The Astrophysical Journal, 753:83 (22pp), 2012 July 1 doi:10.1088/0004-637X/753/1/83
C
2012. The American Astronomical Society. All rights reserved. Printed in the U.S.A.
A STATISTICAL APPROACH TO RECOGNIZING SOURCE CLASSES FOR UNASSOCIATED
SOURCES IN THE FIRST FERMI-LAT CATALOG
M. Ackermann
1
, M. Ajello
2
, A. Allafort
2
, E. Antolini
3,4
, L. Baldini
5
, J. Ballet
6
, G. Barbiellini
7,8
, D. Bastieri
9,10
,
R. Bellazzini
5
, B. Berenji
2
, R. D. Blandford
2
, E. D. Bloom
2
, E. Bonamente
3,4
, A. W. Borgland
2
, A. Bouvier
11
,
T. J. Brandt
12,13
, J. Bregeon
5
, M. Brigida
14,15
, P. Bruel
16
, R. Buehler
2
, T. H. Burnett
17
, S. Buson
9,10
, G. A. Caliandro
18
,
R. A. Cameron
2
, P. A. Caraveo
19
, J. M. Casandjian
6
, E. Cavazzuti
20
, C. Cecchi
3,4
,
¨
O. ¸Celik
21,22,23
, E. Charles
2
,
A. Chekhtman
24,68
,A.W.Chen
19
, C. C. Cheung
25,68
, J. Chiang
2
,S.Ciprini
4,26
, R. Claus
2
, J. Cohen-Tanugi
27
,
J. Conrad
28,29,69
,S.Cutini
20
, A. de Angelis
30
, M. E. DeCesar
21,31
, A. De Luca
32
, F. de Palma
14,15
, C. D. Dermer
33
,
E. do Couto e Silva
2
, P. S. Drell
2
, A. Drlica-Wagner
2
, R. Dubois
2
, T. Enoto
2
, C. Favuzzi
14,15
, S. J. Fegan
16
,
E. C. Ferrara
21
, W. B. Focke
2
, P. Fortin
16
, Y. Fukazawa
34
, S. Funk
2
,P.Fusco
14,15
, F. Gargano
15
, D. Gasparrini
20
,
N. Gehrels
21
, S. Germani
3,4
, N. Giglietto
14,15
, F. Giordano
14,15
, M. Giroletti
35
, T. Glanzman
2
, G. Godfrey
2
,
I. A. Grenier
6
, M.-H. Grondin
36,37
, J. E. Grove
33
, L. Guillemot
38
, S. Guiriec
39
, M. Gustafsson
9
, D. Hadasch
18
,
Y. Hanabata
34
, A. K. Harding
21
, M. Hayashida
2,40
, E. Hays
21
, S. E. Healey
2
,A.B.Hill
41
, D. Horan
16
,X.Hou
42
,
G. J
´
ohannesson
43
,A.S.Johnson
2
,T.J.Johnson
25,68
, T. Kamae
2
, H. Katagiri
44
, J. Kataoka
45
, M. Kerr
2
,
J. Kn
¨
odlseder
12,13
,M.Kuss
5
, J. Lande
2
, L. Latronico
46
,S.-H.Lee
47
, M. Lemoine-Goumard
48,70
, F. Longo
7,8
,
F. Loparco
14,15
, B. Lott
48
, M. N. Lovellette
33
, P. Lubrano
3,4
,G.M.Madejski
2
, M. N. Mazziotta
15
,J.E.McEnery
21,31
,
J. Mehault
27
, P. F. Michelson
2
,R.P.Mignani
49
, W. Mitthumsiri
2
, T. Mizuno
34
, C. Monte
14,15
, M. E. Monzani
2
,
A. Morselli
50
, I. V. Moskalenko
2
, S. Murgia
2
, T. Nakamori
45
, M. Naumann-Godo
6
, P. L. Nolan
2,71
, J. P. Norris
51
,
E. Nuss
27
, T. Ohsugi
52
, A. Okumura
2,53
, N. Omodei
2
, E. Orlando
2,54
,J.F.Ormes
55
, M. Ozaki
53
, D. Paneque
2,56
,
J. H. Panetta
2
, D. Parent
57,68
, V. Pelassa
39
, M. Pesce-Rollins
5
, M. Pierbattista
6
,F.Piron
27
, G. Pivato
10
, T. A. Porter
2
,
S. Rain
`
o
14,15
, R. Rando
9,10
,P.S.Ray
33
, M. Razzano
5,11
,A.Reimer
2,58
,O.Reimer
2,58
, T. Reposeur
48
,R.W.Romani
2
,
H. F.-W. Sadrozinski
11
, D. Salvetti
19
, P. M. Saz Parkinson
11
, T. L. Schalk
11
, C. Sgr
`
o
5
,M.S.Shaw
2
,E.J.Siskind
59
,
P. D. Smith
60
, G. Spandre
5
, P. Spinelli
14,15
,D.J.Suson
61
, H. Takahashi
52
, T. Tanaka
2
, J. G. Thayer
2
, J. B. Thayer
2
,
D. J. Thompson
21
, L. Tibaldo
9,10
, O. Tibolla
62
, D. F. Torres
18,63
, G. Tosti
3,4
, A. Tramacere
2,64,65
, E. Troja
21,72
,
T. L. Usher
2
, J. Vandenbroucke
2
, V. Vasileiou
27
, G. Vianello
2,64
, N. Vilchez
12,13
, V. Vitale
50,66
,A.P.Waite
2
,
E. Wallace
17
,P.Wang
2
,B.L.Winer
60
, M. T. Wolff
33
,D.L.Wood
67,68
,K.S.Wood
33
,Z.Yang
28,29
, and S. Zimmer
28,29
1
Deutsches Elektronen Synchrotron DESY, D-15738 Zeuthen, Germany
2
W. W. Hansen Experimental Physics Laboratory, Kavli Institute for Particle Astrophysics and Cosmology, Department of Physics
and SLAC National Accelerator Laboratory, Stanford University, Stanford, CA 94305, USA; monzani@slac.stanford.edu
3
Istituto Nazionale di Fisica Nucleare, Sezione di Perugia, I-06123 Perugia, Italy
4
Dipartimento di Fisica, Universit
`
a degli Studi di Perugia, I-06123 Perugia, Italy
5
Istituto Nazionale di Fisica Nucleare, Sezione di Pisa, I-56127 Pisa, Italy
6
Laboratoire AIM, CEA-IRFU/CNRS/Universit
´
e Paris Diderot, Service d’Astrophysique, CEA Saclay, 91191 Gif sur Yvette, France
7
Istituto Nazionale di Fisica Nucleare, Sezione di Trieste, I-34127 Trieste, Italy
8
Dipartimento di Fisica, Universit
`
a di Trieste, I-34127 Trieste, Italy
9
Istituto Nazionale di Fisica Nucleare, Sezione di Padova, I-35131 Padova, Italy
10
Dipartimento di Fisica “G. Galilei,” Universit
`
a di Padova, I-35131 Padova, Italy
11
Santa Cruz Institute for Particle Physics, Department of Physics and Department of Astronomy and Astrophysics,
University of California at Santa Cruz, Santa Cruz, CA 95064, USA
12
CNRS, IRAP, F-31028 Toulouse Cedex 4, France; vilchez@cesr.fr
13
GAHEC, Universit
´
e de Toulouse, UPS-OMP, IRAP, Toulouse, France
14
Dipartimento di Fisica “M. Merlin” dell’Universit
`
a e del Politecnico di Bari, I-70126 Bari, Italy
15
Istituto Nazionale di Fisica Nucleare, Sezione di Bari, 70126 Bari, Italy
16
Laboratoire Leprince-Ringuet,
´
Ecole polytechnique, CNRS/IN2P3, Palaiseau, France
17
Department of Physics, University of Washington, Seattle, WA 98195-1560, USA
18
Institut de Ci
`
encies de l’Espai (IEEE-CSIC), Campus UAB, 08193 Barcelona, Spain
19
INAF-Istituto di Astrofisica Spaziale e Fisica Cosmica, I-20133 Milano, Italy; salvetti@lambrate.inaf.it
20
Agenzia Spaziale Italiana (ASI) Science Data Center, I-00044 Frascati (Roma), Italy
21
NASA Goddard Space Flight Center, Greenbelt, MD 20771, USA; elizabeth.c.ferrara@nasa.gov
22
Center for Research and Exploration in Space Science and Technology (CRESST) and NASA Goddard Space Flight Center, Greenbelt, MD 20771, USA
23
Department of Physics and Center for Space Sciences and Technology, University of Maryland Baltimore County, Baltimore, MD 21250, USA
24
Artep Inc., 2922 Excelsior Springs Court, Ellicott City, MD 21042, USA
25
National Academy of Sciences, Washington, DC 20001, USA
26
ASI Science Data Center, I-00044 Frascati (Roma), Italy
27
Laboratoire Univers et Particules de Montpellier, Universit
´
e Montpellier 2, CNRS/IN2P3, Montpellier, France
28
Department of Physics, Stockholm University, AlbaNova, SE-106 91 Stockholm, Sweden
29
The Oskar Klein Centre for Cosmoparticle Physics, AlbaNova, SE-106 91 Stockholm, Sweden
30
Dipartimento di Fisica, Universit
`
a di Udine and Istituto Nazionale di Fisica Nucleare, Sezione di Trieste, Gruppo Collegato di Udine, I-33100 Udine, Italy
31
Department of Physics and Department of Astronomy, University of Maryland, College Park, MD 20742, USA
32
Istituto Universitario di Studi Superiori (IUSS), I-27100 Pavia, Italy
33
Space Science Division, Naval Research Laboratory, Washington, DC 20375-5352, USA
34
Department of Physical Sciences, Hiroshima University, Higashi-Hiroshima, Hiroshima 739-8526, Japan
35
INAF Istituto di Radioastronomia, 40129 Bologna, Italy
36
Max-Planck-Institut f
¨
ur Kernphysik, D-69029 Heidelberg, Germany
37
Landessternwarte, Universit
¨
at Heidelberg, K
¨
onigstuhl, D 69117 Heidelberg, Germany
1
The Astrophysical Journal, 753:83 (22pp), 2012 July 1 Ackermann et al.
38
Max-Planck-Institut f
¨
ur Radioastronomie, Auf dem H
¨
ugel 69, 53121 Bonn, Germany
39
Center for Space Plasma and Aeronomic Research (CSPAR), University of Alabama in Huntsville, Huntsville, AL 35899, USA
40
Department of Astronomy, Graduate School of Science, Kyoto University, Sakyo-ku, Kyoto 606-8502, Japan
41
School of Physics and Astronomy, University of Southampton, Highfield, Southampton SO17 1BJ, UK
42
Centre d’
´
Etudes Nucl
´
eaires de Bordeaux Gradignan, IN2P3/CNRS, Universit
´
e Bordeaux 1, BP120, F-33175 Gradignan Cedex, France
43
Science Institute, University of Iceland, IS-107 Reykjavik, Iceland
44
College of Science, Ibaraki University, 2-1-1, Bunkyo, Mito 310-8512, Japan
45
Research Institute for Science and Engineering, Waseda University, 3-4-1, Okubo, Shinjuku, Tokyo 169-8555, Japan
46
Istituto Nazionale di Fisica Nucleare, Sezioine di Torino, I-10125 Torino, Italy
47
Yukawa Institute for Theoretical Physics, Kyoto University, Kitashirakawa Oiwake-cho, Sakyo-ku, Kyoto 606-8502, Japan
48
CNRS/IN2p3, Centre d’
´
Etudes Nucl
´
eaires de Bordeaux Gradignan, Universit
´
e Bordeaux 1, 33175 Gradignan, France
49
Mullard Space Science Laboratory, University College London, Holmbury St. Mary, Dorking, Surrey RH5 6NT, UK
50
Istituto Nazionale di Fisica Nucleare, Sezione di Roma “Tor Vergata,” I-00133 Roma, Italy
51
Department of Physics, Boise State University, Boise, ID 83725, USA
52
Hiroshima Astrophysical Science Center, Hiroshima University, Higashi-Hiroshima, Hiroshima 739-8526, Japan
53
Institute of Space and Astronautical Science, JAXA, 3-1-1 Yoshinodai, Chuo-ku, Sagamihara, Kanagawa 252-5210, Japan
54
Max-Planck Institut f
¨
ur extraterrestrische Physik, 85748 Garching, Germany
55
Department of Physics and Astronomy, University of Denver, Denver, CO 80208, USA
56
Max-Planck-Institut f
¨
ur Physik, D-80805 M
¨
unchen, Germany
57
Center for Earth Observing and Space Research, College of Science, George Mason University, Fairfax, VA 22030, USA
58
Institut f
¨
ur Astro-und Teilchenphysik and Institut f
¨
ur Theoretische Physik, Leopold-Franzens-Universit
¨
at Innsbruck, A-6020 Innsbruck, Austria
59
NYCB Real-Time Computing Inc., Lattingtown, NY 11560-1025, USA
60
Department of Physics, Center for Cosmology and Astro-Particle Physics, The Ohio State University, Columbus, OH 43210, USA
61
Department of Chemistry and Physics, Purdue University Calumet, Hammond, IN 46323-2094, USA
62
Institut f
¨
ur Theoretische Physik and Astrophysik, Universit
¨
at W
¨
urzburg, D-97074 W
¨
urzburg, Germany
63
Instituci
´
o Catalana de Recerca i Estudis Avan¸cats (ICREA), Barcelona, Spain
64
Consorzio Interuniversitario per la Fisica Spaziale (CIFS), I-10133 Torino, Italy
65
INTEGRAL Science Data Centre, CH-1290 Versoix, Switzerland
66
Dipartimento di Fisica, Universit
`
a di Roma “Tor Vergata,” I-00133 Roma, Italy
67
Praxis Inc., Alexandria, VA 22303, USA
Received 2011 August 4; accepted 2012 March 28; published 2012 June 15
ABSTRACT
The Fermi Large Area Telescope (LAT) First Source Catalog (1FGL) provided spatial, spectral, and temporal
properties for a large number of γ -ray sources using a uniform analysis method. After correlating with the most-
complete catalogs of source types known to emit γ rays, 630 of these sources are “unassociated” (i.e., have
no obvious counterparts at other wavelengths). Here, we employ two statistical analyses of the primary γ -ray
characteristics for these unassociated sources in an effort to correlate their γ -ray properties with the active galactic
nucleus (AGN) and pulsar populations in 1FGL. Based on the correlation results, we classify 221 AGN-like and
134 pulsar-like sources in the 1FGL unassociated sources. The results of these source “classifications” appear to
match the expected source distributions, especially at high Galactic latitudes. While useful for planning future
multiwavelength follow-up observations, these analyses use limited inputs, and their predictions should not be
considered equivalent to “probable source classes” for these sources. We discuss multiwavelength results and
catalog cross-correlations to date, and provide new source associations for 229 Fermi-LAT sources that had no
association listed in the 1FGL catalog. By validating the source classifications against these new associations, we
find that the new association matches the predicted source class in ∼80% of the sources.
Key words: catalogs – galaxies: active – gamma rays: general – methods: statistical – pulsars: general
Online-only material: color figures, machine-readable tables
1. INTRODUCTION
Astrophysical sources of high-energy γ rays (photon energies
above 10 MeV), although inherently interesting as tracers of
energetic processes in the universe, have long been hard to
identify. Only four of the 25 sources in the second COS-B
catalog had identifications (Swanenburg et al. 1981), and over
half the sources in the third EGRET catalog had no associations
with known objects (Hartman et al. 1999). A principal reason
68
National Research Council Research Associate; resident at Naval Research
Laboratory, Washington, DC 20375, USA.
69
Royal Swedish Academy of Sciences Research Fellow, funded by a grant
from the K. A. Wallenberg Foundation.
70
Funded by contract ERC-StG-259391 from the European Community.
71
Deceased.
72
NASA Postdoctoral Program Fellow, USA.
for the difficulty of finding counterparts to high-energy γ -ray
sources has been the large positional errors in their measured
locations, a result of the limited photon statistics and angular
resolution of the γ -ray observations and the bright diffuse
γ -ray emission from the Milky Way. In addition, a number of
the COS-B and EGRET sources were determined to be spurious
by follow-up analysis and observations.
A major step forward for detection and identification of high-
energy γ -ray sources came when the Gamma-ray Large Area
Space Telescope was launched on 2008 June 11. It began its sci-
entific operations two months later, and shortly thereafter, it was
renamed the Fermi Gamma-ray Space Telescope. Its primary
instrument is the Large Area Telescope (LAT; Atwood et al.
2009), the successor to the Energetic Gamma-Ray Experiment
Telescope (EGRET) on the Compton Gamma-Ray Observatory
(Thompson et al. 1993). The LAT offers a major increase in
2
The Astrophysical Journal, 753:83 (22pp), 2012 July 1 Ackermann et al.
sensitivity over EGRET, allowing it to study the 100 MeV to
∼300 GeV γ -ray sky in unprecedented detail.
The high sensitivity, improved angular resolution, and nearly
uniform sky coverage of the LAT make it a powerful tool for
detecting and characterizing large numbers of γ -ray sources.
The Fermi-LAT First Source Catalog (1FGL; Abdo et al.
2010a) lists 1451 sources detected during the first 11 months
of operation by the LAT, of which 821 were shown to be
associated with at least one plausible counterpart. Of these,
698 were extragalactic (mostly active galactic nuclei or AGNs)
and 123 were Galactic (mostly pulsars and supernova remnants
(SNRs), but also pulsar wind nebulae (PWNe) and high-mass
X-ray binaries). After the publication of the 1FGL catalog, the
association panorama evolved very quickly with the release of
the catalog of AGNs (1LAC; Abdo et al. 2010b)aswellasa
catalog of PWNe and SNRs (Ackermann et al. 2011).
Here, as a starting point for our multivariate classification
strategy, we consider the entire original list of 630 1FGL sources
that remain unassociated with plausible counterparts at other
wavelengths. A plausible counterpart is a member of a known or
likely γ -ray emitting class located close to the 95% uncertainty
radius of a given 1FGL source, with an association confidence of
80% or higher (Abdo et al. 2010a). The 95% uncertainty radii for
1FGL source locations are typically 10
. While greatly improved
over the degree-scale uncertainties of previous instruments,
these position measurements are still inadequate to make firm
identifications based solely on location.
We have taken a multi-pronged approach toward understand-
ing these unassociated 1FGL sources, using all the available
information about the γ -ray sources. Information about loca-
tions, spectra, and variability has been combined with proper-
ties of the established γ -ray source classes and multiwavelength
counterpart searches.
Here we look in depth at the properties of the 1FGL
unassociated sources, and investigate the implications of those
characteristics. Specifically, this paper addresses five primary
questions.
1. What do the γ -ray properties of the unassociated 1FGL
sources reveal about these sources (Section 2)?
2. What does our understanding of the γ -ray properties of the
associated sources suggest about the possible source class
for each of the 1FGL unassociated sources (Section 3)?
3. What new associations or multiwavelength counterparts
have been found beyond those from the first LAT catalog
(Section 4)?
4. Do the new classifications properly predict sources that
have been associated since the release of the 1FGL catalog
(Section 5)?
5. What do the new classifications and associations imply
about the existence of unknown new γ -ray source classes
(Section 6)?
Although the 2FGL catalog (Nolan et al. 2012) was being
developed in parallel with the present work, we focus on the
1FGL results, where some follow-up results are available for
comparison with the methods of this work. Such follow-up
observations for 2FGL have yet to be done.
2. GAMMA-RAY PROPERTIES OF UNASSOCIATED
FERMI-LAT SOURCES
In the 1FGL catalog (hereafter “1FGL”; Abdo et al. 2010a),
source identifications and associations were assigned through
an objective procedure. For a source to be considered identified
in the 1FGL catalog, detection of periodic emission (pulsars or
X-ray binaries) or variability correlated with observations at
other wavelengths (blazars) was required. Additionally, mea-
surement of an angular extent consistent with observations at
other wavelengths was used to declare identifications for a few
sources associated with SNRs and radio galaxies (Abdo et al.
2009a, 2010c, 2010d, 2010e). Associations were reported only
for sources with positional correlations between LAT sources
and members of plausible source classes (based on Bayesian
probabilities of finding a source of a given type in a LAT error
box). This automated procedure was based on a list of 32 cat-
alogs that contain potential counterparts of LAT sources based
either on prior knowledge about classes of high-energy γ -ray
emitters or on theoretical expectations. In addition, it indicated
coincident detections at radio frequencies and TeV energies,
as well as positional coincidences with EGRET and AGILE
sources.
In total 821 of the 1451 sources in the 1FGL catalog (56%)
were associated with at least one counterpart by the automated
procedure, with 779 being associated using the Bayesian method
while 42 are spatially correlated with extended sources based
on overlap of the error regions and source extents. From the
simulations in 1FGL we expect that ∼57 among the 821 sources
(7%) are associated spuriously in 1FGL. We found the initial list
of unassociated sources by simply extracting the list of 1FGL
sources without any association from the 1FGL catalog. These
sources are spread across the sky, with about 40% located within
10
◦
of the Galactic plane.
Sources without firm identifications that are in regions of
enhanced diffuse γ -ray emission along the Galactic plane or are
near local interstellar cloud complexes (like Orion), sources that
lie along the Galactic ridge (300
◦
<l<60
◦
, |b| < 1
◦
), and
sources that are in regions with source densities great enough
that their position error estimates overlap in the γ -ray data are
called c-sources, as their 1FGL designator has a “c” appended
to indicate “caution” or “confused region.” The remainder of
the unassociated sources did not have a “caution” designator in
1FGL, and here are called “non-c” sources.
The positions, variability, and spectral information given in
the catalog provide an important starting point for the charac-
terization of LAT unassociated sources. We can easily compare
intrinsic properties of the 1FGL sources such as spectral in-
dex, curvature index, and flux in different energy bands for both
associated and unassociated populations, potentially providing
insight into the likely classes of the unassociated sources.
For the 1FGL catalog, the limiting flux for detecting a source
with photon spectral index Γ = 2.2 and Test Statistic of 25
(TS = 2Δ log(likelihood); Mattox et al. 1996) varied across the
sky by about a factor of five (see Figure 19 of 1FGL). This non-
uniform flux limit is due to the non-uniform Galactic diffuse
background and non-uniform exposure (mostly arising from the
passage of the Fermi observatory through the South Atlantic
Anomaly).
As discussed in 1FGL, when the variability and spectral
curvature properties of Fermi-LAT sources are compared against
each other, a clear separation is visible between bright sources
with AGN associations and those with pulsar associations. In
Figure 1 (top panel), pulsars lie in the lower right-hand quadrant
and AGNs lie in the upper half. However, the two classes mix in
the lower left-hand quadrant, making it difficult to distinguish
between them. This region of parameter space is home to much
of the unassociated source population (bottom panel). A closer
look at these and other properties of the known sources gives
3
The Astrophysical Journal, 753:83 (22pp), 2012 July 1 Ackermann et al.
Figure 1. Comparison of the 1FGL Variability Index vs. Curvature Index for the
associated sources (top panel) and unassociated sources (bottom). A separation
between the AGN (crosses) and pulsar (circles) populations is evident. However
the unassociated sources mainly lie in the region where those two populations
overlap.
(A color version of this figure is available in the online journal.)
clues to methods of separating the two major types, allowing us
to classify some of the unassociated sources as likely members
of one of these two source types (Section 3).
2.1. Source Locations and Flux Distributions
The spatial distributions of the major source types (AGNs,
pulsars, unassociated sources) are given in Table 1.Itisclear
that there is a significant excess of unassociated sources at
low Galactic latitudes (|b| < 10
◦
) where 63% of the detected
sources have no formal counterparts, compared with only 36%
unassociated at |b| > 10
◦
.
Figure 2 shows the spatial distribution of LAT unassociated
sources, with the positions of non-c sources shown as crosses
and the c-source positions given by circles. As for the EGRET
(3EG) catalog sources, the distribution is clearly not isotropic
(Hartman et al. 1999). One consideration when interpreting the
distribution of unassociated 1FGL sources is that a number of
Tab le 1
Spatial Distribution of Various Source Associations from the
1FGL and 1LAC Catalogs
Source Sources at Sources at Ridge
a
Class |b| > 10
◦
|b| < 10
◦
Sources
Associated 670 151 31
AGN 642 51 1
Pulsars 16 47 11
SNRs/PWNe 1 45 19
Other 11 8 0
Unassociated 373 257 88
Non-c sources 354 139 0
c-sources 19 118 88
Notes.
a
Here, the Galactic ridge is defined as sources with |b| < 1
◦
and
|l| < 60
◦
. This value is a subset of the previous column of |b| < 10
◦
sources.
Bold indicates the totals in each column for associated and unassociated sources.
The unbolded values are subsets.
the remaining unassociated sources are in low Galactic latitude
regions where catalogs of AGNs have limited or no coverage,
reducing the fraction of AGN associations. If we bin the different
source types by Galactic latitude (Figure 3), we see a clear
absence of AGN associations in the central 10
◦
of the Galaxy
(|b| < 5
◦
), while in the same region there is a spike in the
number of unassociated sources.
The unassociated sources have an average flux of 3.1 ×
10
−9
photons cm
−2
s
−1
(E>1 GeV), while the associated
population averages are 5.5 × 10
−8
photons cm
−2
s
−1
for pulsars
and 2.7 × 10
−9
photons cm
−2
s
−1
for AGNs.
In the Galactic plane a γ -ray source must be brighter than
at high latitudes in order to be detected above the strong
Galactic diffuse emission. Figure 4 (left panel) shows the
1FGL source flux distribution versus Galactic latitude for three
longitude bands. It is clear that the Galactic plane (|b| <
2.
◦
5) is dominated by Galactic diffuse emission, raising the
flux detection threshold to >5 × 10
−9
photons cm
−2
s
−1
.
This is reflected in the average flux of the unassociated
c-sources which at 8.2 × 10
−9
photons cm
−2
s
−1
is significantly
higher than that for the non-c unassociated source population
(1.7 × 10
−9
photons cm
−2
s
−1
). Outside the central region
of the Galaxy, the flux threshold is lower than that shown in
Figure 4.
As was the case for COS-B and EGRET, it is likely that a
subset of the unassociated sources are spurious, resulting from
an imperfect Galactic diffuse model. Such sources probably
have very low significance, poor localization, and a spectral
shape that mimics that of the Galactic diffuse emission itself.
The c-sources in the 1FGL catalog are candidates to be sources
of this type.
As discussed in Section 4.7 of 1FGL, the latitude distribution
of the Galactic ridge (300
◦
<l<60
◦
) unassociated sources
shows a sharp narrow peak in the central degree (|b| < 0.
◦
5)
of the Galaxy (Figure 4, right). If this feature is not an artifact,
and we assume these sources originate in a Galactic population,
then the scale height for this population must be ∼50 pc, to keep
the average distance to the sources within the Galaxy. Such a
scale height does not correspond to any known population of
γ -ray sources, making it likely that a number of the sources in
the Galactic ridge are spurious.
2.2. Spectral Properties
The 1FGL catalog provides spectral information that may
be useful for distinguishing between different source classes.
4
The Astrophysical Journal, 753:83 (22pp), 2012 July 1 Ackermann et al.
Figure 2. 1FGL sky map with the positions of the unassociated sources marked. Here, the non-c unassociated sources are indicated by crosses, the c-sources by circles.
(A color version of this figure is available in the online journal.)
Figure 3. Distribution of 1FGL source types by Galactic latitude. The sources associated with AGNs (blue line) show a clear deficit at low latitudes, while the same
region hosts a large number of unassociated sources (yellow line) and identified pulsars (red line).
(A color version of this figure is available in the online journal.)
As part of the 1FGL analysis all sources were fit with a
power-law spectral form and the spectral indices were included
in the catalog. In addition, the catalog includes a “curvature
index,” which measures the deviation of the spectrum from
the simple power-law form for each source. This means the
curvature index is more a measure of the quality of the power-
law spectral fit than of the intrinsic spectral shape. Figure 5
shows the distributions of the spectral index (top panel) and
curvature index (middle panel) with respect to flux. Neither of
these parameters appears to discriminate well between the AGN
and pulsar populations. In addition, the relationship is nearly
linear for the curvature index, indicating that this parameter
is strongly correlated with flux. That is, fainter sources have
relatively poorly measured spectra that cannot be measured to
be significantly different from power laws. This means that faint
γ -ray sources provide less discriminating information than
bright sources.
The majority of γ -ray AGNs are blazars, which are relativistic
jet sources with the jets directed toward Earth. An important
property of blazars is their typical γ -ray spectral index, which
offers some discrimination power between Flat Spectrum Radio
Quasars (FSRQs) and BL Lac objects (Abdo et al. 2010f). The
spectra of blazars in both of these sub-classes are typically
well described as broken power laws in the LAT energy range,
and the distributions of spectral indices for FSRQs and BL
Lac objects are compatible with Gaussians (Abdo et al. 2009b,
2010b). However, because pulsar spectra are not well described
by power laws, the spectral index of a power-law fit is not a
good discriminator between pulsar and AGN classes.
As mentioned before, Figure 5 (middle) shows that the
curvature index for pulsars is strongly correlated with flux.
This is primarily because many of the pulsars detected in the
1FGL catalog are strong γ -ray sources, with brighter pulsars
having a more significant spectral curvature than fainter pulsars.
5
