Hi Everyone, I agree with Andrei comment below, the Ef generated by SIESTA cannot be used to calculate the work function. To determine the work function you need to apply the "Bulk Plus Band LineUp" method. For details please see "Van de Walle and Martin, 1987, Phys Rev B, no. 35, pp.8154". The method involves solving bulk system and a slab to pin down the relative energy between the vacuum and Ef (as Andrei mentioned below "a difference energy"). In general DFT and I believe SIESTA can reproduce the work function of metals with good accuracy (with 5% or so).
Thanks Ahmed Ahmed Huzayyin PhD Candidate ECE Department University of Toronto ________________________________ From: "[email protected]" <[email protected]> To: [email protected] Sent: Sat, December 25, 2010 12:04:56 PM Subject: Re: Re: [SIESTA-L] Metal Fermi Levels > Dear apostnik, > Thanks for your instruction on this issue. > As you have said,we should find out the vacuum energy and then use the got > Fermi level from SIESTA to subtract it and this is the work function. > > Am I right? This is not exactly what I wanted to say; anyway I am not an expert in calculating work functions. There should be enough works done on this subject before, not necessarily with Siesta. In any case, the problem should reduce to extracting some energy difference between reasonably selected reference cases. This granted, the absolute numbers of these energies would not play a role. I intended to comment only on this narrow issue, being worried that the colleagues tended to take an absolute value of Ef, provided by Siesta, for a hint whether the calculation is good or bad. In fact you cannot decide based on this information only. Best regards Andrei Postnikov > > Thanks a lot. > > 2010-12-25 > > > > Guangping Zhang > > > > 发件人: [email protected] > 发送时间: 2010-12-25 15:46 > 主 题: Re: [SIESTA-L] Metal Fermi Levels > 收件人: [email protected] > > > > Dear BR/Yujia / Ronaldo, > it must have been discussed already a couple of times, but still: > The absolute value of the Fermi energy as it comes out of the calculation > is largely method-dependent. It makes sense to discuss its correct / wrong > position with respect to the band structure, but not relative to > zero level of energy, that is, typically, a somehow > averaged value of some potential. (For instance, in a calculation with > a FLAPW method, a "typical" Fermi energy might be +6 eV or so, so what?). > > In order to get the work function, you need to make comparison > with the potential in the vacuum far away from your sample. > The values from bulk calculation don't tell you much about this, > so be careful. > > Best regards > > Andrei Postnikov > > >> Dear BR/Yujia, >> >> The values of IP and EA, and of the work function as a consequence, is >> sensitive to the confinement imposed on basis orbitals. Using the >> default >> value of energy shift (0.01 Ry) >> I also obtained a value of fermi level for Au(111), -4.1eV, which >> greatly >> differs from those previously reported. I have obtained a better result, >> -5.1eV, decreasing the value of energy shift to 0.001 Ry. As it is >> possible >> to see in Nanotechnology *21* (2010) 065705, the value of Au(111) fermi >> level converges to that obtained using plane waves (-5.25eV) in the >> limite >> of very small basis set confinement. >> >> Best, >> >> Ronaldo >> >> On Fri, Dec 24, 2010 at 5:33 AM, zjuyangyj <[email protected]> wrote: >> >>> Dear all, >>> Recently I am calculating the work functions or fermi levels of >>> different >>> metals. However, my results seem to be incorrect. For example, I got >>> the >>> fermi level for Au as -3.68eV, which greatly differs from previous >>> theoretically or experimentally reported values, which range from >>> -4.8eV >>> to >>> -5.3eV. Also I got the fermi level for Pt as -3.99eV, also largely >>> differing >>> from correct values. >>> >>> I want to know what might be the possible reasons for incorrect fermi >>> levels, and when calculating fermi levels, what are the >>> important/relevant parameters? I am calculating very thin metal films >>> (tens >>> of atom layers), and I found the number of layers an important >>> parameter. >>> Also I want to know whether the fermi levels of bulk and thin film >>> metal >>> differs a lot. >>> >>> Thank you for your attention. >>> >>> BR/Yujia >>> >>
