Journal of The Korean Astronomical Society 38: 385 ∼ 414, 2005 INFRARED SUPERNOVA REMNANTS IN THE SPITZER GLIMPSE FIELD Ho-Gyu Lee Astronomy Program,SEES, Seoul National University, Seoul 151-742,Korea E-mail: [email protected] (Received October 5, 2005; Accepted December 7, 2005) ABSTRACT 6 We have searched for infrared emission from supernova remnants (SNRs) included in the Spitzer 0 Galactic Legacy Infrared Mid-Plane Survey Extraordinaire (GLIMPSE) field. At the positions of 100 0 known SNRs, we made 3.6, 4.5, 5.8, and 8.0 µm band images covering the radio continuum emitting 2 areaofeachremnant. In-depthexaminationsoffourbandimagesbasedontheradiocontinuumimages n of SNRs result in the identification of sixteen infrared SNRs in the GLIMPSE field. Eight SNRs show a distinct infrared emission in nearly all the four bands, and the other eight SNRs are visible in more J thanoneband. WepresentfourbandimagesforallidentifiedSNRs,andRGB-colorimagesforthefirst 3 eightSNRs. Theseimagesarethe firsthighresolution(<2′′)imageswithcomparativeresolutionofthe 2 radio continuum for SNRs detected in the mid-infrared region. The images typically show filamentary emission along the radio enhanced SNR boundaries. Most SNRs are well identified in the 4.5 and 5.8 2 µm bands. We give a brief description of the infrared features of the identified SNRs. v 8 Key words : ISM: supernova remnants – infrared 1 5 0 1 I. INTRODUCTION plane survey performed by the Midcourse Space Ex- 5 periment (MSX). The spectral analysis including the 0 Infrared emission from supernova remnants (SNRs) IRAS data suggested that the line contribution could / is important in the evolution of SNRs; furthermore, be significant in the near/mid infrared. The Infrared h it also delivers unique information pertaining to SNR Space Observatory (ISO) conducted spectroscopic and p physics. DustgrainssweptupbySNRs areheatedand imaging observations of nine SNRs covering 3–8 µm - o disrupted, and emit infrared radiation that dominate bands of of the GLIMPSE field. The majority of them r the cooling during most of the SNR lifetime. Their in- werebrightyoungSNRssuchasCasA,Tycho,Kepler, st frared emission contains the most direct informations and Crab (e.g., Douvion et al. 1999, 2001a). Several a on the physical processes associated with dust grains radiative SNRs interacting with molecular clouds such v: behind the SNR shocks (Dwek & Arendt 1992; Draine as IC 443 and 3C 391 show spatially separated molec- i & Mckee 1993). In young SNRs, the thermal infrared ular and ionic lines (Cesarsky et al. 1999;Reach et al. X emission may reveal the composition of newly con- 2002). r densed dust grains from ejecta, while the non-thermal GLIMPSE (Benjamin et al. 2003) is one of the a infrared synchrotron emission is important for under- largest surveys performed in the infrared region. The standing the acceleration and evolution of high-energy survey provides fully-sampled, confusion-limited 3.6, particles in SNRs. Ionic and H emission lines in in- 2 4.5, 5.8 and 8.0 µm Spitzer IRAC data for the inner frared emission serve as a very important diagnostic two-thirds region of the Galactic plane with a spatial tool for understanding the underlying physics of SNR resolution of ∼2′′. Most star formation regions, galac- shocks. This is particularly true for distant SNRs that tic molecular rings, and more than 4 spiral arms are suffer large extinction. However, despite such useful- located in the surveyed area. The GLIMPSE project ness, the infrared is the least-exploited waveband in focusesonGalacticstructureandstarformation. Con- the study of SNRs. currently, the wide, deep, and unbiased survey pro- The all-sky survey by the Infrared Astronomical vides a unique opportunity to investigate diffuse ob- Satellite (IRAS) detected infrared emission from a sig- jects,whichoccupylargeareasinthesky. Inthispaper, nificant (∼30%) fraction of the known SNRs in the we studied infrared emission from all SNRs contained Galaxy(Arendt1989;Sakenetal. 1992). Thusfar,the intheGLIMPSEField. Section2explainsthedataand IRAS surveyis the onlyinfrareddatasetthatincludes identification method of the infrared emission from in- theentireensembleofGalacticSNRs. Althoughitsres- dividual SNRs. Section 3 reports the identified SNRs olution, ofthe orderofa few-arcminutes,is useful only and their characteristics. A brief discussion and sum- to the extent that large SNRs do not suffer from any mary is included in Section 4. An independent study confusioneffect,itshowedthatthermallyheatedgrains to search for SNRs in the GLIMPSE field (Reach et are responsible for the far-infrared emission of SNRs. al. 2005) exists, and we provide a note on their study In the mid-infrared region, Kim et al. (2005) reported at the end of this section. Finally, we briefly describe infrared emission from eight SNRs using the Galactic the infrared features of the identified SNRs in the Ap- – 385 – 386 LEE pendix. the MOST† Supernova Remnant Catalog (Whiteoak & Green 1996) for the southern SNRs. The identified II. DATAANDIDENTIFICATIONOFSNRS SNRs were comparedin detail with high-resolutionra- dio continuum images for confirmation. The GLIMPSE data cover two thirds of the inner Galactic plane (10◦< |l|< 65◦, |b|<1◦) in four wave- III. RESULTS length bands centered at 3.6, 4.5, 5.8, and 8.0 µm (Benjamin et al. 2003). More than 310,000 frames Among the 100 SNRs examined, eight bright SNRs were taken by the Infrared Array Camera (IRAC) aredistinctlyidentified. TheyarelistedinTable1. The across several days between March 9 and November position, size, and type in columns (2)–(5) are quoted 1, 2004. Each IRAC frame has approximately the from the Green’s catalog. These SNRs show infrared same field of view of 5′.2 × 5′.2 with a pixel size of emissioninallfourbandsatthepeakposition,andthe ∼1′′.2 (Fazio et al. 2004). Their point response func- extent of the emission is comparable to the remnant tion varies within the array. The in-flight mean val- size. Figures 1 to 8 show their images. The gray-scale ues of the FWHM are 1′′.66, 1′′.72, 1′′.88, and 1′′.98 range varies from the median of the image pixels to in the 3.6, 4.5, 5.8, and 8.0 µm bands, respectively. threetimesoftheirstandarddeviations. This,however, The total on-source integration time per position was does not hold in the case of two SNRs, which show a 4 seconds after combining the overlapped and revis- variationofuptofivetimesforW49Bandseventimes ited frames. The 5 σ sensitivities were estimated as forG349.7+0.2. Ineachfigure,aradiocontinuummap 0.2 mJy for the two shorter bands and 0.4 mJy for is also included at the bottom left to assist in the ver- the longer two bands. The basic calibration and data ification of the associated infrared emission features. handlingwereperformedbytheSpitzerScienceCenter The position and scale of radio continuum maps are (SSC)Pipeline versionsfromS9.5.0toS10.5.0depend- set to match those of the infrared images. The radio ing on the observed date. Further image processing, continuum maps have angular resolutions of 2′′–5′′ ex- such as column pulldown, banding corrections, mask- cept several SNRs in the southern sky and the large ing pixel, andremovingstraylight,wereperformedfor SNR W 44. Thus, in most cases, thin filamentary fea- the image products by the GLIMPSE Pipeline. We tures in the infrared can be directly compared to the used the GLIMPSE Atlas-Mosaicked Images for each corresponding features in radio. Figures 9 to 16 show band archived by SSC. Each mosaicked image was RGB-color images of these SNRs created from 8.0 µm given by a 1◦.1 × 0◦.8 tile in Galactic coordinates. (Red),5.8µm(Green),and4.5µm(Blue)images. The Among the 231 Galactic SNRs listed in Green’s cat- dynamicrangeofeachcolorisidenticaltoitsgray-scale alog(2004),100SNRsexistintheGLIPMSEfield. We image. obtained the images of these 100 SNRs in all the four In addition to the above eight SNRs, we have iden- bands. The images use the equatorial coordinate sys- tified another eight SNRs that are marginally bright temandhaveapixelsizeof1′′.2. ForlargeSNRsorfor enough to be discernible from the background. They SNRsspreadingacrosstheboundariesoftheGLIMPSE arealsolistedinTable1,andtheirimagesareshownin Atlas-Mosaickedimages,wehavecoaddedadjacentim- Figures17to24. Thescalerangeofthefiguresarenar- ages. The image processing was made by using the rower than (typically half of) those for the distinctly- SWARP software package distributed by TERAPIX identified SNRs to show the associated emission more (http://terapix.iap.fr/soft/swarp). The imaging cen- clearly. AlltheseSNRsareidentifiedinthe4.5and5.8 ters were adopted from Green’s catalog, and the sizes µm bands except G54.1+0.3, which is identified only were chosen to cover areas larger than their respective in the 8.0 µm band. G11.2–0.3 and 3C 396 are also radio continuum extents: 12′ for SNRs smaller than visible inthe 8.0µmband. G298.6–0.0hasabrightfil- 8′, 30′ for those between 9′ and 20′, and 60′ for those amentvisibleinallfourbands;however,itsassociation larger than 21′. is questionable. The identification of SNRs requires careful inspec- The GLIMPSE images show that the infraredemis- tionbecausetheyusuallyappearasfaintfilamentsem- sion of SNRs usually appear as partial filaments or bedded in a complex field with confusing sources such almost complete thin shells along the SNR boundary. as H II regions and planetary nebulae. For each SNR, Kes 17,G311.5–0.3,and RCW 103 aregood examples. we examined all four band images simultaneously to All marginally-identified SNRs except G54.1+0.3 be- search for infrared features coincident with the radio long to this category. In W 44 and W 49B, the fila- continuum emission. Two radio continuum data sets ments are also seen inside the remnants, which might were used for the comparison, i.e., NRAO∗ VLA Sky be located on the shell in a projected manner. The Survey(Condonetal. 1998)forthenorthernSNRsand partial filaments are found typically at the positions of enhanced radio emission and are aligned along the ∗The National Radio Astronomy Observatory is a facility of the National Science Foundation operated under cooperative agree- †TheMOSTisoperatedbytheUniversityofSydneywithsupport mentbyAssociatedUniversities,Inc. from the Australian Research Council and the Science Founda- tionforPhysicswithintheUniversityofSydney. Table1. SNRsIdentifiedintheSpitzerGLIMPSEfield I N F Name RAa Deca Sizea Typea IRASFluxb (Jy) IRACMorphology R (hms) (◦ ′) (′×′) 12 25 60 100 A R DistinctlyidentifiedSNRs E G31.9+0.0(3C391) 184925 −0055 7×5 S 3 14 180 160 Thickshellbroken-outtowardsoutheast D (2) (7) (90) (80) G34.7–0.4(W44) 185600 +0122 35×27 C 410 440 4400 13000 Brightthinfilamentsinsidetheentireremnant S U (100) (110) (1100) (6500) P G43.3–0.2(W49B) 191108 +0906 4×3 S 15 140 1100 2000 Brightthinfilamentsinsidetheentireremnant E (3) (28) (220) (400) R G304.6+0.1(Kes17) 130559 −6242 8 S 17 24 160 440 Brightthinfilamentsalongthewesternshell N (12) (17) (110) (310) O G311.5–0.3 140538 −6158 5 S <4 <12 <190 <310 Almostcompletethinfilamentaryring V A G332.4–0.4(RCW103) 161733 −5102 10 S <200 <260 <3000 <5400 Brightthinfilamentalongthesouthernshell R E G348.5+0.1(CTB37A) 171406 −3832 15 S <150 <250 <2500 <8000 Thickshellbroken-outtowardsouthwest M N G349.7+0.2 171759 −3726 2.5×2 S 12 48 370 <550 Verybrightfilamentwithbrightspotatcenter A (4) (16) (120) N T MarginallyidentifiedSNRs S G11.2–0.3 181127 −1925 4 C 44 189 1400 3100 Thinpartialfilamentattheradio-brightportion I (4) (19) (140) (310) ofsoutheasternshell N G21.8–0.6(Kes69) 183245 −1008 20 S 106 121 1370 6300 Thinfilamentatthemiddleofradioshell T (16) (18) (210) (950) H G39.2–0.3(3C396) 190408 +0528 8×6 C <56 <82 <530 <960 Filamentaryemissionspreadoverthewesternpart E ofremnant S G41.1–0.3(3C397) 190734 +0708 4.5×2.5 S <12 <23 <224 <485 Veryfaintfilamentattheradioenhancedportion P ofnortheasternshell I T G54.1+0.3 193031 +1852 1.5 F? <21 18 26 <600 Diffuseemissionatwesternboundarywithextended Z (6) (9) tailtowardnorth E G298.6–0.0 121341 −6237 12×9 S <110 <230 <1940 <3650 Brightbutsuspectedfilamentalongsouthernshell R G G340.6+0.3 164741 −4434 6 S <23 <32 <370 <1100 Thinfilamentatsouthoverlappedwithunrelated L confusingemission I M G346.6–0.2 171019 −4011 8 S <160 <180 <1500 <5900 Thinfilamentatsouthoverlappedwithunrelated P confusingemission S E aReferredfromtheGreen’scatalog(2004). F bThefirstrowisthemeasuredfluxandthesecondrowinparenthesesistheuncertaintyofmeasuringflux(Arendt1989). I E L D 3 8 7 388 LEE radio shell. However, the coincidence decreases for the other large-scale IR survey is the Two Micron All Sky detailedstructures. Forexample,theradio-brightwest- Survey (2MASS), which surveyed the whole sky at J ernshellhasno infraredcounterpartinW 44. Inaddi- (1.25 µm), H (1.65 µm), and Ks (2.17 µm). No sys- tion, with regard to the fields under investigation, all tematic study of SNRs using the 2MASS data exists; the radio-bright SNRs do not have associated infrared however, four SNRs (Crab, IC 443, Cas A, and RCW emission. Two SNRs, - 3C 391 and CTB 37A - have 103) have been reported as visible. Only RCW 103 rather thick shells, and they appear to have broken- exists in the GLIMPSE field. We checked 2MASS im- outshellmorphologylikeradiofeatures. G349.7+0.2is ages at the position of the other GLIMPSE SNRs and uniqueonaccountofitsverybrightinfraredemissionin were barely able to trace the emission associated with the centralareaofthe remnant. This remnantis inter- W 49B and W 44. actingwithamolecularcloudatthecenter(Lazendicet In general,the infraredemission from SNRs is com- al. 2005),andtheinfrared-brightregioncoincideswith posed of a thermal continuum from shock-heated dust the position of the molecular cloud. For each SNR, we grains, line emission from shock-heated gas, and syn- give a brief description of their infrared features in the chrotron emission. Among the 16 GLIMPSE SNRs, Appendix. only one is Crab-like (or filled-center) and three are The infrared emission of SNRs are better identified composite with a centralpulsar wind nebulae (PWNe) in the 4.5 and 5.8 µm bands than in the other two (see Table 1). We have not detected any IR emis- bands. In the 3.6 µm band, the infraredemission from sion corresponding to these synchrotron nebulae. Ex- the SNRs is weak. Concurrently, the emission from cept G54.1+0.3, all the 16 SNRs show shell-like or fil- stars makes it difficult to identify the SNR emission. amentary IR emission. If we consider the steep slope Ontheotherhand,the8.0µmbandisseverelyaffected of the synchrotron emission and the low sensitivity of by Polycyclic Aromatic Hydrocarbon (PAH) emission the GLIMPSE data, it is highly unlikely that the syn- at 6.2, 7.7, and 8.6 µm, which is dominant in H II re- chrotron emission contributes to the observed IR flux gions,planetarynebulae,andotherdiffusePhotodisso- evenintheshortestIRACwaveband(3.6µm). Onthe ciation Regions (PDRs) (van Dishoeck 2004). There- otherhand,previousISOspectroscopicstudiesofSNRs fore, the emission from SNRs is easily confused with showedthationiclines[FeII]5.340µmand[ArII]6.985 those fromothersources. Hence,mostofthe identified µmintheIRACbandsareparticularlystronginradia- SNRs show blue (4.5 µm) or green filaments (5.8 µm) tiveshocks(Arendt1999;Olivaetal. 1999;Douvionet in the RGB images. al. 1999, 2001a, 2001b; Reach et al. 2002). For SNRs interacting with molecular clouds, e.g., 3C 391 and IC IV. DISCUSSION AND SUMMARY 443, shock-excited H2 lines are also found to be strong (Cesarsky et al. 1999; Reach & Rho 2000; Reach et Wehaveidentified16(8distinctlyand8marginally) al. 2002). Therefore the ionic lines produced by the infrared SNRs in the GLIMPSE field. This is 16% of atomic J-type shock (Hartigan et al. 2004)and/or the the100radio/X-raySNRsinthefield. Forcomparison, H linesproducedbythemolecularC-typeshock(Hol- 2 previousIRASstudiesdetectedinfraredemissionsfrom lenbach et al. 1989) can be the main contributors in 22SNRsoutofthe70(Arendt1989)or18SNRsoutof IRAC images. the 75(Sakenetal. 1992),whichwereidentifiedinthe Independent of this work, there exists a study on GLIMPSE field at the time of the respective studies. the search for SNRs in the GLIMPSE field (Reach et Among the 16 GLIMPSE SNRs, 9 SNRs are detected al. 2005). They also detected the associated infrared by IRAS. IRAS fluxes or limits are listed in columns emission by visually inspecting the IRAC images with (6)–(9) of Table 1. A study was conducted to search an overlaidradio map. 18 SNRs were reportedto have for the infrared emission associated with SNRs from anassociatedinfraredemission(score1). Alleightdis- the Midcourse Space Experiment (MSX) data (Kim et tinct SNRs, as well as the five of the eight marginally al. 2005). MSX surveyed the whole Galactic plane detected SNRs were included in their list. The others (|b|<5◦) in four bands centered at 8.28, 12.13, 14.65, (G54.1+0.3,G298.6–0.0,G340.6+0.3)wereregardedas and 21.34 µm. Kim et al. (2005) generated and ex- SNRs that were not convincingly detected due to the amined the MSX images of over 200 SNRs in the field confusion with unrelated emissions. For the detected and identified 8 SNRs, five of which are included in SNRs, the originofthe infraredemission,e.g., molecu- the GLIMPSE field. These five SNRs, - G11.2–0.3, W lar or ionic shocks or the PAH emission, was discussed 44, W 49B, G54.1+0.3, and G349.7+0.2 - are iden- using the relative intensities between the IRAC bands. tified in this work. The higher detection rate of the The GLIMPSE images of SNRs presented in this GLIMPSE survey is due to its higher sensitivity and paper are the first high-resolution infrared images for angular resolution. The 0.4 mJy sensitivity of the 8 most SNRs. They show that SNRs appear as almost µm band in the GLIMPSE field is seven times better complete shells, partial filaments, or a combination of than the 30 mJy sensitivity of the 8.28 µm band of filaments. The infrared emission generally coincides MSX (Benjamin et al. 2003). Furthermore, the 18′′.3 with the radiocontinuum,e.g.,the infraredemissionis resolution of MSX increases the chance to dilute thin identified with the radio-brightregions,although some filaments and confuse them with other sources. An- displacement is a feature of the detailed distribution. INFRARED SUPERNOVA REMNANTS IN THE SPITZER GLIMPSE FIELD 389 Mostarerelativelywellidentifiedinthe4.5and5.8µm Green, D. A., Gull, S. F., Tan, S. M., & Simon, A. J. B., bands. The dominant contributionmay originate from 1988, MNRAS,231, 735 the line emission of shocked ambient gas. We expect Green,D.A.,2004, BulletinoftheAstronomicalSocietyof that the infrared imaging and spectroscopic observa- India,32, 335–370. tions using the SST and the future ASTRO-F mission Hartigan,P.,Raymond,J.,&Pierson,R.,2004,ApJL,614, will extend the understanding of the nature of the in- L69 frared emission in individual sources. Hollenbach D. J., Chernoff, D. F., & McKee, C. F., 1989, InInfraredSpectroscopy inAstronomy,ESASP-290,pp. 245–58, Noordwijk, ESTEC ACKNOWLEDGEMENTS Kim et al., 2005, in preparation This work is based on archival data obtained with Lazendic, J. S., Slane, P. O., Hughes, J. P., Chen, Y., & the Spitzer Space Telescope, which is operated by the Dame, T. M., 2005, ApJ, 618, 733 JetPropulsionLaboratory,CaliforniaInstituteofTech- Moffett, D.A. & Reynolds, S.P., 1994, ApJ, 425, 668 nology under a contract with NASA. I wish to thank theSpitzerGLIMPSEteamfortheirefforts. Thiswork Oliva, E., Moorwood, A. F. M., Drapatz, S., Lutz, D., & was supported by the Korea Science and Engineering Sturm,E., 1999, A&A,343, 943 Foundation Grant ABRL 3345-20031017. I wish to Reach, W. T. & Rho,J., 1999, ApJ, 511, 836 thank the Institute ofSpace andAstronauticalScience Reach, W. T. & Rho,J., 2000, ApJ, 544, 843 (ISAS) for providing their facilities during part of this work. I wish to thank Bon-Chul Koo for extensive ad- Reach, W. T., Rho, J., & Jarrett, T. H., 2002, ApJ, 564, 302 vice on the entire study. I also wish to thank David Moffett, Stephen Reynolds, Kristy Dyer, David Green, Reach, W. T. et al., 2005, AJ, in press JohnDickel,andCrystalBroganforprividingthehigh- Saken,J. M., Fesen,R.A.,& Shull,J.M., 1992, ApJS,81, resolution radio continuum images, and Chris Pearson 715 for the advice to improve the presentation. van Dishoeck, E. F., 2004, ARAA,42, 119 Whiteoak, J. B. Z. & Green, A. J., 1996, A&AS, 118, 329 REFERENCES (on-lineversionathttp://www.physics.usyd.edu.au/ast- rop/wg96cat) Arendt,R.G., 1989, ApJS, 70, 181 Arendt,R.G., Dwek,E., &Moseley, S.H.1999, ApJ,521, 234 Benjamin, R.A., et al., 2003, PASP,115, 953 Brogan,C.L.,Frail,D.A.,Goss,W.M.,&Troland,T.H., APPENDIX. IDENTIFIED SNRS 2000, ApJ,537, 875 Cesarsky,D.,Cox,P.,PineaudesForˆets,G.,vanDishoeck, E. F., Boulanger, F., &Wright, C.M., 1999, A&A,348, G11.2–0.3– A faintwispy filamentis seenalong the 945 SE SNR boundary at 4.5, 5.8, and 8.0 µm. The ∼20′′- long filament centered at (18h11m31s.7,−19◦27′14′′) Condon, J. J., Cotton, W. D., Greisen, E. W., Yin, Q. F., is particularly noticeable. The bright emission at Perley, R. A., Taylor, G. B., & Broderick, J. J., 1998, AJ, 115, 1693 (18h11m19s.3,−19◦26′19′′) is not associated with the remnant but with a K1/K2 III star (HD 166422). Dickel, J. R., Green, A., Ye, T., & Milne, D. K., 1996, AJ, 111, 340 G21.8–0.6(Kes 69) – A faint ∼6′-long filament cen- Douvion,T., Lagage, P.O.,&Cesarsky,C.J.,1999, A&A, tered at (18h33m05s.9,−10◦12′′53′) is seen along the 352, L111 incomplete SNR shell at 4.5 and 5.8 µm. There is a Douvion,T.,Lagage,P.O.,&Pantin,E.,2001a,A&A,369, complex, extended emission at 5.8 and 8.0 µm in the 589 NE boundary of the field. The SE boundary of this Douvion, T., Lagage, P. O., Cesarsky, C. J., & Dwek, E., structure appears to form a concave filament that is 2001b, A&A,373, 281 roughlyparallelwiththeSNRradioshell. Therelation Draine, B. T., & Mckee, C. F., 1993, ARAA,31, 373 of this emission to the remnant needs to be explored. Dwek, E., 1987, ApJ,322, 812 G31.9+0.0(3C391)–Abrightincompleteshellcoin- Dwek, E., & Arendt,R. G., 1992, ARAA,30, 11 cidentwiththeradiostructureisseen. TheNWpartof Dyer,K. K.& Reynolds,S. P., 1999a, ApJ, 526, 365 theremnantisbrightprimarilyat5.8µm,whiletheNE Dyer,K. K.& Reynolds,S. P., 1999b, BAAS,31, 9735 and SW parts are bright at 4.5, 5.8, and 8.0 µm; this Fazio, G. G. et al., 2004, ApJS, 154, 10 explains their different colors in the true-color image. The bright region at (18h49m23s.3,−00◦57′49′′) shows 390 LEE a regionwhere the remnant interacts with a molecular clump (Reach & Rho 1999). G332.4–0.4 (RCW 103) – A sharp filament corre- sponding to the southernpartofthe remnantis visible G34.7+0.4 (W44) – A bright filamentary elliptical in all the four bands. It is particularly bright at 4.5 shell coincident with the radio shell is seen in all the µm. The bright portion of the filament is at the po- four bands. It is most clearly seen at 4.5 µm, which sition where the radio brightness has a sharp cutoff in yieldsablue colorinthe true-colorimage. Theeastern contrasttotheotherpartsoftheremnant. Thereisalso edgeoftheinfraredshellismissingalthoughitisbright a faint infrared filament corresponding to the western in the radio image. radio-bright region. G39.2–0.3 (3C396) – A faint ∼2′-long filament is G340.6+0.3 – A complex filamentary structure su- seen along the SW SNR boundary where the radio perposedonthenorthernpartoftheremnantisvisible emission is enhanced. It is most clearly visible at 4.5 at 5.8 and 8.0 µm. However, it is not correlated with µm. Theemissionat(19h03m56s.2,+05◦25′43′′)isalso the radio structure of the remnant, and its association be identified in other bands. needs to be explored. Some diffuse emissioncoincident with the southern SNR shell also exists. G41.1–0.3(3C397)– Very faint filaments coincident with the NE SNR boundary are visible at4.5,5.8, and G346.6–0.2 – A very faint thin filament delineating 8.0 µm. The identified filaments are partly affected by the SE boundary of the SNR shell is visible at 4.5 and the emission from a bright confusing star located at 5.8 µm. There exists some faint emission coincident (19h07m33s.6,+07◦09′24′′). with the northern SNR shell; however, its association isnotclearbecauseoftheconfusingextendedstructure G43.3–0.2(W49B)–Abrightfilamentaryshellmatch- to the west of the remnant. ing the radio structure is seen in all the four bands. The SW portion of the shell is very bright as it is in G348.5+0.1–Alarge(∼10′)incompleteshellisseen theradio. Theinterioroftheremnantisfilledwithdif- mainly at 5.8 and 8.0 µm. The remnant is composed fuseemissionincludingseveralbrightfilamentsrunning of a small, bright NE portion and a large, diffuse SW along the NE-SW direction. A bow-shaped structure portion in the radio, and the infrared shell coincides brightat8.0 µm exists just outside the easternbound- with the SW portion. The infraredshell is brighter to- aryoftheremnant;itsassociationneedstobeexplored. ward the NE where the radio structure starts to blow out. Theremnantappearstobeconnectedtoadiffuse, G54.1+0.3 – A partially complete, circular shell is extended structure at the NE. seen mainly at 8.0 µm. The westernhalf of the shell is bright and thick, whereas the eastern half is faint and G349.7+0.2 – Bright filamentary emission is visible fuzzy. The remnant is Crab-like in the radio and the in the interior of the remnant in all the four bands. A shell appears to surround the remnant. In addition, a verybright∼40′′-long,”v”-shapedfilamentatthecen- diffuse, extended emission emanate from the north of ter and a bright ∼45′′-long filament perpendicular to the remnant. theformerinthesouthareprominent. Thesefilaments coincidewiththepositionofthemolecularcloudsswept G298.6–0.0 – A ∼5′-long filament coincident with up by the SNR shock (Lazendic et al. 2005). Another the southern, radio-bright SNR shell is seen in all the bright filament stemming out from the eastern end of four bands. However,the remnantis located in a com- the ”v”-shaped filament is visible along the NE SNR plex field, and the possibility of a chance coincidence shell. A long filamentary emission is connected to the cannot be ruled out. southof the remnantat5.8and8.0µm. This emission stretches out ∼3′ toward the east. G304.6+0.1(Kes 17) – Bright filaments are distinct inallthefourbands. Theyaredistributedinthewest- ern part of the remnant along its boundary. At 5.8 and 8.0 µm, the filaments appear to be connected to the edge of a large cloud in the NE by a diffuse, ex- tended emission. No infrared counterpart is detected at the south of the remnant where the radio emission is stronger than the western shell. G311.5–0.3 – An almost complete thin shell with a diameter of ∼ 3′ is seen in all the four bands. It is brightest at 4.5 and 5.8 µm. The western part of the shell is brighter than the eastern part. INFRARED SUPERNOVA REMNANTS IN THE SPITZER GLIMPSE FIELD 391 Fig. 1.— IRAC images ofG31.9+0.0(3C 391)at3.6 µm (top-left), 4.5µm (top-right),5.8 µm (middle-left), and 8.0 µm (middle-right). The gray-scale range is shown at the left of each image. VLA 1.46 GHz continuum image with 5′′.0×4′′.6 beam (Moffett & Reynolds 1994)is alsoshownat the bottom-left. The contourlevels are 1, 2.5,5, 7.5, 10, 20, and 40 mJy beam−1. 392 LEE Fig. 2.— IRAC images of G34.7–0.4 (W 44) at 3.6 µm (top-left), 4.5 µm (top-right), 5.8 µm (middle-left), and 8.0 µm (middle-right). The gray-scale range is shown at the left of each image. VLA 1.44 GHz continuum image with30′′.0×30′′.0beamisalsoshownatthebottom-left. Thecontourlevelsare50,100,150,and200mJybeam−1. INFRARED SUPERNOVA REMNANTS IN THE SPITZER GLIMPSE FIELD 393 Fig. 3.— IRAC images of G43.3–0.2(W 49B)at 3.6 µm (top-left), 4.5 µm (top-right), 5.8 µm (middle-left), and 8.0 µm (middle-right). The gray-scale range is shown at the left of each image. VLA 1.45 GHz continuum image with 5′′.2×4′′.8 beam (Moffett & Reynolds 1994) is also shown at the bottom-left. The contour levels are 3, 5, 10, 20, 30, 40, and 50 mJy beam−1. 394 LEE Fig. 4.— IRAC images of G304.6+0.1 (Kes 17) at 3.6 µm (top-left), 4.5 µm (top-right), 5.8 µm (middle-left), and 8.0 µm (middle-right). The gray-scalerange is shown at the left of each image. MOST 0.843 GHz continuum image with 43′′×43′′ beam is also shown at the bottom-left. The contour levels are 5, 10, 20, 40, 80, 160, and 320 mJy beam−1.