University of Chicago Reconstruction of Cherenkov Light With Precision Timing Matt Wetstein Enrico Fermi Institute, University of Chicago presenting work by Ioana Anghel, Erika Cantos, Mayly Sanchez, Matt Wetstein, Tian Xin Water Based Scintillator Workshop May 18-19, 2014
Full Track Reconstruction: A TPC Using Optical Light? first 2 radiation lengths of a 1.5 GeV π 0 → γ γ first 2 radiation lengths of a 1.5 GeV π 0 → γ γ first 2 radiation lengths of a 1.5 GeV π 0 → γ γ first 2 radiation lengths of a 1.5 GeV π 0 → γ γ mm WbLS Workshop -May 18-19, 2014 2
Full Track Reconstruction: A TPC Using Optical Light? first 2 radiation lengths of a 1.5 GeV π 0 → γ γ reconstructed “Drift time” of photons is fast compared to charge in a TPC! ~225,000mm/microsecond Need fast timing and new algorithms mm Image reconstruction, using a causal “Hough Transform” (isochron method) WbLS Workshop -May 18-19, 2014 3
Full Geant (WCsim) Study of Track Reconstruction Work by I. Anghel, E Cantos, M. Sanchez, M Wetstein, T. Xin How does vertex l γ path length = 10 m resolution scale with Λ = 250 nm γ path length = 30 m Λ = 365 nm γ path length = 50 m Λ = 445 nm Λ = 545 nm timing? Detector size? Λ = 575 nm Understanding the answer requires a complete understanding of optical effects like chromatic dispersion. time resolution = 2.0 ns time resolution = 1.0 ns We performed a full MC time resolution = 0.5 ns time resolution = 0.2 ns study using WCSim, using a time residual approach, capturing these effects. WbLS Workshop -May 18-19, 2014 4
Results Work by I. Anghel, E Cantos, M. Sanchez, M Wetstein, T. Xin 16 1.2 GeV muon in a 200 kton WCh 1.2 GeV electron in a 200 kton WCh 16 transverse vertex resolution (cm) transverse vertex resolution (cm) 14 14 12 with muon scattering 12 10 10 8 8 muon scattering o fg 6 6 4 4 2 2 0 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 time resolution (nsec) time resolution (nsec) • Vertex resolution is most sensitive to timing, in the plane transverse to the track direction • We see significant improvements in transverse vertex sensitivity with improvements in timing. • Diminishing returns on fast photosensors don’t set in until well below 500 psec WbLS Workshop -May 18-19, 2014 5
Results 8 transverse vertex resolution (cm) 7 6 Vertex resolution scales with 5 coverage consistent with sqrt(N). 4 3 2 20 1 transverse vertex resolution (cm) 18 0 4 6 8 10 12 14 16 18 20 16 percent photocathode coverage 14 12 10 Resolution losses, even going 8 from a 200 kton to 500 kton 6 detector are small. 4 2 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 time resolution (nsec) Work by I. Anghel, E Cantos, M. Sanchez, M Wetstein, T. Xin WbLS Workshop -May 18-19, 2014 6
Background Separation Centimeter level resolutions may allow better identification of the two forward gammas from a boosted pion. γ Two boosted gammas γ overlap. Unable to BOOST π 0 π 0 distinguish two separate γ rings. Looks like a single electron γ ~1 radiation length ~1 radiation length ~37 cm ~37 cm 16 transverse vertex resolution (cm) π 0 14 12 10 8 6 4 approx. vertex separation at LBNE energies at 7 degrees (median): ~4.5 cm 2 at 15 degrees (mean) : ~9.7 cm 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 time resolution (nsec) WbLS Workshop -May 18-19, 2014 7
Conclusion • Timing not only helps to separate the prompt component in scintillation detectors - it helps significantly improve the granularity of the reconstruction using that Cherenkov light. • Di fg erential measurements (transverse vertex or separating between two vertices) are most sensitive to timing, • These capabilities are scalable to very large detectors • Diminishing returns on timing don’t set in until the few hundred picosecond regime • This work does not yet look at granularity of the photosensors which are expected to bring in even more capabilities. WbLS Workshop -May 18-19, 2014 8
Backup Slides WbLS Workshop -May 18-19, 2014 9
DIGITAL Photon Counting LAPPDs are essentially digital photon counters • One can separate between photons based on spatial and time separation in a single photosensor (charge not even very necessary) s 2 WbLS Workshop -May 18-19, 2014 10
DIGITAL Photon Counting LAPPDs are essentially digital photon counters • One can separate between photons based on spatial and time separation in a single photosensor (charge not even very necessary) with conventional PMTs with hires imaging tubes • Measure a single time-of-first- • Measure a 4-vector for each light and a multi-PE blob of individual photon charge • Likelihood based on simultaneous • Likelihood is factorized into fit of space and time light separate time and charge fits m • one can separately test each • History of the individual photons photon for it’s track of origin, is washed out color, production mechanism (Cherenkov vs scintillation) and propagation history (scattered vs direct) WbLS Workshop -May 18-19, 2014 11
Timing and Cherenkov Geometry Based on pure timing, vertex position along the direction parallel to the track is unconstrained, due to ambiguity between vertex position and unknown T 0 . All Cherenkov light is forward wrt the track direction, so di fg erential timing is not possible. s 1 s 2 d casually consistent vertex hypothesis true vertex: point of (albeit non-physical) first light emission T 0 ’= T 0 - dn/c Must used additional constraint: fit the “edge of the cone” (first light) WbLS Workshop -May 18-19, 2014 12 Project funded under NSF CAREER
Timing and Cherenkov Geometry Position of the vertex in the direction perpendicular to the track is fully constrained by causality. Ambiguity in T 0 cancels when looking at light spreading outward in opposite directions. s 1 s 2 casually consistent vertex hypothesis For single vertex true vertex: point of (albeit non-physical) first light emission fitting, we expect the T 0 ’= T 0 - dn/c transverse resolution to improve significantly with photosensor time-resolution! WbLS Workshop -May 18-19, 2014 13 Project funded under NSF CAREER
Timing and Cherenkov Geometry Fortunately, multi-vertex separation is a di fg erential measurement. The ambiguity in T 0 cancels out. Causality arguments are fully su ffj cient to distinguish between one and two vertices. s 1 s 1 s 2 Only one unique solution that can satisfy the subsequent timing of both tracks 100 picoseconds ~ 2.25 centimeters WbLS Workshop -May 18-19, 2014 14 Project funded under NSF CAREER
Results 1.2 GeV muon in a 200 kton WCh 35 30 total vertex resolution (cm) parallel 25 20 15 transverse 10 5 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 time resolution (nsec) • transverse component is most sensitive to timing (as expected) Work by I. Anghel, E Cantos, M. Sanchez, M Wetstein, T. Xin WbLS Workshop -May 18-19, 2014 15
Do These Approaches Scale Up? 0.1 √ √ 0.9 Smeared RiseT [ns] ● ● Time resolution [cm] ] Over large length scales, optical transport of light in time resolution = 2.0 ns water becomes a problem. So does the cost of time resolution = 1.0 ns time resolution = 0.5 ns instrumenting large volumes with photosensors. time resolution = 0.2 ns How well can the concept of detailed track reconstruction scale to detectors with many 100s of kilotons of water (and low coverage)? I. Anghel, E. Cantos, G. Davies, M. Sanchez, M. Wetstein, T. Xin WbLS Workshop -May 18-19, 2014 16
The Challenge: We Are Technologically Limited Neutrino experiments often face tough choices. Water Cherenkov tradeo ff curves of constant sensitivity mass Liquid Argon e ffj ciency WbLS Workshop -May 18-19, 2014 17
The Challenge: We Are Technologically Limited The Limits of Thinking Bigger The limits of thinking bigger Neutrino experiments often face tough choices. Water Cherenkov one particular budget mass Liquid Argon e ffj ciency WbLS Workshop -May 18-19, 2014 18
New Technology Can Have a Transformative Impact on Physics Sensitivity • The development of new technology could push this frontier forward. • New technology can create intermediate Water Cherenkov options. • New capabilities drive new physics. same budget mass new technology Liquid Argon e ffj ciency WbLS Workshop -May 18-19, 2014 19
Do These Approaches Scale Up? But, will the tools and techniques developed for ANNIE scale? Where do we go from there? WbLS Workshop -May 18-19, 2014 20
What about low energy and heavy particles that don’t make Cherenkov? Liquid scintillators, especially water soluble LAB, provide many advantages • Sensitivity to charged particles below Cherenkov threshold • proton recoils • p → K + ν • Much improved lepton energy resolution Water(Cherenkov( Liquid(Argon(TPC( � � � � Efficiency( Background( Efficiency( Background( p) → )e + π 0) 45%) 0.2) 45%)?) 0.1) � � p) → ) ν K + # 14%) 0.6) 97%) 0.1) p) → ) µ + K 0) 8%) 0.8) 47%) 0.2) nWnbar)) 10%) 21) ?) ?) • Very clear Cerenkov ring even without cut However, they also come with disavantages • loss of fast timing response • loss of the directional information contained in Cherenkov light WbLS Workshop -May 18-19, 2014 21
Recommend
More recommend
