Re: [sdpd] Recent Congress and comiing ones about SDPD

[daniel chateigner <daniel.chateigner...@ismra.fr>]

Hi all,

May be some hints after Armel's general thoughts on past congresses:

>    - pushing the limits of SDPD by transforming powder data into
>      single-crystal-like data (using texture, anisotropic thermal
>      expansion, anomalous scattering).

Texture indeed may be interesting if samples ... are textured ! Not a 
question of transformation, but only to take account of the orientation 
distribution in a comprehensive physical way rather than adding free 
parameters like in March-Dollase models or whatsoever.
"Textured samples" means you are between the pure powder (that practically 
never exists) and the perfect crystal (which does exist too). Unless the 
texture is rather simple, you need to refine the orientation distribution 
function, and need to acquire like 1000 patterns in as many sample 
orientations. Care, if you are allowed to destroy your samples (rarely my 
case unfortunately), you sometimes cannot treat the "powderised" pattern: 
too much low-symmetry phases, coupled with needed QPA and microstructural 
determinations ... makes a single pattern giving quite few information to 
avoid any program to diverge.
On another hand, on textured samples you can make use of the 
crystallographic constraint given by angular relationship of crystal 
directions, as implicitely done in single crystal structure determinations. 
This is obviously done through the ODF refinement, associated with a 
Rietveld-like treatment, which actually takes the name of "combined 
approach". For those that can read french, 
http://www.ensicaen.ismra.fr/~chateign/texture/combined.pdf is a place 
where to read about it. Sorry, soon in english. Interestingly, we observed 
that the texture treatment seems to partially remove program's divergency, 
which afterall looks coherent since many information and constraints are 
added treating it.
But you need to acquire a lot of patterns, which is now accessible in a 
reasonable acquisition time (say from a day to a week on regular samples 
including films) with lab diffractometers 
(http://www.ensicaen.ismra.fr/~chateign/texture/gonio.htm)
The advantage is that you can then access all parameters accessible to 
diffraction: texture, structure, residual strains, microstrains, particle 
sizes, phase ratio, film thicknesses and roughnesses, layering ..., without:
- destruction of the sample
- loosing the correlation between parameters (e.g. texture influences QPA, 
structure determination or residual stress)
- loosing the possibility of analysing the parameters (e.g. powderising 
means often residual stresses are relaxed)
- loosing the possibility of characterising your sample (e.g., did you try 
to make a powder sample out of a 100 nm thick film deposited on a 1mm thick 
substrate ?)

Of course you then need a program to analyse all these ! I only know two: 
MAUD and GSAS. May be Fullprof (Juan did you implement the ODF 
determination ?) ? For sure Beartex don't do this. It is able to refine ODF 
and calculate some macroscopic properties from the ODF and microcrystalline 
tensors, plus other texture-related things, but does not know anything 
about Rietveld procedures. Again, you need to combine Rietveld, ODF, 
Residual stresses and sometimes reflectivity algoritms, which I think only 
MAUD can do.

Hope this has not been too long !
Daniel


------------------------ Yahoo! Groups Sponsor ---------------------~-->
Buy Ink Cartridges or Refill Kits for your HP, Epson, Canon or Lexmark
Printer at MyInks.com. Free s/h on orders $50 or more to the US & Canada.
http://www.c1tracking.com/l.asp?cid=5511
http://us.click.yahoo.com/mOAaAA/3exGAA/qnsNAA/UIYolB/TM
---------------------------------------------------------------------~->

 

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