GSH 91.5+2-114: A large HI shell in the outer part of the Galaxy
aa r X i v : . [ a s t r o - ph . GA ] N ov Astronomy&Astrophysicsmanuscript no. 15355-astro-ph c (cid:13)
ESO 2018June 5, 2018
GSH 91.5+2 − I shell in the outer part of the Galaxy S. Cichowolski , and S. Pineault , Instituto de Astronom´ıa y F´ısica del Espacio (IAFE), CC 67, Suc. 28, 1428 Buenos Aires, Argentina D´epartement de physique, de g´enie physique et d’optique, Universit´e Laval, Qu´ebec, G1V 0A6 Canada, and Centre de rechercheen astrophysique du Qu´ebec (CRAQ) Instituto Argentino de Radioastronom´ıa (IAR), CC 5, 1894, Villa Elisa, ArgentinaReceived date / Accepted date
ABSTRACT
GSH 91 . + −
114 is a large H I shell located in the outer Galaxy at a kinematic distance of about 15 kpc. It was first identified in theCanadian Galactic Plane Survey (CGPS) by Pineault et al. (2002) as being possibly associated with objects possessing infrared colorswhich indicates strong stellar winds. The H I shell has no obvious continuum counterpart in the CGPS radio images at 408 and 1420MHz or in the IRAS images. We found no evidence for early-type massive stars, most likely as a result of the large extinction thatis expected for this large distance. An analysis of the energetics and of the main physical parameters of the H I shell shows that thisshell is likely the result of the combined action of the stellar winds and supernova explosions of many stars. We investigate whether anumber of slightly extended regions characterized by a thermal radio continuum and located near the periphery of the H I shell couldbe the result of star formation triggered by the expanding shell.
Key words.
ISM: bubbles — ISM: kinematics and dynamics — ISM: structure — supernova remnants
1. Introduction
A massive star possessing a strong stellar wind (SW) can injectas much energy into the interstellar medium (ISM) during itslifetime as it does during the final supernova explosion. If thestellar spatial velocity with respect to the ambient ISM is nottoo large (i.e. <
30 km s − ), the SW is expected to evacuate alarge cavity around the star (called a stellar bubble) surroundedby a ring or shell of enhanced density. These structures havebeen observationally detected as optical and infrared (IR) nebu-lae around Wolf-Rayet (WR) and Of stars ( ?? ).The situation with respect to radio observations, reviewedby ? , has evolved considerably with the advent of large Galacticneutral hydrogen (H I) surveys, namely the Canadian GalacticPlane Survey (CGPS; ? ), the Southern GPS (SGPS; ?? ) and theVLA (VGPS; ? ). These surveys at ∼ ???? ). Interestingly enough, the largemajority of shells detected by their neutral hydrogen emissionhave low inferred expansion velocities, typically less than or onthe order of 10 km s − ( ? ).In parallel to these developments on the observational scene,recent theoretical studies, building up on the initial work of ? andothers, have considerably increased our understanding of the in-teraction of stellar winds with their surrounding ISM. The e ff ectsof the di ff erent evolutionary phases and of a large peculiar mo-tion of the star have been modeled in detail in a number of newstudies (e.g., ???? , and references therein). Concerning the oftenobserved low expansion velocities of ∼
10 km s − , ? have shownthat velocity dispersion within the shell and the role of the localISM background may significantly a ff ect the appearance of anexpanding H I shell in velocity space and thus its inferred pa-rameters (in particular, mass and expansion velocity). In the case of moderate-size shells, a puzzling aspect is theapparent lack of a radio continuum counterpart or of a candi-date progenitor star ( ???? ). This suggests that either one or sev-eral basic ingredients are missing in the predictions of the theoryand / or that some detected H I shells are not real.Increasing the sample of well studied H I shells is a first natu-ral step in elucidating some of the current puzzles. In an attemptat diversifying the sample of known H I shells (consisting mostlyof shells around objects known for their optically interesting fea-tures), ? initiated a project aimed at discovering new SW sourcecandidates by using first IRAS colors to extract potential can-didates and then the CGPS database to look for a morphologyindicative of a SW, i.e. shells, rings, bubbles, cavities, or voids.An obvious advantage of this procedure is that potential candi-dates su ff er much less from the selection e ff ects associated withoptically chosen targets, for example, WR or Of stars, the distri-bution of which is severely biased by absorption. A positionalcoincidence between one or more candidates and a shell-likemorphological structure is nevertheless not a proof of a phys-ical association, and a detailed analysis is required before anyfirm conclusion can be drawn.In this paper, we focus our attention on a very large (nearly1 . ◦ l, b ) = (91 . ◦ + . ◦
0) and observed at a velocity v lsr ≈ −
114 km s − , whichwas identified by ? . A flat rotation curve model for the Galaxygives for the shell a distance d =
15 kpc, a galactocentric radius R g =
17 kpc, a distance z ∼
525 pc from the Galactic plane, anda diameter D =
400 pc, placing it in the outermost part of theGalaxy.This shell thus o ff ers the opportunity to explore the envi-ronment of remote regions of the Galaxy where many physi- All velocities are with respect to the local standard of rest (lsr) S. Cichowolski & S. Pineault: A large H I shell in the outer part of the Galaxy
Table 1.
Observational parameters for the H I data
Parameter ValueSynthesized beam 1 . ′ × . ′ − ) 0.824Velocity resolution (km s − ) 1.32Velocity coverage (km s − ) 224 cal parameters such as metallicity, density, smaller or negligi-ble perturbations from spiral arms, greatly di ff er from those innearby regions of the Galaxy. Despite this remoteness, there isevidence that star formation in the outer Galaxy may be com-mon, as shown by the discovery of a considerable number ofembedded star clusters in molecular clouds up to galactocen-tric radii R g ≈
17 kpc ( ?? ). A particularly interesting caseis the suggestion by ? that star formation in Digel’s Cloud 2( R g ≈
19 kpc) could have been triggered by the huge supernovaremnant GSH 138-01-94 previously discovered by ? . These hugeshells imply that massive stars form in the outer Galaxy, empha-sizing the importance of studying these objects. At these largedistances, optical obscuration is very severe, so that one has toresort to radio or infrared observations.The plan of the paper is as follows. In Section 2 we describethe observational data used, in Section 3 we briefly review theinitial data used by ? and present detailed neutral hydrogen (H I)and continuum images of the object. The results are analyzedand discussed in Section 4. Section 5 is a summary of the mainconclusions.
2. Observational data
Radio continuum data at 408 and 1420 MHz and 21-cm spectralline data were obtained at DRAO as part of the CGPS survey( ? ). A detailed description of the data processing routines canbe found in ? . At the position of GSH 91.5 + − . ′ × . ′
82 and 3 . ′ × . ′ ? ) IRAS imagesproduced at the Infrared Processing and Analysis Center (IPAC) were also used. The images used are the result of 20 iterationsof the algorithm, giving an approximate resolution ranging fromabout 0 . ′ ′ . At the position of GSH 91.5 + − ′ .
3. Results
Figure 1 shows an H I image averaged between v = − . − . − , showing the large H I shell discovered by ? withthe four IRAS sources which they suggested might be physicallyassociated.The interior of the shell seems to have been entirely clearedof neutral hydrogen. Though the shell is generally quite well de-fined, its northern part is essentially absent, or at least it does notshow a clear outline. Indeed, the general morphology suggeststhat the shell is open to the north in a direction away from the The Infrared Processing and Analysis Center (IPAC) is funded byNASA as part of the
Infrared Astronomical Satellite (IRAS) extendedmission under contract to the Jet Propulsion Laboratory (JPL) BA Fig. 1.
H I emission distribution averaged between − . − . − . The plus symbols indi-cate the position of the four DWCL source candidates,IRAS 21147 + + + + + + + l, b ) = (92 . ◦
1, 0 . ◦
9) and lo-cated on the larger shell structure. The inset in the bottom rightcorner of Fig. 1 shows this region in more detail. Table 2 showsthe IR colors of these four sources.A comparison of these colors with the values given by ? inhis Table 4 allows us to conclude that the IRAS sources areprobably dust shells related to WC8-9 stars (dubbed dusty late-type WR or DWCL by Cohen). In their preliminary analysis ofthe then-incomplete CGPS, ? had concluded that the surface dis-tribution of the stellar wind candidates chosen on the basis oftheir IR colors was on the average of 1 source per 10 squaredegree area. Note that there are no other sources with the sameproperties in the entire field shown in Fig. 1. The position ofthe four candidate sources inside or near the boundary of the H Ishell prompted a more detailed analysis. Figure 2 shows the H I distribution in the LSR velocity rangefrom − . − . − . The original images were Cohen’s (1995) definitions are of the form [ i ] − [ j ] = k ij + . F j / F i ) where the constants k ij are 1.56 and 1.88 for the colors[12] − [25] and [25] − [60], respectively.. Cichowolski & S. Pineault: A large H I shell in the outer part of the Galaxy 3 Table 2.
Parameters of the DWCL source candidates
IRAS source l b C C F ir (Jy) L ir (10 L ⊙ )21147 + . ◦
05 0 . ◦ − .
34 0.21 9 . ± . . ± . + . ◦
56 1 . ◦ − .
32 0.02 11 . ± . . ± . + . ◦
25 1 . ◦ − .
34 0.16 9 . ± . . ± . + . ◦
34 2 . ◦ − . − .
33 27 . ± . . ± . C ij = log( F j / F i ), where i and j run from 1 to 4 and correspond to 12, 25, 60 and 100 µ m, respectively and F i is the flux density (Jy) in band i . L ir / L ⊙ = . F ir d , where the integrated flux in Jy is F ir = . F + F ) + . F + F ) + . F + F ) ( ? ). Subscripts are wavelengthsin microns . Luminosity values are for d =
15 kpc. smoothed to a 3-arcmin resolution to increase the signal-to-noiseratio. At v = − . − the shell, which we shall now referto as GSH 91.5 + − v = − . − . This structure is centered at ( l, b ) = (91 . ◦ + ◦ ) and has an angular size of about 1 . ◦ − could form part of the so-called receding cap, im-plying an expansion velocity of some 7 km s − . In an ideal situ-ation, one would then expect to find an approaching cap at a ve-locity of about –121 km s − . No obvious excess H I emission isseen near this velocity. The lack of confusing emission at this ve-locity suggests that either expansion on the near side took placein an extremely low-density medium, making its detection be-low the sensitivity of the telescope, or that the shell was veryincomplete and nearly absent on the near side. ? had already noted that no continuum (radio or IR) counterpartseemed to be present. Figure 3 shows the CGPS 1420 MHz radiocontinuum and 60 µ m HIRES images of the same field of view asthe preceding figures. Discrete point sources have been removedfrom the radio continuum image and the shadings chosen so as tohighlight the low-level emission centered on GSH 91.5 + − + − l , b ) = (90 . ◦ , . ◦
6) is the H IIcomplex BG 2107 +
49 discussed by ? and located at a kinemat-ical distance of 10 kpc. The circles delineate the approximateinner and outer boundaries of the H I shell, revealing that thereis no obvious radio continuum or infrared emission related toGSH 91.5 + − ff use emission is more or less distributed ho-mogeneously over the northern part of the image, although thereappears to be a slight excess of emission within the shell bound-ary, in particular in its northernmost part. In the south, the loca-tion where radio emission becomes fainter corresponds roughlywith the H I shell boundary.Two interesting and slightly extended (6 ′ − ′ ) structures(named G91.56 + + ? , who concluded that all polar- ization structures observable in the CGPS are generated and / orFaraday rotated closer than 2.5 kpc.
4. Discussion
Some of the questions we shall try to answer are: can the ob-served H I shell be a stellar wind shell formed by the windsfrom one or more of the four candidate DWCL sources? Ifnot, what are the alternative shell-formation scenarios? Wasthe smaller H I cavity (the one embedded in the wall of thelarger H I shell) caused by IRAS 21147 + + ff erent possible formation sce-narios based on the determined size, mass, age, and energy in-volved. From here on we assume a distance of 15 ± + − ? .We note however that this distance is likely an upper limit.Indeed, using a new method ( ? ) based on H I column densi-ties for the determination of distances within the disk of theGalaxy, ? obtained a new distance for the outer Galaxy H IIregion CTB 102. As this object is at l = . ◦ b = . ◦ ? ’s Fig. 5 to extrapolate their velocity-distance rela-tion to obtain a distance estimate of approximately 11.5 kpc forGSH 91.5 + − R g ≈ . z ∼
400 pc. Whenever possible, we shall show the distance depen-dence of simple parameters by defining d = d /
15 kpc.
Error estimates for all the parameters are based on our abilityto estimate the spectral extent of the shell in velocity space, itsspatial extent, and distance.The H I ring is clearly observed over some 14 km s − .Following the procedure described by ? , we derive a mass forthe shell of M sh = (2 . ± . × d m ⊙ . Note that thisvalue agrees with that obtained using ? equation for the massswept up by a shell, namely M sh ( m ⊙ ) = . R (pc) which, tak-ing R sh = (200 ± d pc, gives M sh = . × d m ⊙ . Usingour determined value for the mass and considering a sphericalcavity, we obtain for the initial (i.e. before the gas was swept upin the shell) neutral gas particle density n o = (0 . ± . d − cm − .Adopting an expansion velocity equal to half the velocityinterval where the structure is observed, v exp = ± − ,we obtain for the kinetic energy stored in the expanding shell E kin = M sh v / = (1 . ± . × d erg. The kinematic ageof the shell is given by t (Myr) = α R sh (pc) / v exp (km s − ), where α = .
25 for a radiative SNR shell and α = . t = (7 ± d Myr ( α = .
25) and t = (17 ± d Myr
S. Cichowolski & S. Pineault: A large H I shell in the outer part of the Galaxy
Fig. 2.
CGPS gray-scale images showing the H I distribution in the LSR velocity range from − . − . − . Velocityresolution is 1.3 km s − , and spatial resolution is 3 ′ . The LSR central velocity of each image is indicated in the top left corner. Thestar symbols indicate the position of the four DWCL source candidates. For all images, darker shading indicates higher brightness.( α = . ff erent formationscenarios for GSH 91.5 + − In this section we assume for simplicity that the DWCL objectsare located at the same distance as GSH 91.5 + − The long kinematic timescale of 17 Myr, if GSH 91.5 + − / or that contributions from oneor more SN explosion are to be expected. Nevertheless, as a firststep, it is of interest to estimate the contribution that the winds . Cichowolski & S. Pineault: A large H I shell in the outer part of the Galaxy 5 Fig. 2. cont. of one or more of the DWCL candidates could make to the for-mation of this large H I shell.We first calculated the total integrated IR luminosity of eachcandidate. The derived parameters are shown in Table 2. If, fol-lowing ? , we assume the objects to be WC8-9 stars, then theirtotal luminosity is about ( ? ) L / L ⊙ = . × . The IR luminosi-ties thus represent only about from 3 to 10 % of the total stellarluminosity.Ignoring any intrinsic absorption by dust surroundingDWCL stars, we may ask whether a WC8-9 star would be eas-ily identified at a distance of 15 kpc along the line of sight toGSH 91.5 + − . × cm − , which corresponds to a reddening of E B − V = . E B − V = N H I / . × cm − ; ? ). This gives avisual absorption A v of 7.3 mag. Taking for the absolute magni-tude of WC8-9 stars ( ? ) M v = − .
5, we obtain m v = .
7. A starof this magnitude would clearly not stand out among field stars,although it would be detectable. In their study of GSH 90 + ? had estimated a visual extinction of 6.4 mag toward l ∼ ◦ andconcluded that deep measurements were needed to detect early-type stars at a distance, for their object, of 13 kpc. S. Cichowolski & S. Pineault: A large H I shell in the outer part of the Galaxy
Fig. 3. µ m images from the CGPS survey. Resolution for both is about 1 ′ . Point-like sourceshave been removed from the radio continuum image. Circles indicate the approximate inner and outer boundary of the H I shell seenin Figure 1. The two insets at the bottom of the radio image show a close-up view of G91.56 + + ? ) to produce a H − K s vs J − H diagram (Figure 4). There are 1573 2MASS sources ina circular area centered at ( l , b ) = ( 91 . ◦
7, 1 . ◦
83) within a radius of400 ′′ . These sources are indicated by green dots in Fig. 4. Thepositions of the dereddened early-type main sequence and gi-ant stars are indicated with blue and red solid lines, respectively.The blue and red dashed lines show the reddening curve for O9V and M0 III stars, respectively. Thus, normally reddened mainsequence stars lie between the two dashed lines. Because absorp-tion is proportional to distance, we can infer that statistically themost distant sources have a visual absorption of about 13 magalong this direction in the Galaxy. Hence, as we are dealing witha structure located in the outer part of the Galaxy at about 15kpc, it is likely that the visual absorption is significantly higherthan 7.3 mag. If absorption is as high as 13 mag, the expected ap-parent magnitude of WC8-9 stars could be as faint as m v = . E w = ˙ Mv w t /
2. Adopting ( ? ) for a single WC8-9 a mass lossrate of ˙ M = − m ⊙ yr − , appropriate for a solar metallic-ity Z ⊙ , and a wind velocity of v w = − , we obtain E w = . × t (yr) erg. If the WC phase lasts ∼ yr, eachstar would impart E w = . × erg to its local ISM duringthis phase. But only a fraction ǫ of this energy gets transferredto the gas. According to evolutionary models of interstellar bub-bles, the expected energy conversion e ffi ciency ǫ = E kin / E w is on the order of 0.2 or less ( ? ). Observationally, the values of ǫ derived from optical and radio observations can be as low as 0.02( ? ), indicating that there are cases where severe energetic lossesmay occur.Furthermore, given the location of GSH 91.5 + −
114 in theouter Galaxy, the e ff ect of metallicity should be taken into ac-count. ? suggested that the mass-loss rate depends on metallicity Z as ˙ M ∝ Z m , with m = .
5. The velocity of the wind is alsolower at lower metallicity ( ? ). ? concluded that the wind energy E w decreases by a factor of about 3 when Z decreases by a factorof 10 for stars with the same luminosity. Thus, for Z = . Z ⊙ ,each WC8-9 star would simply inject E w ∼ . × erg inthe ISM during the WC phase. This implies that even with arelatively high e ffi ciency ǫ = .
2, the injected kinetic energy isonly E kin = ǫ E w = . × erg. Summarizing, neither for Z = Z ⊙ nor for Z = . Z ⊙ could the H I shell have been cre-ated only by the stellar winds of the DWCL sources in their WCphase. Even by considering a comparable contribution from theprevious evolutionary phase of each star, the available kinetic en-ergy falls considerably short of the observed shell kinetic energy E kin = (1 . ± . × d erg. We have failed in finding any massive star in the area. Howeverwe have seen in Section 4.2.1 that distance and absorption con-spire to make even bright early-type stars inconspicuous. At adistance of 15 kpc and with an absorption of 7.3 mag, basedon the H I column density, an O3V star ( M v = − .
0) wouldhave m v = .
2, whereas a B1V star ( M v = − .
2) wouldhave m v = .
0. For an absorption of 13 mag, derived fromthe 2MASS color-color diagram, the corresponding magnitudeswould be 22.9 and 25.7, for an O3V and a B1V star respectively.As mentioned above, a shorter distance of 11.5 kpc leads to es- . Cichowolski & S. Pineault: A large H I shell in the outer part of the Galaxy 7
H-Ks J - H Infrared Excess Sources N o r m a ll y R e dd e n e d S t a rs Cool Giants
Av=10 Av=20Av = 5 Av=15
Fig. 4. ( H − K , J − H ) color-color diagram of the 2MASS sourcesfound in a 400 ′′ circular area centered at ( l , b ) = (91 . ◦
7, 1 . ◦ ? . The par-allel dashed lines are reddening vectors with the crosses placedat intervals corresponding to five magnitudes of visual extinc-tion. We assumed the interstellar reddening law of ? , A J / A V = .
282 and A H / A V = . + − + −
114 is indeed the action ofmany massive stars, from the estimated value of the kinetic en-ergy of the shell ( E kin = (1 . ± . × erg, see Section 4.1) wecan estimate how many massive stars would have been needed tocreate it. Adopting mean stellar wind parameters for O-type stars(i.e. ˙ M = × − m ⊙ yr − and v w = − , ? ) and consid-ering that the main sequence phase lasts for at least 3 × yr, theinjected wind stellar energy is about 1 . × erg. Bearing inmind that the conversion e ffi ciency is about or less than 0.2 andconsidering the possible e ff ects of a lower metallicity environ-ment, the energy released by the stellar winds should be greaterthat 5 . × erg. Thus, at least four O-type stars were neces-sary to create GSH 91.5 + − . × erg, implying that at least three O-typestars were required. ? proposed and interesting procedure for setting various con-straints on an SNR origin, which they applied to their H I shelland another one previously discovered by ? . Their procedure es-sentially rests on the expression giving the maximum observableradius of an SNR before it merges with the ISM ( ? ): R merge = . E / n − / Z − / pc , where E is the explosion energy in units of 10 erg, n the mean ambient particle density in cm − , and Z the metal-licity normalized to the solar value. The input parameters to the above equation are the observed radius and particle den-sity, R sh = d pc and n = . d − cm − . The results forGSH 91.5 + −
114 are shown as the last entry in Table 3, wherethe numerical values are for d =
15 kpc. The first two entries areidentical to Table 2 of ? (with the addition of two parameters anda slight change in notation for added clarity).The column labelled R ∗ merge (column 5) gives the merging ra-dius assuming a canonical value of E ∗ = R sh is about twice as large as the predicted merging radius R ∗ merge ,which means that the shell should have vanished well beforereaching the observed radius. This implies that either E is largerand / or the object is closer. The column labelled E ∗ (column6) gives the energy required for R merge to be equal to the ob-served radius of the shell. The morphology of GSH 91.5 + − E ∗ is a lower limit. The last columnis the maximum distance GSH 91.5 + −
114 would have to be inorder to be the result of a single SN explosion with E =
1. Thisdistance corresponds to a systemic velocity of about −
52 km s − which, as is the case for the other two shells, is totally unrealistic,given the observed systemic velocity of −
114 km s − .If we compare the three shells, we see that the first and lastone are very similar in radius and mass, because they are vir-tually at the same distance and evolving in an ISM of com-parable particle density. The second one, GSH90 + ? , neither of thethree shells could have been caused by a single SN explosion.Furthermore, as emphasized by ? , this argument applies as wellto a single O- or B-type stellar wind bubble. Using 11.5 kpc asthe distance would only change R ∗ merge and d max by about 10%and E ∗ o by about 40%, thus not altering the general conclusion. From Figure 3 it is clear that there is no evidence of enhancedradio continuum emission associated with GSH 91.5 + − ′ is equivalent to 4.5 pc at a distance of 15 kpc):inner and outer radius, R i =
160 pc, R o =
200 pc, shell thickness ∆ R =
40 pc and initial particle density n o = . − . These areaverage quantities because obviously the shell radii and thick-ness vary significantly with azimuthal angle. With these param-eters, we deduce that the particle density within the shell is now n s = . × n o = . − . For simplicity, we shall neglect thepresence of helium in the following simple estimates. We assume that an ionizing star is located at the center of the H Ishell and ask to what extent the inner side of the H I shell shouldbe ionized, assuming that negligible material is present insidethe shell. This is the classical Str¨omgren sphere problem where,instead of integrating from 0 to some maximum radius R m , we S. Cichowolski & S. Pineault: A large H I shell in the outer part of the Galaxy
Table 3.
Constraints on a single SNR origin
Object n R sh Z R ∗ merge E ∗ d max Reference(cm − ) (pc) ( Z ⊙ ) (pc) (10 erg) (kpc)GSH 138–01–94 0.27 180 0.1 93.3 8.0 5.9 ? + ×
110 0.1 44.7 6.3 5.2 ? + −
114 0.3 200 0.1 90 12.6 6.9 This paper1.0 80 18.2 6.1 R ∗ merge (column 5): merging radius for E = E ∗ (column 6): value of E o required for R merge to equal the observed radius. d max (last column): maximum distance for a single SN explosion with E = Table 4.
Constraints from radio continuum + Stellar type N u , S R R m (s − ) (mJy) (pc) (pc)O3V 741 3580 141 198O6V 219 1060 42 173B0V 14.5 70 3 161 + For a distance of 15 kpc. integrate from R i to R m : R N u d S u ( r ) = − R R m R i π r n e α (2) d r ,where S u ( r ) is the rate of ionizing photons at distance r , N u is the total number of ionizing photons emitted by the star, n e the electron number density (assumed equal to n s ) and α (2) = . × − cm s − is the recombination coe ffi cient excludingcaptures to the ground level. Solving for R m , we obtain R m = R i (1 + R / R i ) / , R = N u π n e α (2) R i . (1)If R m > R o , then the H I shell is fully ionized (and shouldnot exist!) and some photons escape freely. For our parameters, R = . N u , pc, where N u , is the ionizing photon rate in unitsof 10 s − . The total number N u of UV ionizing photons alsofixes the total observed flux density S ν at a given frequency andis given by (e.g., ?? ) N u = . × T − . ν . S ν d , (2)where T is the electron temperature in units of 10 K, ν GHz = .
42 the frequency in GHz and S ν is in Jy. Using T = . d kpc =
15, we obtain S . = . N u , mJy. This estimate for S ν is an upper limit, because photons can escape through inho-mogeneities of the shell (on scales smaller than the beam) orthrough the opening in the north.Table 4 shows the results of these simple calculations forthree di ff erent stellar types, chosen to illustrate di ff erent possi-bilities. Evidently the observations are totally incompatible withthe presence of an O3V star (or any star with the same value of N u ) as the H I shell would be essentially totally ionized. In orderfor the neutral shell to be actually present, there would then haveto be a substantial amount of ionized gas within R i to absorbthe ionizing photons. This goes against the canonical view of ashell or supershell as surrounding an essentially empty cavity.Figure 3 also fails to show any significant amount of continuumemission in excess of the di ff use emission seen in this generaldirection which, given the large distance of the shell, is likely tobe foreground emission. At the other extreme, a B0V star (or any star with N u ≤ . × s − would ionize a negligible part of the H I shell, herehardly 1 pc. We can use eq. (1) to estimate the ionizing photonrate, which would be enough to ionize a thin layer 1 beam sizein thickness (1 ′ or about 5 pc). This gives log N u = .
90, cor-responding to an O8.5V star. Note that WR stars have log N u inthe range 48.6 to 49.4 (Crowther 2007).As for the intermediate case of an O6V star, Table 4 showsthat it would ionize about 13 pc of the inner H I shell. Thisionized gas would have an angular thickness of about 3 ′ anda flux density of about 1 Jy. An upper limit to the brightnesstemperature T B can be obtained by assuming this flux to origi-nate from an annulus 35 ′ in radius and 3 ′ in thickness (givingan area of about 660 arcmin or 390 CGPS beam areas at 1420MHz). We obtain T B < . / beam ≈ . σ . An alternative means ofestimating T B is to use an approximate emission measure givenby EM = R n e d l ∼ n e R o ∼
100 pc cm − , from which the bright-ness temperature is T B = ( EM /
566 pc cm − ) K = .
17 K. Giventhe spatially varying continuum surface brightness within the in-ner shell boundary, such a faint ionized layer would not be easilydetected by our observations.Summing up, if the H I shell is relatively homogeneous,which is far from certain, we can rule out the presence ofany star more luminous than O6V or with more than about2 × ionizing photons s − . Placing the shell at the slightlycloser distance of 11.5 kpc would have the e ff ect of making itslightly thinner and closer to any central star with the result thatfor any given spectral type, more of the shell would be ionized. In addition to an obvious opening in the north, the H I shell doesappear to be inhomogeneous on scales comparable to the beamsize (about 1 ′ or 4.5 pc at a distance of 15 kpc). If it were inho-mogeneous on scales smaller than the beam (thus undetectablein the CGPS image), the H I shell could let a significant num-ber of photons escape, thus leading to a more diluted and fainterionized layer.The existence of inhomogeneities, whether they are calledclumps, filaments, clouds, or cloudlets, has been invoked in anumber of contexts. As a possibly extreme case (with regards tosize), ? recently postulated the existence, within the stellar windcavity of the 2 . ′ cm − and of a radius as small as 0.05 pc. Thesewould provide the dust needed to explain the 24 µ m emission ob-served within the N49 cavity through erosion and evaporation. . Cichowolski & S. Pineault: A large H I shell in the outer part of the Galaxy 9 We have no way of knowing the degree or type of inho-mogeneities which could be present, however, for illustrativepurposes, we consider inhomogeneities in the form of clumpswith a radius r = n ′ = cm − ,thus each clump would have a mass m = . m ⊙ . With atotal H I shell mass of 2 . × m ⊙ , this implies that thereare N = . × such clumps. The volume filling factor is f ≈ (4 π r N / / (4 π R i ∆ R ) = r N / (3 R i ∆ R ) ≈ . × − . To es-timate the transparency of the shell, we note that as viewed fromthe center of the shell, the clumps cover a fraction ǫ of the shellarea where ǫ ≈ N π r / (4 π R i ) = f ∆ R / (4 r ) ≈ . × ionising photons s − could contribute to partially ion-ize the clumps from R i to R o , resulting in an undetectable radiocontinuum surface brightness.In summary, if the H I shell is basically homogeneous, therecan be no star producing more than about 2 × ionizingphotons per second. Depending on the degree of clumpiness ofthe shell however, such powerful stars could still be present.Nevertheless, given the possibly large kinematical age of theshell (up to 17 Myr), it is most likely that there are no O starsleft inside the shell and that any B star, if present, contributesnegligibly to ionization. Shocks in expanding supershells are widely believed to be theprimary mechanism for triggering star formation ( ? ). Shellsbehind shock fronts experience gravitational instabilities thatmay lead to the formation of large condensations inside theswept-up material, and some of these may produce new stars( ? ). An increasing body of observational evidence confirms theimportance of this mechanism (e.g. ???? ). Given the size ofGSH 91.5 + − The 1420 MHz radio continuum image (Fig. 3) shows that a fewextended, yet relatively compact objects are found near the in-ner periphery of the H I shell. Apart from G91.11 + +
49 discussed by ? , there are only two such objects within the boundary of theH I shell. Could these be H II regions whose formation was trig-gered by the expanding H I shell? An unambiguous answer tothis question requires a determination of both the distance andthe radio spectral index.We can obtain an estimate of the spectral index α ( S ν ∝ ν − α )by measuring the flux densities at 1420 and 408 MHz. Yet forboth G91.56 + + S and S , see insets of Fig. 3). Whereas this poses no problemat 1420 MHz, the larger beam at 408 MHz results in a partialblend of these point sources with the nearby extended structure,particularly severe for G92.24 + S and S were successfully subtracted atthis frequency, we then used the imview program to measure This program is part of the DRAO export software package the flux density of the remaining extended source. The error wasestimated by using slightly di ff erent background levels.On the 408 MHz image however, the larger beam size ( ≈ ′ × ′ ) results in the Gaussian-fitting routine which finds both S and S to be extended, in contradiction with the fact that both areunresolved even at 1420 MHz. Inspection of the 408 MHz imageshowed however that the program did successfully subtract thecombined flux density of the extended structure plus the nearbycompact source, for both G91.56 + + α = .
75, which is representative of the spectral indexof extragalactic radio sources between 178 and 1400 MHz ( ? ).We then created a Gaussian source with this flux density with the408 MHz beam size that we removed from the original image.We varied α until the best artefact-free subtraction was found.The flux density of the extended structure was then estimatedas the di ff erence between the combined flux density and that ofthe nearby point source. The 408 MHz flux densities of the twopoint sources, S and S , were cross-checked by comparing our408 MHz values with the 327 MHz values from the Westerborksurvey ( ? ). The values for S and S at 327 MHz are 57 mJy and90 mJy, respectively, in satisfactory agreement with our determi-nations.Table 5 summarizes the obtained flux densities and spectralindices, together with some independent measurements. Bothsources S and S have a non-thermal spectrum consistent withtheir origin as extragalactic radio sources. As for G91.56 + + ff erence in background removal and / or in-clusion of nearby point sources, they show a spectral index con-sistent with a thermal nature.Without a distance estimate it is impossible to ascertainwhether they are associated or not with the H I shell, howevergiven their position on or near the H I shell, the possibility thatthey might be H II regions whose formation was triggered by theexpanding H I shell is worth pursuing.Both sources are unfortunately too faint for a significant ab-sorption spectrum to be obtained, which would enable us to seta limit on the distance. Another way to try to estimate theirdistances is to look for signatures in the H I emission distri-bution around these sources (since unfortunately no moleculardata is available for this region). An inspection of the entire H Idata cube shows a minimum in the velocity range from about–46.0 to –50.0 km s − in the area of G91.56 + + + + − + + +
49 (van der Werf & Higgs1990), also lies near the periphery of the large H I shell. Couldthe two be associated? ? obtained an H103 α recombination linespectrum, which showed this source at a velocity of − . ± . − . The resolution of their NRAO 43-m observations was 5 ′ .Van der Werf & Higgs (1990) obtained a similar value of − ± − , using DRAO H I absorption spectra and an H112 α re-combination line spectrum obtained with the 100-m E ff elsberg Fig. 5.
H I emission distribution averaged between –46.0 and –49.3 km s − .The black contour corresponds to the 8.5 K level ofthe 1420 MHz radio continuum emission.telescope at a resolution of 2 . ′
6. These observations would seemto place G91.11 + ff er-ent distances. However, the VLA 5 ′′ radio continuum image ofHiggs et al. (1987) shows the head of BG 2107 +
49 to consist ofa set of discrete knots of emission about 10 ′′ in size and a moreextended and di ff use region about 3 ′ in diameter. Although wecannot completely rule out the possibility that some of the knotscould be at the larger distance of our H I shell, the fact that theH103 α recombination line spectrum of Higgs et al. (1987) isextremely well fitted by a single line at − . − with es-sentially no residuals near −
114 km s − would seem to rule thisout. We feel that the case for triggered formation is stronger forIRAS 21147 + ? and ? . In both cases the authors found a small cavityimmersed inside a larger HI shell. However, their interpretationabout the origin of the structures di ff ered. ? concluded that boththe large HI shell and the smaller bubble had been created by thesame star, HD 197406, during di ff erent stages of its evolution.On the other hand, ? found an HI cavity around the OB associ-ation Bochum 7 and concluded that this association might havebeen born as a consequence of the evolution of the large shellGS263-02 + . ◦
3. At the distance of GSH 91.5 + − − (see Fig. 2), a lower limitfor the expansion velocity can be assumed as 5 km s − . Underthese conditions we derived an upper limit of about 5 Myr forthe dynamical age if a stellar wind origin is considered. Thisage is significantly smaller than the one obtained for the largeshell (7 - 17 Myr), but larger than the duration of the WC phase(10 yr), suggesting that the O-phase of the current WC star would also have contributed to its formation. Taking into ac-count that progenitors with a mass of some 40 to 50 M ⊙ aresuggested for late WC stars ( ? ), the dynamical age estimatedfor the small cavity is consistent with the time that these starsstay on the main sequence, between 3.7 and 4.9 Myr ( ? ). Basedon the age estimate di ff erence, the most probable scenario is onewhere the large shell, GSH 91.5 + − +
5. Conclusions
The measured kinetic energy of expansion of GSH 91.5 + − ff -center suggests that if they are in-deed inside the shell, their shaping influence is minimal. We havefound no evidence for the presence of other massive stars, butabsorption would preclude the detection of these objects at theinferred distance.An interpretation as the H I shell of a single SNR isalso not tenable. As for GSH 138-01-94 and GSH 90 + + −
114 has likely been caused by the combined windsand SN explosions of a number of massive stars.The H I shell appears open to the north, i.e. in a directionaway from the Galactic plane, suggesting that the shell has burstout of the Galactic disk. A filament displaced from the north-ern boundary (Filament B, Fig. 1) could be the remains of thetop (now missing) part of the shell. The faint radio continuumemission, consisting of filaments more or less aligned in a di-rection perpendicular to the Galactic plane, lends support to thebreakout hypothesis.Two relatively compact thermal sources (Table 5), seen inprojection near or on the boundary of the H I shell, couldhave formed in gas compressed by the expanding shell. One ofthe DWCL sources, IRAS 21147 + Acknowledgements.
The research presented in this paper has used data fromthe Canadian Galactic Plane Survey, a Canadian project with international part-ners, supported by the Natural Sciences and Engineering Research Council. Thework of S.P. and S.C. was supported by the Natural Sciences and EngineeringResearch Council of Canada and the Fonds Qu´ebecois pour la Recherche surles sciences de la Nature et la Technologie. SP acknowledges the hospitality ofthe Instituto Argentino de Radioastronom´ıa where part of this work was carriedout. We are grateful to the referee, whose suggestions led to the improvementof this paper. This project was partially financed by the Consejo Nacional deInvestigaciones Cient´ıficas y T´ecnicas (CONICET) of Argentina under projectsPIP 01299 and PIP 6433, Agencia PICT 00812 and UBACyT X482. . Cichowolski & S. Pineault: A large H I shell in the outer part of the Galaxy 11
Table 5.
Flux densities and spectral indices of compact sources + ( S ν ∝ ν − α ) F (mJy) σ (mJy) F (mJy) σ (mJy) α ∆ α Notes a S
37 2 76 9 0.58 0.10 This paper, see text2 G91.56 + − .
15 0.16 This paper, see text3 261 11 346 28 0.09 0.06 vdWH, HvdW b − .
88 0.03 Kerton et al. (2007)5 S
25 2 59 4 0.69 0.08 This paper, see text6 G92.24 + − .
19 0.13 This paper, see text7 99.4 6.9 168 31 0.42 0.16 Kerton et al. (2007) c + Sources S and S are the nearby point sources seen in the insets of Fig. 3 a HvdW = Higgs & van der Werf (1991), vdWH = van der Werf & Higgs (1990) b Flux density at 408 MHz includes contribution from source S (29P62 at 1420 MHz – HvdW). Note that the spectral index listed there comesfrom a linear fit using measurements at five frequencies between 408 MHz and 4.85 GHz c Likely includes source S1