \Phi_{\alpha \beta}] are given by [image: F_\alpha(jl) = \frac{\partial V }{\partial r_\alpha(jl)}] where l is an unit cell and [image: \Phi_{\alpha\beta}(jl, j'l') = \frac{\partial^2 V}{\partial r_\alpha(jl) \partial r_\beta(j'l')} = \frac{\partial F_\beta(j'l')}{\partial r_\alpha(jl)},] For single unit cell it will be l = l' VASP gives Force on atoms although at single qpoint (by DFPT method) also mentioned in phonopy manual. phonopy calculates Force constant matrix from the Forces. Then Dynamical Matrix form will be D (jj', q) = 1/√(mj mj' ) [image: \Phi_{\alpha \beta}] (j0, j'0) exp( iq . [r(j'0)  r(j0)]) And then from here we get dynamical matrix for pair interactions and then solve the dynamical eqn of motion. We get then phonon frequencies at other qpoints also. Could any one please tell me if this is correct explanation or not. Thanks. Santu Baidya On 23 August 2013 11:19, Santu Baidya <santubaidya2009@...> wrote: > Dear Alejandro, > > Thanks for your reply . Ofcourse it helps me to understand. > > Can I ask you then if I use original unit cell (not supercell) whose > lattice parameters ~ 8 Angstrom (large enough) and do not make any > supercell and do a VASP calculation (GAMMA point ) to get forces within > that range. > > And then I use phonopy I shall get frequencies even at other Kponts. > > Is that right? > > Could you please explain it little more. > > Thanks again. > > Santu Baidya > > >  *"The happiest people do not always have the best of all,* * they simply appreciate what they find on their way!!!" SANTU *  *"The happiest people do not always have the best of all,* * they simply appreciate what they find on their way!!!" SANTU * 
From: Atsushi Togo <atz.togo@gm...>  20130825 13:11:06
Attachments:
Message as HTML

Hi, Alejandro explains all. In the case of supercell approach, finite displacement of an atom appears periodicity of supercell, so it has to be considered though. It's considered as an interpolation scheme, i.e., except for commensurate points (e.g. q=(0.5, 0, 0) for 2x1x1 supercell), they are not exact. This approximation works because usually atomic interaction is short range. If the unit cell is huge, the interaction from an atom to an atom outside of unit cell is almost zero. In this case, even using unit cell only, this approach works well. But electrostatic polarization is an exception and would be treated differently anyway. Togo On Fri, Aug 23, 2013 at 4:16 PM, Santu Baidya <santubaidya2009@...>wrote: > > >  Forwarded message  > From: Santu Baidya <santubaidya2009@...> > Date: 23 August 2013 12:44 > Subject: Re: [Phonopyusers] How from a single unit cell (not supercell) > phonon dispersion appears? > To: Alejandro Rébola <arebol2@...> > > > Dear phonopy users, > > I have got an explanation for the single unit cell phonon calculation > using phonopy and I need to know if that is correct or not. > > From phonopy manual I found that > A force and a secondorder force constant [image: \Phi_{\alpha \beta}]are given by > > [image: F_\alpha(jl) = \frac{\partial V }{\partial r_\alpha(jl)}] > where l is an unit cell and > > [image: \Phi_{\alpha\beta}(jl, j'l') = \frac{\partial^2 V}{\partial > r_\alpha(jl) \partial r_\beta(j'l')} = \frac{\partial > F_\beta(j'l')}{\partial r_\alpha(jl)},] > For single unit cell it will be l = l' > > VASP gives Force on atoms although at single qpoint (by DFPT method) also > mentioned in phonopy manual. > > phonopy calculates Force constant matrix from the Forces. > > Then Dynamical Matrix form will be > > D (jj', q) = 1/√(mj mj' ) [image: \Phi_{\alpha \beta}] (j0, j'0) exp( > iq . [r(j'0)  r(j0)]) > > > > And then from here we get dynamical matrix for pair interactions and then > solve the dynamical eqn of motion. > > We get then phonon frequencies at other qpoints also. > > > Could any one please tell me if this is correct explanation or not. > > > Thanks. > > Santu Baidya > > > > > > > > > > > On 23 August 2013 11:19, Santu Baidya <santubaidya2009@...> wrote: > >> Dear Alejandro, >> >> Thanks for your reply . Ofcourse it helps me to understand. >> >> Can I ask you then if I use original unit cell (not supercell) whose >> lattice parameters ~ 8 Angstrom (large enough) and do not make any >> supercell and do a VASP calculation (GAMMA point ) to get forces within >> that range. >> >> And then I use phonopy I shall get frequencies even at other Kponts. >> >> Is that right? >> >> Could you please explain it little more. >> >> Thanks again. >> >> Santu Baidya >> >> >> > > >  > *"The happiest people do not always have the best of all,* * they simply > appreciate what they find on their way!!!" SANTU > * > > > >  > *"The happiest people do not always have the best of all,* * they simply > appreciate what they find on their way!!!" SANTU > * > > >  > Introducing Performance Central, a new site from SourceForge and > AppDynamics. Performance Central is your source for news, insights, > analysis and resources for efficient Application Performance Management. > Visit us today! > http://pubads.g.doubleclick.net/gampad/clk?id=48897511&iu=/4140/ostg.clktrk > _______________________________________________ > Phonopyusers mailing list > Phonopyusers@... > https://lists.sourceforge.net/lists/listinfo/phonopyusers > >  Atsushi Togo http://atztogo.github.com/ atz.togo@... 