Far UV and Optical Emissions from Three Very Large Supernova Remnants Located at Unusually High Galactic Latitudes
Robert A. Fesen, Marcel Drechsler, Kathryn E. Weil, Xavier Strottner, John C. Raymond, Justin Rupert, Dan Milisavljevic, Bhagya M. Subrayan, Dennis di Cicco, Sean Walker, David Mittelman, Mathew Ludgate
DDraft version February 26, 2021
Typeset using L A TEX twocolumn style in AASTeX63
Far UV and Optical Emissions from Three Apparent Supernova RemnantsLocated at Unusually High Galactic Latitudes
Robert A. Fesen, Marcel Drechsler, Kathryn E. Weil, Xavier Strottner, John C. Raymond, Dan Milisavljevic, Bhagya M. Subrayan, Dennis di Cicco, Sean Walker, and David Mittelman Baerenstein Observatory, Feldstrasse 17, D-09471 Baerenstein, Germany Department of Physics and Astronomy, Purdue University, 525 Northwestern Avenue, West Lafayette, IN 47907 USA Montfraze, 01370 Saint Etienne Du Bois France Harvard-Smithsonian Center for Astrophysics, 60 Garden St., Cambridge, MA 02138, USA MDW Sky Survey, New Mexico Skies Observatory, Mayhill, NM 88339, USA
ABSTRACTGalactic supernova remnants (SNRs) with angular dimensions greater than a few degrees are rel-atively rare, as are remnants located more than ten degrees off the Galactic plane. Here we reportthe results of a UV and optical investigation of two previously suspected SNRs that are more than (cid:39)
10 degrees in both angular diameter and Galactic latitude. One is the proposed G354-33 remnantdiscovered in 2008 through 1420 MHz polarization maps. GALEX far UV (FUV) emission and H α mosaics show the object’s radio emission coincident with a nearly continuous 11 ◦ × ◦ shell of thinUV filaments which surround a broad H α emission ring. Another proposed high latitude SNR is theenormous 20 ◦ × ◦ Antlia nebula (G275.5+18.4) discovered in 2002 through low-resolution all-sky H α images and ROSAT soft X-ray emission. GALEX FUV image mosaics along with deep H α imagesand optical spectra of several filaments indicate the presence of shocks throughout the nebula withestimated shock velocities of 70 to over 100 km s − . We conclude that both of these nebulae arebona fide SNRs with estimated ages less than 10 yr despite their unusually large angular dimensions.Using FUV and optical spectra and images, we also report finding an apparent new, high latitude SNR(G249.2+24.4) approximately 2 . ◦ × . ◦ in size based on its UV and optical emission properties. Keywords:
SN: individual objects: ISM: supernova remnant INTRODUCTIONIn spite of an increasing variety of supernovae (SNe)sub-types, the majority of SNe are generally believed torelease 0 . − × erg, although superluminous SNemay release substantially more (Howell 2017). Only asmall percentage of this comes out as visible light, withmost of a SN’s energy initially carried away in the formof kinetic energy. It is this enormous point deposition ofenergy that has a significant impact on the structure andenergy content of a galaxy’s interstellar medium throughan expanding remnant that can last largely intact up to (cid:39) yr (Cox & Smith 1974; McKee & Ostriker 1977;Blondin et al. 1998).Currently, there are some 300 confirmed Galactic su-pernova remnants (SNRs) catalogued with dozens ofother suspected SNRs and more added every few years(Ferrand & Safi-Harb 2012; Green 2019). Most GalacticSNRs are less than a degree in angular size, more than1 kpc distant and well evolved, with typical estimatedages between 10 and 10 yr.Out-sized attention relative to their population per-centage has been directed to the handful of the Milky Way’s young remnants with ages less than 5000 yr.These include the Crab Nebula, Cassiopeia A, the rem-nants of Tycho and Kepler, the Vela remnant, and Pup-pis A. This increased attention is due to their highexpansion velocities, metal-rich ejecta, bright emissionacross the entire electromagnetic spectrum, and clearerconnections to the various core-collapse and thermonu-clear SN sub-types.The majority of Galactic SNRs were first detectedthrough radio observations due to their characteristicsynchrotron nonthermal radio emission associated withshocked gas (Downes 1971; Chevalier 1977; Green 1984,2004). Such nonthermal emission leads to a power lawflux density, S, with S ∝ ν − α where α is the emissionspectral index with typical SNR values between -0.3 and-0.7. Rarer pulsar wind-driven SNRs exhibit a muchflatter radio spectrum, with spectral indices around zero.While multi-frequency radio surveys have been histor-ically the dominant tool for finding SNRs, X-ray studieshave also led to the discovery of several additional SNRs.These include RX J1713.7-3946 (G347.3-0.5; Pfeffer-mann & Aschenbach 1996) and the young “Vela Junior”SNR (G266.2-1.2; Aschenbach 1998) coincident with the a r X i v : . [ a s t r o - ph . H E ] F e b Figure 1.
GALEX FUV intensity map of half of the sky in Galactic coordinates, showing locations of two apparent SNRs,G249+24 and G354-33, and the exceptionally large Antlia SNR. White areas denote regions with no GALEX imaging data. much larger Vela SNR. Recent examples of X-ray con-firmed or discovered remnants include several new SNRsin the LMC (Maggi et al. 2014) and X-ray confirmationof the suspected Galactic remnant G53.4+0.0 (Driessenet al. 2018).Despite the fact that less than 50% of Galactic SNRsexhibit any appreciable associated optical emission, aremnant’s optical emission can be useful in confirmingthe presence of high-velocity shocks and in defining aremnant’s overall size and morphology. Although dis-coveries of new SNRs in the optical is relatively rare,there have been several recently reported discoveries(Stupar et al. 2007; Boumis et al. 2009; Stupar & Parker2012; Fesen et al. 2015; Neustadt et al. 2017; Stuparet al. 2018; How et al. 2018; Fesen et al. 2019) along withmany proposed SNR candidates (Stupar et al. 2008; Stu-par & Parker 2011; Boumis et al. 2009; Alikakos et al.2012; Sabin et al. 2013).The main criterion for optical SNR identification isa line ratio of I([S II ])/I(H α ) ≥ . II ] line emissions at 1.27 and 1.64 µ m havealso been used to detect dust obscured SNRs in nearbygalaxies (see review by Long 2017).One wavelength regime that has not yet been fullyexploited to search for new Galactic SNRs is the ultra-violet (UV). Here we present results of an initial studyof SNRs located far away from the Galactic plane us-ing wide field-of-view (FOV) UV images assembled fromthe Galaxy Evolution Explorer (GALEX) All-Sky sur-vey (Bianchi 2009). We began this research by investi-gating two unusually high latitude suspected supernovaremnants including the exceptionally large Antlia rem-nant (McCullough et al. 2002). During this work, wealso found a new apparent SNR. UV and optical im-ages plus some follow-up optical spectra are describedin §
2, with results presented in §
3. We discuss the gen-eral properties of these SNRs in §
4, with our conclusionsand discussion of helpful follow-up observations given in § DATA AND OBSERVATIONS2.1.
GALEX UV Images
The GALEX satellite was a NASA science mission ledby the California Institute of Technology who operated
Table 1.
Locations and Dimensions of SNRsSNR ID Approximate Center (J2000) Galactic Coordinates Diameter Distance for Dia. = 100 pcG354-33 RA = 20 h m Dec = − ◦ (cid:48) l = 353.9 b = − . ◦ × ◦ ∼
500 pcG249+24 RA = 09 h m Dec = − ◦ (cid:48) l = 249.2 b = +24 . . ◦ × . ◦ ∼ h m Dec = − ◦ (cid:48) l = 275.5 b = +18 . ◦ × ◦ ∼
250 pc it from July 2003 through February 2012. The maininstrument was a 50 cm diameter modified Ritchey-Chr´etien telescope, a dichroic beam splitter and astig-matism corrector, and two microchannel plate detectorsto simultaneously cover two wavelength bands with a1.25 degree field of view with a resolution of 1 . (cid:48)(cid:48) pixel − .Direct images were obtained in two broad bandpasses:a far-UV (FUV) channel sensitive to light in the 1344 to1786 ˚A range, and a near-UV (NUV) channel covering1771 to 2831 ˚A. Resulting images are circular in shapewith an image FWHM resolution of ∼ . (cid:48)(cid:48) and ∼ . (cid:48)(cid:48) in the FUV and NUV bands, respectively.Being mainly a mission to study the UV propertiesof galaxies in the local universe, GALEX survey imageslargely avoided the complex and external galaxy poorGalactic plane. However, even its All-Sky Survey pro-gram focused away from the plane of the Milky Way didnot cover all possible areas, leading to numerous gaps(see Fig. 1). Nonetheless, some 26,000 square degreeswere imaged to a depth of m AB = 20.5 (Bianchi 2009).Using the on-line GALEX images, large FUV and NUVimage mosaics were examined of several regions typicallymore than 10 degrees off the Galactic plane.2.2. H α images Wide-field, low-resolution H α images of regionsaround suspected SNRs obtained as part of the MDWHydrogen-Alpha Survey were examined. This surveyuses a 130 mm telescope at the New Mexico Skies Ob-servatory, with a FLI ProLine 16803 CCD and 3 nm fil-ter centered on H α . This telescope-camera system hasa field-of-view (FOV) of approximately 3 . ◦ × . ◦ . (cid:48)(cid:48)
17. Each field was observed 12 times eachwith an exposure of 1200 s.Follow-up, higher-resolution images of interestingFUV filaments were obtained with a 1.3m telescope atthe MDM Observatory at Kitt Peak, Arizona using a1024 × . (cid:48) . Withon-chip 2 × . (cid:48)(cid:48) .Two exposures of 900 s each were taken at four positionswhere the FUV images of G249+24 exhibited optical fil-aments and where long slit spectra were also obtained.We have also made large mosaics from Southern H α Sky Survey Atlas (SHASSA; Gaustad et al. 2001). This wide-angle robotic survey covered δ = +15 ◦ to − ◦ with 13 ◦ square images with an angular resolution of (cid:39) . (cid:48) . Narrow passband continuum filters on blue and redsides of H α centered at 6440 and 6770 ˚A allow for stellarand background subtraction. The survey had a nativesensitive level of 1 . × − erg cm − s − arcsec − whichcould be significantly lowered with smoothing.2.3. Optical Spectra
Low-dispersion optical spectra of filaments in the morenorthern suspected SNR were obtained with the MDM2.4m Hiltner telescope using Ohio State Multi-ObjectSpectrograph (OSMOS; Martini et al. 2011). Employinga blue VPH grism (R = 1600) and a 1.2 arcsec wide slit,exposures of 2 ×
900 s were taken covering 4000–6900 ˚Awith a spectral resolution of 1.2 ˚A pixel − and a FWHM= 3.5 ˚A. Spectra were extracted from regions clean ofappreciable emission along each of the 15 arcminute slits.For southern hemisphere nebulae, spectra were takenof suspected SNR filaments using the Robert StobieSpectrograph (RSS) on the 10 m SALT telescope inSouth Africa. Using a 900 lines per mm and a 1.5 arcsecwide slit, spectra were obtained covering 4050 to 7120˚A region with a FWHM resolution of 5 ˚A and a disper-sion of 1.0 ˚A pix − . Exposures ranged from 2 × × in Astropy.MDM spectra were further reduced using PYRAF andL.A. Cosmic (van Dokkum 2001) to remove cosmic raysand calibrated using a HgNe or Ar lamp and spectro-scopic standard stars (Oke 1974; Massey & Gronwall1990). RESULTSA few large and unusually high Galactic latitude neb-ulae have been recently proposed as previously unrecog-nized SNRs. These objects have angular diameters wellin excess of the largest confirmed SNRs, namely the 8 ◦ diameter Vela SNR. If these objects were to be foundto be true SNRs, they would help expand the notion of https://github.com/jrthorstensen/thorosmos PYRAF is a product of the Space Telescope Science Institute,which is operated by AURA for NASA.
Figure 2.
GALEX FUV intensity map of G354-33 showing a roughly spherical shell of UV emission filaments. Approximateshell center: 20 h m s − ◦ (cid:48) . Note: Individual circular GALEX images are 1 . ◦ in diameter. how large an identifiable remnant can be and how faroff the plane should SNR researchers be looking.The proposed new SNRs include the huge ∼ ◦ di-ameter Antlia nebula (G275.5+18.4; McCullough et al.2002) and the ∼ ◦ diameter radio remnant G354-33(Testori et al. 2008). Due to limited data on it, Antliais listed only as a possible SNR in the most recent cat-alogue of Galactic supernova remnants (Green 2019),whereas no mention is made of G354-33 in either this2019 SNR catalogue or in online SNR lists . Below we discuss GALEX far UV (FUV) and optical prop-erties of these two suspected SNRs, plus a third object,G249.2+24.4, which we believe is likely a new SNR. Thepresentation order follows that of our research work.3.1. The Suspected Radio SNR G354-33
Using a 1420 MHz linear polarization survey of thesouthern hemisphere sky, Testori et al. (2008) identifieda large ∼ ◦ depolarized shell at α (J2000) = 20 h δ (J2000) = − ◦ , roughly corresponding to Galactic co-ordinates l = 353 ◦ b = − ◦ . This shell could also beseen in a 1.4 GHz polarization all-sky survey map (Reich& Reich 2009). Due to the object’s spherical morphol-ogy and radio properties, Testori et al. (2008) suggesteda SNR identification, with a size suggesting a distance Figure 3. Top:
Blow-up of a small section of the GALEXFUV map of G354-33 along the remnant’s northern rim.
Bottom:
Blow-up of a small section along G354-33’s west-ern limb. between 300 and 500 pc, a physical diameter of 17.4 pc × d and a z distance of 57.4 pc × d . Testori et al.(2008) noted that the object could be seen in 408 MHzdata (Haslam et al. 1982) and weakly in the SHASSAH α images (Gaustad et al. 2001). This object also showsup in gradients of linearly polarized synchrotron radioemission (Iacobelli et al. 2014). Recently, Bracco et al. (2020) briefly commented onthe presence of some GALEX FUV filaments coincidentwith this radio shell. However, to our knowledge, thereis no paper showing this suspected SNR’s FUV emis-sions. Below we present large mosaics of GALEX FUVimages along with a SHASSA H α image and comparethese to the low resolution 1420 MHz radio images.3.1.1. Far UV and H α Images
Wide-field mosaics of GALEX far UV (FUV) and nearUV (NUV) images show a large, faint UV emission shelllargely coincident with the 1420 MHz radio shell. TheUV emission consists of over a dozen thin emission fila-ments. Because these filaments appear brightest in theGALEX FUV images compared to NUV images, we willconcentrate mainly on the shell’s far UV emission struc-ture.The strong shock-like morphology of these filamentsand their arrangement in a nearly continuous shell lendssupport to Testori et al. (2008)’s suggestion for it to bea previously unrecognized Galactic supernova remnant,albeit with unusually large angular dimensions and dis-tance from the Galactic plane.Figure 2 shows a mosaic of GALEX FUV images ofG354-34 smoothed by a nine point Gaussian . As seenin this FUV mosaic, the object exhibits a large and co-herent set of sharp UV emitting filaments arranged in anearly complete shell-like structure. Multiple and over-lapping filaments are common along the northern andwestern limbs. Although many filaments are unresolvedat GALEX’s 4.6 (cid:48)(cid:48) FWHM resolution, some filaments ap-pear partially resolved in places, especially along itsnorthernmost edge. It is not clear if this appearanceis due to closely spaced multiple filaments or resolutionof a single emission filament.Based on its brighter FUV filaments, we estimatesomewhat larger angular dimensions than the 10 ◦ citedby Testori et al. (2008). Instead, we find dimensions of (cid:39) ◦ East-West and (cid:39) ◦ North-South, with a distinctindentation along its south-western limb. Its southernextent is poorly constrained in our GALEX FUV mosaicbut FUV filaments extend south to at least a Declina-tion of − ◦ .Due to its large size, many of the filamentary detailsvisible in the full resolution FUV mosaic are lost in Fig-ure 2. Consequently, we show in Figure 3 blow-ups oftwo regions, one along G354-33’s northern rim and onealong its western limb to give a better sense of its fine-scale morphology. It should be noted that some individ-ual filaments are two to three degrees in length, equiva-lent to the entire lengths of some of the largest GalacticSNRs. Because of distortions due to the large size of this emission struc-ture, coordinates shown are only accurate to a few arc minutes.
Figure 4.
GALEX FUV (left) and continuum subtracted SHASSA H α (right) images of G354-33. Figure 5.
Matched GALEX FUV and SHASSA H α images for a bright filament along G354-33’s western limb (upper panels)and northeastern limb (lower panels). Field centers are: 20 h m − ◦ (cid:48) (top) and 19 h m − ◦ (cid:48) (bottom). Continuum subtracted SHASSA H α image mosaics ofthe remnant region reveal a large and thick diffuse shellof emission located within the boundaries of the FUVfilaments. Figure 4 shows the GALEX FUV along sideof the SHASSA Halpha image covering the same skyregion. The H α emission structure is far more diffusethan that seen in the GALEX FUV image with very fewfilaments. We estimate an H α flux of around 1 − × − erg cm − s − arcsec − for most of the diffuse emissionand about 3 times times this for the area of brighteremission along the nebula’s northwestern limb.While its H α emission’s location matches that seenin the FUV image, weak H α emission is seen to extendfarther to the east and south than is readily visible inthe GALEX FUV mosaic image. However in general,the FUV filaments appear to mark the outer edges ofthe diffuse H α emission shell.This is shown in Figure 5 where we compare FUV vs.H α images for western and northeastern sections alongits limb. The upper images show the presence of weakH α emission coincident with the bright main section ofthe FUV filament, with diffuse H α emission extendingtoward the east bordered by the long FUV filament. Thelower panels show a section along the nebula’s northeast-ern limb where again the UV filaments are seen to markthe extent of the diffuse H α emission. The SHASSAimage also gives a hint of filamentary structure that re-sembles that seen in the FUV image.3.1.2. Radio and X-rays
As already noted, Testori et al. (2008) used the 1420MHz Villa Elisa radio data (Testori et al. 2001) to iden-tify the radio emission ring, G354-33, as a possible SNR.The upper panel of Figure 6 shows a 60 ◦ × ◦ widesection of the Villa Elisa 1420 MHz polarization south-ern sky survey data roughly centered on the remnant.The emission ring appears distinct in shape and sep-arate from other emission features along the southernedge of the Galactic plane (i.e., the right-hand side ofthe image).The images in the lower panel of Figure 6 show identi-cal regions of a GALEX FUV image mosaic (left panel)and the 1420 MHz Villa Elisa radio image (right panel).Overall, there is good agreement in both size and lo-cation for the FUV emission shell and radio ring seenin these images. This includes the greater extent to-ward the southern limb area. We note that the brightestportion of this radio emission shell lies on the object’swestern limb, the side facing the Galactic plane.While coincident in location, the radio shell has asmaller diameter of (cid:39) ◦ than the roughly 11 ◦ East-West diameter seen in the FUV images and is centeredaround 20 h m s − ◦ (cid:48) ( l =353.5 b = − .
2) whichis slightly different from that of the GALEX FUV emis-sion shell (see Table 1). Thus, although we have usedthe name G354-33 above following Testori et al. (2008), we believe the object’s correct Galactic coordinate nameshould be G353.9-33.4 (see Table 1).Examination of ROSAT on-line data showed no obvi-ous associated X-ray emission with this nebula. This isnot surprising given its location 33.4 ◦ below the Galacticplane. Moreover, in view of the lack of any bright FUVemission nebulae along its limbs, this suggests that itlies in a region with few or none large-scale interstellarclouds which might generate significant X-ray flux.3.2. G249.2+24.4
Using GALEX FUV mosaics, we discovered a regioncontaining several long, thin, shock-like morphology fil-aments aligned mostly N-S and arranged in a roughly el-liptical shape ∼ . × ◦ in size. These filaments appearcentered at α (J2000) = 9 h m δ (J2000) = − ◦ (cid:48) ,which corresponds to Galactic coordinates l = 249.2 ◦ , b = +24.4 ◦ . We call this object G249+24. Below, wepresent images of this object’s UV and H α emissionstructure along with optical spectra of four of its fila-mentary regions.3.2.1. Far UV Mosaic Image
Figure 7 presents a 5 . ◦ × . ◦ mosaic of GALEX FUVimages centered on G249+24. The image shows sev-eral UV filaments along with a few smaller diffuse emis-sion patches. Although GALEX imaging of this regionis fairly incomplete, enough imaging exists to indicateG249+24’s minimum size and extent .The nebula’s brightest filaments are concentrated inits southern region where there is a bright, long gen-tly curved filament along its southwestern edge. Thisfeature itself consists of several separate, closely alignedfilaments. Although a smaller but a similarly bright fil-ament caught on the western edge of a more northernGALEX image might give the impression that this fila-ment’s full length and extent is partially missing in thismosaic, this is not the case based on H α images (seebelow).Along the northern portion of the FUV emission mo-saic lies a fairly diffuse emission patch plus a few shortfaint filaments one of which is curved toward the brightSW filament. These features, taken together with thecomplex of southern filaments near δ = − ◦ , suggeststhe nebula has an angular N-S dimension around 4.2 ◦ .In contrast, its E-W dimensions are quite a bit less.A faint filament seen in G249+24’s southeastern regionsuggests an E-W dimension of ≈ ◦ .3.2.2. H α Emission
Whereas the object’s FUV emission filaments arereadily apparent in the GALEX image mosaic shown The bright feature at 9 h m , − ◦ (cid:48) is HD 82093, anAp(EuSrCr) star; V = 7.08. Figure 6.
Upper P anel ◦ × ◦ region centered onG354-33. Lower P anels : Comparison of the G354-33 region in the GALEX FUV image with the 1420 MHz image showingexcellent size and positional agreement of the 1420 MHz radio emission ring with the shell of UV emission filaments. in Figure 7, its optical emissions are relatively faint. Al-though some of its brighter filaments are weakly visibleon broad red passband Digital Sky Survey images, theyare so faint and scattered over such a large area thatit is not surprising that this nebula had not attractedprior attention.The H α images of the MDW All-Sky survey revealmore fully G249+24’s optical emissions. In Figure 8,we show a side-by-side comparison of G249+24’s centralemission structure in GALEX FUV and the MDW H α images. In general, there is a good agreement betweenthe nebula’s UV and H α emissions. The bright FUV andcurved southwestern filament shows up strongly in H α as does a shorter filament a degree to its east. CorrelatedUV – H α emission is also seen along the upper left-handportion of Figure 8, where diffuse and filament type ofemission is seen in both. (Note: Faint diffuse emissionseen along the lower portion of the H α image extends several degrees to both the east and west of the regionand hence does not seem to be connected to G249+24’sother optical features.)However, little in the way of H α emission can be seenalong the object’s southeast limb, where in contrast onefinds considerable FUV emission. A similar differencebetween FUV and H α fluxes is seen for the south-centralarea around 9 h m , − ◦ (cid:48) . Consequently, it is clearthat this nebula’s overall structure is more readily visiblein the FUV than in even fairly deep H α images.A few higher resolution H α images of selected regionsof the nebula were also obtained. Although these im-ages were taken in preparation of follow-up spectralobservations, they showed considerably more detail toG249+24’s filamentary emission than seen in Figure 8.Unfortunately given its large angular size, these compar-atively small images serve only as a hint of the nebula’srich and detailed optical emission structure. Figure 7.
FUV intensity map of G249+24 showing an elliptical shaped shell of UV emission filaments.
Optical Spectra
Low-dispersion, exploratory slit spectra were taken atat four positions (P1 through P4) shown in Figure 9 toinvestigate the emission nature G249+24’s optical emis-sion filaments. Figure 10 shows the resulting spectracovering the wavelength region 6200 to 6900 ˚A.These spectra reveal evidence for its filaments as shockheated ISM like that commonly seen in SNRs. Spectraobtained at P2, P3 and P4 exhibited [S II ]/H α ratios of1.26 ± .
20, 0.45 ± .
04, and 0.91 ± .
15, respectively, thuswell above the standard criteria ratio of 0.40 indicativeof shock emission. Added support for shocks is the de-tection of [O I ] λ I ] is a secondary indicatorfor shock emission commonly observed in evolved SNRs(Fesen et al. 1985; Kopsacheili et al. 2020).The [S II ] λ λ + recombination zoneand is nearly independent of electron temperature (Os- terbrock & Ferland 2006). The λ λ e <
100 cm − ; 1 . ± .
04 for P2, and 1 . ± .
03 forP4. A bit surprisingly, this ratio for P3 was observedto be (cid:39) e density much higher thanthe other locations of around 500 cm − . Interestingly,the spectrum seen at P3 also showed much weaker [N II ] λ α /[O III ] can be seen to vary considerablydown stream from the shock front.Figure 11 also shows the 2D image of the backgroundsubtracted P1 spectrum for the areas around H α , and[O III ] λ λ β . This figure shows a clear0 Figure 8.
G249+24’s FUV emissions (left) compared to its H α emission as seen in MDW’s All-Sky H α Survey images (right).
Figure 9.
Left: MDW H α image of G249+24 showing the four locations where long slit spectra were taken. Right: Higherresolution of filaments and slit positions. Slit lengths as shown are 1.5 (cid:48) . Rest Wavelength (Å) F l u x ( e r g c m s Å ) Figure 10.
Spectra of four slit positions in G249+24.
Figure 11.
Sections of the long slit spectra for Position 1. Both [O
III ] λ λ III ] emission resulting in large[O
III ] λ α ratios. The arrows indicate the sectionwhere the P1 spectrum shown in Figure 10 was taken.Behind the leading edge of the shock front markedby start of line emissions, H α emission becomes strongwith little or no emission in other lines. The intensityof H α then briefly drops in strength within a distanceof a few arcseconds, then increases for a short distancefollowed by a long, gradual decline. The [O III ] emissionis barely detectable at the shock front where the H α starts but substantially increases some 10-20 arcsecondsbehind the shock front to where it dominates the down-stream spectrum. This is consistent with an extendedpost-shock cooling zone. The presence of such strong[O III ] emission behind the shock front suggests shockvelocities at least 100 km s − are present in certain fil-aments. In contrast, spectra at P2, P3, and P4 showedno appreciable [O III ] emission.Similar 2D emission structures were observed in theGalactic Halo SNRs G70.0-21.5 (Raymond et al. 2020)and G107.0+9.0 (Fesen et al. 2020). They can be in-terpreted as emission from a ∼
100 km s − shock inpartially neutral gas. Neutral hydrogen atoms swept upby the shock are quickly ionized, but before that hap-pens, some of them are excited to produce H α . Thisemission region is very thin. It is followed by a thickerionization zone where O II is ionized to O III . As thegas cools, the [O
III ] fades and H α brightens. In thecase of G107.0+9.0, the 10” gap between H α and [O III ]and the ∼
10” wide diffuse patch of [O
III ] emission arewell matched by models of 100 km s − shocks in a gasof density 0.1 cm − at a distance of 1 kpc (Fesen et al.2020). The wider patch of [O III ] diffuse emission at Po-sition 1 of G249+24 suggests a longer cooling length, aswould occur in a slightly faster shock, perhaps 120 kms − . These structures are spatially resolved thanks tothe low density of the Galactic halo and the relativelysmall distances to these SNRs.The flux level of H β was either undetected or too faintat P2, P3, and P4 to accurately measure an H β /H α ratio which could be used to estimate extinction. Onlythe spectrum P1 showed a weak but measurable level ofH β flux. There the measured H α /H β ratio of 2 . ± . α /H β ratio of 2.87 for 10 K we estimatea foreground E ( B − V ) < .
07. However, in view of theweakness of the H β detection, this estimate comes witha large uncertainty. Nonetheless, such a low interstellarextinction is not unexpected due to the remnant’s highGalactic latitude of 24.4 ◦ .If this low extinction estimate is correct, then the lackof [O III ] emission at P2, P3, and P4 suggests a fairlylow shock velocity, namely 70 km s − or less, in contrastto that indicated at P1. Consequently, it appears thata variety of shock speeds, from <
70 to over 120 km s − Figure 12.
A 408 MHz intensity image of the G249+24region from the Haslam et al. (1982) All-Sky survey. appear to be present throughout the nebula’s structure,probably driven by an inhomogeneous local ISM.3.2.4.
Associated Radio and X-Ray Emissions
Examination of several on-line radio maps includingthe NRAO VLA Sky Survey (NVSS; Condon et al. 1998)did not show obvious radio emission in the G249+24’sregion. However, the 1420 MHz Bonn survey (Re-ich 1982) did show some emission along G249+24’sFUV and optical filaments near α (J2000) = 9 h m , δ (J2000) = − ◦ (cid:48) .The 408 MHz all-sky survey (Haslam et al. 1981, 1982)also showed enhanced emission in a roughly sphericalshape approximately 4 ◦ in diameter and centered at 9 h m , − ◦ (cid:48) , a position that is nearly the same indi-cated by FUV and optical images (see Fig. 12). A seem-ingly unrelated broad, vertical emission feature bordersthe emission shell on its eastern side and complicatesthe assessment of this emission. If this emission shell isassociated with the nebula, it would suggest a far morespherical SNR than indicated by either the FUV or H α images.Regarding possible X-ray emission, we note that therecently published low-resolution SRG/eRosita all-skyX-ray image (Predehl et al. 2020) shows a faint shell ofsoft X-ray emission at an usually high Galactic latitudein a location consistent with it possibly being associated The Exceptionally Large Suspected Antlia SNR
McCullough et al. (2002) reported a discovery of avery large 24 ◦ diameter H α emission shell with interiorROSAT 0.25 keV X-ray emission located in the south-ern hemisphere and at a high Galactic latitude of +19 ◦ .They proposed it to be a previously unrecognized SNRand named it Antlia for the constellation it lies in. Mc-Cullough et al. (2002) suggested it was an extremely oldremnant with an estimated age of ∼ III λ
977 andC IV λλ GALEX FUV and MDW H α Images
In the top panel of Figure 13, we present a mosaic ofSHASSA H α images of the Antlia remnant at a higherresolution than the VTTS H α image (Finkbeiner 2003)which led McCullough et al. (2002) to discover it. Itshows a ∼ ◦ × ◦ H α emission shell with a well de-termined boundary roughly centered at α (J2000) = 10 h m , δ (J2000) = − ◦ (cid:48) corresponding to Galactic co-ordinates l = 275 . ◦ b = +18 . ◦ . (Note: These Galacticcoordinates differ slightly from those given by McCul-lough et al. 2002.)A mosaic of FUV GALEX images covering just theremnant’s northern limb is shown in the lower panel ofFigure 13. This image reveals a long and nearly contin-uous set of thin, bright FUV emission filaments and fila-ment clusters extending approximately 20 degrees alongthe remnant’s northeastern limb. The line of FUV fil-aments starts at RA = 10 h m and Dec = − ◦ andextends southward to the bottom left of the figure atRA = 11 h m and Dec = − ◦ .Because the morphology of Antlia’s FUV filaments issimilar to that seen in the proposed SNRs G354-33 andG249+24, we examined MDW H α images of the FUV fil-aments and adjacent regions above the − ◦ Declinationlimit of the MDW survey. Overall, we found excellentpositional agreement between FUV and H α filaments.To illustrate this agreement Figure 14 shows a com-parison of GALEX FUV and MDW survey H α imagesfor three sections along the Antlia’s northeastern bound- ary. While we find there is a close correlation of theouter FUV filaments with many H α filaments, there arealso considerable differences in terms of the nebula’s in-ternal emissions. That is, far more H α emission is seen‘behind’, i.e., to the west, of the sharp, outlying fila-ments than seen in the FUV images. This is most strik-ing in the lower panel images of Figure 13 for the rem-nant’s southeastern limb region. While the SHASSA im-age shows considerable internal diffuse and filamentaryemission, the brightest H α emissions are found alongAntlia’s northwestern, western and southeast limbs.3.3.2. Optical Spectra
If McCullough et al. (2002) is correct in their assess-ment that the Antlia H α nebula is a SNR, as seem-ingly supported by the GALEX FUV and H α imagesdescribed above, the remnant would then be both manytimes larger than any confirmed SNR and be locatedat an unusually high Galactic latitude. On the otherhand, considering its location immediately adjacent tothe enormous Gum Nebula with its extended emission,outer ‘filamentary wisps’, and numerous H I wind blownbubbles (Brandt et al. 1976; Chanot & Sivan 1983; Pur-cell et al. 2015), there is real uncertainty about its trueorigin.Optical spectra are a powerful tool for confirming thepresence of shock emission and hence can be used to in-vestigate the SNR origin of optical nebulosities. Conse-quently, we obtained long-slit, low-dispersion spectra ofseven filaments and emission clumps distributed acrossthe whole of the Antlia nebula to explore the nature ofits optical emissions.Figure 15 shows the location of the seven filaments ob-served in the Antlia remnant, along with the resultingspectra. Table 2 lists the relative emission line strengthsuncorrected for extinction. Listed relative line strengthsare believed accurate to 10%. Based on the observedH α /H β ratios of 3.0 to 4.30, we find a variable degree ofextinction across the Antlia nebula, with E ( B − V ) val-ues ranging from 0.0 to 0.35 assuming an intrinsic ratioof 3.0 and an R value of 3.1. Here we have chosen anH α /H β value greater than the theoretical value of 2.87for 10 K due to the likelihood of significant collisionalexcitation of the n = 3 level at postshock temperaturesseen in SNRs. Such a range of extinction values is notunexpected in view of the nebula’s large dimensions.These optical spectra show clear evidence that theseven filaments observed distributed across the wholeof the Antlia nebula exhibit line ratios indicative ofshock emissions. In fact, emissions at Positions 4 and6, located along Antlia’s southeastern limb display non-radiative, pure Balmer line emission. Such shock spec-tra are usually seen in situations where the shock ve- Note: The [O I ] emission seen in the spectrum for Position 4 isresidual imperfect [O I ] sky emission. Figure 13.
T op : Continuum subtracted SHASSA H α image mosaic of the Antlia remnant. Bottom : GALEX FUV mosaicshowing a nearly unbroken line of UV emission filaments along Antlia’s northern and eastern boundary. Figure 14.
Comparison of GALEX FUV emission vs. MDW H α image mosaic of the northern, eastern, and southeasternportions of the Antlia SNR showing a close agreement of UV and H α emission filaments along its eastern boundary. There isconsiderable H α emission but little FUV emission in the nebula’s interior. P1 P2 P3 P4 Rest Wavelength (Å) F l u x ( e r g c m s Å ) P6 P7 P8 Rest Wavelength (Å) F l u x ( e r g c m s Å ) Figure 15.
Top: Continuum subtracted SHASSA H α image mosaic of the Antlia remnant showing the locations of sevenpositions where SALT spectra were obtained. Bottom panels: Observed optical spectra at these seven positions. >
500 km s − ) as seen in youngType Ia SNRs like Tycho and SN 1006. However, similarBalmer dominated spectra has also been seen in mucholder remnants (G107.+9.0; Fesen et al. 2020). Rela-tively slow shocks can remain non-radiative if the am-bient density is sufficiently low, like that expected inthe Galactic halo. This requires that the cooling timeof the shocked gas should be long enough to preventthe forbidden lines from becoming bright. Scaling thecooling times of Hartigan et al. (1987) with n − , thenon-radiative shocks at the edge of a 100,000 year oldremnant in a density of 0.1 cm − would be at least 170km s − .The other five filaments show optical spectra com-monly seen in evolved SNRs like the Cygnus Loop andIC 443. In all these cases, the [S II ] λ λ III ] λλ III ] emission indicates shockvelocities less than about 90 km s − . The high H α /H β ratio in Position 1 of the Antlia remnant suggests a sub-stantial contribution of collisional excitation at fairly lowtemperatures to the Balmer lines, and this is borne outby the low [N II] and [S II] to H α ratios. That suggestsshock speeds around 70 km s − . Given these slow shockspectra along with pure Balmer emission spectra alongthe remnant’s leading edges, there must be a consider-able range of shock velocities present in the nebula. Thisis not too surprising given the large size of Antlia and anearly 18 ◦ range in Galactic latitude. DISCUSSIONBased on the data presented above, we find thattwo high latitude suspected SNRs, namely G354-33 andAntlia, are likely bona fide SNRs. Moreover, their UVemission filaments appear best at marking these objects’forward shock front locations. Below we discuss sup-porting evidence for our conclusions, followed by esti-mates regarding their true physical dimensions and gen-eral properties. We then briefly comment of the useful-ness of far UV imaging for SNR detection particularlyin areas far off the Galactic plane.4.1.
G353.9-33.4
There are several observations that support the SNRidentification of this large FUV, H α , and radio nebula.At its Galactic latitude of − . ◦ , there are no brightand nearby early type stars near the center of this UVshell that might have generated its UV emission shellthrough stellar winds. There is also no known nearbyrecurrent nova projected inside, and the emission shellis at least two orders of magnitude larger and does nothave a similar appearance to known nova shells.Then there is the large 1420 MHz radio emission shellroughly centered on and lying entirely within the bor- ders of the FUV emission filaments. The FUV filamentsalso exhibit a shock-like morphology such as commonlyseen in Galactic SNRs. Furthermore, its nearly continu-ous ring of thin filaments surrounds a broad, diffuse H α emission shell. So in summary, its shock-like filamentaryappearance and the positional agreement between UV,H α and radio emissions for such a large object locatedfar off the Galactic plane leaves few viable options otherthan a SNR origin.While we have not yet been able to obtain opticalspectra of its H α and UV filaments to test this con-clusion, we suspect this remnant may resemble that ofthe high-latitude remnant, G65.3+5.7 (Gull et al. 1977;Boumis et al. 2004). That remnant also exhibits strongfar UV emissions (Kim et al. 2010) with few H α brightfilaments across the whole of its 3 ◦ × ◦ shell. Becausemost of the G65.3+5.7 filaments are brightest in [O III ]line emission, if G354-33 is similar, then its spectrummight also show strong [O
III ] line emissions and hencebe most apparent in wide-angle [O
III ] imaging.Centered at a Galactic latitude around − . ◦ , G354-33 ranks as the highest Galactic latitude SNR foundyet. In addition, if not for the Antlia remnant, thisobject would also rank first in angular dimensions amongthe roughly 300 SNRs in the Green (2019) catalogue.Because of its ∼ ◦ × ◦ size and the fact that thelargest known SNRs have physical dimensions ∼
100 pc,we can make some crude estimates of its distance andproperties.As shown in Table 1, if its diameter is taken to be100 pc, then its distance is around 400 pc. Based onthe presence of its stronger FUV than NUV emission,its shock velocity is likely greater than 100 km s − butprobably less than 150 km s − , according to Figure 4 ofBracco et al. (2020). That shock speed is too high forthis remnant to be in the snowplow phase of SNR evolu-tion, but the presence of radiative shock waves aroundmuch of the rim indicates that it is no longer in the Se-dov phase, at least in those places. That suggests that itis in the pressure-driven shell phase, when the recentlyshocked gas has cooled, but there is still relatively hot,high-pressure gas in the interior.From the calculations of Cioffi et al. (1988), this veloc-ity range and a diameter of around 90 pc would be con-sistent with an age (cid:39) years and a preshock density of0.1 cm − (see Fig. 14 of Fesen et al. 2020). One mightexpect detectable X-ray emission during the pressure-driven shell phase, but Shelton (1998) has discussed haloSNRs in which the interior gas is too cool to produce X-rays, but still rich in high ionization states such as O VI .4.2. G249.2+24.4
The evidence for identifying this nebula as a SNR iscompelling. Both GALEX FUV and SHASSA H α im-ages display a highly filamentary morphology like thatseen in SNRs, and optical spectra that show emissionline ratios consistent with the presence of shocks. The8 Table 2.
Observed Emission Line Fluxes for Antlia FilamentsEmission Line Filament Position(˚A) P1 P2 P3 P4 P6 P7 P8H β III ] 5007 < < < < < < < I ] 5200 · · ·
23 29 · · · · · · · · · [O I ] 6300 · · ·
93 98 · · · · · ·
101 122[N II ] 6548 57 89 97 · · · · · ·
137 181H α II ] 6583 167 247 275 · · · · · ·
403 542[S II ] 6716 146 239 247 · · · · · ·
406 442[S II ] 6731 91 167 184 · · · · · ·
283 307F([S II ])/F(H α ) 0.55 1.35 1.27 · · · · · · II ] 6716/6731 1.60 1.43 1.34 · · · · · · E ( B − V ) 0.35 0.00 0.12 0.23 0.20 0.01 0.19F(H α ) erg cm − s − Figure 16.
Left : VTTS H α image of the Galactic plane around the Gum Nebula with the location of the Antlia SNR marked. Right : The continuum subtracted SHASSA H α image of the Antlia remnant showing the presence of bright filaments at theoverlapping Antlia and Gum Nebula region suggestive of physical interaction. added presence of coincident radio and possibly X-rayemissions leaves little doubt that this object is likely atrue SNR.With a size of 2 . ◦ × . ◦ , this object exhibits angu-lar dimensions among the largest of previously knownSNRs. In addition, it being situated at Galactic lati-tude of over 24 degrees, G249+24 lies farther from theMilky Way’s plane than any other confirmed SNR – thatis, other than G354-33.Although we do not know its distance, we can makeestimates of its shock velocities based on our opticalspectra which, in turn, can be used to constrain both itsdistance and age. Based on its optical emissions, along with the strength of [O III ] emission and the presence ofBalmer dominated emission detected at slit P1, shockvelocities of at least 70 −
100 km s − appear present inthe remnant. Following the discussion above for G354-33, we estimate a radius ∼
30 - 40 pc assuming E o = 1 × erg, n o = 1 cm − , and a blast velocity of100 - 150 km s − . Adopting a radius of 35 pc meansG249+24’s angular diameter of 4.2 degrees suggests adistance around 1 kpc. As with G354-33, this SNR islikely to be in the pressure-driven shell phase. Its phys-ical size is probably smaller and the shock speed larger,so Figure 14 of Fesen et al. (2020) suggests an age closerto 80 × years.94.3. The Nature of the Antlia Nebula
McCullough et al. (2002) claimed this extraordinar-ily large emission nebula was a likely SNR based on itsappearance on the deep H α image of the VTTS surveyand on the presence of diffuse soft X-ray emission in itsinterior. However, this conclusion does not appear tobe widely accepted, as measured by the remnant havingattracted little subsequent attention.This situation might in part be due to a reluctanceby SNR researchers to accept its huge 20 ◦ × ◦ size,more than 3-5 times larger than the largest known con-firmed SNRs, plus its location so close to the even largerGum Nebula with its complex of large emission shellsand wind-blown bubbles and the huge Vela SNR. Con-sequently, except for a far UV study by Shinn et al.(2007), the Antlia remnant has not yet be studied inany detail, leaving open its true nature.Our GALEX FUV mosiacs show a well-defined shellin H α with many individual and overlapping filamentsthat closely resemble shocks. In addition, the locationsof these sharp UV filaments along the boundary of theobject’s H α emission are consistent with a SNR wheresuch UV filaments mark the location of the remnant’sshock front.Importantly, results of our seven optical spectra ofAntlia’s filaments argue for shock emissions throughoutthe nebula thereby leading to it being a SNR. Thesespectra include two textbook cases of non-radiativeBalmer dominated spectra (Positions 4 & 6), plus severalother filaments exhibiting high [S II ]/H α line ratios wellabove the 0.4 value distinguishing shocked from pho-toionized nebulae. If the Antlia remnant was not solarge, it would be a easy case for supernova remnantclassification.Unfortunately, the physical size of the Antlia SNR isunknown due to its unknown distance. McCulloughet al. (2002) estimated a wide range of possible dis-tances, from 60 pc to 320 pc, and believed the remnantto be extremely old, at least 1 Myr, and hence in thefinal snowplow phase of SNR evolution. However, ouroptical spectra do not support such an old object or adistance less than 200 pc. A 1 Myr old SNR is expectedto have a a very low expansion velocity of around 10 to20 km s − (Chevalier 1977).However, our optical spectra show line emissions in-dicating shock velocities of 50 to 150 km s − , making itfar younger ( < yr) and in the pressure-driven shellphase. Moreover, there are additional data supportinghigh-velocity gas inside a much younger Antlia remnant.Before the Antlia nebula was discovered, Bajaja et al.(1989) found high-velocity clouds in the remnant’s direc-tion, and Penprase & Blades (1992) reported detectinghigh-velocity Ca II absorption lines in the spectrum ofthe B9/A0 III/IV star HD 93721 which lies in the di-rection of the Antlia SNR (see Fig. 15). This star hasa estimated Gaia Early Data release distance of 512 ± II absorption components with v lsr ranging from −
65 kms − to +75 km s − . Because a second star, HD 94724(d = 210 pc) lying in the same general direction didnot show any high-velocity components, they concludedthat the high-velocity absorbing cloud’s distance was be-tween 200 and 500 pc. However, because not all starsbehind a SNR display high-velocity absorption lines, alack of high-velocity components does not provide a ro-bust minimum distance estimate.An alternative means of estimating the distance tothe Antlia remnant is its apparent collision with the 36 ◦ diameter Gum Nebula (Gum 1952, 1955; Brandt et al.1971; Sivan 1974; Chanot & Sivan 1983). Although thecoincidence of the remnant’s southwestern rim with theGum Nebula’s bright northeastern emission cloud hasbeen noted in a few studies of the Gum Nebula (Iacobelliet al. 2014; Purcell et al. 2015), no mention has beenmade about a physical interaction between the two.However, our SHASSA imaging mosaic strongly indi-cates such collision has occured between the Antlia rem-nant and the Gum Nebula. To begin with, the left panelof Figure 16 presents a section of the VTTS H α imageof the Galactic plane showing the Antlia emission shell,similar to the VTTS image presented in the McCulloughet al. (2002) Antlia discovery paper. This shows Antlia’sposition relative to the Gum Nebula. The remnant’sbright southern limb coincides with the northern rim ofthe Gum Nebula.The right panel of Figure 16 shows a low contrast ver-sion our SHASSA image of Antlia. A line of long, brightfilaments can be seen in the projected overlap region ofboth shells (see also Fig. 15). The simplest explana-tion for such bright filaments situated only in the exactAntlia/Gum overlap region is that the Antlia remnanthas collided with the outer northeastern rim of the GumNebula.If that is correct, then the Antlia’s distance is roughlythe same as for the Gum Nebula. Unfortunately, the dis-tance to the Gum Nebula is poorly known, with valuesranging from 200 −
500 pc (Brandt et al. 1971; Woer-mann et al. 2001; Howarth & van Leeuwen 2019). How-ever, at the least, distances below 200 pc like suggestedby McCullough et al. (2002) would appear to be ruledout. If we adopt a distance of 300 pc, like that estimatedfor some of the Gum Nebula’s ionizing stars (Howarth &van Leeuwen 2019), then, the Antlia remnant’s physicaldiameter is ∼
120 pc.4.4.
The Power of FUV Emissions for Finding SNRs
A few large Galactic SNRs and supperbubbles havebeen recently studied in in terms of their far UV emis-sions. For example, studies of far UV emissions havebeen reported for the Vela SNR (Nishikida et al. 2006),the Cygnus Loop (Seon et al. 2006), the Lupus Loop(Shinn et al. 2006), and the Orion-Eridanus Superbub-ble (Kregenow et al. 2006). Many of these made use0
Figure 17.
Comparison of MDW H α and GALEX FUV images of the high-latitude remnant G70.0-21.5. North is up, East tothe left. of the SPEAR imaging spectrograph (Edelstein et al.2006).However, there have been few papers reporting discov-eries of large interstellar emission structures using FUVemissions. One such paper is that of Bracco et al. (2020)who reported finding a 30 ◦ long UV arc in Ursa Majorusing GALEX images. That work was a follow-up toan earlier detection of a much shorter 2.5 ◦ filament byMcCullough & Benjamin (2001) using deep H α imag-ing. However, only through the GALEX’s FUV imageswas the full extent of this faint, long interstellar filamentfinally revealed.The FUV emission of moderate velocity shocks is dom-inated by the hydrogen 2-photon continuum, resonancelines of C IV and Mg II, and intercombination lines suchas C III] and Si III]. Bracco et al. (2020) computed theGALEX FUV and NUV count rates for shocks fromthe Sutherland & Dopita (2017) MAPPINGS models.Shocks slower than about 100 km s − are dominated bythe 2-photon continuum, while faster shocks have strongcontributions from C IV and other lines. The predictedratios range from about 0.1 to 0.9. Based on our ownmodel calculations, a reddening E ( B − V ) ∼ α emissions.Both our investigations and that of Bracco et al.(2020) indicate that broad, far UV imaging can be anespecially useful means for detecting and distinguishinginterstellar shocks and, in some cases, is more sensitivecompared to H α imaging. As noted by Bracco et al.(2020), both line emissions and two photon continuumemissions contribute to FUV emission in shocks withvelocities above (cid:39)
50 km s − .To illustrate the usefulness of far UV emission imag-ing for detecting SNRs, Bracco et al. (2020) noted thepresence of networks of thin FUV filaments in both theAntlia SNR and the recently discovered Galactic rem-nant G70.0-21.5 (Boumis et al. 2002; Fesen et al. 2015;Raymond et al. 2020). In Figure 17 we show a compar-ison of MDW’s H α image and the GALEX FUV imageof G70.0-21.5. Until the study of the SNRs discussedhere, this remnant at b = − . ◦ had been the remnantwith the highest Galactic latitude. This figure amplydemonstrates how that the FUV image makes the rem-nant much easier to detect and helps to define its fulldimensions despite the many missing individual GALEXimages. Knowing about the existence of these GALEXFUV images could have helped Fesen et al. (2015) andRaymond et al. (2020) in their analysis of this remnantin regard to the remnant’s true physical size. In sum-mary therefore, far UV images appear to be an espe-cially useful tool for identifying interstellar shocks likethose found in SNRs, albeit though best suited for highGalactic latitude searches. CONCLUSIONS1We have investigated the nature of two large and sus-pected supernova remnants located at unusually highGalactic latitudes through GALEX far UV emission mo-saics and optical images and spectra. This research alsohas uncovered one new Galactic SNR. Our findings in-clude:1) The large ∼ ◦ radio emission shell, G353.9-33.4,seen in 1420 MHz and 1.4 GHz radio polarization mapsis very likely a SNR. The remnant exhibits numeroussharp FUV emission filaments in a thin, unbroken shellwith angular dimensions of 11 ◦ × . ◦ . It also exhibitsa coincident H α emission shell.2) A group of bright, sharp FUV emission filamentsextending some 2 . ◦ × . ◦ in size and coincident withnumerous but faint H α filaments appears to be a previ-ously unrecognized SNR. Optical spectra of several fil-aments show evidence for the presence of 50 - 150 kms − shocks. Deep H α images reveal a highly filamentarymorphology like that seen in evolved SNRs. Coincidentdiffuse 408 MHz radio emission lends additional supportfor its SNR identification.3) Despite its enormous angular dimensions (20 ◦ × ◦ ), GALEX FUV mosaic images, plus wide-angle H α images and optical spectra strongly support a SNR ori-gin for the Antlia nebula. This conclusion is in linewith ±
70 km s − Ca II absorptions in one backgroundstar. We estimate an age ∼ yr, which is an order ofmagnitude less than the earlier estimate of 1 Myr. Wealso find the remnant is in likely physical contact alongits southwestern rim with the Gum Nebula which couldhelp constraint its distance.4) Our investigation of suspected SNR located at un-usually high Galactic latitudes ( > ◦ ) highlights thevalue of UV images to detect interstellar shocks.Follow-up work on these three objects could includeoptical spectra of the FUV and H α filaments of theG354-33 remnant, and wide-field [O III ] λ III ] filaments. Higher resolution H α images and optical spectra of the series of long and verybright filaments seen in southwestern limb of the Antliaremnant could also provide a test of our conclusion re-garding Antlia’s collision with the Gum Nebula.Although we have found that all three nebula in ourstudy are SNRs, none have accurate distance estimates,leaving us with only approximate physical parameters and evolutionary status. This problem can be ad-dressed especially well in the era of accurate Gaia paral-laxes through high-dispersion spectra looking for high-velocity Na I λλ II λ | b | > ◦ range, generally viewed as unproductive.Finally, we note that although far UV imaging appearsto be a sensitive new tool for uncovering the presence ofinterstellar shocks, it is most useful in searching uncom-plicated regions, like the objects discussed here locatedin the Galactic halo. However, the number of similar butunrecognized high latitude Galactic remnants is proba-bly pretty small. Nonetheless, given the dozens of un-confirmed but seemingly likely or suspected SNRs in theliterature (see list in Green 2019’s catalogue), many newSNR discoveries may be aided by the use of UV imaging.We thank Justin Rupert, Eric Galayda and the wholeMDM staff for making the optical observations possible,and the SALT Observatory and Resident Astronomerstaff for obtaining the excellent RSS spectra despite dis-ruptions due to COVID-19 restrictions. We also thankR. Benjamin for helpful discussions. This work made useof the Simbad database, NASA’s Skyview online dataarchives, and the Max Planck Institute for Radio As-tronomy Survey Sampler. This work is part of R.A.F’sArchangel III Research Program at Dartmouth. DMacknowledges support from the National Science Foun-dation from grants PHY-1914448 and AST-2037297.REFERENCES Alikakos, J., Boumis, P., Christopoulou, P. E., & Goudis,C. D. 2012, A&A, 544, A140,doi: 10.1051/0004-6361/201219118Aschenbach, B. 1998, Nature, 396, 141, doi: 10.1038/24103Bajaja, E., Cappa de Nicolau, C. E., Martin, M. C., et al.1989, A&AS, 78, 345 Bianchi, L. 2009, Ap&SS, 320, 11,doi: 10.1007/s10509-008-9761-3Blair, W. P., Kirshner, R. P., & Chevalier, R. A. 1981, ApJ,247, 879, doi: 10.1086/159098Blondin, J. M., Wright, E. B., Borkowski, K. J., &Reynolds, S. P. 1998, ApJ, 500, 342, doi: 10.1086/305708 Boumis, P., Mavromatakis, F., Paleologou, E. V., & Becker,W. 2002, A&A, 396, 225,doi: 10.1051/0004-6361:20021365Boumis, P., Meaburn, J., L´opez, J. A., et al. 2004, A&A,424, 583, doi: 10.1051/0004-6361:20040410Boumis, P., Xilouris, E. M., Alikakos, J., et al. 2009, A&A,499, 789, doi: 10.1051/0004-6361/200811474Bracco, A., Benjamin, R. A., Alves, M. I. R., et al. 2020,A&A, 636, L8, doi: 10.1051/0004-6361/202037975Brandt, J. C., Roosen, R. G., Thompson, J., & Ludden,D. J. 1976, ApJ, 208, 109, doi: 10.1086/154585Brandt, J. C., Stecher, T. P., Crawford, D. L., & Maran,S. P. 1971, ApJL, 163, L99, doi: 10.1086/180676Cha, A. N., & Sembach, K. R. 2000, ApJS, 126, 399,doi: 10.1086/313306Chanot, A., & Sivan, J. P. 1983, A&A, 121, 19Chevalier, R. A. 1977, ARA&A, 15, 175,doi: 10.1146/annurev.aa.15.090177.001135Cioffi, D. F., McKee, C. F., & Bertschinger, E. 1988, ApJ,334, 252, doi: 10.1086/166834Condon, J. J., Cotton, W. D., Greisen, E. W., et al. 1998,AJ, 115, 1693, doi: 10.1086/300337Cornett, R. H., Jenkins, E. B., Bohlin, R. C., et al. 1992,ApJL, 395, L9, doi: 10.1086/186476Cox, D. P., & Smith, B. W. 1974, ApJL, 189, L105,doi: 10.1086/181476Danks, A. C., & Sembach, K. R. 1995, AJ, 109, 2627,doi: 10.1086/117476Dopita, M. A., Binette, L., Dodorico, S., & Benvenuti, P.1984, ApJ, 276, 653, doi: 10.1086/161653Downes, D. 1971, AJ, 76, 305, doi: 10.1086/111124Driessen, L. N., Domˇcek, V., Vink, J., et al. 2018, ApJ,860, 133, doi: 10.3847/1538-4357/aac32eEdelstein, J., Korpela, E. J., Adolfo, J., et al. 2006, ApJL,644, L159, doi: 10.1086/505205Ferrand, G., & Safi-Harb, S. 2012, Advances in SpaceResearch, 49, 1313, doi: 10.1016/j.asr.2012.02.004Fesen, R. A., Blair, W. P., & Kirshner, R. P. 1985, ApJ,292, 29, doi: 10.1086/163130Fesen, R. A., Neustadt, J. M. M., Black, C. S., & Koeppel,A. H. D. 2015, ApJ, 812, 37,doi: 10.1088/0004-637X/812/1/37Fesen, R. A., Neustadt, J. M. M., How, T. G., & Black,C. S. 2019, MNRAS, 486, 4701,doi: 10.1093/mnras/stz1140Fesen, R. A., Weil, K. E., Raymond, J. C., et al. 2020,MNRAS, 498, 5194, doi: 10.1093/mnras/staa2765Finkbeiner, D. P. 2003, ApJS, 146, 407, doi: 10.1086/374411Gaustad, J. E., McCullough, P. R., Rosing, W., & VanBuren, D. 2001, PASP, 113, 1326, doi: 10.1086/323969 Green, D. A. 1984, MNRAS, 209, 449,doi: 10.1093/mnras/209.3.449—. 2004, Bulletin of the Astronomical Society of India, 32,335. https://arxiv.org/abs/astro-ph/0411083—. 2015, MNRAS, 454, 1517, doi: 10.1093/mnras/stv1885—. 2019, Journal of Astrophysics and Astronomy, 40, 36,doi: 10.1007/s12036-019-9601-6Gull, T. R., Kirshner, R. P., & Parker, R. A. R. 1977,ApJL, 215, L69, doi: 10.1086/182480Gum, C. S. 1952, The Observatory, 72, 151—. 1955, MmRAS, 67, 155Hartigan, P., Raymond, J., & Hartmann, L. 1987, ApJ,316, 323, doi: 10.1086/165204Haslam, C. G. T., Klein, U., Salter, C. J., et al. 1981, A&A,100, 209Haslam, C. G. T., Salter, C. J., Stoffel, H., & Wilson,W. E. 1982, A&AS, 47, 1How, T. G., Fesen, R. A., Neustadt, J. M. M., Black, C. S.,& Outters, N. 2018, MNRAS, 478, 1987,doi: 10.1093/mnras/sty1007Howarth, I. D., & van Leeuwen, F. 2019, MNRAS, 484,5350, doi: 10.1093/mnras/stz291Howell, D. A. 2017, Superluminous Supernovae, ed. A. W.Alsabti & P. Murdin, 431,doi: 10.1007/978-3-319-21846-5 41Iacobelli, M., Burkhart, B., Haverkorn, M., et al. 2014,A&A, 566, A5, doi: 10.1051/0004-6361/201322982Jenkins, E. B., Silk, J., & Wallerstein, G. 1976, ApJS, 32,681, doi: 10.1086/190412Kim, I. J., Seon, K. I., Min, K. W., et al. 2010, ApJ, 722,388, doi: 10.1088/0004-637X/722/1/388Kopsacheili, M., Zezas, A., & Leonidaki, I. 2020, MNRAS,491, 889, doi: 10.1093/mnras/stz2594Kregenow, J., Edelstein, J., Korpela, E. J., et al. 2006,ApJL, 644, L167, doi: 10.1086/505196Leonidaki, I., Boumis, P., & Zezas, A. 2013, MNRAS, 429,189, doi: 10.1093/mnras/sts324Long, K. S. 2017, Galactic and Extragalactic Samples ofSupernova Remnants: How They Are Identified andWhat They Tell Us, ed. A. W. Alsabti & P. Murdin,2005, doi: 10.1007/978-3-319-21846-5 90Maggi, P., Haberl, F., Kavanagh, P. J., et al. 2014, A&A,561, A76, doi: 10.1051/0004-6361/201322820Martini, P., Stoll, R., Derwent, M. A., et al. 2011, PASP,123, 187, doi: 10.1086/658357Massey, P., & Gronwall, C. 1990, ApJ, 358, 344,doi: 10.1086/168991McCullough, P. R., & Benjamin, R. A. 2001, AJ, 122, 1500,doi: 10.1086/3220973