Multi-band properties of superluminal AGN detected by Fermi/LAT
T.G. Arshakian, J. León-Tavares, J. Torrealba, V.H. Chavushyan
aa r X i v : . [ a s t r o - ph . C O ] J un Fermi meets Jansky – AGN in Radio and Gamma-RaysSavolainen, T., Ros, E., Porcas, R.W., & Zensus, J.A. (eds.)June 21–23, 2010, Bonn, Germany
Multi-band properties of superluminal AGN detected byFermi/LAT
T.G. Arshakian ⋆ , J. Le´on-Tavares , J. Torrealba , and V.H. Chavushyan Max-Planck-Institut f¨ur Radioastronomie, Auf dem H¨ugel 69, 53121 Bonn, Germanye-mail: [email protected] Aalto University Mets¨ahovi Radio Observatory, Mets¨ahovintie 114, FIN-02540, Kylm¨al¨a, Finlande-mail: [email protected] Instituto de Astronom´ıa, Universidad Nacional Aut´onoma de M´exico, Apartado Postal 70-264, 04510 M´exicoD.F., M´exico, e-mail: [email protected] Instituto Nacional de Astrof´ısica ´Optica y Electr´onica, Apartado Postal 51 y 216, 72000 Puebla, Pue, M´exicoe-mail: [email protected]
Abstract.
We perform a multi-band statistical analysis of core-dominated superluminal active galactic nuclei (AGN)detected with
Fermi
Large Area Telescope (LAT). The detection rate of γ -ray jets is found to be high for opticallybright AGN. There is a significant correlation between the γ -ray luminosity and the optical nuclear and radio(15 GHz) luminosities of AGN. We report a well defined positive correlation between the γ -ray luminosity and theradio-loudness for quasars and BL Lacertae type objects (BL Lacs). The slope of the best-fit line is significantlydifferent for quasars and BL Lacs. The relations between the optical and radio luminosities and the γ -ray loudnessare also examined, showing a different behavior for the populations of quasars and BL Lacs. Statistical resultssuggest that the γ -ray, optical and radio emission is generated at different locations and velocity regimes along theparsec-scale jet.
1. Introduction
Extreme optical variability was a clear signature for someblazars detected at γ -ray energies during the EGRET era(Fitchel et al. 1994). Using the Fermi /LAT improved sen-sitivity, recent multiwavelength variability studies on indi-vidual sources have confirmed the tight connection amongthe γ -ray and optical variable emission (e.g. 3C 273 inAbdo et al. 2010a). So far there was no attempt to com-pare optical continuum properties among γ -ray blazars. Inthis manuscript we take advantage of the unprecedentedsensitivity provided by the first bright source catalogue(1FGL; Abdo et al. 2010b) to study the multi-band cor-relations between optical, radio, and γ -ray emission in su-perluminal AGN detected by Fermi /LAT during its first11 months of operation. We adopt a flat cosmology with H = 71 km s − Mpc − , Ω m = 0 .
27, and Ω Λ = 0 .
2. The sample of AGN detected with Fermi/LAT
The MOJAVE-2 sample consists of about 290 core-dominated AGN from the MOJAVE (Monitoring of Jetsin AGN with VLBA Experiments) program (Lister et al.2009). The sample of AGN studied in this manuscript in-cludes 250 core-dominated AGN (Kovalev et al. 2005)from the MOJAVE-2 sample. Most of these AGN arecurrently monitored with the Very Large Baseline Array(VLBA) at 15 GHz. The 1FGL catalogue includes about190 sources from the MOJAVE-2 sample. Optical nuclear ⋆ Speaker −5 −4 −3 −2Log S opt (Jy)01020304050 N u m b e r Fig. 1.
Distributions of optical nuclear fluxes for AGNfrom the MOJAVE sample (full line), 102 AGN de-tected as the gamma-ray source (filled histogram) andnon-detected with
Fermi /LAT during the 1FGL period(shaded histogram).fluxes and redshift measurements were available for a sam-ple of 102 MOJAVE sources (Arshakian et al. 2010a)identified by
Fermi /LAT (hereafter, M-1FGL sample).Seventy six out of 102 sources in the M-1FGL sample arepart of the statistically complete MOJAVE-1 sample. Wewill refer to these 76 sources as the M1-1FGL sample.The Kolmogorov-Smirnov (K-S) test shows that the dis-1
T.G. Arshakian et al.: Multi-band properties of superluminal AGN detected by
Fermi /LAT
42 44 46log L opt (erg/s)40424446 l og L V L B A ( e r g / s ) GalaxyBLLacQuasar
Fig. 2.
Radio luminosity at 15 GHz against optical nuclearluminosity at 5100 ˚A.tribution of redshifts in the MOJAVE-1 sample and theM1-1FGL sample are drawn from the same parent popu-lation indicating that the M1-1FGL sample is not biasedby redshift.The M-1FGL sample consists of 76 quasars, 24BL Lacs, and two radio galaxies which are excludedfrom further statistical tests. We use the non-parametricKendall’s τ test to analyze correlations between indepen-dent variables and the partial Kendall’s τ test to accountfor the common dependence on redshift in the correla-tions between luminosities. Throughout the paper we as-sume a correlation to be significant if a chance probability P < .
3. Interplay between γ -ray, optical, and radioproperties of superluminal AGN The distribution of optical nuclear fluxes for the MOJAVEAGN detected and non-detected with
Fermi /LAT is pre-sented in Figure 1. At first glance, the
Fermi /LAT de-tected sources seem to have higher optical fluxes thanthose with no-detection. This is further supported by theK-S statistical test: there is a significant difference (ata confidence level of 99 . Fermi /LAT. When using the M1-1FGL sample, the dif-ference is significant at a confidence level of 96 %. We con-clude that the detection rate of γ -ray AGN is high forsources having high optical nuclear fluxes.We derive the optical nuclear luminosities ( L opt ), to-tal VLBA luminosities ( L VLBA ), and the rest-frame radio-loudness ( R ∝ S VLBA /S opt ) using the fluxes given inArshakian et al. (2010a). γ -ray luminosities were com-puted using the Eq. (1) in Ghisellini et al. (2009), where S γ ( ν , ν ) is the energy flux between 0.1 GeV and 100 GeVfrom the 1FGL catalogue. Note that the γ -ray, optical, andradio luminosities are estimated from non-simultaneousobservations. log L γ (erg/s) l og L op t ( e r g s − ) GalaxyBLLacQuasar
44 46 48 50log L γ (erg s −1 )40424446 l og L V L B A ( e r g s − ) Fig. 3.
Optical nuclear luminosity against γ -ray luminos-ity (top panel), and radio (15 GHz) luminosity against γ -ray luminosity (bottom panel) for AGN from the M-1FGLsample. Labels in the top panel denote the correspondingpopulation. Optical–Radio emission.
Arshakian et al. (2010a)found a positive correlation in the L opt − L VLBA rela-tion plane for AGN from the MOJAVE-1 sample. Theyconcluded that the correlation is due to the populationof quasars and that the optical emission is non-thermaland generated in the parsec-scale jet. This is supported bystudies of individual radio galaxies, 3C 390.3 and 3C 120(Arshakian et al. 2010b and Le´on-Tavares et al. 2010) forwhich the link between optical continuum variability andkinematics of the parsec-scale jet was found. It was in-terpreted in terms of optical continuum flares generatedat subparsec-scales in the innermost part of a relativis-tic jet rather than in the accretion disk. We confirm the L opt − L VLBA positive correlation for a larger sample ofM-1FGL quasars as well as no-correlation for BL Lacs (seeFigure 2 and Tables 1 and 2). .G. Arshakian et al.: Multi-band properties of superluminal AGN detected by
Fermi /LAT 3 l og L γ ( e r g / s ) GalaxyBLLacQuasar
Fig. 4. γ -ray luminosity ( L γ ) versus radio-loudness ( R )for M-1FGL AGN. The solid line represents the ordinaryleast-square fit to the data. Gamma-ray–Optical emission.
We find a positive cor-relation between L γ and L opt (Figure 3). The correlationis significant for quasars from the M-1FGL and M1-1FGLsamples and it is stronger for M1-1FGL quasars (Tables 1and 2), suggesting for a single production mechanism for γ -ray and optical nuclear emission. Gamma-ray–Radio emission.
Kovalev et al. (2009) re-ported a significant correlation between γ -ray and radioVLBA (8 GHz) emission for sample of ∼
30 AGN. This cor-relation holds at high confidence level ( > . γ -ray emission leads the radio emis-sion of the parsec-scale jet at 15 GHz with time delay offew months. They interpreted the observed time lag asa result of synchrotron opacity in the jet: the radio and γ -ray emission are generated in the same region (pertur-bation in the jet?) and become observable with some timedelay due to the opacity effects. If this scenario is correctthen the variable optical emission is also generated in theperturbation moving upstream the jet, and we should ex-pect that the optical emission leads the radio emission anddelays with respect to the γ -ray emission.The correlation in the L VLBA − L γ relation plane isapparently stronger than that in the L opt − L γ (Figure 3).One may think that the corrected optical nuclear emis-sion of some AGN (Arshakian et al. 2010a) might be con-taminated by contribution of a stellar component thuscausing the large dispersion in the L opt − L γ diagram.Contamination should be stronger in radio galaxies andweaker in quasars and BL Lacs for which the contributionof optical nuclear emission is dominant. The large disper-sion in the L opt − L γ relation plane can be due to non- simultaneous optical/ γ -ray observations and stronger vari-ability in the optical regime than that in the radio, and/orwider range of Doppler factors in the optical regime com-pared to the range of Doppler factors of the jet at 15 GHz,if the bulk of optical emission is generated in the relativis-tic jet and it is Doppler boosted.We report a significant positive correlation ( >
99 %)between L γ and radio-loudness for quasars and BL Lacs(see Figure 4 and Table 1). The solid line in Figure 5represents the best fit to the data,log L γ = (0 . ± .
11) log R + (42 . ± . . (1)The L γ ∝ R relation suggests that the strong γ -ray jetshave progressively high Doppler factors (or faster speeds)in the radio domain compared to those in the opticalregime. For quasars, the regression line is fitted bylog L γ = (0 . ± .
14) log R + (45 . ± . , (2)while for BL Lacs the best fit is,log L γ = (0 . ± .
21) log R + (42 . ± . . (3)It is noticeable that the slope derived for quasars is shal-lower than the slope fitted for BL Lacs. The L γ − R cor-relation is still significant for all AGN ( >
99 %), quasarsand BL Lacs (99 %) of the M1-1FGL sample (Table 2).The best-fit parameters for the later sample are almostunchanged.
Table 1.
Kendall’s τ correlation analysis between emis-sion characteristics of AGN from the M-1FGL sample.A1 and A2 are the independent variables for which theKendalls τ correlation analysis is performed, τ is the cor-relation coefficient, and P is the probability of a chancecorrelation. The correlations are considered to be signif-icant if the chance probability P < .
05 (or confidencelevel >
95 %).
All Quasars BL LacsA1 A2 τ P τ P τ PL
VLBI L opt L γ L opt L γ L VLBA L γ R L opt S γ /S VLBA -0.3 1e-5 0.2 4e-2 0.5 1e-3
Table 2.
Kendall’s τ correlation analysis between emis-sion characteristics of AGN from the M1-1FGL sample. All Quasars BL LacsA1 A2 τ P τ P τ PL
VLBA L opt L γ L opt L γ L VLBA L γ R L opt S γ /S VLBA -0.3 1e-4 0.2 4e-2 0.5 2e-3
Gamma-ray loudness.
We define the γ -optical loud-ness and γ -radio loudness as the S γ /S opt and S γ /S VLBA , T.G. Arshakian et al.: Multi-band properties of superluminal AGN detected by
Fermi /LAT γ /S opt l og L V L B A ( e r g s − ) GalaxyBLLacQuasar γ /S VLBA l og L op t ( e r g s − ) Fig. 5.
Radio VLBA luminosity against γ -optical loudness (left panel), and optical luminosity against γ -radio loudness(right-panel) for the M-1FGL quasars and BL Lacs. The solid line represents the ordinary least-square fit to the data.respectively. We find that the radio luminosity is inde-pendent of γ -optical loudness (Figure 5, left panel). Onthe other hand, there is a negative correlation in the L opt − S γ /S VLBA relation plane (Figure 5, right panel),which is much stronger for BL Lacs (confidence level > ∼ . L opt ∝ ( S γ /S VLBA ) − . .We interpret this relation as the optically weaker jets tohave higher Doppler factors (or Lorentz factors if outflowsradiating in optical and γ -ray bands have the same view-ing angle) in the gamma domain compared to those in theradio domain. The slope of the best-fit line is steeper forBL Lacs ( − . ± .
21) than for quasars ( − . ± .
1) andthe difference is statistically significant. We suggest that,on the average, the ratio between the Doppler factors ingamma and radio regimes varies slower for quasars thanfor BL Lacs, indicating for different velocity regimes intheir jets.The significant differences found between quasars andBL Lacs are supported in recent studies (e.g. Ghisellini etal. 2009, Sambruna et al. 2010, Tornikoski et al., in thisproceedings) suggesting the presence of different physicalconditions along the jet in quasars and BL Lacs. Thisand other insights into the relationship between the γ -ray, radio, and optical emission will be pursued in a moredetail in further studies.
4. Summary
Using the sample of 100 superluminal quasars and BL Lacsdetected by
Fermi /LAT, we investigate relations betweentheir optical, radio, and γ -ray emission available from non-simultaneous observations. Our main results are summa-rized as follows: – The detection rate of γ -ray superluminal AGN is highfor optically bright AGN. – The known positive correlation between L opt and L VLBA holds for the M-1FGL quasars at a confi-dence level of 98 %. The known correlation between γ -ray flux and radio flux density (measured quasi-simultaneously) is also valid for non-simultaneous mea-surements of L γ and L VLBA for quasars. The L γ − L VLBA correlation is significantly stronger than thatin the L γ − L opt relation plane. – There is a correlation between L opt and L γ which ex-clusively holds for quasars. The correlation is signifi-cant for the M-1FGL quasars (at a confidence level of99 %) and marginally significant for quasars from theM1-1FGL sample (c.l. 95 %). – We report a statistically significant positive correla-tion (c.l. >
99 %) between γ -ray luminosity and radio-loudness for both, quasars and BL Lacs. The slope ofthe L γ − R relation is found to be steeper for the pop-ulation of BL Lacs. – We find that the radio luminosity at 15 GHz is in-dependent of the γ -radio loudness ( S γ /S VLBA ) forquasars and BL Lacs. The γ -optical loudness ( S γ /S opt )and optical nuclear luminosity are negatively corre-lated for quasars (c.l. 96 %) and BL Lacs (c.l. 99 . Acknowledgements.
TGA acknowledges support by DFG-SPPproject under grant 566960. This work was supported byCONACYT research grant 54480 (Mexico).
References
Abdo, A. A. et al. 2010a, Nature, 463, 919Abdo, A. A. et al. 2010b, ApJS, arXiv:1002.2280Arshakian, T. G. et al. 2010a, accepted to A&A,arXiv:1003.2577Arshakian, T. G., et al. 2010b, MNRAS, 401, 1231Fitchel, C. E. et al. 1994, ApJS, 94, 551Ghisellini, G., et al. 2009, MNRAS, 396, L105.G. Arshakian et al.: Multi-band properties of superluminal AGN detected by