Study of the extended radio emission of two supernova remnants and four planetary nebulae associated to MIPSGAL bubbles
Adriano Ingallinera, Corrado Trigilio, Grazia Umana, Paolo Leto, Claudia Agliozzo, Carla Buemi
MMon. Not. R. Astron. Soc. , 1– ?? (2014) Printed 11 September 2018 (MN L A TEX style file v2.2)
Study of the extended radio emission of two supernova remnantsand four planetary nebulae associated to MIPSGAL bubbles
A. Ingallinera , (cid:63) , C. Trigilio , G. Umana , P. Leto , C. Agliozzo , , C. Buemi Universit`a di Catania, Dipartimento di Fisica e Astronomia, Via Santa Sofia, 64, 95123 Catania, Italy INAF-Osservatorio Astrofisico di Catania, Via Santa Sofia 78, 95123 Catania, Italy Millennium Institute of Astrophysics, Santiago, Chile Universidad Andr´es Bello, Avda. Republica 252, Santiago, Chile
Accepted 2014 October 2. Received 2014 September 19; in original form 2014 August 4
ABSTRACT
We present radio observations of two supernova remnants and four planetary nebulae with theVery Large Array and the Green Bank Telescope. These objects are part of a larger sample ofradio sources, discussed in a previous paper, counterpart of the MIPSGAL 24- µ m compactbubbles. For the two supernova remnants we combined the interferometric observations withsingle-dish data to obtain both a high resolution and a good sensitivity to extended structures.We discuss in detail the entire combination procedure adopted and the reliability of the re-sulting maps. For one supernova remnant we pose a more stringent upper limit for the fluxdensity of its undetected pulsar, and we also show prominent spectral index spatial variations,probably due to inhomogeneities in the magnetic field and in its ejecta or to an interactionbetween the supernova shock and molecular clouds. We eventually use the 5-GHz maps ofthe four planetary nebulae to estimate their distance and their ionized mass. Key words: techniques: image processing – planetary nebulae: general – ISM: supernovaremnants.
From the visual inspection of the MIPSGAL Legacy Survey (Careyet al. 2009) mosaic images, obtained with MIPS (Rieke et al.2004) on board of the Spitzer Space Telescope , Mizuno et al. (2010)identified at 24 µ m 428 compact roundish objects presenting dif-fuse emission. These small ( (cid:46) (cid:48) ) rings, disks or shells (hereafterdenoted as ‘bubbles’) are pervasive throughout the entire Galacticplane in the mid-infrared (IR). The main hypothesis about the na-ture of the bubbles is that they are different type of evolved stars(planetary nebulae, supernova remnants, Wolf–Rayet stars, asymp-totic giant branch stars, etc.). However, currently, only about 30 percent of the bubbles are classified.In order to shed light on the nature of the bubbles, we had car-ried out radio continuum observations with the Karl G. Jansky VeryLarge Array (VLA) in 2010 ( C band, configuration D) and 2012 ( L band, configuration C and CnB), on a subset of 55 bubbles (In-gallinera et al. 2014; ‘Paper I’ hereafter). We were able to calculatethe radio spectral index for 31 of them, finding that at least the 70per cent are radio thermal emitters. Among the 55 bubbles observedat 5GHz, ten were classified and, mostly, well known objects. Wehad excluded these objects from the remainder of that work, mainlyaimed at characterising unclassified objects. (cid:63) E-mail:[email protected] The Multiband Imaging Photometer for Spitzer.
In this paper we present an analysis of the VLA radio con-tinuum observations described in Paper I for six of these ten clas-sified bubbles, namely two supernova remnants (SNRs) and fourplanetary nebulae (PNe), addressed to improve the knowledge oftheir physical properties. For the two SNRs, whose extension is ∼ ∼ The interferometric observations and the following data reductionwere described in detail in Paper I. Therefore here we will limit toa brief summary.From the original list of the 428 MIPSGAL bubbles we se-lected only sources visible at the VLA latitude and with a possible c (cid:13) a r X i v : . [ a s t r o - ph . I M ] O c t A. Ingallinera et al. detection in NVSS or in MAGPIS . The final subset consistedof 55 bubbles, that were observed at 5GHz ( C band) in March2010 with the VLA in configuration D. Among these 55 bubbles,40 were subsequently observed at 1 . L band) in March andMay 2012 in configuration C and CnB. For C -band observations,each bubble was observed for slightly less than 10 minutes with atotal bandwidth of 256MHz. The average synthetic beam size wasaround 25 × and the typical map rms ∼ µ Jy / beam.For L -band observations, each bubble was observed for 10 to 20minutes with a total bandwidth of 1GHz. The average beam sizewas 18 × . The presence of conspicuous radio-frequencyinterferences limited the sensitivity reached in these observations,and the typical map rms was about 0 . − / beam.The entire data reduction process was performed using thepackage CASA . For all the observations, the bandpass and flux cal-ibrator was 3C286, while for gain calibration, we used a variety ofstandard calibrators, depending on their distance from the sources(typically within 10 ◦ ).Differently from Paper I, the imaging procedure involved theuse of the ‘multi-scale algorithm’ (Cornwell 2008). While the stan-dard Clark implementation (Clark 1980) of the CLEAN algorithm(H¨ogbom 1974) assumes that the sky is a collection of point sourceson an empty background, the multi-scale CLEAN permits the pres-ence of sources of many different sizes and scales. This algorithmis therefore extremely useful when an extended source is to be im-aged. In particular we defined three different scales for source ex-tensions: 0 pixel to search for point sources (equivalent to a normalCLEAN and necessary to properly consider nearby point sources),5 pixels (about the beam dimension) and 15 pixels.In Table 2.1 we list all the sources studied in this paper, in-dicating their main designation, the Mizuno et al. (2010) name([MGE]) and whether we observed them with the VLA and withthe GBT. An interferometer is an instrument capable to measure the visibil-ity function V ( u , v ) that, under certain circumstances, is the Fouriertransform of the true sky brightness (e.g. Clark 1999). However, theinstrument is able to measure V ( u , v ) only on a discrete set of ( u , v ) pairs, given by the interferometer baseline projections on the uv plane. Even if the Earth rotation is exploited, the uv plane coveragewill be always incomplete. In particular there will be a minimumvalue for √ u + v below which no data can be acquired. This min-imum value is the projection of the interferometer shortest baseline,and the corresponding angular extension in the real plane is calledthe ‘largest angular scale’ (LAS).The impossibility to measure the visibility function at smalldistances from the uv plane origin has important consequences inthe data imaging. The lack of data in this region of the uv planetranslates to a poor interferometer sensitivity to extended sources.There is no a net extension limit between sources that can be wellimaged by interferometer and source that cannot be. We can safelystate that only sources with angular dimension significantly smallerthan the LAS can be reliably imaged by the interferometer, and thatthe more extended a source is the more poorly will be imaged. http://third.ucllnl.org/gps/catalogs.html (Helfand et al. 2006). http://casa.nrao.edu (McMullin et al. 2007). The imaging artefacts produced by the missing low spatial fre-quencies prevent the total flux density recovery of the observedsource. Therefore this issue, known as the ‘short-spacing problem’or the ‘flux-loss problem’, prevents also a reliable reconstruction ofthe source spectral energy distribution.
In order to fill in the gap in the central zone of the uv plane and tocross-calibrate the single-dish and interferometer data, we neededto use a single-dish telescope whose diameter was much greaterthan the VLA minimum baseline, equal to 35m both in configura-tion C and D. For this reason we chose to observe with the GBT,which has an effective diameter of 100m.The observations were carried out in June 2011 at 1475MHz( L band) and at 5100MHz ( C band), with both the receivers collo-cated in the Gregorian focus of the GBT. For the L band the totalbandwidth was 20MHz, while for the C band 80MHz. As back-endwe used the Digital Continuum Receiver.The observing strategy was to map a sky region slightly widerthan the VLA field of view. The GBT is equipped with a single-feed receiver but a map can be still obtained by letting the tele-scope move and scan the desired sky region. In particular we usedthe ‘on-the-fly mapping’ technique: in this mode, the telescope isslewed within a rectangular area of the sky while it acquires thedata and the map is built piling linear scans. Paying attention that,given the telescope beam, an appropriate sampling is made for eachscan and that two adjacent linear scans are taken sufficiently closeto each other, then the Nyquist sampling theorem guarantees that areliable map of the sky can be obtained in the limits of instrumentalerrors. In order to reduce background and instrumental fluctuations,we set a slewing velocity high enough to execute scans of 0 .
3s perbeam (whose width was chosen to be 3 pixels). This allowed us toreach the desired sensitivity of 5 mJy beam − in C band and to ob-serve at the confusion level ( ∼ − ) in L band. The mapdimensions are 25 × for the C band and 75 × for the L band. The scan separation is 50arcsec for C band and150arcsec for L band, with the expected beam FWHM respectively2 . c (cid:13) , 1– ?? adio emission of two SNRs and four PNe Designation [MGE] Right ascension Declination Observations(J2000) (J2000) VLA 5GHz VLA 1 . . − ◦ (cid:48) (cid:48)(cid:48) yes no yesSNR G027.4+00.0 027.3891-00.0079 18:41:19 . − ◦ (cid:48) (cid:48)(cid:48) yes yes yesPN G029.0+00.4 029.0784+00.4545 18:42:46 . − ◦ (cid:48) (cid:48)(cid:48) yes yes yesNGC 6842 065.9141+00.5966 19:55:02 . + ◦ (cid:48) (cid:48)(cid:48) yes no yesPN G040.3-00.4 040.3704-00.4750 19:06:45 . + ◦ (cid:48) (cid:48)(cid:48) yes yes noPN G031.9-00.3 031.9075-00.3087 18:50:40 . − ◦ (cid:48) (cid:48)(cid:48) yes yes no Table 1.
List of the sources studied in this work.
Figure 1.
A schematic illustration of the two intersecting scan series usedin this work. The shadowed area is the sky region to map; first the telescopeis moved along the horizontal path, with a data acquisition at each circle,then the piled scans are performed in the orthogonal direction, followingthe vertical path (crosses represent acquisition points).
The map construction method described in Section 2.3 producesimportant artefacts in the final image. In particular, instrumentaldrifts during mapping time usually give rise to a different baselevelfor each scan and the resulting image appears affected by ‘stripes’,running in the same direction of the scans. The simplest way totake into account these baselevel drifts is to estimate their contri-bution in a region supposed to be free of emission. This can beobtained by letting the scans be more extended than the region tomap and then using these extra-points, at both side of the map area,to linearly fit a baselevel that will be subtracted from each scan.In many circumstances, however, this method may not give a satis-factory result. Emerson & Gr¨ave (1988) (EG88) pointed out threemain limitations: first, there can be emission at the end of eachscan; second, also the scan ends are affected by noise which canhave great effects on the baseline determination; third, it is possi-ble that a baseline cannot be adequately represented by a linear fit.In their paper, they discussed the general problem of the reduction of scanning noise in raster scanned data, with no limitations to thesingle-dish single-feed radio telescope case presented above. Thesolution proposed by EG88 was to filter out the spatial frequenciesresponsible for stripes by means of opportune weights to the datain the Fourier domain.One of the most important problems when one maps a portionof the sky in the Galactic plane, with the resolution of the GBT, isthat there are no large regions free of emission in the neighbour-hood of any source. Therefore it is mandatory to use a destripingalgorithm such as the one by EG88, since no reliable baselevel es-timation at the ends of the scan can be made. However the strategyto create the maps with scans extended only few arcminutes withthe telescope scanning quickly gives us two important advantages:the systematic errors along each scan can be considered constantand the small amount of pixels constituting the map allows to workmore surgically in the Fourier plane. Therefore we modified theEG88 algorithm to better fit our problem. A detailed description ofthe adopted solution is reported in appendix A1.The main advantage of this method, compared to the EG88 al-gorithm, is the preservation of the information from all spatial fre-quencies. In particular the central pixel in the Fourier plane main-tains its value. This is very important because, as we said, its valueis the sum of all pixel values and therefore if it were (almost) zeroseveral negative-valued features would appear in the destriped im-age. Even if in the real image it accounts only for a flat base level(i.e. even a zero central pixel in the Fourier plane does not introducefurther artefacts), it may play a role in the combination process. Thepreservation of the central pixel in the Fourier plane guarantees themap total flux density is conserved after the destriping procedure.However this could be not true for a single source. We tested thispossible flux density alteration comparing sources from pre- andpost-destriped maps (for details see Appendix A2). We found thatthe error introduced by the destriping process is negligible withrespect to the other source of error discussed in this work (like cal-ibration errors or background noise).
As discussed, one of the goal of this work is to combine interfero-metric and single-dish maps. However, despite the presence of sev-eral works regarding the flux-loss problem and the combination ofinterferometric and single-dish data, only a few reliable implemen-tations of the theoretical algorithms found in literature were avail-able. Our choice was to use the new
CASA tool,
CASAFEATHER ,introduced in version 4.1.The theory behind the combination we aimed to perform israther simple. It is based on the fact that a single-dish telescopemeasures not only a single spatial frequency, but a whole range ofcontinuous spatial frequencies up to a maximum one, which cor- c (cid:13) , 1–, 1–
CASAFEATHER ,introduced in version 4.1.The theory behind the combination we aimed to perform israther simple. It is based on the fact that a single-dish telescopemeasures not only a single spatial frequency, but a whole range ofcontinuous spatial frequencies up to a maximum one, which cor- c (cid:13) , 1–, 1– ?? A. Ingallinera et al. responds to the diameter of the dish, D . Hence, a single-dish tele-scope behaves as an interferometer with a continuous range of base-lines, from zero to D . The Fourier transform of a single-dish imageis a two-variable complex function that can be interpreted as the‘intensity’ of each bidimensional spatial frequency. On the otherhand, an interferometer directly observes the source in the Fourierdomain, measuring the visibility function, but it is not able to fillthe entire Fourier plane, and in particular it will not cover the innerregion of this plane up to a spatial frequency corresponding to theshortest baseline. This ‘hole’ is usually filled during the imagingprocess through some kind of interpolation, which, however, can-not provide reliable values for the missing visibility (more preciselyit should be treated as an extrapolation).If the single-dish telescope diameter is greater than the short-est baseline of the interferometer, there would be an annular regionin the Fourier plane inside which we have both interferometric andsingle-dish data for the source, i.e. the single-dish and the inter-ferometer are sensitive to the emission related to some commonspatial frequencies. The trick now is to use this overlapping regionin the Fourier plane to cross-calibrate the data and make them ascoincident as possible in this region. This operation is known as‘feathering’. The CASAFEATHER tool allows the user to controlsome parameters involved in the feathering process. In particularwe set the effective dish diameter to 80m and single-dish imagescaling factor from 1 to 2 .
5. We can then merge the two datasetsto obtain a single visibility function defined from the zero baselineto the longest interferometer baseline. An inverse transformationof the visibility obtained in this way results in an image with thesame resolution of the original interferometric image, but also withwell determined low spatial frequencies. From this combined im-age a total flux density recovery is possible. It is hard to determinethe reliability of this measurement. We found however that the fluxdensities should be corrected within 5 or 10 per cent, depending onthe particular map (see Appendix A2).The actual result of the combination varied appreciably forthe different sources. For the two SNRs the combined maps showa significant improvement in terms of image quality as well as anincreasing in total flux density. In particular the recovered extendedemission is now clearly spatially coincident with the diffuse emis-sion present in the MIPSGAL images. In the GBT images, both thetwo SNRs are characterised by a high total flux density at 5GHz( > We discuss now on the results obtained for the two SNRs and thefour PNe. We report a very brief summary of state-of-art knowledgefor each one of them. We present then our maps and derive fromthem different physical parameters that complement or improve thevalues available in literature.
The SNR G011.2-00.3 is a well studied radio, IR and X-ray source.Very likely, it is the remnant of the historical SN 386 (Reynolds etal. 1994). The radio emission associated to this object has been firstdetected at 2 . α line. The source wassubsequently observed in a wide range of radio frequencies, from30 . . . . . . . ± . The radio emission associated to SNR G027.4+00.0, also knownas KES 73, is located in a particularly crowded region. At radiowavelengths, the entire region was first detected as a single discretesource at 408MHz (Large, Mathewson & Haslam 1961). Higherresolution images at 5 GHz resolved the huge radio emission re-gion in at least four discrete components. One of them presented aclearly non-thermal emission and was proposed as a SNR (Milne1969; Angerhofer, Becker & Kundu 1977).Kriss et al. (1985) gave the first evidence that the SNR har-boured a compact source, by means of X-ray observations. Thiscompact source, designed 1E 1841-045, is an anomalous X-ray pul-sar with a period of 11 .
8s (Vasisht & Gotthelf 1997). It is charac-terised by an extremely intense magnetic field (Gotthelf, Vasisht &Dotani 1999) and shows soft γ -ray repeater like bursts (Kumar &Safi-Harb 2010; Lin et al. 2011). It is now an established magnetar(Kumar et al. 2014). c (cid:13) , 1– ?? adio emission of two SNRs and four PNe D e c ( J ) D e c ( J ) Figure 2.
GBT images of SNR G011.2-00.3 at 1 . . D e c ( J ) Figure 3.
High resolution images of SNR G011.2-00.3 at 5GHz. The imageis the result of feathering VLA (configuration D) and the GBT map.
Studies on H I absorption line toward the source pose the SNRat a distance of about 7 . (cid:46) (cid:38) M (cid:12) (Got-thelf & Vasisht 1997; Tian & Leahy 2008; Kumar et al. 2014).In Figure 4 we report the GBT images we obtained respec-tively at 1 . . . CASAFEATHER . Though the source is less extended than the VLALAS at this frequency we further combined this high resolution im-age with the GBT map. The result is shown in Figure 5 (left). Thecalculation of the total flux density of the SNR proved hard becauseof the peculiar background. We measured a value of 5 . ± . . σ level (previous estimates gave anupper limit of 0 . . ± . . ± . α = − . ± .
18, in perfect agreement withprevious estimates in other spectral region (the turnover is around100MHz; Caswell et al. 1982; Kassim 1989). Starting from the twoimages presented in Figure 5 we were also able to build a spectralindex map. In particular we convolved both the images with an op-portune bidimensional Gaussian in order to obtain for both a cir-cular beam of 25arcsec. The convolved images were subsequentlyregridded to have a pixel-by-pixel correspondence and the spectralindex map was created. The result is shown in Figure 6. Not only aprominent spectral index variation is evident but it is also possibleto notice a flatter index toward the pulsar with respect to the globalvalue, as well as an important asymmetry, with the western regionflatter than the eastern one. A spectral index of about − . c (cid:13) , 1–, 1–
18, in perfect agreement withprevious estimates in other spectral region (the turnover is around100MHz; Caswell et al. 1982; Kassim 1989). Starting from the twoimages presented in Figure 5 we were also able to build a spectralindex map. In particular we convolved both the images with an op-portune bidimensional Gaussian in order to obtain for both a cir-cular beam of 25arcsec. The convolved images were subsequentlyregridded to have a pixel-by-pixel correspondence and the spectralindex map was created. The result is shown in Figure 6. Not only aprominent spectral index variation is evident but it is also possibleto notice a flatter index toward the pulsar with respect to the globalvalue, as well as an important asymmetry, with the western regionflatter than the eastern one. A spectral index of about − . c (cid:13) , 1–, 1– ?? A. Ingallinera et al. D e c ( J ) Figure 6.
Spectral index map of SNR G027.4+00.0. The super-imposedcontours refer to the 5-GHz map convolved to a circular 25-arcsec beam(see text). The pulsar is indicated with a black cross.
X-ray and it is characterised by a higher column density with re-spect to the other regions (Kumar et al. 2014). Being the compactstar a magnetar, it is possible to hypothesise that its strong magneticfield has an influence on the western region, increasing the remnantintensity. Moreover, spatial inhomogeneities in the magnetic fieldintensity lead to variations of the emission turnover. Overlappingsynchrotron contributions with different turnover frequencies couldfinally result in a flatter spectral index. Alternatively, Giacani et al.(2011) discussed a similar picture for the shell-type SNR G344.7-00.1, where a flat spectral index around − . µ m emission.They conclude that the flatter spectral index derives from the SNshock impacting a dense molecular cloud, resulting in a radiativelyenergy loss. In Figure 7 we show a superposition of MIPSGAL 24- µ m image of SNR G027.4+00.0 with our radio maps. We can noticethat in the western region the IR and the 1.4-GHz emission are co-incident. A morphological comparison between 24- µ m and 5-GHzimage is more difficult because of the poor resolution of the radiomap. The 5-GHz emission seems to trace well the 24- µ m image inthe north-west part of the SNR (which is the region with the flattestspectrum) and partially in its centre. A satisfactory morphologicalanalysis would require that also 5-GHz map had about the sameresolution of the IR image. Though the correlation between IR andradio for our SNR appears less stringent than for SNR G344.7-00.1 we cannot rule out the hypothesis of shock impacting densemolecular cloud as a cause of spectral flattening. Finally anotherlimb-flattened spectral index behaviour is reported by Bhatnagar etal. (2011) for the composite-type SNR G016.7+00.1. One of the main conclusions of Paper I was that the great majorityof the MIPSGAL bubbles showing radio emission could be clas-sified as PNe. In that sense, the radio study of these objects hadproved an extremely powerful instrument for finding undiscoveredPNe (currently only about 10 percent of the expected galactic PNehave been found). However the sole IR and radio images were in-sufficient to determine many physical characteristics of these PNe and we could not go further than the mere, sometimes tentative,classification.In this section we present the PNe in our sample. We derivesome physical properties starting from the radio maps, given theirelectron temperature (see, for example, Wilson, Rohlfs & Hutte-meister 2012). For each PN, we consider the 5-GHz map and weassume that at this frequency the nebula is an optically-thin free-free emitter, as derived from their spectral index (see Table 4.3).Under this assumption the nebula brightness B is given by B = B bb ( T e ) τ ν , (1)where T e is the electron temperature, B bb ( T e ) is the brightness ofa black body at a temperature T e and τ ν is the optical depth at thefrequency ν . If T e is known, from the brightness map it is possibleto derive an optical-depth map. The optical depth is related to theemission measure EM through the relation τ ν = . × − T − . e (cid:16) ν GHz (cid:17) − . (cid:18) EMpc · cm − (cid:19) . (2)If also the distance is known it is possible to estimate the dimensionof the nebula. In this case the electron density n e can be derivedfrom the relationEM = (cid:90) s n e ds , (3)where s is the line-of-sight coordinate. Assuming an opportune ge-ometry for each PN and that the nebular gas is completely ionizedhydrogen, it is finally possible to derive the PN ionized mass as avolume integral: M ion ≈ (cid:90) V n e m p dV , (4)where m p is the proton mass.From the equation (4), and considering the previous ones, it ispossible to show that M ion ∝ T . e D / , (5)where D is the distance. Therefore, the ionized mass shows onlya weak dependence on T e , while it is very sensitive to the nebuladistance. The distance values from literature usually suffer of im-portant uncertainties. Given the nebula flux density at 5GHz and itsmean angular radius, we used the formula derived by van de Steene& Zijlstra (1995), which holds only for optically thin sources, tocalculate the PN statistical distance. We adopted this value as thebest distance estimate, except for the last PN (see Section 4.3.4).The less critical dependence on T e allows us to safely assume atypical value of 10 K whenever no previous estimates were avail-able.In the Table 4.3 we report the mean electron density, the dis-tance and the ionized mass that we derived for the four PNe. A briefdiscussion on each one of them follows.
Also known as Abell 48, PN G029.0+00.4 was first discoveredthanks to the National Geographic Society–Palomar ObservatorySky Survey and recognised as PN after its morphology (Abell 1955;Abell 1966). Optical spectroscopy of the central star by Wachter etal. (2010) seemed to dispute its nature of PN in favour of a WR starwith a surrounding nebula. However, recently, Todt et al. (2013)and, independently, Frew et al. (2014) (F14 hereafter) conductedextremely deep spectroscopic studies both on the central source c (cid:13) , 1– ?? adio emission of two SNRs and four PNe D e c ( J ) D e c ( J ) SNR G027.4+00.0 CD Figure 4.
GBT images of SNR G027.4+00.0 at 1 . . C and D are as in Angerhofer, Becker & Kundu (1977). Please note that the scale of the twoimages is different. D e c ( J ) D e c ( J ) Figure 5.
High resolution images of SNR G027.4+00.0 at 1 . . (cid:104) n e (cid:105) D T e M ion α (cm − ) (kpc) (10 K) ( M (cid:12) )PN G029.0+00.4 39.2 1.6 7 . − . − . . Table 2.
Physical parameters for the four PNe. The first column is the meanelectron density as derived from our 5-GHz map. In the second we reportthe distance value used to compute the ionized mass. The third column liststhe electron temperature as reported in literature for the first two PNe andthe assumed value of 10 K for the other two (see text). In the fourth columnwe report the ionized mass while in the last one the radio spectral index. and on the nebula. The two groups came to the same conclusionthat Abell 48 is indeed a PN harbouring an extremely rare [WN5]or [WN4-5] star. The distance of PN G029.0+00.4 has been esti-mated between 1.6 and 1 . Spitzer (both IRAC andMIPS),
WISE and
AKARI (Phillips & Ramos-Larios 2008; Phillips& M´arquez-Lugo 2011). The mid-IR infrared colour, in particular,was used by F14 as one of the indicator of the PN nature. The firstradio detection was made by Cahn & Rubin (1974) at 11 . . ∼ c (cid:13) , 1–, 1–
AKARI (Phillips & Ramos-Larios 2008; Phillips& M´arquez-Lugo 2011). The mid-IR infrared colour, in particular,was used by F14 as one of the indicator of the PN nature. The firstradio detection was made by Cahn & Rubin (1974) at 11 . . ∼ c (cid:13) , 1–, 1– ?? A. Ingallinera et al. D e c ( J ) D e c ( J ) Figure 7.
Superposition of the MIPSGAL 24- µ m image of SNR G027.4+00.0 and radio contours at 1 . µ mand at 1 . necessary. Indeed a test combination was performed and, as expect,no significant variations were revealed. The resolution achieved inFigure 8 (19 . × ) is worse than the resolution of NVSSarchive image at 1 . S C = . ± . . S L = ± α = − . ± .
08, in accordance with a optically-thin free-freeemission. From the 5-GHz flux density we derived a statistical dis-tance of 1 . ± . α flux density. They found M ion ∼ . √ ε M (cid:12) , where ε is the nebula filling factor. We used the mapat 5 GHz to provide a different approach to the same problem. As-suming an electron temperature of 7500K (F14) and a depth equalto its mean diameter (this assumption holds also for the followingPNe), we found M ion ∼ . M (cid:12) . This value is in agreement withthe one from F14 and is compatible with the typical ionized massvalues for PNe (Frew & Parker 2010). First listed in the New General Catalogue (Dreyer 1888), NGC6842 was classified as PN by Curtis (1919) because of its morphol-ogy. It appears as a ring nebula with a diameter of ∼ . D e c ( J ) Figure 8.
VLA image of PN G029.0+00.4 at 5GHz. frequency after the one published by Zijlstra, Pottasch & Bignell(1989). The high sensitivity of our map allowed us to determinean accurate flux density of 37 . ± . ± . ± . T e =
13 200K afterKaler (1983) the ionized mass of NGC 6842 results of order of0 . M (cid:12) (with the greatest uncertainties deriving from the distanceassumption). Note that Lenzuni, Natta & Panagia (1989) calculatedan ionized mass M ion = . M (cid:12) and Gorny, Stasi´nska & Tylenda(1997) for the central star M (cid:63) (cid:39) . M (cid:12) . c (cid:13) , 1– ?? adio emission of two SNRs and four PNe D e c ( J ) Figure 9.
VLA image of NGC 6842 at 5GHz.
PN G040.3-00.4 was first detected and recognised as PN by Abell(1955; 1966). In the optical it appears as a ring nebula with a di-ameter of about 50arcsec. Several studies indicates that this PN islocated at a distance d ∼ . Spitzer (Phillips & Ramos-Larios 2008; Kwok et al. 2008) and
AKARI (Phillips & M´arquez-Lugo 2011). It was first observed at radiowavelength by Milne & Aller (1982) at 14 . . . S C = . ± . . S L = . ± . α = − . ± .
01 (see Paper I). Itis important to notice that the NVSS catalogue lists a flux densityof 33 . ± . . . ± . T e = K (to ourknowledge, there is no real estimates of T e in literature), we cancalculate an ionized mass M ion ∼ . M (cid:12) . Very little is known about PN G031.9-00.3. Discovered and clas-sified by Weinberger & Sabbadin (1981), a tentative distance of4 . . S C = . ± . . S L = . ± . α = . ± .
28, showing that the source is stilloptically thick at 1 . . ± . T e = K, we find M ion ∼ . M (cid:12) if D = . M ion ∼ . M (cid:12) if D = . Radio interferometer observations are an important instrument inthe characterisation of Galactic circumstellar envelopes. Not onlythey permit us to reliably trace the ionized medium surroundingan evolved star, but also to derive different physical parameters ofthese objects.However, the synthesis imaging process from interferometerdata is optimised for point source observations. When extendedsources are to be imaged we found that different algorithms, likethe multi-scale CLEAN, give more satisfactory results in terms ofimage quality. For sources extended more than the interferometerLAS we also complemented the interferometer data with single-dish maps in order to get to a qualitatively and quantitatively reli-able representation of the sky.The combination process, via feathering, resulted fundamentalfor the two SNRs, for which the mere VLA maps were extremelypoor and the recovered flux density was a factor 3 or 4 less than thereal one. The information availability in a very wide range of spatialfrequencies allowed us to create accurate and detailed images ofthese objects. Furthermore, the two SNRs served as a test to verifythat this simple procedure can be used to perform also quantitativeanalyses on very extended sources.For this reason we used SNR G011.2-00.3 mainly to testthe method, and in particular the total flux density recovery. Thissource, in fact, has been deeply studied so far and a good baseknowledge is available in literature. What we found is a perfectagreement between our flux density determination and the litera-ture estimates from single-dish data. The resulting map at 5GHz iscurrently the best representation of the source at this frequency.For SNR G027.4+00.0 we were able to obtain important re-sults. Using the 1.4-GHz map we posed a more stringent upperlimit for the flux density of its yet non-detected pulsar. We did alsoshow that the nebula presents a spatial variation of its radio spectralindex. One hypothesis is that the flatter spectral index characteris-ing the western region may be due to inhomogeneities of the strongmagnetic field from the central object, causing a superposition ofdifferent ‘regular’ synchrotron contributions with different turnoverfrequencies. This, in turn, results in a global flatter spectrum. Al-ternatively it is also possible that the spectrum flattening could beascribed to the impact of the SN shock on dense molecular clouds.The possibility to accurately measure the flux densities alsoallowed us to constraint some physical parameters for the four PNe.In particular we were able to determine their ionized masses andtheir statistical distance. We found that all the four PNe have anionized mass of order of 0 . M (cid:12) . For two of them it is possible tofind in literature ionized mass estimates, via H α or H β fluxes, thatare in agreement with our results from radio. For the other two, asfar as we know, our values are the first estimates. c (cid:13) , 1–, 1–
28, showing that the source is stilloptically thick at 1 . . ± . T e = K, we find M ion ∼ . M (cid:12) if D = . M ion ∼ . M (cid:12) if D = . Radio interferometer observations are an important instrument inthe characterisation of Galactic circumstellar envelopes. Not onlythey permit us to reliably trace the ionized medium surroundingan evolved star, but also to derive different physical parameters ofthese objects.However, the synthesis imaging process from interferometerdata is optimised for point source observations. When extendedsources are to be imaged we found that different algorithms, likethe multi-scale CLEAN, give more satisfactory results in terms ofimage quality. For sources extended more than the interferometerLAS we also complemented the interferometer data with single-dish maps in order to get to a qualitatively and quantitatively reli-able representation of the sky.The combination process, via feathering, resulted fundamentalfor the two SNRs, for which the mere VLA maps were extremelypoor and the recovered flux density was a factor 3 or 4 less than thereal one. The information availability in a very wide range of spatialfrequencies allowed us to create accurate and detailed images ofthese objects. Furthermore, the two SNRs served as a test to verifythat this simple procedure can be used to perform also quantitativeanalyses on very extended sources.For this reason we used SNR G011.2-00.3 mainly to testthe method, and in particular the total flux density recovery. Thissource, in fact, has been deeply studied so far and a good baseknowledge is available in literature. What we found is a perfectagreement between our flux density determination and the litera-ture estimates from single-dish data. The resulting map at 5GHz iscurrently the best representation of the source at this frequency.For SNR G027.4+00.0 we were able to obtain important re-sults. Using the 1.4-GHz map we posed a more stringent upperlimit for the flux density of its yet non-detected pulsar. We did alsoshow that the nebula presents a spatial variation of its radio spectralindex. One hypothesis is that the flatter spectral index characteris-ing the western region may be due to inhomogeneities of the strongmagnetic field from the central object, causing a superposition ofdifferent ‘regular’ synchrotron contributions with different turnoverfrequencies. This, in turn, results in a global flatter spectrum. Al-ternatively it is also possible that the spectrum flattening could beascribed to the impact of the SN shock on dense molecular clouds.The possibility to accurately measure the flux densities alsoallowed us to constraint some physical parameters for the four PNe.In particular we were able to determine their ionized masses andtheir statistical distance. We found that all the four PNe have anionized mass of order of 0 . M (cid:12) . For two of them it is possible tofind in literature ionized mass estimates, via H α or H β fluxes, thatare in agreement with our results from radio. For the other two, asfar as we know, our values are the first estimates. c (cid:13) , 1–, 1– ?? A. Ingallinera et al. D e c ( J ) D e c ( J ) Figure 10.
VLA image of PN G040-00.4 at 1 . D e c ( J ) D e c ( J ) Figure 11.
VLA image of PN G031.9-00.3 at 1 . We remark that this work has shown the importance of reli-able radio data in multiwavelength studies. In fact radio maps canbe successfully used to derive important physical parameters fordifferent types of Galactic sources, regardless of their angular ex-tensions. Eventually we want to stress that the procedure adoptedfor interferometer and single-dish data combination does not de-pend on the particular instruments used (VLA and GBT), since itworks directly on FITS images. Therefore, given the positive re-sults showed in this work, it can be applied as it is to differentother interferometers such as ALMA, where flux-loss issues areeven more important, and, in future, SKA.
ACKNOWLEDGEMENTS
This work is based on observations made with the Very Large Ar-ray and the Green Bank Telescope of the National Radio Astron-omy Observatory, a facility of the National Science Foundationoperated under cooperative agreement by Associated Universities Inc.. Archive search made use of the SIMBAD database and theVizieR catalogue access tool, operated by the Centre de Donn´eesastronomique de Strasbourg.
REFERENCES
Abell G. O., 1955, PASP, 67, 25Abell G. O., 1966, ApJ, 144, 259Altenhoff W. J., Downes D., Goad L., Maxwell A., Rinehart R.,1970, A&AS, 1, 319Angerhofer P. E., Becker R. H., Kundu M. R., 1977, A&A, 55, 11Bhatnagar S., Rau U., Green D. A., Rupen M. P., 2011, ApJ, 739,L20Becker R. H., Kundu M. R., 1975, AJ, 80, 679Cahn J. H., Kaler J. B., 1971, ApJS, 22, 319Cahn J. H., Rubin R. H., 1974, AJ, 79, 128Cahn J. H., Kaler J. B., Stanghellini L., 1992, A&AS, 94, 399Carey S. J., et al., 2009, PASP, 121, 76 c (cid:13) , 1– ?? adio emission of two SNRs and four PNe Caswell J. L., Haynes R. F., Milne D. K., Wellington K. J., 1982,MNRAS, 200, 1143Clark B. G., 1980, A&A, 89, 377Clark B. G., 1999, in Taylor G. B., Carilli C. L., Perley R. A., eds,ASP Conf. Ser. Vol. 180, Synthesis imaging in radio astronomyII. Astron. Soc. Pac., San Francisco, p. 1 Scappini F., Caselli P.,Attolini M. R., 1995, MNRAS, 276, 57Condon J. J., Kaplan D. L., 1998, ApJS, 117, 361Condon J. J., Cotton W. D., Greisen E. W., Yin Q. F., Perley R. A.,Taylor G. B., Broderick J. J., 1998, AJ, 115, 1693Cornwell T. J., 2008, IEEE Journal of Selected Topics in SignalProcessing, Vol. 2, Issue 5, p.793-801Curtis H. D., 1919, PASP, 31, 285Dickel J. R., Milne D. K., 1972, AuJPh, 25, 539Dobrinˇci´c M., Villaver E., Guerrero M. A., Manchado A., 2008,AJ, 135, 2199Downes A., 1984, MNRAS, 210, 845Dreyer J. L. E., 1888, MmRAS, 49, 1Emerson D. T., Gr¨ave R., 1988, A&A, 190, 353Frew D. J., Parker Q. A., 2010, Publ. Astron. Soc. Aust., 27, 129Frew D. J., et al., 2014, MNRAS, 440, 1345Gaensler B. M., Slane P. O., 2006, ARA&A, 44, 17Giacani E., Smith M. J. S., Dubner G., Loiseau N., 2011, A&A,531, A138Giammanco C., et al., 2011, A&A, 525, A58Glushak A. P., 2014, ARep, 58, 6Gorny S. K., Stasi´nska G., Tylenda R., 1997, A&A, 318, 256Gotthelf E. V., Vasisht G., 1997, ApJ, 486, L133Gotthelf E. V., Vasisht G., Dotani T., 1999, ApJ, 522, L49Gregory P. C., Taylor A. R., 1986, AJ, 92, 371Gregory P. C., Scott W. K., Douglas K., Condon J. J., 1996, ApJS,103, 427Helfand D. J., Becker R. H., White R. L., Fallon A., Tuttle S.,2006, AJ, 131, 252H¨ogbom J.A., 1974, A&AS, 15, 417Hromov G. S., Kohomek L., 1968, BAICz, 19, 1Ingallinera A., et al., 2014, MNRAS, 437, 3626Kaler J. B., 1983, ApJ, 271, 188Kassim N. E., 1988, ApJS, 68, 715Kassim N. E., 1989, ApJ, 347, 915Kriss G. A., Becker R. H., Helfand D. J., Canizares C. R., 1985,ApJ, 288, 703Koo B.-C., Heiles C., 1991, ApJ, 382, 204Koo B.-C., Moon D.-S., Lee H.-G., Lee J.-J., Matthews K., 2007,ApJ, 657, 308Kothes R., Reich W., 2001, A&A, 372, 627Kumar H. S., Safi-Harb S., 2010, ApJ, 725, L191Kumar H. S., Safi-Harb S., Slane P. O., Gotthelf E. V., 2014, ApJ,781, 41Kwok S., Zhang Y., Koning N., Huang H.-H., Churchwell E.,2008, ApJS, 174, 426Large M. I., Mathewson D. S., Haslam C. G. T., 1961, MNRAS,123, 123Lenzuni P., Natta A., Panagia N., 1989, ApJ, 345, 306Lin L., et al., 2011, ApJ, 740, L16Maciel W. J., 1984, A&AS, 55, 253McMullin J. P., Waters B., Schiebel D., Young W., Golap K.,2007, ASPC, 376, 127Milne D. K., 1969, AuJPh, 22, 613Milne D. K., 1979, AuJPh, 32, 83Milne D. K., Aller L. H., 1982, A&AS, 50, 209Mizuno D. R., et al., 2010, AJ, 139, 1542 Morsi H. W., Reich W., 1987, A&AS, 71, 189O’dell C. R., 1962, ApJ, 135, 371Ortiz R., 2013, A&A, 560, A85Pazderska B. M., et al., 2009, A&A, 498, 463Phillips J. P., 2004, MNRAS, 353, 589Phillips J. P., Ramos-Larios G., 2008, MNRAS, 383, 1029Phillips J. P., M´arquez-Lugo R. A., 2011, RMxAA, 47, 83Reynolds S. P., Lyutikov M., Blandford R. D., Seward F. D., 1994,MNRAS, 271, L1Rieke G. H., et al., 2004, ApJS, 154, 25Sabbadin F., 1986, A&A, 160, 31Sanbonmatsu K. Y., Helfand D. J., 1992, AJ, 104, 2189Shaver P. A., Goss W. M., 1970, AuJPA, 14, 77Stanghellini L., Corradi R. L. M., Schwarz H. E., 1993, A&A,279, 521Stil J. M., et al., 2006, AJ, 132, 1158Tajitsu A., Tamura S., 1998, AJ, 115, 1989Taylor A. R., Gregory P. C., 1983, AJ, 88, 1784Taylor A. R., Goss W. M., Coleman P. H., van Leeuwen J., Wal-lace B. J., 1996, ApJS, 107, 239Tian W. W., Leahy D. A., 2008, ApJ, 677, 292Todt H., et al., 2013, MNRAS, 430, 2302Torii K., Tsunemi H., Dotani T., Mitsuda K., 1997, ApJ, 489,L145van de Steene G. C., Zijlstra A. A., 1995, A&A, 293, 541Vasisht G., Aoki T., Dotani T., Kulkarni S. R., Nagase F., 1996,ApJ, 456, L59Vasisht G., Gotthelf E. V., 1997, ApJ, 486, L129Wachter S., Mauerhan J. C., Van Dyk S. D., Hoard D. W., KafkaS., Morris P. W., 2010, AJ, 139, 2330Weinberger R., Sabbadin F., 1981, A&A, 100, 66White R. L., Becker R. H., Helfand D. J., 2005, AJ, 130, 586Wilson T. L., Rohlfs K., Huttemeister S., 2012, Tools of RadioAstronomy, fifth edition, SpringerZhang C. Y., 1995, ApJS, 98, 659Zijlstra A. A., Pottasch S. R., Bignell C., 1989, A&AS, 79, 329 c (cid:13) , 1–, 1–