> Silly questions,
they're not silly Armel... you caught the point...
> >From where come the equations used for modeling the instrumental
> component and emission profile in the fundamental parameter
> approach ? For instance, in BGMN, the instrumental function seems
> to be made of sums of squared Lorentzian functions (if I am right ?).
>
> I mean : is there not some remaining empirism in this "fundamental
> approach" ? Are the profiles exactly described (as we can describe
> a circle or a sphere exactly by a mathematical formula) or modeled ?
... what should also be said is that the Fundamental Parameter Approach
(FPA) cannot be used everywhere and with whatever geometry and setup...
It is nearly impossible to have a 0% empirism... at least in our world ;o)
If we do not tolerate any empirism, well, probably we should forget the
Rietveld method and the diffraction techniques in general also... :o)
> If some % of empirism remain in the "Fundamental Parameter
> Approach", is there not some intellectual swindle in this affair ?:-)
... what we can argue is that there is less empirism than using the
traditional Rietveld (U V W) approach... we simply change a set of meaningless
parameters into a set of "meaningful + still some meaningless" parameters
(if any justification on possible meaning of the latter...i.e. size/strain
is not given)... I look at the FPA under this light... as a net result, we
can also have some more info and account for strange behaviours
(anisotropic broadening... funny peak shapes...)
> Moreover, how are described the crystallite size and strain
> broadening , "entered as refinable values". Are they still
> represented by Lorentzian and Gaussian components ? If yes,
> this is nothing more than a semi-empirical approach (i.e. the
> crystallite size and strain distributions and the variation law
> of strain with distance from an arbitrary origin are postulated :
> and this is rather not fundamental...)
you already know my opinion on that... there isn't any indication on the
range of validity of the model, nor an indication on the necessity for a
preliminar evaluation of size/strain broadening sources based on
traditional techniques.... Profile Analysis IS NOT ABSOLUTE! The approach
used in Koalariet, for instance, is too semplicistic... while I haven't
got enough data to say anything about BGMN....
How about size/strain anisotropy...??? How about distinction of actual
sources of broadening??
As regarding to the postulation of size/strain distributions... this is
valid for traditional LPA... but there are also approaches based on
Information Theory or on Monte Carlo methods that can estimate this
distribution... just a matter of implementing them in the Rietveld code
;o)... don't think that noone is working on this... ;o)
> So tell me : really 0.00% empirism ? If not, give an estimation
> of the remaining percentage, please.
That is a good question.... I join you in waiting for an answer from those
who already implemented the FPA... in my opinion, without any furteher
indication on the ranges of validity there is already a 100% empirism in
the choice of the size/strain broadening model...
A further question is: how to check for the percentage of empirism in the
definition of the instrumental function? You can just have a qualitative
assessment of the agreement between the modelled instrumental profile and
that measured on a "standard" (?) like NBS LaB6...
Lot of questions... lot of discussion topics for EPDIC6....
Matteo
wwwww
g( o 0 )g
--oOO--(_)---OOo---------------------------------------oOO-wwwww-OOo-------
Department of Materials Engineering
University of Trento
38050 Mesiano (TN)
dott. ing. Matteo Leoni ITALY
.ooo0 0ooo. Tel +39 461 882417
( ) ( ) Fax +39 461 881977 E-mail: Matteo.Leoni@ing.unitn.it
----\ (---) /--------------------------------------------------------------
\_) (_/ | | | |
.ooo0 0ooo.
( ) ( )
\ ( ) /
\_) (_/