FIRST “Winged” and X-shaped Radio Source Candidates: II. New Redshifts

The majority of double-lobed radio galaxies can be classified into one of two morphological classes: edge-dimmed and edge-brightened, corresponding to the Fanaroff & Riley (1974) type I and II sources, respectively. Some of the most unusual radio galaxies are those with an additional pair of extended low surface brightness ‘wings’, making an overall X-shaped appearance. One popular model posits that these are recently merged systems and that the wings are the inactive lobes marking the pre-merger axis of the central supermassive black hole/accretion disk system (see, e.g., Rottmann, 2001; Merritt & Ekers, 2002; Komossa, 2006). However, this and other models (e.g., buoyant expansion, Leahy & Williams, 1984; Worrall et al., 1995) are not well tested.

Generally, these interpretations are hampered by the fact that a small number of X-shaped radio sources are known (a recent census has it at 16 bona fide examples; Cheung, 2007) and an even smaller subset have been studied in detail. In an attempt to remedy this deficiency, Cheung (2007, hereafter Paper I) compiled a sample of 100 new candidate sources of this type selected from radio maps from the VLA-FIRST survey (Becker et al., 1995). One of the first steps in a systematic follow-up multi-wavelength study of this large sample is to obtain spectroscopic redshifts and identifications. This is necessary to study their demographics, environments, etc., enabling comparisons with radio galaxies without wings (cf., Ulrich & Rönnback, 1996; Zirbel, 1997). For instance, X-shaped radio sources are known to have radio luminosities near the Fanaroff-Riley (FR) type I/II divide (Leahy & Parma, 1992; Dennett-Thorpe et al., 2002), and their morphologies suggest that these may be the long sought-after transition sources (Paper I). To evaluate this interesting scenario, an important first step is to place the X-shaped sources in the Owen-Ledlow (Owen, 1993; Ledlow & Owen, 1996) plane (radio luminosity versus host galaxy absolute magnitude).

As part of our X-shaped source study, we have obtained optical spectroscopic observations of 26 sources included in Paper I, and one additional new example (presented here). The redshifts, optical spectra, and some of the basic results from these observations are discussed. An Owen-Ledlow plot for a 50-source sample of spectroscopically identified X-shaped radio galaxies and best candidates from Paper I is constructed and this is discussed in relation to the Fanaroff-Riley division. A more detailed study of the spectra will be presented in a future paper in this series. Throughout, we assume $h=H_{0}$/(100 km s^-1 Mpc^-1)=0.7, $\Omega_{\rm M}=0.3$ and $\Omega_{\rm\Lambda}=0.7$, as in Paper I.

This paper presents observations with the 9.2 m Hobby-Eberly Telescope (HET) at McDonald Observatory in 2006 and 2007 under queue observing mode and with the 6.5 m Multiple-Mirror Telescope (MMT) at Mt. Hopkins Observatory in a two-night run in Feb. 2007. An observation of one object (J0702+5002) was obtained earlier with the 2.7 m Harlan J. Smith Telescope (HJST) at McDonald Observatory. A summary of the observations is presented in Table 1.

On the HET, we used the Marcario Low-Resolution Spectrograph (LRS; Hill et al., 1998), with grism G1, a $2\hbox{${}^{\prime\prime}$}$ slit, and a Schott GG385 long-pass filter. This provided broad wavelength coverage (from $\sim$4000–9200 Å) and a resolution of $\mathcal{R}\approx 500$. Typically, we obtained $2\times 300$ s exposures for the brightest targets and $2\times 600$ s exposures for fainter ones. Eleven total spectra from the HET are presented here.

The MMT observations utilized the blue channel spectrograph ($\mathcal{R}\approx 300$) with a $1\hbox{${}^{\prime\prime}$}$ slit and wavelength range $\sim$3200–8400 Å. The slit was rotated to the parallactic angle for all observations. To mitigate against second-order effects, we used the L-42 filter on the second night of the observing run. Sky conditions were clear early on with increasing clouds toward the end of each night resulting in several hours where the telescope was idle. The seeing was consistently at the sub-arcsecond level for all 15 successfully obtained spectra.

The observation of J0702+5002, a candidate X-shaped radio galaxy identified early on in the work leading up to Paper I, was obtained by S. E. H. in 2005 with the 2.7 m HJST, using the Imaging Grism Instrument (IGI) and the 6000 Å VPH grism. The redshift was reported in Paper I, and the spectrum is provided here.

Our targets were selected from the list of 100 FIRST X-shaped radio source candidates presented in Paper I (Table 2 therein). We created a prioritized list of targets consisting of candidates with radio morphologies judged most likely to be bona fide X-shaped sources from their VLA-FIRST maps as well as those few known X-shaped radio sources without known redshifts (Paper I, Table 1 therein). To fill in scheduling gaps during the MMT run, we additionally observed several objects from the known sample where we did not have access to the optical spectra in digital format either from SDSS (Adelman-McCarthy et al., 2007) or privately available from other researchers. Lastly, one of us (C. C. C.) is compiling a new set of candidate X-shaped radio sources as a follow-up study to Paper I; a MMT spectrum of one of these was obtained, and the new identification is presented in § B.

All spectroscopic data were analyzed with IRAF¹¹1IRAF is distributed by the National Optical Astronomy Observatories, which are operated by the Association of Universities for Research in Astronomy, Inc., under cooperative agreement with the National Science Foundation. using standard routines, including calibration of the absolute fluxes by comparison with standard stars and removal of telluric absorption features. Redshifts were determined from the narrow emission lines [O II] $\lambda$3727 and/or [O III] $\lambda$5007 for the majority of our sources (19/27 objects). In about one-third of our sources (8/27 objects), no emission lines were detected, and we derived the redshift from (at least two of) the absorption features of the host galaxy (typically Ca II, G band, Mg Ib, and NaD). We regard the redshifts of two sources, J1135–0737 and J1434+5906, as tentative because of the relatively low S/N ratio ($\sim$5) of their spectra.

The spectra are presented in Figure 1, and the redshift measurements are reported in Table 1. New spectroscopic identifications and redshifts have been successfully obtained for 21 sources, and new spectra were obtained for 6 sources with known redshifts (§ 3.1). In 2 of the 21 cases in the former category, new redshifts are measured where only previous estimates were available (§ 3.2).

Since the publication of Paper I, a spectroscopic redshift was found for the candidate ($\#$78 in Paper I) X-shaped radio source J1433+0037 ($z=0.5031$; Cannon et al., 2006). Also, in § A, we describe an optical misidentification of a candidate from Paper I revealed by a new MMT spectrum. We note finally that other radio sources with X-shaped or ‘winged’ type morphologies are being identified (e.g., Landt & Bignall, 2008; Saripalli & Subrahmanyan, 2008). One new X-shaped radio source, FIRST J1018+2914, discovered as part of a follow-up study to Paper I, is described in § B.

In Table 2, we define a sample of 50 spectroscopically identified X-shaped radio galaxies from the known list and the best candidates from Paper I. Of this radio galaxy²²2Based on their optical spectra, in particular the absence of broad emission lines, the optical light is most likely dominated by the host galaxy and little contaminated by relativistically beamed emission from a jet. The broad emission-lined sources are not included at this time because of the possible contamination of the nuclear emission. sample, 15 known sources and 16 VLA-FIRST candidates had redshifts available from the literature; this paper is responsible for the inclusion of the remaining 19 sources with our newly determined redshifts. For this sample, we can determine the source distances, thus allowing estimates of the absolute optical magnitudes of their host galaxies and the (monochromatic) radio luminosities.

We find that the average absolute $R$-band magnitude of the X-shaped radio galaxy sample is $-$23.2 (1$\sigma=0.8$; median = $-$23.1) which is entirely consistent with values for ‘normal’ radio galaxies. Specifically, Govoni et al. (2000) found an average of $M_{\rm R}=-23.33\pm 0.69$ (converting to our adopted cosmology) for their sample which includes both FR I and FR II sources, and McLure et al. (2004) found an average $M_{\rm R}=-23.20\pm 0.09$ (median = $-$23.25) in their sample of powerful $z\sim 0.5$ radio galaxies.

As found in Paper I, we confirm that the average 1.4 GHz radio luminosity of our sample is close to the Fanaroff-Riley division (average log $L_{\rm 1.4}$ [W Hz^-1]= 25.79, 1$\sigma=0.69$). This is an expected result since X-shaped radio galaxies have long been known to have typical radio luminosities near the FR I/II division (Leahy & Parma, 1992; Dennett-Thorpe et al., 2002).

To further explore the relationship of X-shaped radio galaxies with the Fanaroff-Riley division, in Figure 2, we plot the 1.4 GHz radio luminosity vs. parent host galaxy absolute $R$-magnitude diagram for this sample. Owen & Ledlow (Owen, 1993; Ledlow & Owen, 1996) found a clear dividing line between FR I and FR II radio sources in this diagnostic plane, and this line is drawn in Figure 2. As anticipated from their average radio luminosities, the X-shaped sources straddle the Owen-Ledlow FR I/II dividing line, but there is a large dispersion. For the sample of known X-shaped radio galaxies (marked with an ‘X’ in Figure 2), the objects are neatly separated by their Fanaroff-Riley type morphology with the one clear FR I radio galaxy (NGC326 = X02; Ekers et al., 1978; Murgia et al., 2001) clearly below this line. The only potential exception is that of 3C 315 (X14), whose radio morphology on the scale of hundreds of kpc is more typical of an FR I (Leahy & Williams, 1984; Lal & Rao, 2007) but which lies in the FR II regime of the Owen-Ledlow plot (Figure 2). However, its central radio component is remarkable in that it is resolved into a few kpc-scale FR II-like double (de Koff et al., 2000; Saripalli & Subrahmanyan, 2008), although this feature is responsible for only a small fraction of the source total luminosity (see the maps in Leahy & Williams, 1984; Lal & Rao, 2007). A few of the FIRST candidates from Paper I also extend into the FR I portion of the Owen-Ledlow plot. New VLA imaging of the sample (already obtained for many of them) will allow us to classify their FR type to explore this relationship further in a future paper. At present, the connection between the X-shaped source phenomenon and the Fanaroff-Riley division remains unclear.

New optical spectroscopic observations have been obtained for 27 candidate and known X-shaped radio sources. Six had previously known redshifts (§ 3.1), and 21 are newly determined redshifts (19 candidates from Paper I, one previously known, and one new X-shaped source). With only 34/100 X-shaped radio source candidates presented in Paper I with redshifts previously available from the literature, this now brings the sample of candidates over 50$\%$ identified. More importantly, since our spectroscopic targets were selected as the most promising X-shaped source candidates based on their morphologies in the FIRST maps, a large fraction of the most interesting sources are now spectroscopically identified (see § 4 and Table 2).

The sources with redshifts determined in this work are predominantly fainter optically than the 34 X-shaped source candidates with previously available redshifts in Paper I (see Figure 3). As expected, our targets are found to be typically more distant than previously known examples ($z<0.4$; Paper I). With 8 sources newly determined to have redshifts of $z>0.5$ (Figure 3), this extends our census of X-shaped sources to higher redshifts. It is worth mentioning that half of the redshifts taken from the literature were obtained from the SDSS (Adelman-McCarthy et al., 2007), and these were fairly bright ($r$ $\;\buildrel<\over{\sim}\;$18–19) targets; our targets are typically fainter and thus will likely not be observed by SDSS in the future.

Roughly a third of the observed sources have no emission lines detected and the remaining are predominantly strong narrow emission-line objects. The obvious exceptions are J1342+2547 and J1406+0647 (Catalog $\#$67 and $\#$73, respectively), which also have strong broad emission lines indicative of quasars. Broad-line emission from these two targets was indeed anticipated in Paper I based on their bluer optical colors ($g-r=0.4$ and 0.0, respectively). This adds to the 2 known and 4 candidate broad lined X-shaped sources presented in Paper I. Such broad line objects seem to be relatively rare amongst X-shaped radio sources, occurring in only $\sim$10$\%$ of the sample, and these are two excellent examples. The overall rate of detected emission line objects/types here is consistent with that of normal radio galaxies (e.g., Tadhunter et al., 1993).

It is important to identify more broad emission-lined objects as they may provide valuable clues as to the possible formation scenarios for X-shaped radio morphologies. Specifically, the spectra can be used to constrain systematic velocity offsets between the broad and narrow lines, a possible observational signature of gravitational wave radiation recoil following a supermassive black hole binary merger (Bonning et al., 2007; Komossa et al., 2008). These, and other aspects of the spectra, will be explored in a future paper to search further for possible clues as to the origin of the unusual morphologies of these radio sources.