ON THE GAMMA-RAY EMISSION FROM THE CORE AND RADIO LOBES OF THE RADIO GALAXY CENTAURUS A ∗

NAREK SAHAKYAN†,¶, FRANK M. RIEGER‡ and FELIX AHARONIAN‡,§ †ICRANet-Yerevan, Marshall Baghramian Avenue, 24, Yerevan 0019, Republic of Armenia and ICRANet, Piazza della Repubblica 10, I-65122 Pescara, Italy ‡Max-Planck-Institut für Kernphysik, P.O. Box 103980, 69029 Heidelberg, Germany §Dublin Institute for Advanced Studies, 31 Fitzwilliam Place, Dublin 2, Ireland ¶narek@icra.it RUIZHI YANG‖ and EMMA DE ONA-WILHELMI∗∗ ‖Key Laboratory of Dark Matter and Space Astronomy, CAS, Nanjing, 210008, China ∗∗Institut de Cincies de l’Espai (IEEC-CSIC), Barcelona 08193, Spain


Introduction
The prominent radio galaxy Centaurus A (NGC 5128), at a distance of 3.8 Mpc, 1 is our closest active galaxy.Often considered as a prototype Fanaroff-Riley Class I 2 radio source and as a misaligned BL Lac-type object at higher energies, 3,4 it is one of the best-studied extragalactic objects over a wide range of frequencies. 5Observations at radio frequencies show giant structures (so-called 'lobes') with a total angular size of ∼ 10 • , which corresponds to a physical extension of ∼ 600 kpc (d/3.8Mpc).
The Fermi-LAT collaboration has early on reported the detection of HE γ-rays from both the core (i.e., within ∼ 0.1 • ) and the giant radio lobes of Cen A: 6,7 An analysis of the available ten-month LAT data set revealed a point-like HE emission region coincident with the position of the radio core of Cen A, and two large extended emission regions detected with a significance of 5 and 8σ for the northern and the southern lobe, respectively.On the other hand, the HE emission from the extended regions seemed to be morphologically correlated with the giant radio lobes, contributing more than 50 % to the total HE source emission.These regions appeared spectrally well described by a power-law function extending up to 2 or 3 GeV with photon indices of Γ ∼ 2.6.
Here we report on our results obtained using a much larger Fermi-LAT data set, allowing for a detailed investigation of the spectrum and morphology of the 'lobes' and for an improved analysis of the spectral and temporal characteristics of HE core emission.

Fermi-LAT Data Analysis
Here we present the main observational results of a Fermi-LAT analysis of HE core and lobe emission in Cen A. Detailed information on the analysis can be found in Refs.8, 9.

The core of Cen A
The continuum gamma-ray emission of the core of Cen A has first been modeled with a single power law and a binned gtlike analysis been performed.The best-fit power-law parameters for the core of Cen A turned out to be . ( The test statistic is T S = 1978 above 100 MeV, corresponding to a ≈ 44 σ detection significance.Figure 1  that the measured spectrum exhibits clear deviations from a single power-law model with respect to the data above several GeV (χ 2 fit of the power-law model to the data gives a relatively poor fit with χ 2 = 39.7 for 9 dof).In order to analyze this in more detail, the core spectrum has been modeled with a broken power-law model and gtlike tool retried.The best-fit broken power-law parameters are now with Γ 1 = 2.74±0.02and Γ 2 = 2.12±0.14below and above E b = (4.00±0.09)GeV, respectively.The power-law and the broken-power-law models can be compared using a log likelihood ratio test.The test statistic is twice the difference in these loglikelihoods, which gives 9 for this case.Note that the probability distribution of the test statistic can be approximated by a χ 2 distribution with 2 dof, corresponding to different degrees of freedom between the two functions.In this case P (χ 2 ) = 0.011, which again indicates a deviation from a simple power-law function.The results presented above indicate an interesting hardening of the (average) gamma-ray core spectrum towards higher energies.The break in the spectrum at 4 GeV could most naturally be explained by a superposition of different spectral components.If one divides the data set into two energy ranges, i.e., [0.1-4] GeV and [4-100] GeV, modelling each energy range separately with a power-law function, the analysis gives a photon index of Γ 1 = 2.74±0.02and a flux F γ = (1.68±0.04)×10−7 photon cm −2 s −1 for the [0.1-4]GeV interval (test statistics gives TS=1944) and Γ 2 = 2.09 ± 0.2, and F γ = (4.20 ± 0.64) × 10 −10 photon cm −2 s −1 for the [4-100]  GeV range (TS value 124.4), respectively.These components are depicted with a blue and red line in Fig. 1(a).

The lobes of Cen A
For the analysis of Fermi-LAT data of the lobes events with energies between 200 MeV and 30 GeV were selected.Figure 1(b) shows the residual image obtained after subtracting point like sources (including the core of Cen A) and the diffuse background.In order to evaluate the total (extended) HE γ-ray emission in more detail, a template based on the residual map (T1) has been used.The TS values for the south and the north lobe in this template are 411 and 155, respectively.Green contours in Fig. 1(b) show the radio lobes (WMAP, 22 GHz), with the higherfrequency 22 GHz map being used as it better represents the GeV-emitting particles.
A closer inspections reveals that the south lobe of the HE γ-ray image is roughly similar to the south lobe of the radio one, whereas the HE emission in the north seems to extend beyond the radio lobe emission region Fig. 1(b).A morphological analysis in fact indicates some incongruity between the morphology of the radio lobe and γ-ray lobe in the north.This is particularly instructive as HE gammarays directly trace the underlying spatial distribution of energetic electrons (via IC processes) without the degeneracy involved in the radio synchrotron part (i.e., uncertainties in magnetic field topology).Using the template generated with the residual map (T1), the total flux and photon index of the north and the south lobe is derived for the energy range from 100 MeV to 30 GeV.For the north lobe, the integral HE flux is (0.93 ± 0.09) × 10 −7 ph cm −2 s −1 and the photon index is 2.24 ± 0.08, while for the south lobe we find (1.4 ± 0.2) × 10 −7 ph cm −2 s −1 and 2.57 ± 0.07, respectively.

Discussion
The core of Cen A: The analysis of the 4 yr-data set interestingly reveals that the HE core spectrum of Cen A shows an unusual break, with photon index changing from 2.7 to 2.1 at an energy of E b 4 GeV.While spectral breaks are often associated with a situation where the spectrum gets softer, here the opposite occurs.For the component below 4 GeV, the detected photon flux F γ = (1.68 ± 0.04) × 10 −7 photon cm −2 s −1 corresponds to an apparent (isotropic) γ-ray luminosity of L γ (0.1 − 4 GeV) 10 41 erg s −1 (for the distance 3.8 Mpc).On the other hand, the HE luminosity of the component above 4 GeV translates into L γ (> 4GeV) 1.4 × 10 40 erg s −1 .This is an order of magnitude less when compared with the first component, but still larger than the VHE luminosity reported by H.E.S.S. L γ (> 250 GeV) = 2.6 × 10 39 erg s −1 . 10In fact, the spectral hardening seen at GeV energies may allow to account for the (non-simultaneous) H.E.S.S. flux data.
The limited angular resolution (∼ 5 kpc) and the lack of significant variability introduces substantial uncertainties as to the production site of the HE gammaray emission.In principle, the hard HE component could originate from both a very compact (sub-pc) and/or extended (multi-kpc) region(s).The double-peaked nuclear SED of Cen A has in the past been reasonably well-modeled with simple one zone SSC models up to a few GeV. 4,7In this framework, the observed break at 4 GeV would indeed mark the appearance of a physically different component.This additional component could in principle be related to a number of different (not mutually exclusive) scenarios, from magnetosphere VHE emission 11 up to inverse Compton processes in the kpc-scale jet, 12 see Ref. 9 for a full discussion.At the current stage, none of these models can be easily discarded.Definite progress concerning the true origin of the HE gamma-ray could be achieved, however, in case of a significant detection of gamma-ray time variability.

The lobes of Cen A:
The HE γ-ray emission observed from the lobes could in principle be related to leptonic (inverse-Compton scattering) or hadronic (e.g., ppinteraction) processes.In the following, we briefly summarize possible constraints for the underlying radiation mechanism imposed by the observed SEDs (for detailed information see Ref. 8).In a leptonic scenario, both the HE γ-ray and the radio emission could be satisfactorily accounted for.Assuming an electron injection spectrum with an exponential cut-off, the emergent electron distribution, found by solving the kinetic equation describing the energetic and temporal evolution of the radiating electrons, can been used to model the SED through synchrotron and inverse-Compton emission.
The duration of particle acceleration (associated with this the age of the giant lobe emission) is unknown.Dynamical arguments suggest a lower limit > 10 7 yr for the giant radio lobes, while synchrotron spectral ageing arguments indicate an age < ∼ 3 × 10 7 yr. 5,13,14In the following, we therefore discuss the SED implications for an epoch time t between 10 7 yr and 10 8 yr.As it turns out, the modeling of the GeV data provides support for a maximum lobe age of ∼ 8 × 10 7 yr.
Figure 2 shows a SED representation for an epoch time t max = 8 × 10 7 yr (for the SED for t = 10 7 yr see Ref. 8), with a maximum electron Lorentz factor γ max = 2.5 × 10 6 and 1.5 × 10 6 for the north lobe and the south lobe, respectively.Note that in this case the contribution by inverse-Compton scattering of CMB photons alone would be sufficient to account for the observed HE spectrum (see the solid line in Fig. 2).On the other hand, for an epoch time t exceeding t max = 8×10 7 yr, the high-energy part of the SED would no longer be consistent with the data (see the dashed line in Fig. 2 for t = 10 8 yr).These considerations provide additional support for a finite age < 10 8 yr of the lobes.The maximum total energy of electrons in both lobes is found to be ≈ 6 × 10 57 erg, with the total energy in particles and fields comparable to the 10 7 yr-case, thus requiring only a relatively modest mean kinetic jet power of ∼ 10 43 erg/s.In principle, a hadronic origin is conceivable as well: Once protons are efficiently injected, they are likely to remain energetic since the cooling time for pp-interactions is t pp ≈ 10 15 (n/1 cm −3 ) −1 s.High-energy protons interacting with the ambient low-density plasma can then produce neutral pions which decays into two γ-rays.Similar to before, we employed a power-law proton distribution with an exponential cut-off to model the SED.Furthermore, a thermal plasma density n = 10 −4 cm −3 for the giant radio lobes of Cen A has been assumed.This model representation is depicted by a dotted line in Fig. 2. In both lobes, the power-law index of the proton population is α = 2.1, and the high-energy cut-off is E max 55 GeV.The maximum total energy W p is proportional to the gas number density n, so that W p 10 61 (n/10 −4 cm −3 ) −1 erg, obtained here, should be considered as an upper limit.

Fig. 1 .
Fig. 1.(a) Average high-energy gamma-ray (>100 MeV) spectrum of the core of Cen A. The power-law function from the Eq.(1) is depicted with dashed black line.(b) Cen A 'lobes': Excess map after subtraction of diffuse background, point-like sources and Cen A core.

Fig. 2 .
Fig.2.SED for t = 8 × 10 7 yr and for a mean magnetic field value B for the south lobe and the γ-ray excess region in the north lobe of 0.91 µG and 1.17 µG, respectively.The dot-dashed line in (a) represents the IC contribution due to EBL upscattering while the dashed line shows the result for t = 10 8 yr.The possible γ-ray flux expected from pp-interactions given a thermal gas density n = 10 −4 cm −3 is also shown (dotted line).1460182-5Int.J. Mod.Phys.Conf.Ser.2014.28.Downloaded from www.worldscientific.comby 149.217.1.5 on 04/23/14.For personal use only.
γ-Ray Emission from the Core and Radio Lobes of the Radio Galaxy Centaurus A