Dear Riccardo, I think this question still needs to be considered and discussed. If you look closer into the output file of the scantering region, you will find the third number is always 1.But in this case,it is periodical in the z direction in the siesta part. I think the report is aimed at the siesta calculation. So the 1 for the third number I think can not exactly equal to there is no interactions between the next nearest ones. I think this is much stricter.
Regarding the discussion about the number and the effect on the convergence you have given, I want to say a word. I guess transiesta in default use a more general routine to do the diagonalization for Green's function. But as you say, if the third number is 1 (I still think this is not necessary) the matrix will be a exact diagonal matrix and the most important there is no interactions between the two electrodes then you can without worry use the option "TS.TriDiag T". Usually, a much faster performance and good convergence will be seen and the results between with or without this option are hardly any difference. Please feel free to comment on all what I have said. Guangping 2012-01-08 Guangping Zhang 发件人：Riccardo Rurali 发送时间：2012-01-06 03:37 主题：Re: [SIESTA-L] Transiesta: length of electrode unit cell 收件人：siesta-l 抄送： Dear Guangping, ----- Original Message ----- Would you like to explain me what the "Naive supercell factors" and "Internal auxiliary supercell" actually mean? They are puzzling me a lot. I don not think 2 mean there is only interactions between the nearest unitcells. Like what Yunpeng said, he got 1. Is that mean there is not any interactions between the nearest unitcells? That is definitely not. So I wondering what their exact meaning. I have searched much, but have few results. I don't know about Naive supercell factors (actually I never noticed the line so far...), but in the case of Internal auxiliary supercell I am prety sure that this is exactly what it means. If you run a calculation for an isolated atom or molecule, so that "System type = molecule", the Internal auxiliary supercell is 1 x 1 x 1. If Yunpeng has a 1 as third number, my guess is that the coordinates (or most lilely the lattice vectors) of his electrode are not correct. Riccardo
