Structure and Feedback in 30 Doradus I: Observations
aa r X i v : . [ a s t r o - ph . GA ] S e p Structure and Feedback in 30 Doradus I: Observations
E.W. Pellegrini , Department of Astronomy, University of Michigan, 500 Church Street, Ann Arbor, MI 48109
J.A. Baldwin
Physics and Astronomy Department, Michigan State University, 3270 Biomedical Physical Sciences Building, EastLansing, MI 48824, USA
G.J. Ferland
Department of Physics and Astronomy, University of Kentucky, 177 Chemistry /Physics Building, Lexington, KY40506, USA [email protected]
ABSTRACT
We have completed a a new optical imaging and spectrophotometric survey of a 140 x 80 pc region of 30 Doradus centered on R136, covering key optical diagnostic emission lines including H α ,H β , H γ , [O III] λλ λλ λλ λ λ
1. Introduction
A significant share of our knowledge about star-formation rates, chemical abundances and abundancegradients in galaxies comes from studying emissionlines from distant Giant Extragalactic H II Regions(GEHRs). But the observations of GEHRs are gen-erally interpreted in the absence of a quantitative un-derstanding of the structure responsible for the emis- Visiting astronomer, Cerro Tololo Inter-American Observatory,National Optical Astronomy Observatory, which are operated by theAssociation of Universities for Research in Astronomy, under con-tract with the National Science Foundation. Physics and Astronomy Department, Michigan State Univer-sity, 3270 Biomedical Physical Sciences Building, East Lansing, MI48824, USA sion, or of its relationship to the processes of star for-mation. Massive stars heat, fragment and compressthe gas clouds from which they have recently formed.This “stellar feedback” is believed to drive the collapseof gravitationally bound clouds and trigger ongoingstar formation (Zavagno et al. 2010; Oey et al. 2005;Elmegreen & Lada 1977) and also shapes the structureof the gas which in turn affects the emission line spec-trum. For these reasons, it is important to understandhow such processes work.Most GEHRs are distant, with poor spatial resolu-tion. What is needed to better understand the mecha-nisms of stellar feedback is a nearby “Rosetta stone”in which we can resolve spatial details on the scaleover which the feedback effects operate, for example1ver the pc-scale thickness of the ionized layer in anH II region. 30 Dor in the Large Magellanic Cloud,a nearby star-forming region that can be classified asa GEHR, is that key example. It is the largest star-forming region in the Local Group. The central clus-ter emits almost 500 times more ionizing photons thanthe Orion Nebula and 2–4 times more ionizing pho-tons than other large Local Group star-forming regionssuch as NGC 3603 in our Galaxy or NGC 604 in M33.Because 30 Dor is only 48.5 kpc away (Macri et al.2006), it can be studied with high spatial resolution(1” = 0.25 pc).The violent history of 30 Dor has led to the for-mation of vast cavities and shell-like structures aroundits central star cluster NGC 2070 (Meaburn 1984; Sel-man et al. 1999). The walls of the cavities have longbeen seen in optical ionization-front tracers like [S II].The SAGE (Meixner et al. 2006) survey of the LMCincluded 30 Dor. SAGE traces the ionization-frontsin 8 µ m PAH emission with a resolution of 5 arcsec.Filling these cavities is a K gas visible in X-rayemission (Townsley et al. 2006). Many of these struc-tures have H α velocity profiles which show them tobe expanding shells (Meaburn 1984). These shellshave typical kinetic energies equal to erg with atotal kinetic energy in the vicinity of NGC 2070 equalto erg (Chu & Kennicutt 1994). These types ofkinematic arguments suggest SNe are a major compo-nent of the total energy budget. However, within rea-sonable estimates about the timescale for injecting en-ergy from stellar winds into the ISM it is equally likelythat stellar winds are the agent shaping the nebula.The remnants of the molecular gas cloud fromwhich the stars have formed has been observed in op-tically thick CO emission (Poglitsch et al. 1995). Thismolecular cloud runs NE-SW behind NGC 2070. Thebrightest optical arcs, seen in Figure 1, appear to tracethe surface of this molecular cloud, and are bright be-cause of their high density. 30 Dor has been named the“Tarantula Nebula” because images showing emissionlines are dominated by the interwoven pattern of thesebright arcs. However, as we will show here, the ma-jority of the line emission originates in the extendedregions of low surface brightness.This current paper describes our new optical-passband imaging and spectroscopic survey of 30 Dor.Our data set, which is publicly available, is designedto include key emission lines which quantitatively de-scribe feedback processes. One of our goals is to sur-vey the full region on the sky that would be lumped together in observations of a distant GEHR, for directcomparison to such objects. Luminous H II regioncomplexes in distant galaxies typically are about 100pc in diameter (Oey et al. 2003; Hunt & Hirashita2009). The 140 ×
80 pc area covered by our spec-tra is of similar size, and encompasses 50 % of theH α flux from the 10’ diameter region with H α sur-face brightness appreciably above the general back-ground level in this part of the LMC as measured fromthe Magellanic Cloud Emission Line Survey (MCELS;Smith & MCELS Team (1998)). The larger 11’ × α flux.In 30 Dor, the numerous expanding shells arecaused by mechanical energy input (Meaburn 1984;Chu & Kennicutt 1994), and the ionizing energysource for a radius of at least 100 pc is generallyaccepted to be photoionization by the central clusterNGC 2070 (Elliott et al. 1977; Tsamis et al. 2003).However, recent work based on Spitzer observations(Indebetouw et al. 2009; hereafter I09) has questionedthis. In this paper we use our observations to inves-tigate the source of the photons that ionize the gasbeyond the bright arcs.The ionization and shaping need not be caused bythe same mechanism. The application of our new datato studying feedback processes – the question of whathas shaped the nebula – will be presented in a com-panion paper (Pellegrini et al. 2010, hereafter PaperII). There we will also discuss the consequences of theresulting structure on the escape of ionizing radiation.Finally, the second paper will evaluate the results from“strong line techniques” (Pagel et al. 1979; Kewley& Doptita 2002; Denicol´o et al. 2002) applied to ourcomposite spectrum of 30 Dor, by comparing them tothe chemical abundances and other nebular propertiesobtained by a point-by-point analysis of our spatiallyresolved dataset.
2. Observations2.1. Existing optical passband data sets
Most of the strong optical emission lines from star-forming regions trace ionized gas at electron tempera-tures T e ∼ K. A few key lines from [O I], [O II],[N II] and [S II] form in the interface region at theedge of ionization-bounded H II regions. The elementsS, O, N, Ar have transitions visible from the groundthat are sensitive to both T e and the electron density n e , and also to gas-phase abundances. Even though 302or is a key example of a GEHR, the available mea-surements of it in the emission lines of these elements,using either direct imaging or especially spectroscopy,are surprisingly limited.The best existing publicly available narrow-bandoptical image data sets are the MCELS, and archivalHST images. The MCELS survey covers the central8 × of the LMC including 30 Dor, in H α +[N II], [O III] and [S II] (using many of the sameemission line and continuum filters used in this presentstudy). However, the data were taken with the 0.9mSchmidt telescope at CTIO and have a spatial scale of2.3 arcsec pixel − with a resolution closer to 5 arcsecFWHM. This means that structures smaller than 1.2 pcare unresolved, blurring fine details.The HST archival data cover the central 4.5 by 3arcmin of 30 Dor with at least 0.1 arcsec resolution(Scowen 1998; Walborn et al. 2002) in H α , [O III] and[S II], revealing vastly more detail. However the HSTdata are limited to the brighter central region aroundR136, the very compact group of stars at the very cen-ter of NGC 2070. Since the brighter nebula accountsfor only 25 per cent of the total nebular emission inH α the limited spatial coverage of HST misses thebulk of the emission that would be detected if 30 Dorwere viewed from a much greater distance.Turning to the available spectroscopy, Krabbe & Copetti(2002) obtained a set of long-slit observations whichcovered H β , [O III] λ λ α and [N II] λλ − ) ve-locity resolution. They found that about half of thekinetic energy in 30 Dor is contained in shells in thecentral regions which are expanding with characteris-tic velocities v ∼ − km s − , and that the kineticenergy contained in this expansion greatly exceeds thegravitational binding energy.There has also been a limited amount of deepechelle spectroscopy covering a much wider wave-length range (Tsamis et al. 2003; Peimbert 2003). These spectra measured hundreds of emission linesthat can be used for detailed chemical abundance anal-ysis, and also cleanly resolve the density-sensitive[O II] λλ To the existing data sets we added a new set ofnarrow-band images taken with the SOAR Optical Im-ager (SOI) on the 4.1m SOAR Telescope . We usedthe H α × × in fil-ter sets, where the filter names refer to the approximatecentral wavelengths and FWHM bandpasses. A sum-mary of the SOI observations is given in Table 1.In each passband we took grids of 5 × SOIimages, overlapping them to give a 12 ×
13 arcmin field of view with a scale of 0.15 arcsec pixel − . Theindividual images were then combined to create finalmosaic images in each passband. Due to the signif-icant spatial overlap of the individual images, the to-tal integration time in a given filter varies across themosaic. Table 1 includes, for each filter, the result-ing minimum and maximum integration times at anypoint in the mosaic. The data are seeing-limited, withfull width at half maximum intensity FWHM = 0.5 -0.9 arcsec, much better than the 5 arcsec FWHM inthe MCELS survey, and are three times more sensitiveto diffuse nebular emission than the existing HST dataset.As an example of the results, Figure 1a shows themosaic image made with the H α + [N II] filter, prior tocontinuum subtraction. R136 is marked with a white × . The 1 arcmin scale bar is equivalent to 14.1 pcfor an LMC distance of 48.5 kpc (Macri et al. 2006).Figure 1b shows, on the same image, the slit positionsused in the SOAR Telescope spectroscopic observa-tions described below. R line in each filter, according to R line = R narrow − R cont W narrow W cont (1)where R narrow and R cont are equal to the atmospheric-extinction corrected count rate in the narrow band andcontinuum filters, respectively, and the effective filterwidth W i in the i th filter is W i = Z T i ( λ ) dλ (2)where the transmission curve T i ( λ ) was measured bythe SOAR staff.Changes in stellar scattered light, seeing or skybrightness between the time the emission lines weremeasured and the continuum was measured werefound to be a limiting factor in the quality of thecontinuum-subtracted images, so the observationswith the continuum and emission filters were madesequentially one after the other before offsetting thetelescope to each new point in the mosaic grid. Thisensures that the relative flux between emission line andcontinuum images remains constant for each pointing. In Feb 2008 we obtained a grid of long-slit spec-tra of 30 Dor using the RC spectrograph on the 4mBlanco telescope at Cerro Tololo Interamerican Ob-servatory (CTIO). The spectrograph slit was 5 arcminin length, and was positioned at a total of 37 differ-ent locations spanning the nebula (Figure 1b). Twosets of spectra were taken at orthogonal angles. Theslit position angles PA = 13 deg and PA = 103 degwere chosen to maximize the number of key ionizationfronts that could be covered with the slit either crossingthe ionization fronts at an approximately perpendicu-lar angle or running directly along them. The Blancotelescope uses an equatorial mount, which results in aconstant position angle of the slit once the instrumenthas been rotated to the correct PA. We minimized theuncertainty in the PA by rotating the slit only once eachnight, after all observations at the initial PA were com-pleted. As a result we have a high degree of confidence in our stated position angles, and therefore of the map-ping onto the sky of individual points along the slit.Table 2 lists for each slit position the identifyingposition number, the RA and Dec of the slit center inJ2000 coordinates, the PA, and the total exposure time.For simplicity, slit positions with PA = 13 deg are num-bered beginning at 1 at the easternmost position andincreasing to 17 toward the west. These include thetwo most extreme north and south slit positions (posi-tion numbers 16 and 17, respectively). Numbering ofslit positions with PA =103 deg then begins at 20 andincreases towards the south, with even numbered slitpositions centered to the east of R136 and odd num-bered ones to the west. The only exception is posi-tion 31 which has a PA equal to 98 degrees. Positions40-44 were taken with the SOAR Telescope and aredescribed in the following sub-section.The Blanco Telescope observations were taken un-der photometric conditions with a slit width of 3.5arcsec sampled at 0.5 arcsec/pixel, using the 600l/mm grating. This setup delivers a spectral resolu-tion of 7.0 ˚A (FWHM) with a spectral sampling of 1.06˚A/pixel as measured from the [O I] λ λ Six additional locations in 30 Dor were observedwith the Goodman High Throughput Spectrograph4Clemens et al. 2004) on the SOAR telescope, on 5Feb 2009. The instrument was used with a single slit3.9 arcmin long and 0.46 arcsec wide, with its 300 linemm − grating. The wavelength range was 3950 ˚A to9335 ˚A with 1.32 ˚A pixel − sampling and a spectralresolution of 4.9 ˚A FWHM at H α . This setup easilyresolved the [S II] doublet near 6720 ˚A and the [O II]lines at λλ α intensity ratio, lying to the east of R136.Figure 2 shows the SOAR slit positions superimposedon a map of the [S II]/H α intensity ratio made fromour SOAR images. The orientation and spatial cover-age of Figure 2 is the same as that of Figure 1. At thetime of the observations the SOAR Telescope lackedan atmospheric dispersion corrector, which is impor-tant with our broad wavelength coverage and narrowslit. As a result our observations were restricted tonearly the parallactic angle, resulting in the patternshown in Figure 2. The positions of the SOAR spectraare numbered beginning at 40, to minimize confusionbetween the two spectroscopic data sets.To remove night sky contamination we obtained aspectrum of the sky 2 deg north and 2.7 deg east ofR136. The sky spectrum was smoothed with a me-dian filter spanning 15 arcsec along the slit, to decreasenoise and remove stars along the slit. The intensity inthe night sky lines was then scaled to match the objectframes and subtracted. The redshift of the LMC helpsto separate night sky forbidden lines from the samelines from 30 Dor. For example, in the case of the [OI] λ λ λλ λ λ λ U ,the energy distribution of the ionizing radiation, andthe O/H and S/H abundances.
3. The spectroscopic ’data cube’3.1. Emission-line flux measurements
One of our goals was to create a high-quality setof measured emission-line strengths that will be ofgeneral public use, so we will describe our measure-5ent procedures in some detail. Table 3 lists the emis-sion lines measured from the Blanco and SOAR spec-tra. Columns 1 through 3 indicate the observed andrest wavelengths and ionic species. The fourth col-umn lists the extinction coefficient f λ used to dered-den the observations, using a standard Galactic extinc-tion curve with a ratio of total to selective extinctionR=3.1 (Cardelli et al. 1989). The final column indi-cates which data sets include the emission line.We extracted 1D spectra binned over successive 2.5arcsec increments along the slit in the 2D spectra. The2.5 arcsec width of the extraction window along theslit was chosen so that the in the fainter regions theline flux equaled the noise in the continuum. The pro-gram then automatically measured, in each extracted1D spectrum, the emission line fluxes and their uncer-tainty. This produced a ‘data cube’ of emission-lineintensity measurements at 4238 points on the 30 Dornebula.We needed to use a summation approach to measurethe flux in each emission-line profile because of the of-ten complicated shape of the profiles. It is worth notingthat most of this structure in the profiles is not due tovelocity. It is due to spatial structure in the nebula thatis smaller than the 3.5 arcsec slit width used for theBlanco spectra. To illustrate this, Figure 4a shows anenlarged portion of the SOAR [O III] image at the po-sition of a ring-like structure, and Figure 4b shows the[O III] λ λ λλ λ λ β will be underestimated.Placement of bright star clusters in the slit was inten-tionally avoided and the number of regions affected bystellar absorption is small. We estimated the uncertainties in the measured linestrengths by assuming Poisson statistics in the originalphoton count rate and then carrying parallel noise im-ages all the way through the same reduction process asthe data images. At the beginning of the data reductionthe measured counts were converted to photons usingthe known detector gain. An initial noise image wasthen calculated using σ = ( N photons + gain × ReadN oise electrons ) / (3). All multiplicative calibrations such as flat fields, il-lumination corrections and the flux calibration wereapplied to the noise image by multiplying the σ val-ues in the noise image.In the step where the spectra were linearized anddistortion was corrected, there was a re-binning of thedata. In this case σ (the variance of the noise) wasrebinned using the same transformation. The result-ing σ in each pixel is thus summed in quadrature. Wemanually inspected a random set of extracted spectraand verified that these estimated errors do in fact rea-sonably measure the observed pixel-to-pixel scatter inthe continuum points. An exception to this is [O III] λ λ σ values for eachpixel were then used to compute error bars for the mea-sured line fluxes and ratios. The large range in the surface brightness of thenebula meant that the 5 arcmin long slit includedboth bright and faint regions. In order to get a suf-ficiently high signal/noise ratio in the faint regions,long exposures were used and the [O III] λ α lines were often saturated at various places alongthe slit. When necessary, additional short exposureswere made to correct for saturation. All repeated ob-servations of a single position, long and short, weregray-shifted to the highest value according to the to-tal flux of the [O III] λ λ λ β and ([S II] λ λ α ratios mea-sured for these overlapping observations. The his-togram shows that the data do repeat to within the typ-ical range of differences due to the mismatch betweenthe areas sampled on the sky by the two slits, with es-pecially good repeatability in the [O III]/H β ratio. The reddening was determined separately for eachextracted 1D spectrum, using the observed H α /H β in-tensity ratio and assuming an intrinsic ratio of 2.87appropriate for Case B and a gas temperature K(Osterbrock & Ferland 2006) and R=3.1. This is ade-quate even though 30 Dor is in the LMC because (1)although the LMC extinction curve departs consider-ably from the Galactic curve in the ultraviolet, the twoare very similar in the optical passband, and (2) someconsiderable part of the reddening is in any case dueto foreground material within our own Galaxy. Wethen applied the reddening to each measured line fromthe same extracted 1D spectrum, using the extinctioncurve values f λ from Table 3. A check on the validity of this procedure is that the dereddened H γ /H β ratiosagree with the predicted Case B values to within an av-erage of 3 percent, with a 1 σ scatter of 4 percent whichis consistent with the observational errors. At each position along each spectrograph slit we de-termined the electron density n e from the dereddened[S II] λ λ nebular was used to derive the electron densitiesusing a 5 level sulfur atom assuming a gas temperatureof K. The errors in the density measurements werecalculated using the 1 σ uncertainty in the line ratio.The difference between the densities at these extremesand at the nominal value are the reported density un-certainties. If either the nominal or 1 σ [S II] line ratiowas greater than 1.41 the density was only constrainedto be n e ≤ cm − (the low-density limit). Using thepublished contour plots of log( n e ) derived from [S III]in I09 we compared the electron densities measuredfrom [S II] and [S III]. We find they agree to 0.1 to 0.2in log( n e ), noting that the increments in the densitycontours of I09 are 0.1 in log( n e ).The electron kinetic temperature T e was alsomeasured at each position, from the dereddened[O III] ( λ λ λ The Blanco instrument setup was chosen to coverthe strong optical nebular emission lines from 4100 to7400 ˚A including H β , H α and the [S II] doublet at λλ α along the slit to the H α images, (2)using positions, measured from our narrow band con-tinuum images, of stars purposely placed in the slit asreferences and (3) where necessary using 2D nebularfeatures visible in both the spectra and images such as7n the example shown in Figure 4.Table 4 is a sample of the final data product from theBlanco spectra. It lists the dereddened line strengthsand several additional parameters for each extractionwindow along each slit position, and their uncertain-ties. The entire table, which is available in electronicformat, includes measurements at 4238 positions withone row per position. Column 1 lists the slit positionnumber (as used in Figure 1b and Table 2). Columns2-4 contain the RA and Dec offsets in arcsec fromR136 and the central pixel row of the extraction win-dow along the slit. Then columns 5-10 list the electrondensity and its 1 σ uncertainty limits, followed by theelectron temperature with its 1 σ uncertainties. Column11 is the reddening A V deduced from the H α /H β ra-tio. These are followed in columns 12 and 13 by thedereddened H β surface brightness and its 1 σ uncer-tainty. The remaining 44 columns list, in pairs, thedereddened surface brightness of each measured emis-sion line relative to H β and the corresponding uncer-tainty.The last three rows of Table 4, collectively labeledposition 38, list the average properties of 30 Doradusderived from the entire data set. Position 38 row 1 isthe average of the dereddened fluxes, and the physicalproperties derived from those. Row 2 is the averageof the observed values. Row 2 was then dereddenedusing the average H β and H α fluxes to produce row3. Rows 2 and 3 represent the global spectroscopicproperties that would be measured for 30 Dor if it werespatially unresolved.Table 5 is the similar data product from the SOARspectra, computed in exactly the same way as for Table4. The SOAR spectra are presented in a separate tablebecause there are a different number of columns dueto the additional emission lines which were measured,including [S III] λ λ λ λ λ . The complete version ofTables 4 and 5 are available in electronic form at thatsite, and also on the ApJS website.
4. Observational Results4.1. Overview
Figures 7-12 are maps of selected emission lines,nebular diagnostics, and physical conditions derivedfrom those emission lines, interpolated in the spatialplane of the Blanco data cube. The two extreme slitpositions, 16 and 17, are excluded from these maps.These maps are rotated 13 deg with respect to the N-S and E-W directions, and are labeled with offsets inthat rotated coordinate system. An outline of the areacovered by these maps is shown on Figure 2.Figure 7 shows the dereddened H α surface bright-ness and demonstrates the effective resolution of thegrid interpolated from the ’data cube’. It can be com-pared with the SOAR H α image in Figure 1. Most fea-tures, including the bright arcs centered around NGC2070, are visible. To the east the cavity-like region oflow surface brightness as well as the bright rim on itseastern edge are clearly seen.The ionization of the gas is traced by intensity ratioswhich include ([O III] λ β (Figure 9), ([N II] λ α (Figure 10), ([S III] λ λ λ α (Figure 12). Fig-ure 12 shows the ratio derived from the Blanco spectra,as opposed to the ratio made from the SOAR imagesthat is shown in Figure 2. These line ratios consistentlyshow a rough circular symmetry around a point about40 arcsec E of R136. The same distribution was iden-tified by I09 from a Spitzer Space Telescope surveyof 30 Dor using higher ionization lines. This will bediscussed further below. The electron kinetic temperature is computed fromthe [O III] λ λ λ ≤ T e ≤ h T e i for 30 Dor.For the most direct comparison to distant, unresolved8EHRs, we should spatially integrate, over the wholenebula, all of the light in each emission line, and thenapply a single reddening correction based on the ratioof the spatially integrated H α and H β fluxes. Fromthose reddening-corrected line strengths we can thenfind h T e i using the observed [O III] ( λ λ λ R OIII given by R OIII = R F (4959 ˚ A + 5007 ˚ A ) d Ω R F (4363 ˚ A ) d Ω (4). This method uses the total line fluxes listed for Posi-tion 38, row 3 in Table 4, and yields an equivalent tem-perature of h T e i = 10,680 ± K ( R OIII = 172 ± . ).Here we have used the error bars on the measured linestrengths derived as explained in the previous section.An alternative is to use [O III] line fluxes that havefirst been individually dereddened at each point onthe nebula and then spatially integrated. That methodleads to h T e i = 10,760 ± K ( R OIII = 168 ± . ). Thiscorresponds to the values listed for Position 38, row 1in Table 4.When individual measurements of T e have beenmade at multiple locations in a nebula, as is the casehere, another common approach is to compute h T e i = R T e ( ~r ) × F ( Hβ ) d Ω R F ( Hβ ) d Ω (5)where T e is weighted by the H β flux at each posi-tion. Using this method we again find h T e i = 10,680K, equivalent to R OIII = 172.For comparison, using this last method, Krabbe & Copetti(2002) found h T e i = 10,270 K ( R OIII = 195) usingmeasurements along all three of their slit positions. Aneven lower h T e i = 9,990 K ( R OIII = 209) was foundby Tsamis et al. (2003) who summed over the lengthof a 160 arcsec slit. Even though the formal standarderror of the mean of each of these mean temperatures is ∼
10 K, the true uncertainties will be somewhat larger.To explore the source of the discrepancy betweenour results and the previous measurements of h T e i , wecoadded all of our Blanco spectra to produce a singlespectrum with very high signal/noise ratio, excludingonly the regions containing the brightest stars. The en-tire spectrum was then dereddened as in the first ofthe averaging methods described above. We then mea-sured R OIII and computed the associated h T e i in sev-eral ways.First, using our automated line measuring softwareon this spectrum produced h T e i = 10,565 K ( R OIII = 178), in reasonable agreement with the value h T e i = 10,680 K obtained from the first of the averagingmethods described above.We then investigated the way in which our pro-gram sets the continuum level. The automated soft-ware fits the continuum in windows covering the wave-length ranges λλ β ) and λλ λ R OIII and h T e i , verifyingthat our automated routine works correctly. However,inspection of the coadded spectrum showed that the re-gion between H β and [O III] λ λ λ h T e i = 10,280K ( R OIII = 194.6).We conclude that the value h T e i = 10,270 K foundby Krabbe & Copetti (2002) agrees with our corre-sponding value h T e i = 10,673 K to within the measure-ment uncertainties due just to the issue of where thecontinuum is drawn. This uncertainty in the measuredtemperature translates to an uncertainty in the derivedoxygen abundance. There would be a 12 percent (0.05dex) increase in the derived O/H abundance ratio if thetemperature were decreased from 10,680 K to 10,280K. The λ λ σ = 50 per cent (0.29 in the log). Although thisscatter is much larger than the uncertainty computed ateach point from the photon counting statistics, we in-terpret it as being dominated by the true point-to-pointmeasurement uncertainties.We cannot resolve the discrepancy between thetemperatures measured by Tsamis et al. (2003) andthose measured here and by Krabbe & Copetti (2002).However, we are confident that our measurements arecorrect to within the uncertainties discussed above. The observed kinetic temperature provides an im-portant constraint to the energy sources that ionize thegas. The violent history of 30 Doradus is evident in thediffuse X-ray emission that is present throughout thenebula (Townsley et al. 2006), as well as in the high-velocity expansion features with speeds up to 200 km9 − seen in H α emission (Chu & Kennicutt 1994).These show that supernovae and strong winds frommassive O and WR stars have combined to heat gasto the observed 3.5 – 7 × K temperatures derivedfrom Chandra X-ray spectra. Despite these indicationsof high-velocity flows, it has long been known that themoderate K temperatures in the ionized gas do notindicate shock heating as the current source of ioniza-tion.The absence of [O IV] or [Ar V] emission lines(I09) as well as of any significant detection of nebu-lar He II emission has been used to argue that strongshocks are not an important ionization mechanism in30 Doradus. Instead, photoionization is implied as theenergy input mechanism. However, this still leavesopen the question of the source of the ionizing pho-tons, which we now consider.
5. Structural Details
The geometry and ionization structure of 30 Dorare quite complex. The SOAR direct images providea powerful tool for distinguishing edge-on ionizationfronts (IFs), regions ionized by localized sources ofradiation besides the central cluster, optically thickpillars like those found in M16 (Hester 1996), andline-emitting foreground structures. In particular, thenearly reddening-free [S II]/H α image (Figure 2),formed from the ratio of the SOAR images in thosetwo passbands, offers considerable insight into the ion-ization structure on many scales. A key question is the degree to which the centralcluster dominates the ionizing radiation field at differ-ent positions in the nebula. This will depend on theextent to which isolated O stars provide additional lo-cal contributions to the radiation field. Such stars canbe recognized by localized, circular variations in theionization level as indicated in the [S II]/H α maps,in combination with brightness enhancements in theH α emission. We carefully searched the SOAR im-ages for such features, and located the 49 isolated starslisted in Table 6. The spectral types of known starswithin the affected region are listed. Sources identi-fied in previous surveys with with unknown spectraltypes are listed as unknown and previously unidenti-fied objects are left empty. Of the 49 sources identi-fied, 19 have stars with known spectral types. Thereare 4 WR/WN, 7 O and 6 B type stars. The later O and early B type stars produce less ionizing radi-ation than more massive stars and may be illuminat-ing some of the remnant material from their forma-tion still present due to weak winds. The other objectsidentified with unknown spectral types are typically IRsources. Given the fraction of massive stars identifiedas embedded objects, the unknown sources should befollowed up to determine if they too are massive stars.Some percentage of these is likely to have been se-lected by chance, so this estimate should represent anupper limit to the number of locally ionizing stars. Thearea of the nebula in which the ionization level is sig-nificantly affected by the radiation field of these stars,as projected on the sky, covers only 2 per cent of thefull nebula. This means that it will not be an importantcontaminant in our analysis, in Paper II, of the overallproperties of the nebula. There are numerous bright fingers and protrudingIFs scattered over the face of the nebula, but they areparticularly concentrated in the large X-ray emittingcavity to the East of R136. Many of these are prob-ably elephant trunks similar to the famous “Pillars ofCreation” in M16. A small number of these have beencommented on by Scowen et al. (1998). These struc-tures are easily identified using a combination of theH α , [S II] and 8 µ m images. Table 7 catalogs 106 ofthe brightest such features, listing their RA and Decpositions, projected length and PA. However, there arehundreds more that blend together as one looks onfiner and finer scales, until they are indistinguishablefrom small examples of the edge-on ionization frontsdiscussed above. Almost all of these point back toR136, strongly suggesting that the central cluster is thedominant source of ionizing radiation.Many of these features are clearly connected tolarger bodies of molecular gas, as can be seen by com-paring our [S II]/H α ratio image to the 8 µ m im-age from the Spitzer Space Telescope archives, whichshows PAH emission (Figure 13). However, someof them do appear to be isolated tubes or blobs ofgas rather than protrusions from background walls ofmolecular material. Scowen et al. (1998) have com-mented on one bright structure of this latter type. Someof these might correspond to the expanding shells seenin H α emission in the high-resolution echelle spectraof Chu & Kennicutt (1994).10 .3. Edge-on ionization fronts and the source ofthe ionizing radiation Edge-on ionization fronts stand out in the [S II]/H α im-age as narrow linear structures with high [S II]/H α ra-tio, butted up against regions of low [S II]/H α . Table8 lists prominent, isolated, edge-on ionization frontsfound both in our [S II]/H α image and in the Spitzer 8 µ m PAH mosaic. The structures identified in both theoptical and IR passbands are perfectly suited for de-tailed, high angular resolution (e.g. with HST and/orALMA) studies of the neutral and molecular gas be-yond the IF, commonly called the Photo-DissociationRegion or Photon Dominated Region (PDR). A find-ing chart of all IFs listed in Table 8 is shown on a[S II]/H α image in Figure 14.With our new observations there is enough informa-tion available about several of these IFs to test whetheror not they are likely to be photoionized by the centralcluster including R136. The thickness of the ionizedgas layer dr , depends on the density n and the dis-tance of the gas from the source of ionizing radiation r according to dr ∝ Q παn e n p r (6)where Q is the ionizing photon luminosity by num-ber, α = 2 . × − cm − s − is the H + recombina-tion rate, and we assume dr ≪ r . The test consists ofcomputing Q values for different IFs at very differentdistances from R136, assuming that R136 marks thecenter of the ionizing radiation, and seeing if the dif-ferent Q measurements agree with each other. Therewill be considerable uncertainty due to the implicitassumption that the projected distance from R136 toeach IF is the true three-dimensional distance, so wedo not carry error bars through the analysis and willregard agreement at a factor of two level to be satisfac-tory.Two IFs of interest for this test, IF1 and IF2, arefairly close to 30 Dor (see Figure 14). Both have aspatially resolved layer of H + separated from an edge-on IF seen in [S II], similar to the Orion Bar (see P09).In the case of IF1, there is a line of stars coinci-dent with the IF that could be an alternative to NGC2070 as the source of ionization, or which could bestars formed in a region of gas that was compressed byradiation and wind pressure from the cluster, or whichcould simply be foreground stars. Most of these starsfell in one or another of our slit positions. To gauge the influence these stars may have on the ionizationstructure of this region, we have examined the stel-lar spectra and found them to lack any absorption fea-tures. This would indicate they are massive stars; how-ever this result is uncertain due to the brightness of theemission lines coincident with the expected absorptionfeatures. If these stars were embedded in the gas wewould expect the stellar spectra to be significantly red-dened, but with the possible exception of one star, thisdoes not appear to be the case. Additionally, we notethat the overall structure of IF1 is oriented almost ex-actly perpendicular to the direction to R136. There arealso many prominent pillars in this region, all point-ing back up to R136, suggesting the nearby stars areunimportant to the structure.Our SOAR spectrographic slit position number 40cuts directly across IF1, avoiding the pillars noted byScowen et al. (1998) and targeting the larger structure.The emission-line strengths in the region of IF1 weremeasured as described above, with the exception thatthe extraction window is smaller (0.75 arcsec) alongthe slit to better resolve the detailed structure. Figure15 shows the clear separation of the highly ionized gasfrom the lower-ionization gas in IF1. In the top panelare the relative intensities of H α , [O III] λ λ n e ≃ cm − ,increasing to 800 cm − at the peak of the [S II] emis-sion, and then decreasing to a minimum of 200 cm − .There is a bright knot in the [S II] images at the lo-cation of the peak density. Excluding this point, thedensity at the IF is likely to be 560 cm − . Using thisdensity, assuming a central ionizing source, and thatthe 13.6 pc (58 arcsec) projected distance from R136is the true three-dimensional distance and that the 1.18pc projected thickness of the H + zone is the true thick-ness, Eq. 6 requires Q = 5 . × s − . In addi-tion to the uncertainties in projected vs. true distances,this value neglects the effect the observed gradient willhave in lowering the measured average n e .The edge-on front IF2 forms the inner edge of oneof the most prominent bright arcs, located at a pro-jected distance of 17 pc (70 arcsec) NE of R136. Thethickness of the H + layer perpendicular to the direc-tion to R136 is approximately 0.94 pc. The density atthe IF along a ray from R136 into IF2 is n e = 280 ± cm − . Applying Eq. 6 with the same assumption used11ith IF1, we find Q = 2 . × s − .IF4 is a bright rim lying much further out in thenebula, at a projected distance 53 pc (220 arcsec) tothe east of R136 on the far side of the large faint low-density region that is easily seen in figures 7-12. Thislarge wall is remarkably homogeneous in density andhas a thickness between 3.5 and 4 arcsec all along itsprojected length of 33 pc (140 arcsec). The limb showsup as a density enhancement, with n e = 125 ±
30 cm − .It has a nearly constant projected distance from R136.Using the previous technique we find Q = 4 . × s − .The above estimates of Q are all in reasonableagreement with each other and also with Q =4 . × s − estimated by Crowther & Dessart(1998) from adding up the contributions expected fromthe individual stars in the cluster. These results sup-port the conclusion that the central cluster of starsdominates the ionizing radiation field throughout mostof the volume of 30 Dor, and that isolated stars providea negligible contribution out to 100pc.However, there is at least one region, to the SWof R136, where the observed structure indicates ion-ization by a different source. The IFs designated 1*through 5* on Figure 14 show PAH emission closerto R136 than the [S II] emission. The expected strat-ification of PAH and H + emission are consistent withphoto-ionization by stars in the region of the OB as-sociation LH 99 region. These IFs form an elongatedshell structure with a diameter of 15 pc around thatcluster. If this interpretation is correct, the emission-line spectrum produced from a different ionizing SEDwould be different from the bulk of the emission from30 Dor. Unfortunately, our spectroscopic survey doesnot cover this particular region, so we cannot make thistest.Interestingly there are examples of regions that ap-pear in the optical data sets as IFs which have no as-sociated PAH emission. One such example is a shell-like structure located at RA = 05 h m s , Dec = -69 ◦ α images, but is present in the[S II]/H α or [S II]/[O III] ratio images as a prominentbubble-like structure north of IF 13. However there isno associated feature seen in the 8 µ m Spitzer images atthis location. This cannot be due to obscuration, whichwould affect the optical data more severely. Unfortu-nately this region is just beyond the boundary of our spectrophotometric survey and we can draw no furtherconclusions about it. The various surface brightness and ionization mapsshown in Figures 7 to 12 all have roughly circular sym-metry. The center in each case is at a point offset about1 arcmin to the E of R136. While at first sight thismight seem counter to the idea that R136 marks thecenter of the source of ionizing radiation, the offset ismost likely an artifact of the asymmetrical distributionof gas around R136. The extensive bubble of hot x-ray emitting gas extending off in this direction appearsto be a ‘blow-out’ like the ones observed on a smallerscale in M17 (Townsley et al. 2003; Pellegrini et al.2007) and elsewhere. We conclude that the asymmet-rical density distribution has just modestly shifted theapparent center of the ionization structure that other-wise would be centered on R136.
In Paper II we will compare the chemical abun-dances obtained from a point-by-point analysis usingour new data set, to the properties that would be de-rived if 30 Dor were viewed at a cosmological distanceand its integrated light analyzed using the ‘strong-line’analysis techniques (Pagel 1979; Kewley & Doptia2002; Denicol´o et al. 2002) generally used in suchcircumstances.It is important to not be misled by the measuredproperties of the prominent bright arcs that have given30 Dor the name ‘Tarantula Nebula’. The majorityof the line emission originates in the extended re-gions of low surface brightness. This is demonstratedin Figure 16, a histogram of the fraction of nebularflux from regions with a surface brightness ≤ S ( Hα ) for our 12 ×
13 arcmin sky- and continuum-subtractedSOAR H α image. The bright arcs have a character-istic H α surface brightness S ( H α ) ≥ × − ergcm − s − arcsec − , while 50 per cent of the integratedH α emission over our field originates from regionswith a surface brightness an order of magnitude fainterthan the bright arcs.
6. Conclusions
We have obtained and are making publicly availablea dense grid of long-slit optical spectra which measurekey emission lines in the λλ ×
80 pc (10 × ) region of30 Dor. These spectra were taken at 37 slit positionsand then extracted and measured as 4238 individual1D spectra. The resulting ‘data cube’ of emission-lineintensity measurements is also publicly available. Sup-plementary spectra at a few additional slit positions ex-tended the coverage to a wider wavelength range.We also obtained a set of subarcsec-resolution di-rect images in multiple narrow-band filters, coveringa 170 ×
180 pc (12 ×
13 arcmin ) field of view, whichwe have used to identify and catalog a large numberof structures of special interest. These include regionslikely to be locally ionized by embedded stars, as wellas edge-on ionization fronts seen in both the visibleand infrared and ‘elephant trunk’ pillars. With JWSTand ALMA on the horizon and the newly upgradedHST, these well-resolved structures in this nearby ob-ject will provide important calibrations for constrain-ing the physical conditions in the PDRs of much moredistant, unresolved low-metallicity extragalactic H IIregions.We find that the cluster of O stars centered on R136is the dominant source of ionization for almost the en-tire nebula. Our measurements of the [O III] temper-ature does not reveal any shock-heated low-ionizationgas. The observed emission-line intensity ratios indi-cate that photoionization is the predominant ionizationmechanism. The general radial symmetry of these in-tensity ratios around a point on the sky near R136 (Figs9-12) supports the idea that the central cluster is themajor source of this photoionization.Further evidence includes the absence of strongcompetition from other potential sources of ioniz-ing radiation within 15 pc of R136 (Sect. 5.1), thelarge numbers of elephant trunks pointing back to-wards R136 from points all over 30 Dor (Sect. 5.2),and the variation in the measured thickness of edge-onionization fronts as a function of distance from R136(Sect. 5.3). In a follow-up paper we will use theseresults as the basis for detailed models of the structureof 30 Doradus.We also showed that the integrated emission-linespectrum of 30 Dor is dominated by the extensive re-gions of low surface brightness. The bright arcs nearR136 catch the eye when one looks at direct images of30 Dor, but are not what would actually be measuredspectroscopically if 30 Dor were so far away that itcould not be spatially resolved.
7. Acknowledgments
EWP and JAB gratefully acknowledge supportfrom NSF grant AST-0305833 and NASA grants HSTAR-10932 and NNX10AD05G. JAB gratefully ac-knowledges additional support from NASA ADP grantNNX10AD05G. EWP gratefully acknowledges sup-port from NASA grant 07-ATFP07-0124 and partialsupport from NSF grant AST-0806147. GJF gratefullyacknowledges support from NASA grant 07-ATFP07-0124 and from NSF through 0607028 and 0908877.
REFERENCES
Bosch, G., Terlevich, R., Melnick, J., & Selman, F.1999, A&AS, 137, 21Bosch, G., Selman, F., Melnick, J., & Terlevich, R.2001, A&A, 380, 137Breysacher, J. 1981, A&AS, 43, 203Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989,ApJ, 345, 245Chu, Y.-H., & Kennicutt, R. C., Jr. 1994, ApJ, 425, 720Clemens, J. C., Crain, J. A., & Anderson, R. 2004,SPIE, 5492, 331Crowther, P. A., & Dessart, L. 1998, MNRAS, 296,622Denicol´o, G., Terlevich, R., & Terlevich, E. 2002,MNRAS, 330, 69Elliott, K. H., Goudis, C., Meaburn, J., & Tebbutt,N. J., 1997, A&A, 55, 187Elmegreen, B. G., & Lada, C. J. 1977, ApJ, 214, 725Hamuy, M., Suntzeff, N. B., Heathcote, S. R., Walker,A. R., Gigoux, P., & Phillips, M. M. 1994, PASP,106, 566Hunt, L. K., & Hirashita, H. 2009, A&A, 507, 1327Indebetouw, R., et al. 2009, ApJ, 694, 84 (I09)Kewley, L. J., & Dopita, M. A. 2002, ApJs, 142, 35Krabbe, A. C., & Copetti, M. V. F. 2002, A&A, 387,295Macri, L. M., Stanek, K. Z., Bersier, D., Greenhill,L. J., & Reid, M. J. 2006, ApJ, 652, 113313athis, J. S., Chu, Y.-H., & Peterson, D. E. 1985, ApJ,292, 155Meaburn, J. 1984, MNRAS, 211, 521Meixner, M., et al. 2006, AJ, 132, 2268Osterbrock, D. E., & Ferland, G. J. 2006, Astro-physics of gaseous nebulae and active galactic nu-clei, 2nd. ed. by D.E. Osterbrock and G.J. Fer-land. Sausalito, CA: University Science Books,2006Oey, M. S., Parker, J. S., Mikles, V. J., & Zhang, X.2003, AJ, 126, 2317Oey, M. S., Watson, A. M., Kern, K., & Walth, G. L.2005, AJ, 129, 393Pagel, B. E. J., Edmunds, M. G., Blackwell, D. E.,Chun, M. S., & Smith, G. 1979, MNRAS, 189, 95Parker, J. W., & Garmany, C. D. 1993, AJ, 106, 1471Parker, J. W. 1993, AJ, 106, 560Peimbert, A. 2003, ApJ, 584, 735Pellegrini, E. W., Baldwin, J. A., Ferland, G. J., Shaw,G., & Heathcote, S. 2009, ApJ,693, 285 (P09)Pellegrini, E. W., et al. 2007, ApJ, 658, 1119Poglitsch, A., Krabbe, A., Madden, S. C., Nikola, T.,Geis, N., Johansson, L. E. B., Stacey, G. J., & Stern-berg, A. 1995, ApJ, 454, 293Robert, C., Pellerin, A., Aloisi, A., Leitherer, C.,Hoopes, C., & Heckman, T. M. 2003, ApJS, 144,21Schaller, G., Schaerer, D., Meynet, G., & Maeder, A.1992, A&AS, 96, 269Scowen, P. A., et al. 1998, AJ, 116, 163Selman, F., Melnick, J., Bosch, G., & Terlevich, R.1999, A&A, 341, 98Schnurr, O., Moffat, A. F. J., Villar-Sbaffi, A., St-Louis, N., & Morrell, N. I. 2009, MNRAS, 395,823Smith, R. C., & MCELS Team 1998, Publications ofthe Astronomical Society of Australia, 15, 163 Townsley, L. K., Broos, P. S., Feigelson, E. D., Brandl,B. R., Chu, Y.-H., Garmire, G. P., & Pavlov, G. G.2006, AJ, 131, 2140Townsley, L. K., Feigelson, E. D., Montmerle, T.,Broos, P. S., Chu, Y.-H., & Garmire, G. P. 2003,ApJ, 593, 874Tsamis, Y. G., Barlow, M. J., Liu, X.-W., Danziger,I. J., & Storey, P. J. 2003, MNRAS, 345, 186Walborn, N. R., & Blades, J. C. 1997, ApJS, 112, 457Walborn, N. R., Ma´ız-Apell´aniz, J., & Barb´a, R. H.2002, AJ, 124, 1601Zavagno, A., et al. 2010, A&A, 518, L101
This 2-column preprint was prepared with the AAS L A TEX macrosv5.2. α image of 30 Dor taken with the SOAR telescope. North is rotated 13 deg clockwisefrom vertical. The center of R136 is marked as a white “ × ” symbol. (a) The outline of the region covered by our mapsmade from the Blanco spectra. (b) The individual slit positions of our Blanco spectroscopic data set.15ig. 2.— Ratio of [S II]/H α from our SOAR narrow-band images. Darker shades indicate a higher ratio. Theorientation is the same as Figure 1. For reference, the SOAR spectroscopic slit positions are plotted on top of theimage. 16 [O I] 6300Å[S III] 6312Å F l x ( e r g s − c m − a r c s ec − Å − ) Wavelength (Å) 7300 7320 7340 7360 7380 [O II] 7320Å[O II] 7330ÅSky subtractedNebula + Sky
Fig. 3.— A sample SOAR spectrum near [O I] λ λλ λ F l x ( e r g s − c m − Å − ) Wavelength (Å)[S II] 6717Å[S II] 6731ÅHe I 6678Å
Fig. 5.— Two examples of extraction windows used to measure line flux. The dashed horizontal bars show the rangesover which the continuum was fitted. The vertical lines show the ranges over which the line fluxes were summed. HeI λ N / N t o t ratio [O III]/H b [S II]/H a Fig. 6.— Repeatability of results at overlapping points along different Blanco slit positions. The curves are histogramsof the distributions of ratios of intensity ratios; for example the [OIII]/H β measured from one slit position is dividedby the [OIII]/H β measure at the same position on the sky but from a different slit position.19 og(S(H a )-300-200-100 0 100 200 300 D x (arcsec)-150-100-50 0 50 100 150 D y ( a r cs e c ) -14.50-14.00-13.50-13.00-12.50-12.00 Fig. 7.— Interpolated dereddened H α surface brightness in erg s − cm − arcsec − . The region shown is the same asthat outlined in Figure 1 and Figures 8 through 12. R136 is at ∆ x = 0, ∆ y = 0.20 V -300-200-100 0 100 200 300 D x (arcsec)-150-100-50 0 50 100 150 D y ( a r cs e c ) Fig. 8.— Interpolated A V . 21 og([O III]/(H b ))-300-200-100 0 100 200 300 D x (arcsec)-150-100-50 0 50 100 150 D y ( a r cs e c ) Fig. 9.— Log([O III]/H β ). 22 og([N II]/H a )-300-200-100 0 100 200 300 D x (arcsec)-150-100-50 0 50 100 150 D y ( a r cs e c ) -2.00-1.80-1.60-1.40-1.20-1.00-0.80-0.60 Fig. 10.— Log(([N II] λ α ). 23 og([S III]/[SII])-300-200-100 0 100 200 300 D x (arcsec)-150-100-50 0 50 100 150 D y ( a r cs e c ) -1.40-1.20-1.00-0.80-0.60 Fig. 11.— Log([S III] λ λ λ og([S II]/H a )-300-200-100 0 100 200 300 D x (arcsec)-150-100-50 0 50 100 150 D y ( a r cs e c ) -1.80-1.60-1.40-1.20-1.00-0.80-0.60 Fig. 12.— Log(([S II] λ λ α ). 25ig. 13.— Left SOAR [S II]/H α ; Right SPITZER 8 µ m PAH. A selection of bright pillars are shown with arrowsindicating their location and direction. These dense IFs are detected in both optical and IR passbands and show aconnection with the background molecular cloud. 26ig. 14.— H α (upper) and [S II]/H α (lower) images showing the positions of the ionization fronts (IFs) cataloguedin Table 8. The approximate center of each IF is at the tip of the arrow marking it. The arrows point in the directiontowards higher ionization within each I-front, and except for IF16 are approximately perpendicular to the IF. On themain image, darker shades indicate larger values of [S II] /H α , but with the grey scale adjusted in the central box toshow regions with lower [S II] /H α ratios. The white X in each image marks the position of R136.27 .00.20.40.60.81.01.21.4 50 55 60 65 70 75 80H a [OIII]5007[SIII] 90690.00.20.40.60.81.01.21.4 50 55 60 65 70 75 80[SII][OII][NII]02004006008001000 50 55 60 65 70 75 80Projected distance from R136 (arcsec)n e cm -3 Fig. 15.— The profile of IF1. Top: Ionized gas is traced by H α , [OIII] and [S III] emission. Middle: The IF is tracedby [S II], [N II] and [O II]. Bottom: The electron density profile in cm − measured from [S II] as described in the text.28 C u m u l a ti v e F r ac ti on o f T o t a l H a F l ux S(H a ) erg s -1 cm -2 arcsec -2 Fig. 16.— A cumulative histogram of the fraction of total H α flux from pixels with ≤ S ( Hα ) from our sky- andcontinuum-subtracted H α image over a 12 ×
13 arcmin field of view centered on R136.29
ABLE UMMARY OF IMAGING OBSERVATIONS USED IN
30 D
ORADUS OPTICAL MOSAICS .Filter ( ˚ A ) No. of Exp. × Duration No. Pos. FWHM (arcsec) Min/Max Exp Time (s)at each position6563 3 × × × × × ABLE UMMARY OF B LANCO AND
SOAR
SPECTROSCOPIC OBSERVATIONS OF
30 D
ORADUSPos R.A. (2000)(slit center) Dec. (2000) PA(deg E of N) Exp. Time (s)Blanco1 05:39:16.53 -69:06:37.04 13 4052 05:39:11.25 -69:06:31.18 13 4253 05:39:06.29 -69:06:29.68 13 4054 05:39:00.44 -69:06:22.80 13 2505 05:38:56..29 -69:06:11.60 13 5006 05:38:52.21 -69:06:11.81 13 2607 05:38:45.00 -69:06:03.65 13 2608 05:38:41.38 -69:05:55.66 13 3209 05:38:37.98 -69:05:53.44 13 30010 05:38:35.17 -69:05:48.71 13 12011 05:38:30.08 -69:05:38.24 13 40012 05:38:24.82 -69:05:29.57 13 45013 05:38:17.42 -69:05:28.52 13 50014 05:38:08.61 -69:05:18.33 13 30015 05:37:59.70 -69:04:56.19 13 25016 05:38:49.47 -69:01:28.18 13 60017 05:38:37.48 -69:08:52.47 13 60020 05:39:16.50 -69:04:51.76 103 40021 05:38:28.34 -69:03:50.99 103 40022 05:39:15.65 -69:05:28.77 103 40523 05:38:27.16 -69:04:29.19 103 30824 05:39:12.49 -69:05:54.62 103 50025 05:38:25.24 -69:04:55.67 103 35526 05:39:12.16 -69:06:22.34 103 50027 05:38:25.90 -69:05:23.89 103 46028 05:39:06.04 -69:06:36.65 103 60029 05:38:22.74 -69:05:43.05 103 52030 05:39:07.60 -69:07:05.02 103 70031 05:38:24.16 -69:06:13.86 98 45032 05:39:06.61 -69:07:26.38 103 90033 05:38:21.35 -69:06:31.58 103 50034 05:39:06.01 -69:07:58.72 103 40035 05:38:20.45 -69:07:00.61 103 30036 05:39:01.11 -69:08:34.64 103 60037 05:38:15.70 -69:07:38.32 103 450SOAR40 05:38:46.430 -69:05:35.06 10 100041 05:39:03.142 -69:06:40.45 33.9 300042 05:39:02.763 -69:06:45.82 61 200043 05:39:21.240 -69:06:55.44 61 30044 05:39:08.649 -69:07:02.80 80 3000N
OTE .—Columns 1-5 are position number corresponding to Figures 1 and 2, RA and Declination of the slit center, PA and total integratedexposure time. Positions 1-37 are Blanco observations. 40-44 are SOAR observations. ABLE INE
ID’
S AND WAVELENGTHS λ Obs λ ion f λ ( R = 3.1) Data Set4106 4101 H I 1.43 SOAR4344 4340 H I 1.35 Blanco+SOAR4366 4363 [O III] 1.34 Blanco+SOAR4476 4471 He I 1.30 Blanco+SOAR4800 4800 Continuum 1.19 Blanco+SOAR4866 4861* H I 1.16 Blanco+SOAR4926 4922 He I 1.14 SOAR4963 4959 [O III] 1.13 Blanco+SOAR5012 5007* [O III] 1.12 Blanco+SOAR5625 5625 Continuum 0.97 Blanco+SOAR5880 5875* He I 0.93 Blanco+SOAR6308 6300 [O I] 0.86 SOAR6318 6312 [S III] 0.86 Blanco+SOAR6364 6364 NS [O I] - Blanco6552 6548 [N II] 0.82 Blanco+SOAR6570 6563 H I 0.82 Blanco+SOAR6590 6584* [N II] 0.81 Blanco+SOAR6684 6678 He I 0.80 Blanco+SOAR6721 6716* [S II] 0.79 Blanco+SOAR6738 6731* [S II] 0.79 Blanco+SOAR7072 7065* He I 0.74 Blanco+SOAR7142 7135* [Ar III] 0.73 Blanco+SOAR7288 7281 He I 0.71 SOAR7325 7320 [O II] 0.70 SOAR7337 7330 [O II] 0.70 SOAR7758 7751 [Ar III] 0.63 SOAR9075 9069 [S III] 0.48 SOARN OTE .—Measured strengths for each of these lines are listed in Table 4 and/or 5 at every extracted point in the nebula. Rest frame wavelengthswith an asterisk indicate the lines used to fit the Blanco data to models at each point in the nebula, as described in Paper II. ABLE AMPLE OF B LANCO SPECTROSCOPY DATA CUBE1 2 3 4 5 6 7 8 9 10 11 12 13Pos ∆ RA ∆ Dec row log n e -err +err T(O3) -err +err A V F(H β ) I(H β )1 2 3 4 5 6 7 8 9 10 11 12 131 148.1 -181.6 3 1.41 0.41 0.27 13600 1010 1500 1.12 1.94E-15 6.56E-151 148.6 -179.1 8 1.70 0.24 0.16 10900 1280 3210 1.13 1.88E-15 6.38E-15 · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·
37 3.9 -128.5 603 1.88 0.11 0.09 13900 1320 2280 2.86 2.70E-15 6.05E-1438 0.0 0.0 1 2.11 0.001 0.001 10800 8 8 0.00 6.22E-14 6.21E-1438 0.0 0.0 2 2.08 0.001 0.001 0 4 4 1.28 3.55E-15 1.43E-1438 0.0 0.0 3 2.08 0.001 0.001 10700 5 5 0.00 5.75E-16 5.75E-14 σ I ( Hβ ) σ σ σ σ σ σ
14 15 16 17 18 19 20 21 22 23 24 25 265.65E-17 48.42 1.37 4.97 1.07 2.42 0.93 43.52 0.98 114.01 0.83 334.02 1.227.18E-17 44.08 1.82 2.91 1.45 0.07 1.28 48.43 1.34 117.72 1.09 357.57 1.59 · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ABLE Continued σ σ σ σ σ σ · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · σ σ σ σ σ σ σ
40 41 42 43 44 45 46 47 48 49 50 51 520.22 284.89 0.77 16.36 0.28 2.02 0.20 25.73 0.29 18.31 0.27 2.21 0.180.21 284.89 0.67 16.15 0.25 2.51 0.19 23.84 0.26 17.31 0.24 2.71 0.17 · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ABLE Continued σ
53 549.4 0.24 · · · · · ·
OTE .—The complete version of this table, with 4241 rows of data, can be found in the electronic version of the journal. The units of electrondensity and temperature are cm − and K, respectively. The dereddened H β surface brightness are reported in units of erg s − cm − arcsec − .The strengths of the other emission lines are given in units of 100 × S(line)/S(H β ). The entries for position number 38 are average values for thewhole data set, as described in the text. ABLE AMPLE OF
SOAR
DATA SPECTROSCOPY DATA SET1 2 3 4 5 6 7 8 9 10 11 12Pos ∆ RA ∆ Dec row log n e -err +err T(O3) -err +err T(S3) -err1 2 3 4 5 6 7 8 9 10 11 1240 4.65 -72.58 260 2.34 0.06 0.05 13100 582 728 11100 60340 5.03 -70.42 275 2.43 0.08 0.08 14100 1120 1720 9250 100040 5.41 -68.26 290 2.67 0.06 0.06 11500 1050 1890 10900 917 · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · +err A V F(H β ) I(H β ) σ I ( Hβ ) σ σ σ · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · σ σ σ σ σ σ · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ABLE Continued σ σ σ σ σ σ · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · σ σ σ σ σ σ · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · σ σ σ σ σ σ
61 62 63 64 65 66 67 68 69 70 710.067 0.572 0.044 1.181 0.043 1.403 0.044 1.973 0.044 18.540 0.1020.105 0.753 0.075 1.264 0.073 1.459 0.071 2.137 0.073 17.843 0.1450.096 0.613 0.068 1.225 0.066 1.304 0.063 2.066 0.067 18.485 0.134 · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · N OTE .—The H β fluxes are in units of erg s − cm − arcsec − . T(S3) is the temperature derived from the [S III] λ λ ABLE EGIONS OF POSSIBLE LOCAL ENHANCEMENT OF IONIZATION DUE TO NEARBY MASSIVE STARSObject ID RA Dec ∆ RA ∆ Dec radius Spec. Type Ref.(arcsec) (arcsec) (arcsec)1 05:39:03.44 -69:06:35.72 113 -32 3.97 B0.2V · · · · · · · · · · · ·
10 05:38:41.19 -69:02:57.75 -7 186 3.07 O7V
11 05:38:45.39 -69:02:50.84 16 193 3.83 O4V
12 05:38:13.98 -69:07:47.63 -153 -104 10.11 unkown13 05:39:20.55 -69:06:54.42 205 -51 4.06 · · ·
14 05:39:03.36 -69:09:33.58 113 -210 3.02 · · ·
15 05:38:45.08 -69:08:08.06 14 -124 6.26 B1Ia +unknown16 05:38:52.82 -69:06:12.01 56 -8 3.81 O5.5V
17 05:38:55.79 -69:05:24.90 72 39 2.41 unknown18 05:38:04.84 -69:07:34.84 -202 -91 3.54 · · ·
19 05:38:31.75 -69:02:14.28 -57 230 3.59 unknown20 05:38:41.76 -69:01:58.79 -4 245 4.27 unknown21 05:38:53.54 -69:02:00.56 60 243 2.43 WN6
22 05:39:11.63 -69:02:01.02 157 243 3.07 WR
23 05:38:15.64 -69:04:37.64 -144 86 9.94 unknown24 05:38:54.94 -69:08:45.76 67 -162 3.77 B0.5Ia
25 05:38:56.83 -69:08:42.47 77 -159 4.25 A2/3Ia +unknown26 05:38:54.70 -69:07:45.81 66 -102 5.29 G8V
27 05:38:36.43 -69:06:58.70 -32 -55 2.98 WN
28 05:38:36.06 -69:06:47.52 -34 -44 2.04 O8.5V +B0.5III +B1Ia
29 05:38:51.32 -69:06:41.99 48 -38 2.84 unknown30 05:38:49.79 -69:06:44.12 40 -40 1.7 B0.5I
31 05:38:57.40 -69:07:10.75 81 -67 2.73 B3Ia
32 05:38:46.60 -69:04:28.07 22 96 2.3 O4V
33 05:38:36.91 -69:05:08.19 -30 56 5.97 · · ·
34 05:39:01.05 -69:06:30.16 100 -26 2.44 · · ·
35 05:38:58.80 -69:05:24.55 88 39 2.46 · · ·
36 05:38:17.62 -69:05:42.83 -133 21 13.67 · · ·
37 05:39:22.86 -69:07:46.88 217 -103 3.39 · · ·
38 05:39:12.35 -69:06:02.74 161 1 4.78 · · ·
39 05:38:14.88 -69:04:31.80 -148 92 2.31 · · ·
40 05:38:08.53 -69:05:44.36 -182 19 15.11 · · ·
41 05:38:10.64 -69:06:17.52 -171 -14 3.69 · · ·
42 05:38:09.46 -69:06:22.26 -177 -18 3.69 · · ·
43 05:37:49.22 -69:06:14.27 -286 -10 4.41 · · ·
44 05:38:24.70 -69:07:43.55 -95 -100 3.06 · · ·
45 05:38:30.51 -69:06:46.21 -64 -42 3.07 · · ·
46 05:38:29.73 -69:06:57.41 -68 -54 3.07 · · ·
47 05:38:31.02 -69:06:37.47 -61 -34 2.63 · · · N OTE .—Stellar spectral types are from Bosch et al 1999, Schnurr et al. 2009, Wolborn & Blades 1997, Robert et al. 2003, Breysacher1981. Objects found in surveys with no reported spectral type are listed as unknown. Objects with no known counterpart are left blank. ABLE CATALOG OF BRIGHT , DENSE PILLARS AND PROTRUDING IF SPillar ID RA (J2000) Dec ∆ RA ∆ Dec Length PA(arcsec) (arcsec) (arcsec)1 05:37:28.35 -69:01:51.86 -400 252 37 2162 05:37:30.68 -69:04:12.44 -384 111 17 1203 05:37:32.85 -69:04:16.79 -373 107 19 1154 05:37:42.14 -69:02:30.70 -325 213 26 2095 05:37:42.85 -69:04:42.79 -319 81 10 1766 05:37:42.95 -69:04:34.28 -319 90 10 1837 05:37:50.28 -69:06:02.92 -282 1 19 1558 05:37:54.67 -69:05:09.56 -255 54 9 1679 05:37:58.57 -69:03:44.17 -233 140 7 18210 05:37:59.29 -69:05:49.22 -233 15 8 16311 05:38:03.81 -69:05:22.05 -207 42 10 16412 05:38:05.74 -69:06:36.40 -196 -33 6 13813 05:38:08.17 -69:06:51.16 -185 -47 33 16714 05:38:09.49 -69:07:05.29 -180 -61 19 15915 05:38:11.95 -69:07:08.39 -164 -65 26 14216 05:38:13.19 -69:04:25.76 -158 98 13 23617 05:38:14.22 -69:07:00.28 -153 -56 10 13118 05:38:15.89 -69:05:25.19 -142 39 6 26019 05:38:17.03 -69:05:57.00 -137 7 3 11220 05:38:17.56 -69:05:02.66 -131 61 14 20021 05:38:24.12 -69:04:56.88 -99 67 24 20622 05:38:28.52 -69:08:26.14 -72 -142 16 16623 05:38:28.74 -69:03:37.29 -72 147 6 23824 05:38:30.07 -69:08:38.73 -67 -155 24 13825 05:38:33.14 -69:06:02.88 -51 1 5 15326 05:38:34.01 -69:03:33.01 -45 151 12 26827 05:38:34.26 -69:06:26.96 -45 -23 12 15328 05:38:36.05 -69:07:43.52 -35 -100 12 14829 05:38:38.27 -69:08:13.16 -24 -129 9 8130 05:38:40.51 -69:09:58.77 -8 -235 9 11131 05:38:40.97 -69:07:27.27 -8 -83 12 8432 05:38:43.16 -69:07:04.03 3 -60 7 9633 05:38:43.20 -69:01:19.02 3 285 15 17334 05:38:43.23 -69:10:21.48 3 -258 11 13235 05:38:44.16 -69:06:58.89 8 -55 5 9036 05:38:44.25 -69:11:10.68 8 -307 9 12737 05:38:45.18 -69:05:05.74 14 58 8 30438 05:38:45.41 -69:04:19.41 14 104 8 26839 05:38:45.65 -69:10:25.07 19 -261 18 10840 05:38:45.85 -69:09:51.28 19 -227 15 6241 05:38:46.15 -69:07:04.11 19 -60 8 8642 05:38:46.86 -69:09:33.58 25 -210 7 7343 05:38:47.05 -69:07:56.66 25 -113 13 6144 05:38:47.85 -69:07:15.15 30 -71 14 6545 05:38:50.55 -69:10:17.26 46 -253 16 7746 05:38:53.45 -69:08:00.23 57 -116 8 7047 05:38:54.71 -69:05:36.11 68 28 3 34348 05:38:55.01 -69:05:39.94 68 24 3 30949 05:38:55.29 -69:05:40.63 68 23 3 26750 05:38:56.40 -69:07:21.25 73 -77 13 30 ABLE Continued
Pillar ID RA (J2000) Dec ∆ RA ∆ Dec Length PA(arcsec) (arcsec) (arcsec)51 05:38:56.81 -69:05:30.58 78 33 10 29752 05:38:57.29 -69:01:56.18 78 248 5 3253 05:38:57.51 -69:06:27.65 84 -24 16 1654 05:38:57.81 -69:07:56.53 84 -113 19 6855 05:38:58.07 -69:01:51.67 84 252 7 2356 05:38:58.58 -69:02:16.27 89 228 9 23357 05:38:58.84 -69:07:28.60 89 -85 15 4358 05:38:59.48 -69:08:04.34 89 -121 37 6759 05:39:00.01 -69:11:45.42 95 -342 22 8360 05:39:00.14 -69:06:08.15 95 -4 6 34061 05:39:00.29 -69:08:56.01 95 -172 21 5262 05:39:02.13 -69:08:22.12 105 -138 17 4663 05:39:02.47 -69:06:40.10 105 -36 17 2964 05:39:03.49 -69:07:31.02 111 -87 19 3365 05:39:04.13 -69:08:14.74 116 -131 23 5766 05:39:05.07 -69:05:54.83 121 9 7 27967 05:39:07.00 -69:10:49.34 132 -286 7 3968 05:39:07.51 -69:10:26.01 138 -262 14 1269 05:39:08.05 -69:08:47.38 138 -164 18 5370 05:39:11.46 -69:08:15.27 154 -131 9 4971 05:39:11.76 -69:10:53.42 159 -290 14 5572 05:39:12.36 -69:07:44.75 159 -101 13 2673 05:39:13.25 -69:08:56.22 164 -172 9 9174 05:39:14.28 -69:01:45.16 170 259 23 30875 05:39:14.63 -69:10:54.29 175 -290 24 5476 05:39:14.72 -69:07:39.20 175 -95 13 3777 05:39:16.70 -69:05:28.48 186 35 8 35078 05:39:17.98 -69:10:36.02 191 -272 11 3279 05:39:18.95 -69:05:56.61 197 7 11 35580 05:39:18.97 -69:07:45.44 197 -102 8 6381 05:39:19.05 -69:07:45.59 197 -102 8 5582 05:39:21.91 -69:04:31.65 213 92 15 33583 05:39:22.10 -69:05:27.13 213 37 18 34184 05:39:22.91 -69:06:08.13 218 -4 7 35985 05:39:23.79 -69:07:00.26 224 -56 24 1486 05:39:24.53 -69:04:45.88 229 78 9 33987 05:39:26.37 -69:05:59.02 234 5 15 2788 05:39:26.59 -69:04:42.32 240 81 4 34489 05:39:26.62 -69:07:32.13 240 -88 22 2890 05:39:27.52 -69:06:11.28 245 -7 9 34991 05:39:27.79 -69:04:04.10 245 120 28 32092 05:39:30.83 -69:05:55.85 261 8 20 2593 05:39:30.85 -69:10:11.75 261 -248 19 4494 05:39:31.05 -69:05:29.34 261 34 8 34195 05:39:31.14 -69:05:24.43 261 39 6 33096 05:39:31.23 -69:05:20.09 261 44 16 35797 05:39:34.55 -69:05:25.43 283 38 6 32498 05:39:37.13 -69:07:28.57 293 -85 13 1299 05:39:37.64 -69:07:19.20 299 -75 10 13100 05:39:37.86 -69:07:09.53 299 -66 8 8 ABLE Continued
Pillar ID RA (J2000) Dec ∆ RA ∆ Dec Length PA(arcsec) (arcsec) (arcsec)101 05:39:39.19 -69:11:25.06 304 -321 7 240102 05:39:39.53 -69:08:34.63 310 -151 5 29103 05:39:39.73 -69:08:31.87 310 -148 4 58104 05:39:39.77 -69:11:25.05 310 -321 12 258105 05:39:44.91 -69:06:49.21 336 -45 17 2106 05:39:47.81 -69:08:49.79 353 -166 39 19 ABLE ROMINENT IONIZATION FRONTS (IF S ) SUITABLE FOR FOLLOW - UP MULTI - WAVELENGTH STUDIES . IF ID RA (J2000) Dec ∆ RA ∆ Dec Length PA(arcsec) (arcsec) (arcsec)1 05:38:44.15 -69:06:59.71 9.3 -55.9 31.0 257.72 05:38:55.02 -69:05:42.02 67.7 21.8 39.8 317.93 05:38:36.07 -69:06:27.99 -34.1 -24.2 92.2 349.14 05:39:22.75 -69:07:17.56 216.8 -73.8 87.3 34.75 05:39:11.80 -69:08:14.33 157.9 -130.5 49.0 68.96 05:38:59.15 -69:05:14.79 89.9 49.0 38.2 120.07 05:38:51.70 -69:05:06.06 49.9 57.7 38.2 70.08 05:37:55.71 -69:05:43.70 -251.1 20.1 75.2 321.99 05:37:54.88 -69:05:11.42 -255.6 52.4 59.5 306.110 05:37:45.34 -69:05:14.08 -306.8 49.7 28.5 0.011 05:38:06.22 -69:07:43.53 -194.6 -99.7 37.3 344.112 05:38:11.38 -69:08:06.38 -166.9 -122.6 37.3 283.813 05:39:37.31 -69:07:17.48 295.1 -73.7 28.0 26.014 05:39:37.57 -69:08:25.22 296.5 -141.4 96.0 63.015 05:39:23.05 -69:08:33.57 218.4 -149.8 31.0 310.016 05:39:19.76 -69:08:21.01 200.7 -137.2 14.4 29.217 05:38:58.60 -69:09:36.36 87.0 -212.6 63.8 16.918 05:38:32.82 -69:09:17.55 -51.6 -193.8 63.8 340.719 05:38:58.56 -69:08:43.34 86.8 -159.5 2.4 324.720 05:39:03.35 -69:08:04.44 112.5 -120.6 13.5 67.921 05:39:09.79 -69:04:36.61 147.1 87.2 18.3 142.722 05:38:52.60 -69:04:40.27 54.7 83.5 20.3 128.723 05:38:44.74 -69:04:19.25 12.5 104.5 20.3 120.024 05:38:37.67 -69:04:27.59 -25.5 96.2 13.6 45.025 05:38:39.17 -69:03:31.66 -17.5 152.1 25.9 90.026 05:38:28.95 -69:02:54.60 -72.4 189.2 29.8 25.727 05:38:23.40 -69:02:24.23 -102.2 219.6 62.5 111.428 05:38:12.58 -69:02:16.68 -160.4 227.1 50.4 66.229 05:38:10.17 -69:02:44.58 -173.4 199.2 11.4 61.030 05:38:07.60 -69:03:16.01 -187.2 167.8 15.7 39.131 05:38:14.24 -69:03:29.95 -151.5 153.8 38.8 53.032 05:38:38.16 -69:00:34.55 -22.9 329.2 28.8 63.833 05:38:04.92 -69:04:38.91 -201.6 84.9 9.2 26.334 05:38:34.69 -69:01:05.06 -41.6 298.7 20.8 32.51* 05:37:57.50 -69:07:44.65 -241.5 -100.9 107.0 322.22* 05:38:09.50 -69:08:58.47 -177.0 -174.7 83.8 298.53* 05:38:19.48 -69:09:55.73 -123.3 -231.9 54.8 343.74* 05:38:06.23 -69:10:47.75 -194.5 -284.0 54.8 60.85* 05:37:58.58 -69:08:32.56 -235.7 -148.8 18.4 317.2N
OTE .—IDs with an asterisk identify IFs with PAH emission closer to R136 than the [S II] emission which indicate an ionization source otherthan the central cluster..—IDs with an asterisk identify IFs with PAH emission closer to R136 than the [S II] emission which indicate an ionization source otherthan the central cluster.