Interpretation of USRBDX sum.lis total response

Scoring
1

Versions

FLUKA: 4-5.2
Flair: 3.4-5.4

Description

Hello, I am currently performing what I expected to be a simple simulation to design a degrading foil for use in an experiment at CERN’s TELMAX irradiation facility, and in the process I got some results from USRBDX scorings that have me questioning if I properly understand how to score fluence from USRBDX and apply that as a total survival percentage downstream of the degrading foil.

The included flair project shows the setup in more detail (and NB that the usual step length and mulsopt cards for particle transport in a very thin material are used, along with lower transport thresholds), but, in short: a 100 keV antiproton beam impinges on a thin mylar foil (with thicknesses ranging between 500 nm and 1500 nm, varied from run to run with the “thickness” define statement). The region of the foil has a 1-way USRBDX fluence scoring, scoring the outgoing fluence from the FOIL region to the VOID region. There are also some track length density USRBINs along the way, which are used to measure the growth of the beam sigma (and backsolve for RMS angular kick and position-angle correlation). Finally, 10 cm downstream of the foil, there is another region, TARG, with another 1-way USRBDX fluence scoring from VOID to TARG.

Here’s where my question comes in: when I post-process my results in flair and check the USRBDX sum.lis file for the FOIL region, it lists a total integrated response given in parts per primary. When I have a zero thickness foil (obtained by setting the foil to vacuum), this value is very, very close to 1 (1.000014), as I would have expected. Similarly, taking a surface integral of the USRBIN tracklength densities is very close to one, too. The problem comes in when I have a thickness on the order of 500nm - 1100nm, since paradoxically this total response increases appreciably above 1 (in the case of 1100nm, 1.225988 part/pr). I would only expect this number to be close to 1 or decrease, so I’m quite confused, even after re-reading the slides and manual for USRBDX several times.

My understanding of the USRBDX scoring option has always been slightly shaky, unfortunately. But the reasons I expect this number to only decrease is that:

  1. This is a 1-way scoring, so we do not double count the antiprotons entering and exiting, and the total surface area of the region is specified in the scoring card, so it’s not a normalization problem (unless I made a real basic error here, but the fact the zero thickness case is right casts doubt on that).

  2. The integral of a USRBIN tracklength scoring or USRBDX that scores current can be larger than 1 if there are large angle scattered particles (the deviation from fluence is orientation dependent of the region), but since the USRBDX fluence is inverse cosine weighted, the total integrated fluence response should be faithful and just be the number of anti-protons exiting per initial primary

  3. antiprotons are stable and do not decay, and they will only stop, annihilate, or elastic scatter, and do not have enough kinetic energy (only 100 keV) to produce additional antiprotons by inelastic scattering

  4. the only non-vacuum material in the geometry is the foil, so antiprotons could not somehow scatter back into the foil after exiting

So, basically, I think the most likely explanation is that I am misunderstanding what this reported number means, since from the above there is (seemingly) no way that a degrading foil at this low of energy would increase the number of antiprotons, unless there is something very basic I am forgetting about annihilation physics, haha.

Any advice? At the end of the day, I’m basically hoping to know how many antiprotons exit the foil, and how many make it to a downstream region, so that I can determine the percent transmittance of the foil. I also want to use the USRBDX to find the mean energy and spread of the outgoing anti-protons using a post-processing script I am writing, but this should be understood first.

Input files

Here is the .flair file, pbarfoil.flair, and an input deck that yields anomalously high integrated response, 1100nm_01.inp.

pbarfoil.flair (13.2 KB)

1100nm_01.inp (4.7 KB)

Thanks very much for your time and help, and hopefully I just made a basic mistake here.

--Will

3

Dear Will,

It’s the opposite: The current should be faithful to the number of outgoing particles, not the fluence which is inverse-cosine weighted as you say.

Find attached a small modification of your geometry that illustrates the difference. I have created another USRBDX scorer exCur that - except from being a current scorer - has the same parameters as your exFlu scorer. Furthermore, I have idealized the beam (no transverse spread), given it a small angle, and set the foil material to vacuum.

cosx = 0.1 in the BEAMPOScard corresponds to an angle of about

α =90 deg - arccos(0.1) ~ 5.74 deg

w.r.t. the beam (z) axis. Since no interactions occur in the “vaccum foil”, this is the angle at which all particles exit the foil. For an exiting angle of about 5.74 degrees, the fluence-weighting factor would be about

1.0/cos(α) ~ 1.0050392777811088

Indeed, the fluence scorer returns 1.005051 exiting antiprotons per primary, while the current scorer returns 1.000014. The latter is the same value that you obtained with a fluence scorer and a zero-degree incidence angle.

Although the uncertainty estimate is practically zero for this scenario, both the fluence and the current are off from the known true value by about 0.001 %. I would brush this aside as a rounding issue.

Cheers,

Udo

1100nm_01_01_modified.inp (4.9 KB)

4

Hi Udo, thank you very much, this explains everything! It was a rookie mistake, but the upshot is it’s very easy to fix then, so I’ll try to suspend my embarrassment, haha!

And thank you for updated input file, that concrete example is very illustrative. So, what I gather from this then is that in the sum.lis file, the total response (part/pr) in the case of a fluence USRBDX scoring will still have the averaged inverse cosine weight, and total response in this sense is still not particle counting, BUT the point of that reported number is that energy and angular dependencies have been properly integrated out. In the case of current, it really is particle counting, and the orientation dependence means something different than I thought, whoops!

Thanks again Udo, I really appreciate it, and now I can just turn the crank for all these sims!

--Will