163 Acta Polytechnica CTU Proceedings 1(1): 163–169, 2014 163 doi: 10.14311/APP.2014.01.0163 Modelling the Multifrequency SED of AGN Candidates among the Unidentified EGRET and Fermi Gamma-Ray Sources Pieter J. Meintjes1, Pheneas Nkundabakura2,1, Brian van Soelen1, Alida Odendaal1 1Department of Physics, University of the Free State, P.O. Box 339, Bloemfontein, 9300, South-Africa 2Kigali Institute of Education, P.O. Box 5039, Kigali, Rwanda Corresponding author: MeintjPJ@ufs.ac.za Abstract Of the 271 sources in the 3rd EGRET catalogue, 131 were reported as unidentified, i.e. not associated with any particular class of point source in the sky. Since the largest fraction of the EGRET sources were extragalactic, a sample of 13 extragalactic unidentified sources have been selected for multi-wavelength follow-up studies. Five of the selected EGRET sources coincide with gamma-ray flux enhancements seen in the Fermi-LAT data after one year of operation. In this article, we report the multi-wavelength properties of, among others, the 5 sources detected by Fermi-LAT from our sample of high galactic latitude unidentified EGRET sources. Recent spectroscopic observations with the Southern African Large Telescope (SALT) confirmed one of the unidentified EGRET sources as a possible Seyfert 2 galaxy, or alternatively, a narrow line radio galaxy. The detected gamma-ray emission (Eγ > 30 MeV) of the 5 coinciding EGRET/Fermi-LAT sources are fitted with external Compton and Synchrotron Self Compton (SSC) models to investigate the energetics required to produce the EGRET/Fermi gamma-ray flux. In all the models the inclination angle of the jet with respect to the observer is θjet ≈ 60◦, between those of Seyfert 1 and Seyfert 2/radio galaxies. These results confirm the possibility of Seyfert and radio galaxies sources are constituting a new class of γ-ray source in the energy range Eγ > 30 MeV. Keywords: radiation mechanisms: non-thermal - line: identification - techniques: spectroscopic - galaxies: jets - BL Lacertae objects. 1 Introduction The Energetic Gamma Ray Telescope Experiment EGRET (30 MeV - 10 GeV) provided the highest gamma-ray window on board the Compton Gamma- Ray Observatory (CGRO). EGRET detected 271 gamma-ray sources above 100 MeV, 92 % of which were blazars. Of the 271 sources detected, 131 re- mained unidentified, i.e. could not be associated with any specific point source of gamma-ray emission (Hart- man et al., 1999). The large number of unidentified EGRET sources above and below the galactic plane in- spired a search for possible extra-galactic radio loud Active Galactic Nuclei (AGN) counterparts that could possibly be associated with these unidentified sources. To avoid confusion with possible galactic sources, espe- cially molecular cloud distributions, the search for coun- terparts was restricted to those unidentified sources at galactic latitudes | b | > 10◦. Source Selection Criteria: The candidate coun- terparts should be inside the error box associated with the EGRET detection (Hartman et al., 1999), con- firmed as extragalactic in the NASA Extragalactic Database (NED), possess a radio brightness above 100 mJy at 8.4 GHz (Sowards-Emmerd, Romani & Michel- son, 2003), exhibit hard spectra with spectral indices | α | < 0.7 (Sowards-Emmerd, Romani & Michelson, 2003) and display variability (e.g. Fan, 2005) that may be associated with an inner accretion disc or jet. Based upon these criteria, 13 sources have been selected for further follow-up study (Meintjes & Nkundababura, 2012; Nkundabakura & Meintjes, 2012). Of the 13 candidate sources selected, 5 have con- firmed gamma-ray excesses in the Fermi-LAT cata- logue containing the first year’s observations (Abdo et al., 2010a). The Fermi Large Area Telescope (Fermi- LAT) is a pair conversion γ-ray telescope sensitive to photon energies between 20 MeV and 300 GeV. Launched on 11 June 2008, Fermi-LAT started to collect data in August 2008. The data are made available on a daily basis and can be accessed on- line at the official website of the Fermi Science Sup- port Centre (http://fermi.gsfc.nasa.gov/cgi-bin/ssc /LAT/LATDataQuery.cgi). The results of de- tections made within the first 24 months of operation were released in June 2011, in the form of the Second Fermi-LAT catalog 163 http://dx.doi.org/10.14311/APP.2014.01.0163 Pieter J. Meintjes et al. (2FGL; see Nolan et al., 2012) 1. Five sources from our previous sample of unidenti- fied EGRET sources have been detected with Fermi- LAT, namely 2FGL J0727.0-4726, 2FGL J1304.3-4353, 2FGL J1703.2-6217, 2FGL J1709.0-0821 and 2FGL J1815.6-6407, which are counterparts of 3EG J0724- 4713, 3EG J1300-4406, 3EG J1659-6251, 3EG J1709- 0828 and 3EG J1813-6419 respectively. The Fermi-LAT sources that coincided with the EGRET sources are pre- sented as circled crosses in Fig. 1. The EGRET (30 MeV-10 GeV) gamma-ray spec- tra of our chosen sample of sources that were observed between April 1991 - October 1995 (cycles 1, 2 3 and 4 of the mission) have been determined. The photon spectral index distribution is displayed in Fig. 2. The spectral distribution of these unidentified sources cor- responds remarkably well with the gamma-ray blazar photon spectral index distribution observed by Fermi- LAT (Abdo et al., 2010a). b=−90° b=+90° l=0° b=+10° b=−10° b=0°,l=+360° b=+30° b=+60° b=−30° b=−60° Figure 1: Galactic distribution of the unidentified EGRET sources (circles) and their Fermi-LAT coun- terparts (crosses). 1 1.5 2 2.5 3 3.5 4 3E G J0159-3603 3E G J0500+ 2529 3E G J0702-6212 3E G J0706-3837 3E G J0724-4713 3E G J0821-5814 3E G J1300-4406 3E G J1659-6251 3E G J1709-0828 3E G J1800-0146 3E G J1813-6419 3E G J1822+ 1641 3E G J1824+ 3441 γ- ra y ph ot on i nd ex HSP-BLLacs L and ISP-BLLacs FSRQs γ=2.03 γ=2.38 Figure 2: Gamma-ray photon spectral indices of the EGRET sources between 30 MeV - 10 GeV. The gamma-ray spectra of the sample of EGRET sources with Fermi-LAT counterparts were determined. The spectra are presented in Fig. 3. Noticeable is the apparent change in the spectral index between the EGRET and the Fermi-LAT gamma-ray data, which may point to a transition in the gamma-ray production process. -100 0 100 10 100 1000 10000 100000R es id ua ls ( % ) E (MeV) 1e-15 1e-14 1e-13 E 2 ( dN /d E )( W .m -2 ) 3EG J0724-4713 dN(E)=const E-2.6dE a.) 3EG J0724-4713 -100 0 100 10 100 1000 10000 100000 1e+06R es id ua ls ( % ) E (MeV) 1e-15 1e-14 1e-13 E 2 ( dN /d E )( W .m -2 ) 3EG J1300-4606 dN(E)=const E-3.07dE c.) 3EG J1300-4406 -100 0 100 10 100 1000 10000 100000R es id ua ls ( % ) E (MeV) 1e-15 1e-14 1e-13 E 2 ( dN /d E )( W .m -2 ) 3EG J1659-6251 dN(E)=const E-2.83dE 1e-15 1e-14 1e-13 E 2 ( dN /d E )( W .m -2 ) 3EG J1659-6251 dN(E)=const E-2.83dE b.) 3EG J1659-6251 -100 0 100 10 100 1000 10000 100000R es id ua ls ( % ) E (MeV) 1e-15 1e-14 1e-13 E 2 ( dN /d E )( W .m -2 ) 3EG J1709-0828 dN(E)=const E-3dE d.) 3EG J1709-0828 -100 0 100 10 100 1000 10000 100000R es id ua ls ( % ) E (MeV) 1e-15 1e-14 1e-13 E 2 ( dN /d E )( W .m -2 ) 3EG J1813-6419 dN(E)=const E-2.85dE e.) 3EG J1813-6419 Figure 3: The gamma-ray spectra of the EGRET (30 MeV - 10 GeV) and coinciding Fermi-LAT (20 MeV - 300 GeV) sources. 2 Gamma-Ray Variability The multi-wavelength emission of accretion driven sources like AGN is characterized by very high lumi- nosity, assuming isotropic emission, and variability over several time scales (e.g. Fan, 2005). This is recon- ciled with the fact that the bulk of the Spectral En- ergy Distribution (SED) of these sources is produced in a non-homogeneous and variable jet. Gamma-ray flux 1http://fermi.gsfc.nasa.gov/ssc/data/access/lat/2yr catalog/ 164 Modelling the Multifrequency SED of AGN Candidates among the Unidentified EGRET... variability has also been confirmed from AGN-blazars (Mattox et al., 1997). Aperture photometry of the Fermi-LAT data over a time span of 4.6 years (from August 2008 to March 2013) has been performed to investigate possible long term variability. Photons in the energy interval 100 MeV to 200 GeV were counted in an area with a ra- dius of 1 degree centered on the source and monthly averages were determined. Two of the sources, 2FGL J1304.3-4353 and 2FGL J1703.2-6217, showed definite signs of variability, quantified in terms of a variability index V (Nolan et al., 2012), which implies that for V > 41 there is a < 1% probability of the source be- ing steady. Although variability was detected in 2FGL J1304.3-4353 (V = 47) and 2FGL J1703.2-6217 (V = 167) respectively, no periodicity could be detected (see Fig. 4). 1e-07 1.5e-07 2e-07 2.5e-07 3e-07 3.5e-07 54500 55000 55500 56000 56500 P ht on f lu x (p ho to ns .c m -2 .s -1 ) Time (MJD) 2FGL J1304.3-4353 (3EG J1300-4406) Average a.) 2FGL J1304.3-4353 1e-07 1.5e-07 2e-07 2.5e-07 3e-07 3.5e-07 4e-07 4.5e-07 5e-07 5.5e-07 54500 55000 55500 56000 56500 P ho to n fl ux ( ph ot on s. cm -2 .s -1 ) Time (MJD) 2FGL J1703.2-6217(3EG J1659-6251) Average b.) 2FGL J1703.2-6217 Figure 4: The gamma-ray lightcurves of the two Fermi-LAT sources, 2FGL J1304.3-4353 and 2FGL J1703.2-6217, which displayed variability over a period of 4.6 years. The vertical error bars indicate the 68% confidence level. 3 Optical Follow-Up Studies The optical counterparts of two unidentified EGRET sources have already been identified (Meintjes & Nkundabakura, 2012; Nkundabakura & Meintjes, 2012). The spectrum of PKS J0820-5705 (3EG J0821- 5814) resembles that of a FSRQ at redshift z = 0.06, while the spectrum of PMN J0710-3850 (3EG J0706-3837) shows broad and narrow lines resembling the spectrum of a LINER or Seyfert I galaxy at redshift z = 0.129. What distinguishes the spec- trum of PKS J0820-5705 from that of a normal ra- dio galaxy is the shallow K4000 depression of only 8.8 ± 2.5 %, indicating substantial non-thermal activity, while the corresponding value for PMN J0710-3850 is 80 ± 1 % (Meintjes & Nkundabakura, 2012; Nkund- abakura & Meintjes, 2012), in agreement with the value expected for a LINER-Seyfert 1 galaxy (e.g. Caccianiga et al., 1999). Spectroscopic observations of other sources from our sample have been performed with the Southern African Large Telescope (SALT) (see Fig. 5), equipped with the Robert Stobie Spectrograph (RSS), during 2012, in or- der to determine the redshift and to identify the class of AGN. The spectrum of one of the unidentified EGRET sources, 3EG J 0159-3603, could be determined (see Fig. 6), showing distinct narrow emission lines of O ii, O iii and He ii redshifted by z = 0.35. The spectrum resem- bles that of a typical Seyfert 2 galaxy, or alternatively, a narrow line radio galaxy. This implies the possible as- sociation of two Seyfert galaxies with the unidentified sources 3EG J0706-3837 and 3EG J 0159-3603 respec- tively. Figure 5: The Southern African Large Telescope (SALT). −5e−16 0 5e−16 1e−15 1.5e−15 2e−15 4500 5000 5500 6000 6500 7000 7500 F lu x (e rg s − 1 cm − 2 Å − 1 ) Wavelength (Å) [N eI II ]+ H ε [O II ] [N eI II ] [O II I] H γ PMN J0156−3616 (z=0.35) [O II I] * * * * H eI I Figure 6: SALT spectrum of the optical counterpart of the unidentified EGRET source 3EG J0159-3603, show- ing emission lines redshifted by z=0.35 4 The Spectral Energy Distributions (SED) The spectral energy distributions (SEDs) of all the counterparts corresponding to the unidentified EGRET 165 Pieter J. Meintjes et al. sources have been pieced together through multi- wavelength archival data as well as multi-wavelength observations from radio to optical, using the Harte- beesthoek Radio Astronomy Observatory (HartRAO) 26 m telescope, as well as various optical telescopes at the South African Astronomical Observatory (SAAO) in South Africa. Infrared data were obtained from the Two Micron All Sky Survey (2MASS) conducted be- tween June 1997 and February 2001. The SEDs of the sample of EGRET sources and the associated Fermi- LAT counterparts are presented in Fig. 7. 1e-18 1e-17 1e-16 1e-15 1e-14 1e-13 1e-12 1e-11 1e+10 1e+15 1e+20 1e+25 νF ν (W m -2 ) ν (Hz) Radio Opt. FERMI EGRET 3EG J0724-4713 a.) 3EG J0724-4713 1e-18 1e-17 1e-16 1e-15 1e-14 1e-13 1e-12 1e-11 1e+10 1e+15 1e+20 1e+25 νF ν (W m -2 ) ν (Hz) Radio Opt. ROSAT EGRET 3EG J1300-4406 FERMI c.) 3EG J1300-4406 1e-18 1e-17 1e-16 1e-15 1e-14 1e-13 1e-12 1e-11 1e+10 1e+15 1e+20 1e+25 νF ν (W m -2 ) ν (Hz) Radio Opt. NIR EGRET 3EG J1659-6251 FERMI IRAS b.) 3EG J1659-6251 1e-18 1e-17 1e-16 1e-15 1e-14 1e-13 1e-12 1e-11 1e+10 1e+15 1e+20 1e+25 νF ν (W m -2 ) ν (Hz) Radio Opt. EGRET 3EG J1709-0828 FERMI d.) 3EG J1709-0828 1e-18 1e-17 1e-16 1e-15 1e-14 1e-13 1e-12 1e-11 1e+10 1e+15 1e+20 1e+25 νF ν (W m -2 ) ν (Hz) Radio Opt. NIR EGRET 3EG J1813-6419 FERMI e.) 3EG J1813-6419 Figure 7: The multifrequency SEDs for the sample of EGRET sources. The association of some of the unidentified EGRET sources with Seyfert galaxies/LINERS poses a possibil- ity of non-aligned AGN and radio galaxies constituting a new class of gamma-ray sources. Before discussing the modeling of the SEDs presented above, a brief discus- sion of the gamma-ray properties of some radio galaxies are presented to illustrate that these sources possess the required energetics to produce measurable γ-ray emis- sion in the EGRET and Fermi-LAT energy domain 5 Non-Aligned AGN: A New Class of Gamma-Ray Source At High Energies (HE; Eγ > 100 MeV) Fermi-LAT re- ported about ten misaligned radio galaxies (Abdo et al., 2010b,c; Rieger, 2012a). At Very High Energies (VHE; Eγ > 100 GeV), four radio galaxies have been detected, Cen A (d ≈ 3.8 Mpc), M87 (d ≈ 16.7 Mpc) and the Perseus Cluster (d ≈ 77 Mpc; z ≈ 0.018) sources NGC 1275 and IC 310 (e.g. Rieger, 2012a). Cen A was the only non-blazar detected at MeV to GeV energies by CGRO (see Steinle, 2010 for a review). The HE emis- sion from Cen A reported by Fermi-LAT seems to come from both the extended radio lobes and the core region (Abdo et al., 2010b,c). The detected HE gamma-rays from the extended lobe regions suggests that particle ac- celeration up to VHEs occurs in the disrupted jet region (Bordas, 2012). The reported TeV emission from Cen A (Aharonian et al., 2009) provided further evidence of a very effective particle accelerator in Cen A. The spectrum up to 5 TeV is consistent with a power-law with a photon index ∼ 2.7 ± 0.5, with no apparent variability. The HE (Eγ > 0.2 GeV) emission in Cen A is explained in terms of inverse Compton (IC) up- scattering of the Cosmic Microwave Background (CMB) photons (Eph = 8 × 10−4 eV) and infrared (IR) back- ground photons by relativistic electrons with Lorentz factors γe = 6 × 105, in a jet with bulk flow veloci- ties between βΓ ∼ 0.1-0.5. Another peculiar aspect of Cen A is a rather low, sub-Eddington inferred accre- tion rate, ṁ ∼ 10−3ṁEdd, resulting in a rather low bolometric luminosity Lb ∼ 1043 erg s−1 (Whysong & Antonucci, 2004). The inferred equipartition magnetic field in the radio lobes is B ≈ 9 µG (Rieger, 2012b), with the equipartition magnetic field near the black hole (BH) ranging between B ∼ 103 − 104 G (Rieger, 2012b). The detection of non-aligned AGN in the HE and VHE regime poses interesting theoretical challenges regarding particle acceleration and associated gamma- ray emission in the jets of AGN. For example, Cen A shows that the radio lobes of radio galaxies may pos- sess the required energetics to accelerate electrons to VHEs producing the HE gamma-rays through IC up- scattering the CMB photons, even though the bulk flow Lorentz factor of the jet is fairly low. The detection of TeV gamma-rays is explained in terms of γe ∼ 107 electrons up-scattering disc photons to the TeV domain (Rieger & Aharonian, 2009). The nuclear SED of Cen A, based on non-simultaneous data, shows two peaks, one around ∼ 1013 Hz and another around 0.1 MeV (e.g. Chiaberge et al., 2001; Abdo et al., 2010c). The SED below a few GeV seems to be satisfactorily explained by a one-zone Synchrotron self-Compton (SSC) model (Chiaberge et al., 2001) but the same approach fail to account for the TeV emission observed by H.E.S.S. (Abdo et al., 2010c). 166 Modelling the Multifrequency SED of AGN Candidates among the Unidentified EGRET... The discussion presented above underlines the fact that normal, non-aligned AGN do possess the required en- ergetics to accelerate leptons to VHE energies. The production of sub-GeV gamma-rays through a SSC pro- cess in the nuclear region, combined with the HE emis- sion in the radio lobes and TeV emission in the inner disc region close to the black hole (BH), presents a new paradigm in particle acceleration and gamma-ray pro- duction in AGN. With this in mind, the SED of the Fermi-LAT counterparts of the unidentified EGRET sources were modelled. This is a first attempt to ex- plain the SED of these sources within the framework of a Synchrotron self-Compton (SSC) and External Comp- ton (EC) model. These sources are all treated as mis- aligned AGN, with a fixed inclination of θjet ≈ 60◦ be- tween jet and observer. 6 SED Modelling The multi-wavelength data from radio to gamma-rays have been combined to create the SED over more than 15 decades in energy (Fig. 7). The EGRET and Fermi- LAT gamma-ray data are evaluated with the theoretical framework of a single zone Synchrotron self-Compton (SSC) (e.g. Katarczynski, Sol & Kus, 2001) and Exter- nal Compton (EC) model (e.g. Moderski et al., 2003; Sikora, Begelman & Rees, 1994), i.e. where relativistic jet electrons up-scatter infrared (IR) photons from the disc torus and optical photons from the broad emis- sion line (BEL) regions to high energies. The model parameters for the SSC and EC models are presented in Table 1 (SSC) and Table 2 (EC). The corresponding SED model fits are presented in Fig. 8. Table 1: Parameters related to SSC models for the respective sources. Source 3EG J0724-4713 3EG J1659-6251 3EG J1709-0828 3EG J1813-6419 Parameter Units 1 radius (m) 1.5E+14 1.0E+14 1.0E+14 1.0E+14 2 B (T) 7.0E-04 7.5E-04 2.5E-04 2.5E-04 3 p1 -2 -2 -2 -2 4 p2 -3 -2.6 -2.6 -2.6 5 γmax 1.6E+03 6.2E+02 1.6E+03 2.0E+03 Table 2: Parameters related to the EC models for the respective sources. Source R θobs γbr Ke νBEL αBEL νIR (1) (2) (3) (4) (5) (6) (7) 3EG J0724-4713 1.0E+17 1 2.0E+03 4.0E+55 10 -1 0.1 3EG J1300-4406 (BEL) 1.0E+17 1 2.0E+04 6.0E+53 1 -1 0.1 3EG J1300-4406 (IR) 1.0E+17 1 3.0E+03 4.0E+54 1 -1 0.1 3EG J1659-6251 (BEL) 1.0E+17 1 2.0E+03 6.0E+54 10 -1 0.1 3EG J1659-6251 (IR) 1.0E+17 1 2.0E+03 3.0E+55 10 -0.5 0.1 3EG J1709-0828 1.0E+17 1 2.0E+03 3.0E+55 10 -1 0.1 3EG J1800-0146 2.5E+16 1 5.0E+02 3.0E+55 1 -1 0.1 3EG J1813-6419 1.0E+17 1 2.4E+03 1.3E+55 10 -1 0.1 (1): Radius of the emitting region (in cm), (2): Viewing angle (in radians), (3): Electron Lorentz factor at spectral break, (4): Electron normalization constant, (5): Characteristic frequency of BEL (in eV), (6): BEL distribution photon index, (7): Radiation characteristic frequency of IR (in eV). 167 Pieter J. Meintjes et al. 1e-18 1e-17 1e-16 1e-15 1e-14 1e-13 1e-12 1e-11 1e+10 1e+15 1e+20 1e+25 νF ν (W m -2 ) ν (Hz) Radio Opt. FERMI ROSAT EGRET 3EG J0724-4713 EC (IR) SCC Sync SCC EC (BEL)IC (tot) a.) 3EG J0724-4713 1e-18 1e-17 1e-16 1e-15 1e-14 1e-13 1e-12 1e-11 1e+10 1e+15 1e+20 1e+25 νF ν (W m -2 ) ν (Hz) Radio Opt. ROSAT EGRET 3EG J1300-4406 FERMI Sync SCC IC (tot) c.) 3EG J1300-4406 1e-18 1e-17 1e-16 1e-15 1e-14 1e-13 1e-12 1e-11 1e+10 1e+15 1e+20 1e+25 νF ν (W m -2 ) ν (Hz) Radio Opt. NIR EGRET 3EG J1659-6251 FERMI IRAS EC (IR) SYNC SSC EC (BEL) b.) 3EG J1659-625 1e-18 1e-17 1e-16 1e-15 1e-14 1e-13 1e-12 1e-11 1e+10 1e+15 1e+20 1e+25 νF ν (W m -2 ) ν (Hz) Radio Opt. EGRET 3EG J1709-0828 FERMI EC (IR) SSC Sync EC (BEL) IC (tot) d.) 3EG J1709-0828 1e-18 1e-17 1e-16 1e-15 1e-14 1e-13 1e-12 1e-11 1e+10 1e+15 1e+20 1e+25 νF ν (W m -2 ) ν (Hz) Radio Opt. NIR EGRET 3EG J1813-6419 FERMI EC (IR) SSC Sync IC (tot) EC (BEL) e.) 3EG J1813-6419 Figure 8: The modeled SEDs for the various EGRET/Fermi-LAT sources. From these results (see Fig. 8) it can be seen that the EGRET and Fermi-LAT data are consistent with an EC model, i.e. the IC upscattering of IR photons from the disc and UV/optical photons from the line emitting clouds (Broad Line Regions) to the EGRET and Fermi-LAT energy domain. In all the models the electron Lorentz factors were γe ∼ few × 103, rather modest in comparison with the Lorentz factors re- quired to explain the γ-ray emission from, for example, Cen A. 7 Conclusions We report the discovery of 13 flat spectrum extragalac- tic sources within the error boxes of some high galactic latitude, unidentified EGRET sources. Five of these EGRET sources have been detected with Fermi-LAT within the first 11 months of operation. In all cases the EGRET and Fermi-LAT gamma-ray emission could be successfully explained in terms of the IC upscattering of BEL photons, as well as IR photons from the disc, to EGRET and Fermi-LAT energies. The adopted elec- tron energy is of the order of γe ∼ few × 103, which is rather moderate compared to the electron energies of γe ∼ 106 − 107, inferred from the HE and TeV emis- sion from Cen A. These preliminary results definitely confirm that Seyfert and radio galaxies could be associ- ated with a significant fraction of the still unidentified extragalactic EGRET and Fermi-LAT sources. Acknowledgement The authors thank the organisers for the invitation to present this work at this conference. Some of the ob- servations reported in this paper were performed by the Southern African Large Telescope (SALT). The SALT observations were performed under the Proposal: 2012- 1-RSA-003 (PI: B van Soelen). The authors thank Dr. Petri Vaisanen from the SALT science team for his guid- ance with the data reduction. References [1] Abdo, A.A., et al.(FERMI): 2010a, ApJ, 710, 1271 doi:10.1088/0004-637X/710/2/1271 [2] Abdo, A.A., et al. (FERMI): 2010b, Science, 328, 725 doi:10.1126/science.1184656 [3] Abdo, A.A., et al. (FERMI): 2010c, ApJ, 719, 1433 doi:10.1088/0004-637X/719/2/1433 [4] Aharonian, F.A., et al. (H.E.S.S.): 2009, ApJ, 695, L40 doi:10.1088/0004-637X/695/1/L40 [5] Bordas, P.: 2012 in Multifrequency Behaviour of High Energy Cosmic Sources, Vulcano (May 23- 28, 2011), MEMORIE S.A.It, Vol 83 (1), 87 (eds. F. Giovanelli & L. Sabau-Graziati) [6] Caccianiga, A., della Ceca, R., Gioi, I.M., Maccac- aro, T. & Wolter, T.: 1999, ApJ, 513, 51 [7] Chiaberge, M. et al.: 2001, MNRAS, 324, 33 doi:10.1046/j.1365-8711.2001.04642.x [8] Fan, J-H.: 2005, Chin. J. Astron. Astrophys. Suppl., 5, 213 doi:10.1088/1009-9271/5/S1/213 [9] Hartman, R., et al.: 1999, APJS, 123, 79 [10] Katarzynski, K., Sol, H. & Kus, A.: 2001, A&A, 367, 809 [11] Mattox, J., et al.: 1997, ApJ, 476, 692 [12] Meintjes, P.J. & Nkundabakura, P.: 2012, in High Energy Gamma-Ray Astronomy, Proc of the 5th International High Energy Gamma-Ray Astron- omy meeting, Heidelberg (Germany), AIP Con- ference proceedings, Vol. 1505, p. 554 (eds. F.A. Aharonian, W. Hoffman & F. Rieger), Publisher, Melville (New York) 168 http://dx.doi.org/10.1088/0004-637X/710/2/1271 http://dx.doi.org/10.1126/science.1184656 http://dx.doi.org/10.1088/0004-637X/719/2/1433 http://dx.doi.org/10.1088/0004-637X/695/1/L40 http://dx.doi.org/10.1046/j.1365-8711.2001.04642.x http://dx.doi.org/10.1088/1009-9271/5/S1/213 Modelling the Multifrequency SED of AGN Candidates among the Unidentified EGRET... [13] Moderski, R., et al.: 2003, A&A, 406, 855 [14] Nkundabakura, P. & Meintjes, P.J.: 2012, MNRAS,427, 859 doi:10.1111/j.1365-2966.2012.21953.x [15] Nolan, P.L., et al. (FERMI): 2011, ApJ Suppl., 199, 31 doi:10.1088/0067-0049/199/2/31 [16] Rieger, F. & Aharonian, F.A.: 2009, A&A, 506, L41 [17] Rieger, F.: 2012a, in High Energy Gamma-Ray Astronomy, Proc of the 5th International High Energy Gamma-Ray Astronomy meeting, Heidel- berg (Germany), AIP Conference proceedings, Vol. 1505, p. 80 (eds. F.A. Aharonian, W. Hoffman & F. Rieger), Publisher, Melville (New York) [18] Rieger, F.: 2012b, in Multifrequency Behaviour of High Energy Cosmic Sources, Vulcano (May 23-28, 2011), MEMORIE S.A.It, Vol 83 (1), 127 (eds. F. Giovanelli & L. Sabau-Graziati) [19] Sikora, M., Begelman, M.C. & Rees, M.J.: 1994, ApJ, 421, 153 [20] Sowards-Emmerd, D., Romani, R.W. & Michelson, P.F.: 2003, ApJ, 590, 109 [21] Steinle, H.: 2010, PASA, 27, 431 [22] Whysong, D. & Antonucci, R.: 2004, ApJ, 602, 116 169 http://dx.doi.org/10.1111/j.1365-2966.2012.21953.x http://dx.doi.org/10.1088/0067-0049/199/2/31 Introduction Gamma-Ray Variability Optical Follow-Up Studies The Spectral Energy Distributions (SED) Non-Aligned AGN: A New Class of Gamma-Ray Source SED Modelling Conclusions