Figure 1. Fits of Schwarzschild (solid lines) and Kerr (dotted lines) disk models to the Fe K-alpha line profiles of two Seyfert 1 galaxies, from Nandra et al. (1996 ApJ in press).

Figure 2. A model fit (Bromley et al. 1996 ApJ in press) to the Fe K-alpha line profile of MCG-6-30-15 (Tanaka et al. 1996 Nature 375, 659). The model disk has inner radius 7.2 r_g, outer radius 15.8 r_g, inclination angle i = 30 deg, and power-law emissivity index q = 3.

Not only does this result provide the only direct view of the innermost, relativistic accretion disk, but it is also a dynamical demonstration of the existence of supermassive black holes more important than the recent evidence from spatially resolved water vapor masers in Seyfert galaxies, and from the emission-line rotation velocities in the cores of AGNs as seen by the HST. Furthermore, the inclination angles of ~ 30 deg required to fit the ASCA line profiles are consistent with orientation-dependent unification schemes, in which Seyfert 1 galaxies are the unobscured variety, viewed near the axis of the molecular torus, and assumed to be the same as that of the inner accretion disk. (In the non-relativistic limit the line profile from an accretion disk is double-peaked and symmetric about the rest energy. However, for small radii and small inclination angles, the Doppler-boosted blueshifted peak is also gravitationally redshifted, so that it emerges at roughly the rest energy. The redshifted peak is smeared to even lower energy, forming the sloping red wing down to 4 keV.) Continuing studies may be able to distinguish the even more extreme effects of rotating (Kerr) black holes.

Similarly broad Fe K-alpha lines have been seen in at least two intermediate (type 1.9) Seyfert galaxies (IRAS 18325-5926: Iwasawa et al. 1996 MNRAS 279, 837; MCG-5-23-16: Weaver et al. 1996a ApJ in press). In these cases, the profiles require disks of larger inclination angle than in the Seyfert 1 galaxies, exactly as would be expected in orientation-dependent unification schemes. An additional, narrow-line Fe K-alpha component is sometimes seen. The long delay in the response of the strong, narrow Fe K-alpha line to continuum variations in the Seyfert 1.9 galaxy NGC 2992 (Weaver et al. 1996b ApJ 458, 160) implies that is probably fluorescence in the cold molecular torus.

Obscured Seyfert Galaxies

All the other Seyfert 2 galaxies observed by ASCA look considerably different from NGC 1068. A typical example is Mrk 3 (Iwasawa et al. 1994 PASJ 46, L167). It shows a highly absorbed (N_H = 3-7 x 10^23 atoms/cm^2 hard X-ray continuum plus an unabsorbed soft component. The latter contains numerous emission features characteristic of both thermal and photoionized plasma, while the former is coming to us directly from the optically obscured Seyfert nucleus. There is also a strong 6.4 keV Fe K-alpha line of EW ~ 900 eV. While the hard X-ray continuum varied on a time scale of years, the soft X-ray remained constant. Since the Fe K-alpha line varied as well, we conclude that some of it must be coming from a region smaller than that responsible for the steady soft X-ray flux. All of these results lead to a model in which the soft X-rays come to us from an extended region parsecs in size, possibly scattered from the central continuum source, while the Fe K-alpha line is at least partly produced in the nucleus or in high column density cold gas that is much smaller than the scattering region. Mrk 3 and most other Seyfert 2 galaxies probably differ from NGC 1068 in their lack of substantial competing starburst activity, as well as having obscuring gas which is not as thick, at least not along our line of sight.

Figure 3. ASCA GIS spectrum of NGC 6552 fitted with a model dominated by reflection, plus a highly absorbed power law, from Reynolds et al. (1994 MNRAS 268, L55).

Figure 4. ASCA SIS spectrum of the Circinus galaxy, from Matt et al. (1996 MNRAS in press). Above 2 keV, it is dominated by reflection from cold gas, while below 2 keV the lines are identified with ionized species of Ne, Mg, Si, and S, which could be coming from the warm scattering mirror. There is also a line (F) at 7 keV which is attributed to Fe K-beta.

Even more highly obscured Seyfert 2 galaxies have been studied by ASCA. Most notable are NGC 6552 (Fukazawa et al. 1994 PASJ 46, L141; Reynolds et al. 1994 MNRAS 268, L55), the ``Circinus galaxy'' (A1409-65: Matt et al. 1996 MNRAS in press) and NGC 4945 (Done, Madejski, & Smith 1996 ApJ 463, L63). All have extremely strong ``cold'' Fe K-alpha lines. NGC 6552 was discovered serendipitously as a Seyfert 2 galaxy in the ASCA GIS, with an absorbing column density of ~ 9 x 10^23 atoms/cm^2. The entire ASCA spectrum of NGC 6552 can be interpreted in terms of reflection from cold gas (Reynolds et al. 1994 MNRAS 268, L55), which might be a mechanism for explaining the spectrum of the X-ray background. The Circinus galaxy is similar, but with an even stronger Fe K-alpha line of EW ~ 2 keV, and intermediate-ionization species of Ne, Mg, Si, and S. The case of NGC 4945 is even more extreme, with an absorption column of ~ 5 x 10^24 atoms/cm^2. In combination with a GRO OSSE observation, Done et al. (1996 ApJ 463, L63) have shown that NGC 4945 is actually the second brightest hard X-ray Seyfert in the sky, second only to NGC 4151.

If Seyfert 2 galaxies are as common in the X-ray as they are optically, then it is only their large obscuring column densities that hamper their detection in hard X-ray surveys. However, they might become easier to detect at high redshift. Therefore, it is very interesting that ASCA discovered serendipitously a narrow-line AGN at z = 0.9 that has an X-ray luminosity of ~ 1.7 x 10^44 ergs/s (Ohta et al. 1996 ApJ 458, L57), which the authors somewhat optimistically propose as a rare example of the elusive ``type 2 QSO''. Whatever the interpretation of this source AX J08494+4454 turns out to be, it is clear that ASCA has the ability to discover hard X-ray selected sources at cosmological redshifts.

Warm Absorbers

Observed variability of the absorption edges in MCG-6-30-15 on time scales of hours (Otani et al. 1996 PASJ 48, 211) confirms the essential features of the model. The O^+7 edge decreased in depth as the ionizing continuum flux increased, thus stripping the single electron from this hydrogen-like ion. The short response time, in combination with the ionization parameter required to explain the ionization balance, was used to set a lower limit on the density (n > 2 x 10^7 cm^-3) and an upper limit on the distance of the absorbing gas from the nucleus (R < 1.4 x 10^17 cm). Both of these values are consistent with a medium interspersed among the ordinary broad-line region clouds and in pressure equilibrium with them. The interpretation of the variability of warm absorption edges is complicated by the possible existence of additional diffuse absorbing structures. In the same object, MCG-6-30-15, Otani et al. (1996 PASJ 48, 211) found that the depth of the O^+6 edge does not vary, requiring them to postulate the existence of a much more diffuse medium, possibly associated with a wind from the hypothesized molecular torus, or the warm scattering medium responsible for the observed properties of Seyfert 2 galaxies.

The subject of intrinsic absorption in AGNs has undergone a major revitalization as a result of the detection of ionized oxygen edges in roughly half of the Seyfert galaxies observed by ASCA. Dubbed the ``warm absorber,'' this phenomenon was first postulated (Halpern 1984 ApJ 281, 90) to explain the peculiar shapes and variability of some Einstein spectra. In a photoionized plasma of suitable ionization parameter, 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 edges seen in absorption gives additional information about the density and location of the warm absorbing gas, since absorption features respond to changes in X-ray flux on the ionization and recombination time scale.

Observed variability of the absorption edges in MCG-6-30-15 on time scales of hours (Otani et al. 1996 PASJ 48, 211) confirms the essential features of the model. The O^+7 edge decreased in depth as the ionizing continuum flux increased, thus stripping the single electron from this hydrogen-like ion. The short response time, in combination with the ionization parameter required to explain the ionization balance, was used to set a lower limit on the density (n > 2 x 10^7 cm^-3) and an upper limit on the distance of the absorbing gas from the nucleus (R < 1.4 x 10^17 cm). Both of these values are consistent with a medium interspersed among the ordinary broad-line region clouds and in pressure equilibrium with them. The interpretation of the variability of warm absorption edges is complicated by the possible existence of additional diffuse absorbing structures. In the same object, MCG-6-30-15, Otani et al. (1996 PASJ 48, 211) found that the depth of the O^+6 edge does not vary, requiring them to postulate the existence of a much more diffuse medium, possibly associated with a wind from the hypothesized molecular torus, or the warm scattering medium responsible for the observed properties of Seyfert 2 galaxies.

Figure 5. Comparison of absorption edges in MCG-6-30-15 at two times less than 1 day apart, from Otani et al. (1996 PASJ 48, 211). The change above 0.9 keV is due to variation of the depth of the O^+7 edge at 871 eV.

If the covering fraction of the warm absorber is large enough, then one would expect to detect emission lines of the same highly ionized ions that are seen in absorption (Netzer 1993 ApJ 411, 594). In the Seyfert 1 galaxy NGC 3783, emission lines at 0.57 and 0.65 keV are seen in addition to the ionized oxygen edges (George, Turner, & Netzer 1995 ApJ 438, L6). These lines are identified with the primary transitions of O^+6, the He-like ion, and the Ly-alpha line of O^+7, respectively.

Considerable progress in understanding the multiwavelength absorption properties of AGNs has been made by simultaneous ultraviolet and X-ray observations. It has long been known that many Seyfert galaxies exhibit UV resonance absorption lines that are variable in depth, as well as possessing outflow velocities of a few thousand km/s . Photoionization modelling of the UV and X-ray features has led to a consistent explanation of all the observed absorption in three objects, NGC 5548, 3C 212, and 3C 351 (Mathur, Elvis, & Wilkes 1995 ApJ 452, 230), as occurring in material just outside the broad-line region. However, it is not clear that such a one-zone absorber is a sufficient model for all Seyfert galaxies. In the case of NGC 3516, Kriss et al. (1996a ApJ in press) made ASCA observations simultaneously with the Hopkins Ultraviolet Telescope, and argued convincingly that a single absorber could not simultaneously account for both the high-ionization X-ray absorption edges, and some of the low-ionization UV resonance lines that were seen. Not surprising, a range of densities and ionization parameters are probably required. The same conclusion was drawn by Kriss et al. (1996b ApJ in press) about the prototype Seyfert 1.5 galaxy NGC 4151, which is also rich in UV absorption lines. Apparently, there are additional absorbing structures that affect our view of these lower-luminosity Seyfert galaxies, either because of our viewing direction or their intrinsic properties. A promising model that tries to incorporate all of the UV and X-ray absorption phenomenon involves a radiatively driven wind from the accretion disk (Murray & Chang 1995 ApJ 454, L105).

At the extreme end of intrinsic absorption properties are the BALQSOs. Characterized by broad UV absorption troughs of width 0.1-0.2c, the BALQSOs were known from Einstein and ROSAT observations to be relatively weak X-ray sources. But it was not known whether this was an intrinsic property, or the result of strong absorption. An ASCA observation of the prototype BALQSO PHL 5200 at z = 1.98 has resolved the issue in favor of strong absorption (Mathur, Elvis, & Singh 1995 ApJ 455, L9). The observed intrinsic column density of 1.3 x 10^23 atoms/cm^2 is 100-1000 times larger than previously inferred from the ultraviolet absorption. Apparently, most of the BAL material is highly ionized. Another observation that has been related to the warm X-ray absorbing material in QSOs is the discovery of a broad Ne VIII 774 Angstrom emission line, which requires the same sorts of column densities and ionization parameters as does the X-ray warm absorber (Hamann, Zuo, & Tytler 1995 ApJ 44, L69; Hamann et al. 1995 ApJ 443, 606).

High-Redshift Quasars

The spectra of quasars at energies above 10 keV are of crucial importance to the understanding of their intrinsic spectral shape, and the search for a possible Compton reflection component which would peak around 30 keV. Such a component could be consistent with the spectrum of the X-ray background. Curiously, our best data on the spectra of quasars in the 10-50 keV range comes not from bright, low-redshift objects, but from observations of high-redshift quasars by ASCA. The spectrum of the z = 3.384 quasar S5 0014+813 in the ASCA GIS is consistent with a power law of energy index alpha = 0.63 +/- 0.03, or thermal bremsstrahlung of kT = 40 +/- 4 keV in the quasar frame (Elvis et al. 1994 ApJ 436, L55). Similar results were obtained on two other quasars at z = 2.852 and 3.275 (Serlemitsos et al. 1994 PASJ 46, L43). The highest redshift quasar observed by ASCA is Q1508+5714 at z = 4.30 (Moran et al., in preparation). Its power-law energy index is alpha = 0.40 +/- 0.05 over a detected rest-frame energy interval 3-53 keV. No significant intrinsic or intervening absorption is detected. An upper limit on the equivalent width of an Fe K-alpha line is 106 eV. Superficially, at least, these results would seem to indicate that AGNs can produce the X-ray background.

The Weakest AGNs

Turning from the most luminous to the least luminous AGNs, ASCA has also made contributions to the detection of low-level nuclear activity in weak Seyferts and LINERs. Einstein and ROSAT have left us with considerable remaining uncertainty as to whether the low-luminosity X-ray sources (L_X < 10^42 ergs/s) in many galaxies are to be attributed to AGNs, or instead to some manifestation of starburst activity. At the very lowest luminosity (L_X <= 10^40) ergs/s , confusion with individual luminous binaries, i.e., black-hole transients, becomes a problem. The nearby spiral galaxy NGC 3147, previously not known to harbor an active nucleus, was detected by ROSAT at the level of 10^42 ergs/s . On the basis of its weak optical emission lines, a Seyfert 2 classification was then proposed by Moran, Halpern, & Helfand (1994 ApJ 433, L65). A subsequent observation by ASCA detected a power-law continuum energy index alpha = 0.80 +/- 0.09 and L_X ~ 5 x 10^41 ergs/s , confirming its Seyfert activity (Ptak et al. 1996 ApJ in press). It is comforting that sensitive optical and X-ray spectroscopy are in agreement that this is indeed a weak AGN. The low-level active nucleus of the LINER galaxy M106 (NGC 4258), now famous for its spatially resolved maser disk, was first detected by ASCA, with power-law energy index alpha = 0.78 and L_X = 4 x 10^40 ergs/s in the 2-10 keV band (Makishima et al. 1994 PASJ 46, 77). Previous X-ray instruments had not been able to demonstrate the existence of an AGN in NGC 4258 because of its large column density, N_H = 1.5 x 10^23 atoms/cm^2 and extended thermal X-rays. ASCA clearly sees both the compact and the extended components. The latter has emission lines with a characteristic temperature of 0.5 keV.

A more troublesome case is that of NGC 3628. Originally classified as a starburst galaxy, it was discovered to have a variable, compact nuclear source in ROSAT observations. The ASCA spectrum of the nuclear source (Yaqoob et al. 1995 ApJ 455, 508) is very hard, with power-law energy index alpha ~ 0.2, comparable to that of a high-mass X-ray binary. But its luminosity of 3.4 x 10^40 ergs/s is much too high for Eddington-limited accretion onto a ~ 1 Solar mass object. No standard model, either AGN or black hole binary, is entirely satisfactory. But the existence of such an object is certainly something new and fascinating.

Blazars

While blazars are generally devoid of spectral features in the X-ray and elsewhere, ASCA makes significant contributions via its studies of the correlation of flux and spectral shape. Perhaps the most dramatic example is the BL Lac object Mrk 421, a GeV emitter (Lin et al. 1992 ApJ 401, L61), but also the first one to be detected at TeV energies by the Whipple Observatory (Punch et al. 1992 Nature 358, 477). 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 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 ASCA data show a ``soft lag'' (or ``hard lead''), such that the variability in the hard X-rays always leads that in the soft X-rays (Takahashi et al. 1996 ApJ in press). With the lifetime of radiating electrons to synchrotron losses expressed as t_s ~ 1.2 x 10^3 B^-3/2 (E/1 keV)^1/2 delta^-1/2 s, the ~ 1/2 hr lead of the 5 keV as compared to the 1 keV photons implies t_s (1 keV) of 6000 s, and thus, with delta = 5, B ~ 0.2 Gauss. This, in turn, yields gamma_e ~ 5 x 10^5 (E/1 keV)^1/2, nicely consistent with the value obtained by 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 next CGRO cycle, ASCA observations are a crucial component for the understanding the structure of these enigmatic sources of the highest energy photons observed in the Universe.

Radio Galaxies

ASCA has made the long-awaited detection of inverse-Compton X-ray emission from the lobes of extragalactic radio sources. This emission due to microwave background photons scattering of relativistic electrons is important because in combination with radio measurements it allows the determination of both electron energy and magnetic field energy in the lobes. Radio measurements alone cannot separate these so up to this point theorists have had to work with assumptions like that of equipartition, where the electrons and magnetic fields are assumed to have the same energy density. ASCA data now shows that Fornax A (Kaneda et al. 1995 ApJ 453 L13) is in equipartition however Centaurus B (Tashiro et al. 1996 Waseda meeting) is not, with the electron energy density exceeding the magnetic. Further, in Centaurus B the ratio of energy densities changes with distance from the central source implying that the magnetic fields are compressed towards the lobe edges.

The spectral resolution, spatial resolution, and bandpass of ASCA have also led to improved knowledge of Cygnus-A, the archetypal double-lobed radio galaxy. AGN unification theories suppose that such sources have their jets pointed away from our line of sight with an obscuring gas torus blocking our view to the central engine. ASCA data confirm that the power-law source absorbed by > 10^23 atoms/cm^2, detected by earlier missions, is indeed from the core of the radio source. This data also reveals a cold Fe fluorescence line whose equivalent width is consistent with production in a gas torus surrounding the hard X-ray source (Arnaud et al. 1996 in preparation).

ASCA science highlights

---

Last modified: Tuesday, 26-Jun-2001 14:22:33 EDT

---

If you have any questions concerning ASCA, visit our Feedback form.

This file was last modified on Tuesday, 26-Jun-2001 14:22:33 EDT