Astron. Astrophys. 338, 405-412 (1998)

4. Results and discussion

The central part of NGC 1068 covered by the slit configuration we used is shown together with our CASPEC echelle spectrogram observations in the same spatial scaling in Fig. 4. The l-v-map of the [OIII] 5007 emission line profile is clearly structured on scales of less than and the extension in velocity space amounts to km s^-1 (Fig. 1 and Fig. 4). The dominant feature is a double peaked structure located at the center with v -250 km s^-1 and an emission triplett at v 500 km s^-1 in the southwest direction. A comparison between the observation and the reconstruction model is shown for representative parts of the l-v map in the Fig. 4 (right four panels). Even in the outer parts of the l-v map of the [OIII] 5007 emission line profile the model fit is close to the observed intensity distribution.

The results of the reconstruction of the [OIII] 5007 emission line, i.e. the spatial location of individual emission complexes have been compared with HST-observations obtained by Macchetto et al. (1994). The HST image shown in Fig. 4 was taken with the F501N narrow-band filter corresponding to [OIII] 5007. It was smoothed with a gaussian filter of FWHM = to emphasize the dominant [OIII] 5007 emission features. Six bright emission line knots are visible in the smoothed HST image and at least 10 weaker structures. In the original NGC 1068 HST-frame more than 30 cloud complexes can be identified.

Due to the orientation of the slit the central have been observed twice while the components located in the SW and NE are covered only for PA = . The location of components showing ambiguities in this regions of the l-v map have been determinated using the intensity ratio. For the central part both slit orientations have been used to recover the location of the components in spatial coordinates. The analysis of the l-v intensity distribution have been perfomed independently for the individual 2 D spectrograms. The fitting parameters of the gaussian curves which we have derived to reconstruct the l-v-maps of the [OIII] 5007 emission line are given in Table 1.

We measured the relative distance of the cloud complexes with respect to cloud 1, i.e. cloud D following the nomenclature of Evans et al. (1991). The location of the clouds parallel to the major axis were derived from 2 D spectrograms with PA = and the location perpendicular to the major axis were provided by 2 D spectrograms obtained at PA = . For component 1 - 5, i.e. B - F (Evans et al. 1991) and component 15 and 16 we compared the positions calculated from our 2 D spectrograms with the measured positions of the high resolution HST image of the central region of NGC 1068. The mean deviation amounts to . The identification of component 9, 10, and 11 in the southwest direction of the nucleus can be taken as reasonable as well as the identification of 6, 7, 8, and 14 northwest from the nucleus which based on the intensity ratios. The mean deviation taking these components in addition into account yields . Therefore, the components given in Table 1 can be identified within with the discrete emission line knots visible in the HST image.

The relative velocities of the fitted gaussian components range from v -1000 km s^-1 to v +950 km s^-1. The uncertainty of the position in the velocity-space which has been estimated from the independent reconstruction of the individual echelle spectra is of the order of v = (118) km s^-1. Shifts of this order produce already characteristic sharp spike-like features in the l-v-maps. The location of those components in velocity space which have been observed for both position angles are nearly identical within 15 km s^-1 (Table 1). Most of the components have a velocity dispersion of the order of 300 km s^-1 or even less given by the FWHM of the fitted gaussian curve (Table 1). The structures with a FWHM 600 km s^-1 can be ascribed more complex regions which might consist of at least two individual emission line complexes. Especially the components 3, 4, 5, 12, and 13 provide evidence for a complex sub-structure or high internal motions regarding their extension in velocity space of FWHM = (62363) km s^-1.

The l-v-map of the observed central region of the NLR of NGC 1068 provide evidence for a complex velocity field structure (Fig. 5). The emission line components in the southwest (9, 10, 11) show velocities in a range of +300 km s^-1 to +700 km s^-1 while the emission line regions in the northeast show a wide spread of velocities ranging from +500 km s^-1 to -1000 km s^-1 (6 - 8 and 12 - 16). Studies with lower spectral resolution provide some evidence for an emission line complex in the southwest with v 400 km s^-1 and FWHM 400 km s^-1 (Pecontal et al. 1997). These structure has been resolved into the components 9 - 11 based on the 2 D echelle spectra we have taken.

Fig. 5. The location of the components which have been used to deconvolve the [OIII] 5007 line profile in a rotation curve for PA = . The asterisk indicates the location of the nucleus.

Capetti et al. (1995b, 1997a) have determinated the position of the nucleus of NGC 1068 using HST polarimetry measurements of the NLR. The derived position of the nucleus can not be associated with cloud B, i.e. component 4 or with the center of the twin crescent structure described by Macchetto et al. (1994). The derived location of the nucleus indicates that the radio emission structure now appears to be more closely associated with the emission morphology of the NLR than previuosly. If the nucleus is assigned the systematic velocity of NGC 1068 it can be added to the rotation curve plot displayed in Fig. 5. It is plotted as an asterisk.

For some of the emission-line regions spectra were taken with HST (P.I. Ford). The spectra corresponding component 1 and 2 were recorded at June, 61991, spectra of component 6 and 9, 10, 11 were recorded with the HST on March, 1 1993. An HST spectrum of component 3 were observed on June, 25 1991 and of component 4 on October, 29 1990. We have retrieved the spectra from the HST archive and transformed them into velocity space in the same way as the 2 D echelle spectrograms which we have observed. The shape of the line profiles have been compared. The profiles of these HST [OIII] 5007 emission line spectra are consistent with the shape and velocities which we have derived from our 2 D echelle spectrograms.

Recently the NLR of NGC 1068 has been observed with HST in the longslit spectroscopy mode (Axon et al. 1998). The results of our reconstruction of the [OIII] 5007 line has been compared with the spectrum they have obtained for POS 3 which corresponds spatially to component 6, 7, 8 of our study (Fig. 4). They found that at the jet axis the emission line split up into two velocity systems separated by 1500 km s^-1. The region east of the jet axis shows v -1300 km s^-1 while at the west side v +150 km s^-1. The velocities we derived for component 6 and 7 are in good agreement with these measurements as well as for component 8 (see Table 1). A similar result has been obtained by Pecontal et al. (1997) who found -800 km s^-1 for cloud G. Furthermore, the radial velocity we detected for component 16 is nearly identical with the radial velocity provide by Axon et al. (1998). Axon et al. explained the large velocity split which we also found in the region of cloud G (component 6, 7, 8) as the result of the interaction of the radio jet with the NLR gas in terms of an expanding and cooling cocoon around the jet (cf. Taylor et al. 1992; Steffen et al. 1997). The interaction of the radio jet with the NLR has been already mentioned by Wilson & Ulvestad (1983) who found hints for a velocity split in the NE region of 1100 km s^-1.

In addition to the cloud G complex there is another region in the NLR showing a strong velocity split in the line profile. The features 3, 4, 5 (B, C, E following the nomenclature of Evans et al. 1991) show a steep gradient ( 30 km s^{- 1} pc^-1) of increasing velocity in north-south direction (Fig. 5, Table 1). The velocity split between component 3 and 4 amounts to 1000 km s^-1. A similar velocity split was found by several authors (e.g. Wilson & Ulvestad 1983; Cecil et al. 1990). Meaburn & Pedlar (1986) derived from their echelle spectra of the central NLR evidence for two components with v₁ -760 km s^-1 (FWHM₁ 870 km s^-1) and v₂ +340 km s^-1 (FWHM₂ 640 km s^-1). These measurements correspond well with component 3 and 4 of our investigation. We associated component 3 and 4 with the twin crescent structure mentioned by Macchetto et al. (1994). The kinematical behaviour of these components is a further indication that the nature of these cloud complexes is different from the regular NLR gas. The polarisation of this structure amounts to 100 % in the ultraviolet after correction for dilution (Capetti et al. 1995a). Therefore, it might be a reflection image of the obscured central region of NGC 1068. Due to the large velocity split it might be possible that this structure also indicates interaction of the radio jet with the NLR gas. The importance of the influence of the radio jet to the emission line gas of the NLR has been shown based on the close correspondance of the morphology even at smaller scales (e.g. Gallimore et al. 1996a,b; Steffen et al. 1997). In the radio domain is has been shown that this structure is located close to the point where the orientation of the radio jet axis changes from PA = to PA = . Since the change of the orientation of the jet axis can be expected for an interaction of the jet with a cloud these structure is suggested to be also explained within a shock model, too (Gallimore et al. 1996b).

The components 1, 2, 9, 10, 11, 15, 16 seem to follow a regular rotation with regard to the location of the nucleus of NGC 1068 (Fig. 5).

Some evidence for the distinction of the different components regarding their kinematics and even their possible excitation mechanism is provided regarding the velocity dispersion of the line emitting gas. These results suggest a picture of the NLR of NGC 1068 where shocks introduced by the radio jet - NLR gas interaction play an important role. The components 6, 7, 8, 14 which can be associated with the shock proposed by Axon et al. (1998) show a FWHM = (35060) km s^-1. The emission line complexes corresponding to component 3, 4, 5 can be characterized with FWHM (60050) km s^-1. Finally, structures whose kinematics seem to be dominated by gravitational forces show smaller velocity dispersions of FWHM (21065) km s^-1. Component 1 and 2 (i.e. D and F) show a complex morphology and might consist of more individual cloud systems. The value of the velocity dispersion is reduced to FWHM = (16510) km s^-1 taking into account only component 9, 10, 11, 15, 16.

Since the highly structured [OIII] 5007 line profile can be reconstructed with a small number of components with FWHM less than 600 km s^-1 it seems not necessary to ascribe the large profile width of [OIII] to a single blue shifted broad component of FWHM 1700 km s^-1 (Pelat & Alloin 1980). Furthermore, Cecil et al. (1990) provide evidence from Fabry-Perot interferometry study of the inner region of NGC 1068 that the profile of the [NII] emission lines is better described with individual components in the blue and red wing than with a single broad component.

Comparison with several studies of the inner region of NGC 1068 provide evidence for the identification of the components we derived from the reconstruction of the [OIII] 5007 line profile. The physical state of the NLR gas and the excitation mechanisms will be studied making use of the red part of the spectrum of NGC 1068 which we already obtained as 2 D spectrogram.

© European Southern Observatory (ESO) 1998

Online publication: September 14, 1998
helpdesk@link.springer.de