Astron. Astrophys. 354, 411-422 (2000)

1. Introduction

1.1. Narrow-line Seyfert 1 galaxies

X-ray and optical observations of the last decade revealed a new sub-class of active galaxies that shows a number of unusual properties which are still not well understood. The subgroup of Narrow-line Seyfert 1 galaxies (NLSy1s hereafter) was recognized by Osterbrock & Pogge (1985) based on optical properties, namely, the small widths of the lines emitted from the broad line region (BLR). Puchnarewicz et al. (1992) made the interesting observation that many optical spectra of a sample of ultrasoft X-ray AGN discovered during the Einsteinsurvey turned out to be NLSy1s, confirming the suggestion of Stephens (1989) that `X-ray selection may be an efficient way to find NLSy1 galaxies'. Many more galaxies of this type were identified in the course of optical follow-up observations of ROSATX-ray sources (e.g., Bade et al. 1995, Greiner et al. 1996, Becker et al. 1996, Moran et al. 1996, Wisotzki & Bade 1997, Grupe et al. 1998, Xu et al. 1999).

Correlation analyses performed in the last few years confirmed and quantified the trend that was already present in the study of Puchnarewicz et al. (1992): the correlation of steep X-ray spectra (measured at soft X-ray energies) with small widths of the BLR Balmer lines (e.g., Laor et al. 1994, 1997, Boller et al. 1996 (BBF96 hereafter), Brandt et al. 1997, Grupe et al. 1999a). ¹ Further, there are some correlations between the optical emission line properties in the sense that small widths of BLR lines appear to go hand in hand with strong FeII complexes and weak [OIII]/H ratios (e.g., Gaskell 1985, Puchnarewicz et al. 1992, Boroson & Green 1992, Laor et al. 1994, Lawrence 1997, Lawrence et al. 1997, Grupe et al. 1998).

A detailed analysis of the UV spectra of a number of NLSy1 galaxies was carried out by Rodriguez-Pascual et al. (1997) who detected a broad component in the permitted UV lines (FWHM km/s) but its absence in optical lines and favored an optically thin BLR à la Shields et al. (1995) as explanation. They also collected IR - X-ray fluxes and conclude that Seyferts and NLSy1s are generally very similar concerning luminosities in different energy bands except that NLSy1s tend to be underluminous in the UV.

The causes for (a) the very soft X-ray spectrum in the ROSAT energy band, and for (b) the correlations among the optical emission lines and with the X-ray properties are still under discussion. Whereas most of the spectral steepness in, e.g., the NLSy1 galaxy NGC 4051 is caused by the presence of a warm absorber, strong soft excesses have been observed in other sources (e.g., TON S180; Fink et al. 1997, Comastri et al. 1998). In particular, a model that explains in detail all properties of NLSy1s within one scenario seems to be still lacking. Several suggestions have been made to explain individual aspects, e.g., (i) a special geometry, i.e., a disk-like BLR that is viewed face-on (Osterbrock & Pogge 1985, Stephens 1989, Puchnarewicz et al. 1992) or (ii) selective absorption of the high-velocity component of the BLR by dust (Halpern & Oke 1987) to account for the small width of H; (iii) partial shielding of the NLR by a thick BLR (Boroson & Green 1992) to explain the anti-correlation of FeII and [OIII]; (iv) a removal or hindrance of a multi-phase BLR equilibrium by a steep X-ray spectrum à la Guilbert et al. (1983) (Brandt et al. 1994, see also Komossa & Fink 1997d) or (v) a scaling of BLR radius with X-ray spectral slope as in Wandel (1997) to explain the correlation of with . BBF96 tentatively favored (vi) low-mass central black holes to produce a `hot' soft excess in combination with a shielded NLR as suggested by Boroson & Green (1992). Komossa & Greiner (1995) and Komossa & Fink (e.g., 1997a,d,e) studied the possibility that (vii) the steep X-ray spectra in the ROSAT band and/or the optical high-ionization iron lines are predominantly caused by the presence of warm absorbers, and find, for the case of NGC 4051, that warm absorber and coronal line region are likely of different origin. Komossa & Janek (1999) examined the influence of various EUV-X-ray spectral shapes on the optical emission line ratios of NLSy1s.

1.2. Warm absorbers

Warm absorbers, highly ionized matter in the central region of active galaxies (AGN), are an important new diagnostic tool for investigating the conditions within the nuclei of AGN (see Fabian 1996, Komossa & Fink 1997d for overviews). The presence of an ionized absorber was first discovered in Einsteinobservations of the quasar MR 2251-178 (Halpern 1984). With the improved spectral resolution of ROSAT and ASCA, many more were found. They have been observed in 50% of the well-studied Seyfert galaxies as well as in some quasars (e.g., Pan et al. 1990, Nandra & Pounds 1992, Turner et al. 1993, Fiore et al. 1993, Mathur 1994, Done et al. 1995, Cappi et al. 1996, Ulrich-Demoulin & Molendi 1996, Komossa & Fink 1997b,c, Schartel et al. 1997a). Signatures of ionized absorbers have also been detected in quite a number of NLSy1 galaxies (e.g., Brandt et al. 1994, Pounds et al. 1994, Leighly et al. 1996, 1997, Guainazzi et al. 1996, Brandt et al. 1997, Komossa & Fink 1997a, Hayashida 1997, Iwasawa et al. 1998).

1.3. The present study

Given the enigmatic properties of NLSy1s, their detailed study is important. The sources discussed below all show some particularly extreme behavior in terms of spectral slope or variability. X-ray analyses of them were either not published previously, or with different emphasis (for details see below).

Part of the original selection criterion also was to check for the presence of a warm absorber (WA), since WAs suggest themselves as explanation for both extreme spectral steepness in the soft X-ray band and strong spectral variability. ² However, we do not only focus on this scenario. Alternatives are discussed in some detail. In particular, we examine the influence of different EUV-X-ray spectral shapes on BLR multi-phase equilibrium following the suggestion of Brandt et al. (1994).

This paper is organized as follows: The data reduction is described in Sect. 2. In the next two sections we present the general assumptions on which the data analysis is based (Sect. 3) and results for the individual objects (Sect. 4). In Sects. 5.1-5.5 we give a discussion of the properties of the individual galaxies while in Sects. 5.6-5.7 consequences for NLSy1s in general are addressed. The concluding summary is given in Sect. 6. Throughout this paper, we assume = 50 km/s/Mpc and the galaxies to follow the Hubble flow.

© European Southern Observatory (ESO) 2000

Online publication: February 9, 2000
helpdesk.link@springer.de