Hi, Den søn. 31. mar. 2019 kl. 22.02 skrev Rodrigo Garcia Amorim < rgamo...@id.uff.br>:
> Dear developers, > > I saw in the manual that is possible calculate the phonon > transmission. Regarding this topic, my questions are: > > 1) Could I use similar setup of transport calculation? > Yes and no. Generally it would be extremely costly to perform phonon transmission calculations using siesta since it requires a huge system to ensure the phonons are "relaxed" in the electrode region. This is why we (in the paper) did phonon transmissions using GULP which can calculate the dynamical matrix. > > 2) For electrodes I need run FC with transiesta to obtain the dynamic > matrix and frequencies? > If you wish to use siesta (transiesta is not needed) then yes, you need the electrodes in a big enough cell such that you can decouple the cell interactions. The problem is that phonons are long ranged and thus intrinsically requires a very large system to get correct dynamical matrices. I.e. for the graphene system in the paper we used a system of 128 atoms which reduces to an 8 atom big electrode and 32 big device region. Since this is for the GULP MD scheme you may need an even bigger system in siesta. > > 3) For scatter region I also need calculate using FC and transiesta? > See above. > > 4) In the paper (Improvements on non-equilibrium and transport Green > function techniques: The next-generation transiesta) > > The authors claim that is possible use Nc>=1. For Nc=1. Should I consider > just left electrode? > Do you mean number of electrodes? N_e? If so yes. You can do as many (or as few as 1) electrodes. If you want transmission you need at least 2 electrodes. > > > Thanks > > Rodrigo > > -- > -- > Dr. Rodrigo G. Amorim > > =================================== > Universidade Federal Fluminense - UFF > Departamento de Física, ICEx > > R. Des. Ellis Hermydio Figueira, 783 > Aterrado, Volta Redonda - RJ, Brazil > 27213-145 > --------------------------------------------------------------- > webpage : http://rgamorim.wix.com/rodrigo > e-mail : rgamo...@id.uff <rodrigo.amo...@physics.uu.se>.br > ==================================== > -- Kind regards Nick
