Miguel, thanks for getting back so quickly. Actually, I use what the ph.x docs call "line of input" feature failrly often to do "quicky" calculations of the "gamma point" e-p interaction. I'm familiar with example 2, but note, for Si, G and X are two separate calculations. I don't think you can specify two different q-points, e.g., at -kf, +kf as separate lines of input (I could be wrong). What I sort of have in mind is whether one could somehow specify as input to ph.x a "q-vector" spanning the points -kf to kf and calculate the net e-p coupling. For example, suppose kf = (0.25, 0.25, 0)(2pi/a units) and 2kf "connects" two mostly planar Fermi surfaces on opposite sides of the BZ. Physically, this an extremely unstable situation would result in a CDW phase transition doubling the unit cell (dimerization) and gapping the Fermi surface and should in principle yield a very large e-p coupling constant beforehand. I suspect ph.x, at present, doesn't have the capability to simulate this situation.
Right now, I have a run going with a "line of input" for q(or kf) = (0.37, 0.37, 0) but it's taking a long time to converge. I will also try later 2kf (0.74, 0.74, 0) just to see what happens. Note these are not "high symmetry" points. Again, thanks for your feedback. -Paul -----Original Message----- From: pw_forum-bounces at pwscf.org [mailto:[email protected]] On Behalf Of Miguel Mart?nez Sent: Sunday, March 06, 2011 3:22 AM To: pw_forum at pwscf.org Subject: Re: [Pw_forum] ph.x and Nesting Fermi Surfaces Dear Paul, If I understood you correctly, you want to run a ph.x calculation at a q of your choice such that q=2kf, right? In that case, you only need to specify ldisp =.false., not specify any nq1, nq2, nq3 and *after* the slash in the ph.x input, specify the q points at which you want the calculations done. Example 02 has ph.x calculate the phonon frequencies at X (and gamma) for Fm-3d Si and C, and fcc Ni at X. Hope it helps, Miguel On Sat, 5 Mar 2011 18:08:30 -0800 "W2AGZ" <w2agz at w2agz.com> wrote: > To All: > > > > I'm running some exercises on electron-phonon interactions in narrow > band metal monolayers (actually tetragonal unit cells with "large" c/a > parameters...and think doped transition metal oxides). In this > quasi-2D system, several of the Fermi surfaces approach planarity > (actually, "Fermi lines"), suggesting strong electron-phonon > interaction should be driven by the "nesting" wave vector (think CDW > and/or Peierls-Froehlich). > > > > However, I'm having difficulty seeing how to structure the input to > ph.x to select this wavevector alone (or indeed if this is even > possible with ph.x). Imagine a square-planar BZ with dimensions (0.5, > 0.5) in 2pi/a units, with a "Fermi line" perpendicular to, and kf > distance along the (1,1) direction from the BZ origin. Now, as far as > I can see from the ph.x documentation, nq1, etc., is only for defining > the Monkhorst-Pack grid for the desired phonon grid-mesh (and are > integers anyway). In principle, one could set the "line of input" > xq's to the desired kf (fractional values are presumably allowed), but > which would have its origin at BZ center, whereas one would like to > "drive" or "excite" the phonons at -kf, +kf together. > > > > I hope I've made the issue/question clear and haven't said anything > egregriously stupid. > > > > Thanks in advance, > > > > Paul Grant > > IBM Research Staff Member Emeritus > > Visiting Scholar, Stanford > -- ---------------------------------------- Dr. Miguel Martinez Canales Department of Physics & Astronomy University College London Gower Street, London, WC1E 6BT (UK) Fax: +44 207 679 0595 Tlf: +44 207 678 3476 ---------------------------------------- "If you have an apple and I have an apple and we exchange these apples then you and I will still each have one apple. But if you have an idea and I have an idea and we exchange these ideas, then each of us will have two ideas." George Bernard Shaw _______________________________________________ Pw_forum mailing list Pw_forum at pwscf.org http://www.democritos.it/mailman/listinfo/pw_forum
