Dear all, I guess that this string of emails calls for some clarification.
The SIESTA team together with myself worked for some time to have the SOC facility implemented into the main trunk, but for a variety of reasons the project never reached the desired end. As a consequence, I have two private versions of SIESTA+SOC that I am happy to share with anyone having a SIESTA license. This comes under an old agreement between the SIESTA authors and me. I believe that perhaps these versions could be made publicly available in the SIESTA Web site, provided that there is a critical mass of plausible users, but this decision is beyond my reach. The two versions have each advantages and disadvantages, that I summarize: 1. Serial version, based on siesta-1.4. This one is fully tested, and works with LDA and GGA, either non-magnetic, collinear magnetic and non-collinear magnetic. It also computes orbital moments. However, as far as I remember, it does not output SOC bands - although it would be easy to fix this. 2. Parallel version, based on a beta release of siesta-3.0. This one is fully tested for LDA: non-magnetic and any sort of collinear and non-collinear magnetic systems. For GGA, it works only for non-magnetic systems. It provides wrong results for non-collinear magnetic systems. This version outputs also bands, DOS and I kind of remember that also wave-functions. Again orbital moments are computed and written. Two warnings are in order now: 1. The SIESTA-SOC version converges slowly, and swallows more RAM resources than conventional simulations. 2. You have to test carefully your pseudopotential parameters, because ghost states may appear. However, a new open-access paper has just appeared where Pablo Rivero, Salvador Barraza-Lopez, myself and collaborators have tested thoroughly the pseudo potentials of 20 elements with focus on those having strong SOC -having in mind topological insulators -. We have used for testing the parallel SOC version. We have benchmarked the SIESTA results against VASP and in quite a few cases against an all-electrons code, and the pseudos work quite well. I would advise those interested in SOC to use any of the above two flavors together with the pseudos written down in the paper. The reference to the paper is: http://www.sciencedirect.com/science/article/pii/S0927025614007940# <http://www.sciencedirect.com/science/article/pii/S0927025614007940#> A last warning: I am no longer maintaining the SIESTA-SOC. What I can hand is what I wrote up to 2009. OK. I guess that this is all for now. With kind regards, Jaime Ferrer > El 8/12/2014, a las 18:05, I. Camps <[email protected]> escribió: > > Hello, > > Since 2006, in the paper bellow there is an implementation of the spin-orbit > coupling in SIESTA. Unfortunately, until now, I didn´t see this implemented > in any version of the code. > > Paper: On-site approximation for spin–orbit coupling in linear combination of > atomic orbitals density functional methods, L Fernandez-Seivane, M A > Oliveira, S Sanvito and J Ferrer, J. Phys.: Condens. Matter 18 (2006) 7999 > <tel:18%20%282006%29%207999>–8013 > > Abstract > We propose a computational method that drastically simplifies the inclusion > of the spin–orbit interaction in density functional theory when implemented > over localized basis sets. Our method is based on a well-known procedure > for obtaining pseudopotentials from atomic relativistic ab initio calculations > and on an on-site approximation for the spin–orbit matrix elements. > We have implemented the technique in the SIESTA (Soler J M et al 2002 > J. Phys.: Condens. Matter 14 2745–79) code, and show that it provides accurate > results for the overall band-structure and splittings of group IV and III–IV > semiconductors as well as for 5d metals. > > DOI: http://dx.doi.org/10.1088/0953-8984/18/34/012 > <http://dx.doi.org/10.1088/0953-8984/18/34/012> > > > []'s, > > @mps > > On Mon, Dec 8, 2014 at 2:52 PM, Suman Chowdhury <[email protected] > <mailto:[email protected]>> wrote: > Yes, I completely agree with you. As far as I know, spin-orbit coupling is > not included in siesta code. Even I posted about this issue in past in siesta > forum. > > On Mon, Dec 8, 2014 at 10:01 PM, Seyed Mohammad Tabatabaei <[email protected] > <mailto:[email protected]>> wrote: > Dear SIESTA Users, > > I want to implement spin orbit coupling for mono-layer MoS2 but as far as I > know this is not implemented directly in the code. Any help or ideas in this > regard is appreciated. > > Kind regards, > Mohammad, > > > > -- > Senior research fellow > Dept. of Physics, > University of Calcutta > Kolkata- 700009, West Bengal, India. > Ph no-+91-9830512232 <tel:%2B91-9830512232> > >
