The MicroBooNE Continuous Readout Stream for the Detection of - - PowerPoint PPT Presentation

▶
the microboone continuous readout stream for the
SMART_READER_LITE
LIVE PREVIEW

The MicroBooNE Continuous Readout Stream for the Detection of - - PowerPoint PPT Presentation

Oct 28, 2022 119 likes •474 views

The MicroBooNE Continuous Readout Stream for the Detection of Supernova Neutrinos Iris Ponce Columbia University On Behalf of the MicroBooNE Collaboration 1 What is MicroBooNE? MicroBooNE is an 89 active ton liquid-argon time

slide-1
SLIDE 1

The MicroBooNE Continuous Readout Stream for the Detection of Supernova Neutrinos

Iris Ponce – Columbia University On Behalf of the MicroBooNE Collaboration

1

slide-2
SLIDE 2

What is MicroBooNE?

  • MicroBooNE is an 89 active ton

liquid-argon time projection chamber (LArTPC)

  • Along the Booster Neutrino

Beamline (BNB) located at Fermilab

  • Collecting data since 2015

Iris Ponce - Columbia University 2

slide-3
SLIDE 3

LArTPC Technology

Iris Ponce - Columbia University 3

1. 2 4 3 3 5 5 4

  • 1. Neutrinos enter the TPC
  • 2. Neutrino interacts with an

Argon nucleus and creates charge particles

  • 3. The particles created ionize

the argon and release ionization electrons

  • 4. The electric field inside the

TPC drifts the electrons toward the wire planes

  • 5. The electrons are recorded as

signals in the wires

slide-4
SLIDE 4

MicroBooNE Research Goals

  • Investigate the Low Energy

Excess (electron-like events) seen in MiniBooNE

  • Measurement of neutrino-Argon

cross-sections

Iris Ponce - Columbia University 4

1 10

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6

  • A. Schukraft, G. Zeller

00 150 200 250 300 350

X

  • µ

→ N

µ

ν X

+

µ → N

µ

ν

100

(GeV)

ν

E

/ GeV)

2

cm

  • 38

(10

ν

/ E

CC

σ

CDHS, ZP C35, 443 (1987) GGM-SPS, PL 104B, 235 (1981) GGM-PS, PL 84B (1979) IHEP-ITEP, SJNP 30, 527 (1979) IHEP-JINR, ZP C70, 39 (1996) MINOS, PRD 81, 072002 (2010) NOMAD, PLB 660, 19 (2008) NuTeV, PRD 74, 012008 (2006) SciBooNE, PRD 83, 012005 (2011) SKAT, PL 81B, 255 (1979) T2K (Fe) PRD 90, 052010 (2014) T2K (CH) PRD 90, 052010 (2014) T2K (C), PRD 87, 092003 (2013) ArgoNeuT PRD 89, 112003 (2014) ArgoNeuT, PRL 108, 161802 (2012) ANL, PRD 19, 2521 (1979) BEBC, ZP C2, 187 (1979) BNL, PRD 25, 617 (1982) CCFR (1997 Seligman Thesis)

uB

  • Phys. Rev. Lett. 121, 221801(2018)
slide-5
SLIDE 5

MicroBooNE Research Goals

  • R&D opportunities for current and

future LArTPC experiments

  • Longest operating LArTPC
  • Searches for astroparticle and exotic

physics

  • Galactic supernova bursts neutrinos

Iris Ponce - Columbia University 5 ESA/Hubble & NASA

Begins operation in 2026 Begins operation in 2021

slide-6
SLIDE 6

Supernova Neutrinos

  • Supernova (SN) core-collapse radiates

neutrinos

  • Short timescale ~ 10 seconds
  • Low Energy ~ 10 MeV
  • MicroBooNE could detect ~O(10)

events for SN at 10 kpc.

Iris Ponce - Columbia University 6

Image: CERN Courier, M. Nakahata SN1987A

  • M. Koshiba et al., 1988

Kamiokande

slide-7
SLIDE 7

Supernova Neutrinos

  • Supernova (SN) core-collapse radiates

neutrinos

  • Short timescale ~ 10 seconds
  • Low Energy ~ 10 MeV
  • MicroBooNE could detect ~O(10)

events for SN at 10 kpc.

  • Complication -> MicroBooNE as a

surface detector cannot self-trigger

  • An alternative approach is needed
  • A second data stream which

records data continuously in parallel to the trigger (neutrino beam) stream

Iris Ponce - Columbia University 7

Image: CERN Courier, M. Nakahata SN1987A

  • M. Koshiba et al., 1988

Kamiokande

slide-8
SLIDE 8

Example of the cosmic background in MicroBooNE

Iris Ponce - Columbia University 8

Finding a supernova interaction from all the cosmic activity will be non trivial

slide-9
SLIDE 9

Two Main Challenges for MicroBooNE

  • Is it possible to maintain the data rates needed for a continuous

readout?

  • If we use a lossy compression algorithm, are we sensitive to such

low energy events?

Iris Ponce - Columbia University 9

slide-10
SLIDE 10

Two Main Challenges for MicroBooNE

  • Is it possible to main the data rates needed for a continuous

readout?

  • If we use a lossy compression algorithm, are we sensitive to such

low energy events?

MicroBooNE has shown we can!

Iris Ponce - Columbia University 10

slide-11
SLIDE 11

SN and Trigger Data Scheme

Iris Ponce - Columbia University 11

slide-12
SLIDE 12

MicroBooNE’s Continuous (Supernova) Readout Stream

  • The supernova stream reads out the data

continuously

  • Relies on delayed external trigger –Supernova Early

Warning System (SNEWS)

  • Saving data continuously is non trivial
  • Without compression:

2 Msamples/s * 2 B/sample * 8256 channels = 33 GB/s 3.7GB/s per DAQ server

  • Set the target writing speed of ~50 MB/s for the DAQ

servers

  • Apply lossy compression to reduce

data by a factor of 80

Iris Ponce - Columbia University 12

Save data for 2 days

slide-13
SLIDE 13
  • Implemented in the Front End Module FPGA (developed at Columbia University)
  • Save regions which are above a threshold relative to the channel baseline.

Iris Ponce - Columbia University 13

Saved Pulses - ROIs What we achieved

Compression Output

ADCs ADCs

Zero Suppression Algorithm

slide-14
SLIDE 14
  • Implemented in the Front End Module FPGA (developed at Columbia University)
  • Save regions which are above a threshold relative to the channel baseline.

Iris Ponce - Columbia University 14

Saved Pulses - ROIs What we achieved

Compression Output

ADCs ADCs

Zero Suppression Algorithm

We can zoom in

  • nto one of the

Regions-of-Interest

slide-15
SLIDE 15

Zero Suppression Algorithm

  • Implemented in the Front End Module FPGA (developed at Columbia University)
  • Save regions which are above a threshold relative to the channel baseline.

Iris Ponce - Columbia University 15

Dynamic baseline: calculated by FPGA Or Static baseline: loaded on the FGPA at beginning of the run

slide-16
SLIDE 16

Zero Suppression Algorithm

  • Implemented in the Front End Module FPGA (developed at Columbia University)
  • Save regions which are above a threshold relative to the channel baseline.

Iris Ponce - Columbia University 16

Dynamic baseline: calculated by FPGA Or Static baseline: loaded on the FGPA at beginning of the run Pre and post samples are needed for reconstruction

slide-17
SLIDE 17

Our Current Data Rates after Compression

Iris Ponce - Columbia University 17

Supernova Stream Run III uses static baseline and low thresholds

Supernova TPC

DAQ Servers

(Crate) compression factor

With Run III, the rates are stable ~50 Mb/s

slide-18
SLIDE 18

The Supernova Continuous Readout

  • Supernova Stream finalized commissioning on October 2017
  • To test the performance of the readout we created a software emulation which

can recreate the existing data reduction algorithms implemented in LArSoft

  • We can make compressed data sets from the non-compressed stream (trigger

stream) and simulation (Monte Carlo)

  • We can simulate the response of the electronics

Iris Ponce - Columbia University 18

slide-19
SLIDE 19

Trigger Stream Data Same Trigger Stream Data + Zero Suppression

Iris Ponce - Columbia University 19

M i c h e l e Stopping muon M i c h e l e Stopping muon

Dead Wires

Michel Electron Candidates in Collection Plane

Blue Background = Data at Baseline values White Background = No Data

slide-20
SLIDE 20

Michel Electron Candidates in Collection Plane

Trigger Stream Data Same Trigger Stream Data + Zero Suppression

Iris Ponce - Columbia University 20

M i c h e l e Stopping muon M i c h e l e Stopping muon

Noise passing zero suppression -> Set thresholds to keep as much information as possible

We keep the important features while discarding most of the background signals

Dead Wires White Background = No Data Blue Background = Data at Baseline values

slide-21
SLIDE 21

Testing the performance of the Readout

  • Use Michel electrons from a decaying cosmic muon

to probe the SN readout sensitivity to low energy EM activity.

Iris Ponce - Columbia University 21

slide-22
SLIDE 22

Testing the performance of the Readout

  • Use the reconstruction and

selection algorithms used follow MicroBooNE’s previous publication.

Iris Ponce - Columbia University 22

slide-23
SLIDE 23

Testing the performance of the Readout

  • Use the reconstruction and

selection algorithms used follow MicroBooNE’s previous publication.

Iris Ponce - Columbia University 23

slide-24
SLIDE 24

Reconstructing the Michel Energy Spectrum

  • Continuous Stream data: 53.31

minutes taken on September 21st, 2018

  • Trigger Stream: 58.85 minutes

taken between December 1st, 2017 and July 7th, 2018

  • Data set without offline zero

suppression

  • Data with offline zero suppression

Iris Ponce - Columbia University 24

slide-25
SLIDE 25

The Michel Energy Spectrum

Iris Ponce - Columbia University 25

MicroBooNE In-Progress

  • Both spectra were generated

using the reconstruction detailed before

  • The spectra are relatively

normalized

  • Discrepancies between the

Trigger Stream and SN stream are caused by the zero suppression

slide-26
SLIDE 26

The Michel Energy Spectrum

Iris Ponce - Columbia University 26

MicroBooNE In-Progress MicroBooNE Preliminary

Zero suppressing the trigger stream files offline

slide-27
SLIDE 27

The Michel Energy Spectrum

Iris Ponce - Columbia University 27

MicroBooNE Preliminary

  • Both spectra are relative

normalized

  • The shapes of the spectra match

very well!

slide-28
SLIDE 28

Next Steps

Iris Ponce - Columbia University 28

Start with MicroBooNE’s suppressed data “ROIs” Create Trigger Primitives (will be used in DUNE) and remove hit noise

https://www.nevis.columbia.edu/reu/2019/reports/Hinrichs_report.pdf

slide-29
SLIDE 29

Next Steps

Iris Ponce - Columbia University 29

MicroBooNE’s default reconstruction Resulting TPs

https://www.nevis.columbia.edu/reu/2019/reports/Hinrichs_report.pdf Max Peak TP Tot TP Integral TP Default Reco

slide-30
SLIDE 30

Michel Energy Reconstruction using TPs

  • Both spectra were ran using

MicroBooNE’s Michel reconstruction module.

  • The TPs recreate the tracking

and calorimetry information.

  • The TPs can reconstruct the

michel energy spectrum reasonably.

Iris Ponce - Columbia University 30

Gausshit – default reconstruction TP integral

https://www.nevis.columbia.edu/reu/2019/reports/Hinrichs_report.pdf

slide-31
SLIDE 31

Michel Energy Reconstruction using TPs

  • Both spectra were ran using

MicroBooNE’s Michel reconstruction module.

  • The TPs recreate the tracking

and calorimetry information.

  • The TPs can reconstruct the

michel energy spectrum reasonably.

Iris Ponce - Columbia University 31

Gausshit – default reconstruction TP integral

https://www.nevis.columbia.edu/reu/2019/reports/Hinrichs_report.pdf

This is can be developed further to have online clustering for a trigger for MicroBooNE or SBND

slide-32
SLIDE 32

Conclusions

  • MicroBooNE has commissioned the only LArTPC readout to detect

neutrinos from a supernova core-collapse

  • We have accomplished a stable compression rates ~80
  • We validated the performance of the Continuous Readout with

Michel Electrons

  • Future Goals : Work on evaluating MicroBooNE’s sensitivity to

Supernova neutrinos with simulations Observe a galactic supernova burst!

Iris Ponce - Columbia University 32

slide-33
SLIDE 33

Thank you!

Iris Ponce - Columbia University 33

slide-34
SLIDE 34

Back-Up Slides

Iris Ponce - Columbia University 34

slide-35
SLIDE 35

Calculating Channel Thresholds

  • The goal with is to keep as much data as possible
  • We want to get rid of noise but maintaining the physics intact.
  • Channel thresholds allow us to optimize the compression algorithm

and keep a steady data rate

Iris Ponce - Columbia University 35

The limits of the red region are the thresholds Red shaded regions are the samples which will be zero-suppressed. This corresponds to 98.5% of the integrated distribution symmetric to the baseline value