ESO-VLT optical spectroscopy of BL Lac objects: II. New redshifts, featureless objects and classification assessments. B. Sbarufatti1, A. Treves Universita` dell’Insubria, Via Valleggio 11, I-22100 Como, Italy 6 0 R. Falomo 0 2 INAF, Osservatorio Astronomico di Padova, Vicolo dell’Osservatorio 5, I-35122 Padova, Italy n a J. Heidt J 3 Landenssterwarte Heidelberg, Ko¨nigstuhl, D-69117 Heidelberg, Germany 2 1 J. Kotilainen v Tuorla Observatory, University of Turku, Va¨isa¨la¨ntie 20, FIN-21500 Piikkio¨, Finland 6 0 5 R. Scarpa 1 0 European Southern Observatory, 3107 Alonso de Cordova, Santiago, Chile 6 0 / h p - ABSTRACT o r We report on ESO Very Large Telescope optical spectroscopy of 42 BL Lacertae objects of t s unknown redshift. Nuclear emission lines were observed in 12 objects, while for another six we a detected absorptionfeatures due to their host galaxy. The new high S/N spectra therefore allow : v us to measure the redshift of 18 sources. Five of the observed objects were reclassified either as i starsorquasars,andone is ofuncertainnature. Forthe remaining18the opticalspectraappear X without intrinsic features in spite of our ability to measure rather faint (EW ∼0.1 ˚A) spectral r a lines. For the latter sources a lowerlimit to the redshift was setexploiting the veryfact that the absorption lines of the host galaxy are undetected on the observed spectra. Subject headings: BL Lacertae objects: general 1. Introduction grouped together in the blazar class. From the spectroscopical point of view BL Lacs are char- BL Lac objects (hereinafter BL Lacs or BLL) acterized by quasi featureless optical spectra. In are active galactic nuclei (AGN) characterized facttheirspectraareoftendominatedbythenon– by luminous, rapidly variable UV–to–NIR non– thermal continuum that arises from the nucleus. thermal continuum emission and polarization, Tothisemissionitissuperimposedathermalcon- strong compact flat spectrum radio emission and tribution due to the stellar componentofthe host superluminal motion. Similar properties are ob- galaxy. Like in other AGN, emission lines could servedalsoinflatspectrumradioquasars(FSRQ) be generated by fluorescence in clouds surround- and the two types of active nuclei are often ingthecentralblackhole. Moreover,asithappens for high z quasars in some cases absorption lines 1alsoatUniversita`diMilano-Bicocca due to intervening gas in the halo of foreground 1 galaxiescanbeobservedinthespectraofBLLacs perwe adoptedthe followingcosmologicalparam- and one can derive a lower limit to the redshift of eters: H = 70 km s−1 Mpc−1, Ω =0.7, Ω =0.3. 0 Λ m the object. The detectability of spectral features dependsonthebrightnessofthenuclearsource: in 2. The sample factduringlowbrightnessstates,intrinsicabsorp- ThesampleofBLLacobjects(andcandidates) tion features can be more easily revealed, while observed with the VLT telescopes was selected during high states one can better discover inter- from two extended lists of BL Lacs: the Padovani vening absorption systems. Because of the strong & Giommi (1995a) collection of BL Lacs and the contribution from the continuum the equivalent Sedentary Survey (Giommi et al. 1999, 2005, in width (EW) of all these spectral features is of- the following addressed as SS). The Padovani & ten very small and their detection represents a Giommi(1995a)listcontainsallobjects identified challenging task. asBL Lacsbelonging to the complete samples ex- In the past decade a number of projects were isting at the time of its compilation, selected in carried out to derive the redshift of BL Lac ob- the radio, optical and X-ray bands (e.g.: 1 Jan- jects. Most of these works were based on op- sky survey – 1-Jy, Stickel et al. (1991), Palomar- tical spectra collected with 4 m class telescopes, Greensurvey–PG,Greenetal.(1986),Extended and are therefore limited by relatively low signal- Medium Sensitivity Survey – EMSS, Gioia et al. to-noise ratio (S/N), low spectral resolution and (1990),Slewsurvey,Perlmanetal.(1996),White- limited wavelengthrange (e.g.Falomo et al.1993; Giommi-Angelini catalogue – WGA White et al. Stickel & Ku¨hr 1993; V´eron-Cetty & V´eron 1993; (1994)). ItincludesalsosourcesfromtheHewitt& Bade, Fink, & Engels 1994; Falomo, Scarpa, & Burbidge(1993)andV´eron-Cetty&V´eron(1993) Bersanelli1994;Falomo 1996;Marcha˜et al. 1996; catalogues (in the latter case we checked that the Drinkwater et al. 1997; Laurent-Muehleisen et al. source was still included in the 2001 version), for 1998; Landt et al. 2001; Rector & Stocke 2001; a total of 233 objects. The criteria used to define Londish et al. 2002; Carangelo et al. 2003; Hook a BL Lac object in Padovani & Giommi (1995a) etal.2003). Recently,however,someobservations dependonthesampleoforigin. Inmostcases,the with 8 m class telescopes were carried out (Heidt EWofthelinesisrequiredtobe≤5˚A,butalsoUV et al. 2004; Sowards-Emmerd et al. 2005). De- excess, optical polarization and variability, radio- spite these efforts, a significant fraction of known to-optical spectral index are used as selecting cri- BL Lacs (e. g. 50 % in V´eron-Cetty & V´eron teria. The SS was obtained cross-correlating the (2003) catalogue) have still unknown redshift. National Radio Astronomy Observatory (NRAO) In order to improve the knowledge of the red- VeryLargeArray(VLA)SkySurvey(NVSS)data shift of BL Lacs we carried out a project to ob- (Condon et al. 1998) with the ROSAT All Sky tain optical spectra of sources with still unknown Survey–Bright Source Catalogue (RBSC) list of or uncertain redshift using the European South- sources (Voges et al. 1999). SS selected a com- ernObservatory(ESO)8-mVeryLargeTelescopes plete sample of 150 High energy peaked BL Lacs (VLT). This allows one to improve significantly (HBL, see Padovani & Giommi 1995b, for defini- the S/N of the spectra and therefore the capabil- tion) down to a 3.5 mJy radio flux limit. BL Lac ity to detect faint spectral features. A first report classification in the SS is based on the position of onthis work,givingthe redshiftof12objects,has the sources on the α −α plane. OX RO been presented by Sbarufatti et al. (2005a, Paper The Padovani & Giommi (1995a) and SS I), and here we refer on the results for the full datasets lead to a combined list containing 348 sample of 42 observed sources. objects. The distribution of the V magnitude for The outline of this paper is the following. In these objects is reported in Fig. 1. The bulk of section 2 we give some characterization of the 42 themhaveVbetween15and20,andthe the frac- observed objects. The observation and analysis tion of objects with unknown redshift increases procedures are described in section 3. In sections with the apparent magnitude and reaches ∼ 50% 4 and 5 we report the results of our spectroscopic at the faintest magnitudes. Note, however, that study. Finallyinsection6asummaryandconclu- also at V ∼ 15-17 about 20% of the sources have sions ofthis study aregiven. Throughoutthis pa- 2 not known redshift. The total number of objects Although this programdid not require optimal with unknown redshifts is 105. photometric conditions, most of the observations Fromthecombinedlistweselectedsourceswith were obtained with clear sky. This enables us to δ <+15◦, for observability from the VLT site. perform a spectrophotometric calibration of the Moreover to grant a sufficiently high S/N level of acquireddatausingstandardstars(Oke1990)ob- the optical spectra we required V<22. Thus we served in the same nights. From the database of gathered a list of 59 objects. During three obser- sky conditionsatParanalweestimate thata pho- vationalcampaigns,performedinservicemode,we tometric accuracy of 10% was reached during our completed this optical spectroscopy program, ob- observing nights. The spectra were also corrected taining data for ∼70%ofthe sample (42 sources). for Galactic extinction, using the law by Cardelli, Our sample is similar to the parent sample of 348 Clayton, & Mathis (1989) and assuming values of objectsintermsofmeanapparentmagnitudeand EB−V from Schlegel, Finkbeiner, & Davis (1998). subdivisioninLow(LBL)andHighenergypeaked BL Lacs. 4. Results In Fig. 2 we give the optical spectrum of each 3. Observations and data analysis source. In order to show more clearly the contin- Optical spectra were collected in service mode uum shape and the faint features we report both with the FOcal Reducer and low dispersion Spec- the flux calibrated and the normalized spectrum trograph(FORS1,Appenzeller et al.1998)onthe for each object. The main emission and absorp- VLT. The observations were obtained from April tion features are identified. Those due to the 2003 to March 2004 with UT1 (Antu) and from galactic interstellar gas are indicated as “ISM” April to October 2004 with UT2 (Kueyen). We and“DIB”(Diffuse InterstellarBands,seesection used the 300V+I grism combined with a 2” slit, 4.2.3),while telluric absorptionsaremarkedas⊕. yieldingadispersionof110˚A/mm(corresponding to 2.64 ˚A/pixel) and a spectral resolution of 15– 4.1. The continuum emission 20 ˚A covering the 3800−8000 ˚A range. The see- Inafirstapproximation,the opticalcontinuum ingduringobservationswasintherange0.5−2.5”, of a BL Lac object is due to the superposition of withanaverageof∼1”. Relevantinformationson two components: the non-thermal emission of the the sample objects are given in Table 1. active nucleus, Doppler-enhanced because of the DatareductionwasperformedusingIRAF1(Tody alignment of the jet with the line of sight, and 1986, 1993) following standard procedures for theemissionofthehostgalaxy. Dependingonthe spectral analysis. This includes bias subtraction, relativestrengthofthenucleuswithrespecttothe flat fielding and cleaning for bad pixels. For each galaxy light, the spectral signature of the latter target we obtained three spectra in order to get can be either easily detected or diluted beyond a good correction of cosmic rays and to check the the point of recognition. Taking into account the reality of weak features. The individual frames robust evidence that the host galaxies are giant were then combined into a single average image. ellipticals (e.g. Urry et al. 2000), to describe the Wavelength calibration was performed using the continuumandderivetheopticalspectralindexof spectra of a Helium/Neon/Argon lamp obtained thenon-thermalcomponent,wefittedapowerlaw during the same observing night, reaching an ac- (F ∝λ−α,the spectralindicesaregiveninTable λ curacy of ∼ 3 ˚A(rms). From these images we 1) plus the spectrum of a typical elliptical galaxy extracted one-dimensional spectra adopting an asdescribedbythe Kinneyetal.(1996)template. optimal extraction algorithm (Horne 1986) to While in most cases the contribution of the host improve the S/N. galaxywas negligible, in 6 sources it was not, and theluminosityofthehostcanthusbederived. For 1IRAF (Image Reduction and Analysis Facility) is dis- these six sources (three of them were presented tributedbytheNationalOpticalAstronomyObservatories, in Paper I) we give the best fit decomposition in which are operated by the Association of Universities for ResearchinAstronomy,Inc.,undercooperativeagreement Fig.3 and report the parameters in Table 2. The withtheNationalScienceFoundation. derived absolute magnitudes of the host galaxies 3 are consistent with the distribution of M of BL 1319+019),while object0841+129remains ofun- R Lac hosts given by Sbarufatti et al. (2005b). certain nature. 4.2. Spectral features and redshifts 4.2.3. Lineless BL Lacs. The detection and the measurement of very InspiteofthehighS/N18objectsexhibitspec- weakspectralfeaturesisdifficulttoassessbecause tra lacking any intrinsic feature. In several spec- itdependsonthechoiceoftheparametersusedto tra we clearly see absorptionfeatures fromthe in- define the spectral line and the continuum. In or- terstellar medium (ISM) of our Galaxy. In par- der to apply an objective method for any given ticular, we are able to detect CaII λλ3934,3968, spectrum we evaluate the minimum measurable NaI λ5892 atomic lines, and a number of DIBs equivalent width (EW ) defined as twice the λλ 4428,4726,4882,5772, generated by complex min rms ofthe distributionofallEWvalues measured molecules in the ISM (e.g. Galazutdinov et al. dividing the normalized spectrum into 30 ˚A wide 2000, and references therein). In Fig. 4 we re- bins (details for this automatic routine are given port the average spectrum of the interstellar ab- inPaperI).WecheckedthattheS/Nratiodepen- sorptions. In three cases absorptionlines from in- denceinsidetheconsideredspectralrangevariesat tervening gas are detected, leading to lower limits mostby20%,remaining<10%overalargewave- on the redshift of the objects (0841+129,z>2.48; lengthrange. This reflects into a similarvariation 2133–449,z>0.52; 2233–148,z>0.49). of EWmin. The procedure for calculating EWmin For these 18 sources we have estimated a red- was applied to all featureless or quasi-featureless shiftlowerlimitbasedontheEW oftheirspec- min spectra to find faint spectral lines. All features traandtheapparentmagnitudesofthenuclei. We above the EWmin threshold, ranging from ∼ 1 ˚A reportthese in Table 1. The procedure to obtain to 0.1 ˚A in our data, were considered as line can- these limits is described in section 4.2.4. didates and were carefully visually inspected and measured. The results are summarized in Tab. 1. 4.2.4. Redshift lower limits procedure. Based on the detected lines and the shape of the Inthissectionwedescribetheproceduretoob- continuumweconfirmtheBLLacclassificationfor tainredshiftlowerlimits forBL Lacswithlineless 36 objects, while 6 sources were reclassified. De- spectra(seeTable1)fromtheEW ofthespec- pending on the observed spectral properties the min trum and the observed magnitude of the object. objects can be assembled in three groups. Under the assumption that the host galaxy lumi- nosityisconfinedinanarrowrange(Sbarufattiet 4.2.1. Confirmed BL Lacs with measured z. al. 2005b) it is in fact possible to constrain the Twelve objects belonging to this group were position of the source on the nucleus-to-host flux reported in paper I. Six more are presented here ratio (ρ) vs redshift plane. (Table 1). Three have redshift derived from emis- We assume that the observed spectrum of a sionlines(0723–008,z=0.128;2131–021,z=1.284; BL Lac object is given by the contribution of two 2223–114, z=0.997) and three from absorption components: 1- a non-thermal emission from the lines (1212+078, z=0.137; 1248–296, z=0.382; nucleus that can be described by a power law 2214–313, z=0.460). Details on each source are (F(λ)=Cλ−α,whereC isthe normalizationcon- given in section 5. stant); 2 - a thermal component due to the host galaxy. Depending on the relative contribution of 4.2.2. Misclassified and uncertain nature ob- the two components the optical spectrum will be jects. dominated by the non-thermal (featureless) emis- Despite their classification as BL Lac objects sionorbythespectralsignatureofthehostgalaxy. in one or more input catalogues, six sources Theobservedequivalentwidth(EWobs)ofagiven havespectraincompatiblewiththisidentification. spectral absorption line is diluted depending on Five of them were reclassified either as quasars theratioofthetwocomponents. Detectionofthis (0420+022,1320+084)orstars(1210+121,1222+102, spectral feature requires a spectrum with a suffi- cientlyhighS/N.ThisisillustratedinFig5,where 4 a simulated spectrum (ρ=5, z=0.5) is reproduced Ontheotherhandthequantityρ dependsalso 0 withtwodifferentS/Nratios. TheS/N=300spec- on the observed magnitudes of the object, since trum grants a secure detection of the CaII fea- tures,whilewithS/N=30thelinesareundetected. log(ρ0)=−0.4[Mn(z)−Mh(z)] (4) In order to estimate the redshift of an object where M is the nucleus absolute magnitude and fromtheEW weneedtoknowthe relationbe- n min M is the host absolute magnitude, and tween EW and the nucleus-to-hostflux ratio ρ. h obs For a spectral absorption line of intrinsic equiva- M (z)=m +5−5logd (z)−k (z) (5) n n l n lent width EW the observed equivalent width is 0 given by the relation (see also Sbarufatti 2005): where m is the nucleus apparent magnitude, n d (z) is the luminosity distance and k (z) is the L n (1+z)×EW 0 nucleusk-correction,computedfollowingWisotzki EW = (1) obs 1+ρ×A(z) (2000). The absolute magnitude of the host is The nucleus-to host ratio ρ can be represented M (z)=M∗ −E(z) (6) by h h ∗ whereM =–22.9istheaverageRbandmagnitude F(λ) h ρ(λ)= (2) of BL Lac hosts at z=0 and E(z) is the evolution G(λ) correction, as given by Bressan et al. (1998). where G(λ) is the giant elliptical spectral tem- An example of the procedure described above plate by Kinney et al. (1996,see also section 4.1), isgiveninFig. 8,wheretherelationshipsbetween and A(z) is a correction term that takes into ac- log(ρ0)andtheredshiftforagivenvalueofEWmin countthelossoflightinsidetheobservedaperture. and mn are shown. The intersection of the two Inthis workthe apertureis a2”×6”slitthatcap- curves yields a lower limit to the redshift of the tures&90%ofthenuclearlight,butnotthewhole target. Whenitgoesbeyondtheobservedspectral surroundinggalaxythatismoreextendedthanthe range, we set the redshift limit to the value cor- aperture(inparticularforlowztargets). Inorder responding to the considered feature reaching the to estimate this effect we evaluated the amount upperlimitoftheobservedwavelengthrange(z∼1 of light lost from the galaxythrough the aperture in the case of CaII λ3934 line). The uncertainty in use from simulated images of BL Lacs (point of this procedure depends mainly on the spread source plus the host galaxy). The main parame- of the distribution of the host galaxy luminosity. tersinvolvedaretheshapeandthesizeofthehost This issueisdiscussedinUrryetal.(2000)andin galaxy. According to the most extensive imaging Sbarufatti et al. (2005b), where it is shown that studies of BLL (Falomo 1996; Wurtz et al. 1996; the 64 BL Lacs hosts of known redshift resolved Falomo&Kotilainen 1999;Heidtetal.1999;Nils- with HST are well represented by an elliptical of sonet al.2003;Urry et al.2000)we assumedthat MR=-22.9,with 68% of them in the interval -23.4 the host galaxy is a giant elliptical of effective ra- – -22.4. dius R = 10 kpc. The fraction of starlight lost This procedure can be used for any absorption e then depends on the redshift of the object and is line belonging to the host galaxy and for which particularly significant at z< 0.2, producing the an estimate of the un-diluted EW is available. In bending of the curves in Fig. 6. thisworkweconsideredtheCaIIabsorptionlineat Since we want to refer the observed equivalent λ=3934 ˚A(EW0=16 ˚A), we assumed a power law width to the nucleus-to-hostratio ρ =ρ(λ ) at a spectral index α=0.7 (Falomo et al. 1993), and 0 0 fixedwavelengthλ ,equation(1)canberewritten we referred to the effective wavelength of the R 0 as: band (λ0=6750 ˚A) to compute ρ0 (which implies (1+z)×EW0 ∆=4.3). EW = (3) obs 1+ρ0×∆×A(z) In order to test this procedure we considered eight BL Lacs for which the CaII line of the host where∆(λ)isthenucleus-to-hostrationormalized galaxy has been measured. Five of these objects to that at λ (∆(λ)=ρ(λ)/ρ(λ ); see Fig. 7). 0 0 derive from the observations discussed here and 5 in paper I, three others are from observations ob- (1993) catalogue. Falomo (1996) detected the tainedattheESO3.6(Carangeloetal.2003;Sbar- host galaxy with ground-based imaging, propos- ufatti 2005). ThesespectraarereportedinFig. 9 ing z∼0.2–0.3. The optical spectrum taken by and the relevant parameters are given in Table 4. Falomo,Scarpa,&Bersanelli(1994)isfeatureless. The comparison between the redshifts estimated Our spectrum (S/N=230)does not showevidence by our procedure with the spectroscopic ones in- for intrinsic spectral features from the host, sug- dicatesareasonablegoodagreement(seeFig. 10). gesting a very high N/H ratio. Interstellar ab- sorptions from NaI λ5892 and DIBs at 5772 and 5. Notes to individual objects. 4726 ˚A are well detected. Based on EW , we min estimate z>0.31. 0047+023 This compact and flat spectrum ra- dio source was classified as a BL Lac by Hewitt 0627–199 Hooketal.(2003)obtainedalineless & Burbidge (1993) on the basis of UV color and spectrum for this radio selected BL Lac object. featureless spectra. Further featureless optical Our VLT spectrum, of moderate S/N (50), shows spectraobtainedbyAllington-Smithetal.(1991); no spectral features. From EW we set z>0.63. V´eron-Cetty & V´eron (1993) confirmed the BL min Lac. Even in our S/N∼ 80 spectrum no spectral features were found. Based on the minimum de- 0723–008 Wills & Wills (1976) classified this tectable EW the sourceis most likely at z > 0.82. source as a Narrow Emission Line Radio Galaxy based on an optical spectrum, giving z=0.127. Rusk & Seaquist (1985) report an optical polar- 0048–097 Previous optical observations of this ization of 1.5 %, classifying the source as a Low well known BL Lac object belonging to the 1-Jy Polarization QSO. V´eron-Cetty & V´eron (2001) sample,reportedafeaturelessspectrum(Stickelet reportthissourceasaBLLacobject. Henriksenet al.1991;Falomo,Scarpa,&Bersanelli1994). Rec- al. (1984) gives broad band indices α =0.7 and tor & Stocke (2001),however,suggested the pres- RO enceofanemissionlineat6092˚A(possiblyidenti- αOX=1.0, which are compatible with a BL Lac or a FSRQ classification. Our optical spectrum fiedwith[OII]λ3727atz=0.634or[OII]λ5007at is clearly dominated by a non thermal emission z=0.216). Falomo (1996) proposed z>0.5, based with spectral index α=0.7. Superposed to this, on the non detection of the host galaxyin the op- strong narrow emission lines and absorption lines tical images of the source. Our S/N=250 opti- from the underlying host galaxy at z=0.127 are cal spectrum does not confirmthe presence of the emissionlineat6092˚A,andapartofsometelluric detected, confirming the redshift. From the val- ues of the spectral indices and the measured EW lines and a number of Galactic absorptions it is for the spectral lines, we suggest that this object found featureless. From our EW estimate, we min is of intermediate nature between a BL Lac and a infer that this source is at z>0.3. quasar. 0420+022 Fricke et al. (1983) classified this 0841+129 This source, first identified by C. source as a BL Lac candidate on the basis of a Hazard (see Jaunsen et al. 1995, and references featureless (although noisy) optical spectrum. El- therein),isaDampedLymanαAbsorption(DLA) lison et al. (2001) through an unpublished optical QSOatz>2.48asderivedfromthetwoverystrong spectrum propose a redshift z=2.277 and classify DLAsat∼4100and∼4225˚A(seeforexamplePet- the source as a radio loud QSO. In our optical tinietal.1997;Prochaskaetal.2001;Warrenetal. spectrum we are able to clearly detect emission 2001, and references therein). The classification linesLy λ1419OVI]λ1034,CIVλ1549andCIII] α as a BL Lac object was motivated by the absence λ1909, at z = 2.278. A recent spectrum obtained of prominent emission lines (Hewitt & Burbidge byHooketal.(2003)alsoconfirmourfindingsand 1993). the classification of the object as a QSO. Ourspectrum,inadditiontoseveralabsorption lines, exhibit three possible broad emission struc- 0422+004 Thisobjectisawellknownradiose- tures at∼4310, ∼4850 and ∼5370 ˚A. These could lectedBLLac,includedinthe Hewitt&Burbidge 6 be interpreted as NV λ1240, SiIV λ1397 and CIV vation of a featureless spectrum. Burbidge (1996) λ1549 at z∼2.47. This is consistent with z∼2.5, estimatesthisobjectapossiblecandidateofexpul- deduced from the observed position of the onset sionfromagalacticnucleus. Thesharpabsorption of the absorption of the Ly forest Warren et al. lines detected in our spectrum clearly indicate a α (2001). An alternative explanation is, however, stellar origin. The measured colors lead to a tem- that these structures are pseudo-emissions result- perature of ∼10000 K. If the object were a main ing from the depression of the continuum caused sequence or a supergiant star, the corresponding by the envelope of many unresolved narrow ab- distance willput it outside the Galaxy,but notat sorptionfeatures. Higherresolutionspectraofthe the distance of NGC 4380. We are therefore led object in the spectral range 4200 to 5800 ˚A are toconsiderawhitedwarf,whichwouldbeat100– needed to distinguish between the two possibili- 200 pc. The absence of H lines indicates a DQ ties. or DXP white dwarf (Schmidt et al. 2001, 2003). SomeofthelinesarereferabletoHeIandCItran- sitions. The object clearly deserves further study; 1210+121 Hazard&Murdoch(1977)proposed in particular polarizationmeasurements would be that this object was the optical counterpart of interesting. a radio source in the Molongo Catalogue (MC2 Sutton et al. 1974); the separation was however 16”. Zotov & Tapia (1979) reported large optical 1248–296 Perlman et al. (1996) obtained a low variability and polarization, apparently reinforc- S/N spectrum of this source, and proposed a BL ing the identification. Baldwinetal.(1973)found Lac at z=0.487based on the possible detection of afeaturelessopticalspectrum. OurVLTspectrum the host galaxy features. In our VLT spectrum clearly shows that the source is a type B star in CaII,Gband,H areclearlydetectedatz=0.382, β our Galaxy. confirmingthefindingsofWooetal.(2005),while inthe blue partthe contributionofanon-thermal component is clearly visible. The best fit decom- 1212+078 OurVLTspectrumclearlyshowsthe positiongivesα=0.92forthe non-thermalcompo- presence of a strong thermal component due to nent visible below 5000 ˚A, and M =-22.7 for the the host. We detected CaII λλ3934, 3968, G R host, in good agreement with result from the di- band λ4305, MgI λ5175 and H λ6563 in emis- α rect detection of the host in HST imaging (Urry sionatz=0.137,confirmingtheredshiftestimated et al. 2000; Sbarufatti et al. 2005b). by Perlman et al. (1996). The contribution of thenon-thermalcomponentisvisibleinthebluest part of the spectrum. The best fit decomposition 1319+019 This object was initially selected as of the spectrum gives α=1.17 for the non-thermal a BL Lac candidate onthe basis of the University component and M =-22.0 for the host. Though of Michigan objective prism survey (MacAlphine R this is somewhat fainter than expected for a BL & Williams 1981) designed to find AGN and it is Lac host galaxy, we can not exclude that given included as BL Lac in the V´eron-Cetty & V´eron the low redshift and the consequent large appar- (2001) catalogue. No radio counterpart for this ent size of the host, partofthe light did not enter sourcehasbeenfoundinliterature. LaterThomp- in the slit. son&Djorgovski(1990)proposeditsclassification asaBLLac,basedonalowS/Nopticalspectrum that was found featureless. In our much better 1222+102 This is a blue stellar object in the quality spectrum we clearly see many absorption direction of the Virgo-Coma cluster. Its apparent features that characterize the object as a galactic position in the sky is very close to the center of star of spectral type ∼A. Our findings are also in the galaxy NGC 4380, still well inside the galaxy agreement with the spectral classification of the boundaries. The projected separation to the nu- cleusattheredshiftofthegalaxyis∼10kpc. The 2dF QSO Redshift survey (2QZ, see Croom et al. 2004). object is considered a candidate BL Lac in the Burbidge & Hewitt (1987) list, selected because of its UV excess. Arp (1977) reports the obser- 7 1320+084 This source is part of the BL Lac 1722+119 Griffithsetal.(1989)reportedaten- sampleextractedfromtheEINSTEINSlewSurvey tative redshift z=0.018 for this X-ray selected, and a radio counterpart was reported by Perlman highly polarized BL Lac. This estimate was not et al. (1996). Our VLT data show the source has confirmed by more recent observations (V´eron- a QSO like spectrum at z=1.5, in contrast with Cetty & V´eron 1993; Falomo et al. 1993; Falomo, a featureless spectrum observedby Perlmanet al. Scarpa, & Bersanelli 1994). Our VLT spectrum (1996). Several intervening absorption lines, in (S/N=350)shows only absorptionfeatures due to particular MgII at z=1.347 were also detected. ourgalaxyISM:CaIIλλ3934,3968,NaIλ5892and DIBs at 4428˚A, 4726˚A, 4882˚A, 5772˚A, with no evidence of intrinsic features. From the minimum 1349–439 The spectrum of this X-ray selected EW we derive z>0.17. BL Lac (della Ceca et al. 1990), shows a number min of absorption lines from the interstellar medium: CaIIλλ3934,3968,the5772˚ADIB,NaIλ5892. No 2012–017 Consistently with previous observa- intrinsic features were detected, and the deduced tions of this radio selected BL Lac (White et redshift lower limit is z>0.39. As already pointed al. 1988; V´eron et al. 1990; Falomo, Scarpa, & out by V´eron (1996), the value z=0.05 sometime Bersanelli 1994), also our S/N=130 VLT spec- reportedfor this object is consequence of a confu- trum is featureless. The optical spectral index is sionwith the nearbySeyfert1galaxyQ 1349-439. α=0.49,inmarginalagreementwithα=0.33±0.12 reported by Falomo, Scarpa, & Bersanelli (1994). From EW we derive z>0.94. 1442–032 ThisX-raysource,the radiocounter- min part of which was found in the NVSS survey, was first classified as a BL Lac in the RBSC-NVSS 2128–254 The spectrum of this X-ray selected samplebyBaueretal.(2000),andthenconfirmed BLLaccandidateisreportedasfeaturelessbySS. by the SS. There are no published optical spec- We confirm this result and set a lower limit of tra for this source. Our optical spectrum is fea- z>0.86 for the redshift. tureless, with the exception of the NaI λ5892 ab- sorptionfeaturefromourgalaxyISM.TheEW min 2131–021 Rector&Stocke(2001)andDrinkwa- value for this objects leads to z>0.51. ter et al. (1997) proposed a redshift of 1.285 for this source, based on the detection of CIII] 1500-154 This X-ray selected BL Lac is part of λ1909, MgII λ2798 and [OII] λ3727, opposed to the RSBC-NVSS sample (Bauer et al. 2000) and the z=0.557 suggested by Wills & Wills (1976). entersinSS.Nopreviousopticalspectroscopyhas While [OII] falls outside our spectral range, we beenfoundintheliterature. Ourspectrumiscom- confirm the presence of CIII] and MgII emission pletelyfeatureless,leadingtoz>0.38fromtheob- lines at z=1.283, also detecting the fainter CII] tained EW . λ2326 feature at the same redshift. min 1553+113 This source is an optically selected 2133–449 This source was discovered because BL Lac from the Palomar-Green survey. The of its optical variability by Hawkins et al. (1991). redshift estimate z=0.360 given in the Hewitt & OpticalspectroscopybyHawkinsetal.(1991)and Burbidge (1993) catalogue was disproved by later Heidt et al. (2004) led to completely featureless spectroscopy (Falomo & Treves 1990; Falomo, spectra. Our VLT observations clearly show the Scarpa, & Bersanelli 1994). While no intrinsic presence of an intervening absorption feature at featuresweredetectedinourS/N=250VLTspec- 4250˚A,atentativeidentificationofwhichisinter- trum, a number of absorption lines due to our vening MgII at z=0.52(see Churchill et al. 2005). galaxyISMwererevealed: CaIIλλ3934,3968,NaI ThelowerlimitonzderivedfromEW isz>0.98 min λ5892 and DIBs at 4428,4726,4882,5772 ˚A. The . EW estimate for this object gives a limit min z>0.09. 2136–428 The spectrum obtained by Hawkins et al. (1991), who discovered this source study- 8 ing its optical variability, is completely feature- z=0.492, while, using the EW estimate from min less. Our VLT observations shows severalabsorp- the spectrum, z>0.65 is found. tionfeaturesduetotheISMofourgalaxy: DIBat 4428˚A,4726˚A,4882˚Aand5772˚A,CaIIλλ3934, 2254–204 Previousopticalspectroscopy(V´eron- 3968 and NaIλ 5892 atomic lines. The feature at Cetty & V´eron1993;Hook et al. 2003)of this BL 5942 ˚A could be CaII λ3968 at z=0.497. Since Lacobjectfromthe1Jysampleshowedcompletely at this redshift the CaII λ3934should fall at 5890 featureless spectra. With VLT we are able to de- ˚A, where it will be strongly contaminated by the tect faint interstellar absorptions of CaII λλ3934, interstellar NaI absorption, the redshift estimate 3968 and NaIλ 5892, but no intrinsic or interven- is only tentative. The lower limit deduced by the ing spectral lines are found. The inferred redshift minimum measurable EW is z>0.24. limit is z>0.47. 2214–313 Our VLT spectrum of this object 2307–375 This source was first classified as a clearly shows the typical spectral signature of the BL Lac in the RSBC-NVSS sample (Bauer et al. hostgalaxy(CaIIλλ3934,3968andGbandλ4305) 2000). The classification was then confirmed by atz=0.46. Thebestfitdecompositiongivesα=0.9 the SS.Nopreviousopticalspectroscopyhasbeen for the non-thermal component and M =–22.3 R published. Our VLT spectrum is featureless, al- for the host. Previous optical spectroscopy per- lowing us to set only a lower limit to the redshift formed by Bade, Fink, & Engels (1994) with the of z>1. ESO 3.6m telescope failed to detect any spectral feature. 2342–153 ThissourceispartoftheEMSSsam- ple of BL Lac objects. Our VLT data, as well 2223–114 Optical observations of this radio as previous optical spectroscopy with the 6.5 m source obtained by V´eron-Cetty & V´eron (1993) telescope of Multi Mirror Telescope Observatory did not show any intrinsic spectral feature. In (Rectoretal.2000)showedafeaturelessspectrum. our spectrum, that extends further in the red, From EW we derive z>1. min we detect a single narrow emission line at λ7367 (EW = 5 ˚A). This is a realfeature since it clearly 2354–021 This object was discussedin paper I. appears on each of the 3 individual spectra (see Here we report only the spectrum, in Fig. 2. section 3). A possible identification of this line is [OII]λ 3727 at z=0.977, while MgII λ2798 gives z=1.633. We discarded this second classification 2354–175 This X-ray source from ROSAT All because the line FWHM (1200 km s−1), is typi- SkySurvey,isclassifiedasBLLaccandidateinthe cal for a narrow line such [OII], while for MgII a RBSC-NVSSsample(Baueretal.2000)andinthe larger value would be expected. Moreover, with SS.Nopreviousspectroscopywaspublishedinlit- a MgII identification both CIV λ1549 and CIII] erature. Our S/N=150 VLT spectrum is feature- λ1909 broad lines would be expected inside the less, allowing only to set a lower limit of z>0.85 observedspectral range,but no other features are to the redshift. detected. 6. Summary and conclusions 2233–148 The redshift z=0.325 reported by Out of 42 objects observed we confirm the BL Johnston et al. (1995) is due to confusion with Lac classification for 36 sources and for 18 of the source HB89 2233+134 in Schmidt & Green themweareabletomeasure/confirmtheredshift. (1983). Drinkwater et al. (1997) report an in- This information allows us to derive the luminos- tervening system at z=0.609, but without giving ity of the objects. The distribution in the V band an identification of the corresponding absorption luminosity-distance plane is indeed fully consis- feature. We detect several absorption features on tent with what observed for BL Lacs of known the spectrum. In particular we propose to iden- redshift in the combination of the Padovani & tify the features at 4165 and 4183 ˚A as MgII at Giommi (1995a) and the SS sample (see Fig. 11). 9 We note that the sources in this combined list REFERENCES are affected by the typical selection effect of in- Abraham, R. G., Crawford, C. S., & McHardy, complete, flux limited samples: the envelope of I. M. 1991, MNRAS, 252, 482 the objects follows in fact the expected behavior for sources with constant V magnitude, centered Allington-Smith, J.R., Peacock,J.A., & Dunlop, aroundV=18,with a spreadof∼ 3 mag. The ob- J. S. 1991, MNRAS, 253, 287 jectsdiscussedhere(filledcirclesandlowerlimits) follow the same distribution, with absolute mag- Appenzeller et al., 1998, Messenger 94, 1 nitudes ranging between -21.5 and -27.5, slightly Arp, H. 1977, Coll. Intl. No. 263, Paris-Centre increasing with the redshift. Nat. Recherche Sci, p. 377 In 18 cases the optical spectra remain lineless in spite of the high S/N of the obtained optical Bade, N., Fink, H. H., & Engels, D. 1994, A&A, spectra. This indicates that if they are hosted by 286, 381 galaxies of standard luminosity they have likely Bauer,F.E.,Condon,J.J.,Thuan,T.X.,&Brod- very luminous or extremely beamed nuclei (see erick, J. J. 2000, ApJS, 129, 547 also Sbarufatti et al. 2005b). In the latter case one may expect to see the most extreme cases of Baldwin,J.A.,Burbidge,E.M.,Hazard,C.,Mur- relativistic beaming, making these sources ideal doch,H.S., Robinson,L.B.,& Wampler,E.J. targets for milli-arcsec resolution radio observa- 1973,ApJ, 185, 739 tions. Alternatively, if the host galaxies were under-luminous,theseobjectscouldberareexam- Bressan, A., Granato, G. L., & Silva, L. 1998, ples of dwarf galaxies hosting an AGN (see Sbar- A&A, 332, 135 ufatti et al. 2005b). Burbidge, G., & Hewitt, A. 1987,AJ, 93, 1 The high S/N of most of the optical spectra obtained at VLT represents a frontier for the de- Burbidge, G. 1996, A&A, 309, 9 termination of the redshift of BL Lacs with cur- rent instrumentation and further improvement of Carangelo,N., Falomo, R., Kotilainen, J., Treves, the issue will not be easy to get. In particu- A., & Ulrich, M.-H. 2003, A&A, 412, 651 lar the four brightest objects (R<16: 0048–099, Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1553+113, 1722+119, 2136-428) we observed at 1989,ApJ, 345, 245 the VLT have values of EW smaller than 0.25 min ˚A. These objects belong to an interesting sub- Churchill, C. W., Kacprzac, G. G. & Steidel, population of BL Lac objects with extreme nu- C. C., 2005, proceedings for IAU colloquium clear (and/or host) properties for which it is ac- 199: ProbingGalaxiesthroughQuasarAbsorp- tually not possible to derive the intrinsic physical tion Lines, Cambridge University Press, 2005 parameters. Onepossibilityistoimagethesource when it is particularly faint in order to improve Condon,J.J.,Cotton,W.D.,Greisen,E.W.,Yin, the detection of the host galaxy and to derive an Q.F.,Perley,R.A.,Taylor,G.B.,&Broderick, imaging redshift (Sbarufatti et al. 2005b). Deep J. J. 1998, AJ, 115, 1693 spectroscopicobservationsinthenear-IRmayalso Croom, S. M., Smith, R. J., Boyle, B. J., Shanks, prove to be effective in the determination of the T., Miller, L., Outram, P. J., & Loaring, N. S. redshift considering this region of the spectrum is 2004,MNRAS, 349, 1397 poorly known. della Ceca, R., Palumbo, G. G. C., Persic, M., We would like to thank the referee, Dr. John Boldt, E. A., Marshall, E. E., & de Zotti, G. Stocke, for his accurate and helpful comments, 1990,ApJS, 72, 471 which allowed us to improve the quality of this Drinkwater, M. J., et al. 1997, MNRAS, 284, 85 work. Ellison, S. L., Yan, L., Hook, I. M., Pettini, M., Wall, J. V., & Shaver, P. 2001,A&A, 379, 393 10