[ase-users] Superimpose a structural model onto slices through a CHGCAR file

[Gaël Donval]

> Dear all,
> 
> I want to superimpose a structural model onto slices through a
> CHGCAR 
> file or differences of CHGCAR files.
> Theses slices are displayed with the contour plot from matplotlib.
> My first question is, if there is a script in existence that does
> this 
> which somebody would be willing to share?
There might be but I am not aware of it. Alternatively, you could try
to export your structure + density as a cube file and read it in
Paraview for instance. Paraview would show you the structure and allow
you to slice your data and do many more things.

But I am quite sure other tools can do it.

> 
> At the moment i am trying to plot circles based on the atoms object
> i 
> get from my data. I do want to limit the
> amount of atoms from the structure to only include those which are 
> necessary and roughly of the same hightas the slice in question. This
> is 
> were i currently am
> wicht my code:
> 
> import  numpyas  np
> from  matplotlibimport  pyplot
> from  aseimport  Atom
> from  ase.atomsimport  Atoms
> from  ase.dataimport  covalent_radiias  radii
> from  ase.data.colorsimport  jmol_colors
> from  matplotlib.patchesimport  Circle
> from  ase.ioimport  read,  write
> fig,  ax=  pyplot.subplots()
> 
> 
> atoms=  read('CHGDIFF',  format="vasp")
> Z_POS=25.5
> maxi=  max(a.zfor  ain  atoms)
> mini=  min(a.zfor  ain  atoms)
> percent=40
> add=  ((maxi-mini)/100)*percent
> 
> 
> if  min(a.zfor  ain  atoms)  <  Z_POS<  max(a.zfor  ain  atoms):
> 	filter  =  [a.indexfor  ain  atomsif  Z_POS-
> add<  a.z<=  Z_POS+add]
> 	atoms_filtered=  Atoms([afor  ain  atomsif  a.indexin  filter])
> 	
> 	
> 	
> for  atomin  atoms_filtered:
> 	color=  jmol_colors[atom.number]
> 	#color = "None"
> 	radius=  radii[atom.number]
> 	circle=  Circle((atom.x,  atom.y),  (radius/2.5),  facecolor=co
> lor,edgecolor='k',  linewidth=0.4)#alpha=10)
> 	ax.add_patch(circle)
> 	
> 
> 
> ax.axis('equal')
> ax.set_xticks([])
> ax.set_yticks([])
> ax.axis('off')
> 
> fig.savefig('out.png',dpi=300)

You could directly use matplotlib markers instead of drawing circles
(you can select the round ones and specify a size).

> 
> 
> 
> 
> 
> 
> 
> 
> There are a few problems, however.
> 
> a) Some atoms are not in the unit cell any more but appear as mirrors
> on 
> the other side  after the calculation? Can i shift them back somehow?
atom.wrap() in general, otherwise you will have to play with positions.


> b) if i limit with the "percent" variable the amount of atoms i
> display 
> to not include the slab any more, the plotted picture does not have
> the 
> size of the unicell any more?
It would... if only you gave it in your atom constructor. ASE is not
magical, if you give it:
    Atoms([a for  a in  atoms if  a.index in filter])
it won't relate that to what you did before because the cell
information is contained in the `Atoms` object, not in the `Atom` one.

>  Can i somehow force it to this size anyways?
So either specify the cell size upon instanciation or use slicing on
your `atoms` variable instead of building everything yourself.

> How do i make sure, that i can to a superposition in a way that the 
> atoms are always at the corect position with in the slice thru the 
> CHGCAR file?
You can make sure of it by checking it.  

> c) i use jmol_colors to color my structure i would however want to 
> change only the colour of the atoms in my slab ( Cu) to be
> transparent. 
> Can i change the corresponding entry in the array
> color = jmol_colors[atom.number] somehow?
Yes you can, but I suggest instead that you color the structure
yourself. That way you can easily use predefined colors in jmol_colors
while special-casing slab Cu.

> d) Can i somehow connect only the centres of those circles from a 
> certain specie such that the molecular structure becomes visible?
Yes you can: Python is a proper programming language and matplotlib a
very capable plotting library. But if your question is actually "can I
do that without providing any effort", the answer is most likely "no".
ASE is a toolkit: it is more powerful that almost anything else I know
of as long as one is willing to dive into it. That's difficult at first
but then becomes easier and easier.


> e) I suppose there are much better ways to do this? Could somebody
> point 
> my in the right direction should this be the case?

As I already said, Paraview could be a solution, though it also
requires (different) work and training to get there. Note also that it
is not a tool dedicated to chemistry and might seem weird because of
that. As well, I suppose Vesta, Avogadro, VDM, PyMOL and others are all
capable to do exactly what you want but it is going to require some
work.

Regards,
Gaël

> 
> All the best and thank you in advance
> 
> Fabian
> 
> 
> 
> 
> 
> _______________________________________________
> ase-users mailing list
> ase-users at listserv.fysik.dtu.dk
> https://listserv.fysik.dtu.dk/mailman/listinfo/ase-users
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 3199 bytes
Desc: not available
URL: <http://listserv.fysik.dtu.dk/pipermail/ase-users/attachments/20170518/3cb6001d/attachment-0001.bin>