Re: unit cell refinement

Armel Le Bail ( armel@fluo.univ-lemans.fr )
Wed, 18 Mar 1998 15:15:48 +0100

Nita DRAGOE wrote :
>Lately I started to add a unit cell refinement routine to the program
Convert.
>It seems to work now but I want to test it on some "real world cells". So I
>would be grateful if you could send me some data for testing (preferably not
>simulated). The program accepts internal standard, zero and wavelength
>corrections, the data should be in 2theta, theta or d.

Almost about the same subject :
Before refining cell parameters, one needs to identify reflections.
This is generally done by comparing the d(obs) or 2-theta(obs) to
a list of values calculated from some program generating hkls,
knowing the cell. Let me announce here a new simple software
doing that in a simple way (sorry Brian, I gave it the same name
as your HKLGEN software, but the space group seems easiest to
enter than in your own version):

http://fluo.univ-lemans.fr:8001/ftp/hklgen.zip (zipped by winzip)

It is a Win95 32 bits software compiled with Digital Visual Fortran
and showing (useless in such a case) nice windows as User
Interface (UI). It is User Friendly (UF) in my opinion (input as
free format with few data to enter, simple output).

Now, it seems that this step of refining cell parameters is really
obsolete (sorry for ICDD :-). Cell parameters are better defined by
a Rietveld refinement. Even if the structure is unknown, cell
parameters are better defined by a Pawley or Le Bail methods
application. Evenmore, such an application may allow to confirm
or reject a cell proposition (from indexing programs like TREOR,
DICVOL and ITO). If you have d(obs), then this means that you have a
full powder pattern. So that, it is recommended to use the whole
pattern instead of (more or less bad) extracted "d(obs)".

In fact, ICDD should not accept cell parameters refined in the old
way after December 31, 1999 (or maybe 2000?). Comments ?

Finally, you may not need HKLGEN, after all. Anyway, if you want
to use it at its maximum, and refine cell parameters according to
the old way, the limit with HKLGEN is 8000 hkls, good luck...;-).

Armel
out of lethargy