SNR G349.7+0.2: A gamma-ray source in the far 3 kpc arm of the Galactic center
aa r X i v : . [ a s t r o - ph . GA ] F e b Draft version July 8, 2018
Preprint typeset using L A TEX style emulateapj v. 5/2/11
SNR G349.7+0.2: A γ -RAY SOURCE IN THE FAR 3 KPC ARM OF THE GALACTIC CENTER W.W. Tian , D.A. Leahy Draft version July 8, 2018
ABSTRACTWe analyze the HI absorption profile for TeV Supernova Remnant (SNR) G349.7+0.2 based onupdated knowledge of the inner Galaxy’s structure. We significantly revise its kinematic distancefrom the previous ∼
22 kpc to ∼ ∼ ∼ × ergs. This removes G349.7+0.2 from the set of brightest SNRs in radio and X-ray to γ -ray wavebandsand helps us to better understand γ -ray emission originating from this remnant. However, one needsto use caution when discussing old kinematic distances of Galactic objects (e.g., SNRs, pulsars andHII regions) in the range of -12 ≤ l ≤ with distance estimates of ≥ Subject headings:
ISM:supernova remnants - ISM:lines and bands - Galaxy:center - cosmic rays INTRODUCTION AND DATAIt is widely believed that most Galactic cosmicrays probably originate from supernova remnants(SNRs). Recently the TeV γ γ -ray sources which were likely identified asSNRs/PWNe, and showed that most of them are lessthan 6 kpc away.Many SNRs’ kinematic distances are obtained by ob-serving radial velocities of atomic and molecular linestoward the SNRs and using a circular rotation curvemodel. A circular rotation curve model is generally notsuitable for Galactic center (GC) objects because thegas likely follows oval orbits in the inner Galaxy, i.e.inside the 3 kpc ring. Many studies have shown thatit is extremely difficult to determine precise distancesalong the line-of-sight to objects in the 3 kpc ring. Inthis case, previous kinematic distance estimates to someremnants might contain a large uncertainty or even beincorrect. Recently-finished radio and infrared surveysreveal detailed kinematic images of gas and dust in GCregion (Dame & Thaddeus 2008, McClure-Griffiths et al.2012). Several good gas flow models in the GC have beenpresented based on recent observations (e.g. Rodriguez-Fernandez & Combes 2008).In this letter, we study a newly detected TeV source:SNR G349.7+0.2 (Trichard et al. 2013), which has longbeen believed to be at a large distance of ∼
22 kpcaway, meaning that it resides on the other side of theMilky Way, based on observations of HI, 1720 MHz OHmasers, CO and H CO etc. (Caswell et al. 1975, Frail National Astronomical Observatories, CAS, Beijing 100012,China. [email protected] Department of Physics & Astronomy, University of Calgary,Calgary, Alberta T2N 1N4, Canada et al. 1996, Reynoso & Mangum 2001). A high reso-lution radio continuum image of G349.7+0.2 at 18 cmwas presented previously by Lazendic et al. (2010). Weconstruct an HI absorption line profile and utilize newstudies of the inner Galactic structure in order to con-strain the distance and other key physical parameters ofSNR G349.7+0.2. The 1420 MHz radio continuum and21 cm HI emission data sets are from the Southern Galac-tic Plane Survey (SGPS),using the Australia TelescopeCompact Array and the Parkes 64m single dish telescope(Haverkorn et al. 2006). The continuum observationshave a resolution of 100 arcsec and a sensitivity betterthan 1 mJy/beam. The HI data have an angular res-olution of 2 arcmin, an rms sensitivity of ∼ − . The CO ( J =1-0) datafor G349.7+0.2 is from survey data from the 12 m NRAOtelescope (see Reynoso & Mangum 2000 for details). ANALYSIS AND RESULTS2.1.
Spectra
Methods for extracting HI absorption spectrum havebeen introduced in earlier papers (Tian et al. 2007;Leahy & Tian 2008; Leahy & Tian 2010). The HI emis-sion spectrum of SNR G349.7+0.2 is shown in the upperpanel of Fig. 1. IN this spectrum, there are several HIemission peaks from -113 km s − to 35 km s − . TheHI absorption and CO emission spectra are shown inthe lower panel of Fig. 1. SNR G349.7+0.2 is strongradio source, which results in a strong HI absorption sig-nal. An estimate of the error in exp(- τ ( v )) of Fig. 1 isthe r.m.s. for velocity channels with no emission ( > +50km/s), which is 0.0063. Artifacts in the continuum im-age cause another type of error. We have analyzed ab-sorption spectra both with the same source region andwith different backgrounds. We find that the artifactsonly affect the normalization and not the shape of theHI absorption spectra. We find that the error in normal-ization is about 20% in depth of the HI absorption. Wefind that each CO emission peak has an associated HIabsorption feature except for the 16.5 km s − CO peak.The highest positive velocity absorption feature is at 6 ± − . The lowest negative velocity absorption featureis likely at -110 ±
10 km s − (a weak absorption feature Tian & Leahy -1000100 T B ( K ) G349.7+0.2background-200 -150 -100 -50 0 50 100Velocity (km/s)00.51
CO emission T B (K)/4HI: exp(-Tau) Fig. 1.—
HI and CO spectra of G349.7+0.2 appears at -189 ± − which is real, see section 3.1).We calculate the absorption column density from the HIabsorption spectrum, N HI ∼ × cm − , takinga value of T s = 100 K (using N HI = 1.9 × R τ d v T s cm − , Dickey & Lockman 1990).2.2. HI Channel maps
We use the HI channel maps to confirm the reality ofall above-mentioned absorption features, including the189 km s − absorption feature. Absorption continues toappear from the channel map at velocity ∼ -123 km s − sthrough the map at velocity ∼
13 km s − ;. Fig. 2 showsthe HI channel maps at 8.3 km s − and 16.5 km s − towards the SNR. The left panel of Fig. 2 shows clearabsorption against the SNR. The right one reveals thatthe remnant sits on a bright HI patch which is associatedwith the 16.5 km s − CO cloud. This is consistent withabove absorption spectrum analysis, i.e. no HI absorp-tion associated with the HI and CO clouds at 16.5 kms − . This directly confirms that the 16.5 km s − cloudsare behind the remnant. DISCUSSION AND CONCLUSIONHI absorption spectra towards Galactic radio-brightobjects have been used to estimate the line-of-sight dis-tances based on a rotation curve model. However, it isa challenging job to constrain the distances to Galacticobjects in the inner 3 kpc of the Galaxy because the ob-jects likely follow non-circular orbits in the region (seeWeiner & Sellwood 1999).The situation has improved due to recent studies of theGalactic center’s structure. High resolution HI observa-tions (McClure-Griffiths et al. 2012) of the inner Galaxyshow greater details of the region’s structure. CO obser-vations (Dame & Thaddeus 2008) have detected the sym-metric expanding 3 kpc arms, supporting the existence ofa bar at the center of our Galaxy. Rodriguez-Fernandez & Combes (2008) iroveprovedd a gas distribution modelfor the GC region by simulating the gas dynamics in theinner Galaxy. This model gives us an opportunity tobetter understand the HI and CO spectra and refine thedistance estimation of G349.7+0.2. We adopt the 1985IAU standard for Galactic parameters in this paper, i.e.R o = 8.5 kpc (the distance between Sun and the GC)and V o = 220 km/s (circular velocity at R o =).3.1. Distance of G349.7+0.2
The HI terminal velocity at l =349 to 350 is about -190 km s − (see Fig. 2 of Weiner & Sellwood 1999 and ofBurton & Liszt 1993). We detect a weak absorption fea-ture at -189 km s − in the spectrum of SNR G349.7+0.2,which is likely caused by clumps of HI possibly acceler-ated by a SNR or HII region along this line-of-sight. Thelowest-velocity major absorption feature, except for the-189 km s − feature, is at ∼ -110 km s − in the SNRspectrum (Fig. 2). This velocity is consistent with theexpected radial velocity (101.6 ± − ) of the near3 kpc arm in the line-of-sight to l =349 to 350 (equa-tion (3) of Jones et al. 2013, see also Fig. 1 of Dame& Thaddeus 2008). This is convincing evidence that theremnant is beyond the near 3 kpc arm.In addition, the highest-velocity absorption in thespectrum of G349.7+0.2 is at 6 ± − . This isconsistent with the expected radial velocity (16.3 ± − ) of the far 3 kpc arm toward l ∼ -10.3 (Fig. 1 andEquation (4) of Jones et al. 2013). However, in the SNRspectrum we found no HI absorption associated with theHI and CO emission clouds at 16.5 km s − , which ex-actly lie in the far 3 kpc arm extending at least 20 ◦ inlongitude, starting at l=-12 ◦ (Dame & Thaddeus 2008).These two pieces of evidence argue that the remnant isin the far 3 kpc arm; the cold HI gas at 6 ± − is in front of the remnant, and the HI and CO clouds at16.5 km s − are behind the remnant (see Fig.3).SNR G349.7+0.2 is known to be interacting withnearby molecular clouds because of clear evidence: thereare five 1720 MHz OH masers along the radio emissionridge of the SNR (Frail et al. 1996); shock-excited near-infrared H emission has been found toward the centerof the remnant as well as OH absorption (1665 and 1667MHz) in the remnant (Lazendic et al. 2010), and COexpansion caused by the SNR shock has been observed(Reynoso & Mangum 2001). Because most 1720 MHzOH masers are seen as signposts of SNR-molecular cloudinteraction (Wardle & Yusef-Zadeh 2002), their velocitieshave been used to determine kinematic distances to asso-ciated SNRs based on a rotation curve model. The veloci-ties of the 1720 MHz OH masers (14.3 to 16.9 km s − ) arenicely consistent with the velocity of shocked CO clouds(16.5 km s − ), also with the far 3 kpc arm velocity of ∼ − . Combined with the HI absorption analysis,we conclude that the 16.5 km s − CO clouds are behindbut shocked by SNR G349.7+0.2 and this SNR-cloud in-teraction excites the near-infrared H emission and the1720 MHz OH maser emission in the region.An HI 21-cm absorption line study of G347.9+0.2 haspreviously been done using the Parkes hydrogen line in-terferometer (Caswell et al. 1975). Our absorption spec-trum confirms the previous major absorption featuresnear -110 km s − , -62.5 km s − , and 6 km s − . Butistance to SNR G349.7+0.2 3 Fig. 2.—
HI channel maps at velocities of 8.3 and 16.5 km/s. The contours at 0.2, 2, 6, 8 Jy/Beam are from the 1420 MHz continuumimage of SNR G349.7+0.2. -5 0 5-505
SunSolar CircleG349.7+0.2l=349.7Galactic CenterNear 3 kpc armFar 3 kpc arm
Fig. 3.—
Diagram indicating the location of SNR G349.7+0.2in the far 3 kpc arm of the Galactic center. The Galactic centerstructure is referenced in to Churchwell et al. (2009) we find more fine structure because of the better qualityof the SGPS data (higher continuum sensitivity and HIspectral resolution) and the improved methods. We findnew absorption features at -35 and -189 km s − , no ab-sorption at 20 and 38 km s − , much lower noise for base-lines above 20 km s − and below -125 km s − . Caswellet al. (1975) explained the -62.5 km s − feature as fromthe far ring of 3-4 kpc arm and the 6 km s − feature asfrom local gas, based on 1970s knowledge of the Galacticcenter, so they suggested a distance of 13.7 kpc ≤ d ≤ − to-wards G347.9+0.2 respectively, combined with our newresult of reliable absorption at 6 ± − . Thus wefind the SNR has no absorption associated with the 16.5km s − clouds in the far 3 kpc arm. In this case, the -62.5km s − feature is likely caused by much closer HI cloudswhose motion follows the normal rotation curve model.Our conclusion is strongly supported by other evidence:the HI absorption at -62.5 km/s is also clearly seen inabsorption spectra of the nearby SNRs G348.5+0.1 andG348.5-0.0, which both have distances of less than 9.5kpc (Tian & Leahy 2012); and the remnant is interact-ing with 16.5 km/s clouds in the far 3 kpc arm. Previ-ously, the 1720 MHz OH masers at 14.3 to 16.9 km s − were assumed to be outside of the Solar circle based onthe circular rotation curve model, so SNR G349.7+0.2was suggested to have a kinematic distance of ∼
22 kpc.Now we know that the circular rotation curve model isnot valid near the GC, so we conclude that the SNR andits nearby clouds are located in the far 3 kpc arm, wellwithin the Solar circle.SNR G349.7+0.2 is located at the near edge of the far3 kpc arm because there is no absorption from the mainpart of the far 3 kpc arm (which has center velocity 16.5km s − ). The far 3 kpc arm is at a distance of ∼ ∼ × cm − .This is consistent with the estimate from adding HI(2.8 × cm − , assuming T s =140 K) and H (4.6 × cm − , assuming CO-to-H conversion factor X = 1.8 × cm − K − km s − ) emission (Lazendic et al.2005). Our calculation of the column density againstG349.7+0.2 by HI absorption (using T s =100 K) is ∼ cm − ), similar to Lazendic et al. (2005). The HIcolumn density measures neutral hydrogen, whereas the Tian & LeahyX-ray column density is sensitive to the heavier elementcomponent in the total gas (HI, H and HII) plus dust.So the X-ray column density could easily be 3 or moretimes higher than the HI column density. In addition,it is generally known that the column density value isstrongly related with direction of a source, because theGalactic HI has a different distribution along differentlines of sight. For example, SNR G18.8+0.3 has a dis-tance of 12 kpc and column densities of 2 x 10 cm − from HI and 3.2 10 cm − from H ) (Tian et al. 2007).Generally, there are more clouds towards the GC thanother directions, so there is no inconsistency in columndensities with G349.7+0.2 being at 11.5 kpc.It should be noted that kinematic distances to Galac-tic objects always include uncertainty due to uncertainparameters of different rotation curve models and wide-spread non-circular streaming motions. This uncertaintymay reach up to 20% at the particular longitude of someobjects (Jones et al. 2013). SNR G349.7+0.2 is withinthe far 3 kpc arm: its distance’s uncertainty is from theR o value and the width of the arm. The IAU standard isR o = 8.5 kpc, but recent measurements show a smallervalue, e.g. 8.4 ± ± ± o is about 0.5 kpc. The width ofthe far 3kpc arm is not yet known but should be smallerthan the error of R o . We therefore estimate a distanceof 11.5 ± o ≤ l ≤ o and having distance estimates of ≥ Luminosity, age, density and explosion energy ofG349.7+0.2
G349.7+0.2 has been suggested as one of the brightestGalactic sources in the radio (Shaver et al. 1985), X-ray(Slane et al. 2002) and GeV γ -ray wavebands (Castro& Slane 2010) because of its large distance. We giveits luminosity based on the new distance measurementhere: L GHz ∼ . × Watt Hz − , L X (0 . − . keV ) ∼ . × erg s − , L γ (0 . − GeV ) ∼ × erg s − .The angular diameter of G349.7+0.2 ( ∼ ′ ) from Chan-dra (Lazendic et al. 2005), yields a radius of R = 3.3 pc(d = 11.5 kpc) for the SNR. Lazendic et al.(2005) finda shock velocity of ∼
710 km s − based on the X-raymeasured plasma temperature. The small remnant sizeand fast shock velocity indicate that the remnant is stillevolving in the Sedov phase. So we estimate its age of ∼ . Using the X-ray emission measure from the Chandraspectrum of 9.9 × cm − (Lazendic et al. 2005) and aSedov interior density profile (see Leahy et al. 2013), theISM density is n ∼
10 cm − (d = 11.5kpc). G349.7+0.2is then found to have a low explosion energy of 2.5 × ergs.We thank the anonymous referee for his/her construc-tive comments. WWT acknowledges support from theChina Ministry of Science and Technology, the NSFC,and the CAS. DAL s grateful for the Natural Sciencesand Engineering Research Council of Canada. This pub-lication is partly supported by a grant from the JohnTempleton Foundation and NAOCAS.ergs.We thank the anonymous referee for his/her construc-tive comments. WWT acknowledges support from theChina Ministry of Science and Technology, the NSFC,and the CAS. DAL s grateful for the Natural Sciencesand Engineering Research Council of Canada. This pub-lication is partly supported by a grant from the JohnTempleton Foundation and NAOCAS.