Dear Marcos, I have problem with position of fermi level in density of states profile? .EIG and .DOS files have shifted energy scales, perhaps in .EIG file fermi level is shifted to zero of energy, but not in .DOS file.? Where/which file should i look to *change/shift the fermi level*?
Also, in spin polarized calculations .EIG file gives dos for both spin states in same direction, i mean, states overlap with each other. How can i get the *dos for spin up and down separately using .EIG file*, although i can get it using .DOS file, but there fermi level is not shifted to zero of energy ? With Warm Regards, Both .EIG and .DOS files give the profile of density of states in a required energy window and with accuracy 2010/4/12 Marcos Veríssimo Alves <marcos.verissimo.al...@gmail.com> > Guangping: > > Could you please explain with a bit more detail what you mean by "peak > width" in the DOS and "different profiles"? I suppose I have an idea of what > "peak width" means for you, but I have no idea of what a "DOS profile" could > be. Plus, what is the problem with the code not reading the .KP file? > K-point generation in siesta is automatic - I am not sure that the inclusion > of arbitrary k-points for sampling in the BZ has been implemented in siesta. > Since this part of the calculation is very quick, I guess no one bothered > implemented an option for reading it. > > Gregorio: > > If you don't need a PDOS, you can simply generate a DOS from the EIG file > at the end of the calculation, which is cheaper than calculating the PDOS. > Of course, your sampling should be large enough for you to have a dense set > of eigenvalues for a smoother DOS. The advantage of this is that generating > the DOS for arbitrary values of the gaussian smearing is cheaper, after the > calculation is over. > > The eigenvalues will come out correct no matter what origin of coordinates > you set for your system. In fact, your atomic coordinates need not be set > within the cell limits; as long as the relative positions and orientation of > the atoms with respect to the cell vectors are correct, everything is okay. > The block Atomic CoordinatesOrigin shifts the origin of atomic coordinates > in real space, and this does not influence at all the determination of > eigenvalues. It might be useful only for later visualization purposes; > otherwise, it should have no influence whatsoever in anything in your > calculation. > > Marcos > > 2010/4/12 Gregorio García Moreno <gjgar...@ujaen.es> > > Hi, >> The data needed to plot the DOS profile are in the DOS file. >> I am not sure, but I think that obtain a correct profiel with eig2dos and >> EIG file >> depended of the atomic origin of coordinates, i.e.: >> >> %block Atomic CoordinatesOrigin >> 7.720577 0.014252 0.743077* >> %endblock AtomicCoordiantesOrigin >> >> *Wich must coincide with a kind of recciprocal laticce, >> But I am not sure. >> If you take the data found in DOS file, this profile its correctly, >> regardless the origin. >> I always use DOS file. >> >> I hope help you >> >> >> El 12/04/2010 10:32, Guangping Zhang escribió: >> >> Dear SIESTA user: >>> In a siesta run ,a file *.EIG can be generated,and we can get the DOS >>> from it using eig2dos!But how to define the peak width,since different >>> widths,the DOS have different profiles! >>> Who can tell me ? >>> In additon, the code in not read the *.KP to define the k weight. >>> Thanks in advance! >>> >> > -- Sonu Kumar Phd Student Physics Department Indian Institute of Technology Delhi Delhi-110016 web:-http://www.iitd.ac.in/
