Astronomy & Astrophysics manuscript no. (will be inserted by hand later) − First optical light from the supernova remnant G 17.4 2.3 1 1 2 P. Boumis , F. Mavromatakis , and E. V. Paleologou 2 0 1 University of Crete, Physics Department,P.O. Box 2208, 710 03 Heraklion, Crete, Greece 0 2 Foundation for Research and Technology-Hellas, P.O. Box 1527, 711 10 Heraklion, Crete, Greece 2 n Received 16 November2001 / Accepted 29 January 2002 a J Abstract. Deep optical CCD images of the supernova remnant G 17.4–2.3 were obtained and faint emission has 9 beendiscovered.Theimages,takenintheemissionlinesofHα+[Nii],[Sii]and[Oiii],revealfilamentarystructures 2 intheeast,south–eastarea,whilediffuseemissioninthesouthandcentralregionsoftheremnantisalsopresent. The radio emission in the same area is found to be well correlated with the brightest optical filament. The flux 1 calibratedimagessuggestthattheopticalfilamentaryemissionoriginatesfromshock-heatedgas([Sii]/Hα>0.4), v while the diffuse emission seems to originate from an Hii region ([S ii]/Hα < 0.3). Furthermore, deep long–slit 2 8 spectra were taken at the bright [Oiii]filament and clearly show that the emission originates from shock heated 4 gas. The [Oiii] flux suggests shock velocities into the interstellar “clouds” greater than 100 km s−1, while the 1 [Sii]λ6716/6731 ratioindicateselectron densities∼240cm−3.Finally,theHαemission hasbeenmeasuredtobe 0 between 7 to 20 × 10−17 erg s−1 cm−2 arcsec−2. 2 0 Key words.ISM: general – ISM: supernovaremnants – ISM: individualobjects: G 17.4-2.3 / h p - 1. Introduction an incomplete radio shell, characterized by diffuse shell– o r like emission,anangularsize of∼24′ anda radio spectral st TheGalacticsupernovaremnants(SNRs)havebeeniden- index of ∼0.8 (Green 2001). Case & Bhattacharya(1998) a tified by both radio (non-thermal synchrotron emission) calculatedits surfacebrightnessto be 1.3×10−21 W m−2 v: and optical(optical emissionlines) surveys.New searches Hz−1 m−1.Because,thereisnodirectdistancedetermina- in both wavebands continue to identify galactic SNRs i tion, they have made an estimation by utilizing the radio X (Fesen&Hurford1995,Fesenetal.1997;Green2001and surface brightness – diameter relationship (Σ − D) and r references therein; Mavromatakis et al. 2000, 2001, 2002) a found a distance of 8.5 kpc, but still the uncertainties but since the last few years, observations in X-rays have are large (∼40%). Green et al. (1997), through their sur- also detected new SNRs (e.g. Seward et al. 1995). The vey with the Parkes 64 m telescope, detected maser OH ratio of [S ii]/Hα has become the standard discriminator (1720MHz)emission.Inradiosurveysofthe surrounding usedinopticalSNRobservationsbecausethephotoionized region,nopulsarwasfoundtobe associatedwithG17.4– nebulae (like Hii regions and planetary nebulae) usually 2.3 but another SNR has been discovered in its neigh- exhibit ratios of about 0.1-0.3, while collisionally ionized borhood. G 17.8−2.6 has a very well defined shell, it lies nebulae (like knownGalactic SNRs) show ratiostypically ′ about 30 north–east of G 17.4–2.3 and has an angular greater than 0.4 (Smith et al. 1993). Fesen et al. (1985) ′ diameter of 24 (Reich et al. 1988). Neither of these rem- suggestedthat a division at [S ii]/Hα < 0.5 does not pro- nants has been detected optically in the past and from videclearevidencetodistinguishSNRsfromphotoionized our observations no optical emission has been found in regions and additional observations of the strong forbid- G 17.8–2.6. On the other hand, X–ray emission was not den oxygen lines ([Oi], [Oii] and [Oiii]) are needed to detected from G 17.4–2.3 in the ROSAT All–sky survey, giveacompletediagnostic.Furthermore,theoreticalshock while there is some evidence of X–ray emission from the models , generally predicted [S ii]/Hα ratios of 0.5 to 1.0 neighboring SNR G 17.8–2.6. for SNRs (Raymond 1979, Shull & McKee 1979). G17.4−2.3isnotawellknownSNR,andwasfirstde- In this paper, we report the discovery of faint opti- cal filaments from G 17.4–2.3. We present Hα + [Nii], tected by Reich et al. (1988) in their Effelsberg 2.7–GHz [Sii] and [Oiii] images which show filamentary structure survey,whileitsradioimagewaspublishedbyReichetal. along the south–east edge of the remnant correlated very (1990).Itisclassifiedasacircularsupernovaremnantwith wellwiththeradioemission.Spectrophotometricobserva- Send offprint requests to: P. Boumis, e-mail: tions of the brightest filament were also obtained and the [email protected] emission lines were measured. In Sect.2, we present infor- 2 P. Boumis et al.: First optical light from thesupernovaremnant G 17.4−2.3 mations concerning the observations and data reduction, Table 1. Typically measured fluxes in the area of the while the results ofourimagingandspectralobservations brightest filament are givenin Sect. 3 and 4,respectively.In the lastsection (Sect. 5) we discuss the physicalproperties ofG 17.4–2.3. N1 SW1 Center1 Hii 2 Hα+[Nii] 116.0 111.2 108.0 99.7 [Sii] 16.3 22.1 17.3 8.6 2. Observations [Oiii] 13.0 4.7 22.1 – 2.1. Imagery Fluxesin unitsof 10−17 erg s−1 cm−2 arcsec−2 The observationspresented here were performed with the 1Median values overa 36′′× 21′′ box 0.3 m Schmidt-Cassegrain (f/3.2) telescope at Skinakas 2Median values overa 60′′× 60′′ box Observatory in Crete, Greece in August 20 and 21, 2001. The 1024×1024(with 19×19µm2 pixels) Tektronix CCD camerawasusedresulting ina scaleof4′′.1 pixel−1 anda field of view of 70′ × 70′. Two exposures in [Oiii] and [Sii] of 2400 s each were takenduringthe observations,whileoneexposureof1800 sandoneof2400swereobtainedwiththeHα+[Nii]filter. Note thatthe finalimagesineachfilter arethe averageof the individual frames. The image reduction (bias subtraction, flat-field cor- rection) was carried out using the standard IRAF and MIDASandtheirnegativegray–scalerepresentationusing theSTARLINKKappaandFigaropackages.Theastrome- tryinformationwascalculatedforeachimageindividually usingstarsfromtheHubbleSpaceTelescope(HST)Guide StarCatalogue(Laskeretal.1999).Thespectrophotomet- ricstandardstarsHR5501,HR7596,HR7950,andHR8634 (Hamuyetal.1992;1994)wereusedforabsolutefluxcali- bration.Allcoordinatesquotedinthispaperrefertoepoch 2000. 2.2. Spectroscopy Low dispersion long–slit spectra were obtained with the Fig.1. The field of G 17.4–2.3 in the Hα+[Nii] filter. 1.3 m telescope at Skinakas Observatory in 2001 August The image has been smoothed to suppress the residuals 21. The 1300 line mm−1 grating was used in conjunction from the imperfect continuum subtraction. Shadings run with a 2000×800 SITe CCD (15×15 µm2 pixels) which linearlyfrom0to120×10−17 ergs−1 cm−2 arcsec−2.The resulted in a scale of 1 ˚A pixel−1 and covers the range line segments seen near overexposed stars in this figure of 4750 ˚A – 6815 ˚A. The slit width is 7′.′7 and it was and the next figures are due to the blooming effect. orientedinthesouth–northdirection,whiletheslitlength is7.′9.Thecoordinatesoftheslitcentreareα=18h31m30s and δ = −14◦51′33′′ and two spectra of 3600 s each were In Table 1, we present typical averagefluxes measured in obtained.ThespectrophotometricstandardstarsHR5501, several locations within the field of G 17.4–2.3 including HR7596, HR9087, HR718, and HR7950 were observed in the unknown Hii region which is located south–east of order to calibrate the spectra of G 17.4–2.3. the remnant. A deeper study of these images shows that the emission from the brightest part of the remnant (east filament) originates from shock heated gas since we esti- 3. Imaging of G 17.4–2.3 mate a ratio [S ii]/Hα ∼0.4–0.6, while a photoionization 3.1. The Hα+[Nii] and [Sii] emission line images mechanism produces the south–east Hii region([Sii]/Hα ∼0.2–0.3). The possibility that the Hii emission contam- The major characteristic revealed from our Hα + [Nii] inates the remnant’s emission can not be ruled out since and [S ii] images (Fig. 1 and 2, respectively) seems to be forsomeoftheareasclosetothefilamentaswellastothe thelowsurfacebrightnessofG17.4–2.3.Theweakdiffuse centralregionoftheremnant,weestimate[Sii]/Hα∼0.2– emission is present in the south, south–east and central 0.3. The known dark, extended nebula LDN 379 (Lynds areasof the remnant,while no emissionis detected in the 1965), which is at a distance of ∼200 pc (Hilton et al. north–west part. The most interesting region lies in the 1986) is also visible in the low ionization images. eastpartoftheremnant,whereafilamentarystructureex- ists,whichis verywellcorrelatedwith the radioemission. P. Boumis et al.: First optical light from thesupernovaremnant G 17.4−2.3 3 Fig.2. The [Sii] image of the area around G 17.4–2.3, Fig.3. G 17.4–2.3 imaged with the medium ionization which has been smoothed to suppress the residuals from line of [Oiii] 5007˚A. The image has been smoothed to the imperfect continuum subtraction. Shadings run lin- suppress the residuals from the imperfect continuum sub- early from 0 to 20× 10−17 erg s−1 cm−2 arcsec−2. tractionandtheshadingsrunlinearlyfrom0to30×10−17 ergs−1 cm−2 arcsec−2.Theprojectionofa∼26′ diameter sphere that matches the location and orientation of the 3.2. The [Oiii] 5007˚A image filament is also shown. In contrast to the previous results, the image of the medium ionization [Oiii] line shows clearly filamentary apertures (Ia and Ib, Fig. 5) along the slit. In particu- nature of the observed emission (Fig. 3). This bright fil- lar, apertures Ia and Ib have an offset of 9′′.5 and 118′′.5 ament extends for ∼24′ in the east, south–east, while no northofthe slitcentre,respectively.TheaperturesIaand significant emission was found in other areas of the rem- Ib were selected because they are free of field stars in an nant. We do not detect [Oiii] emission where the diffuse otherwise crowded field. In addition, these apertures in- Hα + [Nii] and [Sii] emission is detected. Table 1 lists clude sufficient line emission, especially in the blue part also typical [Oiii] fluxes measured in different parts of of the spectrum, to allow an accurate determination of the filament. The latter,matches very well with the radio the observed lines. The background extraction aperture maps of G 17.4–2.3at1400MHz and4850MHz, suggest- was taken towards the north end of the slit. We utilized ing their association (Fig. 4). The correlation shows that the flux calibrated images to identify the nature of this the filamentis locatedcloseto the outeredge ofthe radio background emission used in the spectra. The [Sii]/Hα contours but the low resolution of the radio images does ratio of ∼0.2 measuredin the images suggests an Hii ori- not help us to determine the actual position of the fila- gin for the backgroundaperture emission. Assuming that mentwithrespecttotheshockfront.Thegeometryofthe its strength does not change appreciably at the Ia and [Oiii]filamentallowsustodefinea∼26′circleindiameter Ib positions we performed the background subtraction to (dashedcircleinFig.3)fortheremnant,withitscenterat identify the SNR emission. The signalto noise ratios pre- α ≃ 18h30m23.3s, δ ≃ −14◦45′25′′. Note that the optical sented in Table 2 do not include calibration errors,which angular size is in very good agreement with the value of are less than 10%. Both extracted apertures show clearly ∼24′ given in Green’s catalogue (Green 2001). However, thattheobservedopticalemissionoriginatesshockheated alargerangulardiametercannotbe excludedsinceX–ray gas, since the [Sii]/Hα > 0.7. The [Sii]λ 6716/6731 ra- emission has not been detected so far and the radio shell tio of 1.2–1.4 indicates low electron densities (Osterbrock is incomplete. 1989). The very strong [Oiii] emission detected in aperture Iasuggestsashockvelocitygreaterthan100kms−1 (Cox 4. The long–slit spectra from G 17.4–2.3 & Raymond1985),while the sulfur lines ratioindicate an The low resolution spectra were taken on the relatively electron density ∼240 cm−3 (Osterbrock 1989). However, bright optical filament in the east part of the remnant taking into account the statistical errors on the sulfur (its exact position is given in Sect. 2.2). In Table 2, we lines, electron densities up to 400 cm−3 are compatible present the relative line fluxes taken from two different withourmeasurements.Theshockvelocityimpliesbythe 4 P. Boumis et al.: First optical light from thesupernovaremnant G 17.4−2.3 Fig.4. The correlation between the [Oiii] emission and theradioemissionat1400MHz(dashline)and4850MHz (solidline)isshowninthisfigure.The1400MHz(Condon et al. 1998) and the 4850 MHz (Condon et al. 1994) ra- dio contoursscale linearlyfrom1.1×10−3 Jy/beamto 0.1 Jy/beamand3.5×10−2 Jy/beamto0.3 Jy/beam,respec- tively.LDN 379is alsoastrongradiosourcealthoughnot detectedin [Oiii].Note that inthe north–eastareaofthe filament we find the well–defined radio shell of the SNR G 17.8−2.6,which is also not detected in the optical. Ib spectrum could be less but still around 100 km s−1, Fig.5.ThespectraofaperturesIa(top)andIb(bottom). while the electron density is even lower than in Ia (less than 120 cm−3). Hester et al. (1987) suggested that the presence of such inhomogeneities and density variations would mainly af- 5. Discussion fecttherecombinationzonewherethelowionizationlines ThesupernovaremnantG17.4–2.3showsupasanincom- areproducedanditcouldalsoexplainthe[Oiii]/Hαratio pleteshellintheradioboundwithoutanyX–rayemission variations seen in the long–slit spectra. detected so far.The low ionizationimages generally show An interstellar extinction c at positions Ia and Ib (see diffuse emission in the south and south–east areas of the Table 2), of 0.62 (± 0.33) and 1.10 (± 0.17) or an A of V remnant.Incontrast,afilamentarystructurehasbeendis- 1.27 (± 0.67) and 2.26 (± 0.35) were measured, respec- covered in the medium ionization [Oiii] line in the east, tively.We havealsodeterminedthe electrondensity mea- south–east region which is very well correlated with the suring the density sensitive line ratio of [Sii]λ 6716/6731. radioemissionat1400and4850MHzandcoulddefinethe The densities we measure are below 400 cm−3. Assuming remnant’s outer edge. This correlation indicates that the that the temperature is close to 104 K, it is possible to observed emission is associated to G 17.4–2.3. Both the estimate basic SNR parameters. The remnant under in- calibrated images and the long–slit spectra suggest that vestigation is one of the least studied remnants and thus, the detected emission results from shock heated gas since thecurrentstageofitsevolutionisunknown.Thepossibil- the [Sii]/Hα ratio exceeds the empirical SNR criterion ities that the remnant is still in the adiabatic phase or in value of 0.4–0.5. Note that the Hii region found in the theradiativephaseofitsevolutioncannotbeexcludedand lowionizationimagesshowsa[Sii]/Hαratioof∼0.2.The willbeexaminedinthefollowing.Thepreshockcloudden- eastern filament lies very close to this Hii region. The sityn canbe measuredbyusingtherelationship(Dopita c morphological differences between the low and medium 1979) ionization lines provide evidence for significant inhomo- geneities and density variations in the ambient medium. n ≃ 45 n V2 cm−3, (1) [SII] c s P. Boumis et al.: First optical light from thesupernovaremnant G 17.4−2.3 5 Table 2. Relative line fluxes. greater than 7×1020 cm−2, while for n ≈1 cm−3 it be- 0 comesgreaterthan7×1021cm−2.Combiningtheprevious Ia Ib results and assuming that the column density is found in Line(˚A) Fa,b Fa,b the rangeof2−5×1021 cm−2,then the lowerinterstellar 4861 Hβ 22 (3)c 15 (6) density seems to be more probable. 4959 [OIII] 70 (9) 4 (2) Using the results of Cioffi et al. (1988) and the above 5007 [OIII] 221 (28) 23 (9) rangeofparametersforn of0.1and1cm−3,thepressure 6548 [Nii] 39 (13) 21 (16) 0 driven snowplow (PDS) radii are calculated as 38 and 14 6563 Hα 100 (30) 100 (75) pc, respectively. The radii, when compared with the esti- 6584 [Nii] 142 (44) 75 (57) 6716 [Sii] 44 (15) 41 (34) mated SNR radius of 17 D2.3kpc pc may suggest that the PDS phase has not begun yet. The derived velocities of 6731 [Sii] 38 (13) 30 (25) the main shock front at the beginning of this stage are AbsoluteHα fluxd 7.9 19 ∼300–400 km s−1. Their comparison with the estimated Hα/Hβ 4.5 (3) 6.7 (6) shockvelocitiesintotheinterstellarcloudsresultinaden- [Sii]/Hα 0.82 (17) 0.71 (36) sitycontrastof∼10betweentheambientmediumandthe F(6716)/F(6731) 1.2 (10) 1.4 (20) interstellar “clouds”. However, since neither the distance a Uncorrected for interstellar extinction northeinterstellarmediumdensityareaccuratelyknown, b Listed fluxes are a signal to noise weighted average of the we cannotconfidently determine the currentstage ofevo- individualfluxes lution of G 17.4–2.3. c Numbersinparenthesesrepresentthesignaltonoiseratioof thequoted fluxes d In units of 10−17 erg s−1 cm−2 arcsec−2 6. Conclusions Allfluxesnormalized to F(Hα)=100 The faintsupernovaremnantG17.4–2.3wasobservedfor the first time in major optical emission lines. The im- where n is the electron density derived from the agesshowfilamentaryanddiffuseemissionstructures.The [SII] sulfurlineratioandV istheshockvelocityintotheclouds bright[Oiii]filamentisverywellcorrelatedwiththerem- s inunitsof100kms−1.Furthermore,theblastwaveenergy nant’s radio emission at 1400 and 4850 MHz suggesting canbeexpressedintermsofthecloudparametersbyusing theirassociation.Thefluxcalibratedimagesandthelong– the equation given by McKee & Cowie (1975) slit spectra indicate that the emission arises from shock heated gas. The observed optical filamentary structure E =2×10−5β−1n V2 r 3 erg. (2) 51 c s s provides some evidence for significant inhomogeneities in the ambient medium, implying that the main blast wave The factor β is approximately equal to 1 at the blast propagates into an inhomogeneous medium. wave shock, E is the explosion energy in units of 1051 51 erg and r the radius of the remnant in pc. By using the s Acknowledgements. upper limit on the electron density of 400 cm−3, which wasderivedfromourspectra,weobtainfromEq.(1)that The authors would like to thank the referee for his com- ncVs2 < 8.9. Then Eq. (2) becomes E51 < 0.08 D31kpc, ments and suggestions which helped to clarify, and en- where D1kpc the distance to the remnant in units of 1 hance, the scope of this paper. Skinakas Observatory is kpc. a collaborative project of the University of Crete, the An estimated value of NH ∼ 5×1021 cm−2 is given Foundation for Research and Technology-Hellas and the by Dickey & Lockman (1990) for the column density in Max-Planck-Institut fu¨r Extraterrestrische Physik. This thedirectionofG17.4–2.3.Consideringthat(Ryteretal. researchhas made use of data obtained through the High 1975) Energy Astrophysics Science Archive Research Center NH =6.8 (±1.6)×1021 EB−V cm−2, (3) OFlniglihnteCSeenrtveirc.e, provided by the NASA/Goddard Space where AV = 3.1×EB−V (Kaler 1976), we obtain an NH of3×1021cm−2and5×1021cm−2forthetwocvalues References calculated from our spectra, respectively. A column den- sity of 2+1×1021 cm−2 was measured towards an X–ray Case G. L., and Bhattacharya D.1998, ApJ504, 761 −2 ◦ CioffiD.F.,McKeeC.F.andBertschingerE.,1988,ApJ,334, binarysystem,located1.2 eastofG17.4–2.3atadistance of ∼3.1 kpc (Ribo´ et al. 1999). Since, there are no other 252 CondonJ.J.,BroderickJ.J.,Seielstad G.A.,DouglasK.and measurements of the interstellar density n , values of 0.1 0 Gregory P. C., 1994, AJ, 107, 1829 and 1.0 will be examined. Following the result of Eq. (2) CondonJ.J.,CottonW.D.,Greisen E.W.,YinQ.F.,Perley and assuming the typical value of 1 for the supernova ex- R. A., Taylor G. B. and Broderick J. J., 1998, AJ, 115, plosionenergy (E51), we find that the remnant may lie at 1693 distancesgreaterthan2.3kpc.Then,thelowerinterstellar Cox D. P.and RaymondJ. C., 1985, ApJ, 298, 651 density of ∼0.1 cm−3 suggests that the column density is Dickey J. M. and Lockman F. J., 1990, ARAA,28, 215 6 P. Boumis et al.: First optical light from thesupernovaremnant G 17.4−2.3 Dopita M. A., 1979, ApJS,40, 455 Fesen R. A., Blair W. P. and Kirshner R. P., 1985, ApJ, 292, 29 Fesen R.A. and Hurford A. P., 1995, AJ, 110, 747 FesenR.A.,WinklerP.F.,RathoreY.,DownesR.A.,Wallace D. and Tweedy R. W., 1997, AJ, 113, 767 Green D.A.,FrailD.A.,Goss W.M. andOtrupcekR.,1997, AJ, 114, 2058 GreenD.A.,2001,ACatalogofGalacticSupernovaRemnants (2001 version), Mullard Radio Astronomy Observatory Cambridge Hester J. J., 1987, ApJ, 314, 187 HiltonJ.,WhiteG.J.,CroninN.J.andRaineyR.,1986,A&A, 154, 274 HamuyM.,WalkerA.R.,SuntzeffN.B.,GigouxP.,Heathcote S. R. & Phillips M. M., 1992, PASP,104, 533 Hamuy M., Suntzeff N. B., Heathcote S. R., Walker A. R., Gigoux P.& Phillips M. M., 1994, PASP,106, 566 Kaler J. B., 1976, ApJS,31, 517 Lasker B. M., Russel J. N. & Jenkner H., 1999, in the HST Guide Star Catalog, version 1.1-ACT, The Association of Universities for Research in Astronomy, Inc LyndsB. T., 1965, ApJS, 12, 163 Mavromatakis, F., Papamastorakis, J., Paleologou, E. V. and VenturaJ, 2000, A&A353, 371 Mavromatakis, F., Papamastorakis, J., Ventura, J., Becker, W., Paleologou, E. V. and Schaudel, 2001, A&A 370, 265 Mavromatakis, F., Boumis P. and Paleologou, E. V., 2002, A&A,accepted McKee C. F. and Cowie L. L., 1975, ApJ,195, 715 Osterbrock D. E., 1989, Astrophysics of gaseous nebulae, eds. W. H. Freeman & Company Raymond J. C., 1979, ApJS, 39, 1 Reich W., Fu¨rst E., Reich P. and Junkes N., 1988, in IAU Colloq. 101, Supernova Remnants and the Interstellar Medium,eds.RogerR.S.andLandeckerT.L.(Cambridge Univ. Press), 293 ReichW.,Fu¨rstE.,ReichP.andReifK.,1990,A&AS,85,633 Rib´o M., Reig P., Mart´i and Paredes J. M. 1999, A&A, 347, 518 RyterC., Cesarsky C.J.andAudouzeJ.,1975, ApJ,198,103 Seward F. D., Dame T. M., Fesen R. A. and Aschenbach B., 1995, ApJ, 449, 681 ShullJ. M. and McKee C. F., 1979, ApJ,227, 131 Smith R. C., Kirshner R. P., Blair W. P., Long K. S. and Winkler P. F., 1993, ApJ,407, 564