[sdpd] FIDDLE, the new key for SDPD ?

[Armel Le Bail <armel.le_bail...@univ-lemans.fr>]

Hi,

The XXI IUCr meeting, Osaka, was full of novelties.

Rather than to admit them without any discussion, let us have a closer 
look. Feel free to send comments and opinions there...

I could follow a talk about a new software named FIDDLE, including 
everything concerning SDPD in the global optimization : search for the 
cell, the space group, Z, and the atomic positions. An unknown structure 
was solved.

Anyway, I suspect that all the FIDDLE software calculations made outside of 
the correct cell, the correct space group and the correct Z are useless 
(though working in supercells would also provide the solution), and this 
makes a lot of time wasting... Seems that serious things accur when the 
cell is found by the software. So, why not to concentrate first on the 
cell, in order to save time ?-). Oh, yes, indexing first would break the 
project of including everything in the global optimization... But is there 
really any progress ?

Can I say that a super global optimization is "structure prediction" : you 
even include the synthesis step and the chemical formula in the whole 
process, no need for the powder pattern at the beginning.

Armel


Abstract below :

MS.88.5
FIDDLE: A method for simultaneous indexing and structure solution from 
powder diffraction data
Rene De Gelder, Carmen Guguta, Jan M.M. Smits
Radboud University Nijmegen, Molecular Materials, Heyendaalseweg 135, 
Nijmegen, Gelderland, 6525 AJ, The Netherlands
The complexity of crystal structures determined from powder diffraction 
data has steadily increased through further development of traditional 
methods for structure determination in reciprocal space and application of 
global optimization algorithms in direct space. The usual process for 
structure determination from powder diffraction data consists of the 
following steps: (1) indexing of the pattern, (2) space group 
determination, (3) structure solution, (4) structure refinement. The 
currently available powder methods rely on successfully passing the first 
step, powder indexing. Due to a number of fundamental and experimental 
problems, like peak broadening, the presence of impurity phases, dominant 
zones and geometrical ambiguities, powder indexing will remain difficult in 
many cases, thereby hampering the next steps in structure determination. 
There is no fundamental reason to separate, as is usual today, the process 
of unit cell determination and the process of structure solution. Structure 
determination from powder diffraction data can be seen as a process of 
global optimization of all model parameters, including the unit cell 
parameters. This strategy is applied in the FIDDLE program. For the 
simultaneous optimization of the parameters that describe a crystal 
structure genetic algorithms together with a pattern matching technique 
based on auto and cross correlation functions are used. This one-pot 
strategy for indexing and structure solution, as applied in FIDDLE, was 
successfully used for determining the unknown crystal structures of Ethinyl 
Estradiol anhydrate, Naloxone monohydrate and Creatine anhydrate, cases for 
which indexing was problematic.
keywords:
indexing
crystal structure determination X-ray powder data
optimization algorithms


------------------------------------

Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/sdpd/

<*> Your email settings:
    Individual Email | Traditional

<*> To change settings online go to:
    http://groups.yahoo.com/group/sdpd/join
    (Yahoo! ID required)

<*> To change settings via email:
    mailto:sdpd-digest...@yahoogroups.com 
    mailto:sdpd-fullfeatured...@yahoogroups.com

<*> To unsubscribe from this group, send an email to:
    sdpd-unsubscribe...@yahoogroups.com

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/