Showermax Monte Carlo Studies

Transcription

1 Showermax Monte Carlo Studies Dustin McNulty Idaho State University Thanks to: Carlos Bula, Brady Lowe, Kevin Rhine Jan 24, 2015

2 Showermax Monte Carlo Studies Outline Reproduce 2008 stack results (benchmarking MC) New Showermax Design for MOLLER Concept Light guide geometry Baseline performance (MC) Optimizing Design and Prototyping Summary Dustin McNulty Showermax Monte Carlo Studies 1

3 Detector Ring Design Concept Dustin McNulty Showermax Monte Carlo Studies 2

4 Motivation Provides additional measurement of e-e ring flux Weights flux by energy = less sensitive to low energy/low light bkgds Dustin McNulty Showermax Monte Carlo Studies 3

5 MOLLER Showermax Development Benchmark new MC: Start with 2008 stack experience Apply qsim optical MC framework to the stack Try to reproduce Jan2008 data and compare with Piotr s simulation results Create baseline showermax design Modify stack/lg/pmt geometry for MOLLER Start with Piotr s optimal stack configuration (10 pieces: 5mm quartz, 2.4mm tungsten) Apply Mainz testbeam and Peiqing s lightguide experience Optimize baseline design Study dependence on numbers and thicknesses of W and quartz, and energy, position and angles of incident particles Build prototype and test with beam (at Mainz?, SLAC) Dustin McNulty Showermax Monte Carlo Studies 4

6 2008 Showermax stack Slide from Piotr Dustin McNulty Showermax Monte Carlo Studies 5

7 2008 Stack Detector Schematic stack Al mirror PVC Schematic from Piotr Dustin McNulty Showermax Monte Carlo Studies 6

8 Jan2008 Testbeam Setup and Conditions Thin and Stack dets rigidly mounted along 45 angle; installed above rhrs focal plane (between VDC s and S1 scint. plane) E beam = 956 MeV, 5-50µA, 100mg/cm 2 Ta target rhrs at 19, using VDC s and s0 trigger (removable) Counting rates 10 Hz/µA Dustin McNulty Showermax Monte Carlo Studies 7

9 Jan2008 Testbeam Pulse Height Dists Thin Det (Raw, 3686, trig s0) hthin_raw Entries Mean RMS Stack Det (Raw, 3686, trig s0) hstk_raw Entries Mean RMS Thin Det (ped sub, 3686, trig s0) hthin 400 Entries 8915 Mean RMS χ / ndf / Lwidth 1.9 ± MPV ± 0.13 Integral 8381 ± GSigma ± Stack Det (ped sub, 3686, trig s0) 90 Entries Mean RMS χ 2 / ndf / 272 Lwidth 1.69 ± MPV ± Integral 6445 ± GSigma ± hstk Sun Jul 20 22:17: Stack performance: 51.7/144 = (raw) Dustin McNulty Showermax Monte Carlo Studies 8

10 Piotr s MC: Tungsten and Quartz thickness study (for 900MeV electrons) Dustin McNulty Showermax Monte Carlo Studies 9

11 ' $ January 2008 PREx detectors tests, comparison with simulations Thin 5mm Thin 10mm exp: 50±17(34%) exp: 93±22(24%) sim: 56.8±7.7(14%) sim: 109.2±11.3(10%) Stack exp: 434±152(35%) sim: 538.3±140.5(26%) Number of Cherenkov photons reaching PM NPE=0.2Nph (<QE>=0.2) In panels below widths of gaussian fits to the simulated Nph distributions (red lines) are corrected for the PMT resolution according to the formula (for the used PMTs measured value of the δq is 0.23): exp: 0.71±0.24(34%) pc exp: 1.31±0.31(24%) pc exp: 7.74±2.71(35%) pc sim: 0.80±0.26(32%) pc sim: 1.53±0.36(24%) pc sim: 9.61±2.68(28%) pc PM output charge (pc) & Dustin McNulty Slide from Piotr Showermax Monte Carlo Studies % 10

12 Recent Stack Simulations at ISU Dustin McNulty Showermax Monte Carlo Studies 11

13 2008 Testbeam Data and Simulation Summary Results for stack detector were lack-luster: 35% relative width Why? Because energy too low? probably not Or some other reasons?...det alignment, design... more likely Both simulations give reasonable agreement with real data at 15-20% level Simulations are also in reasonable agreement with each other: Piotr found 538 Cer. photons reach PMT per electron with 28% relative width ISU found 495 photons with 32% rel width Experiment yielded 434 photons with 35% rel. width Dustin McNulty Showermax Monte Carlo Studies 12

14 MOLLER Showermax Development Benchmark new MC: Start with 2008 stack experience Apply qsim optical MC framework to the stack Try to reproduce Jan2008 data and compare with Piotr s simulation results Create baseline showermax design Modify stack/lg/pmt geometry for MOLLER Start with Piotr s optimal stack configuration (10 pieces: 5mm quartz, 2.4mm tungsten) Apply Mainz testbeam and Peiqing s lightguide experience Optimize baseline design Study dependence on numbers and thicknesses of W and quartz, and energy, position and angles of incident particles Build prototype and test with beam (at Mainz?, SLAC) Dustin McNulty Showermax Monte Carlo Studies 13

15 Crude Sketch of MOLLER Showermax Concept Dustin McNulty Showermax Monte Carlo Studies 14

16 Showermax Detector Ring Dustin McNulty Showermax Monte Carlo Studies 15

17 Showermax Quartz Spectrosil 2000 each piece has single 45 deg bevel across long end Dustin McNulty Showermax Monte Carlo Studies 16

18 Showermax Lightguide (Bad Examples) *Lightguide design non-trivial and perhaps most important feature Dustin McNulty Showermax Monte Carlo Studies 17

19 Showermax Lightguide (Good) Dustin McNulty Showermax Monte Carlo Studies 18

20 Showermax Lightguide Development Dustin McNulty Showermax Monte Carlo Studies 19

21 Showermax Lightguide Development Dustin McNulty Showermax Monte Carlo Studies 20

22 Showermax MC LG Study 2 GeV electron beam centered on quartz face; normal incidence Dustin McNulty Showermax Monte Carlo Studies 21

23 Baseline Showermax MC Results: Lightguide C 2, 5, and 8 GeV e uniformly sampled over quartz face; normal Dustin McNulty Showermax Monte Carlo Studies 22

24 What if thickness of W is halved? 2, 5, and 8 GeV e uniformly sampled over quartz face; normal Dustin McNulty Showermax Monte Carlo Studies 23

25 What if just use single piece of W and Quartz? Answer: Doesn t work well at low energy (but way cheaper!) Dustin McNulty Showermax Monte Carlo Studies 24

26 Single piece Design versus Baseline Design at 2GeV Dustin McNulty Showermax Monte Carlo Studies 25

27 Single piece Design versus Baseline Design at 8GeV Dustin McNulty Showermax Monte Carlo Studies 26

28 MOLLER Showermax Development Benchmark new MC: Start with 2008 stack experience Apply qsim optical MC framework to the stack Try to reproduce Jan2008 data and compare with Piotr s simulation results Create baseline showermax design Modify stack/lg/pmt geometry for MOLLER Start with Piotr s optimal stack configuration (10 pieces: 5mm quartz, 2.4mm tungsten) Apply Mainz testbeam and Peiqing s lightguide experience Optimize baseline design Study dependence on numbers and thicknesses of W and quartz, and energy, position and angles of incident particles Build prototype and test with beam (at Mainz?, SLAC) Dustin McNulty Showermax Monte Carlo Studies 27

29 Prel. Prototype design using 2008 Stack pieces MC studies needed to understand potential benefits of beamtest Dustin McNulty Showermax Monte Carlo Studies 28

30 Summary and Future Work Optical MC framework for showermax R&D established Reasonable/good agreement with 2008 beamtest data and Piotr s simulations Baseline MOLLER showermax detector design established Gives strong energy dependent light yields with 25% relative width Optimization of baseline design underway Build prototype based on optimized design and test with beam Other questions/considerations: 90 LG or 135 (or 45 ) LG (Need to decide) Need to worry about sensitivity to neutrons, pions PE uniformity/edge effects due to transv. shower leakage Stray electrons, splashback,...need optical MC in remoll Dustin McNulty Showermax Monte Carlo Studies 29