Real-space charge/spin current density

In the step 3, real-space charge/spin current density is calculated optionally. The parameter for this calculation is as follows:

NEGF.tran.CurrentDensity   on        #  default on

In the calculation of the current density, openmx makes the following standard output:

Start Calculation of the currentdensity

  Spin #0
    Sum of current in real space [a.u.] 
      Left(ideal) :    -9.10585e-06
      Right(ideal):    -9.10583e-06
      Left(truncated ):    -8.66971e-06
      Right(truncated):    -8.69926e-06
  Spin #1
    Sum of current in real space [a.u.] 
      Left(ideal) :    -4.54540e-08
      Right(ideal):    -4.54544e-08
      Left(truncated ):    -4.19469e-08
      Right(truncated):    -4.27460e-08

Output: Currentdensity
  Charge-current density along a-axis: ./negf-8zgnr-0.3.curden1.cube
  Spin-current density along a-axis: ./negf-8zgnr-0.3.scurden1.cube
  Charge-current density: ./negf-8zgnr-0.3.curden.xsf
  Spin-current density: ./negf-8zgnr-0.3.scurden.xsf
  Voronoi Charge-current density: ./negf-8zgnr-0.3.curden_atom.xsf
  Voronoi Spin-current density: ./negf-8zgnr-0.3.scurden_atom.xsf

In this case, 6 files,
negf-8zgnr-0.3.curden.xsf, negf-8zgnr-0.3.scurden.xsf,
negf-8zgnr-0.3.curden1.cube, negf-8zgnr-0.3.scurden1.cube,
negf-8zgnr-0.3.curden_atom.xsf, negf-8zgnr-0.3.scurden_atom.xsf,
are generated. These files contain the following quantities:

(Experimentally) When you set

    NEGF.OffDiagonalCurrent   on        #  default off
in a input file for the calculation with the non-collinear magnetism, the following files are outputted. These files contain the spin-offdiagonal component of the current density. Since this quantity becomes generally a complex number, the real part and the imaginary part of that is outputted separately.

For example, the currentdensity in 8-zigzag graphene nanoribbon with an antiferromagnetic junction under a finite bias voltage of 0.3 V in Fig. 34. In the vicinity of boundaries, that shows an unphysical behavior because the basis set is not treated correctly in these regions. If you want to calculate more precisely the current density in these regions, please make a supercell larger.

Figure 34: (a) The currentdensity in 8-zigzag graphene nanoribbon with an antiferromagnetic junction under a finite bias voltage of 0.3 V. (b) Isosurfaces of the a-compontent of that.
\begin{figure}\centering
\epsfig{file=CurDensity.eps,width=14.0cm}
\end{figure}

2016-04-03