Re: Showing icosahedras in VESTA

[letshin]

Dear All,

I just started using Vesta and would like to create some icosahedras based on some glass word I'm doing. 

I'm not so sure how to go about making these in a random arrangement. Could anyone please advice? :)

Thank you very much.

Best regards,

Zhao

[Philip Howie]

Dear Zhao,

To display polyhedra - such as icosahedra - in VESTA, you need to have atoms in the correct positions, connected with bonds.  Bonds can be created by going to Edit | Bonds... and setting the correct interatomic distances.  Polyhedra should then be displayed when the correct display style is selected from the column on the left.  Unfortunately, fine-grained control of which atoms are included in which polyhedron is not possible at the moment, so you may get better results with some atomic arrangements than with others!

VESTA is, to my knowledge, not able to create random arrangements of atoms.  My best suggestion would be to find another way of generating the random arrangement and then import the atom positions into VESTA for visualisation.

Best wishes,

Philip

[letshin]

Thanks Philip for your reply!

Off the top of your head do you have any idea of any methods that could be used for generating said positions? If left to my own devices I would make a matrix in excel and assign random numbers to all the cells and pop that back into Vesta which seems slightly complicated.

Best regards,

Zhao

[Philip Howie]

I don't really know how to do this, but perhaps I can make a couple of suggestions:

Of course, you can't simply generate random atomic positions, because you need them to be arranged in icosahedra.  If the icosahedra are independent and not bonded together, you could generate random positions and orientations for the centres of the icosahedra and then use those to generate the atoms themselves.  You'd have to come up with some way of preventing them from impinging on each other - or simply throw out any that do impinge.  The density of icosahedra you generate will have to be calculated from the density of the material you're trying to model, and I suspect the higher the density, the more difficult it will be to find rooms for everything without impingement.

More likely, however, your icosahedra are in a giant glassy network and share vertices or edges?  In that case, I think I'd start with a seed icosahedron, randomly pick a shared edge, randomly pick a suitable orientation for the adjoined icosahedron and generate its atoms.  Choose a new edge from any of the existing icosahedra and repeat the process.  If the combination of edge and orientation produces an icosahedron which impinges on the existing network, throw it out and start again.

...the more I think about this, the more I realise that it's going to be very difficult!  Unless you can come up with a better way of generating the positions, you might be better off with a program which is designed for drawing glasses, rather than crystals.

Best,

Philip

[letshin]

Thanks Philip!

Best,

Zhao