Dear Norberto,
Thank you for your comments.
1. The differences in cell parameters on using different ranges may
well derive from the use of a “wrong” model, where “asphericity” of
the form factors (their vectorial dependence) is not considered, nor
anharmonicity of the atomic displacements is introduced. Nothing is
said about particle sizes and their distribution, which typically is
not gaussian (just to mention a few effects). This is physics and
goes well beyond instrumental effects. To partially avoid this bias,
one shoud try a parallel unbiased refinement by the Pawley or the Le
Bail method (Armel…where are you?)
Considering the effects you mentioned may further improves the accuracy
of the refinement parameters. In this point, I totally agree with you.
However, note that our claim is that the mismatch of the lattice parameter
in the calculation brings the finite Δ2θana and it lowers Rwp. Consequently,
a proportional unit-cell to the true one is obtained in the Rietveld method.
We think that it is the very fundamental findings. (Not technical one.)
I suppore this effect is much more largher than your suggested ones.
2. If the new criterion still is superior than Rwp, it’s difficult to
claim that it is Necessary. Perhaps useful in many cases, but
unapplicable if more complex phases, where peak broadening and
severe overlap exist.
Note that our proposed criterion is an additional one. Not alternative
to Rwp.
The proposed criterion is used by combining with Rwp. The convergence
is judged by using Rwp. Our criterion is calculated after the refinement.
As for the range of a usefulness of the criterion, you may be right.
But no one knows for now. I hope that time will solve.
The idea that it is a Necessary one calls to the Financial Interest
sentence, where it is explicitely said that a patent on this work is
pending. Does this mean that if someone (in the industrial environment,
I suppose) adopts this criterion in his/her certifications, he/she can
refrain from paying royaltees to Physonit Inc, being the choice of using
it Necessary (i.e. unavoidable?)
I'm sorry but I don't understand what you meant about 'Necessary.'
You think that the word is inproper?
Or you think that our proposed or any other criterion is not needed?
It you mean the former case, would you please tell me a proper word?
I want to improve my English vocabulary/skill.
As for a patent: In Japan, yes. In other countries, no.
Effect of a Japanese patent rights is applied only in Japan.
It means that it has nothing to do with most of you.
So, I wish you not to be nervous or aggressive about the patent.
3. The simple concept of average lattice parameter, when down to the
5th decimal digit, is clearly a nonsense, as it derives from too
much idealization and is a consequence of numerical analysis. Note
that in page 3 the cell parameter of LaB6 is given as 4.15686(0)
angs. Yes, you read it correctly: (0)! Like a perfectly known number
such as sqrt(2) or 3.14159, no matter how bad we are in measuring
them experimentally.
Mathematical/physical point of view, I agree with you.
But (0) is just a rounded numerical number output by a program.
For example, 4.15686 +/- 0.000001 becomes 4.15686 +/- 0.00000
in the present program. It's about a computational issue, isn't it?
Anyway, it doesn't mean that 4.15686 is a perfectly known number.
I have no comment about the following paragraphs.
4. Also, a Le Bail refinement cannot eliminate basic assumptions, such
as cubic symmetry, absence of diffuse scattering, radiation spectrum
shape, etc.
5. Finally, our structural analyses indicate that relative changes by
T, P, stress, doping, etc., drive the differential functionality of
these solids. Modern perovskites are appealing not because a = 6.12
angs, but because vacancies and interhalide dopings relatively
change the overlap integrals of the frontier orbitals.
Best wishes,
Masami
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Please do NOT attach files to the whole list <alan.he...@neutronoptics.com>
Send commands to <lists...@ill.fr> eg: HELP as the subject with no body text
The Rietveld_L list archive is on http://www.mail-archive.com/rietveld_l@ill.fr/
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
