Supernova Pointing

AJ RoethDAQ Physics Performance WG meeting

06-20-19

�1

Motivation for pointing

• Supernova Early Warning System (SNEWS) to warn astronomers

• Finding progenitor in case of black hole formation

�2

Pointing resolution requirements• Any pointing information is useful to

astronomers, but the more precise, the better

• I found preliminary current pointing resolutions of DUNE for SNe • Pointing resolution = Angle at

which 68% of reconstructed SNe directions are closer to truth

�3

?

�4

0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1Energy (GeV)

0

2000

4000

6000

8000

10000

Even

ts

eν

eν

xν

Events vs neutrino energy

• Wrote ES event generator • Chose neutrino interacted energy and flavor

(GVKM model distribution from SNOwGLoBES)

• Chose electron energy and calculated angle • Simulated 500 isotropic supernovae • 260 elastic scattering events each in same

direction (GVKM model 10kpc SN) • No noise or radiologicals • Elastic scattering events = ~7% of total,

charged current ~93%

Supernova samples of elastic scattering events

Alex Nikrant, Ranjan Laha & Shunsaku Horiuchi, Phys. Rev. D 97, 023019 (2018)

Supernova samples of ES + CC events• Used a pre-existing CC sample (made

using MARLEY generator) • Sampled according to GVKM model SN

energy distribution from SNOwGLoBES • Rotated to match ES neutrino directions • Ignoring detector anisotropy for now

• Added to ES sample to get 500 ES+CC SN samples • Each SN has 260 ES and 3350 CC

events (GVKM model 10 kpc SN) • No noise or radiologicals

�5

0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1Energy (GeV)

0

10

20

30

40

50

60

Even

t rat

es

Event rates vs neutrino energy, nueCC

Directionality

• CC events less directional, but still have directional information

• Can look at how CC/ES ratio changes over signal and its effect on pointing resolution in the future

�6

1− 0.8− 0.6− 0.4− 0.2− 0 0.2 0.4 0.6 0.8 1Cos(Angular Difference)

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

Frac

tion

of e

vent

s

CC

ES

Cosine of angular difference, reco energy 24-26 MeV

1− 0.8− 0.6− 0.4− 0.2− 0 0.2 0.4 0.6 0.8 1Cos(Angular Difference)

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

Frac

tion

of e

vent

s

CCES

Cosine of angular difference, reco energy 24-26 MeV

Addition to current reco: daughter flipping

• There is error between true and reconstructed directions of electron tracks

• Angles between true start electron direction and reconstructed direction (θ) - show ambiguity in reconstructed electron direction - two peaks

�7

1− 0.8− 0.6− 0.4− 0.2− 0 0.2 0.4 0.6 0.8 1Cos(Angular Difference)

0

0.05

0.1

0.15

0.2

0.25

0.3

Frac

tion

of E

vent

s

Cos of error angle, selecting longest track, after PM dQ/dx flippingCos of angular difference

• Can use daughter tracks (from bremsstrahlung gammas compton scattering, ionized electrons, etc.) to reconstruct direction of electron track • Correlate with electron direction

• Calculates average cos of angles between reconstructed track direction and location of daughter tracks, for each end of electron track

• Choose starting point of track to be the side for which avg(cos(θ)) is largest

�8

Addition to current reco: daughter flipping

Reconstructed supernova direction• Wrote likelihood function to reconstruct

supernova direction from all electron directions and energies

�9

• Minimum of negative sum of log likelihood function is reconstructed supernova direction

• Used 250 supernovae to make probability distribution functions for likelihood function, then found pointing resolution of remaining 250

• PDFs only represent ES events for now • Adding CC events would make them more realistic

Supernova pointing resolution• 10kpc SNe with no noise or radiologicals • Next step is studying the effect of noise and radiologicals on

pointing resolution • Future step is studying pointing resolution as function of

number of events • The events happen over ~10s (GVKM model)

�10

ESonly

Noextraflipping,maximumbinisSNdirection

ESonly

Daughterflipping,maximumbinisSNdirection

ESonly

Daughterflipping+likelihoodfunction

ES+CC

Daughterflipping+likelihoodfunction

Pointingresolution 147.2° 30.6° 9.7° 7.5°

Noextraflipping,nolikelihoodfunction

1− 0.8− 0.6− 0.4− 0.2− 0 0.2 0.4 0.6 0.8 1Cos(Angular Difference)

00.05

0.10.150.2

0.250.3

0.350.4

0.45

Frac

tion

of S

Ne

Cos of angle btwn true and reco'd SN direction

ES+CC,daughterflipping,usinglikelihoodfunction

1− 0.8− 0.6− 0.4− 0.2− 0 0.2 0.4 0.6 0.8 1Cos(Angular Difference)

00.10.20.30.40.50.60.70.80.9

Frac

tion

of S

Ne

Cos of angle btwn true and reco'd SN direction

Strange star formation• A time dependent signal worth considering:

strange star formation • Hypothesis that strange quark matter is ground

state of nuclear matter • Possible existence of stars made of strange quark

matter • Drago & Pagliara, Physical Review C, 92, 045801

(2015) - model of conversion of neutron star to strange star

• Initial seed of strange quark matter, star combusts from inside out, ending with hybrid star

• Neutrino signal

�11

image:universetoday.com

Strange star formation• Possible neutrino fluxes for strange

star formation (Drago & Pagliara 2015)

• Occurs over ~20-120 s total, depending on aQ*max

• aQ*max - “the maximum value of the strangeness unbalance at the interface between hadronic matter and quark matter in order for the process of conversion to still proceed because of its exothermicity”

�12

Drago & Pagliara, Physical Review C, 92, 045801 (2015)

Events in argon• Digitized plot, ran through

SNOwGLoBES • No flavor or energy spectrum

info yet - Kate contacted authors • Assumed equal flavor

distribution and T = 5 MeV for now

• Total events given on right (for 10 kpc)

• DAQ buffer length determines how many of these are recorded

• Pointing resolution as a function of number of events to be studied in the future

�13

Time (s)0 20 40 60 80 100 120

Even

ts (/

0.25

s)

0

100

200

300

400

500

aQmax = 0.3

aQmax = 0.5

aQmax = 0.7

Core combustion

Events vs Time

Pointing latency• Pointing information is useful at any time, but

the sooner the better • Time we want pointing information depends

on many factors: when other detectors provide info, when the SN is visible via light, when the astronomers find it • Tens of seconds to hours between neutrino

burst and optically observable signal - want to warn astronomers ahead of this

• Even if it’s been found, pointing information still useful as a confirmation of the origin of the burst and to see DUNE's pointing capabilities

�14

Thank you

�15