Title: Correlations between Lag, Duration, Peak Luminosity, Hardness, and Asymmetry in Long GRB Pulses
1Correlations between Lag, Duration, Peak
Luminosity, Hardness, and Asymmetry in Long GRB
Pulses
- Jon Hakkila1 and Renata S. Cumbee2
- 1 Dept. Physics and Astronomy, College of
Charleston - 2 Dept. Physics and Astronomy, Francis Marion
University
- Acknowledgements
- Poster evaluation committee
- Collaborators on various aspects of this project
over the years Jay Norris, Tim Giblin, Chris
Fragile, Jerry Bonnell, students Renata Cumbee
and Mark Wells - Collaborators-to-be Tom Loredo, Rob Preece, and
(hopefully) many others
2There is a Long History of Studying GRB Pulses
- Pulses exhibit a rich phenomenology, including
(1) inherent time asymmetry as characterized by
longer decay than rise rates, (2) hard-to-soft
spectral evolution, and (3) broadening at lower
energies (e.g., Norris et al. 1996, Ramirez-Ruiz
Fenimore 2000, Norris 2002, Ryde 2005). - However, until recently, little work has been
done on fitting pulses at different energies.
3Semi-automated code for identifying and fitting
GRB pulses.
- Bayesian Blocks routine (Scargle 1998)
identifies potential pulses in 64ms, summed
4-channel data (BATSE and Swift).
- Four-parameter pulse model (Norris et al.
2005) plus background Pulses and background
simultaneously fit using MPFIT (Markwardt 1998).
- Dual timescale peak flux (Hakkila et al. 2003)
used to remove statistically significant pulses
process is not biased pulse toward being either
long or short.
- Pulses obtained from summed data used as
initial guesses for pulses in individual energy
channels.
4RESULTS
- Database 307 pulses in 106 Long and 46 Short
GRBs as we move sequentially through the BATSE
catalog.
- The energy-independent pulse-finding process
regularly recovers energy-dependent pulse
properties.
- Every pulse has its own lag lag is a pulse
property rather than a bulk GRB property.
- It is difficult to unambiguously identify and
fit low intensity pulses and/or pulses in crowded
regions.
- Low fluence events (LFEs) are occasionally
present and are difficult to fit. They could be
subpulses or a noise process.
5Pulse-fitting example GRB 950325a (BATSE 3480)
6Pulse-fitting example GRB 930123 (BATSE 2600)
7Results are not always clean GRB 950625 (BATSE
3649)
25 keV - 1 MeV
1
2
3
8Reulsts are not always clean GRB 060418 (Swift)
10 keV - 300 keV
1
2
9What does GRB lag measure (as obtained from the
CCF)?
The CCF lag is dominated by large amplitude,
narrow pulses with short lags. Longer-lag pulses
can smear out this behavior. The GRB peak flux is
not generally the pulse peak flux of the
brightest pulse, due to pulse overlap.
10The durations of pulses from Long (blue diamonds)
and Short (violet squares) BATSE GRBs are plotted
against their positive lags.
Pulse durations vs. lags demonstrate distribution
of lag uncertainties. Short GRB pulse lags are
consistent with a mean near 0 s, while some Long
GRB pulses have extremely negative lags
consistent with hidden pulse contamination.
11Pulse with negative lag BATSE 1807 (2nd pulse
lag -4.2 s)
2nd pulse is apparently composed of two fainter
pulses the second of these is harder than the
first.
12Short GRB pulse peak intensities strongly
correlate with pulse durations Long GRB pulses
only weakly do.
Durations of Long BATSE GRB pulses are plotted
against their asymmetries longer pulses are
typically more asymmetric than shorter ones.
Short GRB pulse asymmetries are difficult to
measure on the 64 ms timescale and are not
plotted.
13Pulse peak fluxes of pulses from Short (violet
squares) and Long (blue diamonds) BATSE GRBs are
plotted against their durations. Although both
pulse types suggest correlated properties, the
relationship is different, suggesting different
mechanisms and/or origins.
14Long GRB Pulse Correlations
- The tables give rank order correlation
probabilities that the pulse characteristics in
question are random, with correlations in black
and anti-correlations in red. A large font
indicates a stronger correlation, and a bold
large font indicates a very strong correlation. - Since lags and durations of Long GRB pulses are
indicators of pulse peak luminosities (Hakkila et
al. 2008), the correlated pulse properties of
spectral hardness and pulse asymmetry are
luminosity indicators as well. - Although there is some overlap in the observed
properties of Short GRB and Long GRB pulses, the
differing luminosities, spectral hardnesses, and
durations suggest that Short GRB pulses are
inherently different than Long GRB pulses, and
that either different mechanisms or different
environments are involved. - The relationships between pulse peak flux, pulse
duration, and (weakly) spectral hardness of Short
GRB pulses could be be partially due to distance
and cosmology (pulses belonging to distant GRBs
will appear to be fainter, longer, and softer
than those belonging to nearer GRBs). However,
Short GRB pulse durations span almost two orders
of magnitude this is too large to be caused by
time dilation alone for Short GRBs typically
found at z 1.
Short GRB Pulse Correlations
15The redshift z of each pulse can be estimated
from its duration (Hakkila et al. 2008) and from
its luminosity distance DL this is similar to
the technique used for burst lags by Kocevski and
Liang (2006). By averaging these redshifts
together over many pulses, an independent
estimate can be made of the mean burst
redshift where d? is the beaming factor,
f256 is the 256 ms peak flux, and the assumed
cosmology is defined by H0 65 km s-1 Mpc-1, ?m
0.3, and ?? 0.7.
16The z-distribution of 106 Long BATSE GRBs, with
external errors of ?z 0.7. The distribution
implies that high redshift bursts are rare in the
BATSE catalog in agreement with Ashcraft
Schaefer (2007). In fact, the GRB at z 8 is
very faint with a short, asymmetric pulse and may
not be a Long GRB.
17Some low-z BATSE bursts
18Some high-z BATSE bursts
19Is BATSE GRB 0809 a Short or a Long burst?
- Consistent with being a Short GRB based on
duration and spectral hardness. - However, overlapping pulses are also consistent
with pulse width, lag, and luminosity relation
suggesting a possible redshift of z 2.1 0.7.
20Conclusions
- The process of identifying and fitting pulses
can be complicated, but the results are
remarkably self-consistent. - Pulse properties are central to explaining GRB
prompt emission, whereas bulk properties turn out
to be constructed by combining and smearing out
pulse characteristics in ways that potentially
lose valuable information. - Pulse properties can be used to estimate GRB
redshifts the results are consistent with
previous redshift distributions, including burst
complexity vs. luminosity findings. - Work still needs to be done on measuring and
accounting for pulse spectral characteristics.
These results will help determine how
well-defined relationships between pulse
properties and luminosity are.