This is a very good question (might actually consider it for a future online tutorial or something similar). Let me explain as plainly as possible:
As an example, I have run a fast device calculation for a carbon nanoribbon, and have done the DDOS, LDDOS, and PLDOS analyses in order to compare them. I also attach 4 images (A, B, C, D).
DeviceDOS: Image A. This is the device equivalent of ordinary density of states. The main difference from ordinary DOS (for bulks) is that spill-in terms from the electrodes must be taken into account. As the images shows, the DDOS results in a plot of the DOS against electron energy, just as with ordinary DOS, so the is no spatial information in the DDOS (i.e. "where in the system is the DOS large/small?") - we only know the total device DOS vs. energies, not vs. position.
LocalDeviceDOS: Images B and C. The LDDOS gives you the spatial information that is not included in the DDOS. The local DOS is simply calculated in small volume elements throughout all of the device central region. This gives you a GridValues type object which can be visualized in 3D real space using the Viewer. As you see in the images (B: cutplane, C: density), the LDDOS gives you the energy-dependent DDOS in real-space volume elements. (You may use the 1D Projector to get 2D plots, but that's not the point here).
ProjectedLocalDOS: The PLDOS simply takes the LDDOS and projects it onto the Z-axis of the device. This makes it possible to plot a contour plot of the energy-dependent DOS along the device transport direction. So the plot has energy and z-position on the axes, and the numerical value of the DOS is represented by contours.
I hope this makes sense. Thanks for asking an excellent question.
- Jess
