SN trigger requirement changes and latency Pierre Lasorak & - - PowerPoint PPT Presentation

SN trigger requirement changes and latency

1

Pierre Lasorak & Simon Peeters

Pierre Lasorak 13/09/2019

Outline

• SN requirement changes in the TDR
• SumADC problem
• Future work

2

Pierre Lasorak 13/09/2019

SN requirement change

• The requirement for SN triggering efficiency was changed:
• Old: > 90% efficiency for SNB within 100 kpc
• New: > 95% efficiency for a SNB producing at least 60 interactions

with neutrino energy > 10 MeV in 12 kT of active detector mass during the first 10 seconds of the burst.

• Getting rid of the model dependance:
• What is the distribution of probability for SN to happen within 100 kpc
• How many neutrino events at 100 kpc

3

Pierre Lasorak 13/09/2019

Sensitivity 60 events - real spectrum

• Sum ADC PDF integrated over 10 seconds:
• Note that these plots are for the full MCC11 signal statistics, not with

the extra high-rate background-only samples

• Note the integral of these histograms is the number of cluster you

expect to see:

• Background rate x 10 seconds (right)
• Background rate x 10 seconds + 60 SN events x efficiency (left)

4

Background only

20 40 60 80 100 120 140 160 180 200 SADC 0.5 1 1.5 2 2.5 3 3.5 4

PDF

20 40 60 80 100 120 140 160 180 200 SADC 0.2 0.4 0.6 0.8 1 1.2 1.4

PDF

Signal+Background

Caveats:

• Neutron background is at 40 Hz

(could be as high as ~400 Hz)

• Neutron spectrum is wrong

(probably low impact)

• Radon quenching is wrong

(probably low impact)

• No Beta-Alphas correlations

(Bismuth/Polonium)

Pierre Lasorak 13/09/2019

Likelihood distributions 60 events - real spectrum

• Throw according to the 2 histograms from previous slide, and build a likelihood to compare with the background-only hypothesis.
• Black line → 1 fake trigger per month
• With 60 events, it’s very hard for the 2 plots on previous slide to look the same.
• Triggering efficiency is excellent.

5

Pierre Lasorak 13/09/2019

20 40 60 80 100 120 140 160 180 200 Sum ADC 0.5 1 1.5 2 2.5 3 PDF

Sensitivity 60 events - 10 MeV neutrinos

• PDF of the Background and Signal+Background.
• Again the normalisation corresponds to the number of event

you’d expect (with the efficiency at 10 MeV).

• You need to be very “unlucky” to get no events after sum ADC =

100 and have Lbackground ~ Lsignal

6

20 40 60 80 100 120 140 160 180 200 SADC 0.2 0.4 0.6 0.8 1 1.2 1.4

PDF

Background only (unchanged) Signal+Background

Pierre Lasorak 13/09/2019

Likelihood distributions

• As can be anticipated the likelihoods look very different

and good separation can be achieved.

7

D e p r e c a t e d

Pierre Lasorak 13/09/2019

Comment

• The reason this works well is because the

distribution of sum ADC doesn’t change much.

• Red: ~10 MeV neutrinos (i.e. ~5MeV electrons)
• Blue: full spectrum
• This is weird as Josh was pointing out:
• Left: SumADC for single electrons at the cathode
• Right: SumADC for neutrino events
• The SumADC should be roughly the same!

8

20 40 60 80 100 120 140 160 180 200 SADC 0.5 1 1.5 2 2.5 3 3.5 4

PDF

Pierre Lasorak 13/09/2019

Investigating this discrepancy

• Neutron gamma spectrum
• True captures on LAr:

9

5 10 15 20 n gamma 100 200 300 400 500 Count

Gamma multiplicity Gamma multiplicity

2 4 6 8 10 12 14 16 18 20 Energy [MeV] 20 40 60 80 100 120 140 160 Count

Highest energy gamma Highest energy gamma

2 4 6 8 10 12 14 16 18 20 Energy [MeV] 200 400 600 800 1000 1200 Count

Summed gamma energy Summed gamma energy

2 4 6 8 10 12 14 16 18 20 Energy [MeV] 100 200 300 400 500 600 700 800 Count

Gamma individual energy Gamma individual energy

• On average:
• 3 gammas
• ~2 MeV
• Summing up to

~6 MeV (or 8.8 MeV if the capture was on a different isotope

• f Argon)
• Highest photon

energy 3-4 MeV.

Pierre Lasorak 13/09/2019

Down the GEANT4 tree…

• Compton scattering producing electrons:
• At least a 3 MeV Compton electron on average for the

neutron capture.

10

2 4 6 8 10 12 14 16 18 20 Energy [MeV] 500 1000 1500 2000 2500 3000 Count

Electron individual energy Electron individual energy

2 4 6 8 10 12 14 16 18 20 Energy [MeV] 100 200 300 400 500 Count

Highest energy electron Highest energy electron

Pierre Lasorak 13/09/2019

Comparing spectra

• Reweighted the SumADC distribution from ELep

to the highest energy Compton electron

• Blue is what neutrons should look like with the

same SumADC vs ELep dependance as SN neutrino events.

• Red is what the neutrino events at 10 MeV are…

11

2 4 6 8 10 12 14 16 18 20 Energy [MeV] 100 200 300 400 500 Count

Highest energy electron Highest energy electron

5 10 15 20 25 30 35 40 45 50 MeV

E 10000 20000 30000 40000 50000 60000 Counts

< 10.5 MeV

ν

9.5 MeV < E

neutron

20 40 60 80 100 120 140 160 180 200 SumADC 0.02 0.04 0.06 0.08 0.1 0.12 0.14 Count

neutron

neutron

neutron

Pierre Lasorak 13/09/2019

Mini production

• I redid a small production to get the SumADC distribution out of 10MeV neutrinos.
• “True SumADC” of all the hits in the events that are backtracked to SN:
• Left: mini production with 1 SN neutrino per event
• Right: MCC11 (3 neutrinos per events) from the analyser
• In the process of understanding where the bug is coming from.
• Mapping of neutrino interaction → hits looks good
• Of course the Likelihood distributions now don’t look so great.

12

SADC Entries 1000 Mean 29.62 Std Dev 12.96 SumADC Entries 425 Mean 59.95 Std Dev 47.82

SumADC Entries 425 Mean 59.95 Std Dev 47.82

<10MeV

E

Pierre Lasorak 13/09/2019

Future work

• The requirement for the energy can be changed (up

to 15 MeV).

• David Rivera has shown at DAQ Physics

performance meeting that you can find at the trigger level 5 MeV electrons with 30% efficiency by taking into account:

• The wire span of the cluster
• The time span of the cluster
• Planning to include these variables in the trigger

and hope to achieve the efficiency that was quoted:

• Creating a “range variable”/size of the track

variable (like David Rivera)

• Using all the variables in the Likelihood (multi-

dimensional PDFs)

• A selection on these variables can achieve good

efficiency for solar neutrinos as well!!

13

DRivera DRivera

Pierre Lasorak 13/09/2019

Conclusion

• Reprocessing the 10 MeV sensitivities needed due to a

problem in the production.

• Investigations are on going to get to the bottom of this.
• There are ways to improve the clustering and burst

trigger that should make it meet the requirement.

14

Pierre Lasorak 13/09/2019

ENu-ELep correlations

15

5 10 15 20 25 30 35 40 45 50 MeV

E 5 10 15 20 25 30 35 40 45 MeV

E 5 10 15 20 25 30 35 40 45 50 MeV

E 5 10 15 20 25 30 35 40 45 50 MeV

E 10000 20000 30000 40000 50000 60000 70000