Is the Amati relation due to selection effects?
Lara Nava
In collaboration with
G. Ghirlanda, G.Ghisellini, C. Firmani
Egypt, March 30-April 4, 2009NeutronStars & GammaRayBursts
INAF - Osservatorio Astronomico di Brera, Italy
Universita’ degli Studi dell’Insubria, Como, Italy
No instr. selection effects
Possible instr.
selection effects
Observational planes
Amati Yonetoku
Epeak,obs – Fluence Epeak,obs – Peak Flux
No instr. selection effects
Possible instr.
selection effects
Fluence [erg] Peak Flux [erg/cm2 s]
Ep
eak
,ob
s [k
eV]
Instrumental Selection Effects
TRIGGER THRESHOLD
SPECTRAL ANALYSIS THRESHOLD
Minimum Energy Peak Flux:
Plim (Epeak,obs)
Peak flux
Epea
kAllowed region
Minimum Fluence:
Flim (Epeak,obs)
Epea
k
Fluence
HETE: Sakamoto et al. 2005 Swift/BAT: Butler et al. 2007 (freq)BATSE: Kaneko et al. 2006 Konus/Wind: Golenetskii et al. (GCNs)
Amati relation in the observational plane
All GRBs
with
redshift
All GRBs
with
published
Epeak and Fluence
Build a complete spectral sample of
BATSE bursts down to ~ 10-6 erg/cm2
Extend the Bright BATSE GRB sample (Kaneko et al. 2006)
to lower fluences
Density contours
Nava et al., 2008, MNRAS
Epeak,obs distribution of BATSE
bursts
bright
faint
Lower fluences
correspond to
lower Epeak
Considering the whole sample we
found:
‹Ep› = 160 keV
Yonetoku relation in the observational plane
HETE: Sakamoto et al. 2005 Swift/BAT: Butler et al. 2007 (freq)BATSE: Kaneko et al. 2006 Konus/Wind: Golenetskii et al. (GCNs)
Ep
eak
,ob
s [k
eV]
Bosnjak et al. 2009 (submitted)
Epeak-Liso
Ep-Liso correlation found with time Ep-Liso correlation found with time integrated spectra holds also within a integrated spectra holds also within a
burst!!burst!!
Summary
Correlation in the observer frameDetermined by selection effectsWithout z same region that with z
Short bursts
AMATI YONETOKU
YES
YES
YES
NONO
NO
(larger)
YES
The Yonetoku correlation is more fundamental and requires more attention and theoretical efforts to explain
its physical origin
NOWithin a single burst
See talk tomorrow by Ghirlanda
Epeak-Fluence or Epeak-Peak Flux show strong correlations
Add bursts without redshifts (+ a complete BATSE sample)
In the future Epeak-Eiso correlation will have a different slope & larger scatter but maybe not for the Epeak-Liso correlation
6% of outliers of the Epeak-Eiso
0.3% of outliers of the Epeak-Liso
Strong Epeak-Fluence correlation
Strong Epeak-Peak Flux correlation
Summary
What about short bursts?
Short bursts populate the same region of long bursts in
the Epeak,obs-Peak Flux plane. They can be consistent with
the Epeak-Liso correlation
Short bursts have similar Epeak but lower fluences in respect to Long bursts. They cannot be consistent with the
Epeak-Eiso correlation
Epeak-Eiso correlation: 6% of BATSE bursts
are outliers
Summary:
The existence of a correlation in the plane Epeak,obs-Fluence is not determined by instrumental biases
However, they affect the slope and the dispersion of the Epeak-Eiso
correlation
Swift/BAT bursts are strongly biased by the spectral and trigger threshold and by the very small energy range → flat correlation in the observational plane → flat correlation in the rest frame plane
Epeak-Eiso Epeak-Liso