FITTING A PHYSICAL MODEL TO FERMI GRB PROMPT EMISSION DATA 1 Bjrn - - PowerPoint PPT Presentation

1

FITTING A PHYSICAL MODEL TO FERMI GRB PROMPT EMISSION DATA 1

Björn Ahlgren 2
 KTH, Fermi, Oskar Klein Centre

1Submitted to MNRAS. Previous

work: Ahlgren et al 2015

2 On behalf of the

Fermi-LAT Collaboration

Gamma-ray bursts

• Liso > 1050 erg/s
• Cosmological distances
• Long and short bursts

pertain to type Ic supernovae and binary compact object mergers, respectively

T90 distribution of bursts from the BATSE 4B catalogue, (BATSE 2001)

Prompt emission

• Origin of GRB prompt

emission remains unknown

• Laboratory for

fundamental physics in extreme conditions

• Increase our

understanding of dying stars

• Using GRBs as

cosmological probes

Artists impression of a long GRB Typical prompt emission origin region Afterglow emission

• rigin region

The fireball model

• Derived using simple assumptions and

thermodynamical scaling laws

• Few parameter model which is easy to expand

upon using more refined emission mechanisms

r0 rph

Γ Schematics of the fireball scenario

GRB data analysis

• Physical models are

complex and non- analytical

• Hard to fit
• Empirical models are

fitted to data and then compared with physical models Empirical function to describe GRB prompt emission

log E log EFE

β α Ep

GRB data analysis

• Spectral shapes are

highly dependent on fitted model

• Hard to reconcile the

empirical functions with physical processes

log E log EFE

Synchrotron Planck function Observed

A way forward

• We create DREAM, a physical model for

subphotospheric dissipation, which can be easily fitted to data

• We fit DREAM to a large sample of Fermi GRBs
• We can successfully describe a third of the fitted

spectra

7

FITTING A PHYSICAL MODEL TO FERMI GRB PROMPT EMISSION DATA

• Subphotospheric dissipation and the DREAM
• Summary of data sample and data analysis
• Results
• Review and conclusions

8

• Scenario based on the fireball model and internal shocks
• Assume most dissipated energy goes to thermal electrons
• Simulated using code from Pe’er et al (2005)

DREAM model scenario

Dissipation with Radiative Emission as A table Model (DREAM)

Photosphere,

rph

Dissipation radius, ε,τ Saturation radius, Γ

L0,52

Compact

• bject

End of dissipation

9

Creating a table model

• Optical depth at dissipation


τ = 1 - 35

• Lorentz factor 


Γ = 100 - 500

• Isotropic fireball luminosity


L0,52 = 0.1 - 300 erg/s

• Fraction of dissipated energy


ε = 0.01 - 0.4

• Main model has negligible

synchrotron radiation

• Table model = lookup table
• f model spectra
• Span a 4 dimensional grid
• f model parameters, using

1350 grid points

• Physically motivated

parameter boundaries

• Xspec interpolates in table

model to perform fits

10

Data sample and analysis

• Fluence > 10-5 erg/s
• Known redshift
• Detected before

2016-06-01 36 long Fermi GRBs

11

Data sample and analysis

• Time resolved Xspec

analysis

• Bayesian blocks

binning with SNR cut

• Avoid spectral evolution

in fitted spectra

• Goodness-of-fit test with

Monte Carlo

−10 10 20 30 40

Time (s)

1000 1500 2000 2500 3000 3500 4000

counts s−1

Light curve and binning for GRB 100414A. Red regions are excluded for low SNR

12

10 100 1000 10 100 1000 104 0.1 1 10

Results

• 36 bursts
• 293 spectral fits
• 96 accepted

fits

• 197 rejected

fits

13

67 % 33 %

Example of a rejected fit Example of an accepted fit

1 10 100 1000 10 100 1000 104 105 106 0.1 1 10

data/model

Ratio E (keV) E F

E

Ratio E (keV) E F

E

1 10 100 1000 10 100 1000 104 105 106 0.1 1 10

data/model

Rejected fits

• Predicted flux too

small

• Due to specific

choice of dissipation scenario

• Possibly sign of

multi-zone emission

• Generally not the

spectral shape which is the issue

Typical rejected fit

E (keV) EFE

14

Ratio

Dissipation radius, ε Saturation radius, Γ

L0,52

Compact

• bject

Accepted fits

15

16

• Strong correlation between

model and observed luminosity

• Clear correlation between

L0,52 - Γ, also found by Lü et al (2012)

• Correlation within some

bursts between Γ and Epeak

• No other correlations to

Band function parameters

Correlations

100 200 300 400 500 Γ 100 101 102

r = 0.75 p= 5.42e-14

1050 1051 1052 1053 1054 1055 1050 1051 1052 1053 1054 1055

r = 0.89 p= 1.57e-21

L0,52 L0,52 Γ Liso,z

Summary and conclusions

• This is the first time a physical model is fitted to a

large sample of GRBs

• The DREAM model shows that, for the prompt

emission of at least some GRBs, photospheric emission is a promising candidate

• Main challenge is inability to reproduce the most

luminous spectra which constrains underlying model assumptions

17