Searching for TeV Gamma-ray Emission from Binary Systems with HAWC - - PowerPoint PPT Presentation

Searching for TeV Gamma-ray Emission from Binary Systems with HAWC

Chang Dong Rho

University of Rochester ICRC 2017 BEXCO, Busan

• γ-ray physics & γ-ray binary systems
• The HAWC Observatory
• Searches for emission from compact TeV

binary systems (with highlighted results)

• Closer look at HESS J0632+057

Overview

2

• γ rays are the most energetic form
• f EM radiation and have no electric

charge.

• There are multiple ways to generate

them:

• 1. π0 è γ + γ (hadronic)
• 2. e- + γ è e- + γ* (leptonic)
• The hadronic process is seen in

very high energy emissions and can tell us about CR.

• Study new physics under extreme

conditions (e.g. GRBs, pulsars, …).

Why High-Energy γ rays?

3

Greg Vance CHANDRA

4

• To observe γ rays at Earth, we can use air

showers since they are absorbed in atmosphere:

• 1. Record Cherenkov light produced by charged

particles in air showers (IACTs).

• 2. Sample charged particles at ground level (HAWC).

Observing γ rays with Air Showers

4

• CR and γ rays are absorbed in the atmosphere,

so we use air showers:

• 1. Record Cherenkov light produced by charged

particles in air showers (IACTs).

• 2. Sample charged particles at ground level (HAWC).

Observing γ rays with Air Showers

nuclear interactions! EM interactions!

High Altitude Water Cherenkov (HAWC) Observatory

5

• Latitude of 19°N, altitude of 4,100m
• Sierra Negra near Puebla, Mexico
• 300 WCDs – effective area of 22,000m2
• 2 sr F.O.V. and >95% duty cycle
• 300 GeV – 100 TeV

High Altitude Water Cherenkov (HAWC) Observatory

5

• Latitude of 19°N, altitude of 4,100m
• Sierra Negra near Puebla, Mexico
• 300 WCDs – effective area of 22,000m2
• 2 sr F.O.V. and >95% duty cycle
• 300 GeV – 100 TeV

• Binaries are unusual since it is rare to have a natural

mechanism that repeatedly accelerates particles.

• There are many confirmed radio and X-ray binaries

but only 5 γ-ray binaries (PSR B1259-63, LS 5039, LS I +61 303, HESS J0632+057, 1FGL J1018.6-5856, HESS J1832-093(?)) have been observed.

• All 5 γ-ray binaries have been observed in TeV as

point-like sources.

• γ-ray binaries consist of compact Galactic objects in
• rbit with massive companion stars.
• Do not fully understand the mechanism of γ-ray

production and have unexplained mismatches in

• bservations at different energy bands.

TeV γ-ray Binary Sources

6

• Binaries are unusual since it is rare to have a natural

mechanism that repeatedly accelerates particles.

• There are many confirmed radio and X-ray binaries

but only 5 γ-ray binaries (PSR B1259-63, LS 5039, LS I +61 303, HESS J0632+057, 1FGL J1018.6-5856, HESS J1832-093(?)) have been observed.

• All 5 γ-ray binaries have been observed in TeV as

point-like sources.

• γ-ray binaries consist of compact Galactic objects in
• rbit with massive companion stars.
• Do not fully understand the mechanism of γ-ray

production and have unexplained mismatches in

• bservations at different energy bands.

TeV γ-ray Binary Sources

6

• I. F. Mirabel

Table of Binary Candidates

7

• We looked at the

following γ-ray binary candidates:

– 3 known γ-ray binaries in HAWC FOV (red) – 28 XRBs with short

• rbital periods
• I calculated TS after

fitting a power law with a fixed idx of -2.7 and Epiv at 7 TeV.

• Then, post-trial

significances are calculated for each of the sources (< 2σ UL; > 2σ LC).

8

95% UL Fluxes for 25 Candidates

PRELIMINARY

9

95% UL vs. Dec with Sensitivity

PRELIMINARY

HESS J0632+057

10

11

• First discovered as a TeV source

by H.E.S.S. in 2007.

• Variability later found in X-rays

(Porb = 321 ± 5 days) then also

• bserved in TeV (Porb = 315 ± 5

days).

• Only γ-ray binary observed by all

three major IACTs (H.E.S.S., VERITAS & MAGIC).

• No HE observation by Fermi/LAT.

HESS J0632+057

11

• First discovered as a TeV source

by H.E.S.S. in 2007.

• Variability later found in X-rays

(Porb = 321 ± 5 days) then also

• bserved in TeV (Porb = 315 ± 5

days).

• Only γ-ray binary observed by all

three major IACTs (H.E.S.S., VERITAS & MAGIC).

• No HE observation by Fermi/LAT.

HESS J0632+057

• A. R.

PRELIMINARY

11

• First discovered as a TeV source

by H.E.S.S. in 2007.

• Variability later found in X-rays

(Porb = 321 ± 5 days) then also

• bserved in TeV (Porb = 315 ± 5

days).

• Only γ-ray binary observed by all

three major IACTs (H.E.S.S., VERITAS & MAGIC).

• No HE observation by Fermi/LAT.

HESS J0632+057

• A. R.

PRELIMINARY

12

17Months – Light Curve of HESS J0632+057 (Porb = 135 days)

PRELIMINARY

Observation of HESS J0632+057

13

PRELIMINARY

Observation of HESS J0632+057

13 Power law: ~ 3 years of data

PRELIMINARY

Observation of HESS J0632+057

13 Power law: ~ 3 years of data Cutoff5TeV : ~ 8 years of data

PRELIMINARY

Summary

• Upper limits for 25 TeV binary candidates < 2

sigma.

• Light curve analysis on 6 candidates > 2 sigma

(no results shown).

• Upper limits for HESS J0632+057 computed

and presented alongside VERITAS results. We expect to see it with ~3 years of data (power law), ~8 years of data (cutoff @ 5 TeV).

• For phase stacking analysis, check C. Brisbois

poster (GA231, board 136).

14

Reference

• 1. “H.E.S.S. observations of LS 5039”, de

Naurois, M. et al., 2007, ApSS, 309, 277-284

• 2. “Long-term TeV Observations of the Gamma-

ray Binary HESS J0632+057 with VERITAS”, Maier, G. et al., 2015, arXiv 1508.05489

• 3. “IRAS observations of SS 433 and W 50.”,

Band, D. L., 1987, PASP, 99, 622, 1269

15

Back up

24

Source Fitting

• To search for γ-ray sources we do spatial +

spectral fits to the map:

• 1. We assume a morphology (shape) for a source

(e.g. Point, Disk).

• 2. We assume a spectrum for a source (e.g. power

law, cutoff power law).

• 3. “Forward fold” a model through detector response.
• 4. Find the free model parameters that maximize

maximum likelihood and calculate the likelihood ratio (TS) and statistical significance.

dN dE = A E Epiv ⎛ ⎝ ⎜ ⎜ ⎞ ⎠ ⎟ ⎟

α

Sig ~ TS

Maps

• To search for γ-ray sources, a data map is generated

and compared with a background map to remove CR that have survived the γ – hadron separation.

• (raw) Data map contains photon counts after the γ –

hadron separation.

• Background map contains CR that passed γ – hadron

cuts.

• One parameter fit for each pixel. The sqrt of the

calculated maximized TS gives significance.

• 25 month HAWC data used for analysis. Only 17

months of data available for daily maps (light curve).

25

Likelihood Calculation

• Events are binned according to

the fHits.

• Logarithm of likelihood is

computed with the binned data:

• θ that maximizes lnL estimates

the optimum set of free

• parameters. TS is used to

compare two hypotheses:

26

lnL θ;Nobs

( ) =

ln f Nobs

( )i θ

( )

( )

i=0 AllBins

∑

TS = 2ln L H1;Nobs

( )

L H0;Nobs

( )

• Take a theoretical spectrum, smear it, and then

compare the result to the data. The best fit gives you the true spectrum.

• Compare observed count to expected count in a bin.
• But, we use fhit bins (energy variable).

– Fundamental problem: data in reconstructed energy space vs. expectations in true energy space

• Hence, fold (convolve) the expected counts distributed

in true energy space using a model spectrum.

• Det res used to compute the expectations in the

reconstructed bins given hypothesis (model spectrum, e.g. power law) about the true energy distribution.

Forward Folding

27

LS 5039

28

29

• Identified in 1997 as a massive X-

ray binary system with OB star.

• Detected in 2005 by H.E.S.S. as

a TeV binary (Porb = 3.90678 ± 0.0015 days).

• Max flux near inferior conjunction.
• GeV observation by Fermi/LAT

(Porb = 3.90532 ± 0.0008 days).

• A. R.

LS 5039

PRELIMINARY

30

17Months – Light Curve of LS 5039

PRELIMINARY

31

17Months – Periodogram of LS 5039

Similar calculations were done for the other 5 candidates but no orbital modulations were observed.

PRELIMINARY

Check C. Brisbois poster (GA231, board 136) for phase stacking analysis.