some questions about the new version

[John]

Dear everyone,
    For the new version, I have some questions that,
(1)Monkhorst-Pack Grid for the calculations on DOS and Transmission spectrum: there are only two indexes (KA and KB) in the vnl panel, while the input file always includes three indexes, like that kpoints=MonkhorstPackGrid(n, m, 1),  whether the third “1” needs to be modified.
(2)It seems that the function “Restore SCF calculation and Only use initial density” is difficult to achieve, like that in 2008.10, how to calculate the SCF and Transmission spectrum under high bias using the low bias initial density, or restore calculation?
(3)If choosing the method SGGA+U or LSDA+U, how to decide the U values. Why I only find two kinds of normconserving pseudopotentials LDA.PZ and GGA.PBE, although I have tested one job using SGGA.PW91 exchange correlation, and there are not errors happened.  

Thank you!

[zh]

1). For the two-probe system, the transport direction is always assumed to be along the z direction. The periodicity exists in the xy plane perpendicular to the transport direction, the size of cell  in this plane usually smaller, and thus k-point sampling technique is used. When the DOS and transmission spectrum of such system are calculated,  basically only 'n' and 'm' in the kpoints=MonkhorstPackGrid(n, m, 1) should been specified reasonably.


3). Although some groups have proposed some computational methods to determine the U value via ab-initio calculations, these techniques are not implemented in the current codes. For the choice of U, one usually tests several U values and then chooses a reasonable U value that can produce some physics properties (such as lattice properties and electronic structures) being closer to the experimental results. Alternatively, you may simply take a value that has been widely used in literature by other people. 

[Anders Blom]

2) For an example on this, see http://quantumwise.com/documents/tutorials/BenzeneSET/XHTML/chap.SET.html#sect1.charged.gatepotential

3') The label LDA.PZ indicates which functional that was used to generate the pseudopotential, but there is usually no reason to make separate pseudopotentials for the different flavors of LDA or GGA, and thus you can use the GGA.PBE pseudpotential with other GGA functionals as well.