1 - INAF, Osservatorio Astronomico di Roma, via Frascati 33, 00040, Monte Porzio Catone (RM), Italy
2 - ASI Science Data Center, c/o ESA-ESRIN, via Galileo Galilei, 00044 Frascati, Italy
3 - Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, MS-29 Cambridge, MA 02138, USA
4 - European Southern Observatory, Karl-Schwarzschild-Str. 2, 85748 Garching bei München, Germany

Abstract
Aims. The multi-frequency sedentary survey is a flux-limited, statistically well-defined sample of highly X-ray dominated (i.e., with a very high X-ray to radio flux ratio) BL Lacertae objects, which includes 150 sources. In this paper, the third of the series, we report the results of a dedicated optical spectroscopy campaign that, together with results from other independent optical follow-up programs, led to the spectroscopic identification of all sources in the sample.
Methods. We carried out a systematic spectroscopic campaign for the observation of all unidentified objects of the sample using the ESO 3.6 m, the KPNO 4 m, and the TNG optical telescopes.
Results. We present new identifications and optical spectra for 76 sources, 50 of which are new BL Lac objects, 18 are sources previously referred as BL Lacs but for which no redshift information was available, and 8 are broad emission-line AGNs. We find that the multi-frequency selection technique used to build the survey is highly efficient ( $\sim$ 90%) in selecting BL Lacs objects. We present positional and spectroscopic information for all confirmed BL Lac objects. Our data allowed us to determine 36 redshifts out of the 50 new BL Lacs and 5 new redshifts for the previously known objects. The redshift distribution of the complete sample is presented and compared with that of other BL Lacs samples. For 26 sources without recognizable absorption features, we calculated lower limits to the redshift using a method based on simulated optical spectra with different ratios between jet and galaxy emission. For a subsample of 38 object with high-quality spectra, we find a correlation between the optical spectral slope, the 1.4 GHz radio luminosity, and the Ca H&K break value, indicating that for powerful/beamed sources the optical light is dominated by the non-thermal emission from the jet.

Key words: galaxies: active - galaxies: BL Lacertae objects: general - surveys

BL Lacertae objects are strong radio-loud sources that constitute a rare subclass of active galactic nuclei (AGN) distinguished by peculiar and extreme properties, namely irregular and rapid variability, strong optical and radio polarization, lack of prominent emission lines, core-dominant radio morphology, and a broad continuum extending from the radio through high-energy $\gamma$ -rays. Their broad-band spectra are characterized, in a $\nu~f_\nu~{\rm vs.}~\nu$ representation, by two emission peaks, the first located at IR/optical frequencies (but in several cases reaching the UV/X-ray band) and the second in the X-ray to $\gamma$ -ray energy band. The physical process that is believed to produce the low energy peak is synchrotron emission from relativistic electrons in the jet, while inverse Compton scattering by the same population of relativistic electrons is thought to be at the origin of the higher energy peak (e.g., Ghisellini & Maraschi 1989). BL Lac objects are often divided into two classes according to the position of the synchrotron peak energy: low-energy peaked BL Lacs (LBLs), with the peak located at IR/optical wavelengths, and high-energy peaked BL Lacs (HBLs) with the synchrotron emission peaking in the UV/X-ray energy band (e.g., Giommi & Padovani 1994; Padovani & Giommi 1995a).

A still open issue is that of the evolutionary properties of BL Lacs. It has in fact been reported (e.g. Morris et al. 1991; Stickel et al. 1991) that BL Lacs have cosmological properties different from those of FSRQs and of all other type of AGNs. Although based on a few samples with rather small sizes, LBLs have been found to be consistent with being a non evolving population (Stickel et al. 1991), while HBLs seem to show a negative cosmological evolution; i.e. they are less numerous and/or less luminous at high redshift (Morris et al. 1991; Rector et al. 2000; Bade et al. 1998).

Because of these extreme physical characteristics and of their unusual cosmological evolution, BL Lacs have been the subject of intense research activity and observation campaigns from radio to TeV energies. HBLs are exceedingly rare with a density of less than one source (with radio flux larger than 3.5 mJy) every 100 square degrees (e.g., Giommi et al. 1999). A classical approach, which requires identifying all sources in a radio-flux limited survey, would only reveal one HBL every $\sim$ 10 000 radio sources and would therefore be nearly impossible to pursue. All existing complete samples typically include less than 10 extreme HBLs, a subset far too small for any meaningful study.

The sedentary survey (Giommi et al. 1999, hereafter referred to as Paper I) introduced a new multi-frequency, highly efficient approach to the discovery of HBLs. Using this method we have been able to assemble a well-defined, radio-flux-limited, sample including 150 objects (Giommi et al. 2005, hereafter Paper II), which is currently the largest existing complete sample of high energy peaked BL Lacs.

In order to identify candidate BL Lacs in the sedentary survey, we carried out a large optical spectroscopic campaign during which we observed all the unidentified sources of the sample, therefore making the survey 100% spectroscopically identified. In this paper, we present new optical spectroscopic data for 76 objects obtained during several observing runs at the KPNO 4 m, the ESO 3.6 m and the TNG 3.6 m telescopes.

The structure of the paper is as follows: in Sect. 2 we briefly describe the sedentary survey and its selection technique, Sect. 3 discusses the results of our optical spectroscopy, Sect. 4 reviews some of the sample properties, in particular the redshift distribution. Our conclusions are summarized in Sect. 5.

Throughout this paper we have assumed cosmological parameters H₀ = 50 km s^-1 Mpc^-1 q₀ = 0. Spectral indices have been defined as $f_\nu \propto \nu^{-\alpha}$ .

2 The sedentary survey

The Sedentary multi-frequency survey was designed to select a large and statistically well-defined sample of HBLs exploiting the fact that no other known source type has been found to possess such extreme spectral energy distribution. By imposing radio, optical, and X-ray flux ratios that are only consistent with the unique spectral energy distribution of HBL sources, it is then possible to statistically select large samples of these rare sources.

In the following we briefly recall the main definition criteria used to select this sample. We refer the reader to Papers I and II for details. The sample was extracted from a large set of radio and X-ray emitting sources obtained by cross-correlating the "ROSAT All Sky Survey Bright Source Catalog'' (RASS-BSC) of soft (0.1-2 keV) X-ray sources (18 811 sources, Voges et al. 1999) with the NRAO VLA Sky Survey (NVSS) catalog of radio sources at 1.4 GHz (1 807 316 sources, Condon et al. 1998). Optical magnitudes of the sources have been obtained from Palomar and UK Schmidt surveys through the APM and COSMOS services (Yentis et al. 1992; Irwin et al. 1994).

The following conditions have been imposed in order to ensure that the sample is complete above a radio flux limit of $f_{\rm r} \ge 3.5$ mJy

Condition 1) limits the survey area to high Galactic latitude regions where absorption due to Galactic $N_{\rm H}$ is low; condition 2) imposes a very high $f_{\rm x}/f_{\rm r}$ flux ratio that, among radio loud sources, can be only reached by HBLs; condition 3) removes radio quiet sources from the sample, such as nearby Seyfert galaxies where the ratio between the unrelated radio and the X-ray flux accidentally satisfies condition 2); conditions 4) and 5) and 6) ensure statistical completeness above $f_{\rm r} \ge 3.5$ mJy.

The sample has also been updated to include a few new BL Lacs that happened to be just below the $\alpha_{\rm ro}$ threshold of 0.2 used to separate BL Lacs from emission-line Seyfert galaxies (Giommi et al. 1999). In addition, a few spurious sources were found (consistent with the $\sim$ $15\%$ expected contamination from non BL Lac objects, Giommi et al. 1999) and removed from the sample.

Although the radio flux-limited sample was spectroscopically identified only at the $\sim$ $40\%$ level when the first results were published (1999), its content was expected to include a high fraction ( $\sim$ $85\%{-}90\%$ ) of HBL. This assumption is now confirmed both by the results of massive identification campaigns of X-ray sources discovered in the RASSBSC (Beckmann et al. 2003; Bauer et al. 2000; Schwope et al. 2000) and by our spectroscopic identification of the remaining unclassified objects.

The sedentary survey sample is now completely identified and includes 150 HBLs (see Paper II). The full catalog is presented in Paper II and it is also available on-line at http://www.asdc.asi.it/sedentary/ where the broad-band spectral energy distributions and the optical finding charts are also provided.

3 Optical identification

The first identifications in the sample were obtained simply cross-correlating the precise NVSS positions with catalogs of known objects of different types. This first approach, discussed in Paper I, led to identification of 58 BL Lacs out of the original 155 HBL candidates.

In 1999 we started a systematic spectroscopic identification campaign to observe all the remaining HBL candidates or to obtain good-quality optical spectra of those objects already identified as BL Lacs but for which no redshift information was available. At the same time, Bauer et al. (2000) and Schwope et al. (2000) published the first results of the optical spectroscopic identification of bright X-ray sources in the ROSAT All Sky Survey, and Beckmann (2000) reports optical identification of part of the Hamburg-RASS bright X-ray AGN sample (later published in Beckmann et al. 2003). As expected, a significant fraction (about 40 objects) of the sedentary survey HBL candidates was found among these new classifications and thus the fraction of identified candidates in our sample significantly increased. In August 2003 our spectroscopic program was completed with the identification of all candidates leading to a final sample of 150 spectroscopically identified HBLs.

3.1 Spectroscopic observations

Spectroscopic observations of the Sedentary sources still unidentified and of BL Lacs without redshift were carried out during the period 1999-2003 at the KPNO 4 m, at the ESO 3.6 m and at the TNG telescopes. The properties of grisms used in these runs are summarized in Table 1.

Finders of all the objects observed were taken from the on-line Digitized Sky Survey (DSS) in which both the X-ray and radio error circles have been plotted. This procedure is illustrated in Figs. 1 where it is shown that the accurate NVSS coordinate ( $\la$ $3^{\prime\prime}$ ) in most cases permits a firm identification of the optical counterpart to be obtained.

Follow-up observations were made of 76 objects, including the unidentified BL Lacs candidates (58) and objects already classified as BL Lac from literature but without a redshift determination (18). In total we discovered 50 new BL Lac objects, according to the classification criteria employed in the classification scheme of Marchã et al. (1996); additionally, we determined 5 new redshifts out of 18 already known BL Lacs. The remaining 8 objects observed by us were broad emission-line AGN that were excluded from the sample (see Table 3 of Paper II).

Adding the identifications from the literature published in Giommi et al. (2005), our technique is $\sim$ $90\%$ efficient at selecting BL Lacs of the HBL type. In Table 2 we list the journal of observations for the 50 new BL Lacs discovered, for the 18 BL Lacs already known, and for the 8 broad emission lines AGNs. Columns are as follows: (1) name of the source, (2) telescope name, (3) date of observation, (4) exposure time (s).

Standard data reduction was performed using different packages in IRAF to obtain 1-dimensional wavelength-calibrated extracted spectra. The data were bias-subtracted and flat-fielded using programs in the IRAF package noao.imred.ccdred, and the spectra were extracted, wavelength-, and flux-calibrated using programs in the package noao.twodspec. A dereddening correction was applied to the data using the IRAF routine noao.onedspec.deredden and assuming Galactic values of extinction derived from 21-cm measurements (Stark et al. 1992). The spectra were wavelength-calibrated using an He-Ar (ESO), He (TNG), or He-Ne-Ar (KPNO) reference spectrum. The photometric standard stars used for the relative flux calibration are: Hiltner600 (TNG 02/2002), HR1544 (TNG 02/2002, TNG 09/2002), LTT3864 (ESO 05/2003), BD3326 (TNG 02/2003, KPNO 02/1999), CD3299 (ESO 07/2001), and LTT7379 (ESO 08/2003).

In general, we took two exposures for each object (except in few cases for which we have only one spectrum) in order to combine them and improve the signal-to-noise ratio (S/N) if necessary; this procedure allowed us to remove bad pixels and spurious features and to check the reliability of detected absorption and/or emission lines. In some cases dead pixels and cosmic rays were removed manually.

Most of the spectra were taken at parallactic angle, except in those cases where the radio/X-ray error circle contained two candidates, so a rotation of the slit was necessary. In those cases (8 objects) where we found two objects in the NVSS error box, we took the spectra of both objects. We always found that one object was a BL Lac and the other a star.

3.2 Positional information

3.3 Optical spectra

In Appendix A, we present the spectra of the optical counterparts of all BL Lacs observed by us (68). In Appendix B, for completeness, we also present the spectra of the 8 AGNs with emission lines excluded from the sample. All spectra were smoothed with a Gaussian filter of 3 pixel width.

The complete list of the observed BL Lacs objects (the 18 BL Lacs already known are marked with d), together with their properties, are given in Table 4, where the columns are as follows: (1) source name, (2) unabsorbed 0.1-2.4 keV X-ray flux; (3) NVSS radio flux at 6 cm; (4) V magnitude estimated from O and E magnitudes obtained from the APM for the northern hemisphere and from the COSMOS B_J magnitudes as given in Paper I; (5) redshift, computed, whenever possible, by taking the mean of the consistent values derived from the absorption features; and (6) optical spectral slope between the rest frame frequencies of $[{\rm O~II}] \lambda3727$ and $[\rm O~III] \lambda5007$ . The Ca H&K break value C is given in Col. (7) and was measured in spectra $f_{\lambda }$ versus $\lambda$ following (Dressler & Shectman 1987) as C = 1- f_-/f₊, where f_- and f₊ are the fluxes in the rest frame wavelength regions 3750-3950 Å and 4050-4250 Å, respectively. We have considered the Ca H&K break to have reached its minimum value of zero when $f_{-}\ge f_{+}$ . Its $1\sigma$ error was calculated based on the S/N blueward and redward of this feature. Finally, in Col. (8) we give the average S/N of the spectrum around 5500 Å measured in several $\sim$ 200 Å intervals and in (9) the $2\sigma$ upper limits on observed emission-line equivalent widths are shown. For the latter we have assumed a rectangular emission line of ${\it FWHM}=2000$ km s^-1 centered at 5500 Å.

3.4 Redshift determination

In the optical band, the spectrum of a BL Lac is made up of two main components: (i) the amplified non-thermal jet emission, which follows a power-law of the form $f_\nu \propto \nu^{-\alpha}$ , with $\alpha$ the spectral index, and (ii) thermal emission from the host galaxy, normally a luminous elliptical (e.g., Urry et al. 2000; Wurtz et al. 1996). The emission-line regions in BL Lacs are, by definition, only very weak or absent (see Sect. 3.5. for more details), which means that their redshift determination relies strongly on the detection of galaxy absorption features. This, however, is only possible if the jet is weak relative to the galaxy, i.e., only for low-luminosity BL Lacs. In strongly beamed sources, the jet with its featureless spectrum will dilute any galaxy absorption features beyond recognition (Landt et al. 2002). In the sample identified by us, the redshift was determined based on emission lines only for 8 objects (Appendix B). For most sources (41 objects) we have used the absorption features typical of ellipticals, the strongest of which are summarized in Table 5. In a considerable fraction of our objects (36% or 27/76 objects), however, we observe only a featureless spectrum for which no reliable redshift determination is possible. This fraction reduces to 23% (39/169 objects), if we consider the entire sample of HBLs (150 sources, Paper II) and emission-line AGN (19 sources, see Paper II).

In order to still be able to conduct meaningful cosmological studies with our sample, we have developed a method of determining lower limits on the redshift of sources without recognizable absorption features. The simulations of Landt et al. (2002) of low-redshift BL Lac spectra (see their Fig. 1) show that BL Lacs are expected to become featureless for jet/galaxy ratios $\ga$ 10 (defined at 5500 Å). However, the absorption features of ellipticals present at larger rest-frame wavelengths (i.e., redward of the Ca H&K break) are considerably stronger than the ones found at smaller rest-frame wavelengths (see Fig. 2), which means that the situation will be different for high-redshift BL Lacs. For these we expect the spectrum to become featureless at even smaller jet/galaxy ratios.

To determine this value we extended the simulations of Landt et al. (2002) to smaller rest-frame wavelengths. Our results are shown in Fig. 2 representative of a jet of optical spectral slope $\alpha =1$ . As soon as the Ca H&K break moves out of the "useful'' optical window (i.e., lies at observed wavelengths $\ga$ $6500~\AA$ , corresponding to redshifts of $z\ga0.65$ , where prominent telluric absorption bands dominate), a redshift determination based on absorption features is expected to be already impossible for sources with jet/galaxy ratios $\ga$ 1. Therefore, a featureless, power-law like spectrum indicates that either the BL Lac is at low redshift ( $z\la0.65$ ) and strongly beamed (i.e., its jet/galaxy ratio is high) or it is at high redshifts ( $z\ga0.65$ ) where it can be both moderately or highly beamed.

A lower limit on the redshift of featureless BL Lacs can then be determined from the estimate of their minimum jet/galaxy ratio using the fact that ellipticals have a rather constant luminosity. The jet/galaxy ratio constrains the apparent magnitude of the host galaxy from the observed total magnitude of the source, which in turn constrains the redshift. For our sample we have assumed jet/galaxy ratios of 1 and 10 and have used the relation $V_{\rm gal} = 5.10 \cdot \log{z} + 21.65$ from Browne & Marchã (1993) to estimate redshifts. If the resulting redshift for a jet/galaxy ratio =10 was higher than z=0.65, we concluded that the source was at high redshifts and that a reasonable lower limit on the redshift could possibly be derived by instead using a jet/galaxy ratio =1. This new redshift limit, however, obviously had to be $\ge$ 0.65. If this was not the case, we concluded that the Ca H&K break was at observed wavelengths $\ga$ $6500~\AA$ and chose a conservative lower limit of z=0.65. In practise this means that for sources with total apparent magnitudes $V\la18.1$ , we derived redshift lower limits assuming a jet/galaxy ratio =10, for sources with $18.1\la V \la 19.8$ we chose z=0.65, and for fainter sources we derived redshift lower limits assuming a jet/galaxy ratio=1. We applied this method only to sources with high S/N ( $\ga$ 20), high-resolution spectra (26/39 objects), since only these can be reliably classified as definitely featureless. The stardard deviation on the relation of Browne & Marchã (1993) is 0.88 in V (see their Fig. 2). This translates into an error of $\sim$ 0.10 in z. In Table 6 we list the featureless objects of our survey (39). The columns are as follows: (1) Sedentary source name; (2) the visual apparent magnitude estimated from the APM for the northern hemisphere and from COSMOS B_J magnitudes, as explained in Paper I; (3) the lower limit redshift computed with the method explained before.

As can be seen in Fig. 2, the spectral optical slope hardens with increasing jet/galaxy ratio. Unfortunately, the measured slope itself cannot be used to improve on the lower limit on z. Since we do not know z, we do not know the rest-frame wavelength and, as can be seen from Fig. 2, the amount of spectral hardening with increasing jet/galaxy ratio differs along the spectrum.

3.5 Classification

$\begin{figure} \par\includegraphics[width=7.3cm,clip]{7086fig6.ps} \end{figure}$	Figure 2: Simulated BL Lac spectra $f_{\lambda }$ vs. $\lambda$ for a jet of optical spectral slope $\alpha =1$ and of increasing flux relative to the underlying host galaxy. The assumed jet/galaxy ratios (defined at 5500 Å) are from bottom to top: 0.2, 0.5, 1, 3, 10, 15.
Open with DEXTER

We have made no effort to separate the weakly from the strongly beamed sources based on the value of the Ca H&K break (as suggested by Marchã et al. 1996; and Landt et al. 2002), since we did not want to bias our sample against low-luminosity BL Lacs. However, we discuss the Ca H&K-break value distribution of our sources in more detail in Sect. 4.2.

We have also investigated how the classification method of Landt et al. (2004) applies to our sample. We considered all the 169 sources (150 BL Lac and 19 emission line AGN) in the "HBL zone'' (see Paper II) and used the [O III] $\lambda 5007-[$ O II] $\lambda 3727$ rest-frame equivalent width plane (see Fig. 4 of Landt et al. 2004) to separate our sources into weak- and strong-lined AGN.

Landt et al. (2004) present evidence of a bimodal [O III] distribution in radio-loud AGN and define the two classes as sources with intrinsically weak- and strong [O III] emission lines, respectively. Of the 19 emission-line AGN in our sample (see Paper II; Table 3), 8 sources have been observed by us, and published equivalent width values for [O II] and [O III] are available for a additional 2 sources, namely 1RXSJ000729.3+02405 (Forster et al. 2001) and 1RXSJ122044.5+69053 (Puchnarewicz et al. 1992). We classify only two emission-line AGN, 1RXSJ122044.5+69053 and 1RXSJ222944.5-27553, as weak-lined AGN. Although no information on emission line equivalent widths could be obtained for the 9/11 sources already known, we are confident that these belong to the class of strong-lined AGN. All of these sources are classified in Bauer et al. (2000) as either Seyfert 1 or Seyfert 1.5, which means that they have strong broad emission lines, a characteristic atypical of weak-lined AGN (Landt et al. 2004). For sources observed by us without emission lines, we derived $2\sigma$ non-detection upper limits as described in Sect. 3.3., and based on these, we can classify all of these objects as weak-lined AGN (assuming the same upper limit for both [O II] and [O III]). We assume that this classification also holds for the sources from the literature without emission lines, since those spectra should have had a similar quality to ours. In summary, following the classification method of Landt et al. (2004), our sample contains 152 weak-lined AGN (which we refer to as BL Lacs) and 17 strong-lined AGN.

3.6 Notes on individual spectra

SHBL J001527.9-353639. We classify this object as BL Lac, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J003334.2-192133. This object was already identified as BL Lac by Bauer et al. (2000). We confirm its identification and found a tentative redshift of z=0.610.

SHBL J004208.0+364112. This object has already been identified as BL Lac by Bauer et al. (2000). We confirm the identification, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J012657.2+330730. We classify this object as BL Lac, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J013632.5+390559. This object has already been identified as BL Lac by Wei et al. (1999). We confirm the identification, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J032350.7-071737. We classify this object as BL Lac, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J041112.2-394143. We classify this object as BL Lac, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J050335.3-111507. We classify this object as BL Lac, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J075124.9+173051. This object was identified as Seyfert 2 by Wei et al. (1999) with redshift z=0.185. We confirm its redshift but classify it as BL Lac, as its optical spectrum does not show emission lines. The Ca H&K break for this object is located in noise, so its measurement is impossible.

SHBL J092401.1+053345. This object has already been identified as BL Lac candidate by Bauer et al. (2000). We confirm the identification of BL Lac, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J095224.0+750213. This object was identified as an early type galaxy by Bauer et al. (2000) with redshift z=0.181. We classify this object as a BL Lac with a redshift of z=0.179. The Ca H&K break for this object is located in noise so its measurement is impossible.

SHBL J095805.9-031740. We classify this object as BL Lac, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J102243.8-011302. This object has already been identified as BL Lac by Bauer et al. (2000). We confirm the identification, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J114535.1-034001. It was classified as cluster of galaxies (z=0.167) by Schwope et al. (2000); since its X-ray emission is not extended we observed it in order to investigate if this object could be a BL Lac in a cluster. From our spectrum we classify it as a BL Lac and confirm the redshift published by Schwope et al. (2000).

SHBL J124149.3-145558. This object has already been identified as BL Lac by Padovani & Giommi (1995b). We confirm the identification, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J125015.5+315559. We classify this object as BL Lac. It was observed under non-photometric conditions and is characterized by a low S/N (see Table 6).

SHBL J140630.1-393509. We classify this object as BL Lac, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J140630.2+123620. We classify this object as BL Lac, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J140659.2+164207. This object has already been identified as BL Lac by Bauer et al. (2000). We confirm the identification, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J143917.4+393243. This object has already been identified as BL Lac by Bauer et al. (2000). We confirm its identification and we also found its redshift (z= 0.344).

SHBL J144506.3-032612. This object has already been identified as BL Lac by Bauer et al. (2000). We confirm the identification, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J150340.6-154113. This object has already been identified as BL Lac by Bauer et al. (2000). We confirm the identification, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J151041.0+333504. We classify this object as a BL Lac with redshift z=0.112. After our observation in 1999, Schwope et al. (2000) classified this object as BL Lac with z=0.113, thus confirming our identification.

SHBL J151618.6-152343. This object has already been identified as BL Lac by Bauer et al. (2000). We confirm the identification, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J153311.3+185428. This object has already been identified as BL Lac by Bauer et al. (2000). We confirm its identification and we found also its redshift (z= 0.305).

SHBL J161204.6-043815. We classify this object as BL Lac, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J175615.9+552217. This object has already been identified as BL Lac by Bauer et al. (2000). We confirm the identification, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J205242.7+081038. We classify this object as BL Lac, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J213135.4-091523. This object has already been identified as BL Lac by Bauer et al. (2000). We confirm its identification and found a tentative redshift of z= 0.449.

SHBL J213151.3-251558. We classify this object as BL Lac, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J213852.5-205348. This object was classified as candidate BL Lac by Bauer et al. (2000); we confirm its identification and found its redshift (z= 0.290).

SHBL J224910.7-130002. This object has already been identified as BL Lac by Bauer et al. (2000). We confirm the identification, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J225147.3-320614. This object has already been identified as BL Lac by Bauer et al. (2000). We confirm the identification, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J230436.8+370507. This object has already been identified as BL Lac by Cao et al. (1999). We confirm the identification, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J230722.0-120518. We classify this object as BL Lac, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J231028.0-371909. We classify this object as BL Lac, but its redshift cannot be determined because of the absence of emission and absorption features.

SHBL J235023.2-243603. It was classified as cluster of galaxies (z=0.193) by Collins & Mann (1998); since its X-ray emission is not extended, we observed it in order to investigate if this object could be BL Lacs in cluster. From our spectrum we classify it as a BL Lac and confirm the redshift published by Collins & Mann (1998).

4 Sample properties

4.1 Redshift distribution

The sedentary survey redshift distribution has been de-convolved with the appropriate sky coverage. Each bin represents $\sum 1/{\rm Area}(f_{\rm x})$ for all the sources in that bin, where Area $(f_{\rm x})$ is the area accessible at its X-ray flux, divided by the total surface density of sources (Landt et al. 2001).

We compared the sedentary survey fractional redshift distribution (see Fig. 3) with various distributions from other BL Lac surveys, namely the complete DXRBS BL Lac sample (Padovani et al. 2007), the 1 Jy (Stocke & Rector 1997; Stickel et al. 1991; Stickel & Kühr 1994), and the EMSS (Rector et al. 2000) samples. The DXRBS and EMSS redshift distributions have been de-convolved with the appropriate sky coverage. Five EMSS redshifts are uncertain, while four 1 Jy redshifts are lower limits (Fig. 3). Five additional 1 Jy sources have a 0.2 lower limit on their redshift based on non-detection of their host galaxies on the optical images (Stickel et al. 1991). Note that the fraction of BL Lacs with redshift information ranges from $93\%$ and $86\%$ for the EMSS and 1 Jy samples, respectively, to $74\%$ and $71\%$ for the sedentary and the DXRBS, respectively.

$\begin{figure} \par\includegraphics[width=7.3cm,clip]{7086fig7.eps}\end{figure}$	Figure 3: Fractional redshift distribution for the 111 sedentary, 17 DXRBS, 32 1 Jy, and 38 EMSS BL Lacs. The sedentary, DXRBS, and EMSS distributions have been de-convolved with the appropriate sky coverages. The hatched areas represent lower limits (1 Jy) and uncertain values (EMSS).
Open with DEXTER

The mean redshift for the four BL Lac samples is $\langle z \rangle = 0.32$ for the sedentary, $\langle z \rangle = 0.26$ for the DXRBS, $\langle z \rangle = 0.46$ for the EMSS, and $\langle z \rangle = 0.63$ (including lower limits) for the 1 Jy. The sedentary, DXRBS and EMSS samples are peaked at z=0.3, z=0.2 and z=0.3-0.4, respectively, and neither sample includes a significant number of z>0.8 objects ( $\sim$ $1\%$ in the sedentary, $\sim$ $9\%$ in the EMSS, and $\sim$ $23\%$ in the DXRBS). By comparison, the 1 Jy BL Lacs have a somewhat surprising flat redshift distribution out to nearly z=1.5, with 10/32 1 Jy BL Lacs at z >0.8 and 5 at z>1

4.2 The Ca H&K break distribution

Figure 4 shows the Ca H&K break distribution of those HBL (61) from the sedentary survey for which we could take measurements from the literature or from our spectroscopic identifications.

Landt et al. (2002) showed that the Ca H&K break value of low-luminosity, radio-loud AGN is a suitable statistical orientation indicator and can be used to roughly separate the strongly from the weakly beamed sources. This feature is on average $\sim$ 0.5 in normal non-active ellipticals (Dressler & Shectman 1987) and is decreased by the beamed non-thermal jet emission in blazars. We measured the Ca H&K break value for our sources (in spectra plotted as $f_{\lambda }$ versus $\lambda$ ) and list these in Table 4. The error represents the 1 $\sigma$ limit and was computed based on the S/N blue ward and red ward of the feature.

$\begin{figure} \par\includegraphics[width=7cm,clip]{7086fig8.eps} \end{figure}$	Figure 4: Ca H&K break distribution of 61 sources from sedentary survey (from literature and from our spectroscopic campaign).
Open with DEXTER

Most of sedentary objects have Ca H&K break =0 or below 0.25. This property was expected, since our sample is constituted by a particular class of BL Lacs, the HBLs, characterized by a synchrotron emission peak located at high energies (UV/X-ray energy band) for which the dilution of host galaxies optical light is very high.

4.3 Optical slopes Vs. radio luminosities

$\begin{figure} \par\includegraphics[width=7cm,clip]{7086fig9.eps}\end{figure}$	Figure 5: Optical spectral slopes ( $\alpha ^{\rm oii}_{\rm oiii}$ ) for 38 sedentary objects versus their 1.4 GHz radio luminosity ( $L_{\rm r}$ ). Different symbols indicate Ca H&K break values in the ranges $C\le 0.05$ (filled circles), 0.05<C<0.25 (open squares), and $C \ge 0.25$ (stars).
Open with DEXTER

In Fig. 5 we have plotted the optical spectral slopes between the rest frame frequencies of [O II] and [O III] ( $\alpha ^{\rm oii}_{\rm oiii}$ ) of 41 sedentary objects (38 sources with spectra observed by us, 2 sources with spectra from the Sloan Digital Sky Survey (2001) and 1 object from the DXRBS, Landt et al. 2001) versus their radio luminosity ( $L_{\rm r}$ ) at 1.4 GHz using different symbols for Ca H&K break values in the ranges $C\le 0.05$ , 0.05<C<0.25, and $C \ge 0.25$ . The remaining 3 objects with redshift observed by us were excluded from this sample because their Ca H&K break is located in noisy parts of the spectra making its measurement impossible.

Landt et al. (2002) showed that the Ca H&K break value decreases with increasing jet powers, concluding that the Ca H&K break value of BL Lacs and radio galaxies is a suitable indicator of orientation. We indeed find that this also applies to the sedentary survey. As Fig. 5 shows, there is a correlation between $\alpha ^{\rm oii}_{\rm oiii}$ and $L_{\rm r}$ , reflecting the fact that for more intrinsically powerful and/or beamed sources (in the radio band), i.e. objects with stronger and/or more beamed non thermal emission, the optical light is dominated by the jet and is therefore characterized by a harder spectrum.

5 Summary and conclusions

We have presented the results of a dedicated optical spectroscopic campaign of the multi-frequency sedentary survey, a flux-limited and statistically well-defined sample of 150 high-energy peaked BL Lacertae objects. Our program, carried out with the ESO 3.6 m, the KPNO 4 m, and the TNG optical telescopes, led to the spectroscopic identification of all sources in the sample.

In this paper we have presented optical spectra for 76 sources, 50 of which are new BL Lac objects, 18 are sources previously known to be BL Lacs but without redshift determination, and 8 are broad emission-line AGNs. We determined 36 redshifts out of the 50 new BL Lacs and 5 new redshifts for the previously known objects. The redshift distribution of the complete sample is presented and compared with that of other BL Lacs samples. For 26 sources without recognizable absorption features, we calculated lower limits to the redshift using a method based on simulated optical spectra with different ratios between jet and galaxy emission.

For a subsample of 38 object with high-quality spectra, we presented the measured Ca H&K break values, and find a correlation between the optical spectral slope, the 1.4 GHz radio luminosity, and the Ca H&K break, indicating that for powerful/beamed sources the optical light is dominated by the non-thermal emission from the jet.

The main cosmological properties, such as the luminosity function and the cosmological evolution of the sample, are studied in detail in Giommi et al. (2007).

Appendix A: 68 sedentary survey spectra

The wavelength in Å is plotted on the x-axis while the y-axis gives the flux $f_{\lambda }$ in units of 10^-17 erg cm^-2 s^-1 Å^-1.

Appendix B: spectra of the 8 AGNs excluded from the sample