Effelsberg 100-m polarimetric observations of a sample of Compact Steep-Spectrum sources
F. Mantovani, K.-H. Mack, F.M. Montenegro-Montes, A. Rossetti, A. Kraus
aa r X i v : . [ a s t r o - ph . C O ] J un Astronomy & Astrophysics manuscript no. eflsg-pol-astro-ph c (cid:13)
ESO 2018October 29, 2018
Effelsberg 100-m polarimetric observations of a sample ofCompact Steep-Spectrum sources
F. Mantovani , K.-H. Mack , F.M. Montenegro-Montes , , , A. Rossetti , and A. Kraus Istituto di Radioastronomia – INAF, Via Gobetti 101, I-40129 Bologna, Italy Dpto. de Astrof´ısica. Universidad de La Laguna, Avda. Astrof´ısico Fco. S´anchez s/n, E-38200 La Laguna (Tenerife),Spain Instituto de Astrof´ısica de Canarias. C/ Via L´actea s/n. E-38200 La Laguna (Tenerife), Spain Max-Planck-Institut f¨ur Radioastronomie, Auf dem H¨ugel 69, D-53121 Bonn, GermanyReceived October 29, 2018; accepted ???
ABSTRACT
Aims.
We completed observations with the Effelsberg 100-m radio telescope to measure the polarised emission from acomplete sample of Compact Steep-Spectrum sources and improve our understanding of the physical conditions insideand around regions of radio emission embedded in dense interstellar environments.
Methods.
We observed the sources at four different frequencies, namely 2.64 GHz, 4.85 GHz, 8.35 GHz, and 10.45 GHz,making use of the polarimeters available at the Effelsberg telescope. We complemented these measurements withpolarisation parameters at 1.4 GHz derived from the NRAO VLA Sky Survey. Previous single dish measurements weretaken from the catalogue of Tabara and Inoue.
Results.
The depolarisation index DP was computed for four pairs of frequencies. A drop in the fractional polarisationappeared in the radio emission when observing at frequencies below ∼ − found for 3C138to 3900 rad m − in 3C119. In all cases, the λ law is closely followed. Conclusions.
The presence of a foreground screen as predicted by the Tribble model or with “partial coverage” asdefined by ourselves can explain the polarimetric behaviour of the CSS sources detected in polarisation by the presentobservations. Indication of repolarisation at lower frequencies was found for some sources. A case of possible variabilityin the fractional polarisation is also suggested. The most unexpected result was found for the distribution of thefractional polarisations versus the linear sizes of the sources. Our results appear to disagree with the findings of Cottonand collaborators and Fanti and collaborators for the B3-VLA sample of CSS sources, the so-called “Cotton effect”, i.e.,a strong drop in polarised intensity for the most compact sources below a given frequency. This apparent contradictionmay, however, be caused by the large contamination of the sample by quasars with respect to the B3-VLA.
Key words. polarisation – galaxies: quasars: complete sample – radio continuum
1. Introduction
Measurements of the polarised emission from Giga-Hertz-Peaked Spectrum (GPS) and Compact Steep-Spectrum(CSS) sources can provide important information aboutthe physical conditions inside and around the region of ra-dio emission. GPS/CSS sources are physically small ob-jects with radio sizes smaller than 1 kpc (essentially on theNarrow Line Region size scale) and 15 kpc, respectively,that reside inside their host galaxies. The most widely ac-cepted interpretation is that GPS/CSS galaxies are youngradio sources (of ages < yr, Fanti et al. 1990). Thisview is supported by measurements of hot spot advancespeeds (e.g., Polatidis & Conway 2003) and by means ofspectral ageing studies (Murgia et al. 1999). GPS sources,which are the youngest, will grow into CSS and eventu-ally into classical extended radio sources. Consequently,the measurement of the physical properties of GPS andCSS sources can provide insight into the conditions at thebirth of a powerful radio source and those of sources de- Send offprint requests to : F. Mantovani,e-mail: [email protected] veloping in dense interstellar environments (see Fanti et al.1995 and Readhead et al. 1996). The observations suggestthat intrinsic distorsions in CSS sources are be caused bytheir interactions with dense and inhomogeneous gaseousenvironments. Asymmetries in terms of flux density, armratio, spectral index, and polarisation suggest that theyare expanding through the dense inhomogeneous interstel-lar medium of their host galaxies (e.g., Saikia et al. 2003;Rossetti et al. 2006).An effect on the synchrotron polarised emission pro-duced by this magnetized thermal plasma is Faraday ro-tation, which is proportional to the product of electrondensity and the magnetic field component parallel to thedirection of propagation integrated along the line of sight.The Rotation Measure (RM) is the amount of Faraday ro-tation expressed in rad m − . The Faraday rotation can un-ambiguously be determined from observations for at leastthree different wavelengths. Strong variations in Faradayrotation across the telescope beam will reduce or depolarisethe observed fractional polarisation. This effect can oftencompletely depolarise regions of emission. F. Mantovani et al.: Effelsberg polarimetry of CSS sources
In many sources, the polarisation increases with increas-ing frequency when observed at similar angular resolutions.The fractional polarisations of CSS sources at 15 GHz (6–7%) indeed tend to be higher than at 5 GHz (1–3 %; vanBreugel et al. 1984; Saikia et al. 1987). This suggests thatlarge Faraday depths are responsible for the depolarisationbetween 15 GHz and 5 GHz rather than magnetic field ge-ometry.The magnetized plasma responsible for Faraday rota-tion and depolarisation can be situated either within thesource, in a foreground screen, or in both. However, thereare several indications, such as the magnitudes of the RMs,the total rotation of the electric vector position angle > ◦ without very high depolarisation, and the lack of correlationbetween high RMs in extragalactic sources and Galacticlatitude (see for example the observations of Cygnus A byDreher et al. 1987), that most of the observable effectsare not internal to the source, but are produced in fore-ground material in the vicinity of the radio synchrotronsource (e.g., Leahy 1990).A model to describe the depolarisation behaviour in ex-ternal screens has been discussed by Burn (1966) and latergeneralized by Tribble (1991) who considered the impor-tance of the sizes of the individual Faraday cells in relationto the observing beam sizes. A variation in the ratio of thesesizes can cause a more or less strong decline in the intrin-sic polarisation percentage towards longer wavelengths. Asuccessful application of the Tribble model, which impliesthe existence of a foreground screen, can be found in Fantiet al. (2004), who analysed multi-frequency observations ofthe B3-VLA CSS sources (Fanti et al. 2001).Many GPS/CSS sources have been observed mainlywith radio interferometers. These investigations have pro-vided very interesting results about the polarised state ofthe radio emission in GPS/CSS sources. A summary can befound in O’Dea (1998) and Cotton et al. (2003). However,apart from the B3-VLA sample (Fanti et al. 2001), mostof the samples observed in polarisation so far (L¨udke et al.1998; Akujor et al. 1995; van Breugel et al. 1984; Rossettiet al. 2008) contained subsets of sources whose incomplete-ness hamper statistical studies or which were observed at asingle or dual frequency.The set up of this paper is as follows. Sample selection,observations and data reduction are described in Sect. 2.Results from the observations are presented in Sect. 3 anddiscussed in Sect 4. In Sect. 5 we draw our conclusions. Plotsof the fractional polarisation m and Rotation MeasuresRM versus wavelength squared are compiled in Appendix1. Comments about individual sources can be found inAppendix 2.
2. Observations and data reduction
The sample of CSS sources subject to the present inves-tigation was constructed by Fanti et al. (1990) from the3CR catalogue (Jenkins et al. 1977) and the Peacock &Wall (1982, hereafter PW) sample. Since there is enoughinformation about the radio structures and the spectralbehaviour of the sources in the 3CR and PW catalogues,it is safe to assume that the sample contains all the CSSsources belonging to those catalogues with projected linearsizes less than 15 kpc, flux density at 178 MHz ≥
10 Jy, log P > − , and with | b | > ◦ , δ > ◦ , and canthus be considered statistically complete.We observed all the sources belonging to this samplewith the Effelsberg 100-m radio telescope at four frequen-cies between 2.64 GHz and 10.45 GHz in a relatively shortperiod of time (i.e., contemporaneously at the three higherfrequencies, and 16 months later at 2.64 GHz). We choosethis approach to avoid time variability effects when mea-suring the percentage of polarised flux density for each ofthese sources, the depolarisation indices, and the RMs withthe aim of understanding more clearly the physical condi-tions in which CSS sources eventually expand to becomelarger sources.Polarisation parameters at 1.4 GHz were also derivedfor the full sample from the National Radio AstronomyObservatory Very Large Array Sky Survey (NVSS; Condonet al. 1998). Because of their small angular sizes, CSSsources are point-like at the NVSS resolution. In the anal-ysis of sources that were found polarised at two or moreof the five frequencies above, we complemented the polari-sation measurements with those listed by Tabara & Inoue(1980).For consistency with previous work, we used the cos-mology H = 100 km s − Mpc − and q = 0 . The Effelsberg 100-m telescope was used to observe thecomplete sample of GPS/CSS sources constructed by Fantiet al. (1990). To minimize the occurrence of possible n π ambiguities in the determination of RMs, we observed thesources at four independent frequencies, namely 2.64 GHz,4.85 GHz, 8.35 GHz, and 10.45 GHz, making use of the po-larimeters available at the Effelsberg telescope. The ob-servations were carried out in the period January 26 toFebruary 1, 2005 at 4.85 GHz, 8.35 GHz, and 10.45 GHz andJune 24 to 26, 2006 at 2.64 GHz. Since all the target sourcesare point-like to the Effelsberg telescope beams, we usedcross-scanning to determine the total intensity and polar-isation characteristics. All sources in the sample are verybright, thus standard cross-scans along the azimuth and el-evation axes were used, with 4 to 8 subscans, according tothe source flux densities. Table 1 summarises the observingparameters. For further details about the observation modeand a description of the receivers, we refer to Montenegro–Montes et al. (2008), Klein et al. (2003), and referencestherein. The calibration sources 3C 286 and 3C 295 were Table 1.
Observing parameters and estimated integrationtimes
Centre Frequency [GHz] 2.64 4.85 8.35 10.45Bandwidth [MHz] 80 500 1100 300System Temp.(zenith) [K] 17 27 22 53Scan length [ ′ ] 16 12 8 6Scan speed [ ′ /min] 45 45 40 30 regularly observed to correct for time-dependent gain in-stabilities and to bring our measurements onto an absoluteflux density scale (Baars et al. 1977). The quasar 3C 286was also used as a polarisation calibrator to obtain the po-larisation degree m and the polarisation angle χ in agree-ment with values in the literature (Tabara et al. 1980). The . Mantovani et al.: Effelsberg polarimetry of CSS sources 3 unpolarised planetary nebula NGC 7027 was also observedto estimate the instrumental polarisation. We found an in-strumental polarisation ( p instr ) of 0.5% at 2.64 GHz, 0.5%at 4.85 GHz, 0.3% at 8.35 GHz, and 0.8% at 10.45 GHz.The measurement of flux densities from the single-dishcross-scans was done by fitting Gaussians to the signal ofthe polarimeter output channels (Stokes I, Q and U) andidentifying the Gaussian amplitudes with the flux densities S I , S Q , and S U . For all sources with significant S Q and S U contributions, the polarised flux density S P , the degreeof linear polarisation m , and the polarisation angle χ werecomputed.We consider three main contributions to the flux densityerror as in Klein et al. (2003). These are(i) the calibration error ∆ S c , which is estimated to bethe dispersion in the different observations of the flux den-sity calibrators, i.e., about 2% at all observing frequencies;(ii) the error introduced by noise, ∆ S i ∼ i =I,Q,U), which is estimated from the noise at the scan edges;(iii) the confusion error ∆ S conf caused by backgroundsources within the beam area, estimated to be 1.5 mJyat 2.64 GHz, 0.45 mJy at 4.85 GHz, 0.17 mJy at 8.35 GHz,and 0.08 mJy at 10.45 GHz (Klein et al. 2003; the valueat 8.35 GHz was extrapolated from these existing measure-ments. The confusion limits can be neglected in the calcu-lation of the total error in Stokes Q and U .These contribute to the total error in the following way:∆ S i = q ( S i ∗ ∆ S c ) + ∆ S i + ∆ S conf Since we are dealing with relatively bright targets, instru-mental polarisation is an issue in many cases. This has beenincluded in the error of the fractional polarisation in thefollowing way:∆ S P = s ( S Q ∗ ∆ S Q ) + ( S U ∗ ∆ S U ) S P + ( p instr ∗ S I ) For all other errors we follow the definitions given by Kleinet al. (2003). The errors associated with the position angles χ also account for the distribution, assumed to be Gaussian,in the values of χ obtained for the calibrator 3C286 in thecalibration process. The Effelsberg measurements were complemented withdata of the NRAO VLA Sky Survey (NVSS) at 1.4 GHz(Condon et al. 1998). In all these measurements, our tar-gets are point-like to the corresponding beams, thus beamdepolarising effects are avoided. The polarised flux densitiesin Table 2 are given for sources with polarised flux densitiesgreater than three times the rms error. The rms uncertaintyis computed following Eq. 49 in Condon et al. (1998) σ P ≈ σ Q,U + ǫ P A P where ǫ P is the residual instrumental polarisation, whichis about 0.12% for a large sample of sources stronger than1 Jy and A P is the fitted peaked amplitude. In Table 3, wereport the fractional polarisation values for these sources.
3. Results from the observations
In Table 2, we present both total and polarised flux den-sities derived by analysing the results of the observationscompleted with the Effelsberg 100-m radio telescope at thefour frequencies plus those extracted from the NVSS. Thesemeasurements plus those derived from the literature areshown in Appendix 1.
The sample constructed by Fanti et al. (1990) contains 47sources. We considered them to be polarised when the in-tensity of the polarised emission is 3 times the rms errorestimated source by source for the polarised emission. Thisprovides in general lower limits of 3% at 2.64 GHz, 2% at4.85 GHz, 1% at 8.35 GHz, and 1% at 10.45 GHz. Table 3summarizes the values of the polarisation parameters forsources with polarised flux densities above the detectionlimits.A high fraction of sources have polarised emission,if any, below the detection limits of our Effelsberg ob-servations. We found that 22 of them are polarised at10.45 GHz, 16 sources are polarised at 8.35 GHz, 10 sourcesare polarised at 4.85 GHz, and 10 sources are polarised at2.64 GHz. The number of sources with detected polarisedemission clearly decreases from high to low frequency. Thepercentage of polarised emission m ranges between ∼ ∼
12% for the most polarised sources 3C138, 3C286,and 3C454.Because most of the sources in the sample have po-larised emission below the detection limits of our obser-vations, the median values of m for the entire sample arebelow 3% at 2.64 GHz, below 2% at 4.85 GHz, below 1% atboth 8.35 GHz and 10.45 GHz, and below 1.58% at 1.4 GHz. Table 4 summarizes the values of the depolarisation indexand of the observed RM (RM obs ) and source rest frameRM (RM rf = RM obs × (1 + z ) ) both given in rad m − .The depolarisation index DP = m l /m h , is defined to bethe ratio of the percentages of polarised emission at thelower ( m l ) to the higher ( m h ) frequency. The present ob-servations complemented with those taken from Tabara &Inoue (1980), allow us to determine the values of the RMsfor 16 sources in the list. The observing frequencies of theEffelsberg plus NVSS observations are suitably separatedfor a proper determination of the RMs. In particular, theyallow us to apply unambiguous nπ de-rotation to the ob-served polarisation E-vector position angle χ at the variousfrequencies. When applied, these rotations always yield alinear regression with a least squares fit very close to 1, ex-pect for an optimal fit, for the λ rotation. We also note thatthere is no significant difference between the values derivedfrom our Effelsberg observations and those from Tabara &Inoue (1980), which were acquired about 30 years earlier.
4. Discussion
In the following we discuss trends of the depolarisation in-dex, the achieved values of the rotation measure, and thedistribution of the percentage of polarised emission versusthe linear sizes of the observed sources.
F. Mantovani et al.: Effelsberg polarimetry of CSS sources
The depolarisation index DP has been computed for fourpairs of frequencies. The mean and the median values of theDP are reported in Table 5. Clearly a drop in the fractional
Table 5.
The mean and the median values of the depolar-isation index
Depolarisation index DP mean DP median DP . / . ± − . . DP . / . ± − . . DP . / . ± − . . DP . / . ± − . . polarisation appears in the radio emission at frequenciesbelow ∼ m derived from our Effelsberg observa-tions complemented with those extracted from the NVSSat 1.4 GHz and those taken from the Tabara & Inoue (1980)catalogue. The general trend is a quick decrease in the frac-tional polarisation with increasing wavelength. However, m does not always drop to zero at longer wavelengths ( λ >
49 cm) as predicted by the Burn (1966) model. The Burnmodel (dashed line) and the Tribble (1991) model (solidline) have been plotted in Figs. 2 to 17, according to Eqs.(1) and (2) in Fanti et al. (2004). For both models, we haveadopted as the intrinsic fractional polarisation m the onemeasured by us at 10.45 GHz. For the Tribble model, whichassumes RM randomly distributed and a distribution of cellsizes, the ratio of the characteristic scale representing thelargest cell scale to the observing beam is taken to be equalto 1. With these assumptions, the model corresponds tothe highest possible polarisation at long wavelengths, thusan upper limit to the expected fractional polarisation be-haviour.Very low fractional polarisation for sources belonging tothis class is also found with interferometric observations.For example, Peck & Taylor (2000) did not detect any lin-ear polarisation with VLBI observations at 8.35 GHz on asample of 21 Compact Symmetric Objects (CSOs are CSSsources that exhibit core and lobe emission on each sideof the core). While this strengthens the assumption thatbeam depolarisation plays a minor rˆole it also means thatthe polarisation characteristics of CSS sources are not yetfully clarified. Rotation Measures were calculated for about 1/3 of thesources in our list using use of our own measurements,NVSS and Tabara & Inoue data. The observed RMs werecalculated for 16 sources with at least two measurementsof the position angles achieved by Effelsberg and NVSS ob-servations. Plots of the position angle of the electric vectorin degrees versus the λ in m are shown in Figs. 2–17.In all cases, the λ law is closely followed for leastsquares estimates of the linear regression close to 1. Tosearch for possible variability, we determined the RM both from our new measurements only and from all availablemeasurements, including those of Tabara & Inoue, whichwere measured about 30 years ago. No significant differ-ence was found.The observing frequencies were suitably separated to en-able an unambiguous rotation of the polarisation E-vectors.By analysing the combination of RM and fractional polar-isation, we measured λ rotation and depolarisation for allsources, which corresponded to depolarisation produced in-side the source, and rotation produced outside. However,since δχ > π , we actually observed not mixed-in gas buta foreground screen. A revised Tribble model such as thatproposed by Rossetti et al. (2008) with a “partial coverage”of the source of radio emission by NLRs may account forthe depolarisation behaviour.Comments are necessary in a few cases.- 3C67: the RM values calculated making use of the threeposition angles of Table 3 give an excellent least squaresfit the by means of linear regression after rotating by both+and − π the nominal value of 59 ◦ . − and –71 rad m − . The last value isclose to that ( −
67 rad m − ) calculated also using the RMslisted in the Tabara & Inoue catalogue. 3C138: the RMvalue calculated with the position angles listed in Table 3is –1 rad m − with a very poor mean least squares fit.- 3C286: the RM value is not given as the position angleof the electric vector (33 ◦ ) for this source is taken as areference to calibrate the position angles at all the observingfrequencies.- 3C455: the least squares fit obtained with the values ofTable 3 is worse than that achieved using all the availabledata.In conclusion, we adopt in all cases the RMs listed inCol. 11 of Table 4 calculated using the present measure-ments of the position angles plus those taken from the cat-alogue by Tabara & Inoue. The RMs in the source restframes range between –20 rad m − for 3C138 and ∼ Fig. 1.
Percentage of polarised emission versus source lin-ear sizes at 8.35 GHz. Dots are Quasars; open circles areGalaxies. . Mantovani et al.: Effelsberg polarimetry of CSS sources 5 rad m − for 3C119. Seven sources have a RM >
400 radm − .Examples of CSS sources with high values of | RM | ( ≥ − in the source rest frame) have been found inthe past (O’Dea 1998 and references therein).Three of them are in our list with measured integratedRMs, namely 3C119, 1442+101 and 3C318 for which wemeasured 3900 rad m − , –1450 rad m − , and 2260 rad m − respectively. For the source 1442+101 (OQ172), observedby Udomprasert (1998) with the VLBA, RMs up to 22400rad m − were found. For 3C119, Kato et al. (1987) found3400 rad m − using single dish measurements. For 3C318,Taylor et al. (1992) determined a RM of 1400 rad m − usingVLA observations. There are however CSS sources in oursample with smaller values of RM.As described in more detail in Appendix 2, the source3C298 shows evidence of temporal variability in the frac-tional polarisation. As an example, we show the percentage of polarised emis-sion versus the source’s linear size in kpc at a frequency of8.35 GHz in Fig. 1. Linear dimensions for sources labelledwith “*” in Table 4, for which z became available after 1990,were calculated. We note that in the plot sources with lin-ear sizes greater than 15 kpc are omitted and sources withpolarised emission below the detection limits are plottedassuming m = 0.The most unexpected result that we find is the distri-bution of points drawn in the plots at the Effelsberg andthe NVSS observing frequencies. Polarised sources are dis-tributed all over the m – LS space. This seems to clash withthe findings of Cotton et al. (2003) and Fanti et al. (2004)for the B3-VLA sample of CSS sources. Cotton et al. (2003)making use of NVSS at 1.4 GHz found that sources smallerthan 6 kpc are weakly polarised, and that polarised sourceshave linear sizes greater than 6 kpc (“Cotton effect”). Boththe jet and the counter-jet are included in the source lin-ear size. This implies a drastic change in the interstellarmedium at about 3 kpc. This result was later confirmed byFanti et al. (2004), who observed the same sample withthe VLA at 4.9 GHz and 8.5 GHz and with the WSRT at2.64 GHz (Rossetti et al. 2008).The main difference between the two samples, i.e., theB3-VLA and the present 3CR&PW sample of CSS sources,is represented by the number of quasars, 5-10% and 49%,respectively. Excluding the quasars, we find that the Cottoneffect is clearly evident again. However, we also note threeoutliers, the galaxies 3C93.1, 3C268.3, and 3C318. On theother hand, it has been found by Rossetti et al. (2008)that the Cotton effect is not obeyed by all sources. Thisindicates that the current depolarisation scenarios mightnot fully explain the observed behaviour. This is also fur-ther indication that CSS sources optically identified withquasars may represent a separate class of objects.
5. Conclusions
With the present observations of a complete sample of CSSsources at 4 different frequencies made with the Effelsberg100-m radio telescope and from archived NVSS data, thefollowing results have been reached: a) It is confirmed that CSS sources are weakly po-larised, low values of the median fractional polarisation be-ing found. Where m can be compared with those obtainedby Klein et al. (2003) for the steep spectrum extended ra-dio sources selected in the B3-VLA sample, larger valuesare found, which are 2.2% at 1.4 GHz, 3.7% at 2.64 GHz,5.2% at 4.85 GHz, and 5.8% at 10.45 GHz. The more sen-sitive VLA observations at 1.4 GHz of the CSS sources inthe 3CR&PW sample extracted from the NVSS also showa median value for the fractional polarisation of ∼ ∼ − for 3C138and ∼ − for 3C119. Seven sources show a RM >
400 rad m − , confirming the high values of the RM usu-ally found for CSS sources.e) In all cases the λ law is followed closely. The observ-ing frequencies were suitably separated to enable an unam-biguous rotation of the polarisation E-vectors. Analysingthe combination of RM and fractional polarisation, we iden-tified λ rotation and depolarisation for all sources. A re-vised Tribble model such as that proposed by Rossetti etal. (2008) with a “partial coverage” of the source of radioemission by NLRs may account for the depolarisation be-haviour.f) The m - LS diagram contains sources polarised atany linear size and at all available frequencies, includingthe NVSS 1.4 GHz data, i.e., also for sources smaller than6 kpc in linear size, in contrast to the so-called “Cottoneffect” later confirmed by Fanti et al. (2004) for the B3-VLA sample of CSS sources. However, plotting m versus LS for objects optically identified with galaxies, the so-called“Cotton effect” is reproduced. Discussing the interesting,large difference in the percentage of objects identified asquasars in the two samples was not a subject of the presentinvestigation.g) CSS sources optically identified with quasars mayrepresent a separate class of objects. Acknowledgements.
This work is based on observations with the 100-m telescope of the MPIfR (Max-Planck-Institut f¨ur Radioastronomie)at Effelsberg. It has benefited from research funding from theEuropean Community’s sixth Framework Programme under RadioNetR113CT 2003 5058187. FM likes to thank Prof. Anton Zensus,Director, for the kind hospitality at the Max-Planck-Institut f¨urRadioastronomie, Bonn, for a period during which part of this workwas done. The National Radio Astronomy Observatory is a facility ofthe National Science Foundation operated under cooperative agree-ment by Associated Universities, Inc.
References
Akujor, C.E. & Garrington, S.T. 1995, A&AS 112, 235Aller, M.F., Aller, H.D. & Huges P.A. 2003, ApJ 518, 33
F. Mantovani et al.: Effelsberg polarimetry of CSS sources
Baars, J.W.M., Genzel, R., Pauliny-Toth, I.I.K. & Witzel A. 1977,A&A 61, 99Burn, B.J. 1966 MNRAS 133, 67Cohen, A.S., Lane, W.M., Cotton, W.D. et al. 2007, AJ 134, 1245Condon, J.J., Cotton, W.D., Greisen, E.W. et al. 1998, AJ 115, 1693Cotton, W.D., Dallacasa, D., Fanti, C. et al. 2003, PASA 20, 12Dreher, J.W., Carilli, C.L. & Perley, R.A. 1987, ApJ 316, 611Fanti, R., Fanti, C., Schilizzi, R.T. et al. 1990, A&A 231, 333Fanti, C., Fanti, R., Dallacasa, D. et al. 1995, A&A 302, 317Fanti, C., Pozzi, F., Dallacasa, D. et al. 2001, A&A 369, 380Fanti, C., Branchesi, M., Cotton, W.D. et al. 2004, A&A 427, 465Hes, R., Barthel, P.D. & Fosbury, R.A.E. 1996, A&A 313, 423Jenkins, C.J., Pooley, G.G.& Riley, J.M. 1977, MNRAS 84, 61Kato, T., Tabara, H., Inoue, M. & Aizu, K. 1987, Nat 329, 223Klein, U., Mack, K.-H., Gregorini, L. & Vigotti, M., 2003, A&A 406,579Labiano, A., Barthel, P. D., O’Dea, C. P. et al. 2007, A&A 463, 97LLaing, R. 1984, Proceeding NRAO Workshop No.9, 90Leahy, J.P. 1990,
Parsec-Scale Radio Jets, eds. Zensus, J.A. &Pearson, T.J., Cambridge University Press, CambridgeMontenegro-Montes, F.M., Mack, K.-H., Vigotti, M. et al. 2008,MNRAS 388, 1853Murgia, M., Fanti, C., Fanti, R. et al. 1999 A&A 345, 769O’Dea, C. 1998, PASP 110, 493Peacock, J.A. & Wall, J.V. 1982, MNRAS 198, 843Peck, A.B. & Taylor G.B. 2000, ApJ 534, 90Polatidis, A.CG., Conway, J.E. 2003, PASA 20, 69Readhead, A.C.S., Taylor, G.B., Xu, W. et al. 1996, ApJ 460, 612Rossetti, A., Mantovani, F., Dallacasa, D. et al. 2005, A&A 434, 449Rossetti, A., Fanti, C., Fanti, R. et al. 2006, A&A 449, 49Rossetti, A., Dallacasa, D., Fanti, C. et al. 2008, A&A 487, 865Saikia, D.J., Singal, A.K. & Cornwell, T.J. 1987, MNRAS 224, 379Saikia, D.J. & Gupta N. 2003, A&A 405, 499Stanghellini, C., O’Dea, C.P., Baum, S.A. & Laurikainen, E. 1993,ApJS 88, 1Stanghellini, C., O’Dea, C.P., Dallacasa, D. et al. 2005, A&A 443, 891Tabara H. & Inoue M. 1980, A&AS 39, 373Taylor, G.B., Inoue, M. & Tabara, X. 1992, A&A 264, 421Tribble, P.C. 1991, MNRAS 250, 726Udomprasert, P.S., Taylor, G.B., Pearson, T.J. & Roberts, D.H. 1997ApJ 483, L9van Breugel, W., Miley, G. & Heckman, T. 1984, ApJ 276, 79Willott, C.J, Rawlings, S., Blundell, K.M. & Lacy, M. 1998, MNRAS300, 625 . Mantovani et al.: Effelsberg polarimetry of CSS sources 7
Table 2.
Flux densities and polarised flux densities from Effelsberg 100-m and NVSS measurements.
Name S . S . S . S . S . S p . S p . S p . S p . S p . [mJy] [mJy] [mJy] [mJy] [mJy] [mJy] [mJy] [mJy] [mJy] [mJy]3C43 0127+23 29414 ±
88 1787 ±
37 1109 ±
22 698 ±
14 578 ± < < < ± ± ±
481 9373 ±
191 5534 ±
113 3243 ±
65 2495 ±
52 70.8 ± < ± ± ± ±
82 1552 ±
33 878 ±
18 494 ±
10 376 ± < < < < < ±
91 1757 ±
36 997 ±
20 575 ±
12 419 ± ± < < ± ± ±
87 2483 ±
51 2234 ±
46 1452 ±
30 1212 ± < < < < < ±
241 5038 ±
103 2937 ±
60 1563 ±
32 1187 ± < < < < < ±
67 1704 ±
35 1791 ±
37 1527 ±
31 1549 ±
32 71.8 ± ± ± ± ± ±
71 1390 ±
29 926 ±
19 477 ±
10 383 ± < < < < ± ±
169 4042 ±
83 3355 ±
69 2087 ±
42 1648 ± < < < < < ±
112 3166 ±
65 2761 ±
57 1542 ±
31 1289 ± < < < < < ±
295 6126 ±
126 4722 ±
97 2677 ±
54 2260 ± < < < ± ± ±
258 5780 ±
119 4119 ±
84 2552 ±
53 2160 ±
45 619.0 ± ± ± ± ± ±
686 13436 ±
276 10120 ±
207 4735 ±
96 3823 ± < < < < ± ±
37 615 ±
14 271 ± ± ± < < < < < ±
82 1359 ±
28 746 ±
16 426 ±
10 360 ± < < < < < ±
127 1446 ±
50 1731 ±
35 1413 ±
29 1337 ± < < < < < ±
196 3608 ±
74 1971 ±
42 1034 ±
21 823 ± < < < < ± ±
51 799 ±
17 350 ± ±
33 112 ± < < < < < ±
89 1751 ±
36 1030 ±
21 545 ±
11 441 ± < < < < ± ±
112 1947 ±
39 1153 ±
23 618 ±
13 479 ± < < < ± ± ±
63 1434 ±
30 753 ±
16 353 ± ± < < < < < ±
69 1343 ±
30 790 ±
16 481 ±
10 384 ± < < < ± ± ±
146 3369 ±
70 2293 ±
47 1540 ±
31 1293 ± < ± < < < ±
447 10607 ±
219 7430 ±
152 5179 ±
104 4474 ±
94 999.9 ± ± ± ± ± ±
212 4697 ±
97 3130 ±
64 2050 ±
42 1748 ±
36 45.6 ± ± ± ± ± ±
129 2765 ±
57 1701 ±
35 1080 ±
22 888 ± < < < < < ±
56 1488 ±
30 1059 ±
22 735 ±
15 621 ± < ± < < < ±
183 2888 ±
60 1432 ±
30 789 ±
17 616 ± < < < ± ± ±
111 1721 ±
35 922 ±
20 502 ±
10 400 ± < < < < < ±
72 1398 ±
38 1032 ±
21 641 ±
14 516 ±
11 31.7 ± ± ± ± < ±
66 882 ±
19 413 ± ± ± < < < < < ±
50 837 ±
18 412 ± ± ± < < < < < ±
224 5129 ±
105 3552 ±
73 2568 ±
52 2196 ±
46 91.1 ± ± ± ± ± ±
81 1417 ±
29 764 ±
16 417 ± ± < < ± ± ± ±
90 1980 ±
41 1433 ±
29 1049 ±
21 915 ± < < < < < ±
147 3140 ±
65 1728 ±
35 902 ±
18 681 ± < < < < < ±
150 2858 ±
59 1503 ±
31 808 ±
16 622 ± < < < < ± ±
138 2384 ±
49 1190 ±
24 631 ±
13 479 ± < < < < < ±
110 2260 ±
47 1426 ±
29 915 ±
19 773 ±
17 80.6 ± < < < ± ±
105 1871 ±
42 938 ±
19 463 ±
10 337 ± < < < < < ±
413 8139 ±
167 5073 ±
105 3523 ±
71 3055 ±
63 62.2 ± < < < < ±
88 1945 ±
40 1168 ±
24 679 ±
14 533 ± < < < < < ±
216 5695 ±
118 4135 ±
85 3396 ±
69 3209 ±
67 114.0 ± ± ± ± ± ±
64 1265 ±
26 761 ±
16 469 ±
10 383 ± ± ± ± ± ± ±
47 696 ±
15 287 ± ± ± < < < < < ±
81 1491 ±
30 772 ±
16 421 ±
93 312 ± ± ± < < ± ±
113 2275 ±
47 1308 ±
27 747 ±
15 579 ± < < < < < The values at the various observing frequencies (in GHz) are organised as follows: column 1, source name; column 2,other name; columns 3 to 7, flux density S; columns 8 to 12, polarised flux density S p . F. Mantovani et al.: Effelsberg polarimetry of CSS sources
Table 3.
Percentage of polarised flux density and position angle of the electric vector at the five frequencies.
Name
LS m . m . m . m . m . χ . χ . χ . χ . χ . [kpc] [%] [%] [%] [%] [%] [deg] [deg] [deg] [deg] [deg]3C43 0127+23 12.85 2.1 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± + + ± ± ± ± ± ± ± ± ± ± ± ± + OF247 0428+20 0.493C119 0429+41 0.9 5.9 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± + ± ± ± ± + ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± + ± ± ± ± ± ± + ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± < + ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± + Table 3 is organized as follows: column 1, source name; column 2, other name; column 3, source linear size. Asterisksmark sources for which new redshifts have been obtained and new linear sizes were calculated; + signs mark sourceswith improved measurements of their angular sizes (Rossetti et al. 2005; 0319+12 Mantovani in prep.; 1442+10Udomprasert et al. 1997); columns 4 to 8, percentage of polarised emission m ; columns 9 to 13, position angle of theelectric vector χ . . Mantovani et al.: Effelsberg polarimetry of CSS sources 9 Table 4.
Depolarisation indices and RMs.
Name z Id DP DP DP DP RM obs lm-1 RM obs + TI lm-2 RM rf . / .
45 4 . / .
35 2 . / .
85 1 . / . [rad m − ] [rad m − ] [rad m − ]3C43 0127+23 1.46 Q 1.00 −
69 0.9965 − −
79 0.9994 −
64 0.9900 − − −
67 0.9945 − − −
12 0.9521 −
19 0.7698 − − − − −
59 0.9961 − −
148 0.9957 −
146 0.9954 − −
66 0.9994 −
71 0.9376 − −
39 0.9694 −
38 0.9894 − −
88 1.000 −
88 0.9997 − −
68 0.9402 82 0.9923 1942342+82 0.74 Q
Table 4 is organized as follows: column 1, source name; column 2, other name; column 3, redshift; redshifts markedwith “*” are taken from the following pubblications: 0223+34 Willott et al. (1998); 0316+16 Labiano et al. (2007);1225+36 Lawrence et al. (1994); 1600+33 photometric redshift from Stanghellini et al. (2005); 1607+26 Stanghellini etal. (1993); column 4, optical identification. The source 3C455 has been re-classified as galaxy by Hes et al. (1966);columns 5 to 8, Depolarisation Indices DP; column 9, Rotation Measure RM derived from our new Effelsbergobservations; column 10, ls-1: the least square fit of the linear regression of the values in column 9; column 11, RotationMeasure RM derived from our new Effelsberg observations, complemented with data from Tabara & Inoue (1980);column 12, ls-2: the least square fit of the linear regression of the RM values in column 11; column 13, rest frame RMcalculated using the RM obs + T I values.
Appendix 1
In this Appendix, we present plots of the fractional po-larisation m derived form our Effelsberg observations com-plemented with those extracted from the NVSS at 1.4 GHzand those taken from Tabara & Inoue (1980). The values of m predicted by the models of Burn and Tribble are shownfor comparison. Also plotted are polarisation angles versus λ with their corresponding linear best fit. . Mantovani et al.: Effelsberg polarimetry of CSS sources 11 Fig. 2.
Position angles of the electric field vector χ in deg (dots) and fractional polarisation m in % (triangles) versus λ in m for the source 3C43, for the full range ( left ), and for a narrow range ( right ) of wavelengths. The solid line representsthe Tribble model, the dashed line the Burn model. Fig. 3.
Position angles χ and fractional polarisation m for the source 3C48. Layout as in Fig. 2. Fig. 4.
Position angles χ and fractional polarisation m for the source 3C67. Layout as in Fig. 2. Fig. 5.
Position angles χ and fractional polarisation m for the source 0319+12. Layout as in Fig. 2. . Mantovani et al.: Effelsberg polarimetry of CSS sources 13 Fig. 6.
Position angles χ and fractional polarisation m for the source 3C119. Layout as in Fig. 2. Fig. 7.
Position angles χ and fractional polarisation m for the source 3C138. Layout as in Fig. 2. Fig. 8.
Position angles χ and fractional polarisation m for the source 3C268.3. Layout as in Fig. 2. Fig. 9.
Position angles χ and fractional polarisation m for the source 3C277.1. Layout as in Fig. 2. . Mantovani et al.: Effelsberg polarimetry of CSS sources 15 Fig. 10.
Position angles χ and fractional polarisation m for the source 3C277.1. Layout as in Fig. 2. Fig. 11.
Position angles χ and fractional polarisation m for the source 3C277.1. Layout as in Fig. 2. Fig. 12.
Position angles χ and fractional polarisation m for the source 1442+10. Layout as in Fig. 2. Fig. 13.
Position angles χ and fractional polarisation m for the source 3C309.1. Layout as in Fig. 2. . Mantovani et al.: Effelsberg polarimetry of CSS sources 17 Fig. 14.
Position angles χ and fractional polarisation m for the source 3C318. Layout as in Fig. 2. Fig. 15.
Position angles χ and fractional polarisation m for the source 4C11.69. Layout as in Fig. 2. Fig. 16.
Position angles χ and fractional polarisation m for the source 3C454. Layout as in Fig. 2. . Mantovani et al.: Effelsberg polarimetry of CSS sources 19 Fig. 17.
Position angles χ and fractional polarisation m for the source 3C455. Layout as in Fig. 2. Appendix 22.1 Sources following the Tribble model
The model proposed by Tribble reproduces the dataof about one third of our sample, namely of 3C48, 3C67,3C119, 3C138, 3C268.3, 3C277.1, and 3C318. Sources suchas 3C48, 3C67, 3C138, and 3C277.1 also show RM << ∼ few10 rad m − , which is an indication of an unresolved fore-ground screen, presumably the halo of our own Galaxy.The source 3C119, which has a very high RM and thusa fast decline in its fractional polarisation, might alreadybe affected by significant depolarisation at wavelengthsshorter than 2.8 cm. A higher m would lead to depolari-sation according to the Tribble law. Eight sources (3C43, 0319+12, 3C287, 1442+10,3C309.1, 4C11.69, 3C454, and 3C455) show indications ofrepolarisation, i.e., an increase in fractional polarisationwith decreasing frequency, generally at short wavelengths < ∼
10 cm; this corresponds to a relatively strong increasein fractional polarisation followed by a decline that canstill be described by the Tribble law. This effect is visi-ble despite possible instrumental effects of different tele-scopes and also considering that observations were madeat different epochs. In three of these sources (3C309.1,4C11.69, 3C454), the repolarisation effect is indeed con-fused with possible time variability. Focusing on our si-multaneous measurements only, we still find constant orslightly increasing fractional polarisation with increasingwavelength, which is not predicted by any depolarisationmodel. In particular, the galaxy 3C455 shows a measuredrepolarisation greater than the 3 σ level between 10.45 GHzand 2.64 GHz together with a small value of the RM rf (194rad m − ), which is an indication of the influence of a fore-ground screen, possibly an extended cloud with [OII] emis-sion detected by Hes et al. (1996). A plausible mechanismwould be the effect of shear layers caused by the interac-tion between the surface of an expanding source and thesurrounding medium (Burn 1966).The source 3C455 was observed with the VLA at8.35 GHz by Bogers et al. (1994). It shows a triple structurewith the indication of a jet joining the core with the southwestern lobe. The three components are almost alignedalong the source major axis extending up to about 4 ′′ .However, 3C455 appears slightly resolved by the NVSS,which has a restoring beam of 45 ′′ , suggesting that thethree components imaged by Bogers et al. are actually em-bedded in a more extended region of low brightness emis-sion. This can be seen in the image available in the VLALow-Frequency Sky Survey (VLSS; Cohen et al. 2007) at74 MHz, which shows an even more extended structure ofabout 2 ′ in size. The existence of this extended emissionis supported by the source spectral index, which indicatesan upturn towards higher flux density above ∼
100 MHz.Therefore, a second possible interpretation is that by ob-serving at 2.64 GHz or lower frequencies, we have integratedthe flux density and polarised flux density from that region.At 10.45 GHz, the steep spectrum extended structure is be-low the detection limit.A similar case of repolarisation at a lower frequencywas pointed out by Montenegro-Montes et al. (2008) forthe source 1159+01.2.3 Polarisation variability