On Fri, 2008-06-06 at 13:59 -0600, M Bharat Kumar wrote: > Hi Siesta users, > > I need help in making siesta work with fractional electrons in valence > orbitals. I am attaching input for atom program and input & output of > siesta. I was not able to understand the warning messages in siesta > output. > Any help is appreciated and thanks in advance. > > Bharat
I'm not sure what you mean by fractional electrons ... do you mean you want to set NetCharge in your system to some non-integer (which is totally allowed and I do it all the time), or some kind of excited charge-neutral state with a fraction of an electron excited? Anyway, I made a different pseudopotential and your input .fdf worked, so the problem is most likely in your pseudo definition. Here's what I tried: pe Gold tm2 4.0 n=Au c=pbr 0.0 0.0 0.0 0.0 0.0 0.0 12 4 6 0 1.00 0.00 6 1 0.00 0.00 5 2 10.00 0.00 5 3 0.00 0.00 2.00 2.50 1.00 2.00 1.00 1.40 #23456789012345678901234567890123456789012345678901234567890 I changed: * c=pbr (was c=pb, you probably want relativistic corrections enabled for gold). * Changed the population of the valence orbitals (you might not need to mess around with exciting the pseudopotential populations to get a decent excited state calculation, your mileage may vary and I've never calculated gold before, let alone in bizarre charge states). * Messed around with the r_c values. Your pseudofunctions were overshooting the all-electron charge density peaks. ATOM produces .gps files which you can feed into gnuplot to get .ps files with pretty graphs showing how well each L channel in your pseudo corresponds to their all-electron analogs. I picked new r_c values from these graphs. Any combination of these changes may be what you need to make your calculation work, so good luck with that. Edan.