6.1 Compton Observations of AGN
Before the launch of Compton, only four AGNs have been detected at gamma-ray energies: 3C 273, Cen A, NGC 4151 and MCG 8-11-11 (see, e.g., Gehrels & Cheung 1991; Bassani et al. 1985). 3C 273 was the only quasar that has been detected above 10 MeV. It was during one of the early Compton pointings which included 3C 273 in the field of view that another quasar, 3C 279, was first detected by the EGRET instrument from 30 MeV to over 5 GeV (Hartman et al. 1992). The photon flux intensity of 3C 279 was very high, comparable to that of the Crab or Geminga above 100 MeV. The differential photon spectrum is well represented by a single power law of photon index ~ 2 over the entire observed energy range. The lack of detection by COS-B in 1976 implies that 3C 279 is highly variable. Later observations by EGRET confirmed that 3C 279 was initially detected in a flaring state and it is variable on the time scale of days with no significant change in spectral index (Kniffen et al. 1993). Subsequently, many more AGNs have been detected by EGRET. Table 2 lists the AGNs discovered as of November 1992 together with their characteristics (see e.g., Fichtel 1993, Bertsch et al. 1993 and references therein). All the AGNs detected in high-energy gamma rays are core-dominated, flat spectrum radio quasars or radio-loud BL Lac objects (or "blazars"). Blazars generally exhibit strong variability, significant optical polarization and superluminal flow. The rapid variability indicates that the radiation is likely to be beamed in a relativistic jet, which also produces the radio emission. The BL Lac object Mrk 421 has also been detected in TeV gamma rays by the Whipple Observatory ground-based Cerenkov telescope (Punch et al. 1992). The data are consistent with a single power law of photon index ~ 2 from the 100 MeV to the TeV energy range, though it must be kept in mind that the Whipple observations were not simultaneous with the EGRET observations (Lin et al. 1992). Note that no Seyfert galaxies have been detected by EGRET in the 100 MeV range.
|Super||Radio||Flat||Pol.||Position||(10-6 cm-2 sr-1)||Spectral||Relative|
|Name||OVV||BL Lac||Lum.||Loud||Spect. 1||> 3%||Diff. 2||Uncert. 3||(E > 100 MeV)||Index||Lum. 4|
|0202+149 (4C+15.05)||0.3°||0.4°||0.3 ± 0.1|
|0208-512 PKS||0.13°||0.13°||0.4 to 0.9||-1.7± 0.1||1.00||2|
|0235+164 (OD+160)||0.10°||0.3°||0.8 ± 0.1||-2.0 ± 0.2||0.94||2.0|
|0420-014 (OA 129)||0.5°||0.4°||0.4 ± 0.1||0.92||0.4|
|0454-463 PKS||0.27°||0.38°||0.25 ± 0.1||0.86||0.3|
|0528+134 PKS||0.13°||0.15°||0.4 to 1.6||-2.4 ± 0.1||2.06||4 to 13|
|0537-441 PKS||0.4°||0.6°||0.3 ± 0.1||-2.0 ± 0.2||0.894||0.2|
|0716+714||?||0.47°||0.4°||0.20 ± 0.06||-1.8 ± 0.2|
|0836+714 (4C+71.07)||0.58°||0.50°||0.15 ± 0.04||2.17||1.1|
|1101+384 (Mrk 421)||0.3°||0.4°||0.14 ± 0.03||-1.9 ± 0.1||0.031||0.0002|
|1226+023 (3C273)||0.2°||0.5°||0.30 ± 0.05||-2.4 ± 0.1||0.158||0.008|
|1253-055 (3C279)||0.083°||0.08°||0.6 to 4.9||-2.0 ± 0.1||0.54||0.3 to 2|
|1406-076||0.2°||0.4°||1.0 ± 0.2||1.49||2|
|1606+106 (4C+10.45)||0.42°||0.50°||0.5 ± 0.2||1.23||1.6|
|1633+382 (4C+38.41)||0.08°||0.15°||0.4 to 1.4||-2.0 ± 0.1||1.81||3 to 11|
|2052-474||0.35°||0.5°||0.3 ± 0.1||1.489||0.6|
|2230+114 (CTA 102)||?||0.3°||0.4°||0.24 ± 0.07||-2.4 ± 0.07||1.037||0.5|
|2251+158 (3C454.3)||0.25°||0.22°||0.8 ± 0.1||-2.0 ± 0.1||0.859||0.5|
|1. Flat spectrum radio sources: r > 0.5 (2-5 GHz band)|
2. Difference between gamma-ray determined position and known position of identified source. Most are preliminary.
3. There is a 68% probability that the source is within a circle of this radius. Most are preliminary.
4. The source luminosity (> 100 MeV) in f x 1048 erg/s, with f = beaming factor.
In the low-energy range of 50 keV to 10 MeV, the OSSE instrument has observed 21 AGNs and reported ten firm detections and four marginal ones (Cameron et al. 1993). Of these, nine are Seyfert 1 galaxies, with a detection rate close to 100%. Upper limits have been obtained for two Seyfert 2 galaxies (NGC 1068 and NGC 4593). The spectral indices for energies > 60 keV range from 1.6 to 3.2, with an average of 2.6. The spectrum of NGC 4151 obtained in July 1991 by OSSE (Fig. 8) is steeper than most previous observations (Maisak et al. 1993). Such a steep spectrum has been observed once before by the GRANAT instruments SIGMA and ART-P (Jourdain et al. 1992; Apal'kov et al. 1992). The data can be fitted by a broken power law or a thermal Comptonization spectrum (Sunyaev & Titarchuk 1980). There is evidence that the spectrum steepens at times of source brightening (Perola et al. 1986; Yaqoob et al. 1989). It is possible that Seyfert galaxies in general have variable spectral states correlated with luminosity (Paciesas et al. 1993). These results together with observations from Ginga (e.g., Williams et al. 1992) and ROSAT (e.g., Turner et al. 1993) at lower energies indicate that AGNs have more complex spectral characteristics than can be represented by a simple "canonical" power-law spectrum. For example, Ginga observations have revealed an additional flat component in the > 10 keV portion of Seyfert 1 spectrum (e.g., Pounds et al. 1990; Matsuoka et al. 1990). This has been attributed to reflection or Comptonization of the incident power-law spectrum by an accretion disk.
Figure 8. The OSSE spectrum of NGC 4151 compared to selected previous observations. The solid curve is the best fit to OSSE data of a Sunyaev- Titarchuk thermal Comptonization model (Maisack et al. 1993).
A picture is emerging that the gamma-ray emitting AGNs can be divided into two classes: (1) the radio-quiet Seyferts, in which we observe gamma rays emitted from accretion disks; and (2) the radio-loud blazers, in which we observe gamma rays emitted from the relativistic jets and which are highly variable.
6.2 Implications for GRB
The Compton observations firmly identify AGNs as gamma-ray emitters that can contribute to the GRB. Several groups have proposed schemes to explain the low-energy GRB spectrum based on the "reprocessed" AGN model (Zycki et al. 1993; Zdziarski et al. 1993; Madau et al. 1993; Rogers & Field 1991; Terasawa 1991; Mereghetti 1990). The intrinsic AGN spectrum can be generated either through thermal Comptonization of low-energy photons by hot plasma or by non-thermal pair cascade process in a compact plasma (Svensson 1987). Both mechanisms produce a spectral break at ~ 80 keV. Zdziarski et al. (1993) used a Seyfert 1 thermal Comptonization model and the X-ray luminosity function derived by Boyle et al. (1993). They integrated the contributions over redshift and obtained a good fit to the GRB spectrum from 2 to 100 keV, including the 30 keV bump (Fig. 9). The dominant contribution comes from the AGNs at the highest redshift. However, the model fails to explain the MeV bump as the Seyfert spectrum cuts off above 80 keV. Zycki et al. (1993) examined the contribution by non-thermal pair models with different compactness. They found that a good fit to the MeV bump can be obtained by a population of unobserved low-compactness AGNs (Fig. 10), though they cannot explain the entire 30 to 100 MeV GRB radiation with the model.
Figure 9. Contribution to the GRB by thermal AGNs. The solid and dot-dashed curves are Comptonization models without and with absorption. The dashed curve gives the thermal component, and the dotted curve the reflection component (Zdziarski et al. 1993).
Figure 10. Contribution to the GRB spectrum by a population of low-compactness non-thermal AGNs (Zycki et al. 1993).
Padovani et al. (1993), Dermer and Schlickeiser (1992), and Stecker et al. (1993) estimated the contribution of blazars to the high-energy GRB. Since the gamma-ray luminosity function of blazars has not yet been determined, they used the radio luminosity evolutionary function by Dunlop and Peacock (1990) instead. This is based on a significant (possibly linear) correlation between the 100 MeV gamma-ray luminosity and the 5-GHz radio luminosity seen in the gamma-ray blazars (Fig. 11). The computed background falls an order of magnitude below that observed by SAS-2 in the 100 MeV range (Stecker et al. 1993) and so fails to be a major contributor to the GRB in this energy range. Moreover, the average blazer spectrum is a power law of photon index ~ 2, which is much harder than the observed GRB (Fig. 7). However, if the hard spectrum continues unattenuated to higher energies, blazars could be important contributors to the GRB in the multi-GeV energy range.
Figure 11. Gamma-ray luminosity vs. radio luminosity plot for blazars in the EGRET survey. The filled circles represent solid detections, the hollow circles represent marginal detections, and the crossed arrows represent the upper limits for those radio sources which EGRET did not detect (Stecker et al. 1993).
It may well be that the GRB is the sum of a number of different components with different origins. Each component's contribution varies as a function of energy. We need to keep in mind that when we study the gamma-ray background, we are looking at an energy span of over six decades. A particular physical mechanism probably dominates in a specific energy range, for example: the radio-quiet AGNs at < 1 MeV; the diffuse matter-antimatter annihilations at ~ 1 to 100's MeV; the SN Ia's at ~ 1 MeV (see Section 7.2); the primordial black holes at 100's MeV; and blazars at 100's GeV.