A&A 457, 61-70 (2006)
DOI: 10.1051/0004-6361:20065291

A 0.8-2.4 $\mu $m spectral atlas of active galactic nuclei[*]

R. Riffel1 - A. Rodríguez-Ardila2,[*] - M. G. Pastoriza1


1 - Departamento de Astronomia, Universidade Federal do Rio Grande do Sul, Av. Bento Gonçalves, 9500 Porto Alegre, RS, Brazil
2 - Laboratório Nacional de Astrofísica - Rua dos Estados Unidos 154, Bairro das Nações, CEP 37504-364, Itajubá, MG, Brazil

Received 27 March 2006 / Accepted 15 May 2006

Abstract
Aims. We present a near-infrared spectral atlas of 47 active galactic nuclei (AGN) of all degrees of activity in the wavelength interval of 0.8-2.4 $\mu $m, including the fluxes of the observed emission lines. We analyze the spectroscopic properties of the continuum and emission line spectra of the sources.
Methods. In order to exclude aperture and seeing effects we used near-infrared spectroscopy in the short cross-dispersed mode (SXD, 0.8-2.4 $\mu $m), taking the JHK-bands spectra simultaneously.
Results. We present the most extensive NIR spectral atlas of AGN to date. This atlas offers a suitable database for studying the continuum and line emission properties of these objects in a region full of interesting features. The shape of the continuum of QSOs and Sy 1's are similar, being essentially flat in the H and K bands, while a strong variation is found in the J band. In Seyfert 2 galaxies, the continuum in the $F\lambda~ \times~ \lambda$ space smoothly decreases in flux from 1.2 $\mu $m redwards in almost all sources. In J, it smoothly rises bluewards in some sources, while in others a small decrease in flux is observed. The spectra are dominated by strong emission features of H I, He I, He II, [S III] and by conspicuous forbidden lines of low and high ionization species. Molecular lines of H2 are common features of most objects. The absence of O I and Fe II lines in Seyfert 2 galaxies and the smaller FWHM of these lines relative to that of H I in the Seyfert 1 give observational support to the fact that they are formed in the outermost portion of the broad-line region. The[P II] and coronal lines are detected for all degrees of activity. The [Fe II] 12 570 Å/16 436 Å line ratio becomes a reliable reddening indicator for the narrow-line region of Seyfert galaxies.

Key words: atlases - galaxies: active - galaxies: Seyfert - quasars: emission lines - galaxies: starburst

1 Introduction

From the spectroscopic point of view, active galactic nuclei (AGNs) have been poorly studied in the near-infrared (NIR) spectral region, particularly in the interval between 1 $\mu $m and 2.4 $\mu $m. This region has been systematically absent in most surveys mainly because it does not fall within the spectral coverage of optical CCD detectors or infrared satellites (i.e., ISO, Spitzer). As a result, very little is known about the spectroscopic properties of AGNs in a transition zone that contains interesting features, in both the continuum and emission lines that can help to put firm constraints on the physical properties of the nuclear emitting gas and its environment.

With the new generation of IR arrays and their improved sensitivity, it is now possible to carry out spectroscopy at moderate resolution on faint and extended targets, such as galaxies and quasars. In addition, with the availability of cross-dispersed spectrographs offering simultaneous wavelength coverage in the interval 0.8-2.4 $\mu $m, it is now possible to study the NIR region avoiding the aperture and seeing effects that usually affect JHK spectroscopy done in long-slit mode and single-band observations.

There is manifold interest in the NIR range. At $\sim $1.1 $\mu $m (J-band), the nuclear continuum emission that dominates the UV and optical spectral energy distribution of quasars and Seyfert 1 galaxies no longer dominates (Barvainis 1987; Kishimoto et al. 2005). At the same time, reprocessed nuclear emission by hot dust starts becoming an important source of continuum emission, mainly from the K-band and longer wavelengths (Barvainis 1987; Rodríguez-Ardila & Mazzalay 2006; Glikman et al. 2006). Moreover, because the NIR is less affected by extinction than the optical, the detection of highly reddened objects with buried AGN activity, usually associated to starburst and ultra-luminous infrared galaxies, increases. A better understanding of the AGN-starburst connection can then be made. Last but not least, NIR spectroscopy on AGNs of the local universe allows the construction of spectral templates to study the commonest features and the physical processes that originate them. These templates, in turn, are essential for understands the true nature of high-redshift objects discovered using Spitzer, for instance. In this sense, Glikman et al. (2006) recently published an NIR template for AGNs, made from observations of 27 quasars in the redshift range $0.118 < {\it z} < 0.418$. They studied the emission lines in that region, revealing the Pashen series lines, as well as oxygen helium and forbidden sulfur emission.

With the above in mind, here we present the most extensive spectroscopic atlas in the 8000-24 000 Å region to date for a sample composed of 47 AGNs in the redshift range $0.0038 < {\it z} < 0.549$. It is aimed at constructing a homogeneous database for these objects at good S/N and spectral resolution, allowing the study of the continuum and line emission properties of the individual sources and the comparison of these properties among the different types of AGN. Moreover, most of the sources have no previous spectroscopic information in the literature covering the whole NIR interval. Therefore, this atlas is also intended to fill the existing gap in the SED observations of known sources and at the same time to increase the number of spectral features common to AGN that can be used to put additional constraints on the modelling of the physical properties of the nuclear gas emission.

This paper is structured as follows. In Sect. 2 we describe the sample selection, observations and data reduction process. In Sect. 3 we present the results. Comments about the main features found in the spectra are in Sect. 4. The final remarks are presented in Sect. 5. Throughout the text, a Hubble constant of 75  $\rm km~ s^{-1}$ Mpc-1will be employed.

  
2 Observations

2.1 Sample selection

The 47 AGNs that compose our sample are divided into 7 quasars, 13 narrow-line Seyfert 1 galaxies, 12 classical Seyfert 1s, and 15 Seyfert 2s. Note that the above classification was based on published optical spectroscopy of these sources made by different authors. In addition, 4 starburst galaxies were included for comparison purposes, giving a total of 51 spectra available. The dominance of type 1 objects is not by chance. Originally, we were aimed to select type 1 objects because most NIR spectroscopy published previously was done on samples dominated by Seyfert 2 galaxies/LINERS (Veilleux et al. 1997; Goodrich et al. 1994; Sosa-Brito et al. 2001) and very little was known about the NIR spectra of type 1 sources, except probably for those works on some individual sources and for the recent NIR spectroscopy on quasars (Glikman et al. 2006). Moreover, to avoid the effects of strong blending produced by the broad components of the permitted lines that could mask or dilute weak emission lines, emphasis was given to some narrow-line Seyfert 1 galaxies (NLS1). This sub-sample was selected on the basis of their singular behavior in the ultraviolet and/or soft X-ray energy bands. The list of Boller et al. (1996) was used to this purpose. We then increased our sample with classical Seyfert 1 and 2 galaxies. The selection of these objects was based on the CfA sample (Huchra & Burg 1992). Finally, our list of objects was complemented with quasars selected from the Palomar Bright quasar survey (PG) of Schmidt & Green (1983).

The main criterion in the selection of the final sample was to include, as much as possible, well-known studied sources in the optical/UV and X-ray regions that would allow us to establish correlations between the NIR emission and that in other wavelength intervals. Other criteria, such as the K-band magnitude, limited to K<12, was also applied in order to keep the exposure time under reasonable values to reach S/N>50 in the continuum emission in that band. After compiling a list of 102 AGN that matched the above conditions, objects that have a declination ${<}{-}35^{\circ}$or were already extensively studied in the NIR region, were cut out from the list. The final output was a list of 47 AGNs, plus the additional four starburst galaxies, included for comparison purposes.

Based on the above, although our sample of AGNs is not complete in any sense, we consider it as representative of the class of AGNs in the local universe (most sources have ${\it z}<0.1$), because it is composed of well-known and studied objects in other wavelength bands. Note that most of the targets have already been studied in the NIR by imaging techniques.

Columns 2 and 4 of Table 1 list the final sample of objects and the corresponding redshift, respectively. The latter value was taken from the NED database and confirmed by the position of the most intense lines in the individual spectra. Errors of less than 1% were found between our redshift determination and that published in the NED.

2.2 Observations and data reduction

The NIR spectra were obtained at the NASA 3 m Infrared Telescope Facility (IRTF) from April/2002 to June/2004. The SpeX spectrograph (Rayner et al. 2003), was used in the short cross-dispersed mode (SXD, 0.8-2.4 $\mu $m). A complete journal of observations is in Table 1. The galaxies are listed in order of right ascension. In all cases, the detector employed consisted of a 1024$\times$1024 ALADDIN 3 InSb array with a spatial scale of 0.15''/pixel. A $0.8''\times15''$ slit was employed giving a spectral resolution of 360  $\rm km~ s^{-1}$. This value was determined both from the arc lamp spectra and the sky line spectra and was found to be constant with wavelength within 3%. During the different nights, the seeing varied between 0.7''-1''. Observations were done nodding in an ABBA source pattern with typical integration times from 120 s to 180 s per frame and total on-source integration times between 35 and 50 min. Some sources were observed on multiple nights. In these cases, these data were combined, after reduction, to form a single spectrum. During the observations, an A0 V star was observed near each target to provide a telluric standard at similar airmass. It was also used to flux calibrate the corresponding object.

The spectral reduction, extraction and wavelength calibration procedures were performed using SPEXTOOL, the in-house software developed and provided by the SpeX team for the IRTF community (Cushing et al. 2004)[*]. No effort was made to extract spectra at positions different from the nuclear region even though some objects show evidence of extended emission.

The 1-D spectra were then corrected for telluric absorption and flux calibrated using Xtellcor (Vacca et al. 2003), another in-house software developed by the IRTF team. Finally, the different orders of each galaxy spectrum were merged to form a single 1-D frame. It was later corrected for redshift, determined from the average z measured from the positions of [S  III] 0.953 $\mu $m, Pa$\delta $, He  I 1.083 $\mu $m, Pa$\beta $, and Br$\gamma $. A Galactic extinction correction as determined from the COBE/IRAS infrared maps of Schlegel et al. (1998) was applied. The value of the Galactic ${\it E(B-V)}$ used for each galaxy is listed in Col. 5 of Table 1. Final reduced spectra in laboratory wavelengths, in the intervals 0.8-1.35 $\mu $m (left panels), 1.35-1.8 $\mu $m (middle panels), and 1.8-2.4 $\mu $m (right panels), are plotted in Figs. 1 to 8. Because the blue region of SpeX includes the wavelength interval 0.8 $\mu $m-1.03 $\mu $m, which does not belong to the standard J-band, in the rest of this text we will refer to that region as the z-band, following the SpeX naming convention of the different orders.

  
3 Results

3.1 Final reduced spectra

Final reduced spectra are presented from Figs. 1 to 8, sorted in the order of increasing right ascension. For each galaxy, the left panel displays the z+J bands, the middle panel the H band, and the right panel the K band. The abscissa represents the monochromatic flux in units of $\rm 10^{-15}~ erg ~ cm^{-2} ~ s^{-1}$ Å-1. For reference, we marked (dotted lines) the brightest emission lines, usually $\lbrack$S III] 9531 $\AA$, Pa$\delta $, He I 10830 $\AA$, [P II] 11 886 $\AA$, [Fe II] 12 570 $\AA$, Pa$\beta $ (left panel), [Si X 14 300 $\AA$, [Fe II] 16 436 $\AA$ (middle panel), Pa $\alpha $ H2 19 570 $\AA$, H2 21 213 $\AA$, and Br$\gamma $ (right panel). The two high redshift sources Ton 0156 and 3C 351 were drawn in a separate panel (Fig. 8) because the blue edge of their spectra starts at $\sim $5800 Å, in laboratory wavelengths.

Emission line fluxes for each object of the sample were measured by fitting a Gaussian function to the observed profile and then integrating the flux under the curve. The LINER software (Pogge & Owen 1993) was used to this purpose. The results are listed in Tables 2 to 5. We consider 3$\sigma$ level errors. For the large majority of our targets, these measurements represent the most complete lists of NIR fluxes made up to date in AGNs. The line fluxes of Mrk 1210 are already reported by Mazzalay & Rodríguez-Ardila (2006) and that of Mrk 766 in Rodríguez-Ardila et al. (2005).

  
3.2 The continuum spectra

The NIR spectra of AGN have been studied mostly via broad-band photometry. One of the most important results reported is that the continuum shape is correlated with the Seyfert type, in the sense that flatter spectral energy distributions (SEDs) tend to be found in Sy 1's and steeper ones in Sy 2's, in accordance with the unified model (e.g. Alonso-Herrero et al. 2003,2001, and references therein). However, no systematic study of the continuum characteristics in a representative sample of AGN have been made yet by means of spectroscopy. Previous works, concentrated on individual or on a small sample of objects, report a continuum described well by a broken power-law, with a flattening of the continuum slope at $\sim $1.1 $\mu $m (Thompson 1996; Rudy et al. 2000; Thompson 1995; Rodríguez-Ardila et al. 2002a; Rudy et al. 2001). It means that there would be a minimum in the continuum emission around 1.1 $\mu $m, probably associated to the red end of the optical power-law distribution associated to the central engine and the onset of the emission due to reprocessed nuclear radiation by dust (Rudy et al. 2000; Barvainis 1987). Boisson et al. (2002), in a H-band spectroscopic study of 5 AGN, report that the Sy 2 nuclear spectra are dominated by stars, while evidence for dilution of the nuclear stellar components by hot dust and/or power-law AGN are found in Sy 1.

From what is said above, the main goal of this section is to characterize the NIR continuum observed in our sample and compare it to the different types of AGNs and to other data in the literature. To this purpose we normalized to unity the continuum emission of all spectra at $\lambda$ 12 230 $\AA$, except for quasars, where the normalization was done at $\lambda$ 11 800 $\AA$. The region around these two positions are free of emission lines. The normalization point for quasars is different than for the other objects because in the former, after the spectra are converted to rest frame, the first position falls in a region of bad atmospheric transmission. In order to help in the comparison, we grouped the spectra according to the type of nuclear activity. For each type of AGN, the data were sorted according to the spectral shape, from the bluest ones (top) to the reddest ones (bottom). These plots are shown from Figs. 9 to 13.

Overall, it is easy to see from the normalized spectra (Figs. 9 to 13) that the continuum shape of quasars, NLS1s and Sy 1's are rather similar in the H and K-bands, where it is essentially flat or decreases smoothly in flux with wavelength. In contrast, in the z +J bands, the continuum shape varies from that which remains nearly flat, as in Mrk 334 and Mrk 124, to that displaying a strong blue continuum from 1.2 $\mu $m bluewards, as is the case for most quasars, such as Mrk 509 and NGC 4151. In most cases, it seems to be a break in the continuum form at $\sim $1.1 $\mu $m. At first sight, when looking at the nearby sources, one is tempted to state that the blue NIR excess is very similar in form and strength to the so called small blue bump (SBB) that is usually observed from 4000 Å bluewards in the optical spectra of Seyfert 1 galaxies and quasars. The SBB, modelled in detailed by Wills et al. (1985), was described in terms of Fe II, Mg II, and high-order Balmer lines and the Balmer continuum. However, the blue end of the NIR region does not contain that large number of Fe II emission features, as in the optical, able to create an excess of emission over the underlying continuum. Likely, the Pashen continuum and high-order Pashen lines can contribute to the NIR bump.

The quasars Ton 156 and 3C 351, which at rest wavelengths include a large portion of the optical region in the z-band, provide us with important clues for studying the actual shape and extension of the blue NIR excess. Clearly, the continuum emission in these two high-redshift sources decreases steadily in flux with wavelength with no appreciable change in the steepness from the optical up to $\sim $1.2 $\mu $m, where a clear break in the continuum shape is observed, then becoming flat redwards. This situation can be easily extended to the nearby sources, as the location of the turnover point is rather similar in most objects, being the steepness in the blue continuum the only difference among the different objects. This confirms that the break at $\sim $1.1 $\mu $m is a common characteristic of type 1 sources. What we have called the "blue NIR excess'' could simply be the red end of the continuation of the power-law optical continuum typical of type 1 sources for the NIR. Our results agrees with the findings of Glikman et al. (2006), who report a broken power-law function in the interval 0.57-2.23 $\mu $m with the breaking point at 1.085 $\mu $m, to describe the continuum of a NIR composite quasar spectrum. We should add here that type 1 sources display a nearly featureless continuum in the NIR, with only a few sources showing absorption features. In only a few sources, the 2.3 $\mu $m CO bandheads are relatively prominent. Arp 102B and NGC 1097, which were classified as Seyfert 1s because of their broad double-peaked Balmer lines, are among the type 1 sources with conspicuous stellar features. CO absorption lines are also seen in the H-band, with equivalent widths of just a few Angstroms.

The exception to all the trends for Seyfert 1 galaxies mentioned above is Mrk 1239, whose continuum emission is outstanding because it is dominated by a strong bump of emission peaking at 2.2 $\mu $m, with a strength not reported before in an AGN. In this object, the continuum does not becomes flatter at 1.1 $\mu $m, as in most Seyfert 1s but rather steep, reaching a maximum of emission at 2.2 $\mu $m and then declining again in flux with wavelength. This extreme case was the subject of a separate publication by Rodríguez-Ardila & Mazzalay (2006). They found that a blackbody of $T\sim1200$ K was needed to account for the strong excess of emission over a featureless continuum wich a power-law form. The blackbody component was interpreted in terms of very hot dust ( $T_{\rm d}=1200$ K) near its sublimation temperature, very likely located both in the upper layers of torus and close to the apex of a hypothetical polar scattering region in this object. It is worth mentioning that Mrk 766 and Mrk 478 display an emission bump similar in form to that of Mrk 1239, although wich much lower intensity.

In contrast to Seyfert 1 galaxies, none of our Seyfert 2s display the blue rise of the continuum shortward of 1.1 $\mu $m. Moreover, all objects show prominent absorption lines and bands in H and K. Indeed, the 2.3 $\mu $m CO bandheads are present in all sources but NGC 1275 and NGC 262. In J, most Seyfert 2s display an absorption band at 1.1 $\mu $m, not reported before in AGNs, and these we tentatively associated with CN (Maraston 2005). According to Maraston (2005), that band, prominent in the NIR region, is indicative of thermal-pulsing AGB stars with ages $\sim $1 Gyr. The association of young stellar population and the CN feature can be strengthened if we consider that the four starburst galaxies of the sample display this absorption (see below). The contribution of stellar population to the observed continuum is further supported by the detection of CaII triplet absorption features in the large majority of these objects.

Overall, the continuum emission of type 2 objects can be divided into two groups based on its shape: one that decreases in flux with wavelength across the NIR and that can in a first approach be approximated by a power-law function. Twelve out of 15 Seyfert 2s belong to this category. Another is dominated by a red continuum, with the flux increasing with wavelength up to 1.2 $\mu $m. From that point redwards, the flux decreases with wavelength. Two objects, namely Mrk 1066 and NGC 2110, share these characteristics. At this point we should comment on the continuum in NGC 7674, which does not fit in any of the above two categories. From 0.8 $\mu $m up to $\sim $1.4 $\mu $m, the continuum decreases in flux with wavelength as in most Seyferts 2s. In H and K, however, it displays a clear excess of emission, similar to that reported for Mrk 1239. It should also be noted that, although NGC 7674 is classified as a Seyfert 2 from its optical spectrum, in the NIR region it displays broad emission components in the permitted lines, similar to what is observed in classical Seyfert 1s.

Finally, the continuum emission of the Starburst galaxies, from 1.3 $\mu $m redwards, is rather similar for the four objects analyzed, decreasing smoothly in flux with wavelength. For NGC 3310 and NGC 7714, this same behavior is found in the blue portion of the spectrum. No upturns or breaks are found in the NIR. In contrast, the continuum in NGC 1614, is strongly reddened in the interval 0.8-1.2 $\mu $m, becoming flat in the region between 1.2-1.3 $\mu $m. Also, this source displays the most prominent absorption lines of the four galaxies. The CN absorption feature at 1.1 $\mu $m is also conspicuous in the three objects. The detection of this feature in the spectra of Seyfert 2 galaxies that display prominent circumnuclear starburst activity, such as Mrk 1066, suggest that it can be a useful tracer of young stellar populations.

We conclude this section by noting that the continuum in the NIR displays significant differences between type 1 and type 2 sources. In the former, the continuum can be characterized by a broken power-law, with the break located almost invariably at $\sim $1.1 $\mu $m. Shortwards to the break, the continuum is blue, and its steepness can be associated to the spectral index of the power-law that dominates the optical continuum emission. Redwards, the continuum is rather flat or else displays a smooth decrease in flux with wavelength. Overall, the composite power-law continuum is featureless, although absorption lines can be identified in some sources. The continuum emission of type 2 sources, on the other hand, can be grouped into two classes: one that follows a single power-law function across the NIR and another displaying a red spectrum bluewards of 1.2 $\mu $m and then decreasing steeply in flux with wavelength. The objects in the latter category display prominent absorption bands of CO and CN. They likely are dominated by circumnuclear starburst activity as told from the similarity with the spectra of genuine starburst galaxies. A quantitative approach of the analysis of the continuum emission is beyond the scope of this paper, but is left for a future publication.

3.3 The NIR emission line spectrum

The 51 NIR spectra presented in this work offer a prime opportunity for identify the most common emission features found in AGNs in a region not yet observed in such details. For completeness, the emission line fluxes of these lines, listed in Tables 2 to 5, form the largest and most complete database in the interval 0.8-2.4 $\mu $m published so far for these objects.

From our data, it is easy to see that, independent of the Seyfert class, NIR AGN spectra are dominated by strong emission features of H I, He I, He II, and [S III]. Moreover, conspicuous forbidden low-ionization lines of ions such as [Fe II], [S II], and [C I], as well as molecular H2 lines are detected in the large majority of objects. Also detected in an important fraction of the targets are coronal lines of [S VIII], [S IX], [Si VI], [Si X], and [Ca VIII]. This set of lines need ionization energies of up to 360 eV for the production of the parent ion. Their detection is considered an unambiguous signature of nuclear activity, and it increases the number of coronal line species available to study the origin, location, and physical conditions of the gas that emits them. Overall, the fluxes listed in Tables 2 to 5 can be use to add firm constraints to model the physical state of the emission gas, both from the broad line and narrow line regions.

In this section we will describe the commonest NIR emission lines detected in the galaxy sample according to the Seyfert type. To start with, and summarizing what is said in the paragraph above, Fig. 14 shows the frequency with which the most important NIR emission lines appear in the different spectra. A detailed discussion of the main spectral characteristics observed in each source can be found in Sect. 4.


  \begin{figure} \par\includegraphics[width=7cm,clip]{5291fig1.eps} \end{figure} Figure 14: Histogram showing statistics of the commonest NIR emission lines.
Open with DEXTER

  
3.3.1 Seyfert 1 galaxies

According to Osterbrock (1989), the emission-line spectrum of Seyfert 1 galaxies is characterized by optical permitted broad H I, He I and He II lines, with FWHM of order of 5000 km s-1, and narrow permitted and forbidden emission lines with FWHMs of $\sim $500 km s-1. Lines with similar characteristics are also observed in the NIR, as can be seen in Figs. 15 and 9.

We found that the forbidden [S III] $\lambda\lambda$ 9069, 9531 $\AA$ lines are present in all the Sy 1 galaxies of our sample. The permitted He I $\lambda$ 10 830 Å line is detected in 91% of the sources. H I emission lines such as Pa$\alpha $, Pa$\beta $, Pa$\gamma $ and Br$\gamma $ are common to 83% of the spectra. Moreover, exclusive BLR signatures like the Fe II and O I lines were detected in in 67% of the Sy 1 galaxies. Forbidden low ionization species were also detected in the Sy 1 spectra. The commonest are [Fe II] $\lambda\lambda$ 12 570, 16 436 $\AA$, which are present in 67% of the galaxies. Mrk 334, NGC 7469, NGC 3227 and NGC 4151 display [P II] $\lambda$ 11 886 $\AA$ line, corresponding to 33% of the Sy 1 sample. The carbon emisson line [C I] $\lambda$ 9850 Å is identified in 50% of the sources. The molecular H2 2.121 $\mu $m line is observed in 75% of the objects. Finally, the coronal line [Si VI] $\lambda$ 19 641 $\AA$ is present in 50% of the galaxies, while [Si X] $\lambda$ 14 300 Å is common in 42% of the objects.


  \begin{figure} \par\includegraphics[width=8cm]{5291fig2.eps} \end{figure} Figure 15: Histogram, showing statistics of the commonest NIR emission lines, according to each group of nuclear activity.
Open with DEXTER

3.3.2 Narrow-line Seyfert 1 galaxies

The Narrow-line Seyfert 1 galaxies are a peculiar group of Sy 1 sources first identified by Osterbrock & Pogge (1985). Among other properties, they are characterized by optical spectra displaying broad permitted lines with ${\it FWHM} < 2000$  $\rm km~ s^{-1}$ and strong Fe II emission. Our NLS1 subsample of objects, composed of 13 galaxies, is the largest set of AGN belonging to this category already observed in the NIR region and published in the literature, allowing the study of the most important emission features detected in their spectrum. Moreover, our NLS1 list is composed of well-studied objects in other spectral regions.

As can be observed in Figs. 15 and 10, the most conspicuous emission lines identified in the spectra are the first three lines of the Pashen series (Pa$\alpha $, Pa$\beta $, and Pa$\gamma $) and the He I $\lambda$ 10 830 Å line, all of which are observed in all objects. In addition, exclusive BLR features of O I, Fe II, and Ca II, free from contamination of the NLR, are common to all the NLS1 galaxies. The presence of these three features, in particular Fe II, represents a firm advantage of the NIR region compared to the optical in the study of that emission. The large number of Fe II multiplets and its proximity in wavelength in the optical leads to the formation of a pseudo-continuum that usually hampers the detection of individual Fe II lines, even in NLS1. In the NIR, the larger separation in wavelength among the different Fe II multiples in combination with the small FWHM of broad features displayed by NLS1 allows the identification of iron lines that put firm constraints on the mechanisms that creates them. This is the case, for example, for the Fe II lines located in the 9200 Å region, detected in the majority of the NLS1 sources (see Tables 2 to 3), and these are considered as primary cascading lines following Ly$\alpha $ fluorescence (Sigut & Pradhan 1998,2003).

Aside from the lines mentioned above, the forbidden [S III] $\lambda$ 9531 Å line is also detected in all the NLS1 galaxies. Other conspicuous features, such as [Fe II] and molecular hydrogen, are found in 85% and 92% of the galaxies, respectively. Three of the NLS1s, (Ark 564, 1H1934-063, and Mrk 766), display [P II] $\lambda\lambda$ 11 460, 11 886 $\AA$ lines representing 23% of the sample. The forbidden [C I] $\lambda$ 9850 $\AA$ line is clearly identified in 54% of the objects. As in the Sy 1 galaxies, the coronal lines [Si X] $\lambda$ 14 300 $\AA$ and [Si VI] $\lambda$ 19 641 $\AA$are observed in 38% of the sample.

3.3.3 Quasi stellar objects

Overall, the emission line spectrum of quasars are similar to that of Seyfert 1s and NLS1s (see Fig. 11). The only appreciable difference is in the intensity of the forbidden lines, which are weak or absent in a large fraction of the objects studied. It must be recalled, however, that the small number of targets (7) only allow us to establish trends about the frequency of the most important emission features. The advantage here is that our statistics can be compared with the results found by Glikman et al. (2006), who studied a larger sample of quasars in the NIR. We recall that the Glikman et al. sample is composed of more distant quasars than ours.

As expected, the NIR spectrum of quasars is dominated by broad permitted lines of H I, He I 1.083 $\mu $m, O I, and Fe II. These features are identified in all objects except in 3C 351, which lacks Fe II. This can be a dilution effect if we consider that 3C 351 displays extremely broad permitted lines, with FHWM reaching $\sim $12 000  $\rm km~ s^{-1}$. Any weak-to-moderate Fe II emission that broad would either be diluted in the continuum or heavily blended with nearby features turning them very difficult to isolate and identify. The lack of Fe II can also be explained on physical grounds. It is well known from the work of Boroson & Green (1992) that steep radio sources display weak or no Fe II emission and that would be the case of 3C 351.

Regarding the detection of signatures revealing the presence of a NLR, it is interesting to note that the forbidden [S III] $\lambda\lambda$9069, 9531 $\AA$ is found in all quasars. In addition, [C I] is clearly identified in 57% of the objects. The high ionization line [Si X] $\lambda$ 14 300 $\AA$ is detected in two sources (PG 1612 and PG 1126), representing a frequency of 28%. [Si VI] $\lambda$ 19 641 $\AA$ is the commonest coronal line. It was detected in 43% of the objects. Similarly, molecular hydrogen is clearly present in PG 1448, PG 1612, and PG 1126, corresponding to 43% of the galaxies. We should note that the spectra of the high redshift QSOs, Ton 0156 and 3C 351, display the presence of H$\alpha $. The measured line fluxes of these two objects are presented in Table 6.

Our results agree very closely with those reported by Glikman et al. (2006). The only lines that appear in our data, but seems to be missed in their composite quasar spectrum correspond to the [C I] and the coronal lines. This, however, needs to be looked at with caution because the spectrum that they present corresponds to a composite one instead of individual sources.

Table 6: Observed fluxes for the two high redshift QSOs in units of $\rm 10^{-15}~ erg ~ cm^{-2} ~ s^{-1}$. The fluxes of the permitted lines are the total flux of the line.

3.3.4 Seyfert 2 galaxies

The spectrum of Sy 2 galaxies is dominated by strong emission features of permitted and forbidden lines, with FWHM rarely exceeding $\sim $600  $\rm km~ s^{-1}$. By far, the strongest emission lines observed are [S III] 9531 $\AA$ and He I 1.083 $\mu $m, detected in allmost all the sources (see Figs. 15 and 12). Permitted H I is clearly identified in 87% of the objects. Low ionization lines of [Fe II] and molecular H2 are found in all spectra. Phosphorus and carbon are also identified. At least one of the phosphorus forbidden transitions either [P II]$\lambda$ 11 460 $\AA$ or [P II] $\lambda$11 886 $\AA$ is detected in 60% of the Sy 2 sample. [C I] is detected in 67% of the Sy 2s. Forbidden high ionization lines are also detected. The [Si X] $\lambda$ 14 300 $\AA$ line is common to 40% of the objects, and the [Si VI] $\lambda$ 19 641 $\AA$ line is found in 60% of the spectra. Broad permitted lines of H I were found in NGC 7674 and Mrk 993, leading us to consider that they are obscured Seyfert 1 objects. Both sources display broad emission lines in polarized lines (Miller & Goodrich 1990).

  
3.3.5 Starburst galaxies

For comparison purposes, four starburst galaxies, namely NGC 34, NGC 1614, NGC 3310, and NGC 7714 were include in our survey, been the last three genuine SB, while NGC 34 has an ambiguous classification (see Sect. 4). Their spectra are dominated by unresolved permitted lines of Pa$\alpha $, Pa$\beta $, Pa$\gamma $, Br$\gamma $ and He I $\lambda$10 830 $\AA$ (see Fig. 13). The forbidden emission lines of [S III] $\lambda\lambda$9069, 9531 $\AA$, and [Fe II] $\lambda\lambda$12 570, 16 436 $\AA$ are also conspicuous in the four objects. Molecular hydrogen lines are clearly visible in the data. No high ionization lines were found. All four objects display a remarkably similar emission line spectrum, nearly indistinguishable from each other. The continuum emission is different in NGC 1614 and NGC 34.

3.4 Reddening in Seyfert galaxies by means of NIR line ratios

The flux ratios between hydrogen emission lines are very often used as diagnostics of the reddening affecting the emitting gas of an AGN. This approach, however, is subject to large uncertainties when trying to determine the extinction in type 1 sources, either because the H I lines are strongly blended with nearby features, as in classical Seyfert 1, or because of the intrinsic difficulties in deblending the contribution from the NLR and the BLR in NLS1s. Another major problem is the fact that the intrinsic line ratios may depart significantly from Case B because of the high-density environment of the BLR and radiation transfer effects (Collin-Souffrin et al. 1982; Osterbrock 1989). The alternative is to use forbidden line ratios, but this method is very limited because the lines involved need to be from the same ion, have a large separation in wavelength, and must share the same upper limit, so that the line ratio is insensitive to the temperature over a wide range of densities been only function of the transition probabilities.

For all the above, the NIR region opens a new window to explore this issue. First, observational studies based on small samples of objects indicate that the line ratios of Pa$\beta $/Br$\gamma $ are not only comparable in both BLR and NLR but also consistent with Case B recombination, confirming that this ratio is less affected by collisional effects than by optical lines (Rhee & Larkin 2000), which is expected because the NIR lines have smaller optical depths. Second, suitable pairs of forbidden lines can be found, allowing an alternative route for determining the extinction, as is the case of [Fe II] lines at 1.257 $\mu $m and 1.643 $\mu $m (see, for example, Rodríguez-Ardila et al. 2004).

In order to see if the reddening determined by means of the H I line ratios and that found from the [Fe II] are similar, we have plotted the ratios Br$\gamma $/Pa$\beta $ vs. [Fe II] 1.2 $\mu $m/1.6 $\mu $m in Fig. 16. In the calculation, we assumed that the intrinsic ratios are 0.17 and 1.33, respectively (Hummer & Storey 1987; Bautista & Pradhan 1998). Note that for Seyfert 1 galaxies, we used the total flux of the lines to avoid uncertainties introduced in the deblending of the broad and narrow components, mainly in the NLS1. The dashed line corresponds to a reddening sequence, from E(B-V)=0(diamond with arrow) up to E(B-V)=2, in steps of E(B-V)=0.5 mag. The Cardelli et al. (1989) [CCM] reddening law was employed for this purpose.

A first inspection on Fig. 16 shows that Seyfert 2s tend to have a much narrower distribution in the [Fe II] flux ratio than do Seyfert 1s. Also, Seyfert 2s tend to lie close to the locus of points of the reddening curve, with E(B-V) in the interval 0.25-1 mag, implying that the regions emitting the H I and [Fe II] lines are affected by similar amounts of extinction. Seyfert 1s, in contrast, appear to be divided into two groups. One is populated predominantly by broad-line Seyfert 1s, which display extinction values near to zero for both ratios, and the second group, composed mostly of NLS1s, displays high values of extinction for the [Fe II] gas but close to zero for the H I region. Moreover, a few Seyfert 1s have lower Br$\gamma $/Pa$\beta $ ratios than the intrinsic Case B.


  \begin{figure} \par\includegraphics[width=9.5cm]{5291fig3.eps} \end{figure} Figure 16: Reddening diagram involving the [Fe II] 12 570 $\AA$/16 436 $\AA$ line ratio and Br$\gamma $/Pa$\beta $. Stars represent the Seyfert 1 galaxies, filled circles the Seyfert 2 galaxies, open circles represent the starburst galaxies, and filled boxes are the NLS1 galaxies of our sample. The dashed line corresponds to a reddening sequence, from E(B-V)=0 (diamond with arrow) up to E(B-V)=2. The diamonds are the theoretical values reddened in steps of E(B-V)=0.5 mag, assuming the Cardelli et al. (1989) law.
Open with DEXTER

Keeping in mind that the total flux of the Br$\gamma $ and Pa$\beta $ lines plotted in Fig. 16 for the Seyfert 1s is likely to be dominated by the one emitted by the BLR component, we propose that the lack of significant reddening for the H I gas in these objects can be explained if Case B intrinsic values are ruled out for the NIR lines, as happens in the optical region. Density and radiation transport effects modify them so that they are not a reliable source of information for the reddening. The alternative is that the region emitting the H I lines, particularly the BLR, is little or not affected by dust. This hypothesis is highly plausible, as the environment of the BLR is rather turbulent and very close to the central source making the environment unfavorable for dust grain survival.

In Sect. 3.2 we already noted that the NIR continuum within the same type of AGN was rather homogeneous, particularly in the H and K-bands, being the major appreciable difference the steepness of the continuum in the z+J band. Is that steepness related to a measurable parameter such as extinction? In order to investigate if such a relationship can be established, we plotted in Fig. 17 the reddening indicators $\rm Pa\beta /Br\gamma $ (top) and [Fe II] 12 570 $\AA$/16 436 $\AA$(bottom) vs. NIR color indices derived from the flux ratio of continuum emission integrated in windows of $\sim $100 Å. The regions chosen for integration are free of line emission contribution and are meant to be representative of the form of the continuum across the NIR region. The measured continuum fluxes are presented in Table 7.


  \begin{figure} \par\includegraphics[width=11.5cm,clip]{5291fig4.eps}\par\includegraphics[width=11.5cm,clip]{5291fig4b.eps} \end{figure} Figure 17: Plot of the reddening indicators $\rm Pa\beta /Br\gamma $ ( top) and [Fe II] 12 570 $\AA$/16 436 $\AA$ ( bottom) vs. the flux ratio of continuum emission integrated in windows of $\sim $100 Å, free from line emission contributions. Stars are NLS1 galaxies, crosses are the Sy 1, filled circles represent Sy 2s, open circles are the SB of our sample. Filled triangles are the intrinsic values of the line ratios ( $\rm Pa\beta /Br\gamma $ and [Fe II] 12 570 $\AA$/16 436 $\AA$) and the dereddened continuum ratios of the SB galaxy NGC 3310. These triangles joined by a dashed line represent the reddening curve, in steps of E(B-V)= 0.5 mag (Mod. SB). The open triangles represents a reddening sequence starting from the continuum ratios measured in Mrk 493 (Mod. Sy 1), $I\rm _c 0.9~\mu m$ represents the mean continuum in the range 9700-9800 $\rm \AA$, $I\rm_ c 1.2~\mu m$ represents the mean continuum in the range 12 230-12 330 $\rm \AA$, and $I\rm _c 2.1~\mu m$ for the range 20 900-21 000 $\rm \AA$. The measured continuum fluxes are presented in Table 7. For more details see text.
Open with DEXTER

Table 7: Mean continuum fluxes on selected ranges, in units of $\rm 10^{-15}~ erg ~ cm^{-2} ~ s^{-1}$.

As reference, we compared the observed ratios in Fig. 17 with two reddening sequences: one that starts from points representing the intrinsic values of the line ratios ( $\rm Pa\beta /Br\gamma $ and [Fe II] 12 570 $\AA$/16 436 $\AA$) and the dereddened continuum ratios taken from the SB galaxy NGC 3310, assumed to be representative of a continuum typical of a young stellar population. The value of E(B-V) for the deredening was taken from Rodríguez-Ardila et al. (2005). For this object, the dereddened continuum was then reddened in steps of E(B-V)=0.5 mag (Mod. SB, filled triangles joined by a dashed line) up to a E(B-V)=2 mag. The CCM reddening law was employed. We also plot a reddening sequence for the type 1 galaxies in Fig. 17 (open triangles connected by a solid line, Mod. Sy 1), using as zero points in the abscissa axis the continuum ratios measured in Mrk 493, an NLS1 galaxy whose continuum is considered to be affected by extinction and stellar population very little or not at all (see for example Crenshaw et al. 2002).

A first inspection to the two upper panels of Fig. 17 allows us to state that H I ratios constrain the reddening in Seyfert 1 objects poorly, while they are useful diagnostics for Seyfert 2 galaxies. This conclusion is based on the fact that the former type of objects are concentrated in the region close to the point corresponding to E(B-V)=0. The two upper plots also confirm that the continuum emission of Seyfert 1s is rather homogeneous from object to object as the ratio between the continuum and line emission display little scatter. In contrast, Seyfert 2 seems to be divided into two groups. One follows the theoretical reddening curve, suggesting that their continuum emission can be reproduced by means of a reddened starburst component, and another whose H I ratios and continuum emission seem to be dominated by emission from the central engine, as these objects share similar continuum and emission line ratios of Seyfert 1s. The two outliers, identified with the numbers 2 (NGC 34) and 34 (Mrk 291), may represent extreme cases of highly reddened sources. For instance, NGC 34 is a luminous infrared galaxy with strong water megamaser emission, suggesting both strong thermal emission by dust and starburst activity. Mrk 291 is a NLS1 galaxy whose emission line spectrum is closer to that of a Seyfert 2.

The two lower panels of Fig. 17, which involves the reddening sensitive line ratio [Fe II] 12 570 $\AA$/16 436 $\AA$, confirm that an important fraction of the Seyfert 2s of our sample display a continuum emission that is dominated by reddened stellar emission, most likely emitted by circumnuclear starburst activity. The [Fe II] provides a reliable measurement of the NLR extinction, as most points are close to the reddening sequence. Seyfert 1 galaxies display a large scatter in reddening for the NLR (measured through the [Fe II] ratio), although the slope of the continuum emission varies little. It means that the continuum emission it little or not affected by dust. The fact that some Seyfert 1s display continuum flux ratios compatible with a highly reddened starburst component may be artificial. Rather, these objects can have an important stellar contribution to the observed continuum emission, where part of the [Fe II] may be emitted.

Overall, the panels shown in Fig. 17 reveal the complex nature of the NIR continuum in AGNs, but proves to be useful in detecting objects with important starburst activity. Moreover, they show that there are strong differences in the form of the NIR continuum emission between Seyfert 1s and 2s objects.

  
5 Final remarks

We have presented the most extensive NIR spectral atlas of AGN to date. This atlas offers a suitable database for studying the continuum and line emission properties of these objects in a region full of interesting features. Ionization codes and models built to study the physical properties of AGNs need to include the constraints provided here in order to fully describe the state of the emitting gas.

The continuum and line emission properties of each subtype of active nucleus are described. In addition, we provide flux measurements of the lines detected in each of the 51 sources, distributed as follows: 12 Seyfert 1, 13 narrow-line Seyfert 1, 7 quasars, 15 Seyfert 2 and 4 starburst.

We found that the continuum of quasars, Seyfert 1s, and NLS1s are rather similar and well-described by a broken power-law. At 1.2 $\mu $m, most objects display a clear turnover in the continuum, changing from a steep blue continuum shortwards of the breaking point to one being essentially flat or nearly flat redwards. The steepness of the continuum bluewards of 1.2 $\mu $m changes from source to source and we associate it to the extrapolation of the power-law that characterizes the UV/optical continuum of type 1 sources. The exception to this trend is Mrk 1239, which displays a remarkable bump of emission over the underlying power-law, peaking at 2.2 $\mu $m. This bump is accounted for by emission from hot dust at $T\sim1200$ K. The Mrk 766 and NGC 7674 (which optically is classified as Seyfert 2) show evidence of a much weaker but similar bump. The continuum of all quasars is featureless. Some Seyfert 1s and NLS1s show evidence of underlying stellar population as told from the absorption features, mostly of CO, present in the H and K bands.

In contrast to type 1 objects, the continuum of the Seyfert 2s displays a strong young stellar component. In most objects, it decreases steeply in flux with wavelength across the NIR, similar in shape to the continuum observed in the starburst galaxies NGC 3310 and NGC 7714. Mrk 1066 and NGC 2110 are somewhat peculiar as the continuum in the z+J first increases in flux, then becomes flat for a small wavelength interval and decreases in flux from 1.4 $\mu $m redwards, resembling the continuum seen in the starburst galaxies NGC 1614 and NGC 34 . In NGC 1275, NGC 262, NGC 7674, and Mrk 1210, the continuum in the K-band rises with wavelength, suggesting the presence of hot dust. Strong absorption bands of CO were found in the H and K bands, except in the last four Seyfert 2s. The Ca II triplet in absorption, as well as the CN band at 1.1 $\mu $m are also seen in the vast majority of objects. An atlas of absorption lines and a study of the stellar populations of these galaxies will be carried out in a separate publication (Riffel et al. 2006, in preparation).

The NIR emission line spectrum varies from source to source and according to the type of activity. We found that [S III] 9531 $\AA$ and He I 1.083 $\mu $m are, by far, the strongest lines in the NIR. They were detected in all 51 objects of the sample. Neutral oxygen, permitted Fe II transitions and the Ca II triplet in emission are features seen only in type 1 sources (Seyfert 1s, NLS1s and quasars). These lines are absent in the spectra of the Sy 2, even in those objects that display in the NIR genuine broad line components in the He I lines. It confirms previous suggestions that they are exclusive BLR features (Rodríguez-Ardila et al. 2002b, and references therein). Therefore, they are useful indicators of the Seyfert type. Note, however, that the Fe II seems to be absent or rather weak in radio-loud sources.

Molecular H2, as well as [Fe II] lines is present in almost all targets, including quasars and Seyfert 1 galaxies. Moreover, H2 2.121 $\mu $m is more intense relative to Br$\gamma $ in Sy 2s than in starburst galaxies (see Figs. 12 and 13), suggesting that the AGN may play an important role in the excitation of the molecular gas in AGNs. Other NLR features worth mentioning are the forbidden high ionization lines, which were detected in the spectra of both type 1 and type 2 objects but are completely absent in the starburst galaxies. Therefore, their detection is a clear signature of AGN activity. The commonest coronal lines are [Si X] 1.43 $\mu $m and [Si VI] 1.963 $\mu $m. The presence of the coronal lines in the spectra of Sy 2 galaxies (i.e. Mrk 573, NGC 591, NGC 1275 and NGC 7674) suggests that the coronal line region is located very likely in the inner portion of the NLR.

We found that the ratio Pa$\beta $/Br$\gamma $ rules out Case B recombination values in some type 1 sources and it is very close to its intrinsic value in a large fraction of these objects. This result shows that hydrogen recombination lines are not a suitable indicator of reddening for broad-line AGN. In contrast, the flux ratio between the forbidden [Fe II] lines 1.257 $\mu $m/1.644 $\mu $m agrees, within errors, with the extinction measured by means of the Pa$\beta $/Br$\gamma $ in most Seyfert 2s, allowing us to conclude that the former can also be applied, with confidence, in type 1 objects. We also found that the steepness of the continuum in type 1 sources is not correlated with the extinction measured by means of the [Fe II] lines. It suggest that the contribution of the BLR in the NIR continuum is still larger than initially thought. In comparison, the form of the continuum in a fraction of Seyfert 2 galaxies appears to be related to the amount of extinction measured for the NLR.

Acknowledgements
This paper was partially supported by the Brazilian funding agency CNPq(304077/77-1) to ARA. This research made use of the NASA/IPAC Extragalactic Database (NED), which is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration. The authors thank the anonymous referee for useful comments about this manuscript.

References

 

  Online Material

    4 Notes on individual objects

In this section we describe the most important spectral features found in the AGN sample. It is motivated by the fact that for a large fraction of objects (44/51), no previous NIR spectra covering the $\it JHK$ bands simultaneously are available in the literature. In fact, only NGC 4151 (Thompson 1995), Mrk 478 (Rudy et al. 2001), Ark564 (Rodríguez-Ardila et al. 2002c,a), 1H1934-063 (Rodríguez-Ardila et al. 2000,2002a) , Mrk766 (Rodríguez-Ardila et al. 2005), Mrk 1210 (Mazzalay & Rodríguez-Ardila 2006), and Mrk1239 (Rodríguez-Ardila & Mazzalay 2006) have been observed before in this interval.

Table 1: Observation log and basic galactic properties for the sample.

Table 2: Observed fluxes, for type 1 objects, in units of $\rm 10^{-15}~ erg ~ cm^{-2} ~ s^{-1}$.

Table 3: Observed fluxes, for type 1 objects, in units of $\rm 10^{-15}~ erg ~ cm^{-2} ~ s^{-1}$.

Table 4: Observed fluxes, for type 1 objects, in units of $\rm 10^{-15}~ erg ~ cm^{-2} ~ s^{-1}$.

Table 5: Observed flux for type 2 and Starburst galaxies, in units of $\rm 10^{-15}~ erg ~ cm^{-2} ~ s^{-1}$.


  \begin{figure} \par\includegraphics[angle=-90,scale=0.75]{5291fig5.eps} \end{figure} Figure 1: Final reduced spectra in the Earth's frame. In the left panel we present the z+J band, in the middle panel the H band, and in the right panel the K band. The abscissa is the flux in units of $\rm 10^{-15}~ erg ~ cm^{-2} ~ s^{-1}$. The dotted lines are: $\lbrack$S III] 0.9531 $\mu $m, Pa$\delta $, He I 1.0830 $\mu $m, [P II] 1.1886 $\mu $m, [Fe II] 1.2570 $\mu $m, Pa$\beta $ ( left panel), [Si X 1.4300 $\mu $m, [Fe II] 1.6436 $\mu $m ( middle panel), Pa$\alpha $ H2 1.9570 $\mu $m, H2 2.1213 $\mu $m, and Br$\gamma $ ( right panel).
Open with DEXTER


  \begin{figure} \par\includegraphics[angle=-90,scale=0.75]{5291fig6.eps} \end{figure} Figure 2: Same as Fig. 1.
Open with DEXTER


  \begin{figure} \par\includegraphics[angle=-90,scale=0.75]{5291fig7.eps} \end{figure} Figure 3: Same as Fig. 1.
Open with DEXTER


  \begin{figure} \par\includegraphics[angle=-90,scale=0.75]{5291fig8.eps} \end{figure} Figure 4: Same as Fig. 1.
Open with DEXTER


  \begin{figure} \par\includegraphics[angle=-90,scale=0.75]{5291fig9.eps} \end{figure} Figure 5: Same as Fig. 1.
Open with DEXTER


  \begin{figure} \par\includegraphics[angle=-90,scale=0.75]{5291fi10.eps} \end{figure} Figure 6: Same as Fig. 1.
Open with DEXTER


  \begin{figure} \par\includegraphics[angle=-90,scale=0.75]{5291fi11.eps} \end{figure} Figure 7: Same as Fig. 1.
Open with DEXTER


  \begin{figure} \par\includegraphics[angle=-90,scale=0.7]{5291fi12.eps} \end{figure} Figure 8: Final reduced spectra for the two high-redshift galaxies in the Earth's frame. In the left panel we present the observed z+J band, in the middle panel the observed H band, and in the right panel the observed K band. The abscissa is the flux in units of $\rm 10^{-15}~ erg ~ cm^{-2} ~ s^{-1}$. The dotted lines are: H$\alpha $ ( left panel), $\lbrack$S III] 0.9531 $\mu $m, Pa$\delta $, He I 1.0830 $\mu $m ( middle panel), and Pa$\beta $ ( right panel).
Open with DEXTER


  \begin{figure} \par\includegraphics[angle=-90,scale=0.6]{5291fi13.eps} \end{figure} Figure 9: Plot of normalized Sy 1's galaxies spectra ordered according to their shapes from a stepeer spectrum ( top) to a flatter one ( bottom). Some emission lines are also identified. For more details see text.
Open with DEXTER


  \begin{figure} \par\includegraphics[angle=-90,scale=0.6]{5291fi14.eps} \end{figure} Figure 10: Plot of normalized NLS1's galaxies spectra ordered according to their shapes from a stepeer spectrum ( top) to a flatter one ( bottom). Some emission lines are also identified. For more details see text.
Open with DEXTER


  \begin{figure} \par\includegraphics[angle=-90,scale=0.6]{5291fi15.eps} \end{figure} Figure 11: Plot of normalized QSO's galaxies spectra ordered according to their shapes from a stepeer spectrum ( top) to a flatter one ( bottom). Some emission lines are also identified. For more details see text.
Open with DEXTER


  \begin{figure} \par\includegraphics[angle=-90,scale=0.6]{5291fi16.eps} \end{figure} Figure 12: Plot of normalized Sy 2's galaxies spectra ordered according to their shapes from a stepeer spectrum ( top) to a flatter one ( bottom). Some emission lines are also identified. For more details see text.
Open with DEXTER


  \begin{figure} \par\includegraphics[angle=-90,scale=0.6]{5291fi17.eps} \end{figure} Figure 13: Plot of normalized Starburst galaxies spectra ordered acording to their shapes from a stepeer spectrum ( top) to a flatter one ( bottom). Some emission lines are also identified. For more details see text.
Open with DEXTER


Copyright ESO 2006