THE THIRD CATALOG OF ACTIVE GALACTIC NUCLEI DETECTED BY THE FERMI LARGE AREA TELESCOPE

To define the criteria that a source must fulfill to be considered an AGN, the ingredients are primarily the optical spectrum and to a lesser extent other characteristics such as radio loudness, flat/steep radio spectrum between 1.4 and 5 GHz, broadband emission, flux variability, and polarization.

We stress that we are classifying the candidate counterpart to a 3FGL source. If available, the earlier classification in the literature of each reported candidate counterpart was checked.

3.1.1. Optical Classification

3.1.2. SED Classification

To better characterize the candidate counterparts of the 3FGL sources that we consider to be candidate blazars or more generally radio-loud AGNs, we studied their broadband SEDs by collecting all data available in the literature.⁷⁶

We use the estimated value of the (rest-frame) broadband-SED synchrotron peak frequency ${\nu }_{\mathrm{peak}}^{{\rm{S}}}$ to classify the source as either a low-synchrotron-peaked blazar (LSP, for sources with ${\nu }_{\mathrm{peak}}^{{\rm{S}}}\lt {10}^{14}$ Hz), an intermediate-synchrotron-peaked blazar (ISP, for 10¹⁴ Hz $\;\lt \;{\nu }_{\mathrm{peak}}^{{\rm{S}}}\lt {10}^{15}$ Hz), or a high-synchrotron-peaked blazar (HSP, if ${\nu }_{\mathrm{peak}}^{{\rm{S}}}\gt {10}^{15}$ Hz). We refer the reader to the 2LAC paper for the list of broadband data used in this procedure.

The estimation of ${\nu }_{\mathrm{peak}}^{{\rm{S}}}$ relies on a 3rd-degree polynomial fit of the low-energy hump of the SED performed on a source-by-source basis, while in previous catalogs (1LAC, 2LAC) an empirical parameterization of the SED based on the broadband indices ${\alpha }_{\mathrm{ro}}$ (radio-optical) and ${\alpha }_{\mathrm{ox}}$ (optical-X-rays) was used (see Abdo et al. 2010a). In this new method, some sources changed SED classification with respect to the 2LAC (see below).

This new procedure allows more objects to be assigned peak parameters than the empirical method since there is no need for a measured X-ray flux if the curvature is sufficiently pronounced in the IR-optical band. Even though a scrupulous check was performed for each individual source, caution is advised in using these ${\nu }_{\mathrm{peak}}^{{\rm{S}}}$ values that were determined using non-simultaneous broadband data. Significant contamination from thermal/disk radiation may result in overestimation of the ${\nu }_{\mathrm{peak}}^{{\rm{S}}}$ values of FSRQs, while the contribution of the host galaxy may bias the peak estimate toward lower frequencies in BL Lacs. Comparing the two procedures indicates that the new procedure leads to an average shift of +0.26 (rms: 0.49) and −0.05 (rms: 0.64) in $\mathrm{log}{\nu }_{\mathrm{peak}}^{{\rm{S}}}$ relative to the previous one for FSRQs and BL Lacs, respectively, which we take as typical systematic uncertainties.

In the electronic tables, we report the so-obtained observer-frame values of ${\nu }_{\mathrm{peak}}^{{\rm{S}}}$, as well as the rest-frame values (i.e., corrected by a $(1+z)$ factor). For BL Lac and BCU sources without measured redshifts, a redshift z = 0 was assumed for the SED classification, but we omit these sources in figures showing ${\nu }_{\mathrm{peak}}^{{\rm{S}}}$. Assuming a redshift of 1 for these sources as suggested by Giommi et al. (2013) would lead to a shift in the rest-frame $\mathrm{log}{\nu }_{\mathrm{peak}}^{{\rm{S}}}$ of +0.3, taken as an additional systematic uncertainty.

The ${\nu }_{\mathrm{peak}}^{{\rm{S}}}$ distributions for FSRQs and BL Lacs are displayed in Figure 3. The FSRQ distribution is sharply peaked around $\mathrm{log}{\nu }_{\mathrm{peak}}^{{\rm{S}}}$ = 13 while BL Lacs span the whole parameter space from low (LSP) to the highest frequencies (HSP). The BCU distribution resembles that of BL Lacs with an additional fairly weak component akin to FSRQs at this low ${\nu }_{\mathrm{peak}}$ end.

Zoom In Zoom Out Reset image size

3.2.1. The Bayesian Association Method

3.2.2. The Likelihood-ratio Association Method

The adopted threshold for the association probability is 0.80 in either method. This value represents a compromise between association efficiency and purity. As in previous LAC catalog versions, we define a Clean Sample as 3LAC single-association sources free of the analysis issues mentioned in Section 2. Table 1 compares the performance of the two methods in terms of the total number of associations, the estimated number of false associations N_false, calculated as ${N}_{\mathrm{false}}={\displaystyle \sum }_{i}(1-{P}_{i})$, where P_i is the association probability for the ith source, and the number of sources associated solely via a given method, N_S, for the full and Clean samples.

The fraction of sources associated by both methods is 71% (1150/1591), 379, and 62 sources being solely associated with the Bayesian and LR methods, respectively. Among the former, 177 sources are associated due to the list of WISE gamma-ray blazar candidates only (over 1000 3FGL sources have counterparts in that catalog). The overall false-positive rate is 1.9%. The estimated number of false positives among the 571 sources not previously detected in 2FGL and previous LAT catalogs is 12.0 (2.1%).

Figure 4 displays the distributions of separation distance between the gamma-ray sources and their assigned counterparts, normalized to $\sigma ={\theta }_{95}/\sqrt{-2\mathrm{log}(0.05)}$, for the whole sample and for the newly detected sources. Both agree well with the distributions expected for real associations, as expected from the overall low false-positive rate.

Zoom In Zoom Out Reset image size

In this section, we point out blazar candidates derived from the recent work of Petrov et al. (2013) but not all included in 3LAC. Using the Australia Telescope Compact Array (ATCA) at 5 GHz and 9 GHz, Petrov et al. (2013) detected 424 sources in the LAT error ellipses of southern unassociated 2FGL sources. They found that 84 of them have radio-source counterparts with a spectral index flatter (i.e., greater) than −0.5.

The 424 sources are characterized by weak radio fluxes ($\lt 100$ mJy), and were thus missing from the previous AT20G. Flat spectrum radio sources cannot be directly associated with extragalactic sources like blazars, as peculiar Galactic objects (like, for example, η Carinae, microquasars, compact H ii regions, planetary nebulae) can also exhibit a flat radio spectrum. On the other hand a steep radio spectrum does not rule out an extragalactic nature. A total of 24 sources among the 84 flat-spectrum ones are included in 3LAC, as they now fulfill the required criterion (association probability greater than 0.8). An additional 21 sources listed in Table 2 show double-humped radio-to-gamma-ray SEDs resembling those of BCU, but they have association probabilities below threshold. More data may help secure these associations in the future.

Table 2.  List of ATCA Blazar Candidates

3FGL Name	Counterpart name	R.A. radio	Decl. radio	Class count	Log(${\nu }_{\mathrm{peak}}^{{\rm{S}}}$[Hz])	2FGL Name
		(°)	(°)
J0102.1+0943	NVSS J010217+094407	15.57133	9.73622	BCU II	14.419	J0102.2+0943
J0437.7–7330	SUMSS J043836–732921	69.65392	−73.48994	BCU III	⋯	J0438.0–7331
J0725.7–0550	NVSS J072547–054832	111.44867	−5.80753	BCU III	⋯	J0725.8–0549
J0737.8–8245	SUMSS J073706–824836	114.47621	−82.73703	BCU III	⋯	J0737.5–8246
J0937.9–1435	NVSS J093754–143350	144.47783	−14.56414	BCU II	17.150	J0937.9–1434
J1016.6–4244	1RXS J101620.6–424733	154.08650	−42.78975	BCU II	15.600	J1016.4–4244
J1038.0–2425	NVSS J103824–242355	159.59987	−24.39869	BCU II	12.550	J1038.2–2423
J1117.2–5338	MGPS J111715–533816	169.31279	−53.63783	BCU II	14.755	J1117.2–5341
J1115.0–0701	NVSS J111511–070238	168.79832	−7.04417	BCU III	⋯	J1115.0–0701
J1207.2–5052	SUMSS J120719–505350	181.79211	−50.86061	BCU III	⋯	J1207.3–5055
J1240.3–7149	MGPS J124021–714901	190.08821	−71.81653	BCU III	⋯	J1240.6–7151
J1249.1–2808	NVSS J124919–280833	192.33118	−28.14239	BCU II	15.080	J1249.5–2811
J1424.3–1753	NVSS J142412–175010	216.05145	−17.83611	BCU II	15.750	J1424.2–1752
J1539.2–3324	NVSS J153911–332209	234.79825	−33.36822	BCU III	⋯	J1539.2–3325
J1704.4–0528	NVSS J170433–052839	256.14075	−5.47753	BCU II	15.200	J1704.6–0529
J1747.3+0324	NVSS J174733+032703	266.88860	3.45119	BCU III	⋯	J1747.6+0324
J1757.7–6030	SUMSS J175734–603032	269.39413	−60.50794	BCU III	⋯	J1757.5–6028
J2034.6–4202	SUMSS J203451–420024	308.71274	−42.01044	BCU II	15.640	J2034.7–4201
J2046.7–4259	SUMSS J204643–425711	311.68353	−42.95358	BCU III	⋯	J2046.2–4259
J2134.5–2131	NVSS J213430–213032	323.62580	−21.50858	BCU II	15.410	J2134.6–2130
J2258.2–3645	NVSS J225815–364433	344.56195	−36.74264	BCU II	15.150	J2257.9–3646

Download table as:  ASCIITypeset image

Table 3 summarizes the 3LAC source statistics. The 3LAC includes 1591 objects, with 467 FSRQs, 632 BL Lacs, 460 BCUs, and 32 non-blazar AGNs. Their properties are given in Table 4.

Table 4.  High Latitude ($| b| \gt 10^\circ$) 3LAC Full Sample

3FGL Source	Counterpart	R.A.	Decl.	AngSep	${\theta }_{95}$	Optical	SED	log(${\nu }_{\mathrm{peak}}^{{\rm{S}},\mathrm{meas}}$)	log(${\nu }_{\mathrm{peak}}^{{\rm{S}}}$)	z	Prob.	Rel.	Rel.	Compton
Name	Name	(°)	(°)	(°)	(°)	Class	Class				Bay.	LRRG	LRXG	Dominance
(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)	(15)
J0001.2−0748a	PMN J0001−0746	0.32510	−7.77411	0.042	0.075	BLL	ISP	14.486	14.486	⋯	0.9978	0.859	0.00	⋯
J0001.4+2120a	TXS 2358+209	0.38502	21.22679	0.113	0.199	FSRQ	ISP	14.163	14.486	1.10600	0.924	0.00	0.000	2.40
J0002.2−4152a	1RXS J000135.5−415519	0.38642	−41.92367	0.137	0.174	BCU II	HSP	15.800	15.800	⋯	0.972	0.000	0.000	−0.59
J0003.2−5246a	RBS 0006	0.83121	−52.79103	0.017	0.065	BCU II	HSP	16.850	16.850	⋯	0.998	0.000	0.900	−0.52
J0003.8−1151	PMN J0004–1148	1.02048	−11.81622	0.076	0.114	BCU II	LSP	12.515	12.515	⋯	0.995	0.869	0.000	⋯
J0003.8−1151	PKS 0001−121	0.92848	−11.86372	0.030	0.114	BCU I	⋯	⋯	⋯	1.30999	0.988	0.871	0.000	⋯

Note. Columns 1 and 2 are the 3FGL and counterpart names, columns 3 and 4 are the counterpart J2000 coordinates, column 5 gives the angular separation between the gamma-ray and counterpart positions, column 6 is the 95% error radius on the gamma-ray position, column 7 lists the optical class, column 8 is the spectral energy distribution (SED) class (depending on the synchrotron peak frequency given in column 9), column 10 is the synchrotron peak frequency corrected for the redshift shown in column 11, and columns 12–14 report the probability for the Bayesian method and the two reliability values, LR_RG and LR_XG, for the radio–gamma-ray match and the X-ray–gamma-ray match respectively. Column 15 reports log(CD).

^aRefers to sources in the Clean Sample.

Only a portion of this table is shown here to demonstrate its form and content. A machine-readable version of the full table is available.

Download table as:  DataTypeset image

A total of 1563 gamma-ray sources have been associated with radio-loud AGNs among 2192 $| b| \gt 10^\circ$ 3FGL sources, corresponding to an overall association fraction of 72%. The fraction changes substantially between the northern and southern celestial hemispheres (843/1136 = 74% and 731/1056 = 69% respectively), an effect essentially entirely driven by unassociated southern-hemisphere BL Lacs as discussed below.

Only sources in the Clean Sample will be used in the following tallies and figures unless stated otherwise. It includes 1444 objects with 414 FSRQs, 604 BL Lacs, 402 BCUs, and 24 non-blazar AGNs.

A comparison of the results inferred from the 3LAC and 2LAC enables the following observations:

• 1.  
  The 3LAC Clean Sample includes 619 more sources than the 2LAC Clean Sample, i.e., a 75% increase. Of these, 477 sources are new (81 FSRQs, 146 BL Lacs, 240 blazars of unknown type, 10 non-blazar objects); the other sources were present in previous Fermi catalogs but not included in Clean Samples for various reasons (e.g., the corresponding gamma-ray sources were not associated with AGNs, had more than one counterpart or were flagged in the analysis). The fraction of new sources (not present in 1FGL or 2FGL) is slightly higher for hard-spectrum (i.e., ${\rm{\Gamma }}\lt 2.2$) sources than for soft-spectrum ones (i.e., ${\rm{\Gamma }}\gt 2.2$), 51% versus 47%, respectively, but the relative increase reaches 72% for very hard-spectrum (i.e., ${\rm{\Gamma }}\lt 1.8$) sources.
• 2.  
  The fraction of BCU has increased notably between the two catalogs (from 20% to 28%). The number of these sources in the 3LAC Clean Sample has increased by more than a factor of 2.5 relative to that in the 2LAC Clean Sample, being almost equal to the number of FSRQs. This increase is mainly due to the lower probability of having a published high-quality spectrum available for these fainter sources because of the lack of optical/near-infrared observing programs. The census of the BCU sources in the Clean Sample is: 49 BCU I, 308 BCU II, 45 BCU III.
• 3.  
  The relative increase in BCUs drives a drop in the proportions of FSRQs and BL Lacs, which only represent 29% and 41% of the 3LAC Clean Sample, respectively (38% and 48% for 2LAC). The relative increase in the number of sources with respect to 2LAC is 34% and 42% for FSRQs and BL Lacs respectively.
• 4.  
  Out of 827 sources in the 2LAC Clean Sample, a total of 69 are missing in the 3LAC Clean Sample (42 in the full sample), some of them probably due to variability effects. A few others are present in 3FGL but with shifted positions, ruling out their association with their former counterparts.

The loci of sources in the Clean Sample are shown in Figure 5, both in Galactic and celestial coordinates. The deficit in classified AGNs in the region of the celestial south pole already reported in 2LAC is clearly visible, while a relative excess is seen in the region of the celestial north pole. This anisotropy is mainly driven by BL Lacs, with 51% more sources in the northern Galactic hemisphere (362) than in the southern one (242). This effect is ascribed to the relative incompleteness of the counterpart catalogs in the southern hemisphere (for instance, NVSS only covers the $\delta \gt -40^\circ$ sky, where δ is the declination). It is very partially offset by an observed relative excess (54 sources) of associations with BCU in the south relative to the north.

Zoom In Zoom Out Reset image size

Blazars represent the overwhelming majority of 3LAC AGNs, with non-blazar AGNs only constituting 2% of the sample. In 2LAC, eleven sources were classified as AGNs, i.e., were neither confirmed blazars nor blazar candidates (such as BCUs). Although there may have been evidence for their flatness in radio emission or broadband emission, our intensive optical follow-up program did not provide clear evidence for optical blazar characteristics. Nine of them remain in 3LAC, and are now all classified as BCUs, except for one now classified as a BL Lac.

Misaligned AGNs (MAGNs), with jets pointing away from the observer, are not favored GeV sources. By MAGNs we mean radio-loud AGNs with jets directed at large angles relative to the line-of-sight that display steep radio spectra (${\alpha }_{r}\geqslant 0.5$, with the usual convention that ${F}_{\nu }\propto {\nu }^{-\alpha })$ and bipolar or quasi-symmetrical structures in radio maps. The larger jet inclination angle relative to blazars means the observed radio emission from the relativistic jet is not significantly Doppler boosted, making it less prevalent over other radio components such as synchrotron radiation from mildly relativistic outflows or extended radio lobe emission (Abdo et al. 2010e).

Table 5 summarizes the non-blazar objects and MAGNs in the 3FGL/3LAC, also noting their previous appearances in the 2FGL/2LAC and 1FGL/1LAC. All the 1FGL sources, detected in 11 months of exposure, were subsequently studied with 15 months of data (Abdo et al. 2010e).

Table 5.  Non-blazar Objects and Misaligned AGNs

Name	3FGL	2FGL	1FGL	Type	Photon Index	Notes
NGC 1218	J0308.6+0408a	⋯	J0308.3+0403a	FRI	2.07 ± 0.11	⋯
IC 310	J0316.6+4119a	J0316.6+4119	⋯	FRI/BLL	1.90 ± 0.14	Neronov et al. (2010)
NGC 1275	J0319.8+4130a	J0319.8+4130a	J0319.7+4130a	FRI	2.07 ± 0.01	Abdo et al. (2009c); Kataoka et al. (2010)
1 H 0323+342	J0325.2+3410a	J0324.8+3408a	J0325.0+3403a	NLSy1	2.44 ± 0.12	⋯
4C+39.12	J0334.2+3915a	⋯	⋯	FRI/BLL?	2.11 ± 0.17	Giovannini et al. (2001)
TXS 0348+013	J0351.1+0128a	⋯	⋯	SSRQ	2.43 ± 0.18	⋯
3C 111	J0418.5+3813	⋯	J0419.0+3811	FRII	2.79 ± 0.08	Abdo et al. (2010e); Kataoka et al. (2011); Grandi et al. (2012)
Pictor A	J0519.2−4542a	⋯	⋯	FRII	2.49 ± 0.18	Brown & Adams (2012); Kataoka et al. (2011)
PKS 0625−35	J0627.0−3529a	J0627.1−3528a	J0627.3−3530a	FRI/BLL	1.87 ± 0.06	⋯
4C+52.17	J0733.5+5153	⋯	⋯	AGN	1.74 ± 0.16	Part of a duplicate association. Most probable counterpart is a BCU III.
NGC 2484	J0758.7+3747a	⋯	⋯	FRI	2.16 ± 0.16	quasar SDSS J075825.87+374628.7 is 08 away
4C+39.23B	J0824.9+3916	⋯	⋯	CSS	2.44 ± 0.10	⋯
3C 207	J0840.8+1315a	J0840.7+1310	J0840.8+1310	SSRQ	2.47 ± 0.09	⋯
SBS 0846+513	J0849.9+5108a	⋯	⋯	NLSy1	2.28 ± 0.04	⋯
3C 221	J0934.1+3933	⋯	⋯	SSRQ	2.28 ± 0.12	⋯
PMN J0948+0022	J0948.8+0021a	J0948.8+0020a	J0949.0+0021a	NLSy1	2.32 ± 0.05	⋯
PMN J1118–0413	J1118.2–0411a	⋯	⋯	AGN	2.56 ± 0.08	⋯
B2 1126+37	J1129.0+3705	⋯	⋯	AGN	2.08 ± 0.13	Part of a duplicate association. Most probable counterpart is a BLL.
3C 264	J1145.1+1935a	⋯	⋯	FRI	1.98 ± 0.20	⋯
PKS 1203+04	J1205.4+0412	⋯	⋯	SSRQ	2.64 ± 0.16	Part of a duplicate association. The other counterpart is an FSRQ.
M 87	J1230.9+1224a	J1230.8+1224a	J1230.8+1223a	FRI	2.04 ± 0.07	Abdo et al. (2009d)
3C 275.1	J1244.1+1615	⋯	⋯	SSRQ	2.43 ± 0.17	⋯
GB 1310+487	J1312.7+4828a	J1312.8+4828a	J1312.4+4827a	AGN	2.04 ± 0.03	⋯
Cen A Core	J1325.4−4301a	J1325.6−4300	J1325.6−4300	FRI	2.70 ± 0.03	radio core
Cen A Lobes	J1324.0−4330e	J1324.0−4330e	J1322.0−4515	FRI	2.53 ± 0.05	giant lobes detected (Abdo et al. 2010b)
3C 286	J1330.5+3023a	⋯	⋯	SSRQ/CSS	2.60 ± 0.16	⋯
Cen B	J1346.6−6027	J1346.6−6027	⋯	FRI	2.32 ± 0.01	Katsuta et al. (2013)
Circinus	J1413.2−6518	⋯	⋯	Seyfert	2.43 ± 0.10	Hayashida et al. (2013)
3C 303	J1442.6+5156a	⋯	⋯	FRII	1.92 ± 0.18	⋯
PKS 1502+036	J1505.1+0326a	J1505.1+0324a	J1505.0+0328a	NLSy1	2.61 ± 0.05	⋯
TXS 1613−251	J1617.3−2519	J1617.6−2526c	⋯	AGN	2.59 ± 0.10	Part of a duplicate association. Most probable counterpart is a BCU II.
PKS 1617−235	J1621.1−2331a	J1620.5−2320c	⋯	AGN	2.50 ± 0.23	⋯
NGC 6251	J1630.6+8232a	J1629.4+8236	J1635.4+8228a	FRI	2.22 ± 0.08	⋯
3C 380	J1829.6+4844a	J1829.7+4846a	J1829.8+4845a	SSRQ/CSS	2.37 ± 0.04	⋯
PKS 2004−447	J2007.8−4429a	J2007.9−4430a	J2007.9−4430a	NLSy1	2.47 ± 0.09	⋯

Notes. SSRQ implies FRII. The table includes the 34 non-blazar objects and MAGNs at all latitudes associated with 3FGL sources (Cen A Core and Cen A Lobes constitute a single object).

^aRefers to sources included in the Clean Sample of a given catalog.

Download table as:  ASCIITypeset image

M 87 was one of the first new Fermi-LAT detections (Abdo et al. 2009d) of a source classified as a non-blazar object, being a low-power Fanaroff & Riley (1974) type-I (FRI) radio galaxy. Many of the newly associated non-blazar objects are nearby FRIs—J0758.7+3747 (3C 189, a.k.a., B2 0755+37) and 3C 264. The gamma-ray detection of the latter case was recently reported in a study of its parent cluster Abell 1367 (Ackermann et al. 2011a), although the gamma-rays likely originate from the AGN. We remark that 08 away from 3C 189 lies the quasar SDSS J075825.87+374628.7 with redshift 1.50. With the resolution of the NVSS, this source cannot be disentangled from the radio emission of 3C 189. This may be the reason why this source is not present in the NVSS catalog, precluding the estimation of the association probability with the gamma-ray source.

NGC 1275 (3C 84, Perseus A) was first detected in the initial LAT bright source list based on 3 months of data (Abdo et al. 2009e). It was probably detected previously with COS-B (Strong & Bignami 1983), but not with EGRET. In the Fermi era, it is a strong source, exhibiting GeV variability (Abdo et al. 2009c; Kataoka et al. 2010). 3C 120 is not listed in any of the FGL catalogs but its detection was reported in a 15 month study (Abdo et al. 2010e). There are indications that 3C 120 undergoes a series of flares with a low long-term average flux. For instance, in 2014 September a flaring source positionally consistent with 3C 120 was detected with a high significance ($\mathrm{TS}\gt 50$; Tanaka et al. 2014). The closest 3FGL source, 3FGL J0432.5+0539, lies $0\buildrel{\circ}\over{.} 35$ away, with an 95% error radius of $0\buildrel{\circ}\over{.} 15$, hampering association with 3C 120 by our methods. This gamma-ray source has a soft spectrum (${\rm{\Gamma }}=2.7\pm 0.1$), comparable with that ascribed to 3C 120 (Abdo et al. 2010e; Kataoka et al. 2011). The possibility of two separate, soft-spectrum sources cannot be excluded. Another known example from previous lists is 3C 78 (NGC 1218; Abdo et al. 2010g).

Cen A was also reported in the initial LAT bright source list (Abdo et al. 2009e), confirming the EGRET source (Hartman et al. 1999; Sreekumar et al. 1999). It remains as the only AGN with a significant detection of extended gamma-ray emission (Abdo et al. 2010b). There is no convincing case of extended emission in other radio galaxies with relatively large radio extensions, such as Cen B (Katsuta et al. 2013), NGC 6251 (Takeuchi et al. 2012), and Fornax A. Fornax A may be a good case to investigate this emission (Cheung et al. 2007; Georganopoulos et al. 2008). The closest 3FGL source is offset from the Fornax A core by 015, while the 95%-contour distance is 0092 (see Figure 6 for a VLA 1.5 GHz image). NGC 6251 (one square degree in solid angle) was also detected by EGRET (Mukherjee et al. 2002). Its location shifted between 1LAC and 2LAC toward the western radio lobe.

Zoom In Zoom Out Reset image size

3C 111 was also a previous EGRET source (Hartman et al. 2008) and also shows apparent variability (e.g., Abdo et al. 2010e; Kataoka et al. 2011; Grandi et al. 2012). It joins the other two FR type-II sources listed in Table 5: 3FGL J1442.6+5156 (3C 303) and 3FGL J0519.2−4542 (Pictor A). The latter are also broad-lined radio galaxies (BLRGs), and are new detections. The LAT detection of Pictor A was reported by Brown & Adams (2012) following⁷⁹ a previous tentative detection (Kataoka et al. 2011).

The previous LAT detections of PKS 0625−35 and IC 310, two radio galaxies with BL Lac characteristics, were reported in 2LAC, and are confirmed. IC 310 has been classified as a head-tail galaxy (Neronov et al. 2010), but recent works have found increasing evidence for blazar-like properties, e.g., blazar-like VLBI jet structure (Kadler et al. 2012a) and extremely fast TeV variability (Aleksić et al. 2014). The source 4C +39.12 (3FGL J0334.2+3915) was classified as a low-power compact source by Giovannini et al. (2001), separate from its Fanaroff–Riley classification. Two new compact steep-spectrum (CSS) sources are detected: 3FGL J1330.5+3023 (3C 286) and 3FGL J0824.9+3916 (4C +39.23B). While both are CSS, the latter is a duplicate association (the other association being the FSRQ blazar 4C+39.23) so it is not in the Clean Sample. The former has the morphology of a medium symmetric object (MSO), like that of the LAT-detected FSRQ 4C +55.17 (McConville et al. 2011).

The gamma-ray detections of 3C 207 and 3C 380 were first reported in 1LAC. They appear in the 3CRR catalog (Laing et al. 1983) by virtue of their bright low-frequency emission due to the presence of kpc-scale extended steep-spectrum radio lobes, and thus are formally classified as SSRQs. However, they contain pronounced flat-spectrum radio cores with superluminal motions measured in their parsec-scale jets, indicating that they are the most well-aligned sources to our line of sight among the SSRQs in the 3CRR (e.g., Wilkinson et al. 1991; Hough 2013; Lister et al. 2013). New ones to highlight are 3C 275.1 (3FGL J1244.1+1615), TXS 0348+013 (3FGL J0351.1+0128), and 4C +39.26 (3FGL J0934.1+3933). The SSRQ 4C+04.40 is part of a double association (with the FSRQ MG1 J120448+0408) of 3FGL J1205.4+0412.

GB 1310+487 is a gamma-ray/radio-loud narrow-line AGN at z = 0.638, showing a gamma-ray flare in November 2009 and located behind the disk of an unrelated emission-line galaxy at z = 0.500 (Sokolovsky et al. 2014).

Circinus, a type-2 Seyfert galaxy located at b = −38 and thus not in 3LAC, was recently detected (Hayashida et al. 2013). Other Seyfert detections were investigated (Teng et al. 2011; Ackermann et al. 2012b; Lenain et al. 2010), but were found to be starburst galaxies (Ackermann et al. 2012a).

The detections of NGC 6951 (classified as a Seyfert 2 galaxy and a LINER, reported in 1LAC but missing in 2LAC), 3C 407 (a source with broad emission lines but with a fairly steep radio spectrum and reported in 2LAC), and NGC 6814 (type 1.5 Seyfert galaxy, also reported in 2LAC) are not confirmed. The same conclusion applies to PKS 0943−76 (studied in Abdo et al. 2010e). The previous claim that it has a FRII morphology was based on a low-resolution radio map from Burgess & Hunstead (2006). The offset between the 4 year source and PKS 0943−76 is $0\buildrel{\circ}\over{.} 22$, while the radius of the source location region at the 95% confidence level is $0\buildrel{\circ}\over{.} 12$. ESO 323−G77 (type 2 Seyfert galaxy), and PKS0943−76 (radio galaxy), both reported in 2LAC, were actually both mis-associated because of an error in the LR association method (Ackermann et al. 2015).

Five sources are associated with NLSy1. Four of them were included in 2LAC: 3FGL J0325.2+3410 (BZU J0324+3410), 3FGL J0948.8+0021 (PMN J0948+0022), 3FGL J1505.1+0326 (BZQ J1505+0326), and 3FGL J2007.8−4429 (BZQ J2007−4434), while 3FGL 0849.9+5108 (SBS 0846+513) was first reported by Donato & Perkins (2011) and further studied by D'Ammando et al. (2012, 2013).

The highest redshift reported in 2LAC for an HSP-BL Lac was 0.7. The 3LAC lists seven (six in the Clean Sample) HSP-BL Lacs with redshifts greater than 1, six (five in the Clean Sample) of which were included in 2LAC but with other classifications or redshifts. They are briefly discussed below.

3FGL J0008.0+4713 is associated with MG4 J000800+4712. The redshift reported in 2LAC was 0.28 and its SED classification was LSP. Shaw et al. (2013) derived a redshift of 2.1 from the clear onset of the Lyα forest and their new procedure for estimating SED class together with WISE data classified this source as an HSP.

3FGL J0630.9−2406 is associated with TXS 0628−240, an HSP-BL Lac for which $z\gtrsim 1.238$ was determined from certain absorption features by Landt (2012).

3FGL J1109.4+2411 is associated with 1ES 1106+244 and new spectroscopy from SDSS changed the redshift to 1.220.

3FGL J1312.5−2155 is associated with PKS 1309−216. In Shaw et al. (2013) a plausible Mg ii feature is found; this single-line identification is in a small allowed redshift range ($z\simeq 1.6$). However, previous data (Massaro et al. 2009) show a questionable redshift of 1.491.

3FGL J2116.1+3339 is associated with B2 2114+33. The redshift quoted in 2LAC was 0.35, but a recent measurement by Shaw et al. (2013) gives z = 1.596, identifying a significant broad emission feature with C iv, consistent with a weak bump in the far blue at Lyα. A lower redshift is possible if the purported Lyα line is not real.

The newly detected source is 3FGL J0814.5+2943, associated with FBQS J081421.2+294021 at z = 1.084 (from SDSS DR3, Ahn et al. 2012).

The highest redshift BL Lac object is 3FGL J1450.9+5200, associated with BZB J1450+5201 with redshift z = 2.41 coming from new observations in Shaw et al. (2013). The presence of the Lyα forest can suppress a part of the optical spectrum, resulting in ISP classification, so the intrinsic synchrotron peak position is probably greater than our estimate.

A low-redshift source, reported as a BCU in BZCAT, has recently been classified as an FSRQ by G. Chiaro & D. Bastieri (2014, private communication): SBS 1646+499 (3FGL J1647.4+4950) with z = 0.0467.

Two HSP-FSRQs have been detected: BZB J0202+0849 (3FGL J0202.3+0851) and NVSS J025037+171209 (3FGL J0250.6+1713) with LAT spectral photon indices of 2.05 ± 0.16 and 1.98 ± 0.19, respectively. 3FGL J0202.3+0851 was classified as a BL Lac in 1LAC but new observations from Shaw et al. (2013) led to a reclassification as an FSRQ. These objects are probably transitional objects that show broad lines in the optical band when the continuum is low (see, e.g., Ruan et al. 2014).

Because of the intrinsic incompleteness of the counterpart catalogs in this sky area ($| b| \lt 10^\circ$), these sources are treated separately and are not included in the 3LAC or in the analyses presented in the rest of the paper. We report associations for 182 blazars (75% more than in 2LAC) located at $| b| \lt 10^\circ$: 24 FSRQs, 30 BL Lacs, 125 BCUs, and 3 non-blazar AGNs. They are listed in Table 6. Extrapolating from the number of high-latitude sources and assuming the same sensitivity, about 340 sources would be expected in this area. The discrepancy between expected and actual source numbers stems from the dual effect of a higher detection threshold due to a higher Galactic diffuse emission background (see Figure 1) and a higher incompleteness of the counterpart catalogs for this area.

Table 6.  Low-latitude ($| b| \lt 10^\circ$) Sample

3FGL Source	Counterpart	R.A.	Decl.	AngSep	${\theta }_{95}$	Optical	SED	log(${\nu }_{\mathrm{peak}}^{{\rm{S}},\mathrm{meas}}$)	log(${\nu }_{\mathrm{peak}}^{{\rm{S}}}$)	z	Prob.	Rel.	Rel.
Name	Name	(°)	(°)	(°)	(°)	Class	Class				Bay.	LRRG	LRXG
(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)
J0012.4+7040	TXS 0008+704	2.88293	70.75878	0.115	0.105	BCU II	LSP	13.075	13.075	⋯	0.912	0.856	⋯
J0014.6+6119	4C +60.01	3.70330	61.29543	0.031	0.061	BCU II	LSP	13.113	13.113	⋯	0.997	0.976	⋯
J0014.7+5802	1RXS J001442.2+580201	3.67471	58.03404	0.009	0.055	BLL	HSP	16.640	16.640	⋯	⋯	⋯	0.936
J0015.7+5552	GB6 J0015+5551	3.91737	55.86226	0.018	0.043	BCU II	HSP	15.791	15.791	⋯	0.998	0.868	0.952
J0035.9+5949	1ES 0033+595	8.96935	59.83459	0.010	0.018	BLL	HSP	17.120	17.120	⋯	1.000	0.935	0.979
J0047.0+5658	GB6 J0047+5657	11.75179	56.96178	0.013	0.040	BLL	⋯	⋯	⋯	0.74700	1.000	0.910	⋯
J0047.9+5447	1RXS J004754.5+544758	11.96611	54.79579	0.010	0.060	BCU II	HSP	15.896	15.896	⋯	⋯	⋯	0.890
J0102.8+5825	TXS 0059+581	15.69068	58.40309	0.023	0.020	FSRQ	LSP	12.725	12.941	0.64400	0.999	0.956	⋯
J0103.4+5336	1RXS J010325.9+533721	15.85816	53.62036	0.006	0.042	BLL	⋯	⋯	⋯	⋯	⋯	0.824	0.934
J0109.8+6132	TXS 0106+612	17.44310	61.55846	0.015	0.033	FSRQ	LSP	13.290	13.541	0.78300	0.999	0.934	⋯
J0110.2+6806	4C +67.04	17.55364	68.09478	0.022	0.023	BLL	ISP	14.864	14.864	⋯	1.000	0.972	0.895
J0131.2+6120	1RXS J013106.4+612035	22.78011	61.34260	0.013	0.022	BLL	HSP	16.300	16.300	⋯	0.999	0.850	0.979
J0131.3+5548	TXS 0128+554	22.80760	55.75361	0.056	0.082	BCU I	⋯	⋯	⋯	0.03649	0.986	0.828	0.806
J0135.0+6927	TXS 0130+691	23.66984	69.41969	0.055	0.095	BCU III	⋯	⋯	⋯	⋯	0.984	0.830	⋯
J0137.8+5813	1RXS J013748.0+581422	24.46032	58.23648	0.005	0.032	BCU II	HSP	16.580	16.580	⋯	0.999	0.921	0.969
J0148.3+5200	GB6 J0148+5202	27.08473	52.03470	0.025	0.039	BCU III	⋯	⋯	⋯	⋯	0.996	⋯	⋯
J0153.4+7114	TXS 0149+710	28.35771	71.25180	0.010	0.037	BCU I	HSP	15.690	15.699	0.02200	1.000	0.953	0.930
J0211.7+5402	TXS 0207+538	32.73495	54.08692	0.120	0.128	BCU III	⋯	⋯	⋯	⋯	⋯	0.827	⋯
J0214.4+5143	TXS 0210+515	33.57473	51.74776	0.034	0.044	BLL	HSP	15.020	15.041	0.04900	0.999	0.905	0.944
J0217.3+6209	TXS 0213+619	34.26049	62.19274	0.056	0.102	BCU III	⋯	⋯	⋯	⋯	⋯	0.800	⋯
J0223.3+6820	NVSS J022304+682154	35.76891	68.36528	0.031	0.037	BCU II	HSP	15.800	15.800	⋯	0.991	⋯	⋯
J0223.5+6313	TXS 0219+628	35.87363	63.12177	0.104	0.152	BCU III	⋯	⋯	⋯	⋯	0.945	⋯	⋯
J0228.5+6703	GB6 J0229+6706	37.34410	67.11042	0.099	0.166	BCU III	⋯	⋯	⋯	⋯	0.953	⋯	⋯
J0241.3+6542	TXS 0237+655	40.34061	65.71988	0.018	0.043	BCU II	HSP	15.500	15.500	⋯	⋯	0.903	0.910
J0250.6+5630	NVSS J025047+562935	42.69830	56.49317	0.030	0.052	BCU II	HSP	16.138	16.138	⋯	⋯	⋯	0.890
J0253.8+5104	NVSS J025357+510256	43.49003	51.04902	0.022	0.075	FSRQ	LSP	12.500	12.936	1.73200	1.000	0.912	⋯
J0302.0+5335	GB6 J0302+5331	45.59473	53.52958	0.081	0.076	BCU II	HSP	15.988	15.988	⋯	0.965	⋯	⋯
J0303.6+4716	4C +47.08	45.89684	47.27119	0.017	0.031	BLL	ISP	14.000	14.000	⋯	1.000	0.965	⋯
J0304.9+6817	TXS 0259+681	46.09168	68.36041	0.082	0.076	BCU II	LSP	12.725	12.725	⋯	0.911	0.920	⋯
J0332.0+6308	GB6 J0331+6307	52.97465	63.13727	0.016	0.051	BCU II	ISP	14.150	14.150	⋯	0.998	0.816	⋯
J0333.9+6538	TXS 0329+654	53.48641	65.61561	0.022	0.034	BLL	HSP	15.200	15.200	⋯	0.998	0.924	0.885
J0352.9+5655	GB6 J0353+5654	58.28989	56.90859	0.032	0.046	BCU II	HSP	16.315	16.315	⋯	0.996	0.820	⋯
J0354.1+4643	B3 0350+465	58.62505	46.72188	0.065	0.118	BCU III	⋯	⋯	⋯	⋯	0.977	0.904	⋯
J0358.8+6002	TXS 0354+599	59.76100	60.08946	0.055	0.121	FSRQ	LSP	12.905	13.068	0.45500	0.993	0.919	0.830
J0418.5+3813	3C 111	64.58866	38.02661	0.198	0.168	RDG	⋯	⋯	⋯	0.04850	0.961	0.949	0.807
J0423.8+4150	4C +41.11	65.98337	41.83409	0.012	0.021	BLL	LSP	13.180	13.180	⋯	1.000	0.980	⋯
J0425.2+6319	1RXS J042523.0+632016	66.35324	63.33486	0.019	0.040	BCU II	HSP	16.050	16.050	⋯	⋯	0.804	0.920
J0444.5+3425	B2 0441+34	71.15083	34.42877	0.014	0.074	BCU II	LSP	13.005	13.005	⋯	0.997	0.880	⋯
J0501.8+3046	1RXS J050140.8+304831	75.42145	30.80727	0.051	0.059	BCU II	HSP	16.100	16.100	⋯	⋯	⋯	0.886
J0502.7+3438	MG2 J050234+3436	75.62478	34.60960	0.064	0.078	BCU III	⋯	⋯	⋯	⋯	0.983	0.820	⋯
J0503.4+4522	1RXS J050339.8+451715	75.91491	45.28320	0.098	0.086	BCU II	HSP	15.645	15.645	⋯	⋯	⋯	0.844
J0512.2+2918	B2 0509+29	78.17586	29.45100	0.170	0.527	BCU III	⋯	⋯	⋯	⋯	0.891	⋯	⋯
J0512.9+4038	B3 0509+406	78.21893	40.69545	0.054	0.073	BCU II	LSP	13.635	13.635	⋯	0.999	0.933	⋯
J0517.4+4540	4C +45.08	79.37041	45.61802	0.050	0.154	FSRQ	LSP	12.900	13.165	0.83900	0.990	0.907	⋯
J0519.3+2746	4C +27.15	79.88761	27.73454	0.051	0.110	BCU III	⋯	⋯	⋯	⋯	0.992	0.944	⋯
J0521.7+2113	TXS 0518+211	80.44152	21.21429	0.008	0.014	BLL	ISP	14.335	14.380	0.10800	1.000	0.969	0.961
J0526.0+4253	NVSS J052520+425520	81.33690	42.92225	0.140	0.150	BCU II	LSP	13.145	13.145	⋯	0.942	⋯	⋯
J0528.3+1815	1RXS J052829.6+181657	82.12341	18.28188	0.048	0.060	BCU III	⋯	⋯	⋯	⋯	⋯	⋯	0.929
J0533.2+4822	TXS 0529+483	83.31611	48.38134	0.007	0.031	FSRQ	LSP	13.040	13.375	1.16200	1.000	0.950	0.876
J0539.8+1434	TXS 0536+145	84.92652	14.56266	0.031	0.071	FSRQ	LSP	12.445	13.012	2.69000	0.999	0.911	⋯
J0601.0+3837	B2 0557+38	90.26196	38.64144	0.017	0.053	BLL	LSP	13.810	13.810	⋯	⋯	0.945	⋯
J0603.8+2155	4C +22.12	90.96482	21.99381	0.066	0.058	BCU II	LSP	13.250	13.250	⋯	0.981	0.955	⋯
J0611.7+2759	GB6 J0611+2803	92.93284	28.06449	0.067	0.107	BCU III	⋯	⋯	⋯	⋯	0.991	⋯	⋯
J0620.4+2644	RX J0620.6+2644	95.16716	26.72524	0.044	0.063	BCU II	HSP	16.085	16.085	⋯	⋯	0.805	0.940
J0622.9+3326	B2 0619+33	95.71759	33.43622	0.014	0.018	BCU II	ISP	14.050	14.050	⋯	0.999	0.938	⋯
J0623.3+3043	GB6 J0623+3045	95.81747	30.74889	0.025	0.065	BCU II	ISP	14.790	14.790	⋯	0.996	0.800	⋯
J0631.2+2019	TXS 0628+203	97.75443	20.34978	0.050	0.106	BCU II	HSP	15.000	15.000	⋯	0.990	0.862	⋯
J0640.0–1252	TXS 0637–128	100.02993	−12.88761	0.015	0.040	BCU II	HSP	16.050	16.050	⋯	0.998	0.915	0.967
J0641.8–0319	TXS 0639–032	100.46305	−3.34683	0.029	0.142	BCU II	LSP	12.760	12.760	⋯	0.987	0.920	⋯
J0643.2+0859	PMN J0643+0857	100.86019	8.96056	0.066	0.063	FSRQ	LSP	13.000	13.275	0.88200	0.975	⋯	⋯
J0648.1+1606	1RXS J064814.1+160708	102.05790	16.11576	0.018	0.045	BCU II	HSP	16.300	16.300	⋯	⋯	⋯	0.904
J0648.8+1516	RX J0648.7+1516	102.19854	15.27355	0.007	0.029	BLL	HSP	15.850	15.922	0.17900	1.000	0.892	0.976
J0648.8–1740	TXS 0646–176	102.11874	−17.73484	0.109	0.155	FSRQ	LSP	12.480	12.829	1.23200	0.995	0.898	⋯
J0650.4–1636	PKS 0648–16	102.60242	−16.62770	0.019	0.094	BCU II	LSP	11.465	11.465	⋯	0.998	0.954	⋯
J0650.5+2055	1RXS J065033.9+205603	102.64681	20.93242	0.003	0.040	BCU II	HSP	15.650	15.650	⋯	⋯	⋯	0.892
J0654.5+0926	RX J0654.3+0925	103.61306	9.42644	0.032	0.231	BCU II	HSP	15.350	15.350	⋯	⋯	⋯	0.840
J0656.2–0323	TXS 0653–033	104.04634	−3.38522	0.009	0.053	FSRQ	LSP	13.495	13.708	0.63400	1.000	0.929	⋯
J0658.6+0636	NVSS J065844+063711	104.68735	6.61943	0.039	0.068	BCU II	HSP	15.000	15.000	⋯	0.999	⋯	⋯
J0700.0+1709	TXS 0657+172	105.00636	17.15603	0.016	0.116	BCU II	LSP	12.725	12.725	⋯	0.999	0.910	⋯
J0700.2+1304	GB6 J0700+1304	105.05963	13.07345	0.013	0.065	BCU II	HSP	15.425	15.425	⋯	0.998	⋯	⋯
J0702.7–1952	TXS 0700–197	105.67875	−19.85612	0.015	0.053	BLL	ISP	14.050	14.050	⋯	0.999	0.937	⋯
J0709.7–0256	PMN J0709–0255	107.43773	−2.92153	0.019	0.039	BLL	LSP	12.830	13.223	1.47200	0.998	0.898	⋯
J0721.4+0404	PMN J0721+0406	110.34963	4.11228	0.041	0.082	FSRQ	LSP	12.700	12.921	0.66500	0.999	0.881	⋯
J0723.2–0728	1RXS J072259.5–073131	110.74895	−7.52649	0.079	0.090	BCU III	⋯	⋯	⋯	⋯	0.976	⋯	0.899
J0725.8–0054	PKS 0723–008	111.46100	−0.91571	0.010	0.044	BCU I	LSP	13.355	13.407	0.12800	1.000	0.967	⋯
J0729.5–3127	NVSS J072922–313128	112.34570	−31.52438	0.078	0.157	BCU II	LSP	13.133	13.133	⋯	0.979	⋯	⋯
J0730.2–1141	PKS 0727–11	112.57964	−11.68683	0.006	0.013	FSRQ	LSP	12.300	12.713	1.58900	1.000	0.989	⋯
J0730.5–0537	TXS 0728–054	112.61849	−5.59636	0.027	0.050	BCU II	HSP	15.200	15.200	⋯	0.997	0.882	⋯
J0744.1–3804	PMN J0743–3804	115.93736	−38.06650	0.080	0.269	BCU III	⋯	⋯	⋯	⋯	0.936	⋯	⋯
J0744.8–4028	PMN J0744–4032	116.15929	−40.53806	0.083	0.078	BCU II	LSP	12.620	12.620	⋯	0.872	⋯	⋯
J0746.6–0706	PMN J0746–0709	116.61456	−7.16379	0.067	0.098	BCU II	ISP	14.230	14.230	⋯	0.983	⋯	⋯
J0747.2–3311	PKS 0745–330	116.83201	−33.17971	0.016	0.033	BCU II	LSP	13.850	13.850	⋯	1.000	0.958	⋯
J0748.0–1639	TXS 0745–165	117.01285	−16.66396	0.009	0.125	BCU II	LSP	11.920	11.920	⋯	0.997	0.915	⋯
J0754.4–1148	TXS 0752–116	118.61024	−11.78804	0.027	0.039	BLL	LSP	13.355	13.355	⋯	1.000	0.953	⋯
J0804.0–3629	NVSS J080405–362919	121.02237	−36.48863	0.008	0.045	BCU II	HSP	15.920	15.920	⋯	0.999	0.852	⋯
J0816.7–2421	PMN J0816–2421	124.16838	−24.35183	0.012	0.073	BCU II	LSP	12.340	12.340	⋯	0.999	0.873	⋯
J0825.8–3217	PKS 0823–321	126.46405	−32.30645	0.023	0.066	BCU II	ISP	14.030	14.030	⋯	0.999	0.914	⋯
J0825.9–2230	PKS 0823–223	126.50655	−22.50756	0.008	0.018	BLL	ISP	14.160	14.441	0.91100	1.000	0.966	0.947
J0828.8–2420	NVSS J082841–241850	127.17383	−24.31403	0.041	0.098	BCU III	⋯	⋯	⋯	⋯	⋯	0.853	⋯
J0841.3–3554	NVSS J084121–355506	130.34017	−35.91823	0.014	0.027	BCU II	HSP	15.956	15.956	⋯	1.000	0.892	⋯
J0845.1–5458	PMN J0845–5458	131.26034	−54.96904	0.021	0.039	BCU II	LSP	13.005	13.005	⋯	1.000	0.981	0.828
J0849.5–2912	NVSS J084922–291149	132.34210	−29.19734	0.043	0.064	BCU II	ISP	14.504	14.504	⋯	0.988	⋯	⋯
J0849.9–3540	PMN J0849–3541	132.44010	−35.68369	0.034	0.052	BCU II	LSP	12.900	12.900	⋯	1.000	0.913	⋯
J0852.6–5756	PMN J0852–5755	133.16136	−57.92495	0.022	0.050	BCU II	LSP	13.076	13.076	⋯	0.999	⋯	0.858
J0853.0–3654	NVSS J085310–365820	133.29384	−36.97236	0.061	0.047	BCU II	HSP	15.660	15.660	⋯	0.883	0.810	⋯
J0858.1–3130	1RXS J085802.6–313043	134.51195	−31.51118	0.029	0.091	BCU II	HSP	16.235	16.235	⋯	⋯	⋯	0.913
J0904.8–3516	NVSS J090442–351423	136.17658	−35.24010	0.053	0.084	BCU II	ISP	14.171	14.171	⋯	0.988	0.864	⋯
J0904.8–5734	PKS 0903–57	136.22158	−57.58494	0.015	0.030	BCU I	ISP	14.664	14.893	0.69500	1.000	1.000	⋯
J0922.8–3959	PKS 0920–39	140.69341	−39.99307	0.017	0.165	BCU II	LSP	13.775	13.775	⋯	0.999	0.948	⋯
J0940.7–6102	MRC 0939–608	145.19733	−61.12455	0.078	0.158	BCU II	LSP	13.671	13.671	⋯	0.984	0.897	⋯
J0956.7–6441	AT20G J095612–643928	149.05075	−64.65781	0.067	0.087	BCU II	LSP	13.285	13.285	⋯	0.928	⋯	⋯
J1005.0–4959	PMN J1006–5018	151.55837	−50.30374	0.370	0.197	BCU II	LSP	12.140	12.140	⋯	⋯	1.000	⋯
J1015.2–4512	PMN J1014–4508	153.70981	−45.14477	0.097	0.101	BCU II	LSP	12.025	12.025	⋯	0.986	0.900	⋯
J1038.9–5311	MRC 1036–529	159.66941	−53.19535	0.040	0.057	BCU II	LSP	12.235	12.235	⋯	0.998	1.000	⋯
J1047.8–6216	PMN J1047–6217	161.92897	−62.28740	0.016	0.044	BCU II	LSP	12.200	12.200	⋯	0.999	1.000	⋯
J1051.5–6517	PKS 1049–650	162.84800	−65.30240	0.017	0.063	BCU II	ISP	14.030	14.030	⋯	0.998	⋯	⋯
J1103.9–5357	PKS 1101–536	165.96759	−53.95019	0.007	0.028	BLL	LSP	13.830	13.830	⋯	0.999	0.984	⋯
J1123.2–6415	AT20G J112319–641735	170.83090	−64.29339	0.034	0.078	BCU III	⋯	⋯	⋯	⋯	0.995	0.931	⋯
J1136.6–6826	PKS 1133–681	174.00874	−68.45162	0.062	0.105	BCU III	⋯	⋯	⋯	⋯	0.987	0.932	⋯
J1229.8–5305	AT20G J122939–530332	187.41637	−53.05894	0.046	0.116	BCU III	⋯	⋯	⋯	⋯	0.991	⋯	⋯
J1233.9–5736	AT20G J123407–573552	188.52933	−57.59803	0.019	0.036	BCU II	ISP	14.700	14.700	⋯	0.998	⋯	⋯
J1256.1–5919	PMN J1256–5919	194.02043	−59.32886	0.013	0.054	BCU III	⋯	⋯	⋯	⋯	0.998	⋯	⋯
J1304.3–5535	PMN J1303–5540	195.95507	−55.67545	0.119	0.132	BCU II	LSP	12.725	12.725	⋯	0.974	0.908	⋯
J1308.1–6707	PKS 1304–668	197.07240	−67.11812	0.012	0.053	BCU II	ISP	14.230	14.230	⋯	0.998	0.973	⋯
J1315.1–5329	PMN J1315–5334	198.76742	−53.57663	0.089	0.069	BCU I	LSP	13.775	13.775	⋯	0.959	0.912	⋯
J1326.6–5256	PMN J1326–5256	201.70512	−52.93990	0.025	0.043	BLL	LSP	12.559	12.559	⋯	0.999	1.000	⋯
J1328.9–5607	PMN J1329–5608	202.25477	−56.13407	0.009	0.022	BCU I	LSP	12.930	12.930	⋯	1.000	0.990	⋯
J1330.1–7002	PKS 1326–697	202.54615	−70.05363	0.008	0.031	BCU II	LSP	13.425	13.425	⋯	1.000	0.977	⋯
J1346.6–6027	Cen B	206.70435	−60.40815	0.052	0.051	RDG	ISP	14.762	14.762	0.01292	1.000	1.000	⋯
J1353.5–6640	1RXS J135341.1–664002	208.41726	−66.66602	0.011	0.037	BLL	HSP	15.700	15.700	⋯	1.000	⋯	0.963
J1400.7–5605	PMN J1400–5605	210.17407	−56.08210	0.009	0.121	BCU II	LSP	12.280	12.280	⋯	0.997	⋯	⋯
J1413.2–6518	Circinus galaxy	213.29172	−65.34571	0.043	0.119	sy	HSP	15.440	15.440	⋯	0.988	⋯	0.886
J1419.1–5156	PMN J1419–5155	214.89685	−51.91627	0.079	0.143	BCU II	LSP	12.550	12.550	⋯	0.993	0.921	⋯
J1424.6–6807	PKS 1420–679	216.23149	−68.13280	0.027	0.059	BCU II	LSP	12.480	12.480	⋯	1.000	1.000	⋯
J1503.7–6426	AT20G J150350–642539	225.95892	−64.42764	0.025	0.046	BCU II	LSP	13.285	13.285	⋯	0.997	⋯	⋯
J1508.7–4956	PMN J1508–4953	227.16227	−49.88398	0.051	0.087	BCU II	LSP	11.780	11.780	⋯	0.999	0.956	⋯
J1514.5–4750	PMN J1514–4748	228.66677	−47.80829	0.032	0.063	FSRQ	LSP	12.515	12.922	1.55120	0.999	0.963	⋯
J1525.2–5905	PMN J1524–5903	231.21301	−59.06103	0.060	0.206	BCU II	LSP	12.655	12.655	⋯	0.986	0.848	⋯
J1558.9–6432	PMN J1558–6432	239.70952	−64.54157	0.012	0.030	BLL	HSP	15.300	15.333	0.07958	1.000	0.977	0.937
J1600.3–5810	MRC 1556–580	240.05157	−58.18416	0.020	0.078	BCU III	⋯	⋯	⋯	⋯	0.998	0.952	⋯
J1603.9–4903	PMN J1603–4904	240.96119	−49.06820	0.012	0.014	BLL	ISP	14.615	14.615	⋯	1.000	0.988	⋯
J1604.4–4442	PMN J1604–4441	241.12925	−44.69221	0.027	0.038	BCU I	LSP	12.947	12.947	⋯	0.999	1.000	⋯
J1610.6–3956	PMN J1610–3958	242.59116	−39.98287	0.061	0.183	FSRQ	LSP	13.088	13.269	0.51800	0.999	0.868	⋯
J1617.4–5846	MRC 1613–586	244.32455	−58.80218	0.041	0.073	FSRQ	LSP	12.550	12.550	1.42200	0.996	1.000	0.844
J1637.6–3449	NVSS J163750–344915	249.46249	−34.82098	0.039	0.042	BCU II	LSP	13.000	13.000	⋯	0.983	0.843	0.879
J1645.2–5747	AT20G J164513–575122	251.30595	−57.85622	0.067	0.109	BCU III	⋯	⋯	⋯	⋯	0.979	⋯	⋯
J1648.5–4829	PMN J1648–4826	252.19968	−48.43856	0.064	0.139	BCU III	⋯	⋯	⋯	⋯	0.993	⋯	⋯
J1650.2–5044	PMN J1650–5044	252.56928	−50.74673	0.004	0.023	BCU I	LSP	12.725	12.725	⋯	1.000	1.000	⋯
J1656.2–3303	Swift J1656.3–3302	254.07025	−33.03633	0.016	0.118	FSRQ	LSP	12.648	13.179	2.40000	1.000	0.883	⋯
J1659.7–3132	NVSS J165949–313047	254.95383	−31.51325	0.036	0.090	BCU II	LSP	13.110	13.110	⋯	0.998	0.858	⋯
J1711.5–5029	PMN J1711–5028	257.92080	−50.47150	0.023	0.079	BCU II	LSP	13.390	13.390	⋯	0.997	⋯	⋯
J1717.4–5157	PMN J1717–5155	259.39455	−51.92553	0.044	0.076	FSRQ	LSP	12.836	13.170	1.15800	0.990	⋯	⋯
J1717.8–3342	TXS 1714–336	259.40012	−33.70245	0.042	0.036	BLL	LSP	12.865	12.865	⋯	1.000	0.922	⋯
J1718.1–3056	PMN J1718–3056	259.52173	−30.93753	0.007	0.062	BCU III	⋯	⋯	⋯	⋯	0.998	0.880	⋯
J1731.8–3001	NVSS J173146–300309	262.94538	−30.05255	0.035	0.035	BLL	⋯	⋯	⋯	⋯	0.994	0.841	⋯
J1741.9–2539	NVSS J174154–253743	265.47687	−25.62872	0.034	0.044	BCU III	⋯	⋯	⋯	⋯	0.994	0.813	⋯
J1744.9–1725	1RXS J174459.5–172640	266.24914	−17.44348	0.011	0.037	BCU III	⋯	⋯	⋯	⋯	⋯	0.837	0.963
J1802.6–3940	PMN J1802–3940	270.67783	−39.66886	0.006	0.017	FSRQ	LSP	12.445	12.810	1.31900	1.000	0.985	⋯
J1823.6–3453	NVSS J182338–345412	275.91079	−34.90334	0.009	0.024	BCU II	HSP	16.140	16.140	⋯	1.000	0.925	0.983
J1828.9–2417	1RXS J182853.8–241746	277.22879	−24.29344	0.015	0.053	BCU I	HSP	16.456	16.456	⋯	⋯	0.872	0.914
J1830.1+0617	TXS 1827+062	277.52475	6.32110	0.032	0.045	FSRQ	LSP	12.305	12.547	0.74500	0.999	0.920	⋯
J1831.0–2714	PMN J1831–2714	277.75019	−27.23505	0.012	0.114	BCU III	⋯	⋯	⋯	⋯	0.994	0.810	⋯
J1833.6–2103	PKS 1830–211	278.41619	−21.06126	0.007	0.014	FSRQ	LSP	12.585	13.130	2.50700	1.000	0.994	0.939
J1835.4+1349	TXS 1833+137	278.89730	13.81354	0.039	0.118	BCU III	⋯	⋯	⋯	⋯	0.991	⋯	⋯
J1844.3+1547	NVSS J184425+154646	281.10567	15.77940	0.022	0.039	BCU II	ISP	14.708	14.708	⋯	0.998	0.867	⋯
J1849.3–1645	1RXS J184919.7–164726	282.33110	−16.78999	0.026	0.053	BCU III	⋯	⋯	⋯	⋯	⋯	⋯	0.907
J1908.8–0130	NVSS J190836–012642	287.15393	−1.44532	0.086	0.074	BCU II	LSP	11.782	11.782	⋯	0.996	⋯	⋯
J1910.8+2855	1RXS J191053.2+285622	287.71764	28.93926	0.012	0.048	BCU II	HSP	16.910	16.910	⋯	⋯	⋯	0.942
J1912.0–0804	PMN J1912–0804	288.02970	−8.07275	0.010	0.077	BCU II	HSP	15.050	15.050	⋯	0.999	0.905	⋯
J1924.9+2817	NVSS J192502+281542	291.25942	28.26172	0.043	0.056	BCU II	HSP	15.850	15.850	⋯	⋯	0.800	0.906
J1925.7+1228	TXS 1923+123	291.42007	12.46058	0.011	0.115	BCU III	⋯	⋯	⋯	⋯	0.992	0.808	⋯
J1931.1+0937	RX J1931.1+0937	292.78819	9.62119	0.010	0.020	BLL	HSP	16.150	16.150	⋯	1.000	0.860	0.980
J1933.4+0727	1RXS J193320.3+072616	293.33459	7.43941	0.030	0.074	BCU II	HSP	15.980	15.980	⋯	0.998	0.823	0.899
J1942.7+1033	1RXS J194246.3+103339	295.69785	10.55753	0.009	0.023	BCU II	HSP	15.435	15.435	⋯	1.000	0.917	0.958
J1949.0+1312	87 GB 194635.4+130713	297.23037	13.24400	0.043	0.051	BCU II	HSP	15.450	15.450	⋯	⋯	0.836	⋯
J1955.1+1357	87 GB 195252.4+135009	298.79821	13.97118	0.017	0.052	FSRQ	LSP	12.865	13.106	0.74300	1.000	0.885	⋯
J2000.1+4212	MG4 J195957+4213	299.99487	42.22965	0.032	0.063	BCU II	LSP	12.550	12.550	⋯	0.997	0.897	⋯
J2001.1+4352	MG4 J200112+4352	300.30364	43.88134	0.005	0.012	BLL	HSP	15.205	15.205	⋯	1.000	0.944	⋯
J2012.0+4629	7C 2010+4619	303.02349	46.48216	0.020	0.026	BLL	ISP	14.958	14.958	⋯	1.000	0.954	0.967
J2015.6+3709	MG2 J201534+3710	303.86971	37.18320	0.038	0.027	FSRQ	LSP	12.743	13.012	0.85900	0.994	0.961	⋯
J2018.5+3851	TXS 2016+386	304.62927	38.85538	0.005	0.048	BCU II	LSP	13.508	13.508	⋯	0.998	0.910	⋯
J2023.2+3154	4C +31.56	305.82924	31.88397	0.028	0.090	BCU I	LSP	13.382	13.514	0.35600	0.998	0.967	⋯
J2025.2+3340	B2 2023+33	306.29518	33.71673	0.053	0.052	BCU I	LSP	12.305	12.391	0.21900	0.999	0.942	⋯
J2029.4+4923	MG4 J202932+4925	307.41614	49.43949	0.055	0.058	BLL	LSP	13.320	13.320	⋯	0.968	⋯	⋯
J2038.8+5113	3C 418	309.65431	51.32018	0.104	0.110	FSRQ	LSP	12.480	12.909	1.68600	0.996	0.962	⋯
J2039.5+5217	1ES 2037+521	309.84799	52.33056	0.043	0.062	BLL	HSP	16.448	16.470	0.05300	1.000	⋯	0.895
J2056.7+4938	RGB J2056+496	314.17808	49.66850	0.027	0.027	BCU II	HSP	15.742	15.742	⋯	0.995	0.894	0.957
J2108.0+3654	TXS 2106+367	317.02275	36.92404	0.018	0.059	BCU II	ISP	14.860	14.860	⋯	⋯	0.837	⋯
J2110.3+3540	B2 2107+35A	317.38283	35.54933	0.203	0.246	BCU II	ISP	14.048	14.048	⋯	0.989	0.862	0.836
J2201.7+5047	NRAO 676	330.43141	50.81566	0.016	0.042	FSRQ	LSP	12.515	12.977	1.89900	1.000	0.955	⋯
J2347.0+5142	1ES 2344+514	356.77015	51.70497	0.005	0.018	BLL	HSP	15.850	15.869	0.04400	1.000	0.953	0.980
J2347.9+5436	NVSS J234753+543627	356.97138	54.60754	0.007	0.066	BCU II	HSP	16.400	16.400	⋯	⋯	⋯	0.925

Note. Columns 1 and 2 are the 3FGL and counterpart names, columns 3 and 4 are the counterpart J2000 coordinates, column 5 gives the angular separation between the gamma-ray and counterpart positions, column 6 is the 95% error radius on the gamma-ray position, column 7 lists the optical class, column 8 is the spectral energy distribution (SED) class (depending on the synchrotron peak frequency given in column 9), column 10 is the synchrotron peak frequency corrected for the redshift shown in column 11, and columns 12–14 report the probability for the Bayesian method and the two reliability values, LR_RG and LR_XG, for the radio–gamma-ray match and the X-ray–gamma-ray match, respectively.

Machine-readable versions of the table is available.

Download table as:  DataTypeset images: 1 2 3 4

The revised 2LAC sample (Ackermann et al. 2015) includes 929 sources, 65 of which are missing in 3LAC (Table 7). Most do not make the TS cut over the 4 year-long period, probably mainly due to variability. On the other hand, 56 unassociated sources in the 2FGL are now associated with blazars, thanks to a more complete set of counterpart catalogs and more precise localizations for the gamma-ray sources (arising from greater statistics and an improved instrument point-spread function). A total of 27 1FGL sources (not necessarily all in 1LAC) that were not listed in 2LAC are now included in 3LAC. Some 51 2LAC sources have changed classifications in 3LAC, mostly due to improved data: 8 AGNs into BCUs, 1 AGN into a BL Lac, 39 BCUs into 34 BL Lacs and 5 FSRQs, one FSRQ into a BL Lac (TXS 0404+075) and two BL Lacs into FSRQs (B2 1040+24A and 4C +15.54).

The 3LAC catalog lists sources detected with high significance during 48 months of observation. Some blazars flare during a limited time only and may be missing in 3LAC. If bright enough, some of them are caught in near-real time by the Fermi Flare Advocate service, also known as Gamma-ray Sky Watcher (FA-GSW), which we briefly describe here.

A day-by-day review of the whole gamma-ray sky, both by a human-in-the-loop and by automated science processing analysis (see, e.g., Chiang et al. 2012), results in the calculation of preliminary source fluxes, tentative localizations, and counterpart associations for any significant source detection. This service serves as an important resource for the scientific community by providing alerts on flaring or transient sources and by producing seeds for follow-up variability and multiwavelength⁸⁰ studies (see, e.g., Ciprini & Thompson 2013).

Since the beginning of the mission, daily reports are compiled internally to the Collaboration, while information and news are communicated via the LAT-MW mailing-list,⁸¹ published in The Astronomer's Telegrams (ATels,)⁸² special GCN notices,⁸³ and weekly summaries in the Fermi Sky Blog.⁸⁴ A total of 201 ATels were posted on behalf of the LAT Collaboration in the 48 month period considered in the 3FGL/3LAC, specifically from 2008 July 24 (the first ATel#1628) to 2012 July 29 (ATel#4285), primarily derived from the FA-GSW service. Some 143 ATels contained alerts and preliminary results about blazars and other AGN targets⁸⁵ referring to 71 different FSRQs, 18 different BL Lac objects, and 9 other AGNs or BCUs detected in flaring, hardening, or enhanced activity states. Only one, PKS 1915−458 (z = 2.47, ATel#2666 and ATel#2679) is not listed in the 3FGL/3LAC or in previous LAT catalogs. This high-redshift FSRQ appears to only emit gamma-rays sporadically within short time intervals.

In addition, three LAT sources announced in ATels and not present in the 3LAC might have extragalactic source associations: Fermi J0052+1110 located at high Galactic latitude), PMN J1626−2426 (FSRQ in the vicinity of 3FGL J1626.2−2428 but outside its error ellipse and located behind an H ii region), and PMN J0623−3350 (flat spectrum radio source reported as Fermi J0623−3351). A fourth LAT ATel source tentatively associated there with the FSRQ PKS 2136−642 is listed as 3FGL J2141.6−6412 in 3FGL but is now associated with the BCU PMN J2141−6411 that is separated $\sim 15^{\prime}$ from the former.

Figure 7 displays the photon index distributions for the different blazar classes both for the sources previously listed in 2LAC and the newly detected sources. The newly detected FSRQs are slightly softer than the 2LAC ones (2.53 ± 0.03 versus 2.41 ± 0.01), indicating that the LAT gradually detects more lower energy-peaked blazars. In contrast, there is no significant spectral difference between the two sets of BL Lacs. For BCUs, the distribution of the new sources extends further out on the high-index end (${\rm{\Gamma }}\gt 2.4$), where the overlap with the BL Lac distribution becomes very small. The corresponding sources seem likely to be FSRQs.

Zoom In Zoom Out Reset image size

Figures 8 and 9 show the photon index versus the photon flux and energy flux, respectively, together with estimated flux limits. As noted in 2LAC, the strong bias observed toward hard sources in the photon-flux limit essentially vanishes when considering the energy-flux limit above 100 MeV instead. (Note that this feature holds only for a lower bound of 100 MeV; other lower energy limits will bring about a dependence of the energy-flux limit on the spectral index.)

Zoom In Zoom Out Reset image size

Figure 10 shows the position of the synchrotron peak ${\nu }_{\mathrm{peak},\mathrm{meas}}^{{\rm{S}}}$ versus the photon spectral index for FSRQs and BL Lacs with measured redshifts. The strong anticorrelation already observed in 1LAC and 2LAC is confirmed. Fitting a linear function ${\rm{\Gamma }}=A+B\mathrm{log}({\nu }_{\mathrm{peak},\mathrm{rest}}^{{\rm{S}}}/{10}^{14}\;\mathrm{Hz})$ yields $A=2.25\pm 0.04$ and $B=-0.18\pm 0.03$. The mean and rms of the Γ distributions are 2.44 ± 0.20, 2.01 ± 0.25, 2.21 ± 0.18, 2.07 ± 0.20, and 1.87 ± 0.20 for FSRQs, the whole BL Lac sample, LSP-, ISP- and HSP-BL Lacs, respectively. FSRQs are overwhelmingly of the LSP class, so no distinction between SED-based classes will be made for them in figures and tallies. Only 37 FSRQs are of the ISP class and only 2 of the HSP class (BZB J0202+0849 and NVSS J025037+171209 associated with 3FGL J0202.3+0851 and 3FGL J0250.6+1713, respectively). As is visible in Figure 10, most ISP-FSRQs have softer spectra than the bulk of ISP-BL Lacs ($\langle {\rm{\Gamma }}\rangle$ = 2.40 ± 0.04 versus 2.07 ± 0.02). In contrast, the two HSP-FSRQs have spectra ($\langle {\rm{\Gamma }}\rangle$ = 2.01) on par with the HSP-BL Lacs and thus much harder than the spectra of most other FSRQs. A similar trend is actually observed for BCUs, as can be seen in Figure 11, where the photon spectral index is plotted versus ${\nu }_{\mathrm{peak},\mathrm{obs}}^{{\rm{S}}}$. In this figure, the orange bars show the average index for different bins in ${\nu }_{\mathrm{peak},\mathrm{rest}}^{{\rm{S}}}$ obtained from the data plotted in Figure 10 for blazars of known types. This comparison supports the idea that BCUs with low ${\nu }_{\mathrm{peak}}^{{\rm{S}}}$ and high Γ are likely FSRQs, while the rest would mostly be BL Lacs.⁸⁶

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Figure 12 compares the redshift distributions for FSRQs and BL Lacs in the 2LAC Clean Sample and those for the new 3LAC Clean-Sample sources (note that 50% of the BL Lacs do not have measured redshifts; see below). The distributions are fairly similar, although the newly detected FSRQs are located at a slightly higher redshift than the 2LAC ones ($\langle z\rangle$ = 1.33 ± 0.08 versus 1.17 ± 0.03). The maximum redshift for an FSRQ is still 3.1 (four FSRQs have $2.94\lt z\lt 3.1$) and has not changed since the 1LAC. This trend allowed the conclusion that the number density of FSRQs grows dramatically up to redshift ≃0.5–2.0 and declines thereafter (Ajello et al. 2012).

Zoom In Zoom Out Reset image size

The redshift distribution of new BL Lacs is somewhat narrower than that of the 2LAC sources, with a maximum near z = 0.3. The redshift distributions gradually spread out to higher redshifts when moving from HSP-BL Lacs to LSP-BL Lacs, a feature already seen in 2LAC. However, the HSP distribution extends to higher redshifts relative to 2LAC, with six HSP-BL Lacs having measured redshifts greater than 1 and one (MG4 J000800+4712) having a redshift greater than 2. Five of these six HSPs were already included in 2LAC but either lacked measured redshifts or were classified differently.

Among BL Lacs, 309 have a measured redshift, while 295 do not. The fraction of BL Lacs without redshift is 55%, 61%, and 40% for LSPs, ISPs, and HSPs, respectively. However, Shaw et al. (2013) have provided redshift constraints for 134 2LAC BL Lacs: upper limits from the absence of Lyα absorption for all of them and lower limits from non-detection of the host galaxy or from intervening absorption line systems for a subset of 54 objects. It was noted by these authors that the average lower limit exceeded the average measured redshift for BL Lacs, indicating that the measured redshifts are biased low. The allowed redshift ranges for the 54 sources with both lower and upper limits are plotted in the bottom panel of Figure 12, confirming that they are in tension with the measured redshift distributions, in particular for HSPs. Kolmogorov–Smirnov tests (K–S) yield probabilities of $2\times {10}^{-2}$, $1\times {10}^{-7}$, and $1\times {10}^{-6}$ that the distributions of measured redshifts and lower limits are drawn from the same underlying population for LSPs, ISPs and HSPs, respectively. The redshift ranges are very similar for the different subclasses and all cluster at high redshifts, with a median around z = 1.2. This is in good agreement with the predictions of Giommi et al. (2013), which posit that most LAT-detected BL Lacs are actually FSRQs with their emission lines swamped by the non-thermal continuum hampering determination of their redshifts.

The gamma-ray luminosity has been computed from the 3FGL energy flux between 100 MeV and 100 GeV, obtained by spectral fitting. Figure 13 displays the gamma-ray luminosity plotted against redshift, together with the sensitivity limits calculated for Γ = 1.8 and 2.2. The Malmquist bias already reported in previous catalog papers is clearly visible. Low-luminosity BL Lacs (<10⁴⁵ erg s⁻¹) cannot be detected at $z\gt 0.4$. Note that sources with a luminosity greater than 5 × 10⁴⁷ erg s⁻¹ (64 are in 3LAC) could still be detected at $z\gt 3.2$.

Zoom In Zoom Out Reset image size

Figure 14 shows the LAT photon index versus the gamma-ray luminosity for the different blazar classes. This correlation has been widely discussed in the context of the "blazar divide" or "blazar sequence" (Ghisellini et al. 2009, 2012; Meyer et al. 2012; Padovani et al. 2012; Finke 2013; Giommi et al. 2013). The features are similar to 2LAC, namely a branch of MAGNs separate from the bulk of blazars and a correlated trend of both luminosity and photon index as ${\nu }_{\mathrm{peak}}^{{\rm{S}}}$ decreases. Figure 15 shows the LAT photon index versus the gamma-ray luminosity for the 57 BL Lacs with both lower and upper limits on their redshifts or only upper limits (134 sources). Because of the bias mentioned above, the HSPs with both limits are more luminous on average than those with measured redshifts, thus populating a previously scarcely occupied area in the ${L}_{\gamma }$–Γ diagram. This observation has profound consequences for the blazar sequence. Note that Ajello et al. (2014) found a small but significant correlation between gamma-ray luminosity and spectral index when including the redshift constraints from Shaw et al. (2013).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

First observed for 3C 454.3 (Abdo et al. 2009b) early in the Fermi mission, a significant curvature in the energy spectra of many bright FSRQs and some bright LSP-/ISP-BL Lacs is now a well-established feature (Abdo et al. 2010f, 2010g). The break energy obtained from a broken power-law fit has been found to be remarkably constant as a function of flux, at least for 3C 454.3 (Abdo et al. 2011). Several explanations have been proposed to account for this feature, including $\gamma \gamma$ attenuation from He ii line photons (Poutanen & Stern 2010), intrinsic electron spectral breaks (Abdo et al. 2009b), Lyα scattering (Ackermann et al. 2010), Klein-Nishina effects taking place when jet electrons scatter BLR radiation in a near-equipartition approach (Cerruti et al. 2013), and hybrid scattering (Finke & Dermer 2010). The level of curvature has been observed to diminish during some flares (e.g., Pacciani et al. 2014).

In the 3FGL analysis, a switch is made from a power-law model to a log-parabola model whenever ${\mathrm{TS}}_{\mathrm{curve}}\gt 16$. The spectrum of the FSRQ 3C 454.3 cannot be well fitted with a log-parabola model, a power-law+exponential cutoff being a better model. A total of 91 FSRQs (57 in 2LAC), 32 BL Lacs (12 in 2LAC) and 8 BCUs show significant curvature at a confidence level >99%. Figure 16 shows the log-parabola β parameter plotted against gamma-ray flux and luminosity. At a given flux or luminosity the spectra of BL Lacs are less curved than those of FSRQs, a feature already reported in 2LAC. Figure 17 compares the TS distributions for sources with curved spectra and those for the whole samples of FSRQs and BL Lacs. All bright FSRQs have curved spectra. For BL Lacs, the situation is more diverse. For BL Lac sources with $\mathrm{TS}\gt 1000$, the fraction of sources with curved spectra is 16/23 (70%) for LSPs, 6/19 (32%) for ISPs, and 5/28 (18%) for HSPs. Note that because the latter have harder spectra than LSPs/ISPs on the average, potential spectral curvature is easier to detect for them. The average β is lower for HSPs (0.05) than for LSPs and ISPs (0.08).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Variability is a key feature of blazars. The 3FGL monthly averaged light curves provide a baseline reference against which other analyses can be cross-checked and enable cross-correlation studies with data obtained at other wavelengths. Although variability at essentially all timescales has been observed in blazars, the monthly binning represents a trade off between a shorter binning needed to resolve flares in bright sources and a longer binning required to detect faint sources. Even so, only 15 sources are detected in all 48 bins with monthly significance $\mathrm{TS}\gt 25$, while this number becomes 46 if a relaxed condition $\mathrm{TS}\gt 4$ is required. The 15 sources include 11 BL Lacs (7 HSPs), only 3 FSRQs (PKS 1510−08, 4C +55.17, B2 1520+31) and the radio galaxy NGC 1275. The 46 sources comprise 28 BL Lacs (14 HSPs), 15 FSRQs, one BCU and two radio galaxies, NGC 1275 and Centaurus A.

We will focus here on the variability index defined in Section 2; a value of 72.44 for this index indicating variability at the 99% confidence level (while the average index for non-variable sources is 47). Recall that this index can be large only for sources that are both variable and relatively bright. This index is plotted versus the synchrotron peak frequency in Figure 18. The features already reported in 2LAC are again visible, with a large fraction of FSRQs found to be variable (69%), with the fraction for BL Lacs much lower on average (23%) and with a steadily decreasing trend as ${\nu }_{\mathrm{peak}}^{{\rm{S}}}$ rises (39%, 23%, 15% for LSPs, ISPs and HSPs respectively). These fractions are quite similar to those reported in 2LAC, despite the larger population and longer time span of the light curves. A similar trend between variability index and ${\nu }_{\mathrm{peak}}^{{\rm{S}}}$ is observed for BCU (Figure 18 bottom), with 21% of them found to be variable.

Zoom In Zoom Out Reset image size

The variability index is plotted versus TS for different bins in the photon spectral index in Figure 19. A distinct trend is visible: for a given TS the mean variability index increases as the spectrum becomes softer (the spectral index increases) up to ${\rm{\Gamma }}=2.4$ where this effect saturates. A net difference between FSRQs and BL Lacs is also apparent, confirming the behavior reported above. For ${\rm{\Gamma }}\gt 2.2$, 72% of FSRQs and 25% of BL Lacs are variable above the 99% confidence level.

Zoom In Zoom Out Reset image size

For each source, we fit the distribution of monthly photon fluxes with a lognormal function

$$\begin{eqnarray}&&{f}_{\mathrm{Ln}}(x)=\displaystyle \frac{{N}_{\mathrm{Ln}}}{x\;{\sigma }_{\mathrm{Ln}}\;\sqrt{2\pi }}\;\mathrm{exp}\left[-\displaystyle \frac{{(\mathrm{log}(x)-{u}_{\mathrm{Ln}})}^{2}}{2{\sigma }_{\mathrm{Ln}}^{2}}\right],\end{eqnarray} \tag{ 1 }$$

treating the flux values returned by the maximum-likelihood algorithm as if they were always significant, for simplicity. The lognormal function has commonly been used to model blazar flux distributions (e.g., Giebels & Degrange 2009; Tluczykont et al. 2010) and provides reasonable fits for most sources of our large sample. This distribution is expected for a process involving a large number of multiplicative, independently varying parameters. Figure 20 compares the distributions of shape parameters ${\sigma }_{\mathrm{Ln}}$ of FSRQs and BL Lacs that have been detected in 48 months above a TS of 1000 and had a monthly TS above 4 in at least 24 monthly periods. These distributions are distinct. The modes are about 0.8 and 0.4 for FSRQs and BL Lacs, respectively, confirming a larger flux variability for the former.

Zoom In Zoom Out Reset image size

To further illustrate the detection variability and how the sample of brightest blazars renews itself, we compare the samples of brightest sources detected during the first and the last three-month periods of the 4 year-long data accumulation time. We applied the same TS cut used to select the LBAS sample (Abdo et al. 2009a), namely $\mathrm{TS}\gt 100$ (simply adding up the monthly TS values). The two samples include similar numbers of sources (128 versus 134), but have only 50% (65) of the sources in common.

A total of 85 3LAC sources are in common with the Swift BAT 70 months survey (Baumgartner et al. 2013) in the 14–195 keV band performed between December 2004 and September 2010 (there were 47 in 2LAC). These 85 sources include 34 FSRQs with an average redshift of 1.37 ± 0.15. Only 9 BAT FSRQs are missing from 3LAC. The average LAT photon index of BAT-detected FSRQs is 2.57, i.e., somewhat softer than the overall average photon index of LAT FSRQs (2.43), a clue that their high-energy hump is located at slightly lower energies than the bulk of the FSRQs. Out of 37 BAT BL Lacs, 30 have now been detected with the LAT. These BL Lacs comprise 3 LSPs, 2 ISPs, and 19 HSPs, while 4 others are still unclassified. The large fraction of HSPs in this sample is not surprising, as the detection of LSPs and ISPs in the hard X-ray band is hampered by their SEDs exhibiting a valley between the low- and high-energy humps in this band (see Böttcher 2007). Figure 21 displays the LAT photon index versus the BAT photon index. Despite large error bars in the BAT photon index and non-simultaneous measurements, a remarkable anticorrelation (Pearson correlation factor −0.69), already noted in 2LAC, is observed. For the HSP-BL Lacs considered here, BAT probes the high-frequency (falling) part of the $\nu {F}_{\nu }$ synchrotron peak while the LAT probes the rising side of the inverse-Compton peak (assuming a leptonic scenario). For FSRQs, which are all LSPs in the common sample, BAT and LAT probe the rising and falling sides of the inverse-Compton peak, respectively.

Zoom In Zoom Out Reset image size

It is also worth noting that 96 3LAC sources (5 Radio Galaxies, 53 FSRQs, 33 BL Lacs, 4 BCUs, 1 NLSy1) are present in the V38 INTEGRAL source catalog⁸⁷ (based on 3–200 keV data taken since 2002), which includes 540 AGNs located at $| b| \gt 10^\circ$.

At the time of this writing, 56 AGNs that have been detected at TeV energies are listed in TeVCat.⁸⁸ Among them, 55 are present in 3FGL (see Table 10), which is a remarkable result underscoring the level of synergy that has now been achieved between the high-energy and VHE domains. Only HESS J1943+213 (a HSP BL Lac located at b = $-1\buildrel{\circ}\over{.} 3$, affecting the possible LAT detection) is still missing from the 3FGL, but an analysis of five years of LAT data resulted in a $\gt 1$ GeV detection (Peter et al. 2014). There are 15 newly detected sources relative to 2FGL and six relative to the first Fermi-LAT catalog of sources above 10 GeV (1FHL, Ackermann et al. 2013, based on 3 years of data): SHBL J001355.9−185406, 1ES 0229+200, 1ES 0347−121, RX J0847.1+1133 (aka RBS 0723), MS 1221.8+2452, and 1 H 1720+117.