Hi Niki,
Looking at your parameters I would not expect significant differences in
any of the models as they should all be dominated by the core, a very thin
20nm diameter disc (of thickness 0.08nm to 0.34nm depending on the
existence or not of faces). Adding a 0.1nm shell (rim) shouldn't make much
difference unless the core is close to contrast matched I would think? For
the face thickness one might expect some oscillations at unrealistically
high q that I would normally expect to be in the background.
Reading the model documentation I now see the problem(s). The main problem
being our interpretation of the model.
1) The core shell bicelle model states that it is a *general* case of the
coreShellCylinder that allows you to have different SLD on the face and the
rim! I then looked at the cylinder documentation (and if you blow up the
picture in the excel sheet you will see it) where the graphic clearly shows
that the shell surronds the core of the cylinder: both a face *and* a rim
but the thickness and sld is the same all around (the bicelle allows both
for a different t and a differnet SLD between the rim and face). hence the
graphic shows that Ltotal= 2H +2t where L(of core)=2H as also shown on the
graphic.
2) I notice that you in fact are going out to 10 1/A. At these qs you are
looking on the fraction of an Angstom lengths and the average SLD
approximation made in all these models starts to break down as implied in
my first paragraph.
3) Ignoring now whether the continuum approximation is correct at these
very high qs the sandwich structure in the normal direction fo a 0.14nm
coating layer followed by the 0.08nm core thicknes followed by the 0.14nm
coating on the other side shows up as the oscillation you see in 1 and 3
(your model graphic is of course wrong for 3 as discussed in 1 above - it
should have a 0.14nm layer on the face as well as the rim with same SLD on
both). Adding just a tiny rim to the giant disk as I suggested in the
beginning does not matter much unless you make the core match the solvent
perfectly .. then you will see some ringing in the pattern. In fact I was
surprised at how close you have to be to matched before seeing anything.
Hope this helps?
Paul
On Thu, Dec 10, 2015 at 7:24 AM, Niki Baccile <niki.baccile@...16...> wrote:
> Dear Paul, Lilin
>
> thank you for your messages.
>
> As attached document a full pic of what I mean (including the numerical
> values). I tried to make it as clear as possible (drawings come from the
> SasView documentation).
>
> I do not understand why Model N°2 (CoreShellBicelle) and Model N°3
> (CoreShellCylinder), which should in principle describe the same object
> according to the choice of the parameters (no face, rim thickness> 0, rest
> is the same), provide a different scattering profile.
>
> I understand even less why Model N°1 (CoreShellBicelle) and Model N°3
> (CoreShellCylinder), which DO NOT describe the same object (face thick> 0
> versus rim thick>0, rest is the same), do provide an alike scattering
> profile.
>
> Does this make any sense? Did I make a mistake somewhere? Or there is a
> problem in the model description for the Core-Shell Bicelle (or for the
> Core Shell Cylinder)? If I made no mistakes, either one or the other is
> wrong.
>
> thank you very much
>
> Niki
>
>
>
