High-resolution radio study of SNR IC443 at low radio frequencies
aa r X i v : . [ a s t r o - ph . GA ] A p r Astronomy&Astrophysicsmanuscript no. castelletti-astroph c (cid:13)
ESO 2018August 22, 2018
High-resolution radio study of SNR IC 443 at lowradio frequencies
G. Castelletti , ⋆ , G. Dubner ⋆ , T. Clarke , and N. E. Kassim Instituto de Astronom´ıa y F´ısica del Espacio (IAFE, CONICET-UBA), CC67, Suc.28, 1428,Buenos Aires, Argentina.e-mail: [email protected] Facultad de Ciencias Exactas y Naturales (Universidad de Buenos Aires). Remote Sensing Division, Code 7213, Naval Research Laboratory, 4555 Overlook Avenue,SW, Washington DC, USA.Received ¡date¿; Accepted ¡date¿
ABSTRACT
Aims.
To investigate in detail the morphology at low radio frequencies of the supernova remnant(SNR) IC 443 and to accurately establish its radio continuum spectral properties.
Methods.
We used the VLA in multiple configurations to produce high resolution radio imagesof IC 443 at 74 and 330 MHz. From these data we produced the first sensitive, spatially resolved,spectral analysis of the radio emission at long wavelengths. The changes with position in the ra-dio spectral index were correlated with data in near infrared (NIR) from 2MASS, in gamma-raysfrom
VERITAS , and with the molecular CO ( J = Results.
The new image at 74 MHz has
HPBW = ′′ , rms =
30 mJy beam − and at 330 MHz HPBW = ′′ and rms = − . The integrated flux densities for the whole SNR are S SNR74MHz = ±
51 Jy and S SNR330MHz = ±
15 Jy. Improved estimates of the integrated spectrumwere derived taking into account a turnover to fit the lowest frequency measurements in the lit-erature. Combining our measurements with existing data, we derive an integrated spectral index α = − . ± .
01 with a free-free continuum optical depth at 330 MHz τ ∼ × − ( τ = . ∼
10 MHz are equally consistent with a power law spec-trum. For the pulsar wind nebula associated with the compact source CXOU J061705.3 + S PWN330MHz = . ± .
05 Jy, S PWN1420MHz = . ± .
04 Jy, and α ∼ .
0. Substantialvariations are observed in spectral index between 74 and 330 MHz across IC 443. The flattestspectral components ( − . ≤ α ≤ − .
05) coincide with the brightest parts of the SNR alongthe eastern border, with an impressive agreement with ionic lines as observed in the 2MASS J and H bands. The di ff use interior of IC 443 has a spectrum steeper than found anywhere in theSNR ( − . ≤ α ≤ − . α ∼ − . VERITAS γ -ray emission strikingly matches the CO distribution, but noclear evidence is found for a morphological correlation between the TeV distribution and radioemission. Conclusions.
The excellent correspondence between the eastern radio flattest spectrum region ⋆ Member of the Carrera del Investigador Cient´ıfico of CONICET, Argentina 1 astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequenciesand NIR ionic lines strongly suggests that the passage of a fast dissociating J-type shock acrossthe interacting molecular cloud dissociated the molecules and ionized the gas. We therefore con-clude that thermal absorption at 74 MHz ( τ up to ∼ .
3) is responsible for the localized spectralindex flattening observed along the eastern border of IC 443. Towards the interior of IC 443the spectrum is consistent with those expected from linear di ff usive shock acceleration, whilethe flatter spectrum in the southern ridge is a consequence of the strong shock / molecular cloudinteraction. Key words.
ISM:individual objects: IC 443, PWN CXOU J061705.3 +
1. Introduction
The importance of radio continuum studies of SNRs for understanding shock acceleration pro-cesses (Reynolds & Ellison 1992; Anderson & Rudnick 1996) as well as intrinsic and extrinsicinteractions with ionized gas and the interstellar medium (ISM) (Dulk & Slee 1975; Kassim 1989)has been long appreciated. However the magnitude of the observable e ff ects are often subtle, re-quiring a large leverage arm in frequency space to discern. The lack, until recently, of su ffi cientangular resolution and sensitivity at the lowest frequencies has made progress in such studies slow.The advent of high resolution, low frequency observations with instruments like the VLA andGMRT are now changing this picture. In particular, VLA sub-arcminute resolution imaging be-low 100 MHz has been important for discerning spatially resolved intrinsic and extrinsic thermalabsorption. Examples now include unshocked ejecta inside Cas A (Kassim et al. 1995), thermalfilaments within the Crab nebula (Bietenholz et al. 1997), and ionized gas in the ISM along the lineof sight towards W49B (Lacey et al. 2001). Later on, Brogan et al. (2005) spatially resolved theionized boundary marking the SNR / molecular cloud (MC) interface in 3C 391, suggesting suchsignatures could be both common and important for delineating elusive SNR / MC interactions.Most recently, Castelletti et al. (2007) detected strong 74 MHz free-free absorption at the interfacebetween the SNR W44 and the photo dissociation region of a neighbouring HII region.In this paper we extend such high resolution, low frequency radio studies to the classicSNR IC 443, one of the clearest a priori known cases of a remnant interacting with its cloudysurroundings. We present the first, high resolution radio studies of this object, and interpret ourresults in context of recent IR and high energy observations. IC 443 is a particularly attractive SNRfor low frequency studies both because of its large angular size and because its outer Galaxy lo-cation leave it relatively well isolated from the confusing e ff ects normally contaminating studiestowards inner Galactic SNR / MC complexes.
Observed in the radio domain IC 443 (G189.1 + ff erent locations (the names “shell A” and “shell B” are some-times used to refer their locations towards the east and west halves of the remnant, Braun & Strom1986b). On its eastern side, IC 443 has a rim-brightened morphology, while towards the western Send o ff print requests to : G. Castelletti2 astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequencies half the surface brightness is dimmer. A third incomplete and faint shell (called by Braun & Strom1986b, “shell C”) is also evident extending beyond the northeast periphery of the remnant whenthe radio images are displayed with very high contrast. Based on both its morphology and soft X-ray spectrum, “shell C” was proposed to be a di ff erent SNR called G189.6 + + ff erent X-ray observations carried out with the Einstein Observatory (Petre et al. 1988),
Ginga (Wang et al. 1992),
ROSAT (Asaoka & Aschenbach 1994),
ASCA (Kawasaki et al. 2002),and
XMM-Newton (Troja et al. 2006) show a bulk of thermal centrally peaked X-ray emissionwithin the radio rim. On the basis of the radio / X-ray morphology and the X-ray properties, severalauthors proposed that IC 443 belongs to the class of “mixed-morphology” or “thermal-composite”SNRs (Rho & Petre 1998).The compact X-ray source CXOU J061705.3 + × yr (Olbert et al. 2001). An alternative estimate of ∼ + observations, forms a ring in the foreground andappears to be interacting with the remnant at several locations from north to south (Dickman et al.1992). This phenomenon, which seems to be responsible for the multiwavelength observationalpicture of the remnant, was unambiguously confirmed by the presence of several OH (1720 MHz)masers as well as IR cooling lines from H (Rho et al. 2001; Hewitt et al. 2006, and referencestherein). Additionally, HI observations show well-defined filamentary structures in the northeasternregion of the remnant, where the brightest optical filaments were located (Lee et al. 2008). Thesefeatures are interpreted as the recombined neutral gas in an atomic shock (Lee et al. 2008).Another interesting characteristic of IC 443 is the presence of associated high energy sources.The EGRET GeV source (3EG J0617 + VERITAS , and
FERMI telescopes (Albert & et 2007; Acciari et al. 2009; Abdo et al. 2010). It has been suggested astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequencies Table 1.
Observational summary
Program Observing VLA Bandwidth Integration Synthesized Largest detectabledates Config. (MHz) time (hr) beam (arcsec) structure (arcsec)74 MHz ParametersAG0697 2006 Apr 8, 9, 10 A 6.3 9 25 ×
20 800AG0735 2006 Oct 27 C 6.3 8.35 228 ×
214 7500AG0697 2007 Oct 21 AB 6.3 6.85 28 ×
30 1300330 MHz ParametersAG0697 2005 Sep 16 C 6.3 4.7 62 ×
60 1800AG0697 2006 Apr 8, 9, 10 A 1.5 9 6 × ×
53 1800AG0735 2007 Apr 15, 16 D 6.3 5.2 238 ×
205 4200AG0697 2007 Oct 21 AB 1.5 6.85 22 ×
17 500 that the correlation of gamma rays with molecular gas arises from the pion decay of hadronic cos-mic rays generated by the interaction of the SNR shock with dense molecular material (Humensky2008; Abdo et al. 2010).
2. Observations and data reduction
The Very Large Array (VLA) was used in multiple configurations to observe the large SNR IC 443at 74 and 330 MHz. A summary of the observations carried out on various dates between 2005 and2007 is given in Table 1.In order to help with the radio frequency interference (RFI) excision and mitigate the ef-fects of bandwidth smearing all the data were acquired in multi-channel continuum mode (64and 16 channels per polarization at 74 and 330 MHz, respectively). At both, 74 and 330 MHz,we first performed a bandpass calibration on either 3C 405 (Cygnus A), 3C 147 (0542 + . The initial antenna-based gain and phase cor-rections at 74 MHz were estimated from observations of either 3C 405 or 3C 274. For the 330 MHzdata, observations of 3C 147 were su ffi cient for both, flux density and initial phase calibration atthis frequency. In all cases the absolute flux scale was set according to the Baars et al. (1977) scale.The NRAO Astronomical Image Processing Software (AIPS) package was used to process all theobservations.Large fields of view are a general characteristic when observing at long radio wavelengthsand it is necessary to avoid distortions introduced in the image caused by the non-coplanarity ofthe baselines of the array. To overcome this, we employed a pseudo-three-dimensional Fourierinversion as implemented in the AIPS task IMAGR, in which the primary beam area of the 74 and330 MHz data, 11 ◦ . ∼ ◦ .
5, respectively, is divided into multiple partially overlapping facets(Cornwell & Perley 1992). Furthermore, to compensate for time variable ionospheric phase e ff ects The Very Large Array of the National Radio Astronomy Observatory is a facility of the National ScienceFoundation operated under cooperative agreement by Associated Universities, Inc. See http: // lwa.nrl.navy.mil / tutorial / to obtain source models in FITS format.4 astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequencies we performed successive rounds of self-calibration and imaging to the data from each configurationseparately at 74 and 330 MHz. Angle-invariant self-calibration (as implemented in the AIPS taskCALIB) is generally inadequate for compensating for ionospheric e ff ects across a large field view,particularly for the 74 MHz B and A configuration (Kassim et al. 2007). An exception is for fieldsdominated by a bright source at the field center, as in the case of IC 443. The final calibratedvisibility data from all of the observations listed in Table 1 for each frequency were then combinedinto a single uv data set, after which a further amplitude and phase self-calibration was performed.For the concatenated 330 MHz data we employed a multi-scale CLEAN algorithm in AIPS, withthree di ff erent scales sizes. This process is e ffi cient to make high resolution images that are alsosensitive to extended structures. All the resulting facet images were stitched together into one largefield using AIPS task FLATN to create a single final image with a synthesized beam of 17 ′′ . × ′′ .
81, PA = − ◦ , which represents an order of magnitude improvement in angular resolution overthe 330 MHz image of Braun & Strom (1986b) and that of Hewitt et al. (2006). The sensitivityachieved in our image after correcting for primary beam attenuation is 1.7 mJy beam − .Concerning the observations at 74 MHz, we noted that the extended radio emission in thefield was more properly imaged using the capability incorporated in the task IMAGR to switchback and forth between the SDI Clean (Steer et al. 1984) and the usual CLEAN deconvolutionstrategy, depending on the contrast between the brightest residual pixel and the bulk of residualpixels after each major cycle. Following this procedure, the final resolution of the first image ofIC 443 obtained at 74 MHz after combining the data sets described in Table 1 and including primarybeam corrections is 36 ′′ . × ′′ .
76, PA = ◦ .
71. The rms noise level in the 74 MHz image is30 mJy beam − .Ionospheric refraction and self-calibration normally introduce arbitrary position shifts on lowfrequency images. These shifts are readily corrected by registering against background small di-ameter sources whose positions are known from higher frequency observations. We corrected theseshifts by measuring several small-diameter sources with respect to their known positions from theNRAO VLA Sky Survey (NVSS) (the latter has an astrometric accuracy better than 1 ′′ in bothR.A. and dec., Condon et al. 1998). We determined and corrected for a mean positional di ff er-ence of 0 s . ± s .
06 in R.A. and 5 ′′ . ± ′′ .
00 in dec. at 74 MHz and 0 s . ± s .
04 in R.A. and3 ′′ . ± ′′ .
42 in dec. at 330 MHz.
3. Results
In Fig. 1 we present a close-up view of the radio continuum emission at 74 MHz from IC 443. Theseobservations provide the first subarcminute image of this remnant created at meter wavelengths.The data at 74 MHz are sensitive to smooth structures up to ∼ ′ in size, larger than the fullextent of the SNR, and thus are sensitive to the largest scale structure present in the SNR.Figure 2 displays with a resolution of ∼ ′′ the new VLA image of the ∼ . . An enlargement showing the detailed330 MHz total intensity morphology of IC 443 is shown in Fig. 3. The combination of data from the The FITS files of the radio images of IC 443 at 74 and 330 MHz are available online athttp: // / snr / SNR.html. 5 astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequencies di ff erent VLA array configurations ensures that all scales of the radio emission are well representedin the 330 MHz image. ( J ) D e c li na t i on IC443
Fig. 1.
Radio continuum image of IC 443 at 74 MHz. This map has been corrected for the atten-uation of the primary beam. The angular resolution of this image was smoothed to a beamsize of50 ′′ . The grayscale is linear ranging from 25 to 250 mJy beam − and the contour levels are 72,120, 160, 200, 240, and 280 mJy beam − . The plus symbol marks the position of the compactsource CXOU J061705.3 + ∼ ′ , or ∼
15 pc at the assumed distance of 1.5 kpc, is measured in our map for thebright radio shell in IC 443 (that, as mentioned in Sect. 1.1 it is referred to in literature as shellA, and is depicted in the inset in Fig. 2). The low surface brightness emission gradually decreasesoutwards to the west forming a more uniform radio shell of about ∼ ′ in diameter or ∼
23 pc (theso called shell B, Fig. 2), as measured on the new 330 MHz view of the remnant.Additionally, the 330 MHz image of IC 443 serves to show that even at low radio frequenciespart of the east outer border of the remnant is quite structured. It is particularly remarkable theindented morphology of the bright rim at R.A. = h m s , dec. = + ◦ ′ , with tenuous radiosynchrotron emission detected ahead of the main shock. Such faint emission is seen as a weak astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequencies Fig. 2.
The image shows the entire field of view of the VLA around the SNR IC 443 at 330 MHz.The image displayed includes primary beam correction. The synthesized beam is 17 ′′ × ′′ witha position angle of − ◦ , and the sensitivity level is 1.7 mJy beam − . This is the first high fidelityand high resolution view of the emission from IC 443 ever obtained at low radio frequencies. Theinset with the image of IC 443 at 330 MHz is included to help in the location of the bright andweak radio shells referred to in the literature as shells A and B, respectively.and irregular radio halo that in our 330 MHz image is notable at a mean level of ∼ σ extendingup to 5 ′ ahead of the bright sharp boundary. It is interesting to note that similar weak, di ff useemission upstream of the main shock is also observed in the cases of Puppis A and W44, alongthe sides where the expanding blast wave encountered molecular gas (Castelletti et al. 2006, andCastelletti et al. 2007). A series of small protrusions are also evident along most of the extension ofthe eastern rim. The largest of these features located at R.A. = h m s , dec. = + ◦ ′ emergesabout 2 ′ radially from the border of the remnant. These structures can be also seen in the imageof IC 443 at 1420 MHz presented by Lee et al. (2008), who refer to them as “spurs”. The authorsinterpret that both, the halo and the “spurs”, are likely originated from the interaction of the SNRwith the surrounding inhomogeneous ambient medium, although an alternative explanation basedon a physical association between IC 443 and the SNR G189.6 + + astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequencies Fig. 3.
A color image of the radio continuum emission from IC 443 at 330 MHz constructed us-ing multiple-configuration VLA observations. The brightness range covered by the scale is be-tween 3 and 35 mJy beam − . The final beam size shown at bottom left is 17 ′′ × ′′ at a posi-tion angle of − ◦ . The noise level is 1.7 mJy beam − after primary beam correction. The colorscale runs between 3 and 45 mJy beam − . The white plus sign marks the location of the sourceCXOU J061705.3 + near R.A. = h m s , dec. = ◦ ′ .Towards the center of the SNR, in the southern half, the most remarkable feature is a brightannular filament placed near R.A. = h m s , dec. = + ◦ ′ , which is immersed in the faintand di ff use emission that dominates the inner part of IC 443. Such annular structure forms thecentral part of a more extended feature commonly referred in the literature as the southern sinuousridge. The large southern ridge is defined below dec. ∼ + ◦ ′ at 330 MHz and appears to bean extension of the southeast border of the remnant. Shocked CO gas with broad lines has beendetected near this region of the SNR (Cornett et al. 1977; Dickman et al. 1992).Integrated flux densities estimates for IC 443 were made using the new observations, yieldingS
74 MHz = ±
51 Jy and S
330 MHz = ±
15 Jy. The quoted values have been corrected for theprimary beam response and for the contribution of unrelated point sources overlapping the remnant. The radio continuum emission from G189.6 + astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequencies The uncertainties in the measurements account for the statistical errors as well as the selection ofintegration boundaries.
The new sensitive image at 330 MHz o ff ers, for the first time, a view of the low frequency coun-terpart of the low-luminosity plerionic nebula powered by the source CXOU J061705.3 + Chandra and
XMM-Newton (Olbert et al. 2001; Bocchino & Bykov 2001; Gaensler et al. 2006).A close up image of the pulsar wind nebula (PWN) at 330 MHz is displayed in Fig. 4. At74 MHz the combined e ff ect of the lower sensitivity, poorer angular resolution, and flatter in-trinsic spectrum, conspire against its visibility at this low frequency. As in X-rays, at low radiofrequencies the source CXOU J061705.3 + − , lies about 0 ′ .25 northeastward from CXOU J061705.3 + ′ .
25 behind CXOU J061705.3 + ′ .21 ahead of it. The integrated flux density over the entire nebula obtained from our imageat 330 MHz is S PWN330 MHz = . ± .
05 Jy. In addition we estimated the flux density of the PWNat 1420 MHz on the basis of the image of IC 443 presented by Lee et al. (2008) performed bycombining VLA observations and single dish data taken from the Arecibo Telescope, obtaining S PWN1420 MHz = . ± .
04 Jy.
Fig. 4.
A close-up image of the nebular emission around CXOU J061705.3 + + − . The contour levels on the image are traced at 19, 21, 23, and 25 mJy beam − . Figure 5 illustrates, with a spatial resolution better than 20 ′′ , the morphological comparison be-tween the optical and radio emission from IC 443 obtained from the combination in a false colorimage of our new VLA observations at 330 MHz (in red) with data from the Second Palomar astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequencies Observatory Sky Survey (in green). In this figure, features where both spectral bands overlap areshown in yellow.The optical emission tracing the low density atomic gas reproduces the east-west asymmetrycharacteristic of the radio total intensity emission as well. The coincidence of the synchrotron-enhanced radio emission with very strong optical filaments observed in H α and [SII] lines towardsthe east, suggests that the optical features delineate the position of cooling post-shock ISM gas.Furthermore, near infrared emitting gas observed in this portion of the SNR (see below Fig. 9)together with abundant neutral gas, correlated both in space and velocity with the optical filaments,indicate the presence of radiative shocks propagating into gas of di ff erent densities.The quality of the new 330 MHz image allows us to identify the close radio / optical correspon-dence in most of the small-scale radio structures observed as extensions from the bright easternportion of the shell. This is in good agreement with the behavior previously noticed by Lee et al.(2008) using radio continuum observations at 1420 MHz. Fig. 5.
A high resolution comparison between radio and optical emission from IC 443 SNR. Thegreen corresponds to optical emission from the Second Palomar Observatory Sky Survey, while inred the 330 MHz radio emission is shown. The yellow regions are areas where emission in bothspectral bands overlap.The southern radio ridge is also mimicked by a quite faint optical counterpart. Absorption dueto the molecular gas mainly located in the foreground of the SNR is probably responsible for theobserved weakness in the optical emission in this part of IC 443. astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequencies In the breakout region toward the western side of IC 443 only few local radio enhancements,immersed in faint di ff use radio emission, are observed at the locations of the optical filaments.
4. Radio spectral properties of SNR IC 443
In this section we analyze the spectral properties of this SNR using the new VLA images presentedhere at 74 and 330 MHz. Estimates of the spectral index variations with position across the remnantrequire imaging the interferometric data of IC 443 at both frequencies using the same uv coverage.To perform this, we reconstructed the interferometric images by applying appropriate taperingfunctions to the visibility data at 74 and 330 MHz. To show more clearly the main spectral features,avoiding any masking e ff ect from small scale variations, we have chosen a final synthesized beamof 70 ′′ . In addition, to avoid any positional o ff sets, the images were aligned and interpolated toidentical projections before calculating spectral indices.The dependency of the spectral index between 74 and 330 MHz with the position within IC 443was determined through the construction of a spatially resolved spectral index map, which is shownin Fig. 6. To produce the spectral map the matched images of IC 443 at both frequencies weremasked at the 3 σ level of their respective noise levels. The error in the determination of the spectralindex from the map is less than 0 .
04 in the high flux regions (the east limb of the SNR and someinterior filaments) and about 0.1 in the di ff use central emission. The uncertainties increase in theweakest emission regions towards the westernmost part of the remnant due to the lower sensitivityof the image at 74 MHz in this region (see below).Figure 6 shows good agreement between total intensity features and the spatial distribution ofthe spectral index. The 74 /
330 MHz spectral map reveals for the first time the morphology of avery flat spectral component running along the eastern side of IC 443 within which the spectralindex varies between α ∼ − .
05 and ∼ − .
25. This flattening towards the brightest filaments(see also Fig. 1 and Fig. 3) indicates that some thermal absorption is present, and would becomestronger at lower frequencies. It is remarkable that such spectral behavior has also counterparts inthe J and H bands as observed by 2MASS (further details for the correlation between local spectralindex variation across IC 443 and the IR emission are described below in Sect. 5).The large-scale di ff use emission in the SNR’s interior has a spectrum that is markedly di ff erentfrom the eastern part of the remnant with steeper components ranging from α = − . α = − .
85, as would be expected under the linear di ff usive shock accelerationmodel for weak shocks with low Mach numbers (Anderson & Rudnick 1993). The southern sin-uous ridge, where most of the interaction with the molecular cloud is taking place, is seen in ourspectral index map as a region with α varying between ∼ − .
25 and ∼ − .
5. These local spectralindices are in good correlation with total intensity features: the brighter synchrotron areas are sys-tematically flatter than the other parts of the ridge. In this case the interpretation of the flattening inthe spectrum is di ff erent from the case noted above in the east rim. Here, the spectral behavior mightbe a signature of Fermi shock acceleration at the sites where stronger post compression shock den-sities, accompanied by higher local Mach numbers, and / or higher magnetic field strength result dueto the impact of the SNR blast wave on denser ambient medium (Bell 1978; Anderson & Rudnick1993). astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequencies - . - . - . - . - . - . - . - . . S pe c t r a l I nde x Fig. 6.
Spectral index map constructed using VLA observations at 74 and 330 MHz (matched toa common angular resolution of 70 ′′ ). The 5.5, 15, 26, and 40 mJy beam − contours from the330 MHz image are included to facilitate the comparison between spectral continuum and totalpower features. To create this map the 74 and 330 MHz images were masked at 3 σ .Towards the western side of IC 443, which is dominated by di ff use and faint radio emission, thespectral indices are mostly steep ( α ≃ − . = h m s , dec. =+ ◦ ′ ′′ and 06 h m s , + ◦ ′ ′′ confirms their extragalactic nature (Braun & Strom 1986a). A sim-ilar result is obtained for the point radio source located at the position R.A. ∼ h m . s ,dec. ∼ + ◦ ′ ′′ . To update the mean spectral index determination for IC 443, we have included the new integratedflux densities at 74 and 330 MHz in the extensive list of measurements presented in the literature.In Table 2 we list the integrated flux density estimates for the SNR between 10 and 10700 MHz.We applied a correction factor over the wide spectral range between 408 and 10700 MHz in orderto place each value on the flux scale of Baars et al. (1977), except in some cases for which noinformation was available on the flux value considered for the primary calibrators.A plot of the integrated radio continuum spectrum for the SNR IC 443 is shown in Fig. 7.Our new integrated flux measurements at 74 and 330 MHz are indicated by filled circle symbols.From the spectrum it is evident that the flux densities measured at the lowest radio frequencies,and particularly the flux value at 10 MHz, lie below the general trend of the data suggesting thepresence of thermal absorption along the line of sight. In order to fix the integrated spectral indexof IC 443 we first use a single power law spectrum to fit the spectrum well down to our lowest astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequencies Table 2.
Integrated flux densities on the SNR IC443
Frequency Scaled flux References Frequency Scaled flux References(MHz) density (Jy) (MHz) density (Jy)10 . . . . . . . 400 ± (a) Bridle & Purton (1968) 430 . . . . . . 245 ± (c) Kundu & Velusamy (1968)20 . . . . . . . 600 ± (a) Braude et al. (1969) 513 . . . . . . 205 ± (c) Bondar et al. (1965)22 . . . . . . . 615 ± (a) Roger et al. (1986) 610 . . . . . . 215 ± (c) Dickel & McKinley (1969)22.25 . . . . 535 ± (a) Roger et al. (1969) 635 . . . . . . 179 ±
18 Milne & Hill (1969)22.3 . . . . . 529 ± (a) Guidice (1969) 740 . . . . . . 164 ± (c) Bondar et al. (1965)25 . . . . . . . 630 ± (a) Braude et al. (1969) 750 . . . . . . 190 ± (c) Hogg (1964)26.3 . . . . . 600 ± (a) Viner & Erickson (1975) 960 . . . . . . 196 ±
24 Harris & Roberts (1960)26.7 . . . . . 561 ± (a) Guidice (1969) 960 . . . . . . 165 ± (c) Bondar et al. (1965)33.5 . . . . . 582 ± (a) Guidice (1969) 1000. . . . . 160 ± (c) Milne (1971)34.5 . . . . . 440 ± (a) Dwarakanath et al. (1982) 1390. . . . . 177 ±
15 Westerhout (1958)38 . . . . . . . 650 ± (a) Baldwin & Dewhirst (1954) 1400. . . . . 170 ± (c) Hogg (1964)38 . . . . . . . 730 ± (a) Blythe (1957) 1400. . . . . 146 ±
18 Wanner (1961)38 . . . . . . . 460 ± (a) Williams et al. (1966) 1410. . . . . 131 ±
13 Milne & Hill (1969)38.6 . . . . . 547 ± (a) Guidice (1969) 1419. . . . . 130 ±
13 Green (1986)74 . . . . . . . 470 ± (b) This work 1420. . . . . 160 ± (c) Hagen et al. (1955)81.5 . . . . . 420 ± (a) Baldwin & Dewhirst (1954) 1420. . . . . 138 ±
15 Hill (1972)81.5 . . . . . 470 ± (a) Shakeshaft et al. (1955) 2650. . . . . 86 ± ± (a) Kovalenko et al. (1994) 2700. . . . . 104 ±
15 Milne (1971)102 . . . . . . 480 ± (a) Kovalenko et al. (1994) 3000. . . . . 100 ± (c) Hogg (1964)111 . . . . . . 440 ± (a) Kovalenko et al. (1994) 3125. . . . . 100 ± (c) Kuz’min et al. (1960)151 . . . . . . 280 ± (a) Green (1986) 4170. . . . . 100 ± (c) Hirabayashi & Takahashi (1972)159 . . . . . . 270 ± (a) Edge et al. (1959) 5000. . . . . 79 ±
11 Milne (1971)178 . . . . . . 210 ± (a) Bennett (1962) 5000. . . . . 85 ± (c) Kundu & Velusamy (1969)195 . . . . . . 290 ± (a) Kundu & Velusamy (1968) 6640. . . . . 70 ± (c) Dickel (1971)330 . . . . . . 248 ± ( b ) This work 8000. . . . . 90 ±
18 Howard & Dickel (1963)400 . . . . . . 230 ± (a) Davies et al. (1965) 10700 . . . 60 ± (c) Kundu & Velusamy (1972)400 . . . . . . 210 ± (a) Seeger et al. (1965)400 . . . . . . 251 ± (a) Kellermann (1964)408 . . . . . . 289 ±
28 Colla et al. (1971) (a)
No correction to Baars et al. (1977) scale was applied. (b)
Flux density scale from VLA Calibrator Manual, http: // / ∼ gtaylor / csource.html. (c) The correction factor was not available. measurement at 74 MHz, excluding the flux density value at 10 MHz (represented as a solid line inFig. 7). A weighted fit produces a spectral index α = − . ± .
02 ( S ν ∝ ν α ). This result agrees verywell within the error limits with the previous estimates presented by Erickson & Mahoney (1985,and references therein). If we consider the lowest frequencies observations, these measurementscan be fit with a power law plus an exponential turnover using the Eq. 1 (indicated by the dottedline in Fig. 7) astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequencies S ν = S (cid:18) ν
330 MHz (cid:19) α exp[ − τ (cid:18) ν
330 MHz (cid:19) − . ] (1)Here, 330 MHz is a reference frequency at which an integrated flux density S and an opticaldepth τ are measured, α represents the non-thermal integrated spectrum, which is assumed tobe constant throughout the radio band. We have made a weighted fit of the distribution of datapoints over four decades in frequency and find for the whole SNR a single radio spectral index α = − . ± .
01, and an average optical depth τ = (7 ± × − . The free-free continuumoptical depth at 10 MHz derived from the relation τ = τ [10 / − . is τ = .
07, while theoptical depth at 74 MHz is τ = .
02. Our results indicate that although the absorption becomessignificantly stronger at 10 MHz this e ff ect is negligible at 74 MHz. In addition, we note that thespectral index produced by considering free-free absorption is consistent with that derived witha power law. Such a result is not surprising given the results of Sect. 4.1, that clearly indicatedthe presence of thermal absorption, though not at a level to impact the integrated flux at 74 MHz.Future measurements at intermediate frequencies (e.g. ∼
30 MHz) are needed to understand if theturnover inferred from the 10 MHz measurement is related to the subtle absorption revealed in thespectral index analysis presented in Sect. 4.1.
Fig. 7.
Radio continuum spectrum for SNR IC 443 obtained from the flux density values listed inTable 2. The filled circle symbols correspond to the new flux density measurements calculated usingthe VLA data at 74 and 330 MHz presented in this work, the open symbols are for radio observa-tions previously published and, where possible, brought onto the flux density scale of Baars et al.(1977). Solid line represents the linear fit to the flux density values excluding that at 10 MHz, whichproduces a spectral index α = − . ± .
02 ( S ∝ ν α ). Dotted line shows a fit to all of the plottedvalues if absorption were present (eq. 1), which yields a spectral index α = − . ± . ffi culty in separating the nebular emission from its surroundings.A weighted fit to all the flux densities produces a spectral index for the PWN of α = − . ± astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequencies .
05, which is zero within the uncertainty in the fit itself. This result is similar to that obtained byOlbert et al. (2001).
Fig. 8.
Radio continuum spectrum for the pulsar wind nebula (PWN) aroundCXOU J061705.3 + S ∝ ν α ) yields a spectral index α = − . ± .
05. The filled circles symbolsare from the flux density estimates at 330 and 1420 MHz presented in the current work, while theopen symbols are from Olbert et al. (2001).
5. Radio spectral index and the near infrared emission
Based on our accurate, spatially resolved radio continuum 74 /
330 MHz spectral map, we haveinvestigated the correlation between radio spectral features and the near infrared emission (NIR).Figure 9a shows two 74 /
330 MHz spectral index contours (traced at α = − .
05 and α = − .
25) enclosing the east rim where we found regions with very flat spectrum in IC 443 superposedonto the NIR detected in the J (1.25 µ m), H (1.65 µ m), and K s (2.17 µ m) bands as taken from theTwo Micron All Sky Survey (2MASS) (Rho et al. 2001). The 2MASS image of IC 443 showsthe dramatic contrast in near infrared color between the east rim and the southern portion of theremnant. In the color representation of the infrared emission, blue traces the J -band flux, while theinfrared data in the H and K s are shown in green and red, respectively; white thus enlightens regionswhere all the three IR bands overlap. To facilitate the comparison, Fig. 9b displays the 330 MHzcontinuum image of IC 443 with a grayscale selected to emphasize the brightest radio emission.An impressive agreement is observed in location, size and shape between the NIR emissiondetected in the H and J bands and the flattest spectral feature as traced by the α contours alongthe eastern edge of IC 443. This correspondence begins in the northernmost part of the remnant andextends down to positions near dec. ∼ + ◦ ′ . From Fig. 9b it is also notable that the brightestradio synchrotron emission perfectly matches the bright emission in the NIR bands. As noticedby Rho et al. (2001), at this site the predominant constituent of the emission in the J and H bandsis the [FeII] line, with a minor contribution from other multi-ionized species like [NeII], [NeIII],[SiII], [SIII], etc. Rho et al. (2001) proposed a model in which the infrared emission from the ion-ized species in the east bright radio limb of IC 443 comes from shattered dust produced by a fastdissociating J-type shock. The present accurate comparison between radio spectral indices and IR astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequencies emission confirms this model. In e ff ect, the passage of a dissociative shock through a molecularcloud not only dissociates molecules but also ionize the atoms. Such collisional ionization is re-sponsible for the thermal absorbing electrons that produce the peculiar very flat spectrum areasobserved all along the eastern border of IC 443. This interpretation is in agreement with studiesbased on CO and X-ray observations which conclude that the large molecular cloud complex islocated in front of IC 443 (Cornett et al. 1977; Troja et al. 2006). Evidence for a similar situationhas also been observed in the ringlike morphology of 3C 391 (Brogan et al. 2005), another SNRknown to be interacting with a molecular cloud. Fig. 9. (a)
A color representation of the near-infrared emission observed with 2MASS in the J (inblue), H (in green), and K s (in red) bands (Rho et al. 2001). The overlaid contours trace the flattest-spectrum radio structures of the SNR IC443 between 74 and 330 MHz at α = − .
05 and − . (b)
330 MHz image of IC 443 showing the locations of the brightest regions of the remnant.Towards the interior of the bright eastern shell, the spectral index gradually steepens with po-sition in coincidence with a decrease in the intensity of the radio emission. Widely distributed K s band emission is observed in the 2MASS image in this part of the remnant, which was proposed todelineate H shocked gas from the region interacting with the adjacent molecular cloud.In contrast to the excellent agreement between the ionic emitting gas and the flattest spectrumfeatures found in the eastern bright limb, no much obvious correspondence is observed in thesouthern part of IC 443 with the exception of a spectral component located at the northern extremeof the ridge, near R.A. = h m s , dec. = ◦ ′ . This poor IR / radio-spectrum correspondenceis consistent with the hypothesis proposed before, in which the flat spectrum features in this part ofIC 443 have a non-thermal origin.In the remainder of this section we attempt to infer, using our new measurements at low radiofrequencies, the physical properties of the area of thermal absorption seen in Fig. 9a spatiallycoincident with the ionized fine-structure line emitting atoms, i.e. the eastern half of the SNR. Atradio wavelengths, the emission measure (EM) is given by, astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequencies EM = . × − a (T e , ν ) − ν . τ ν T . cm − pc , (2)where a ( T e , ν ) is the Gaunt factor assumed to be 1, a correct value for the range of astrophysicalquantities involved in our calculations; ν and τ ν are the frequency measured in MHz and the free-free optical depth, respectively, and T e is the electron temperature in K of the intervening ionizedgas. By measuring the relative strengths of the [FeII] lines observed in the near and mid infraredemission, Rho et al. (2001) conclude that the emitting region behind the J-type shocks as observedin the J and H bands of the 2MASS has a temperature of 12000 K. Although not quantified, theauthors recognize a large uncertainty associated with this magnitude as a consequence of di ff erentbeam size and possibly di ff erent filling factors in their measurements. If we assume that T e is ina reasonable range between 8000-12000 K (which includes the temperature as estimated from theIR observations) and use the optical depth derived from our radio study particularly for this regionwhere the thermal absorption is stronger ( τ ∼ .
3) we obtain an EM between approximately2 . × and 5 . × cm − pc for the eastern rim. By combining this emission measure withthe postshock electron density, we can roughly calculate the thickness of the molecular gas layerthat has been dissociated and ionized by the SNR shock front. If we assume an electron density of n e ∼
500 cm − as estimated by Fesen & Kirshner (1980) and Reach & Rho (2000) on the basis offorbidden [FeII] lines, we conclude that the dissociation and ionization processes took place in athin screen of about 3.4 to 6.0 × cm ( ∼
6. Comparison with the molecular distribution
In Fig. 10 we present a comparison between our VLA 330 MHz image and new CO ( J = ′′ × ′′ , velocity resolutionof 0.37 km s − , and rms noise level of 0.1-0.3 K at a velocity resolution of ∼ − ). Thecontours superposed on the radio emission depict the CO emission integrated in the range between −
10 and − − , which includes the systemic velocity of IC 443. As described before, themolecular material is preferentially located in the center of the remnant extending in the southeast-northwest direction. The spatial distribution of the molecular gas across IC 443 is clearly non-uniform. Earlier observations have identified the presence of various clumps of molecular gas withbroad line widths, as expected from the interaction with the supernova shock front (clumps labeledfrom A to H in the nomenclature of Dickman et al. 1992).On the basis of the new image at 330 MHz it is possible to recognize details previously un-noticed in the spatial comparison between the radio emission and the molecular gas. In Fig. 10we display the regions where good correlation between radio features and molecular gas distribu-tion is observed. Particularly noticeable is Fig. 10b where it is apparent that the indentation of theeastern border in radio occurs near a region where a significant enhancement in the CO emissionis detected. The molecular complex is transverse to the radio indentation with the maximum ofthe CO emission shifted to the southwest in at least 3 ′ from the border of the SNR (in the regionof the molecular clump E). It is possible that the singular indentation has formed as the result of astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequencies the supernova shock front wrapping around a dense clump. Also, Fig. 10c shows the presence of aconcentration in the CO emission around R.A. = h m s , dec. = + ◦ ′ ′′ (molecular clumpB), that matches a local maximum in the radio emission. The morphological matching between theradio synchrotron emission and the molecular gas is especially remarkable at the northern extremeof the southern ridge of IC 443 near R.A. = h m s , dec. = + ◦ ′ ′′ (as shown in Fig. 10d),where the molecular contours, delineating high density gas in the region of the clump G, are ob-served enclosing the bright radio emission. Fig. 10.
A comparison of the radio continuum emission of IC 443 and the CO ( J = −
10 to − − as taken fromZhang et al. (2010). Close-ups of three interesting areas are displayed around the center image,with the CO contours overlapping. The white letter in each panel corresponds to the designation ofthe molecular clump by Dickman et al. (1992).We also searched for spectral evidence of shock / cloud interaction. Figure 11 displays an over-lay of the CO ( J = CO contoursalong the bright southern ridge and towards the northwest, suggesting that these features must beregions where strong shocks encountered denser material (as discussed by Anderson & Rudnick1993). In the eastern periphery the situation is di ff erent. The very flat spectrum component ex-tends over an area considerably larger than the molecular cloud, precisely because in this region, asshown in Sect. 5, most of the molecules were dissociated and ionized, absorbing the radio emissionat low frequencies. astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequencies Fig. 11.
A comparison of the spectral index distribution with the molecular emission. The colorrepresentation corresponds to the spatially resolved spectral index map between 74 and 330 MHz,while the overlaid contours trace the C0 ( J = −
10 and − − (Zhang et al. 2010). The positions where OH(1720 MHz) maser emission were detected are indi-cated with open white diamonds (Ho ff man et al. 2003; Hewitt 2009).In Fig. 11 we have identified with open white diamonds the regions with OH (1720 MHz) maseremission as detected by Hewitt (2009). In the case of SNR W28, Dubner et al. (2000) demonstrateda clear correspondence between regions of flat spectral index and OH maser emission. We searchfor a similar correlation in IC 443. From Fig. 11 we conclude that there is not a simple associationbetween maser locations and spectral index features but rather both flat and steep components areobserved near masers areas. This fact can be compared with previous results, which suggest thatmaser regions in IC 443 arise from regions with di ff erent shock geometry: a shock mostly propa-gating towards the line of sight in the southern OH maser emission and a transverse shock in thewesternmost OH emitting region (Claussen et al. 1997; Ho ff man et al. 2003; Hewitt 2009). We findthat the OH maser area with transverse shock (R.A. = h m s .
6, dec. = + ◦ ′ ′′ .
7) corre-lates well with flat spectral index ( α = − . = h m s .
29, dec. = + ◦ ′ ′′ .
5) is associated with a steepspectrum ( α = − . α = − .
4) is observed towardsthe easternmost OH maser area with tangential shock (R.A. = h m s .
67, dec. = + ◦ ′ ′′ . ff ects.
7. Comparison with TeV emission
Figure 12 compares the TeV gamma-ray significance contours obtained from
VERITAS obser-vations (Acciari et al. 2009) along with the 74 /
330 MHz spectral index map presented in Fig. 6(Fig. 12a), with the total intensity features of IC 443 at 330 MHz (Fig. 12b), and with the CO ( J = + + astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequencies emission region. On the contrary, it is remarkable the morphological coincidence between the TeVemission and the molecular gas distribution (Fig. 12c), at least up to the level that the available TeVstatistic permits to confirm. The interpretation of this striking correspondence is beyond the scopeof this paper. Fig. 12.
Very high energy gamma-rays contours as taken from
VERITAS (Acciari et al. 2009) su-perposed with (a) the 74 /
330 MHz spectral index map presented in Fig. 6, (b) radio continuumemission at 330 MHz, and (c) C0 ( J = −
10 to − − tracingthe molecular cloud interaction with IC 443 (Zhang et al. 2010). The white plus sign and the whitecross in (a) mark the positions of the source CXOU J061705.3 + +
8. Conclusions
In this work we report on new full-synthesis imaging of the Galactic SNR IC 443 generated frommultiple-configuration VLA observations at 74 and 330 MHz. These high fidelity images constitutethe best angular resolution, low frequency radio study published to date on this classic remnant.Based on these new data we measured integrated flux densities for this object of S
74 MHz = ±
51 Jy and S
330 MHz = ±
15 Jy. On the basis of these new total flux density measurements at 74and 330 MHz together with previously published values, we recalculated the integrated spectrumof IC 443. We have made this analysis, for the first time, taking into account that the spectrum of astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequencies IC 443 has a reported turnover at the lowest radio frequencies. The fit over a wide frequency rangeproduces a radio spectral index α = − . ± .
01 with free-free thermal optical depth τ ∼ × − (or τ ∼ . α isrelatively unchanged from that inferred from a single power law fit ( α = − . ± . α ∼ − . ff use component in the interior of IC 443 with α between − . − .
85, consistent withthose expected from linear di ff usive shock acceleration processes. Finally, another region with non-uniform flattening in the spectrum is identified near the southern ridge of IC 443. We investigatedthe connection between these spatial spectral variations and IR and molecular emission distribution.The comparison with 2MASS NIR data underscored an impressive coincidence between theeastern radio flattest spectrum region and NIR ionic lines which account for most of the infraredemission observed towards this part of IC 443. Based on the presence of this IR emission, whichconfirms the existence of a strong J-type shock dissociating molecules and ionizing atoms, we con-clude that the most likely explanation for the flattest spectrum observed here, is that it is producedby free-free absorption at 74 MHz along the line of sight (with τ up to ∼ . ∼ .
02 pc) that extends all alongthe eastern border. Such a thin layer would be the product of the dissociating / ionizing action of theSNR shock over the adjacent molecular gas. It is important to remark that the low optical depthsderived for the whole SNR ( τ ∼ × − and τ ∼ .
02) indicate that the 74 and 330 MHzintegrated emission is not significantly attenuated, highlighting that the thermal absorption inferredfrom the spatially resolved spectral index map is a relatively subtle and localized e ff ect. Our resultrepresents only the second case, following 3C 391 (Brogan et al. 2005), of spatially resolved ther-mal absorption delineating the interaction of a SNR / molecular cloud shock boundary. Moreoverit confirms such phenomena as common and a rich area of investigation for future, low frequencystudies of Galactic complexes.On the other hand, from the molecular studies it is known that the southern ridge, where the sec-ond region with flat spectrum was identified, is the site in which the most complex interaction be-tween the SNR shock and the external molecular cloud is occurring. We used Zhang et al. (2010)’sobservations of the CO ( J = CO observations revealed spatial irregularities in the denser moleculargas that are well associated with features of bright emission in radio and good agreement betweenthe flat radio spectrum region towards the south and molecular emission. We conclude that herethe flat radio spectrum is predominantly a signature of shock acceleration in a region with strongpostshock densities and enhanced magnetic fields produced after the interaction of the blast wavewith dense ambient medium.Furthermore, from the comparison of the molecular environment mapped by CO data withthe local variations of the spectral index, we find evidence that the shocked molecular gas, as astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequencies illuminated by the OH (1720 MHz) maser emission, is coupled with a flattening in the radio spectralindex in the locations where the shock is transverse to the line of sight, while a steep spectrum isobserved in a OH maser region in which the shocks are propagating along the line of sight.From the new images, we also analyzed the pulsar wind nebula around the sourceCXOU J061705.3 + S PWN330 MHz = . ± .
05. We have also measured the flux density of the PWN in the radiocontinuum image at 1420 MHz presented by Lee et al. (2008) obtaining S PWN1420 MHz = . ± . α ∼ . VERITAS statistics we compared the TeV emission with the CO distribution, finding an excellent morphological correlation between the high energy emissionand the distribution of the molecular gas. No correspondence was found, however, between thegamma-ray emission as observed by
VERITAS and radio spectral or intensity features in radio. Afuture paper will address these findings in connection with the origin of the gamma-ray emissiondetected in IC 443.
Acknowledgements.
We are very grateful to T. B. Humensky for kindly providing us the
VERITAS significance contours,to Z. Zhang for the CO ( J = / California Institute of Technology, funded by the National Aeronautics and Space Administration and the NationalScience Foundation. The optical image used in this work is from the Second Palomar Observatory Sky Survey (POSS-II),which was made by the California Institute of Technology with funds from the National Science Foundation, the NationalGeographic Society, the Sloan Foundation, the Samuel Oschin Foundation, and the Eastman Kodak Corporation. Thisresearch has made use of the NASA’s ADS Bibliographic Services. Data processing was carried out using the HOPE PCcluster at IAFE. This research was partially funded through CONICET (Argentina) grant PIP 112-200801-02166, ANPCYT-PICT (Argentina) grant 0902 /
07, and ANPCYT-PICT (Argentina) 08-0795 grant. Basic research in radio astronomy at theNaval Research Laboratory is supported by 6.1 base funds.
References
Abdo, A. A., Ackermann, M., Ajello, M., et al. 2010, ApJ, 712, 459Acciari, V. A., Aliu, E., Arlen, T., et al. 2009, ApJ, 698, L133Albert, J. & et, a. 2007, ApJ, 664, L87Anderson, M. C. & Rudnick, L. 1993, ApJ, 408, 514Anderson, M. C. & Rudnick, L. 1996, ApJ, 456, 234Asaoka, I. & Aschenbach, B. 1994, A&A, 284, 573Baars, J. W. M., Genzel, R., Pauliny-Toth, I. I. K., & Witzel, A. 1977, A&A, 61, 99Baldwin, J. E. & Dewhirst, D. W. 1954, Nature, 173, 164Bell, A. R. 1978, MNRAS, 182, 147Bennett, A. S. 1962, MmRAS, 68, 163Bietenholz, M. F., Kassim, N., Frail, D. A., et al. 1997, ApJ, 490, 291Blythe, J. H. 1957, MNRAS, 117, 652Bocchino, F. & Bykov, A. M. 2001, A&A, 376, 248Bondar, L. N., Krotikov, V. D., Stankevich, K. S., & Tseitlin, N. M. 1965, Radiophysics and Quantum Electronics, 8, 437Braude, S. Y., Lebedeva, O. M., Megn, A. V., Ryabov, B. P., & Zhouck, I. N. 1969, MNRAS, 143, 289Braun, R. & Strom, R. G. 1986a, A&AS, 63, 345Braun, R. & Strom, R. G. 1986b, A&A, 164, 193Bridle, A. H. & Purton, C. R. 1968, AJ, 73, 717 astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequencies Brogan, C. L., Lazio, T. J., Kassim, N. E., & Dyer, K. K. 2005, AJ, 130, 148Castelletti, G., Dubner, G., Brogan, C., & Kassim, N. E. 2007, A&A, 471, 537Castelletti, G., Dubner, G., Golap, K., & Goss, W. M. 2006, A&A, 459, 535Claussen, M. J., Frail, D. A., Goss, W. M., & Gaume, R. A. 1997, ApJ, 489, 143Colla, G., Fanti, C., Fanti, R., et al. 1971, AJ, 76, 953Condon, J. J., Cotton, W. D., Greisen, E. W., et al. 1998, AJ, 115, 1693Cornett, R. H., Chin, G., & Knapp, G. R. 1977, A&A, 54, 889Cornwell, T. J. & Perley, R. A. 1992, A&A, 261, 353Davies, M. M., Gelato-Volders, L., & Westerhout, G. 1965, Bull. Astron. Inst. Netherlands, 18, 42Dickel, J. R. 1971, PASP, 83, 343Dickel, J. R. & McKinley, R. R. 1969, ApJ, 155, 67Dickman, R. L., Snell, R. L., Ziurys, L. M., & Huang, Y. 1992, ApJ, 400, 203Dubner, G. M., Vel´azquez, P. F., Goss, W. M., & Holdaway, M. A. 2000, AJ, 120, 1933Dulk, G. A. & Slee, O. B. 1975, ApJ, 199, 61Dwarakanath, K. S., Shevgaonkar, R. K., & Sastry, C. V. 1982, Journal of Astrophysics and Astronomy, 3, 207Edge, D. O., Shakeshaft, J. R., McAdam, W. B., Baldwin, J. E., & Archer, S. 1959, MmRAS, 68, 37Erickson, W. C. & Mahoney, M. J. 1985, ApJ, 290, 596Esposito, J. A., Hunter, S. D., Kanbach, G., & Sreekumar, P. 1996, ApJ, 461, 820Fesen, R. A. 1984, ApJ, 281, 658Fesen, R. A. & Kirshner, R. P. 1980, ApJ, 242, 1023Gaensler, B. M., Chatterjee, S., Slane, P. O., et al. 2006, ApJ, 648, 1037Green, D. A. 1986, MNRAS, 221, 473Guidice, D. A. 1969, PhD thesis, THE OHIO STATE UNIVERSITY.Hagen, J. P., Lilley, A. E., & McClain, E. F. 1955, ApJ, 122, 361Harris, D. E. & Roberts, J. A. 1960, PASP, 72, 237Hewitt, J. W. 2009, PhD thesis, Northwestern UniversityHewitt, J. W., Yusef-Zadeh, F., Wardle, M., Roberts, D. A., & Kassim, N. E. 2006, ApJ, 652, 1288Hill, L. E. 1972, MNRAS, 157, 419Hirabayashi, H. & Takahashi, T. 1972, PASJ, 24, 231Ho ff man, I. M., Goss, W. M., Brogan, C. L., Claussen, M. J., & Richards, A. M. S. 2003, ApJ, 583, 272Hogg, D. E. 1964, ApJ, 140, 992Howard, III, W. E. & Dickel, H. R. 1963, PASP, 75, 149Humensky, T. B. 2008, in International Cosmic Ray Conference, Vol. 2, International Cosmic Ray Conference, 835–838Kassim, N. E. 1989, ApJ, 347, 915Kassim, N. E., Lazio, T. J. W., Erickson, W. C., et al. 2007, ApJS, 172, 686Kassim, N. E., Perley, R. A., Dwarakanath, K. S., & Erickson, W. C. 1995, ApJ, 455, L59 + Kawasaki, M. T., Ozaki, M., Nagase, F., et al. 2002, ApJ, 572, 897Kellermann, K. I. 1964, ApJ, 140, 969Kovalenko, A. V., Pynzar’, A. V., & Udal’Tsov, V. A. 1994, Astronomy Reports, 38, 78Kundu, M. R. & Velusamy, T. 1968, MNRAS, 140, 173Kundu, M. R. & Velusamy, T. 1969, ApJ, 155, 807Kundu, M. R. & Velusamy, T. 1972, A&A, 20, 237Kuz’min, A. D., Levchenko, M. T., Noskova, R. I., & Salomonovich, A. E. 1960, AZh, 37, 975Lacey, C. K., Lazio, T. J. W., Kassim, N. E., et al. 2001, ApJ, 559, 954Leahy, D. A. 2004, AJ, 127, 2277Lee, J., Koo, B., Yun, M. S., et al. 2008, AJ, 135, 796Lozinskaya, T. A. 1981, Soviet Astronomy Letters, 7, 17Milne, D. K. 1971, Australian Journal of Physics, 24, 429Milne, D. K. & Hill, E. R. 1969, Australian Journal of Physics, 22, 211Olbert, C. M., Clearfield, C. R., Williams, N. E., Keohane, J. W., & Frail, D. A. 2001, ApJ, 554, L205Petre, R., Szymkowiak, A. E., Seward, F. D., & Willingale, R. 1988, ApJ, 335, 215Reach, W. T. & Rho, J. 2000, ApJ, 544, 843 astelletti et al .: High-resolution radio study of SNR IC 443 at low radio frequencies Reynolds, S. P. & Ellison, D. C. 1992, ApJ, 399, L75Rho, J., Jarrett, T. H., Cutri, R. M., & Reach, W. T. 2001, ApJ, 547, 885Rho, J. & Petre, R. 1998, ApJ, 503, L167 + Roger, R. S., Costain, C. H., & Lacey, J. D. 1969, AJ, 74, 366Roger, R. S., Costain, C. H., & Stewart, D. I. 1986, A&AS, 65, 485Seeger, C. L., Westerhout, G., Conway, R. G., & Hoekema, T. 1965, Bull. Astron. Inst. Netherlands, 18, 11Shakeshaft, J. R., Ryle, M., Baldwin, J. E., Elsmore, B., & Thomson, J. H. 1955, MmRAS, 67, 106Steer, D. G., Dewdney, P. E., & Ito, M. R. 1984, A&A, 137, 159Troja, E., Bocchino, F., & Reale, F. 2006, ApJ, 649, 258Viner, M. R. & Erickson, W. C. 1975, AJ, 80, 931Wang, Z. R., Asaoka, I., Hayakawa, S., & Koyama, K. 1992, PASJ, 44, 303Wanner, J. F. 1961, PASP, 73, 143Westerhout, G. 1958, Bull. Astron. Inst. Netherlands, 14, 215Williams, P. J. S., Kenderdine, S., & Baldwin, J. E. 1966, MmRAS, 70, 53Zhang, Z., Gao, Y., & Wang, J. 2010, Science China Physicis, Mechanics & Astronomy, 54, 1357, in press
List of Objects ‘IC 443’ on page 2‘G189.6 + + +2230’ on page 19