Re: energy deposition in thin silicon layer

From: Alberto Fasso' <fasso_at_slac.stanford.edu>
Date: Fri, 28 Aug 2009 11:40:09 -0700 (PDT)

Dear Marcel,

> Are the Bichsel/FLUKA results really independent, or is the underlying
> model the same?

They are completely independent. The model used by FLUKA is original,
and has never been used in any other code. A description can be found in
http://www.fluka.org/content/publications/1997_sare_1.pdf
You can also find a simple summary in the slides used for the FLUKA
courses: see pages 6-8 of
http://www.fluka.org/content/course/NEA/lectures/EM_2.pdf

> Does anyone know why GEANT4 and FLUKA yield different predictions? Have
> the two been compared in the past?

I don't know why GEANT4 gives a different prediction, and I don't
think there has ever been a comparison of the two models.

> Which is the correct curve? Has the FLUKA curve been compared to data?

Yes: in the two references I have indicated above you will find a comparison
with an experiment (2 GeV/c positrons and protons on a 100 micron Si layer,
by Bak et al., Nucl. Instr. Meth. B288, 681 (1987)).

I would like however to call your attention to the fact that your results
might be dependent of the settings you are using. Since you have no DEFAULTS
card in your input file, and no DELTARAY nor IONFLUCT command, your settings
are the following, as you can see from your output file:
    threshold for delta-ray production: 1 MeV
    accuracy level for ionization fluctuations: 1
You can change both by using DELTARAY and IONFLUCT. Note that there are
4 possible accuracy levels for ionization fluctuations, 1 being the
lowest.

Kind regards,

Alberto

On Tue, 18 Aug 2009, Marcel Vos wrote:

>
> Dear all,
>
> I used the FLUKA input card that you find in attachment to simulate the
> energy loss of 120 GeV pions in 450 microns of silicon.
> Being new to all this, I would appreciate if someone with more expertise
> could check that this procedure is indeed correct (are the default
> cut-offs right for such a thin layer, etc.).
> I analysed the FLUKA output file with some FORTRAN and C++/ROOT code I
> found on the mailing list. Thus, I pick up the event-by-event energy
> deposition in the different regions. In the attached figure I compare the
> resulting "Landau" distribution to several other models. I normalized all
> histograms. I scaled the X-axis to ADC counts for our detector (i.e.
> consider them arbitrary units).
> The yellow line is FLUKA. The red line is a prediction by Hans Bichsel's
> convolution method. These agree perfectly over the full range.
> The remaining lines are different GEANT4 setups (I tried to run
> with/without PAI model, with different step sizes, with/without nuclear
> interactions, adding the energy of escaping delta-electrons). The result
> is not affected very much by these choices and the high energy tail is
> ALWAYS incompatible with the Bichsel/FLUKA result.
> Are the Bichsel/FLUKA results really independent, or is the underlying
> model the same?
> Does anyone know why GEANT4 and FLUKA yield different predictions? Have
> the two been compared in the past?
> Which is the correct curve? Has the FLUKA curve been compared to data?
>
> Regards, Marcel
>
> PS I posted the same question to the GEANT4 authors.
>

-- 
Alberto Fasso`
SLAC-RP, MS 48, 2575 Sand Hill Road, Menlo Park CA 94025
Phone: (1 650) 926 4762   Fax: (1 650) 926 3569
fasso_at_slac.stanford.edu
Received on Fri Aug 28 2009 - 22:12:50 CEST

This archive was generated by hypermail 2.2.0 : Fri Aug 28 2009 - 22:12:53 CEST