[PPT] - Photon Detection System (PDS) and SN triggering Pierre Lasorak 1 PowerPoint Presentation

SLIDE 1

Photon Detection System (PDS) and SN triggering

1

Pierre Lasorak

SLIDE 2

Pierre Lasorak 24/07/2018

Introduction Outline

Introduction
Final aim:
What to expect from the PDS for SN triggering?
Can we motivate from SN/DAQ point of view?
Addition of reflective foil on the cathode
Use of ARAPUCA / higher efficiency/granularity PDS
Photon detection in LAr
PDS hit level information
Clustering
Results
Other info:
MCC10 SN samples and geometry:

snb_timedep_dune10kt_1x2x6_snb_timedep_bkg_reco

Using Jason’s photon backtracker after recent fix (28th June)

2

SLIDE 3

Pierre Lasorak 24/07/2018

Introduction Photon detection in LAr

LAr scintillation, 2 components:
Fast light from singlet Ar2* state
Slow light from triplet state

3

Time [μs] Photoelectrons

Fast Component τ ≈ 8 ns (29%) Intermediate Component τ ≈ 140 ns (8%) Slow Component τ ≈ 1.6 μs (63%)

TallBo measurement

D. Whittington

Neutrino 2014 poster

Ar Ar+ Ar* Ar2*

(singlet)

Ar Ar Ar2*

(triplet)

Ar

e- μ-

Ar2+

e-

50% 50% 35% 65% 1.6 μs 7 ns 128 nm

Self-Trapped Exciton Recombination

ar et s and th This photon signal event time t0 for reconstruction. The ratio of con from the two com depends on βγ and can ai in particle identi

D. Whittington

Neutrino 2014 poster

Effect of HV,

impurities can change (marginally?) the timing of the different components.

SLIDE 4

Pierre Lasorak 24/07/2018

Event Displays (Time and Space)

LArSoft event (full drift

window)

Hit time distribution:
Arrow: true time of generation
Histogram: timing of the optical hits
Hit spatial distribution:
Pink line: wire hits backtracked to SN𝜉
Overlaid histogram: optical hits in the PDS backtracked to

SN𝜉

10 scintillation bars / APA
Red cross: true neutrino interaction position

4

4 − 2 − 2 4 6 8 10 s] µ Time [ 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 n Hits

Timing

X: 116.7 Y: -431.0 Z: 1198.2 energy: 12.5 MeV nHit: 8 ν

Timing

0.5 1 1.5 2 2.5 3 3.5 4

ν SN

200 400 600 800 1000 1200 Z Position [cm] 800 − 600 − 400 − 200 − 200 400 600 800 Y Position [cm]

energy: 24.6 MeV ν X Position: 232 cm N opt hits: 16 N wire hits: 6

ν SN

Other (not BT) ν SN APA CPA Ar39 Neutron Krypton Polonium Radon Ar42 AllBackground All

5 10 15 20 25 30 35

All

200 400 600 800 1000 1200 Z Position [cm] 800 − 600 − 400 − 200 − 200 400 600 800 Y Position [cm]

energy: 24.6 MeV ν X Position: 232 cm N opt hits: 1293 N wire hits: 386 All

… With all the hits:

SLIDE 5

Pierre Lasorak 24/07/2018

Hit level Information

Full drift window
OpHit
Unmatched hits: noise,

dark current

5

Photon Wire

1 −

10 1 10

2

10

3

10 nHit

1 −

10 1 10

2

10

3

10 nEvent

Other (not BT) ν SN APA CPA Ar39 Neutron Krypton Polonium Radon Ar42 AllBackground All

Number of hits per drift window

~1k events

SLIDE 6

Pierre Lasorak 24/07/2018

Hit level information Signal features

Efficiency:

1 or more optical hits from SN / N events

6

400 − 300 − 200 − 100 − 100 200 300 400 X position [cm] 0.5 0.6 0.7 0.8 0.9 1

ne or more hit collection efficiency

600 − 400 − 200 − 200 400 600 Y position [cm] 0.75 0.8 0.85 0.9 0.95 1

ne or more hit collection efficiency

5 10 15 20 25 30 35 40 Energy [MeV] ν 5 10 15 20 25 30 35 40 45 nHit / Event

Drops the further

you get from the APA

Number of hits

scales linearly with E𝜉

Edge effect,

photons escape

APA CPA CPA

0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 s] µ Time [ 0.05 0.1 0.15 0.2 0.25 N optical hits

h_timing_relat

Entries 17332 Mean 1.149 Std Dev 1.349

First hit + other  fast light hits Other  slow light hits

Top Bottom

SLIDE 7

Pierre Lasorak 24/07/2018

Other SNnu APA CPA Ar39 Neutron Krypton Polonium Radon Ar42 AllBackground All 1 −

10 1 10 average number of hits in cluster 1 2 3 4 5 6 7 8 9 10 s] µ Time [ 0.2 0.4 0.6 0.8 1 Fraction of hits

Clustering

7

I went ahead and clustered optical hits:
Reused the code from Alex Booth:
Timing: 800 ns (maybe too small wrt the simulations)
Z position: 300 cm (1 APA)
No Y clustering
Composition of the clusters:
Biggest contributor of hit to tag as SN or not.
Neutron is still the worst background
Background contributions are more evenly spread out → Pile up

is important! (Unlike for wire cluster where the main contributor is neutrons)

signal optical clusters background optical clusters

SLIDE 8

Pierre Lasorak 24/07/2018

signal optical clusters background optical clusters

1 2 3 4 5 6 7 8 9 10 s] µ Time Width [

1 −

10 1 10

2

10

3

10

4

10

5

10 Clusters 200 400 600 800 1000 1200 Z Width [cm]

1 −

10 1 10

2

10

3

10

4

10

5

10 Clusters h_width_sign_opti Entries 2306 Mean 489.2 Std Dev 321.3 200 400 600 800 1000 1200 1400 Y Width [cm]

1 −

10 1 10

2

10

3

10

4

10

5

10 Clusters h_ywidth_sign_opti Entries 2306 Mean 545.8 Std Dev 297.6

1 10

2

10

3

10 n clusters

1 −

10 1 10

2

10

3

10 Events

h_ncluster_sign_opti

Entries 1874 Mean 1.231 Std Dev 0.7394

Cluster properties

8

12m clusters!! ~ Size of the 1x2x6 → not appropriate for these studies

Splitting the SN events!
First pass, still trying to get better at clustering time

properly.

Number of background clusters is large without cuts

Something smarter has to be done for the time

nClusters / Drift Windows

SLIDE 9

Pierre Lasorak 24/07/2018

Clustering

Clustering efficiency (no cut)
Nhit (and n PEs) can be used to suppress

backgrounds

Next: use 10, 12, 15, 17 hits as cut (still missing stats

to go further)

9

5 10 15 20 25 30 [MeV] ν E 0.2 0.4 0.6 0.8 1 Efficiency

10 20 30 40 50 60 70 80 90 n Hits

1 −

10 1 10

2

10

3

10

4

10

5

10 Clusters

h_nhit_sign_opti Entries 2306 Mean 13.61 Std Dev 13.34

2 4 6 8 10 12 14 16 18 20 n PEs

1 −

10 1 10

2

10

3

10

4

10

5

10 Clusters

h_npe_sign_opti Entries 2307 Mean 10.17 Std Dev 6.979

signal optical clusters background optical clusters

SLIDE 10

Pierre Lasorak 24/07/2018

Results

Use a 5s timing window to count the number of

clusters

Can trigger on almost all of the milky way using only

PDS info!

Trying to get to the LMC where you get 10-20 events.

10

Individual Marley Eff & 10kt Bkgd Rate Optical Custers (nHit>= 10): - Eff: 0.57 & Bkgd rate: 100.05 Hz (5 s timing window) Optical Custers (nHit>= 12): - Eff: 0.51 & Bkgd rate: 25.48 Hz (5 s timing window) Optical Custers (nHit>= 15): - Eff: 0.43 & Bkgd rate: 1.85 Hz (5 s timing window) Optical Custers (nHit>= 17): - Eff: 0.38 & Bkgd rate: 0.93 Hz (5 s timing window) Wire Clusters - Eff: 0.58 & Bkgd rate: 0.10 Hz (5 s timing window) Wire Clusters - Eff: 0.58 & Bkgd rate: 0.10 Hz (10 s timing window)

1 10

2

10 Number of Clusters/Time Window

9 −

10

8 −

10

7 −

10

6 −

10

5 −

10

4 −

10

3 −

10

2 −

10

1 −

10 1 10

2

10 Trigger Rate, (Hz)

Number of Clusters in Time Window Required to Trigger vs. Trigger Rate

1/Month 1/Week 1/Day

10 20 30 40 50 SN Distance, (kpc)

5 −

10

4 −

10

3 −

10

2 −

10

1 −

10 Efficiency x SN Probability

Galactic Neighbourhood Coverage, Fake Trigger Rate 1/Month Galactic Neighbourhood Coverage, Fake Trigger Rate 1/Month

SLIDE 11

Pierre Lasorak 24/07/2018

Conclusion Future work

First pass at using PDS for SN trigger
Basic simple clustering implemented (needs improvement)
Currently can trigger on Milky Way SN but not on the 20% of SN coming from LMC
1x2x6 geometry is not really big enough to avoid bias in PDS (light leaks out sides)
Photons travel far → triggering cannot be done so efficiently on the APA level (few APAs at least).
The slow component of the light is important.
Future work
Motivate improved design for the PDS:
Consider the addition of reflective foil on the cathode
More granular/efficient detector can do better?
Combine PDS with wire information at the trigger level?

11