Calculating DOS cp2k tutorial - Density of states wrong?

[Andreas Döll]

Hello,

I want to calculate dos and bandstructure for my research, so in preparation for my system I wanted to follow the cp2k tutorial to calculate the (total) PDOS and afterwards the band structure of WO3, because I am unfamiliar with such calculations.

 

I used the input file for the exercise (linked here: https://www.cp2k.org/exercises:2017_uzh_cmest:pdos). I obtained the pdos files for W and O respectively. Using the python script developed by Tiziano given in the exercise I convoluted BOTH files and summed up for the total PDOS (python pdos_conv.py WO3-pdos-k1-1.pdos WO3-pdos-k2-1.pdos --total-sum  --output WO3-smoothed-total.dat). Before continuing to the band structure of the lattice I noticed that my DOS looks different than the DOS given in the cited paper https://pubs.acs.org/doi/full/10.1021/cm3032225 for cubic WO3 (also the picture for the DOS is in the exercise). 

I tried playing around with the basis set or even replicating to 4 4 4 instead of 2 2 2, also changing the sigma value from the script, which was 0.02 by default, but the shape of my DOS is fundamentally different than the DOS in the paper. I know that in the paper they shift by E_VBM instead of E_F but I cannot find the desired shape in my data. For visualization after convolution I just plot energy vs sum all all corresponding states.

I am a bit lost, because I am unfamiliar with this. Attached is the mostly unchanged input script from the exercise with replicate 4 4 4 and the resulting pdos zoomed into the Fermi region with the flag -s 0.007 for smearing, just to show that smoothening is not the issue (I think). I also provide the original, unsmoothed dos files, as well as the DOS from the paper.

Is this kind of deviation expected? Using different basisset/smoothening/maybe input structure? Any comments or hits are appreciated! 

Cheers, 

Andreas Doell