Problem 2108

Summary: Narrow peaks in secondary pion energies
Product: Geant4 Reporter: Vladimir.Ivantchenko
Component: processes/hadronic/models/cascadeAssignee: dennis.herbert.wright
Status: RESOLVED WONTFIX    
Severity: normal CC: Alberto.Ribon, gandr, genser, resnegfk
Priority: P4    
Version: 10.5   
Hardware: All   
OS: All   
Attachments: Secondary pi-
Secondary pi+
Secondary pi0
macro
Pion production momentum talk

Description Vladimir.Ivantchenko 2018-12-10 11:32:58 CET
Created attachment 523 [details]
Secondary pi-

Krzysztof Genser reported on narrow peaks in energy spectra of secondary pions for 8 GeV proton beam in thick W target. He reported following:

Mu2e is observing a strange phenomenon related to pion production in inelastic interactions.  Namely, one observes some "delta like" peaks with very distinct momenta (one of them being ~88.073 MeV). See the attached plots. If one looks at the creator process, it is

protonInelastic
neutronInelastic
pi+Inelastic
pi-Inelastic
kaon-Inelastic
kaon0LInelastic
hBertiniCaptureAtRest
sigma+Inelastic
sigma-Inelastic
lambdaInelastic
kaon+Inelastic
kaon0SInelastic
xi0Inelastic
xi-Inelastic

The problem has been reproduced in Hadr01 example. Plots are attached as well as macro file.
Comment 1 Vladimir.Ivantchenko 2018-12-10 11:33:54 CET
Created attachment 524 [details]
Secondary pi+
Comment 2 Vladimir.Ivantchenko 2018-12-10 11:34:34 CET
Created attachment 525 [details]
Secondary pi0
Comment 3 Vladimir.Ivantchenko 2018-12-10 11:35:03 CET
Created attachment 526 [details]
macro
Comment 4 Vladimir.Ivantchenko 2018-12-10 11:35:41 CET
Note, that peaks for diffirent pions are at different energies.
Comment 5 dennis.herbert.wright 2019-04-12 19:15:08 CEST
I see that this test was run in 10.5.   Did you see the same thing in 10.4?
Comment 6 Krzysztof Genser 2019-04-12 21:37:35 CEST
Mu2e saw it with Geant4 10.4.p02 and QGSP_BERT physics list.
One gets similar results with ShieldingM or some earlier versions of Geant4
Comment 7 dennis.herbert.wright 2019-04-16 20:25:24 CEST
These peaks are all from the hadronic decay of kaons and hyperons.
Kaons, lambdas and sigmas are produced in the hadronic showers, nearly come to rest and then decay to pions and nucleons.   For pi+, we have

K+ -> pi+ pi0    Tpi+ = 108.4 MeV
Kshort -> pi+ pi-    Tpi+ = 109.2 MeV

This constitutes the peak in histogram 26.

For pi-,

Lambda -> p pi-   Tpi- = 32.7 MeV   as in histogram 27

For pi0,

Lambda -> n pi0    Tpi0 = 35.4 MeV
Kshort -> pi0 pi0  Tpi0 = 113.8 MeV
K+ -> pi+ pi0      Tpi0 = 110.5 MeV
K- -> pi- pi0      Tpi0 = 110.5 MeV

These make up the low energy and high energy peaks in histogram 28.

I did not see a plot for the 88 MeV line, but there is a 92.2 MeV line 
from Sigma+ -> n pi+  and the pi+ may have lost energy in the target.


So I think everything is fine and this is not a bug.
Comment 8 Andrei Gaponenko 2019-04-17 21:09:07 CEST
Created attachment 568 [details]
Pion production momentum talk

pi- distributions for 10M 8GeV protons
Comment 9 Andrei Gaponenko 2019-04-17 21:14:45 CEST
Hello,

Hyperon decays sound like a plausible explanation of the narrow peak
feature, but a more detailed look shows that it is not correct.  The
attached "Pion production momentum talk" zooms on the peak near
p=88.07315 MeV/c in the pi- production momentum.  (The corresponding
kinetic energy T=25.465 MeV.)  It looks like a delta-function, not
just a narrow peak.  But also slides 6 and 7 show that most of the
contribution to the line comes directly from the "protonInelastic"
process with the proton parent, not from a decay.  There are also
contributions from other particles and processes to the same line.
It really does look like a bug.

Andrei
Comment 10 dennis.herbert.wright 2019-04-18 00:18:46 CEST
OK, I'll have another look.
Comment 11 dennis.herbert.wright 2019-04-19 01:36:01 CEST
  As you stated, these lines are indeed coming from the inelastic proton process.
At 8 GeV these interactions are handled by the Bertini cascade.  The lines are, in fact, coming from the decay of hyperons almost at rest in the nucleus.  Strange pairs are produced in hadronic showers; one member of the pair may be a lambda, sigma, etc,
which gets trapped in the nuclear potential.  The correct thing to do in that case would be to create a hypernucleus, and let it decay weakly at sone later time.  However, Geant4 does not support hypernuclei.  
  To handle this, we chose to decay the hyperon hadronically in the nucleus and propagate the decay products.  So for lambda ->  p pi- the pi- momentum at decay is 100.58 MeV/c.  Since the decay happens in the nucleus, the pi- can either climb the nuclear potential, losing energy, or tunnel out, losing no energy.  Hence you see the 100.58 MeV/c line (no energy lost) and the more frequent 88 MeV/c line, which gave up a constant amount of energy getting out of the potential.  Both lines are sharp.  You also see a similar thing at about 182 and 193 MeV/c in the same spectrum: Sigma- -> n pi-, with p = 193 MeV/c.  The larger line at 182 is the pi- that had to climb over the potential.   
  This is not an ideal solution, but I'd have to say it's a feature, not a bug.  The whole issue of dealing with bound hyperons is really a new development project and there is no quick fix.