[sdpd] Re: F1 and Monte Carlo

[Andrzej Grzechnik <andrzej... @servix.mpi-stuttgart.mpg.de>]

At 01:32 10.05.99 +0200, Armel Le Bail wrote:
>Andrzej Grzechnik wrote:
>>In the description of the work by Armel Le Bail (ESPOIR), I can not find
anything
>>about potentials. What are they? My general comment about the Monte Carlo
>>approach is that it is very sensitive to the potential employed and, at
>>least, at the moment is not surpassing the direct methods in dealing with
>>low dimensional structures (not to mention the time of computing). 
>
>No potential at all in ESPOIR. The program is intended to "surpass"
>direct methods in cases where the direct methods fail : bad powder
>data with strong overlapping, leading to few well estimated
>structure factors. The SDPD Round Robin sample II is somewhat
>exemplary on that point. Hard to solve with direct method (no
>participant of the Round Robin proposed a solution). Anyway, a
>solution is obtained from ESPOIR (with time, however INTEL 
>announces already 3GHz for 2003, 64 bits ;-), although I did
>not succeeded with Endeavour. Endeavour may also be used
>without potential by considering lowest approach distances.
>
>What are the complexity of the structures that you were able
>to solve with Endeavour (number of independent atoms, space
>groups, number of reflections) ?
>
>Best,
>
>Armel Le Bail - Université du Maine, Laboratoire des Fluorures,
>CNRS ESA 6010, Av. O. Messiaen, 72085 Le Mans Cedex 9, France
>http://www.cristal.org/
>
>


Hi,

the structures that I'm dealing with and using Monte Carlo for are not as
complicated as the SDPD Round Robin sample II. These are the structures from
high pressure experiments. Usually at high P conditions, it is difficult to
entirely eliminate deviatoric stresses and preferred orientation effects.
For this reason, direct methods are not always successful, although EXPO is
able to estimate preferred orientation at the stage of intensity extraction.
Anyways, with such data, in most cases collected to 2 theta 25-30(lambda:
0.44-0.45), indexing and finding a space group is not a trivial matter. I
used ENDEAVOUR for some structures that I already knew to see how it works
and whether it finds the structure. It actually does the thing. Then I used
it for some unknown cases for which I had the total number of atoms in the
unit cell, lattice parameters, and bad data (peak positions + intensities).
No indexing. No structure factors. The most "complicated" case was with 20
atoms, tetragonal lattice parameters, and 46 reflexions. I got a  structure
(four independent atoms) that I refined with GSAS. 

Anyways, the idea of not using any potentials is a good one, providing that
the minimum distances between atoms can be changed depending on the
conditions of data collection. In my calculations for high pressure
structures, I usually set them up to values 10% to 20% lower than the ones
observed at ambient conditions. 

Regards,

Andrzej Grzechnik


----------------------------------------------------
 Andrzej Grzechnik 
 Max-Planck-Institut fuer Festkoerperforschung   
 Heisenbergstr. 1,  D-70569 Stuttgart,  Germany
 Tel: (++49) 711 689 1404;  Fax: (++49) 711 689 1444
 E-mail:  andrzej... @servix.mpi-stuttgart.mpg.de
----------------------------------------------------