Supernova Neutrino Studies: Progress and Updates Erin Conley March - - PowerPoint PPT Presentation

Supernova Neutrino Studies: Progress and Updates

Erin Conley March 7, 2018 SNB/LE Working Group Meeting

2

Outline

• Position resolution study

– Introduction – Position difference – Position resolution – Event display study

• MARLEY smearing matrix

– Introduction – Steven Gardiner’s fractional energy plot – The behavior of the fractional energy plot

• Summary

3

Position Resolution Study: Introduction

• Determine position resolution for low-energy neutrino

events

– Difference between truth, reconstructed positions (2D and 3D) – Optical flash provides the timing information – If difference ~ 0, then we have good resolution!

• Motivation

– If resolution is bad/not optimized, then significant information about neutrinos will be lost – Calibration requirements – Rejection of background – Correction for drift in z direction

4

Position Difference

• Reconstruction algorithm used: pmtracktc
• Two methods with consistent results:

– First hit in first track

• Advantage: closer to raw data!

– First TrajectoryPoint object in first track

• Advantage: 3D information!
• Found “x” position using these methods, took difference

from MC truth vertex

– Width of position difference distribution = resolution

5 3/7/18

Position Difference Plots

First TrajectoryPoint First Hit 80.25 MeV neutrinos; 1000 events (3 events had no track info); x position only Sharp central Gaussian + wider Gaussian present at all energies

6 3/7/18

Fitting with two Gaussians

• Had ROOT fit position

difference (in x) distributions on two different ranges

• Central peak: [-4.0, 4.0]
• Wider dist.: [-50.0, 50.0]
• Plotted widths/resolutions

against one another

Example plot: 30.25 MeV, 1000 events Blue: Fit on [-4.0, 4.0] Red: Fit on [-50.0, 50.0]

7 80.25 MeV neutrinos; 1000 events (3 events had no track info); x position only

8 3/7/18

Position Resolution Plots

Central Peak Resolution Wide Peak Resolution 80.25 MeV neutrinos; 1000 events (3 events had no track info); x position only

9

3D “position resolution”

• Use TrajectoryPoints in

first track

• Find the distance

between 3D TrajectoryPoint, MC Truth vertex

• “Position resolution”

determined by finding distance in which 68% of the events are contained (example right)

10

11

Event Display Study

• As neutrino energy increases, two separate distributions

visible in the 3D distance distributions

• Second distribution doesn’t seem to appear for lower

energies, i.e., lower energies just have a sharp peak + tail

– Not a lot of charge deposited → not enough reconstruction information for 2+ tracks – Tracks small at lower energies → front/end of track less important

• Significant issues uncovered in study; need to combat

– Determining the front/end of reconstructed track matters – Sometimes the first track is not the primary electron track

12

Current Work

• Determining which track most likely reconstructed

primary electron

– Likelihood function using track/hit information – Machine learning?

• Determining front/end of track

– My current function to determine front/end agrees with truth ~75% of the time (40%-60% performance for energies lower than 13-14 MeV) – Looked into ResidualRange information; only agreed ~40% of the time with truth

13 3/7/18

MARLEY Smearing Matrix Checks: Introduction

• Summer 2017: Worked to

update default LAr smearing matrix in SNOwGLoBES using MARLEY simulations

• As neutrino energy increases,
• ne might expect number of

nucleons (neutrons) to increase; more energy becomes dedicated to nucleons

• Checked that smearing matrix

agrees with simulation in primary MC truth study

• Average fractional truth energy
• vs. neutrino energy plot

14

Current Work

• Smearing matrix with no neutrons

– Currently simulating 10k νe CC events from 4.25 MeV to 99.75 MeV in 0.5 MeV increments – Use simulations to reproduce plots from primary truth study, compare behavior to previous study

• Steven Gardiner and I discussed MARLEY smearing

matrix checks

– Steven reproduced fractional energy plot – Primary truth study consistent with MARLEY physics – Model not as reliable at 60+ MeV νe energies vs. > 60 MeV

15

Steven G.’s reproduced fractional energy plot

16 3/7/18

Comparing the two fractional energy plots

Steven G’s plot My plot

17

De-excitation gammas from nucleus dominate; not enough energy to excite higher levels High-lying nuclear levels become

• accessible. More

phase space becomes available for nucleon emission. No more nuclear matrix elements available in model; all excess energy given to primary electron Neutron emission becomes possible. These events bring the gamma fraction down as the neutron takes away most

• f the available energy.

18

Summary

• Initiated study to determine the position resolution of

low-energy neutrino interactions

– Looked at both 2D and 3D position resolutions – Secondary distributions in the position difference distributions most likely due to reconstruction issues

• Steven Gardiner reproduced the primary fractional

energy plot with similar behavior to mine

– We believe it agrees with the model currently implemented in MARLEY – Updating the model would help with energies > 60 MeV

Backup Slides

20

Event Display Study: Introduction

• Looked at event displays in three (approximate, eye-

balled) ranges:

– Large peak close to zero – Second smaller distribution (if applicable) – Outliers in the tail

• Did this for 5.25 MeV, 28.75 MeV, 95.25 MeV neutrino

energies

21 28.75 MeV – Zoomed In [0, 50.0] Entire distribution is [0, 400.0]; Two distributions visibly seen (wider second distribution); long tail after the two distributions

22 28.75 MeV – Zoomed In [0, 50.0] First track reconstructs primary electron well Significant portion due to incorrect front/end of reconstructed track Other issues include unideal reconstruction (e.g., only half the track was reconstructed); a secondary track reconstructed the primary electron Significant portion due to secondary track reconstructing primary electron

23

Examples of event displays

= Reconstructed Track, = Truth vertex

Distance: 6.16179 cm; example showing incorrect front/end of track; even more significant at higher energies! Distance: 36.9881 cm; example showing secondary track (Track #1) reconstructing the primary electron