Science with the Hubble Space Telescope -- II
Book Editors: P. Benvenuti, F. D. Macchetto, and E. J. Schreier
Electronic Editor: H. Payne

---

Evidence for a Massive Black Hole in the Active Galaxy NGC 4261 from Hubble Space Telescope Images and Spectra

Laura Ferrarese & Holland Ford
Johns Hopkins University and Space Telescope Science Institute, Baltimore, MD 21218 USA

Walter Jaffe
Leiden Observatory, PB 9513, Leiden, 2300 RA, Netherlands

Abstract:

We present Hubble Space Telescope (HST) Wide Field and Planetary Camera 2 (WFPC2) V, R and I images and HST Faint Object Spectrograph (FOS) spectra of the active galaxy NGC 4261. The galaxy hosts a nuclear disk of dust (Jaffe et al. 1993,1996), roughly perpendicular to the radio lobes. We find evidence that the disk is not in an equilibrium configuration: it is not coaxial with the major axis of the galaxy, and it is not centered on either the nucleus nor on the isophotal center of the galaxy. An H map of the nuclear region shows that the ionized gas is concentrated in a resolved region with a FWHM of 0.12 arcsec, corresponding to 17 pc at the distance of NGC 4261 (30 Mpc). FOS spectra were taken with the 0.1 arcsec aperture in the wavelength region between 4570Å and 6870Å, in a grid of 13 aperture positions around and at the nuclear location. The central velocities of the [NII] lines as a function of distance from the center can be accounted for by assuming that the ionized gas is confined in a disk in Keplerian motion around a central mass M. By integrating the unreddened V luminosity density, we find a mass to light ratio M/L within the inner 14.5pc. The large mass to light ratio, and the fact that NGC 4261 is a relatively strong radio galaxy, lead us to conclude that the majority of the central mass is concentrated in a M black hole.

Keywords: black hole physics - galaxies: individual (NGC 4261, 3C270) - galaxies: morphology, kinematics and dynamics - galaxies: nuclei

Introduction

Since the seminal papers by Sargent et al. (1978) and Young et al. (1979), astronomers have tried to establish the existence and masses of central black holes by measuring the rotation and velocity dispersion of stars near the nucleus. An alternative to stellar dynamical studies is provided by the exploration of gas dynamics: in contrast to stellar absorption lines, the bright emission lines of the ionized gas can be easily and accurately measured. Disks of ionized gas and dust have now been detected in a number of cases (e.g., Jaffe et al. 1994, Lauer et al. 1996, O'Neil et al. 1994). It is interesting to note that, for all the radio galaxies in which gas and/or dust disks are detected, the minor axes of the disks are aligned with the radio jets, even when the direction of the jet is perpendicular to the galaxy spin axis (as in NGC 4261, Davies & Birkinshaw 1986). This fact, and the continuity of the ionized gas and/or dust into the active nucleus, suggest that the disks are intimately connected with the mechanism aligning the jets. The disks are obvious reservoirs of fuel for the central engines, and may determine the direction of the jets by defining the angular momentum in the inner accretion disk.

The presence of a disk of ionized gas in the giant elliptical 3C274 = M87, the brightest radio galaxy in the Virgo cluster, has already allowed the successful detection of a black hole (Ford et al. 1994, Harms et al. 1994). This paper deals with another interesting example: the giant elliptical NGC 4261, located at a distance of 30 Mpc (Nolthenius 1993). The galaxy is the optical counterpart of the radio source 3C 270 (Birkinshaw & Davies 1985). The radio structure consists of an unresolved component at the location of the galaxy nucleus, and two radio lobes extending out to 40 arcmin from the nucleus in a direction roughly perpendicular to the major axis of the galaxy. Pre-refurbishment HST/WFPC images of NGC 4261 revealed a sharply defined disk of dust, 240 pc in diameter, whose minor axis is roughly aligned with the radio axis of the galaxy (Jaffe et al. 1993). H, [NII] and [SII] emission lines have been detected in the nuclear region (Kim 1989) and the gas dynamics have been studied in detailed by Jaffe et al. (1996) using high resolution ground-based spectra. The spectra show strong H , [SII] and [NII] emission lines with narrow cores and unusually broad bases, but do not allow an unambiguous measurement of the central mass because of the insufficient spatial resolution of the spectroscopic data and because of seeing effects.

In this paper, we present HST/WFPC2 V, R and I images and HST/FOS spectra of NGC 4261. From the WFPC2 images, we derived a pure emission H image, used to study the morphology and luminosity of the ionized gas in the nuclear region. The HST/FOS spectra reveal that the ionized gas in the nuclear region is in circular motion, and point to the presence of a central mass inside arcsec (15 pc). The corresponding mass to light ratio is M/L, leading us to conclude that we have another--but no less exciting--nuclear massive black hole.

The plan of this paper is as follows: describes the observations and data reduction. discusses the V, R and I morphology, and the properties of the H emission. The reader interested mainly in the calculation of the central mass can skim and read about the FOS observations and the kinematic models in .

Observations

NGC 4261 was imaged on December 12 and 13, 1994, using the HST/WFPC2. The nucleus of the galaxy was centered on the Planetary Camera chip (PC1), which has a pixel size of 46 mas, corresponding to 6.7 parsecs at the distance of NGC 4261 (30 Mpc), and a field of view of 3636 sec ( kpc).

Since a narrow band filter covering the wavelength region of the redshifted H + [NII] emission (--6700Å) is not available in the WFPC2, we used the wide band filter F675W (transformed to Cousins R) for the on-band observations. Two off-band images were obtained in emission free spectral regions on either side of the H emission using the filters F547M (transformed to Johnson V) and F791W (transformed to Cousins I). The exposure times were 2000 seconds for the on-band image, and 800 seconds for each off-band image. All exposures were divided into two equally long images, referred to as CR-split frames, to facilitate cosmic ray removal.

Spectra of NGC 4261 were obtained on August 3, 1995, using the HST/FOS. One of the 0.1-PAIR square apertures (projected dimensions on the sky 0.09 0.09 arcsec) was used to measure the spectrum of the nuclear disk at the 13 positions shown in Figure 2. In the remainder of this paper we will refer to the different aperture positions as labeled in the figure. The grating used, G570H, covers the wavelength region between 4570Å and 6870Å, has a dispersion of 4.36Å per diode, and a spectral resolution of 4.01Å. The exposure times were 2040 seconds at all positions, except for the NE off-axis position, which was exposed for 1640 seconds, the SE and NUC positions, which were both exposed for 2030 seconds.

Further details on the observations, and a complete discussion of the calibration of the WFPC2 images and FOS spectra can be found in Ferrarese, Ford & Jaffe (1996)

Morphology

The fully reduced V, R and I HST/WFPC2 images of NGC 4261 are shown in Figure 1. The scale and orientation is the same in the three images, but the intensity level has been adjusted independently for each, so that relative brightnesses cannot be inferred from the figure. In the WFPC2 R image, which includes the H + [NII] complex, a bright region of line emission surrounds the nucleus. The continuum emission from the nucleus, presumably of non-thermal origin, is itself unresolved in the V and I images.

 
Figure: HST/WFPC2 V, R, and I images of NGC 4261. The galaxy was centered in the PC. Only the inner 3 arcsec are shown. The V and I off band images are shown in the left panels, while the R on band image and the continuum subtracted H images are shown in the panels to the right. The scale and orientation, shown at the left, are the same for the four images.

The most striking feature unveiled by the WFPC2 images is the nuclear dust disk, approximately 1.6 0.7 arcsec ( pc), first detected by Jaffe et al. (1993). The mass of the disk, estimated using the prescriptions by Bohlin et al. (1978), is M. This value is consistent with the findings of Jaffe et al. (1993), which further note that the total mass in the disk, if converted into energy at 1%-2% efficiency, corresponds to the total energy in the radio lobes, about erg, assuming a synchrotron lifetimes of yr. The disk is nearly perpendicular to the axis of the radio lobes, observations that prompted Jaffe et al. (1993) to propose that the disk is responsible for the orientation of the lobes, possibly through the transfer of angular momentum from the disk to the central black hole. The restored resolution capabilities of the WFPC2 allow a more detailed study of the structure of the dust disk than was possible in Jaffe et al. (1993,1996) using WFPC1 data. The disk, which is inclined 26^o with respect to the line of sight, is not perfectly elliptical in shape: a small protuberance is seen extending from the North edge; in addition, the North side of the ellipse is rounder and wider than the South side, possibly suggesting that the disk is slightly warped or twisted along its major axis. A bright `spiral' structure can be detected extending from the nucleus to the North and South side of the disk. These patterns repeat identically in the three photometric bands and cannot be attributed to line emission from ionized gas; Ferrarese, Ford & Jaffe (1996) also proved that the spiral structure corresponds to regions of lower dust content and, therefore, lower optical depth rather than of recent star formation. The spiral structure may provide the means by which angular momentum is transferred from the center of the disk out, allowing the innermost material to sink in. The FOS spectra () show evidence for large random motions and shock waves in the innermost parts of the disk, both of which could be produced by the gravitational energy released by the gas infalling into the black hole potential well.

 
Figure: An enlarged view of the nuclear dust disk, obtained from the V band image. Superimposed onto the image is the ellipse that best fits the edge of the disk (solid white line), and the elliptical isophote just outside the disk (dashed white). The squares represent the position of the 13 HST/FOS aperture pointings, named as indicated in the inset in the lower right. The black ellipse shows the best disk model that reproduces the velocity map obtained from the HST/FOS spectra ()

Further evidence that the disk is not in an equilibrium configuration comes from noticing that the major axis of the disk is misaligned by about 5^o with respect to the major axis of the galaxy. While a misalignment in the projected major axes of the disk and of the galaxy can correspond to an equilibrium configuration in a triaxial galaxy, the intrinsic figure of NGC 4261 is most likely a prolate spheroid, with no or very little trace of triaxiality (Davies & Birkinshaw 1986). As noted by Jaffe et al. (1996), the axis of the radio jets, (Birkinshaw & Davies 1985) projects about 15^o from the minor axis of the disk, implying that the jets and the disk are not exactly perpendicular to each other. Finally, a linear dust feature is seen extending beyond the edge of the disk in a radial direction, at a position angle of ^o, or about 33^o from the minor axis of the dust disk. A similar, but much fainter, feature extends in the opposite direction, beyond the North-East edge of the disk, suggesting that we may be looking at a linear jet of dust ejected from the nucleus (a dust ring or disk seen edge on would produce the same optical depth on both sides of the nucleus, contrary to the observations). The near side of both the linear dust feature and of the radio jet, project onto the West (brighter) side of the disk, which we identify with the far side (Ferrarese, Ford & Jaffe 1996), in agreement with the fact that the radio jet and the presumed dust jet are roughly perpendicular to the disk.

We confirm the Jaffe et al. (1996) suspicion that the disk is not positioned at the isophotal center of the galaxy (defined as the mean of the centers of the isophotes between 1.9 and 4 arcsec), and that the nucleus itself is not at the center of either the disk or the galaxy. With respect to the center of the galaxy, the nucleus is shifted by mas (= pc) to the South-West, while the disk is shifted by mas (= pc) in the same direction. A schematic representation of these results is given in Figure 2, showing an enlarged section of the V band image. The ellipse best fitting the dust disk and the ellipse describing the isophotes just outside the disk are drawn as white solid and dashed lines, respectively; the miscentering and different orientation of the two is readily apparent. The miscentering between the nucleus and the isophotal center, which amounts to less than half a pixel, is less evident but statistically significant, while the offset between the disk and the nucleus is quite obvious. The displacement of the disk with respect to the center of the galaxy may suggest that the dust has an external origin and has not yet reached equilibrium. This hypothesis, already advanced by Jaffe et al. (1996), is further supported by the fact that the disk spin axis, which projects approximately along the minor axis of the galaxy, is different from the galaxy spin axis (the galaxy rotates around its major axis, Davies & Birkinshaw 1986). This could be the case if the dust has been captured from a cannibalized galaxy, rather than being accreted from material internal to NGC 4261 itself.

We do not it find likely that the nucleus, and the associated black hole, have been acquired in a recent capture of a nearby galaxy: the decay time for oscillations of the nucleus in the gravitational potential of the galaxy is of the order of -- years. Since it is improbable that we would have caught this nucleus within only ten million years of its capture, we argue against a recent capture. However, it is possible that the nucleus, which is moved along the direction of the radio jet, could have been displaced by recoil, assuming that the jet is collimated and pushing against the nucleus for a few years (Ferrarese, Ford & Jaffe 1996).

In order to produce a pure line-emission image, the two de-extincted (using the extinction law of Cardelli, Clayton & Mathis 1989) V and I off-band images were averaged, and the resulting image, scaled by an appropriate factor, was subtracted from the de-extincted on-band R image. The line emission is resolved, and concentrated in a region with a FWHM of about 0.12 arcsec around the nucleus. The mass of the ionized gas is of the order of a few hundreds of solar masses (Ferrarese, Ford & Jaffe 1996). The resulting H luminosity is erg s, typical of AGN's with similar X-ray and radio fluxes (Baum & Heckman 1989, Trinchieri & di Serego Alighieri 1991).

Spectral Analysis

The FOS spectrum obtained at the nuclear position is shown in the upper panel of Figure 3. Emission lines of H , H, [OIII] 4459,5007, [OI] 6300,6364, [NII] 6548,6584 and [SII] 6717,6731 are prominent. The lower panel of Figure 3 overplots the [NII]+H spectral region at positions NUC, N1 and S1, i.e., along the major axis of the dust disk. The fluxes at the off-nuclear positions have been scaled to facilitate the comparison. The solid vertical line is at the systemic velocity of the galaxy, 2210 14 (de Vaucouleurs et al. 1991). The dashed lines show the centroids of [NII]6584 at the three aperture positions. The spectra show the clear signature of rotation; the N1 emission is redshifted with respect to the nuclear velocity, while the S1 emission is blueshifted. The large velocity difference observed for the gas 0.09 arcsec away from the nucleus along the major axis of the dust disk ( between the N1 and S1 positions), if ascribed entirely to circular motion of the ionized material in the plane of the disk, points to a total mass inside arcsec of M. The total V luminosity integrated in the same volume is three orders of magnitudes smaller (Ferrarese, Ford & Jaffe 1996) and, therefore, the stellar component in the galaxy does not contribute significantly to the potential in the innermost regions.

 
Figure: The upper panel shows the spectrum of NGC 4261 obtained at the nuclear position. The lower panel overplots the [NII]+H spectral region at positions NUC, N1 and S1, i.e., along the major axis of the dust disk. The vertical line marks the position of the [NII]6584 emission at the systemic velocity of the galaxy, while the three vertical dashed lines mark the centroid of the [NII] 6584 emission at the three different aperture positions.

 
Figure: Predicted velocities versus the observed velocities for the best fitting Keplerian model. The zero point of the y axis corresponds to the velocity measured at the nuclear position. Within the errors, the data can be fit by a straight line, implying that the gas is in Keplerian motion.

The line fluxes and centroids are found using the line synthesis program SPECFIT (Kriss 1994), and are summarized in Table 2 of Ferrarese, Ford & Jaffe (1996). We detect the largest line of sight velocities along the major axis of the disk, while very little velocity is seen along the minor axis. The line widths decrease dramatically from the center outwards, going from 1500 at the nuclear position, to 300 at the outermost points. The large line width at the nuclear position, and the absence of obvious asymmetries in the line profiles, which can be seen more clearly in the Jaffe et al. (1996) data, suggest that the lines are broadened by rotation. The observations argue against the case of a gas outflow: a one-directional outflow would give rise to asymmetric line profiles, which are not observed. A bi-directional outflow roughly perpendicular to the dust disk would produce the largest velocity shift along the minor, rather than the major axis, contrary to the observations. In addition, an outflow is not compatible with the [NII]+H morphology, derived both from the spectra and from the WFPC2 image: no changes in the emission fluxes are observed as a function of the polar angle, as would be expected if the ionized gas had a bi-polar distribution.

The large velocities along the major axis and small velocities along the minor axis of the disk suggests that the gas is approximately in circular motion. In the remainder of this section we will validate this hypothesis by showing that the velocity map from the ten aperture positions in which the [NII] emission lines can be measured is consistent with the prediction of a simple model in which the gas is in Keplerian rotation. Following the procedure outlined by Harms et al. (1994), we compared the observed velocities to the prediction of a simple model in which the gas is confined in a disk in Keplerian motion around a central mass. The model, which is described in more detail in Ferrarese, Ford & Jaffe (1996), depends on three parameters: the central mass M (which plays the role of a simple scaling factor), the angle i between the plane of the disk and the line of sight, and the angle between the projected major axis of the disk and the line connecting the N1 and S1 exposures. In comparing the model with the observed velocities, three more parameters come into play: the systemic velocity of the galaxy (i.e., the zero point of the velocity map), and the offsets and , in both x and y directions, that could have occurred between the center of the FOS nuclear exposure and the center of the H emission due to an imperfect peak-up. The data are best fit by assuming that in the region sampled by the FOS exposures the gas is confined in a disk whose plane forms a angle with the line of sight, and with major axis position angle rotated by about 35^o with respect to the major axis of the dust disk. Figure 4 shows a plot of the predicted velocities versus the observed velocities for the best fitting Keplerian model. Within the errors, the data can be fit by a straight line, implying that the gas is in Keplerian motion. The intercept of this line with the y axis defines the systemic velocity of the galaxy, while the slope of the line is equal to . A least square fit to the data yields , lower than the RC3 value (2210 14 ), and M for a distance to NGC 4261 of 30 Mpc. Integrating the V luminosity density inside this radius gives L, therefore, the mass to light ratio within the inner 14.5 pc is M/L. For a typical elliptical galaxy, to 10, therefore, a normal stellar population cannot be responsible for the observed mass to light ratio. This, combined with the presence of an active galactic nucleus, points to the conclusion that the majority of the mass is concentrated in a central black hole with M.

We conclude by noting that the spectroscopic observations support the hypothesis, already advanced from the study of the morphology of the dust disk (), that the disk is not in an equilibrium state. The inner disk, defined by the spectroscopic data, is rotated with respect to the outer dust disk observed from the WFPC2 images, suggesting that the entire structure is tilted and warped. Within the errors, the minor axis of the inner disk is aligned with the axis of the radio lobes.

References:

Birkinshaw, M. & Davies, R.L. 1985, Ap.J., 291, 32

Bohlin, R.C., Savage, B.D., & Drake, J.F. 1978, ApJ 224, 132

Cardelli, J.A., Clayton, G.C., & Mathis J.S. 1989, ApJ 345, 245

Davies, R.L. & Birkinshaw, M. 1986, Ap.J.Lett, 303, L45

Ferrarese, L., Ford, H.C., & Jaffe , W. 1996, Ap.J., in press

Ford, H.C. et al. 1994, Ap.J.Lett, 435, L27

Harms, R. et al. 1994, Ap.J.Lett, 435, L35

Jaffe, W., Ford, H.C., O'Connell, R.W., Ferrarese, L., & van den Bosch, F. 1993, Nature, 364, 213

Jaffe, W., Ford, H.C., Ferrarese, L., van den Bosch, F., & O'Connell, R.W. 1996, Ap.J., March 20 issue

Jaffe, W., Ford, H.C., O'Connell, R.W., van den Bosch, F., & Ferrarese, L. 1994, AJ 108, 1567

Kim, D.-W. 1989, ApJ 346, 653

Kriss, G. A. 1994, in Astronomical Data Analysis Software and Systems III, ASP Conf. Ser. 61, ed. D. R. Crabtree, R. J. Hanisch, & J. Barnes, (San Francisco: ASP), 437

Lauer T.R. et al. 1996, submitted to A.J.

Nolthenius, R. 1993, 85, 1

O'Neil, E.J., Lynds, R., De Young, D.S, & Idt, W.C. 1994 BAAS, 26, 884

Sargent, W.L.W. et al. 1978, ApJ 221, 731

Trinchieri, G. & di Serego Alighieri, S. 1991, A.A., 101, 1647

Young, P.J., Sargent, W.L.W., Kristian, J., & Westphal, J.A. 1979 ApJ 234, 76