Re: [sdpd] Giant and still so big

[Armel Le Bail <alb...@cristal.org>]

Hello,

>First of all, Thank you for your comments about the compound labelled MIL-100.

This is my job on that SDPD list to track interesting events.

>You will be able to find more details in the papers published in Angew Chem
>(2004, vol 43, page 6285-6301).

I have read the paper, this is why I try to understand more exactly.

- you need to know in advance "well defined and rigid inorganic and
organic building blocks"

- you are "exploring all possible combinations of connections
between inorganic and organic building blocks in the system
experimentally under study by using original global
optimization simulations".

If I understand well, the originality there (if compared with
direct space methods) is that you do not fix the cell and
that your cost function is based on energy calculations.

Several blind tests made recently apply exactly that concept.
See for instance :
Motherwell, W. D. S., et al. (2002). Acta Cryst. B58, 647-661.
Finding the cell by packing considerations and energy
minimizations is not new. See a few computer program names
(some being > 10 years old) below from the SDPD Database:
            http://www.cristal.org/iniref/progmeth.html
- HABIT95
   Molecular packing (Clydesdale, G.; Roberts,
    K. J.; Docherty, R.)
- HARDPACK
   A program for predicting molecular packing
   (Rainer Rudert)
- PMC
   (Dzyabchenko, Belsky & Zorkii) Packing
    calculation from atom-atom potentials (AAP)
- P-RISCON
   Masciocchi et al.)
- PROMET
   (Gavezzotti) Generation of possible crystal
    structure from the molecular structure
- UPACK
   package for crystal structure prediction
   (van Eijck BP.  Kroon J.)
Etc.
>I am afraid to say that this compound is beyond and NOT inside the range
>because we've got a ~0.023° FWHM (first (111) hkl reflection, d= 42.1 Å)
>for this compound using the ID-31 beamline so a V = 387000 A**3 is still
>higher than the estimation limit of V=288000 A**3 for FWHM close to 0.02°.
>Anyway, soon another results will be published with same FWHM BUT with a
>V=706.000 A**3 (Fd-3m S.G.)...

You forget that these limits were given in 1992 for the application
of Direct methods to powder data...

Now, with those methods for solving structures in direct
space (some computer program names : Octopus, FOX, DASH,
PSSP, PowderSolve, ESPOIR, etc), able to move big
building blocks (no matter if there is a mixture of
organic and inorganic blocks) in cells obtained from powder
data indexing, those old limits are no longer valid...
In fact we have no limit now for solving (the 50 first
peak intensities are sufficient for the determination of
the six degress of freedom corresponding to the location
and orientation of any building block of any size).
The only limit we have now is to refine by the Rietveld
method without contraints/ restraints. That is
generally impossible with big organic molecules.
(unless in large cubic cells if there are finally no more
than, say,  <200 atomic coordinates to refine, i.e. less than
70 independent atoms - which is your case).

For that MIL-100, some details are unclear in the published
paper, inclining me to ask for a few stupid questions :

- Did you failed in indexing ?

- Were the simulations started before to have obtained the
compound ? (this would prove the prediction).

- In other words, was the cubic cell parameter
known before the modelling ?

- If yes, why not to use it for restricting the search of the
building units positions (all SDPD experts prefer to do that...)?

- If not, I do not see why the cell indexing would have been
difficult from these synchrotron data (experienced people usually
immediately recognize if a structure is cubic from
the high regularity of the corresponding powder
patten, no matter if the cell is large).

- in which space group is made the simulation
(assembling Cr trimers and BTC, or the intuitively
made super tetrahedron ST) ? P1 space group ?

- Do you fixed the number of ST building units ?

- Do you think you would have solved the structure
with independent Cr trimers and BTC (without the
super tetrahedron hypothesis) ?

Finally, you write in the publication:
   "It is a new method in which conventional direct-
or reciprocal-space methods cannot be used in the
structure-solution process."

I must disagree.

I think that knowing the cubic cell parameter (not a
problem in my opinion), and with the ST hypothesis
(the biggest problem), most direct-space programs
would have solved the structure.

No obligation to answer. You paper passed the
reviewers obstacle (was not me ;-), so it is the truth now.

I agree that prediction is our unavoidable future in
crystallography. So, upload your best models inside
of the PCOD, we will see later if they are confirmed
http://www.crystallography.net/pcod/

Armel



------------------------ Yahoo! Groups Sponsor --------------------~--> 
What would our lives be like without music, dance, and theater?
Donate or volunteer in the arts today at Network for Good!
http://us.click.yahoo.com/TzSHvD/SOnJAA/79vVAA/UIYolB/TM
--------------------------------------------------------------------~-> 

 
Yahoo! Groups Links

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

<*> 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/