Home   Manual

How to build a supercell correctly?

I am trying to build a supercell for a molecular crystal starting from a .cif file that is in the literature (the beta form here: https://scripts.iucr.org/cgi-bin/paper?OS0098). I have been unable to open .cif files in Avogadro2, as I have posted about already on this forum, so I have been using Avogadro 1.2.0.

When I build a supercell by going to Build >> Super Cell Builder, and then entering the repeats (say, 5, 6, 4), I export to an .xyz file. Unfortunately, I see two problems: 1. some of the molecules are missing atoms (for example, there are less hydrogens than there should be based on the number of molecules), and 2. I can’t export the data such that the atoms are grouped by each molecule. How can I get around this?

For #2, I believe I can just write a code to calculate and sort nearest neighbors – though when I move on to non-orthorhombic crystals, this will presumably be more complicated. But #1 seems like a big issue…

Thank you in advance.

If you can post an example or screen shot, it would help - are the hydrogens in the asymmetric repeat unit before replication? Some small molecule CIF sometimes omit the H (because they’re hard to determine from the raw x-ray data).

I did what you described and saw the standard behavior of Avogadro – building a supercell it adds the specified number of atoms that constitute the cell and not more, so molecules at the edges are trimmed.
As for the #2, the grouping of atoms is really very strange. The original cell has NCCHH, and in the supercell coordinates atoms go NCCHH, 4Ns, 7Cs, 10Hs, which is really confusing. Again, since version 1.2.0 is no longer supported I guess it will not be fixed.

If you can point me at a CIF, I’d be happy to look at this for Avo2.
(Based on usernames, I’m assuming that @Iamkaant and @src are different requests?)

It sounds like what you want is:

  • generate the symmetric atoms in the unit cell from the closed-form molecule
  • generate the supercell without “wrapping” atoms
  • keep all bonds intact

That seems like a reasonable separate command… suggestions for a descriptive name would be welcome. There’s often a difference of opinion between inorganic and molecular crystal communities…