At 07:14 16/10/2004 -0500, Bob Hanson wrote:

Bob,
I am sympathetic to your problem but the more information you supply
on "ICSD CIFs" the worse the conformance is. If this is really what ICSD is
emitting there are serious problems and they are doing a disservice to people who have to write readers. There is also a serious likelihood that such programs will have to guess, and will guess wrong, about what the data is.




Peter Murray-Rust wrote:

At 08:06 15/10/2004 -0500, [EMAIL PROTECTED] wrote:

I am in Chicago and don't have full access to my data, but I'd
like to contribute to this discussion.

CIF Symmetry: The key here should be

loop_
_symmetry_equiv_pos_site_id
_symmetry_equiv_pos_as_xyz
 1 '-x, y, -z+1/2'

this is probably illegal. The identity operation is always required

Sorry. I wasn't listing the whole set. I had to work from a hotel computer. The whole set was:

  1     '-x, y, -z+1/2'
  2     '-x, -y, z+1/2'
  3     '-x, y, z'
  4     '-x, -y, -z'
  5     'x, -y, z+1/2'

But, still, 'x, y, z' is not there, is it?

No, and this is not a space group. Even with x,y,z, it is still not a group or a factor group of a group. It should simply throw an error.


A quick look turns up the following
without 'x, y, z':

ic30516.cif

this should be preceded by loop_ and
_symmetry_equiv_pos_number (or something like that

_symmetry_equiv_pos_as_xyz
  1     '-x, y, -z+1/2'
  2     '-x, -y, z+1/2'
  3     '-x, y, z'
  4     '-x, -y, -z'
  5     'x, -y, z+1/2'
loop_

ic26060.cif
_symmetry_equiv_pos_as_xyz
  1     'x+1/2, y, -z+1/2'
  2     'x, -y+1/2, z+1/2'
  3     '-x+1/2, y+1/2, z'
  4     '-x, -y, -z'
  5     '-x+1/2, -y, z+1/2'
  6     '-x, y+1/2, -z+1/2'
  7     'x+1/2, -y+1/2, -z'

this *is* a group if x,y,z, is added, but it should be explicit


loop_

ic66547.cif
_symmetry_equiv_pos_as_xyz
  1     '-z, -x, -y'
  2     '-y, -z, -x'
  3     '-x, -y, -z'
  4     'z, x, y'
  5     'y, z, x'
loop_

Same comment


Oh, I see now. ICSD does not report the (trivial) 'x, y, z'.
I think it's safe to say 'x, y, z' is implied and not necessary.

I think it is extremely unsafe for ICSD to omit it.



I'm guessing that any CIF file showing only fractional coordinates
must contain this record.

Strictly not. All crystals implicit have 'x,y,z' and that would be the default. Rather more difficult, you can specify the spacegroup and look the positions up in Int. Tab. for Cryst. unfortunately there don't seem to be any open collections of space group operators

I'm sure web can get this list from IUCr.

No - I asked on Monday

But it's a BIG list. There are
230 space groups and about 40 optional origins/coordinate systems (see ftp://ftp.iucr.org/pub/cif_sym_1.0.dic).

This is the dictionary governing the symmetry, not the operations themselves.

 Several of the 230 groups have
96 equivalent positions, and some have 192.

Correct.

You want to list all of these somewhere in Jmol code? I think this is going to be a big burden on Jmol.

It isn't unreasonable for a crystallographically aware program (like SHELX and now Jmol) to have access to a lookup-table of space groups. This need only be accessed if the author includes the space group symbol but not the positions. Also if the space group is Hall it is possible to calculate the operators from the symbols.


Personally, I think this is a flaw in the CIF standard.

I'd regard it as a problem in conformance. Theoretically it is perfectly allowable to give just the spacegroup and refer the reader to International tables but it makes it difficult for program writers. I think some authors for some journals do not include the operators and should be strongly urged to do so


This business is SO tricky, it only seems reasonable to include all positions.

I agree. There are shortcuts in some listings where the operators deducible from centre of symmetry and also the lattice centering are omitted. This reduces 192 in (say Fm3m) to 24 (192 = 24*2(centre)*4(lattice)


 Don't atoms with
different equivalent positions have different labels, anyway?

Sometimes, but they are not normally included in the CIF.


Am I understanding correctly that what we are after here is being able
to read CIF files that contain only

loop_
_atom_site_label
_atom_site_type_symbol
_atom_site_symmetry_multiplicity
_atom_site_Wyckoff_symbol
_atom_site_fract_x
_atom_site_fract_y
_atom_site_fract_z
_atom_site_B_iso_or_equiv
_atom_site_occupancy

information for the (x,y,z) identity operation?
For example, buckyball with one single atom listed?

Buckyball would actually require more atoms as crystals do not support icosahedral symmetry. But the logic is correct. Thus an octahedron of B atoms (B6) might just have one atom reported.



I'm not familiar with such files, although I understand that "in the old days" this is how just about all crystal files were delivered.

Correct

Where are they arising from? Peter, maybe you sent us some copies. I think I missed that. Is this possibly something we can avoid?

No. I am currently starting discussions on the CIF discussion group as to how to manage symmetry and disorder


I have implemented code in Molecular Origami that ALMOST does this. But I had trouble with identifying the origins. My solution was to identify a set of "minimum operations" that defined each point
group. It was very compact. For example,


Fd-3m(origin at CxCyCz)=227,192,F 4102/d1 -3111 2023/m
Fd-3m(origin at i)=227,192,F 4131/d -3 2022/m1
Fd-3c(origin at CxCyCz)=228,192,F 4102/d3 -3333 2023/c
Fd-3c(origin at i)=228,192,F 4113/d -3 2020/c1
Im-3m=229,96,I 4/m -3 2/m
Ia-3d=230,96,I 4120/a2 -3 2021/d

(Don't ask what all this means. I simply give these as examples of how 192 or 96
equivalent positions can be coded into a single short expression.)

They can actually be completely encoded in the Hall space group symbol and this should be the preferred way of doing this. If you create code that translates Hall symbols that would be very useful. From what you say, it shouldn't be too hard to adapt your code



Be aware that if you are going to do this right, you also have to take into account (ah, so fun!) trigonal and hexagonal coordinates.

Yes

I have spent enough time on this problem to know it is a VERY serious project.

Yes.

OTOH it is formally completable. I can give you encouragement!

If you have some ICSD CIFs that you think are broken, please send them to me (though I can't do much for ca 2 weeks). Make ABSOLUTELY sure they are direct from ICSD. Then I will send them to David Brown (one of the authors of CIF) and I am sure he will get ICSD to change them. But I have to be sure they came direct from ICSD.

If you want CIFs certified to be correct, look at (say)
http://journals.iucr.org/e/issues/2004/10/00/EAtxtisscontsbdy.html
and click on any link denoted by CIF. The inorganic ones should give you an idea of how to manage highly symmetric structures


P.


Bob


--
Robert M. Hanson, [EMAIL PROTECTED], 507-646-3107
Professor of Chemistry, St. Olaf College 1520 St. Olaf Ave., Northfield, MN 55057
mailto:[EMAIL PROTECTED] http://www.stolaf.edu/people/hansonr





------------------------------------------------------- This SF.net email is sponsored by: IT Product Guide on ITManagersJournal Use IT products in your business? Tell us what you think of them. Give us Your Opinions, Get Free ThinkGeek Gift Certificates! Click to find out more http://productguide.itmanagersjournal.com/guidepromo.tmpl _______________________________________________ Jmol-developers mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jmol-developers

Peter Murray-Rust Unilever Centre for Molecular Informatics Chemistry Department, Cambridge University Lensfield Road, CAMBRIDGE, CB2 1EW, UK Tel: +44-1223-763069



-------------------------------------------------------
This SF.net email is sponsored by: IT Product Guide on ITManagersJournal
Use IT products in your business? Tell us what you think of them. Give us
Your Opinions, Get Free ThinkGeek Gift Certificates! Click to find out more
http://productguide.itmanagersjournal.com/guidepromo.tmpl
_______________________________________________
Jmol-developers mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jmol-developers

Reply via email to