Study of the Gamma-Ray Universe from the GeV to TeV Range Michelle D. Myers 1 University of California, Berkeley August 7, 2015 Nevis REU Summer 2015 Columbia, New York ABSTRACT Over one third of the entries in the third Fermi catalog (3FGL) are unidentified and have no established multi-wavelength counterparts. To help identify these unas- sociated sources, Fermi and VERITAS can be used to characterize sources in a wide high-energy regime (20 MeV < E < 50 TeV). I present a maximum likelihood analysis of two sources (3FGL J1250.2-0233 and 3FGL J2209.8-0450) in the hopes of estab- lishing these to be viable sources for study by VERITAS. The analyses ultimately result in finding possible counterparts in other catalogs. Additionally, I continue an analysis of the BL Lac source B2 1215+30. This source shows correlated variability in the Fermi and VERITAS energy ranges. In studying the variability of the source, a limit on the Doppler factor of its relativistic jet can be derived, thus allowing for a better understanding of the physics of active galaxies. Contents 1 Introduction 2 2 The High-Energy Universe 4 2.1 Emission Mechanisms .................................. 4 2.2 Sources .......................................... 6 2.3 Open Questions ..................................... 6 3 Observations 7 3.1 VERITAS ........................................ 8 3.2 Fermi .......................................... 8 1 [email protected]
20
Embed
Study of the Gamma-Ray Universe from the GeV to TeV … · Study of the Gamma-Ray Universe from the GeV to TeV Range ... from 20 MeV to 300 GeV by the Large Area Telescope ... Space
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Study of the Gamma-Ray Universe from the GeV to TeV Range
Michelle D. Myers1
University of California, Berkeley
August 7, 2015
Nevis REU Summer 2015
Columbia, New York
ABSTRACT
Over one third of the entries in the third Fermi catalog (3FGL) are unidentified
and have no established multi-wavelength counterparts. To help identify these unas-
sociated sources, Fermi and VERITAS can be used to characterize sources in a wide
high-energy regime (20 MeV < E < 50 TeV). I present a maximum likelihood analysis
of two sources (3FGL J1250.2-0233 and 3FGL J2209.8-0450) in the hopes of estab-
lishing these to be viable sources for study by VERITAS. The analyses ultimately
result in finding possible counterparts in other catalogs. Additionally, I continue an
analysis of the BL Lac source B2 1215+30. This source shows correlated variability
in the Fermi and VERITAS energy ranges. In studying the variability of the source,
a limit on the Doppler factor of its relativistic jet can be derived, thus allowing for a
better understanding of the physics of active galaxies.
The Fermi 3FGL catalog (Acero et al. 2015) is the result of four years of observation data
from 20 MeV to 300 GeV by the Large Area Telescope (LAT) aboard the Fermi Gamma-ray
Space Telescope (Thompson et al. 2012). Table 1 shows the population distribution of the types
of sources detected by the LAT. As you can see, unassociated sources comprise 33.29% of the
catalog, followed by BL Lac type blazars. Blazars are a subclass of active galactic nuclei (AGN),
which comprise the next largest source group in 3FGL. Similarly, flat spectrum radio quasars
(FSRQs, the next main shareholders in 3FGL) are a subclass of blazars, implying that the most
common source detected by the LAT could be blazars.
Using the 3FGL catalog in conjunction with the publicly distributed Fermi Science Tools,
the significances of detection, spectral indices, and time variabilities can be determiend to verify
source viability for study by VERITAS (Very Energetic Radiation Imaging Telescope Array
System), a ground-based gamma-ray telescope which can observe sources at higher energies than
the LAT. If there is correlation of a blazar detection with VERITAS, we can characterize time
variabilities associated to help constrain the Doppler factors of their relativistic jets.
I choose two unassociated sources by comparing ther integrated fluxes to that of the Crab
Nebula (generally the strongest persistent gamma-ray source in the sky) and selecting sources
of appropriate galactic latitude. A maximum likelihood analysis allows me to model the sources
to create test statistic (TS) maps to determine the significance of detection and light curves to
determine variability. The analysis ultimately yields unexpected results, with one source having
nearly no detection and the other having strong detection with seemingly featureless variability.
However, archival catalog searches shed more information on the region that warrant further
analyses.
Additionally, I continue a study of B2 1215+30, a BL Lac that was detected in the very-high-
– 3 –
energy regime (E > 100 GeV) by MAGIC in 2011 (Aleksic et al. 2012). VERITAS detected and
performed an analysis using observational data from 2008 to 2012 (Aliu et al. 2013) and detected
variability as long as months. Another analysis was performed on VERITAS observation data
in 2014 in conjunction with LAT data to correlate time variability in both regimes (Zefi et al.
2015), resulting in a limit on the Doppler factor of the relatvistic jet from B2 1215+30.
Source Type Number of Entries Percent of 3FGL
Non-Blazar Active Galaxy 3 0.10%
Active Galaxy of Uncertain Type 573 18.89%
Binary 1 0.03%
BL Lac Blazar 660 21.75%
Compact Steep Spectrum Quasar 1 0.03%
Flat Spectrum Radio-Loud Quasar 484 15.95%
Normal Galaxy 3 0.10%
Globular Cluster 15 0.49%
High-Mass Binary 3 0.10%
Narrow-Line Seyfert 1 5 0.16%
Nova 1 0.03%
Pulsara 143 4.71%
Pulsarb 24 0.79%
Pulsar Wind Nebula 12 0.40%
Radio Galaxy 15 0.49%
Starburst Galaxy 4 0.13%
Seyfert Galaxy 1 0.03%
Star-Forming Region 1 0.03%
Supernova Remnant 23 0.76%
Special Casec 49 1.62%
Soft Spectrum Radio Quasar 3 0.10%
Unassociated 1010 33.29%
Table 1: The population of sources in 3FGL is dominantly unassociated, followed by BL Lac blazars.aIdentified by pulsations ; bNo pulsations seen by LAT ; cPotential association with supernova remnant or pulsar
wind nebula.
This paper will present a brief view of the gamma-ray sky and the types of mechanisms
and sources that occupy it. Open questions that can be explored by the study of gamma-rays
are suggested. Section 3 provides a discussion on instruments relevant to the project follows
to illuminate the utility of ground-based and space-based telescopes in complementary energy
regimes. The methods and parameters used in the analysis are outlined in Section 4, while
Section 5 reveals the results gained from the analysis.
– 4 –
2. The High-Energy Universe
The gamma-energy regime starts in the MeV range and extends to the TeV range, making
it difficult for any single telescope to observe this entire energy range. We typically associate
extreme galactic structures (supernovae, pulsars, quasars, black holes, etc...) with gamma-ray
sources. These exotic environments are capable of conditions that are effectively beyond what
we can recreate on Earth (e.g. extreme Lorentz factors). However, detection of gamma-rays in
the sky is non-trivial due to the attenuation of gamma-rays in space. For example, secondary
processes along the beam path enable us to detect gamma-rays from very distant blazars (Essey
et al. 2011). This section is concerned with the emission processes associated with gamma-ray
sources and the questions we have yet answered regarding them. Sections 3.1 and 3.2 illuminate
the instrumental methods of gamma-ray detection.
2.1. Emission Mechanisms
Cosmic-ray (beams of high-energy charged massive particles) interactions with background
photons that yield gamma-rays can be categorized into hadronic and leptonic processes (Albert
at al. (2008) and Pfrommer (2013)). In the hadronic processes the cosmic ray is made up of
protons or ions. Equation 1, as an example of a hadronic process, shows pion decay: a proton
interacts with another proton, which yields a pion. A neutral pion can then decay into gamma-
ray photons. In the leptonic processes, the cosmic ray is made up of electron or positrons.
Equation 2 depicts synchrotron emission (a leptonic process): an energetic electron or a positron
interacts with a magnetic field, yielding an energetic photon. In either process, the particles are
accelerated to relativistic regimes, with the acceleration mechanisms yet to be determined.
p+ p→
{π0 → γγ
π± → e± + νµ + νµ + νe(1)
e±energetic +B → e± +B + γenergetic (2)
AGN provide different emission mechanisms that can be categorized into thermal and non-
thermal processes (Pfrommer 2013). Thermal processes are associated with disk-dominated AGN
in which the infalling matter from the disk gets Comptonized (i.e. an energetic photon hits a
relatively stationary electron, transferring energy to the electron), thereby allowing the Comp-
tonized electrons to emit thermally in the optical and/or x-ray regime. Models typically require
a beaming effect to detect the gamma-ray emission if it were thermally produced. Currently,
thermal processes have not been unambiguously associated with gamma-ray sources.
An AGN which is dominated by jet energy naturally provides non-thermal processes. Elec-
trons that have been accelerated in the jet of an AGN interact with the jet’s magnetic field,
thereby generating synchrotron radiation in the x-ray to radio regime. These electrons are also
capable of inverse-Compton scattering (i.e. a relativistic electron hits a lower energy photon,
– 5 –
transferring energy to the photon) photons that are either generated by the synchrotron radia-
tion (synchrotron self-Compton) or by some external photon source such as ultraviolet radiation
from the disk. The two emission processes resultant from these accelerated electrons result in
a spectral energy distribution (SED) of the AGN with two distinct peaks (see Figure 1). The
synchrotron radiation photons emitted by the electrons are also capable of interacting with a
proton to create a pion, thus lending some ambiguity into whether or not the second emission
peak is a consequence of hadronic (protons) or leptonic (electrons) processes.
Fig. 1.— The figure on the left is an artistic render of an AGN. Jet orientation relative to our line of sight help
classify AGN into radio galaxies, quasars, and blazars. The figure on the right depicts blazar SEDs studied by
Donato et al. (2001) in attempt to unify blazar emission behaviors into one parameter: bolometric luminosity.
The lower-energy peak corresponds to synchrotron radiation emitted by electrons accelerated by the AGN jet.
The higher-energy peak corresponds to inverse-Compton radiation from electrons interacting with photons or
through pion decay. Blazars can be classified depending on their behavior in the optical regime. Also note how
higher-energy emission peaks correspond to lower luminosities.
The jet which provides for the non-thermal jet allows for the gamma-rays to exhibit a larger
luminosity than we would otherwise observe through aberration, time dilation, and red- (or blue-
) shifts. A measure of these effects, which essentially speaks to the strength of the jet, is the
Doppler factor (Dondi et al. 1995), which is given by
δ = [Γ(1− βcosθ)]−1 (3)
where Γ represents the Lorentz factor (Γ = 1√1−β2
) and β = vc. For an example of how the
Doppler factor can be constrained, see Section 5.3.
– 6 –
2.2. Sources
As mentioned above, AGN with jets are able to provide non-thermal processes for the
detection of gamma-rays. At the center of gamma-ray source galaxies are compact objects which
form twin ultra-relativistic jets from the accretion of matter in the disk (Urry et al. 1995). We
further subdivide radio-loud AGN into radio galaxies, quasars, and blazars depending on the jet
orientation relative to our line of sight (see Figure 1). Radio galaxies have jets perpendicular
to our line of sight, quasars have jets angled to our line of sight, and blazars have jets aligned
nearly directly to our line of sight.
Blazars, the most commonly detected gamma-ray sources in the sky, can be further classified
according to their multi-wavelength behavior. In Figure 1, we can distinguish between two main
subclasses depending on the behavior in the optical region. FSRQs have broad optical emission
lines with synchrotron emission peaking in the infrared regime, subsequently limiting the inverse-
Compton emission to the soft gamma-ray regime. Their lower energy emission also typically
allows for the greater observed luminosities due to energy-dependent gamma-ray attenuation in
space (higher energy gamma-rays are more attenuated). The second subclass of blazars is BL Lac
objects, which can have synchrotron peaks in the far-infrared, optical, or ultraviolet bands (and
sometimes even in the x-ray and gamma-ray bands), subsequently allowing for higher energy
inverse compton emission peaks.
Gamma-ray production is not limited to radio-loud AGN. For example, starburst galaxies
(galaxies with significant amounts of star-forming regions and, in effect, supernovae) with gamma-
ray components were first detected by VERITAS (Galante 2011). They have high densities
of cosmic rays that interact with interstellar gas and radiation, thus providing a non-thermal
mechanism by which gamma-rays are emitted (Acciari et al. 2009). Table 1 can help illuminate
the different types of sources that are capable of gamma-ray emissions. At the heart of all of the
processes lie non-thermal mechanisms to combat the attenuation of gamma-rays in space.
2.3. Open Questions
Aside from constraining the various emission mechanisms of gamma-ray sources (leptonic
vs. hadronic processes, for example), there are many open questions to be answered regarding
high-energy astrophysics that can be understood by studying gamma-ray sources. Constraining
beam properties of blazars, for example, allows us to study cosmological structures. Extra-
galactic background light (EBL) arise from photons that have been emitted by galaxies and other
galactic structures over the course of history. It is akin to the cosmic microwave background
(CMB) radiation with a higher registered energy. Unlike the CMB, the EBL cannot yet be
directly detected due to the abundance of foreground and galactic emission. To circumvent this
obstacle, absorption features of gamma-ray spectra can help reveal EBL properties. Attenuation
of gamma-ray flux in space is, in part, associated with gamma-ray photons pair producing with
EBL photons in the optical to ultraviolet bands, and we expect to see absorption features in
– 7 –
these bands of blazar SEDs. If the SED of an object is known, EBL density along the line of
sight can subsequently be inferred (Schroedter 2005).
Further, ultra-relativistic electron-positron pairs resulting from TeV photons annihilating
on the EBL are thought to lose energy by inverse-Compton scattering off of the CMB, cascading
the TeV emission down to GeV energies (Pfrommer 2013). GeV emission, however, is not ob-
served. Two mechanisms have been posited to be responsible for the absence of GeV emission:
intergalactic magnetic fields (IMFs) and blazar heating ((Arlen 2013) and (Chang et. al 2012),
respectively). IMFs are thought to deflect the electron-positron pairs out of the line of sight,
allowing limits to be placed on IMF strengths. Blazar heating, a more effective mechanism for
”hiding” the GeV emission, involves the propagation of the ultra-relativistic electron-positron
pairs through the intergalactic medium. They are subsequently subjected to plasma instabilities
which ultimately result in heat being transferred to the intergalactic medium.
Blazar time variability also exhibit complex multi-wavelength behavior (Pfrommer 2013)
that cannot be modeled simply. Variability timescales range from the order of minutes (Aharo-
nian et al. 2008) to years (Ruan et al. 2012), with small time-scales correlating to super compact
emission regions and large Lorentz factors (Begelman et al. 2008).
Despite the ambiguity of the emission mechanisms that result in the differing timescales,
quantum gravity theories can also be tested using the variabilities observed. Lorentz invariance
violations (LIVs), for example, should be observed at Planck-scale energies. Looking at the
propagation of gamma-rays at the Crab pulsar at different energies allows constraints to be
placed on LIVs (Zitzer 2013).
The energy regime of gamma-rays also correspond to the energy regime of electroweak force
interactions, which coincides with dark matter particle masses of weakly interacting massive
particles (WIMPs) (Steigman et al. 2012). The relic density of WIMPs is the amount of WIMPs
we currently observe given that they were created thermally. As the universe cooled, WIMP
creation would cease, allowing the population density to decrease as they annihilated with other
WIMPs. A cross-section can be calculated to account for the current WIMP density (i.e. the
probability for WIMP annihilation is very low), which subsequently constrains WIMP masses.
Acciari et al. (2010) outline a method by which observations of gamma-rays from dark-matter
dominated galaxies can constrain cross sections and relative velocities of WIMPs.
3. Observations
VERITAS can observe objects in with energies 50 GeV < E < 50 TeV, while Fermi can
observe at 20 MeV < E < 300 GeV. Together, they form a complementary and broad energy
range for study of gamma-ray sources. The different energy regimes are realized by two different
detection methods: indirect and direct.
– 8 –
Fig. 2.— The figures above depict the Fermi Gamma-Ray Space Telescope on the left and the
four Cherenkov telescopes associated with VERITAS on the right.
3.1. VERITAS
Gamma-rays cannot penetrate the Earth’s atmosphere, but we can detect Cherenkov radi-
ation from incident gamma-rays in the atmosphere (Galbraith et al. 1953). Incident gamma-
rays (and cosmic-rays) trigger particle cascades of relativistic charged particles which create
Cherenkov radiation in the direction of propagation. The shower creates a light pool on the
ground that peaks roughly in the optical range (∼400 nm wavelength which corresponds to blue
light) for high-energy incident events. VERITAS subsequently uses mirrors to reflect the light
onto a focal plane camera (Holder et al. 2006) to image the showers (see Figure 2).
Located at the Fred Lawrence Whipple Observatory in southern Arizona, VERITAS is
comprised of four 12 meter optical reflectors, achieving maximum sensitivity for incident rays
with energies 85 GeV < E < 10 TeV. Both cosmic-rays and gamma-rays can induce particle
showers in Earth’s atmosphere. Cosmic-ray events can be distinguished from gamma-ray events
depending on their image orientation and morphology; cosmic-rays are generally wider and less
regular because the showers generally have several components. When the particle showers are
observed by Cherenkov telescopes, they appear as elongated ellipses from which the original
direction of propagation can be derived (Hillas 1985). By analyzing the Cherenkov yield from
the particle shower of a gamma-ray event, VERITAS is also able to discern the original energy
of the incident gamma-ray.
3.2. Fermi
The Fermi LAT aboard the Fermi Gamma-Ray Space Telescope, which was launched in
June 2008, has a detection range of 20 MeV < E < 300 GeV. Direct detection of gamma-rays
must occur in outer space, which limits the collection area and, therefore, the flux that can be
received by space telescopes. Using adapted accelerator experiment techniques (Thompson et al.
2012), the LAT uses a layer of tungsten to convert incoming gamma-rays into electron-positron
pairs. The pairs then interact with silicon-strip charged-particle trackers to create images of the
pair trajectory, after which a calorimeter measures the energies of the particles. It has an angular
– 9 –
resolution finer than 1◦ for a single gamma-ray, has a large field of view (20◦), and can observe
the entire sky every three hours.
Cosmic-rays also pose a problem to the LAT; there is 105 more cosmic-ray flux than gamma-
ray flux. An Anticoincidence Detecter rejects 99.97% of cosmic-ray signals incident on the LAT
by distinguishing between charged particle events and neutral (gamma-ray) events. Further, the
LAT also filters against gamma-rays that originate in Earth’s atmosphere. Similar to VERI-
TAS, the Fermi LAT determines the incident energies and directions of propagation of incoming
gamma-rays.
4. Fermi Analysis
An unbinned maximum likelihood analysis was performed on specifically chosen sources
from the 3FGL catalog using survey data by the Fermi LAT. The analysis used the Fermi
Science Tools software package, version v10r0p51, in the Python environment. The procedure
is initiated by running gtselect to make cuts on energy and time of the data sets downloaded
from the Fermi database. These cuts take into account the source region (the distance out to
which sources from the 3FGL catalog were included in the model) and the region of interest
(the distance out to which photon counts from sources form the 3FGL catalog were allowed to
have free parameters). Due to the energy-depencies of the point spread function of the LAT, for
analyses with energies E ∼ 1 GeV, it is recommended to use a 15◦ source region with a 5◦ region
of interest. For lower energies (E ∼ 100 MeV), the source region is recommended to be 20◦ with
a 10◦ region of interest2.
Using gtmktime subsequently filters the data set further to create good time intervals (GTIs),
i.e. time periods wherein the data are valid, using spacecraft data from the database. gtbin
then allows for the creation of a counts map, one of the first indicators of a viable data set. The
counts maps is a two-dimensional spatial map that shows photon counts that reflect the various
filteres already placed on the data set. Lack of counts in the area where a detection is expected
may indicate the need for less stringent cuts on time and energy.
The most time-consuming part of the analysis comes from creating exposure maps (gtexpmap)
and livetime cubes (gtltcube) necessary for computing the predicted numver of photons in the
region of interest. In conjunction with those processes make3FGLxml.py3 was used to generate
an XML file to create a model file based on 3FGL catalog sources, using gll_iem_v06 and
iso_P8R2_SOURCE_V6_v064 as the galactic diffuse and isotropic model files, respectively.