> -----Original Message-----
> From: Armel Le Bail [mailto:[EMAIL PROTECTED]]
> Sent: Wednesday, April 18, 2001 12:03 PM
> To: [EMAIL PROTECTED]
> Subject: RE: Size/Strain Round Robin - 1st Report
>
>
>
> >Well, although the specimen was nearly strain-free, strictly
> speaking this
> >was still a size-strain round robin; all the participants extracted (or
> >refined) strain parameters, didn't they :-)
>
> Before to start, the Round Robin participants were said :
> "Although the first sample selected for the Round Robin shows
> isotropic line broadening, future samples will exhibit anisotropic
> line broadening."
>
> Another option would have been to not give that assumption of
> isotropic line broadening, which certainly bias the possibility for more
> spread results (and personally, I have proposed 2 sets of results :
> with isotropic and anisotropic line broadening).
>
> A third option could have been to give 2 assumptions : Although
> the first sample selected for the Round Robin shows isotropic
> line broadening and size-effect only, future samples will exhibit
> anisotropic line broadening and/or both size/strain effects. Then
> the participants would have extracted only the size, isotropic,
> and the spread on the results would have been even narrower.
>
> Round Robin organizers have some latitude, fortunately. The
> participants measured strain because you did not say that the
> sample was almost strain-free. The participants used isotropic
> line broadening because you told them to do so. Anyway, in
> a manuscript submitted for publication, the authors could well
> say : we tried anisotropic and size/strain effects, and concluded
> that the sample shows isotropic line-broadening and is
> strain-free, so that calculations were made again assuming
> isotropy and no strain effect. The R factors were similar in both
> cases. That paper would be accepted for publication (if I was
> the reviewer ;-).
Yes Armel, this choice (to assume isotropic broadening and not strain-free
sample) was purposely made for several reasons:
1. Size and strain values are basic numbers that one gets from
line-broadening analysis. Because physically, they originate from quite
different sources, it is important to establish through the round robin how
easy is to "mix up" the two during the analysis and what is resulting
scatter of results. On the other side, anisotropy is a step further in our
understanding of sample properties, that is, after the crystal-direction
"average" values are established, one can refine the results by allowing for
anisotropy. However, if we try to put too many variables into the first
phase of round robin, the experience tells me that the results are usually
not satisfactory. That is why I thought that more complicated samples should
follow later after we learn what the basic problems are.
2. Asking for anisotropic size broadening results would be technically a
problem, especially for people using Rietveld programs. Although
strain-broadening models became more sophisticated during the last 2-3
years, anisotropic size broadening, if included at all, is still mainly
accomplished through Greaves platelets model (cos(phi) correction). This
would work satisfactory in this case, but not for a more complicated
anisotropic size broadening.
3. The last, but certainly not the least, by importance is that we actually
don't have a definite proof that there is no substantial strain broadening!
That is, although there is a strong indication according to a preparation
route and diffraction line-broadening analysis, there is no independent
confirmation by a different technique (I am sure that Daniel Louer will
correct me if I am wrong). Hence, we will have to make a good case in the
paper why we believe the sample was strain-free.
In view of this, I hope that you'll indeed be the referee or it might get
rejected:-) But, regardless of who the referees are, we could have used some
of this time to work on the manuscript instead:-)
Best regards,
Davor Balzar
******************************************************
National Institute of Standards and Technology
Materials Science and Engineering Laboratory
Div. 853, 325 Broadway
Boulder, CO 80305-3328
Tel: 303-497-3006
Fax: 303-497-5030
E-mail: [EMAIL PROTECTED]
Web: http://www.boulder.nist.gov/div853/balzar
European mirror:
http://www.ccp14.ac.uk/ccp/web-mirrors/balzar/div853/balzar/
******************************************************
