why is my density of states not smooth？

[wot19920302]

    Dear scholars, I have a question about ATK (ver.2015). Now I want to look at the hybrid configuration at spinterface between BDT molecule and Nickle. And I used slab model to simulate interface, which is subjected to present conditions. The trouble matter is as follows:
When I calculate density of state of BDT contacting on Nickle, the image is not smooth.
    I used builder to build its slab model, and then I optimized the bulk and added the analysis of density of state to script.Please look at script1.And follow this I calculate its density of state, but the result is not good. The image of DOS is not smooth prospectively. Please look at image1. I try many methods to change this situation, trying to increase k-polint sampling to 9*9*1, increasing vacuum size to 20, I just hope the DOS of energy-independece image be smooth. However I didn’t find the real reason to avoid this. What good method could solve this trouble?Being more smooth by this slab simulating?
      Best
 

[Jess Wellendorff]

Gaussian broadening of the DOS is needed if you want it to be smooth. In the ATK Manuals you will find an example of how to plot the DOS using a Gaussian spectrum with e.g. 0.2 eV broadening:
https://docs.quantumwise.com/manuals/Types/DensityOfStates/DensityOfStates.html#usage-examples

[Petr Khomyakov]

Another strategy would be to significantly increase the metal slab thickness, or even a better option would be to use Surface Configuration in which the metal is represented with a semi-infinite structure, see https://docs.quantumwise.com/tutorials/gf_surface/gf_surface.html.

I would like to notice that some peaks may correspond to localized or quasi-localized states, and in this case they should be sharp, meaning that the corresponding DOS is not smooth.

[wot19920302]

thanks for your help！