ALICE Software Tracking Update

1. ALICE Software Tracking Update Tom Junk and Jen Raaf GArTPC Meeting March 9, 2018

2. same as last week Sample Definition • I ran a similar set as Jen's example, but changed to: • All events contain one 𝜇- coming out of a particle gun • The initial starting radius is uniformly distributed between 50 and 150 cm. Inner radius is at ~85 cm. The outer radius is about 250 cm • The pT distribution is uniform between 10 MeV and 2 GeV. Jen tells us that alisoft won't work well below 50 MeV so I wanted to map it out. • 𝜂 = -ln(Tan(𝜃/2)) where 𝜃 is the angle wrt the z axis ("beam axix, parallel to B and E") is uniformly distributed between -3 and 3 • 𝜙position and 𝜙momentum are uniformly distributed between 0 and 2𝜋. Jen's script had 𝜙momentum = 𝜙position + 𝜋/2 but now they're independent. • z of particle origin is uniformly distributed from -100 cm to +100 cm • 5000 events, each with one muon (pizeros later in this talk) T. Junk, J. Raaf | ALICE Reco Studies

3. Efficiency Dependency on PT, |𝜂| and R using "beam-off" ALICE reco. arxiv:1507.07577 (high-mult cosmic-ray paper) 50-degree cut wrt zenith: |𝜂|<1 QA plots https://twiki.cern.ch/twiki/bin/view/Main/QA:ChargedTracks don't go out past |𝜂|=1 T. Junk, J. Raaf | ALICE Reco Studies

4. Residual Fits in PT bins: 𝜂 T. Junk, J. Raaf | ALICE Reco Studies

5. Residual Fits in PT bins: 𝜙 T. Junk, J. Raaf | ALICE Reco Studies

6. Residual Fits in PT bins: 1/PT T. Junk, J. Raaf | ALICE Reco Studies

7. Look for Dalitz Decays: 𝜋0 →𝛄e+ e- • Modify original particle gun to shoot 𝜋0 particles, otherwise the same as the original muon sample: • The initial starting radius is uniformly distributed between 50 and 150 cm. Inner radius is at ~85 cm. The outer radius is about 250 cm • The pT distribution is uniform between 10 MeV and 2 GeV. Jen tells us that alisoft won't work well below 50 MeV so I wanted to map it out. • 𝜂 = -ln(Tan(𝜃/2)) where 𝜃 is the angle wrt the z axis ("beam axix, parallel to B and E") is uniformly distributed between -3 and 3 • 𝜙position and 𝜙momentum are uniformly distributed between 0 and 2𝜋. Jen's script had 𝜙momentum = 𝜙position + 𝜋/2 but now they're independent. • z of particle origin is uniformly distributed from -100 cm to +100 cm Use Cosmic Reco Settings from Ruben for all of this. T. Junk, J. Raaf | ALICE Reco Studies

8. Run lots more 𝜋0's • Shoot 200 𝜋0's per event. MC Truth will tell us which particles belong to which 𝜋0 Sample Event T. Junk, J. Raaf | ALICE Reco Studies

9. (x,y) Locations of Conversions and Dalitz Decays Includes Compton Scatters T. Junk, J. Raaf | ALICE Reco Studies

10. Conversion Points in z and (r,z) (may include Compton scatters – any e+ or e- with a photon as a parent) T. Junk, J. Raaf | ALICE Reco Studies

11. Just the Photon Conversions 200 𝜋0 per "event" Some "tridents" Photons convert on atomic electrons. T. Junk, J. Raaf | ALICE Reco Studies

12. Just the Compton Scatters 200 𝜋0 per "event" T. Junk, J. Raaf | ALICE Reco Studies

13. Photon Conversion and Compton Electron Energies T. Junk, J. Raaf | ALICE Reco Studies

14. A 200 𝜋0 Event T. Junk, J. Raaf | ALICE Reco Studies

15. "default" window, TPC Hits turned on A Better Event Display The hits are all there! I think they are sim hits though Hole in the middle T. Junk, J. Raaf | ALICE Reco Studies

16. Getting away from ALICE sw • Export 3D reco points: TPC.RecPoints*.root • Corresponding event to one on previous page Looks like I ran into the segmentation issue. There's a "detector" variable in it which goes form 0:71. (4*18-1). Inner, outer, +z, -z. T. Junk, J. Raaf | ALICE Reco Studies

17. Reading Reco Points in a Standalone Application • ALICE loader and geometry tools willrequire working in ALICE's framework • Export reco points and reconstruct them with a rotation through an angle 𝜙 = (fDetector % 18)x(2𝜋/18) Now do pattern recognition, tracking, vertexing in a convenient development environment! T. Junk, J. Raaf | ALICE Reco Studies

18. Moving Forwards Simplest way forwards to answer immediate questions:1.    What is the tracking efficiency as a function of track angle, especially when the track is along the wire in the case of the STT?2.     What is the energy threshold for detecting protons/pions and how important is this to reduce systematics?3.     Momentum/angular resolution of all particles;4.     Pi0 energy resolution and efficiency in neutrino interactions (overlays, multiple interactions & background from rock); 5.     What are the uncertainties in energy scale for e/mu? 6.     What are the uncertainties in angular resolution? 7.     What are the key performance parameters for the ECAL to detect NC pi0? 8.     Can neutrons be detected in the tracker? 9.     What is the expected performance for key channels (see appendix A) including yields? is to simulate neutrino scatters in ALICE, "avoiding" the hole, using some pattern recognition (to write...) on exported reco points. Questions: What does 10 bar do to the patrec & resolutions? What is the effect of the hole? (we won't have one eventually but ALICE code has it now). T. Junk, J. Raaf | ALICE Reco Studies

19. Moving Forwards (2) • But this is not a long-term solution. Would like our own softwarethat • Fills in the hole • Has 10 bar of gas • Reco the data products made by our own sim • optimized for lower-multiplicity events that start anywhere • Looks more like the rest of DUNE code (built on art) • We support it and do not depend on ALICE • Can steal ideas from ALICE software • But this won't answer the questions quickly. T. Junk, J. Raaf | ALICE Reco Studies