[Pw_forum] Some questions about IFC from PWSCF
75925926E-05 6 2 4 -6.13445370370E-05 1 3 4 -2.3023889E-05 2 3 4 -1.5069444E-06 3 3 4 2.3598333E-05 4 3 4 2.8116667E-06 5 3 4 2.3598333E-05 6 3 4 -1.5069444E-06 1 4 4 -9.10425462963E-05 2 4 4 2.44929629630E-05 3 4 4 -1.40046296296E-06 4 4 4 1.51428240741E-05 5 4 4 -1.40046296296E-06 6 4 4 2.44929629630E-05 1 5 4 -2.3023889E-05 2 5 4 -1.5069444E-06 3 5 4 2.3598333E-05 4 5 4 2.8116667E-06 5 5 4 2.3598333E-05 6 5 4 -1.5069444E-06 1 6 4 2.77759259259E-06 2 6 4 -6.13445370370E-05 3 6 4 7.55775925926E-05 4 6 4 -1.15009259259E-05 5 6 4 7.55775925926E-05 6 6 4 -6.13445370370E-05 1 1 5 -6.843E-06 2 1 5 -6.65476851852E-05 3 1 5 -2.7346667E-05 4 1 5 -1.06576851852E-04 5 1 5 -2.7346667E-05 6 1 5 -6.65476851852E-05 1 2 5 1.7820125E-04 2 2 5 1.0270611E-04 3 2 5 -6.2165278E-05 4 2 5 1.2760972E-05 5 2 5 -6.2165278E-05 6 2 5 1.0270611E-04 1 3 5 -4.8187639E-05 2 3 5 -4.36212962963E-06 3 3 5 -5.6023889E-05 4 3 5 4.51591203704E-05 5 3 5 -5.6023889E-05 6 3 5 -4.36212962963E-06 1 4 5 -2.3023889E-05 2 4 5 -1.5069444E-06 3 4 5 2.3598333E-05 4 4 5 2.8116667E-06 5 4 5 2.3598333E-05 6 4 5 -1.5069444E-06 1 5 5 -4.8187639E-05 2 5 5 -4.36212962963E-06 3 5 5 -5.6023889E-05 4 5 5 4.51591203704E-05 5 5 5 -5.6023889E-05 6 5 5 -4.36212962963E-06 1 6 5 1.7820125E-04 2 6 5 1.0270611E-04 3 6 5 -6.2165278E-05 4 6 5 1.2760972E-05 5 6 5 -6.2165278E-05 6 6 5 1.0270611E-04 1 1 6 2.6899633E-03 2 1 6 6.6021083E-04 3 1 6 3.8196056E-04 4 1 6 -2.6103722E-04 5 1 6 3.8196056E-04 6 1 6 6.6021083E-04 1 2 6 1.07580273148E-03 2 2 6 -8.68419537037E-04 3 2 6 -1.26556851852E-04 4 2 6 4.42628240741E-05 5 2 6 -1.26556851852E-04 6 2 6 -8.68419537037E-04 1 3 6 1.7820125E-04 2 3 6 1.0270611E-04 3 3 6 -6.2165278E-05 4 3 6 1.2760972E-05 5 3 6 -6.2165278E-05 6 3 6 1.0270611E-04 1 4 6 2.77759259259E-06 2 4 6 -6.13445370370E-05 3 4 6 7.55775925926E-05 4 4 6 -1.15009259259E-05 5 4 6 7.55775925926E-05 6 4 6 -6.13445370370E-05 1 5 6 1.7820125E-04 2 5 6 1.0270611E-04 3 5 6 -6.2165278E-05 4 5 6 1.2760972E-05 5 5 6 -6.2165278E-05 6 5 6 1.0270611E-04 1 6 6 1.07580273148E-03 2 6 6 -8.68419537037E-04 3 6 6 -1.26556851852E-04 4 6 6 4.42628240741E-05 5 6 6 -1.26556851852E-04 6 6 6 -8.68419537037E-04 1 1 1 2 1 1 1 -8.74218916667E-03 2 1 1 -8.74218976551E-03 3 1 1 -2.35019119479E-03 4 1 1 -7.60084259259E-05 5 1 1 -7.60088052061E-05 ... -- next part -- An HTML attachment was scrubbed... URL: http://www.democritos.it/pipermail/pw_forum/attachments/20101009/1144fc38/attachment-0001.htm
[Pw_forum] Xspectra calculation
Dear Xspectra users, I met a problem when calculating the X-ray absorption spectra with the Xspectra code. Concern about the polarization, ?// is the polarization along the c axis, while ?? is the polariztion in the plane. It's easy to define the ?// , since normally (0,0,1) is the direction of c axis. But for ?? polarization, what does "in plane" mean? Are any directions in the a,b plane will cause the same result? For instance, will (1,0,0) ,(0,1,0) or (1,1,0) polarization direction have the same spectra? Thank you very much for any help. Cheers, Min Wu 2010-10-8 -- next part -- An HTML attachment was scrubbed... URL: http://www.democritos.it/pipermail/pw_forum/attachments/20101009/5fd9f157/attachment.htm
[Pw_forum] Some questions about IFC from PWSCF
Dear John Bob,
answer to 1. & 2:
IFC have four indices and depend on a lattice vector :
F_(alpha,beta)^(na,nb)(R)
this is the force in direction alpha (cartesian) on atom na in the unit
cell at the origin when atom nb in the unit cell identified by R is
displaced in direction beta.
The header of each block (the four integers 1 1 1 1 for instance) etc.
are just alpha,beta,na,nb and the subsequent lines contain
i j k F_(alpha,beta)^(na,nb)(R)
where R=(i-1)*a + (j-1)*b +(k-1)*c.
There are 6x6x6 lines in your case because your q-points grid had these
dimensions and a grid of q in the BZ translate in a set of R in real space.
In order to compare individual IFC you must take into consideration the
fact that they are tensorial quantities (matrices) and not scalar
For instance if the system is cubic you expect
F_{xx}{11} (0) = F_{yy}{11}(0) = F_{zz}{11} (0) = something
while
F_{xy}{11}(0) = F_{yx}{11}(0) =F_{yz}{11}(0) =F_{zy}{11}(0)
=F_{xz}{11}(0) = F_{zx}{11}(0) = 0
where the R=0 is the first element (i,j,k=1,1,1) after the header of
each block (alpha,beta,na,nb)
you should have
F_{xx}{11}(2 1 1) = F_{yy}{11}(1 2 1) = F_{zz}{11} (1 1 2)
and also
F_{xx}{11}(2 1 1) = F_{xx}{11}(6 1 1) = F_{yy}{11}(1 6 1) = F_{zz}{11}
(1 1 6)
for your 6x6x6 grid.
and many other relations that links rotations of the linking vector R
and rotations of the alpha,beta indices
answer to 3:
The written vectors are normalized displacements i.e. are eigenvectors
rescaled with the sqrt of the atomic mass and remormalized.
In case of doubts the routine that does that is PH/dyndia.f90 and the
corresponding section (look for "Diagonalizing the " in dyndia.f90)
should be quite transparent
stefano
===
John Bob wrote:
> Dear Pwscf users:
> I have some questions about IFC from pwscf. I have read some threads
> about interatomic force constants in PWSCF maillist. I get confused when
> trying to analysis the IFC of NiTi alloy, and also the example06 AsAl.
>
> 1. F(R) and R=(i-1)*a + (j-1)*b +(k-1)*c. It means F(1 2 1) should equal
> to F(1 1 2) and F(2 1 1) for cubic lattice. But in B2 NiTi alloy ( CsCl
> strucuture) F(1 1 2) = F(1 2 1) /= F(2 1 1). and in example06 F(2 1 1) = F(1
> 1 2) /= F(1 2 1). It seems some extra-things added to some direction. Why?
>
>2. How to interpret the "1 1 1 1" section, F_{xx}^{1,1}(0). Does it means
> the force along x axis on Atom1 when Atom1 has a small displacement along
> the x direction?
>
> 3. The section after the "diagonalizing the dynamical matrix". I find the
> vectors in xx.dyn file xx.vec file (after matdyn.x) are different, rotated.
> Are these vectors eigen-displacements or just eigen-vector after
> diagonalizing?
>
> Best regards
>
> Dr. Zhou Bo
> Northwest University, Xi'an China
>
> NiTi.fc
>
> ...
>6 6 6
>1 1 1 1
>1 1 1 4.30811843056E-02
>2 1 1 1.49655475926E-02
>3 1 1 7.4726639E-04
>4 1 1 -2.16783699074E-03
>5 1 1 7.4726639E-04
>6 1 1 1.49655475926E-02
>1 2 1 2.6899633E-03
>2 2 1 6.6021083E-04
>3 2 1 3.8196056E-04
>4 2 1 -2.6103722E-04
>5 2 1 3.8196056E-04
>6 2 1 6.6021083E-04
>1 3 1 -6.843E-06
>2 3 1 -6.65476851852E-05
>3 3 1 -2.7346667E-05
>4 3 1 -1.06576851852E-04
>5 3 1 -2.7346667E-05
>6 3 1 -6.65476851852E-05
>1 4 1 4.4854861E-05
>2 4 1 -1.1179139E-04
>3 4 1 -1.0616667E-05
>4 4 1 -2.2030694E-05
>5 4 1 -1.0616667E-05
>6 4 1 -1.1179139E-04
>1 5 1 -6.843E-06
>2 5 1 -6.65476851852E-05
>3 5 1 -2.7346667E-05
>4 5 1 -1.06576851852E-04
>5 5 1 -2.7346667E-05
>6 5 1 -6.65476851852E-05
>1 6 1 2.6899633E-03
>2 6 1 6.6021083E-04
>3 6 1 3.8196056E-04
>4 6 1 -2.6103722E-04
>5 6 1 3.8196056E-04
>6 6 1 6.6021083E-04
>1 1 2 2.6899633E-03
>2 1 2 6.6021083E-04
>3 1 2 3.8196056E-04
>4 1 2 -2.6103722E-04
>5 1 2 3.8196056E-04
>6 1 2 6.6021083E-04
>1 2 2 1.07580273148E-03
>2 2 2 -8.68419537037E-04
>3 2 2 -1.26556851852E-04
>4 2 2 4.42628240741E-05
>5 2 2 -1.26556851852E-04
>6 2 2 -8.68419537037E-04
>1 3 2 1.7820125E-04
>2 3 2 1.0270611E-04
>3 3 2 -6.2165278E-05
>4 3 2 1.2760972E-05
>5 3 2 -6.2165278E-05
>6 3 2 1.0270611E-04
>1 4 2 2.77759259259E-06
>2 4 2 -6.13445370370E-05
>3 4 2 7.55775925926E-05
>4 4 2 -1.15009259259E-05
>5 4 2 7.55775925926E-05
>6 4 2 -6.13445370370E-05
>1 5 2 1.7820125E-04
>2 5 2 1.0270611E
[Pw_forum] Xspectra calculation
Dear Wumindt2, the fact that different polarization directions can be equivalent or not in a solid depends only on the symmetry of your lattice (thus on the crystal structure) only. So the general response to your question is given by the theory of crystal symmetries. M. > -- > > Message: 1 > Date: Sat, 09 Oct 2010 11:11:16 +0800 > From: "wumindt2" > Subject: [Pw_forum] Xspectra calculation > To: pw_forum at pwscf.org > Message-ID: > Content-Type: text/plain; charset="gb2312" > > Dear Xspectra users, > > I met a problem when calculating the X-ray absorption spectra with > the Xspectra code. > Concern about the polarization, ?// is the polarization along the c > axis, while ?? is the polariztion > in the plane. It's easy to define the ?// , since normally (0,0,1) > is the direction of c axis. > But for ?? polarization, what does "in plane" mean? Are any > directions in the a,b plane > will cause the same result? For instance, will (1,0,0) ,(0,1,0) or > (1,1,0) polarization direction > have the same spectra? > > Thank you very much for any help. > > Cheers, > > Min Wu > 2010-10-8 > -- next part -- > An HTML attachment was scrubbed... > URL: > http://www.democritos.it/pipermail/pw_forum/attachments/20101009/5fd9f157/attachment.html > > -- > > ___ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > > > End of Pw_forum Digest, Vol 40, Issue 16 > > This message was sent using IMP, the Internet Messaging Program.
[Pw_forum] Xspectra calculation
Dear matteo, Thanks. For dichroism compound, do we just need the in-plane and the c-axis polarization vectors? While for trichroism compound, do we need to calculate three different polarization vectors? Am i right? Best regards, Min Wu 2010-10-9 >From: matteo calandra >Reply-To: PWSCF Forum >To: pw_forum at pwscf.org >Subject: Re: [Pw_forum] Xspectra calculation >Date: Sat, 9 Oct 2010 11:30:54 +0200 > >Dear Wumindt2, > > the fact that different polarization directions can be equivalent or not >in a solid depends only on the symmetry of your lattice (thus on the crystal >structure) only. >So the general response to your question is given by the theory of crystal >symmetries. > >M. > > >> -- >> >> Message: 1 >> Date: Sat, 09 Oct 2010 11:11:16 +0800 >> From: "wumindt2" >> Subject: [Pw_forum] Xspectra calculation >> To: pw_forum at pwscf.org >> Message-ID: >> Content-Type: text/plain; charset="gb2312" >> >> Dear Xspectra users, >> >> I met a problem when calculating the X-ray absorption spectra with >> the Xspectra code. >> Concern about the polarization, ?// is the polarization along the c >> axis, while ?? is the polariztion >> in the plane. It's easy to define the ?// , since normally (0,0,1) >> is the direction of c axis. >> But for ?? polarization, what does "in plane" mean? Are any >> directions in the a,b plane >> will cause the same result? For instance, will (1,0,0) ,(0,1,0) or >> (1,1,0) polarization direction >> have the same spectra? >> >> Thank you very much for any help. >> >> Cheers, >> >> Min Wu >> 2010-10-8 >> -- next part -- >> An HTML attachment was scrubbed... >> URL: >> http://www.democritos.it/pipermail/pw_forum/attachments/20101009/5fd9f157/attachment.html >> >> -- >> >> ___ >> Pw_forum mailing list >> Pw_forum at pwscf.org >> http://www.democritos.it/mailman/listinfo/pw_forum >> >> >> End of Pw_forum Digest, Vol 40, Issue 16 >> >> > > > > >This message was sent using IMP, the Internet Messaging Program. > > >___ >Pw_forum mailing list >Pw_forum at pwscf.org >http://www.democritos.it/mailman/listinfo/pw_forum >
