ATOM TYPE category

Discussion of core CIF dictionary definitions. Please choose topic headings so that the first words refer to a particular dataname or category, whether existing or proposed.

Moderators: Brian McMahon, jcbollinger

Post Reply
jamesrhester
Posts: 39
Joined: Mon Sep 19, 2011 8:21 am

ATOM TYPE category

Post by jamesrhester » Wed May 07, 2014 7:08 am

This thread is for discussion of the ATOM TYPE category.

We need a way to handle different isotopes, which have wildly different scattering lengths for neutrons. I propose simply adding _atom_type_atomic_weight to this category in order to allow description of different isotopes.

I note that X-ray laue is covered in the sense that wavelength-dependent dispersion can refer to a table through _atom_type_scat_source and the scattering length can be tabulated in terms of sin theta/lambda

jcbollinger
Posts: 57
Joined: Tue Dec 20, 2011 2:41 pm

Re: ATOM TYPE category

Post by jcbollinger » Wed May 07, 2014 2:43 pm

jamesrhester wrote:We need a way to handle different isotopes, which have wildly different scattering lengths for neutrons. I propose simply adding _atom_type_atomic_weight to this category in order to allow description of different isotopes.

I recommend _atom_type_mass_number or _atom_type_isotope_mass_number instead. However you spell the data name, the mass number is more of an identifier than the isotopic mass; after all, that's what we use in our isotope symbols. Moreover, the mass number is always an integer, whereas the isotopic mass generally is not. Integers are far more suitable for indexing, and are not subject to issues related to variable (sub-unit) precision.

jamesrhester
Posts: 39
Joined: Mon Sep 19, 2011 8:21 am

Re: ATOM TYPE category

Post by jamesrhester » Thu May 08, 2014 7:38 am

Agreed. The integer value is the useful one. Also, as far as I can tell the mmCIF/PDBX dictionary has no equivalent.

Now, how do we describe natural abundance? We would seriously complicate the description if we were to have separate types for every isotope in the sample, and then have to describe a mixed occupation model. So, how about we define an additional dataname '_atom_type_isotope_abundance_index', pointing to a table of isotopes and abundances. The user has the option of
  • ignoring _atom_type_isotope_mass_number completely, in which case natural abundance is assumed and the correctness of this assumption can be checked via the stated scattering length
  • using a pure isotope in which case _atom_type_mass_number is provided in the table
  • using an isotope mix in which case _atom_type_isotope_id points to an entry in the following table, where an atom_type column is not necessary as the index from the main _atom_type table will carry the atom type in any case:

Code: Select all

loop_
    _atom_type_symbol
    _atom_type_oxidation_number
    _atom_type_number_in_cell
    _atom_type_isotope_abundance_id
      C 0 72  2
      H 0 100  .
      O 0 12  .
      B 0 4  1

loop_
_atom_type_isotope_abundance_id
_atom_type_isotope_mass_number
_atom_type_isotope_abundance_percent
1  11 80
1  10 20
2  12 98.9
2  13 1.1


An alternative might be to do away with _atom_type_isotope_mass_number altogether in favour of an index into this table, which would have one line with 100 percent abundance for a pure isotope.

Post Reply