Active Galactic NucleiActive Galactic Nuclei (AGN) are the most distant and luminous objects in the known universe. As such, they represent a unique laboratory for the study of matter in extreme physical conditions. In particular, as they are widely thought to be powered by accretion onto a supermassive black hole, study of the central engines can shed light not only on the nature of the AGN phenomenon, but also on the properties of black holes. AGN have been historically identified and classified by their optical and ultraviolet spectral features. However, it is becoming clearer that the most direct way of investigating the fundamental energy source (the accretion flow onto the black hole) is the study of the X-ray emission. ASCA observations over the last two years have supported the accretion-disk origin of the Fe K-alpha line and yielded important new insight into the conditions in the central regions of the different types of AGN.
Iron K-alpha linesX-ray spectra of AGN show evidence for line emission peaking at a rest energy of 6.4 keV. This is thought to be due to a fluorescence line from the K-shell of iron, produced when the nuclear continuum radiation is reprocessed by circumnuclear material. Early results from ASCA showed the lines in these sources to be extremely broad, with FWHM implying relativistic velocities of order 0.2c. Furthermore, the Fe K-alpha lines showed profiles having a strong asymmetry to the red (MCG-6-30-15 Tanaka et al. 1995 Nature 375, 659; NGC 4151 Yaqoob et al. 1995 ApJ 453, L81), indicative of the gravitational redshifts associated with the inner regions of an accretion disk surrounding the black hole. Thus, the study of Fe K-alpha profiles in AGN represents one of the few cases where general relativistic effects are directly measurable, and are produced in the most intense gravitational field observed. These lines provide the means to probe the immediate environment of a black hole. Furthermore, as they are thought to arise by X-irradiation of the inner accretion disk orbiting the hole, the line profiles can be used to explore the central regions of AGN in detail. A study of archival ASCA data on a sample of Seyfert 1's (Nandra et al. 1997 ApJ 477, 602) has shown that asymmetric profiles are extremely common. Fits to a disk-line model show a very clear preference for low-inclination systems, indicating that the inner regions of Seyfert 1 galaxies tend to be observed face-on. Nandra et al. (1997 MNRAS 284, L7) presented a systematic analysis of a sample of Seyfert galaxies. This study showed that for Seyfert 1 galaxies the line emission comes from a range of radii in the disk; that the surface of the disk is not strongly photoionized (and is therefore very dense); that the pattern of X-ray illumination differs from source-to-source and that no strong component from an origin other than the disk is required by the data. Furthermore, two ASCA observations of NGC3516, separated by 1 year, showed a 60% drop in the flux of the Fe K-alpha line, correlated with a drop in continuum flux (Nandra et al. 1997 MNRAS 284, L7). This measurement constrains the line to originate within 1 pc of the nucleus. Yaqoob et al. (1996 ApJ 470, 27) also present evidence of rapid variability of the shape of the Fe K-alpha line in the narrow-line Seyfert galaxy NGC 7314. The X-ray continuum varies by a factor of 2 on a timescale of hundreds of seconds, and it appears that emission in the red wing of the Fe K-alpha line below ~6 keV responds to these variations on timescales of less than ~3x10^4 s. The response becomes slower and slower toward the line peak near 6.4 keV. An explanation of the less variable narrow core of the iron line is that it originates in distant, cold matter such as a molecular torus.
Figure 1: The mean line profile from ASCA of Seyfert galaxies (Nandra et al. 1997 ApJ 477, 602; Turner et al. 1997 ApJS 113, 23). The Seyfert 1s (top left panel) show a line peak at the rest energy of 6.4 keV, but most of the flux is shifted to the red. The width of the line implies relativistic velocities. The peculiar profiles, with FWHM of order 50,000 km/s, are only easily accounted for by invoking kinematic and gravitational effects close to a central black hole. NELGs and Seyfert 2s observed in transmitted X-rays (bottom left, top right) show remarkably similar profiles. Seyfert 2s observed in scattered X-rays (bottom right) show stronger emission lines and evidence for emission from highly ionized species. The major component is at 6.4 keV, however, and the emission still extends red-wards. This implies that all these classes have iron line emission from the central regions, close to the black hole.
An outstanding question is that of whether or not the black hole is rotating. In the Kerr Metric, the disk can extend closer to the central hole, approaching the gravitational radius. If this is indeed the case, we expect broader, more redshifted profiles in the Kerr Metric compared to the non-rotating (Schwarzschild) case. Iwasawa et al. (1996 MNRAS 282, 1038) found MCG-6-30-15 to show an extremely broad Fe K-alpha profile during a low-state, with significant line variability over the 4-day ASCA observation. Those authors concluded that a Kerr metric was required given the strength and shape of the line emission. Both Reynolds & Begelman (1997 ApJ 488, 109) and Weaver & Yaqoob (1998 ApJ, in press) have proposed alternative explanations. It is striking that this dispute is over what is occurring within a few gravitational radii of a black hole.
Based on Ginga data, Iwasawa & Taniguchi (1993 ApJ 413, L15) suggested an X-ray Baldwin effect in AGN, whereby the equivalent width of the Fe K-alpha line decreased with increasing luminosity. This result was disputed until the advent of ASCA data. Recent analysis of a large sample of QSOs confirms the X-ray Baldwin effect (Nandra et al. 1997 ApJ 488, 91). Using a large sample, those authors were also able to demonstrate that the observed relationship was not solely due to radio power, as the effect persists when only radio-quiet sources are considered. However, the effect can be due solely to luminosity of the central source. In addition to the change in strength, the line profile also appears to be different in higher luminosity sources (Fig. 2). Nandra et al. (1997 ApJ 488, 91) suggest that the accretion disk ionization state is a function of the accretion rate.
Figure 2. Data/Model ratios, constructed as in Fig.1, split into luminosity bins. The vertical dotted lines are at an energy of 6.4 keV. The profile for all sources (upper left) is dominated by the high signal-to-noise objects which are mostly low luminosity. Below L_X = 10^44 erg/s the line profiles are very similar, but above this luminosity there are clear changes. The line strength reduces with increasing luminosity, in both the core and red wing, but the blue flux is enhanced relative to the total line emission. Above L_X = 10^46 erg/s no evidence for line emission is observed at all. This confirms the ``X-ray Baldwin'' effect suggested by Iwasawa & Tanaguchi. The mean redshift is also shown, and demonstrates the strong correlation between redshift and luminosity in our sample.
Analysis of some bright ``intermediate'' Seyferts has allowed detailed analysis of the complex shape of the Fe K-alpha line. Weaver et al. (1997 ApJ 474, 675) demonstrate that the Fe K-alpha line in MCG-5-23-16 cannot be modeled with the profile predicted by standard disk theories. An additional component is required, most likely a narrow line produced in the putative torus or the broad-line-region. This result is the strongest direct evidence to date for two distinct reprocessors within a single Seyfert galaxy.
Recent analysis of ASCA observations of type-2 Seyfert galaxies showed that both Seyfert-2 galaxies and Narrow Emission Line Galaxies (NELGs) have a significant red-wing on the Fe K-alpha profile, at the same level as that observed in the Seyfert-1 sample (Fig 1; Turner et al. 1997 ApJ S 113, 23). A relatively strong core is observed in these lines, suggesting several line components contribute to the observed profile. In the case where we observe a spectrum dominated by scattered X-rays, then the scattering region may see a face-on disk, whose spectrum is scattered towards the observer. This can explain the broad base and red wing in the Fe K-alpha line observed in these sources. Iwasawa, Fabian & Matt (1997 MNRAS 289, 443) suggest another mechanism for producing the red wing in scattering-dominated sources, and the existence of a Compton-shoulder has been suggested in some cases.
In other Seyfert 2s the nuclear X-rays are observed, transmitted through a large column of gas (10^22-10^24 cm^-2). If these sources have their inner regions inclined edge-on, as predicted by AGN Unification Models, then the profiles of the Fe K-alpha lines should be systematically different from those observed for Seyfert~1 galaxies. In fact the mean profiles from several sub-groups of Seyfert 2 galaxies look remarkably similar, leading Turner et al. (1998 ApJ 493, 91) to suggest the orientation of the inner regions may be the same as for Seyfert~1 galaxies. If this were true, factors other than orientation would have to explain the observed differences between the classes. Unfortunately it is difficult to separate the components contributing to the Fe K-alpha line profiles. A narrow line component is likely to arise in circumnuclear material well outside of the accretion disk. Weaver & Reynolds (1998 ApJ, in press) challenge the conclusions of Turner et al. (1998 ApJ 493, 91), asserting that consistency with Unification Models can be achieved if the narrow core of the line is stronger than indicated purely from the column provided by the line-of-sight absorber.
Other X-ray Line Emission in AGNThe attenuation of the primary continuum by the torus reduces the contrast between observed nuclear and reprocessed emission. Thus, absorbed sources provide an opportunity to study some very important aspects of the AGN phenomenon not possible by observation of Seyfert~1 galaxies alone. In particular, it is possible to detect many soft X-ray emission lines in Seyfert~2 galaxies, which are rarely detected in Seyfert~1s. Hydrogen-like and helium-like lines are detected from Fe, Ne, Si, S, Ar and Mg and these cannot be explained solely by the presence of starburst emission (Turner et al. 1997 ApJ S. 113, 23). Netzer et al. (1998 ApJ, in press) find many lines to be consistent with an origin in narrow-line-region clouds. Arguably the most spectacular X-ray result for a Seyfert 2 galaxy to date is provided by the ASCA spectrum of Circinus (Matt et al. 1996 MNRAS 280, 823). The observed spectrum appears to be purely reprocessed radiation, allowing precise measurement of many emission lines.
Warm AbsorbersThe confirmation of so-called ``warm absorbers'' in AGN has been one of the major scientific successes of ASCA. Recent and on-going monitoring campaigns are leading to further discoveries. K-shell absorption edges of warm oxygen (OVII and OVIII) are seen in roughly half of the type 1 AGN observed with ASCA (George, et al. 1998 ApJ Suppl. 114, 73; Reynolds 1997 MNRAS 286, 513). This absorption arises in a photoionized plasma of suitable ionization parameter, in which X-rays above the O^+6 edge at 739 eV or the O^+8 edge at 871 eV are absorbed, while lower energy X-rays are transmitted because of the absence of bound electrons with lower ionization potentials. The variability of the optical depths of the absorption edges yields valuable information about the density and location of the warm absorbing gas, since the absorption features respond to changes in the X-ray flux on timescales of ionization and recombination.
The essential features of this model were supported by the observed variability of the absorption edges in MCG-6-30-15 on time scales of hours (Otani et al. 1996 PASJ 48, 211). However, the variability patterns of the edges in MCG-6-30-15 and of warm absorbers in general suggest a complex absorption structure. In some objects, rapid variability of the warm absorber coupled with a lack of correlated variability of the edges suggests there are at least two absorption ``zones'', the first being highly ionized and located close to the central source (perhaps situated within the broad-line region), and the second being a more diffuse ionized medium, possibly a wind from the molecular torus or the warm scattering medium seen in Seyfert 2s. The highly-ionized zone may have been observed previously in the deep Fe K-shell edges observed with Ginga (Nandra & Pounds 1994 MNRAS 268, 405). In other objects, such as NGC 1365, photoionization modeling of simultaneous far-UV and X-ray spectroscopic observations imply an absorbing region with a broad range of ionization parameters and column densities (Kriss et al. 1996 ApJ 467, 629). Still other objects have significant optical reddening and display deep OVII edges, which suggests the existence of a dusty warm absorber (Reynolds 1997 MNRAS 286, 513; George et al. 1998 ApJ S 114, 73).
Narrow-Line Seyfert~1 Galaxies - NLSy1sSeyfert galaxies with relatively narrow permitted lines are denoted NLSy1s. These sources show distinctive characteristics in the X-ray regime, such as the strong anti-correlation between soft X-ray slope and FWHM H-beta (Boller et al. 1996 A&A 305, 53). ASCA has revealed that this correlation also exists between the hard X-ray slope and FWHM H-beta (Brandt et al. 1997 MNRAS 285, L25). ASCA data also showed greater amplitudes of variability in the hard X-ray band for NLSy1 galaxies than for Seyfert~1 galaxies of comparable luminosity. The distinctions between Seyfert 1 and NLSy1 galaxies may be driven by a fundamental difference between the two, such as accretion rate.
In an exciting new result, Leighly et al. (1997 ApJ 489, L25) report the observation of features near 1 keV in the ASCA spectra from three NLSy1s. Those authors interpret these as oxygen absorption in a highly relativistic outflow. If interpreted as absorption edges, the implied velocities are 0.2--0.3 c, near the limit predicted by "line-locking" radiative acceleration. If instead interpreted as broad absorption lines, the implied velocities are ~0.57 c, interestingly near the velocity of particles in the last stable orbit around a Kerr black hole, although a physical interpretation of this is not obvious. The features are reminiscent of the UV absorption lines seen in broad absorption line quasars, but with larger velocities.
Observation of the bright NLSy1 Ton S~180 shows an Fe K-alpha line which is broad and asymmetric, indicating an origin in the accretion disk. However, that line has an energy close to 7 keV, this, and the detection of spectral features consistent with soft X-ray emission lines suggest the existence of an ionized accretion disk in Ton S~180, consistent with a higher accretion rate in this source, compared to Seyfert 1 galaxies.
Halpern & Moran (1998 ApJ 494, 194) use ASCA data to confirm the NLSy1 nature of IRAS 20181-2244. This source also shows rapid X-ray variability, ruling out the previous supposition that it is a hidden/scattered QSO, or type 2 QSO. The distinction is very important because high-luminosity NLSy1 objects are not uncommon, but type-2 QSO are very rare. This ASCA result leaves no known cases of X-ray emitting QSOs of type-2. The absence of such a population must be accounted for by Unification Models.
Starburst and Ultra-luminous infrared galaxiesCircumnuclear starbursts, with their high kinetic energy deposition of order 10^56-10^59 ergs into relatively small volumes, are capable of disrupting the inner gaseous disks of their host galaxies, which leads to large-scale (of order 10 kpc) outflows into their halos, so-called ``superwinds'' (Chevalier & Clegg 1985 Nature, 317, 44; Heckman, Armus, & Miley 1990 ApJS 74, 833). Such gaseous outflows should produce diffuse soft X-ray radiation and studies with ASCA and ROSAT have now confirmed their existence. This implies that starbursts are not only sites of chemical processing of hydrogen, but also sources of processed material that is fed into intergalactic space if the associated outflows reach the escape velocity of the system. Understanding starbursts not only helps us comprehend the chemical evolution of galaxies, but could also have significant implications for studies of the intergalactic medium.
M82 and NGC 253 are the most well-studied Starburst galaxies in X-rays. ASCA observations (Tsuru et al. 1997 PASJ 49, 619; Moran & Lehnert 1997 ApJ 478, 172; Ptak et al. 1997. AJ 113, 1286; Dahlem, Weaver & Heckman 1998 ApJS, in press) have proven the intricate nature of their spectra and have allowed the physics of these systems to be examined in great detail for the first time. Between 0.1 and a few keV, the spectra are dominated by thermal emission with lines from highly ionized Mg, Si, and S. Above a few keV, there is a hard component extending to >10 keV. The soft X-rays are extended, absorbed by the galaxy disk, and clearly associated with the outflow, while the hard X-rays are pointlike and are most likely due to XRBs (or a possible weak AGN). The data indicate the presence of at least two gas phases, one at ~0.2-0.4 keV and the other at ~0.65-0.9 keV. There is some evidence that the abundances of O, Ne, Mg, Si, S, and Fe are significantly lower than the cosmic value (Tsuru et al. 1997; Ptak et al. 1997), in which case neither type-Ia nor type-II supernova explosions can reproduce the observed abundance ratio. However, if the data are modeled to allow for the absorption of the X-ray emitting gas by the edge-on plane of the host galaxy (seen in the Rosat images), the abundances are not required to be significantly sub-solar (Dahlem et al. 1998).
Ultra-luminous infrared galaxies (ULIRGs) may provide an evolutionaly link between starburst nuclei and active galactic nuclei. If ULIRGs evolve from being starburst-dominated to being powered by an embedded AGN, then we should find examples of both stages of evolution, with the IR emission (indicating the amount of dust) and X-ray emission (indicating a buried AGN) being the best tracers of the change from the starburst-dominated phase to gas-dust enshrouded AGN phase. ASCA, with its broad bandpass coverage of soft energies (where the starburst contribution dominates) to hard energies (where the emission from the buried AGN can leak out) is thus perfectly poised to address the question of whether ULIRGs are powered by star formation and/or a deeply embeddd AGN.
To date, both types of objects have been found, but how they fit into an evolutionary sequence is not clear. IRAS20551-4250 (Misaki et al. 1997, IAU Symp. 186, in press), IRAS 20460+1925 (Ogasaka et al. 1997 PASJ 49, 179) and NGC 6240 (Turner etal 1997; Iwasawa & Comastri 1998 MNRAS, in press) all possess a hard X-ray, absorbed power-law component and an Fe K-alpha emission line that is consistent with fluorescence from neutral Fe. Their X-ray characteristics, which are consistent with Seyfert 2 galaxies, imply that these merging systems contain and are powered by an embedded AGN. On the other hand, NGC3690+IC694(Arp299), NGC1614 (Watarai et al. 1998, IAU Symp. 186, in press), and Arp 220 (Nakagawa et al. 1998, IAU Symp. 186, in press) have spectra dominated by thermal plasma emission with kT ~ 1 keV, similar to starbursts. A lack of any sign of hidden AGN activities implies that these objects are powered by star formation. It is interesting to note that Arp220 and NGC 6240 are at almost the same evolutionary stage. This suggests that the contribution of AGN activity to total luminosity varies from galaxy to galaxy and there is no obvious correlation with evolutionary sequence.
Low-luminosity AGN and LINERsExtensive optical spectroscopic studies reveal that ~1/3 of nearby galaxies have low-level activity classifying them as LINERs (low ionization nuclear emission-line regions). The ionization mechanisms of LINERs, which includes shock excitation and photoionization by low luminosity AGN are still under debate. ASCA observations of LINERs and low-luminosity AGN (Ptak 1988, Ph.D. thesis, UMD; Terashima et al. 1997, IAU Symp. 184, Kyoto Japan) show that in most cases the spectra are well described by a canonical model consisting of a power-law with a photon index, gamma ~ 1.7-2.0, plus soft, optically-thin thermal emission with kT ~ 0.6-0.8 keV. LINERs typically show a hard point-like nuclear source of X-ray luminosity of 10^40 - 10^41 erg/s, which provides strong support for the presence of a low luminosity AGN. The soft component is usually extended and the nuclear, point-like emission is sometimes absorbed by column densities in the range NH = 10^20-10^23 cm^-2. The hard component is similar to the X-ray spectra of quasars and classical Seyfert galaxies; however, no significant short time scale ( < day) variability is seen. Most of the galaxies have L_X/L_H-alpha ratios comparable to those of more luminous AGN. These results suggest a strong connection between classical AGN, LINERs, and starburst galaxies.
BlazarsASCA data on blazars, which are highly variable AGN dominated by emission from a relativistically expanding jet, imply that there are significant differences in the X-ray spectra of blazars exhibiting quasar-like emission lines, and those associated with BL Lac-type objects, generally devoid of spectral features. The former show relatively hard X-ray spectra, with energy indices alpha ~ 0.7 - 1, while the spectra of the latter are significantly softer, with alpha ~ 1.5 (Kubo et al. 1998, ApJ, in press). Very interesting constraints on the physics of the jet and the nature of the radiating particles can be gleaned from the spectral variability associated with the changes in flux. ASCA has now entered a phase where it often conducts long, essentially uninterrupted, observations, lending itself well to the necessarily lengthy monitoring observations.
Some such observations have already been conducted: perhaps the most dramatic example is the BL~Lac object Mrk~421, emitting in the GeV (Lin et al. 1992 ApJ 401, L61) and TeV (Punch et al. 1992 Nature 358, 477) bands. The 1994 multiwavelength campaign showed that while the keV X-ray and TeV gamma-ray fluxes varied simultaneously, flux in other bands remained relatively steady (Macomb et al. 1995 ApJ 449, L99). This suggests that both keV and TeV spectral regimes are produced by the same, most energetic end of the electron population, radiating via the synchrotron process in the keV, and Compton (most likely synchrotron self-Compton: SSC) process in the TeV bands. The rapid variability in the TeV band and the pair opacity limit imply that the source of both keV and TeV emission is expanding relativistically, with Doppler factor delta > 5. Since production of TeV photons requires Comptonization by TeV electrons, both TeV and keV photons are produced by electrons with gamma_e ~ 10^6. The key evidence that this version of the SSC model is correct comes from the spectral variability observed in X-rays. The time-resolved spectral variability observed by ASCA manifests itself as a ``soft lag'' (or ``hard lead'') (Takahashi et al. 1996, ApJ 470, L89), and implies a lifetime of radiating electrons to synchrotron losses t_s (E/1 keV) of 6000~s. This, with delta = 5, implies B ~ 0.2 Gauss, and yields gamma_e ~ 5x10^5 (E/1 keV)^1/2, in excellent agreement with the value obtained by the Compton energy transfer argument above.
With the recent detection of two more TeV-emitting BL Lac objects, and several multi-wavelength campaigns planned for the current cycle, ASCA observations are crucial to our understanding of the structure of these enigmatic sources of the highest energy photons observed in the Universe. A good example is an observation of another TeV blazar, Mkn 501 (Kataoka et al. 1998, ApJ, submitted); however, this observation was clearly undersampled. A continuous ``long look'' is needed to learn more about radiation processes in this object and to allow for a comparison with Mkn 421 to find out whether the behavior described above is a general property of BL Lac - type blazars, or is unique to Mkn 421.
Broad-line Radio GalaxiesASCA continues to make leaps forward in our understanding of radio galaxies and their place within the unification scheme of AGN. The X-ray spectrum of the blazar-like radio galaxy 3C 120 is complex, and consists of an absorbed power law (nH ~ 2x10^21 cm^-2), a steep soft excess, and a broad (sigma ~ 0.4-0.8 keV) iron line at ~6 keV (EW ~ 300 eV) plus a possible hard component, corresponding either to reflection from an accretion disk or to a flatter power law from a jet (Grandi et al. 1997, ApJ, 487, 636). The broad-line radio galaxy 3C 445 also shows an absorbed, relatively flat spectrum (nH ~ 10^22-23 cm^-2) and a broad Fe K-alpha line (sigma ~ 0.2 keV indicating gas with velocities of ~35,000 km/s) with with EW ~ 270 eV (Sambruna et al. 1998 ApJ 495, 749). The significant intrinsic absorption and broad Fe K line indicate that the X-ray emission in the 0.5--10 keV energy band is dominated by a Seyfert-like component. The inverse-Compton X-ray emission from the radio jets appears to become important at > 10 keV. These results support unified schemes for active galaxies, and demonstrate a remarkable similarity between the X-ray properties of powerful radio sources and those of lower luminosity, Seyfert 1 galaxies. The inverse-Compton X-ray emission from the radio jets appears to become important at > 10 keV.
Last modified: Tuesday, 26-Jun-2001 14:22:36 EDT
If you have any questions concerning ASCA, visit our Feedback form.