ABINIT Discussion Forums

I know this is a kinda stupid question but I can't really understand how to obtain wavevectors in reciprocal space in abinit. The help saying:

List of nph1l wavevectors, at which the phonon frequencies will be interpolated. Defined by 4 numbers: the wavevector is made by the three first numbers divided by the fourth one (a normalisation factor). The coordinates are defined with respect to the unit vectors that spans the Brillouin zone. Note that this set of axes can be non-orthogonal and not normed. The normalisation factor makes easier the input of wavevector such as (1/3,1/3,1/3), represented by 1.0 1.0 1.0 3.0 .
The internal representation of this array is as follows : for each wavevector, the three first numbers are stored in the array qph1l(3,nph1l), while the fourth is stored in the array qnrml1(nph1l).

But how to find out wich coourdinates in reciprocal space abinit use?

To calculate rf1 I'm using theese lines in input file:

acell 9.1 9.1 12.3

rprim 1.0000000000E+00 0.0000000000E+00 0.0000000000E+00
0.0000000000E+00 1.0000000000E+00 0.0000000000E+00
3.6672160342E-01 3.6672160342E-01 8.5500323460E-01

Since then I'am able to calculate primitive vector (rprimd). But what next? I have to calculate unit vectors in reciprocal space?

I'm attached BZ for the stucture and there is two set of vectors. Axes (k_x, k_y, k_z) -- blue ones, and (b1, b2, b3) -- red ones. Wich one abinit use?
For example if I need X point? I should use qph1l=(1,1,0) or (0,1,0)? And for Z point qph1l=(0,0,1) or (1,-1,1)?

Hi!

Please, search the forum. There was a post on the general method how to produce a set of qpts.

Igor L.

Thank you Igor! I search Post-processing Tools and respons calculations brances of forum and found two discussions connected whith my topic. The first one pointed to tutorial rf2 http://www.abinit.org/documentation/hel ... rf2.html#1
And the second one is:
viewtopic.php?f=10&t=1347 (qph1l phonon dispersion anaddb)

What can I say. Both have no answer to my question. What I am asking is wich directions abinit use? Please see fig I attached to the first post. The point Z have a coordinate (001) and the vectors on wich Brilluen Zone based are orthogonal (marked blue) this is widly used in other programs. But abinit uses non ortogonal primitive vectors for bodycentered tetragonal lattice in real space (rpirimd matrix I've pointed in first post) thus I don't know how abinit constructs base vectors in reciprocal space. Are they blue ones, or red ones in Fig?

In case someone comes across this page again, the reciprocal lattice vectors in Abinit are defined in the usual way (as in for example (5.3) of Ashcroft and Mermin). You can see this in Lattice.F90 for example (search within the Abinit folder).

Edit: I may be wrong. The reciprocal vectors given in the output files don't have a 2 Pi, as hinted at in http://www.abinit.org/doc/helpfiles/for ... vargs.html, Hexagonal Lattices.

Question: So the reciprocal vectors are given by k = 2 Pi (k1 G1 + k2G2 + k3G3), with an extra 2Pi with respect to http://www.abinit.org/doc/helpfiles/for ... s.html#kpt ??