Illuminating the Tadpole's metamorphosis I: MUSE observations of a small globule in a sea of ionizing photons
Megan Reiter, Anna F. McLeod, Pamela D. Klaassen, Andrés E. Guzmán, J. E. Dale, Joseph C. Mottram, Guido Garay
MMNRAS , 1–18 (2019) Preprint 30 September 2019 Compiled using MNRAS L A TEX style file v3.0
Illuminating the Tadpole’s metamorphosis I: MUSEobservations of a small globule in a sea of ionizing photons
Megan Reiter, (cid:63) Anna F. McLeod, , Pamela D. Klaassen, Andr´es E. Guzm´an, J. E. Dale, Joseph C. Mottram, and Guido Garay UK Astronomy Technology Centre, ROE, Blackford Hill, Edinburgh, EH9 3HJ, UK Department of Astronomy, University of California Berkeley, Berkeley, CA 94720, USA Department of Physics and Astronomy, Texas Tech University, PO Box 41051, Lubbock, TX 79409, USA National Astronomical Observatory of Japan, 2-21-1 Osawa, Mitaka, Tokyo 181-8588, Japan Centre for Astrophysics Research, University of Hertfordshire, College Lane, Hatfield, AL10 9AB, UK Max Planck Institute for Astronomy, K¨onigstuhl 17, 69117 Heidelberg, Germany Departamento de Astronom´ıa, Universidad de Chile, Camino el Observatorio 1515, Las Condes, Santiago, Chile
Accepted XXX. Received YYY; in original form ZZZ
ABSTRACT
We present new MUSE/VLT observations of a small globule in the Carina H ii regionthat hosts the HH 900 jet+outflow system. Data were obtained with the GALACSIground-layer adaptive optics system in wide-field mode, providing spatially-resolvedmaps of diagnostic emission lines. These allow us to measure the variation of thephysical properties in the globule and jet+outflow system. We find high tempera-tures ( T e ≈ K), modest extinction ( A V ≈ . mag), and modest electron densities( n e ≈ cm − ) in the ionized gas. Higher excitation lines trace the ionized out-flow; both the excitation and ionization in the outflow increase with distance from theopaque globule. In contrast, lower excitation lines that are collisionally de-excited atdensities (cid:38) cm − trace the highly collimated protostellar jet. Assuming the globuleis an isothermal sphere confined by the pressure of the ionization front, we computea Bonnor-Ebert mass of ∼ . M (cid:12) . This is two orders of magnitude higher than pre-vious mass estimates, calling into question whether small globules like the Tadpolecontribute to the bottom of the IMF. Derived globule properties are consistent with acloud that has been and/or will be compressed by the ionization front on its surface.At the estimated globule photoevaporation rate of ∼ × − M (cid:12) yr − , the globule willbe completely ablated in ∼ Myr. Stars that form in globules like the Tadpole willemerge into the H ii later and may help resolve some of the temporal tension betweendisk survival and enrichment. Key words:
HII regions, (ISM): jets and outflows, (ISM:) individual: NGC 3372
Most stars are born in large complexes where feedback fromhigh-mass stars will sculpt the parent molecular cloud, andthus the natal cocoon of on-going star formation. Ionizing ra-diation carves elongated dust pillars that protrude into theH ii region, pointing toward the dominant ionizing source(e.g., Hester et al. 1996; Klaassen et al. 2014; McLeod et al.2015, 2016; Pattle et al. 2018). Many pillars have high-density clumps at their tips (e.g., Ohlendorf et al. 2012) thatwere either compressed (e.g., Gritschneder et al. 2010; Daleet al. 2013) or uncovered by the advancing ionization front(e.g., Dale & Bonnell 2011; Tremblin et al. 2013). Initially, (cid:63) E-mail: [email protected] (MR) this high density material may cast a shadow that shieldsthe pillar body from incident ionizing radiation. Eventually,the clumps at the pillar tips may detach, appearing as free-floating globules in the H ii region. If some globules harbor anascent star, these may be the progenitors of the photoevap-orating protoplanetary disks (proplyds) seen in nearby re-gions like Orion (e.g., O’dell & Wen 1994; Mann & Williams2010; Mann et al. 2014). Connecting these stages is criticalto understanding how stellar feedback shapes ongoing starformation.However they are formed, many H ii regions have apopulation of small, opaque globules that remain afloat inthe ionized gas after the surrounding low-density materialis cleared away. The survival and star-forming potential ofthese oases of neutral gas have been a subject of interest for © a r X i v : . [ a s t r o - ph . GA ] S e p M. Reiter et al. decades (Bok 1948; Pottasch 1956, 1958; Dyson 1968; Herbig1974; Schneps et al. 1980; Reipurth 1983). Many theoreticalinvestigations have explored the sensitivity of globule evolu-tion to initial conditions and the shape and intensity of theradiation field (Sandford et al. 1982; Bertoldi 1989; Lefloch& Lazareff 1994; Kessel-Deynet & Burkert 2003; Esquivel &Raga 2007; Miao et al. 2009; Bisbas et al. 2011).Whether this external feedback triggers or starves starformation remains a matter of debate. Plane-parallel radia-tion focuses material whereas the divergent flux of a point-source results in multiple fragments (Esquivel & Raga 2007).The shock driven by the ionization front may dramaticallyenhance the effects of self-gravity (e.g., Miao et al. 2009;Gritschneder et al. 2009; Bisbas et al. 2011). Tremblin et al.(2013) argue that the observed star formation would havehappened anyway, but may benefit from the accumulationof matter being pushed by feedback. In contrast, Kessel-Deynet & Burkert (2003) argue that the redistribution ofkinetic energy into undirected random motions may prolongthe cloud lifetime by making it more robust against collapse.Globules that do collapse may make a unique contribu-tion to the stellar and planetary populations in feedback-dominated regions. Initially, the globule will shield the em-bedded protostar from ionizing radiation that rapidly de-stroys protoplanetary disks (e.g., Winter et al. 2018; Nichol-son et al. 2019). Stars forming in these globules are relativelyclose to the high-mass stars that provide chemical enrich-ment thought to play a central role in the geochemical evo-lution of terrestrial planets (e.g., Cameron & Truran 1977;Grimm & McSween 1993; Hester et al. 2004; Lichtenberget al. 2019). As such, star-forming globules represent an in-termediate case that may help reconcile short disk lifetimeswith the lifetimes of high-mass stars.Smaller globules that collapse may form planetary-massobjects, potentially contributing significantly to the low-mass end of the initial mass function (IMF; see Grenman& Gahm 2014). If this is the case, then feedback-dominatedregions may show evidence for variations in the IMF. Whileobservations of local Galactic star-forming regions suggestthat the IMF is invariant (e.g., Bastian et al. 2010), the sit-uation is less clear in the more extreme environments sam-pled in other galaxies (e.g., Conroy & van Dokkum 2012;Conroy et al. 2013; Zhang et al. 2018). Local regions pro-vide the best laboratory to resolve how feedback affects starformation and the low-mass stellar (and sub-stellar) popu-lation, unlike extra-galactic environments where direct mea-surements of the IMF are impossible.Grenman & Gahm (2014) catalogued a population ofsmall globules in one of the nearest giant H ii regions, theCarina Nebula. These so-called globulettes are larger thanrevealed proplyds but smaller than cannonical Bok globules(Bok & Reilly 1947; Thackeray 1950). Based on a singlewavelength intensity-tracing analysis, they estimated thatmost of the globulettes are roughly planetary mass. If evena small fraction of these globulettes undergo gravitationalcollapse, they may contribute significantly to the low-massend of the IMF. However, Haworth et al. (2015) showed thatthese low-mass globulettes are stable and require an addi-tional perturbation to stimulate collapse. Estimated photoe-vaporation rates (e.g., Smith et al. 2004; Reiter et al. 2015a)suggest that such small, low-mass objects will be rapidly ab-lated. McCaughrean & Andersen (2002) report evidence for star formation in ∼ % of the (larger) globules seen in M16,demonstrating that at least some globules have larger massreservoirs, and thus may survive longer.Determining the fate of small globules and constrainingthe relevant physics requires an estimate of the incident en-ergy and the resulting kinematics in the cold molecular gas.In this paper, we consider a small ( r ∼ (cid:48)(cid:48) , correspondingto ∼ . pc at a distance of 2.3 kpc Smith 2006b) tadpole-shaped globule in the Carina Nebula. The so-called Tadpole(see Figure 1) lies between the young open cluster Tr16 anda large wall of gas and dust that bisects the H ii region. Inoptical images, the globule is seen in silhouette against thebright nebular emission from the H ii region.H α images from the Hubble Space Telescope (HST) re-vealed the unusually wide-angle outflow HH 900 (Smith et al.2010) emerging from the globule and illuminated by themany O-type stars in Tr16. Both the morphology and thekinematics of the H α suggest it traces an outflow entrainedby the fast, collimated jet seen in near-IR [Fe ii ] (Reiter et al.2015a). Protostellar jets seen in other H ii regions often ap-pear to point away from the ionizing sources, bent eitherby their winds, the rocket effect, or both (e.g., Bally et al.2006). However, HH 900 bends toward Tr16, leading Reiteret al. (2015a) to suggest that it is pushed by the photoevapo-rative flow off the dust wall behind it. With a jet dynamicalage of ∼ yr, HH 900 is one of the youngest outflowsin Carina. Jet kinematics require a driving source embed-ded in the small opaque globule, but previous observationsprovided no evidence for a protostar inside the Tadpole.In this paper (Paper I), we present new data from theVery Large Telescope (VLT) using the Multi Unit Spectro-scopic Explorer (MUSE; Bacon et al. 2010) with GALACSIwide-field adaptive optics correction of the Tadpole globuleand HH 900 jet + outflow system. These data provide a suiteof spectral lines to measure the physical properties of the hotgas in this small globule adrift in an H ii region. Togetherwith resolved observations of the cold, molecular gas in theglobule (Paper II; Reiter et al. in prep), this allows us to con-strain the role of external photoionization and internal dis-ruption of the globule by the jet, laying the groundwork forbuilding an evolutionary scenario of feedback carved struc-tures from pillars to globules to photoevaporating protoplan-etary disks. We obtained data with the MUSE visual wavelengthpanoramic integral-field spectrograph on the VLT in ser-vice mode on 03 April 2018. Using the newly-commissionedMUSE+GALACSI Adaptive Optics module in Wide FieldMode (WFM), this provides sub-arcsec resolution over theentire (cid:48) × (cid:48) field-of-view. The AO-assisted angular resolu-tion of the MUSE observations is ∼ . (cid:48)(cid:48) , corresponding to ∼ . pc at the distance of Carina. MUSE provides a spec-tral resolution of R=2000-4000 over a nominal wavelengthrange from − ˚A with a gap between ∼ − ˚Adue to the laser guide stars for the adaptive optics system.This corresponds to a velocity resolution of ∼ − km s − .The Tadpole and full extent of the HH 900 outflow fitwithin a single MUSE pointing that we observed with athree-point dither pattern, rotating ◦ between each ex- MNRAS000
Most stars are born in large complexes where feedback fromhigh-mass stars will sculpt the parent molecular cloud, andthus the natal cocoon of on-going star formation. Ionizing ra-diation carves elongated dust pillars that protrude into theH ii region, pointing toward the dominant ionizing source(e.g., Hester et al. 1996; Klaassen et al. 2014; McLeod et al.2015, 2016; Pattle et al. 2018). Many pillars have high-density clumps at their tips (e.g., Ohlendorf et al. 2012) thatwere either compressed (e.g., Gritschneder et al. 2010; Daleet al. 2013) or uncovered by the advancing ionization front(e.g., Dale & Bonnell 2011; Tremblin et al. 2013). Initially, (cid:63) E-mail: [email protected] (MR) this high density material may cast a shadow that shieldsthe pillar body from incident ionizing radiation. Eventually,the clumps at the pillar tips may detach, appearing as free-floating globules in the H ii region. If some globules harbor anascent star, these may be the progenitors of the photoevap-orating protoplanetary disks (proplyds) seen in nearby re-gions like Orion (e.g., O’dell & Wen 1994; Mann & Williams2010; Mann et al. 2014). Connecting these stages is criticalto understanding how stellar feedback shapes ongoing starformation.However they are formed, many H ii regions have apopulation of small, opaque globules that remain afloat inthe ionized gas after the surrounding low-density materialis cleared away. The survival and star-forming potential ofthese oases of neutral gas have been a subject of interest for © a r X i v : . [ a s t r o - ph . GA ] S e p M. Reiter et al. decades (Bok 1948; Pottasch 1956, 1958; Dyson 1968; Herbig1974; Schneps et al. 1980; Reipurth 1983). Many theoreticalinvestigations have explored the sensitivity of globule evolu-tion to initial conditions and the shape and intensity of theradiation field (Sandford et al. 1982; Bertoldi 1989; Lefloch& Lazareff 1994; Kessel-Deynet & Burkert 2003; Esquivel &Raga 2007; Miao et al. 2009; Bisbas et al. 2011).Whether this external feedback triggers or starves starformation remains a matter of debate. Plane-parallel radia-tion focuses material whereas the divergent flux of a point-source results in multiple fragments (Esquivel & Raga 2007).The shock driven by the ionization front may dramaticallyenhance the effects of self-gravity (e.g., Miao et al. 2009;Gritschneder et al. 2009; Bisbas et al. 2011). Tremblin et al.(2013) argue that the observed star formation would havehappened anyway, but may benefit from the accumulationof matter being pushed by feedback. In contrast, Kessel-Deynet & Burkert (2003) argue that the redistribution ofkinetic energy into undirected random motions may prolongthe cloud lifetime by making it more robust against collapse.Globules that do collapse may make a unique contribu-tion to the stellar and planetary populations in feedback-dominated regions. Initially, the globule will shield the em-bedded protostar from ionizing radiation that rapidly de-stroys protoplanetary disks (e.g., Winter et al. 2018; Nichol-son et al. 2019). Stars forming in these globules are relativelyclose to the high-mass stars that provide chemical enrich-ment thought to play a central role in the geochemical evo-lution of terrestrial planets (e.g., Cameron & Truran 1977;Grimm & McSween 1993; Hester et al. 2004; Lichtenberget al. 2019). As such, star-forming globules represent an in-termediate case that may help reconcile short disk lifetimeswith the lifetimes of high-mass stars.Smaller globules that collapse may form planetary-massobjects, potentially contributing significantly to the low-mass end of the initial mass function (IMF; see Grenman& Gahm 2014). If this is the case, then feedback-dominatedregions may show evidence for variations in the IMF. Whileobservations of local Galactic star-forming regions suggestthat the IMF is invariant (e.g., Bastian et al. 2010), the sit-uation is less clear in the more extreme environments sam-pled in other galaxies (e.g., Conroy & van Dokkum 2012;Conroy et al. 2013; Zhang et al. 2018). Local regions pro-vide the best laboratory to resolve how feedback affects starformation and the low-mass stellar (and sub-stellar) popu-lation, unlike extra-galactic environments where direct mea-surements of the IMF are impossible.Grenman & Gahm (2014) catalogued a population ofsmall globules in one of the nearest giant H ii regions, theCarina Nebula. These so-called globulettes are larger thanrevealed proplyds but smaller than cannonical Bok globules(Bok & Reilly 1947; Thackeray 1950). Based on a singlewavelength intensity-tracing analysis, they estimated thatmost of the globulettes are roughly planetary mass. If evena small fraction of these globulettes undergo gravitationalcollapse, they may contribute significantly to the low-massend of the IMF. However, Haworth et al. (2015) showed thatthese low-mass globulettes are stable and require an addi-tional perturbation to stimulate collapse. Estimated photoe-vaporation rates (e.g., Smith et al. 2004; Reiter et al. 2015a)suggest that such small, low-mass objects will be rapidly ab-lated. McCaughrean & Andersen (2002) report evidence for star formation in ∼ % of the (larger) globules seen in M16,demonstrating that at least some globules have larger massreservoirs, and thus may survive longer.Determining the fate of small globules and constrainingthe relevant physics requires an estimate of the incident en-ergy and the resulting kinematics in the cold molecular gas.In this paper, we consider a small ( r ∼ (cid:48)(cid:48) , correspondingto ∼ . pc at a distance of 2.3 kpc Smith 2006b) tadpole-shaped globule in the Carina Nebula. The so-called Tadpole(see Figure 1) lies between the young open cluster Tr16 anda large wall of gas and dust that bisects the H ii region. Inoptical images, the globule is seen in silhouette against thebright nebular emission from the H ii region.H α images from the Hubble Space Telescope (HST) re-vealed the unusually wide-angle outflow HH 900 (Smith et al.2010) emerging from the globule and illuminated by themany O-type stars in Tr16. Both the morphology and thekinematics of the H α suggest it traces an outflow entrainedby the fast, collimated jet seen in near-IR [Fe ii ] (Reiter et al.2015a). Protostellar jets seen in other H ii regions often ap-pear to point away from the ionizing sources, bent eitherby their winds, the rocket effect, or both (e.g., Bally et al.2006). However, HH 900 bends toward Tr16, leading Reiteret al. (2015a) to suggest that it is pushed by the photoevapo-rative flow off the dust wall behind it. With a jet dynamicalage of ∼ yr, HH 900 is one of the youngest outflowsin Carina. Jet kinematics require a driving source embed-ded in the small opaque globule, but previous observationsprovided no evidence for a protostar inside the Tadpole.In this paper (Paper I), we present new data from theVery Large Telescope (VLT) using the Multi Unit Spectro-scopic Explorer (MUSE; Bacon et al. 2010) with GALACSIwide-field adaptive optics correction of the Tadpole globuleand HH 900 jet + outflow system. These data provide a suiteof spectral lines to measure the physical properties of the hotgas in this small globule adrift in an H ii region. Togetherwith resolved observations of the cold, molecular gas in theglobule (Paper II; Reiter et al. in prep), this allows us to con-strain the role of external photoionization and internal dis-ruption of the globule by the jet, laying the groundwork forbuilding an evolutionary scenario of feedback carved struc-tures from pillars to globules to photoevaporating protoplan-etary disks. We obtained data with the MUSE visual wavelengthpanoramic integral-field spectrograph on the VLT in ser-vice mode on 03 April 2018. Using the newly-commissionedMUSE+GALACSI Adaptive Optics module in Wide FieldMode (WFM), this provides sub-arcsec resolution over theentire (cid:48) × (cid:48) field-of-view. The AO-assisted angular resolu-tion of the MUSE observations is ∼ . (cid:48)(cid:48) , corresponding to ∼ . pc at the distance of Carina. MUSE provides a spec-tral resolution of R=2000-4000 over a nominal wavelengthrange from − ˚A with a gap between ∼ − ˚Adue to the laser guide stars for the adaptive optics system.This corresponds to a velocity resolution of ∼ − km s − .The Tadpole and full extent of the HH 900 outflow fitwithin a single MUSE pointing that we observed with athree-point dither pattern, rotating ◦ between each ex- MNRAS000 , 1–18 (2019) adpole metamorphasis Figure 1.
Left:
MUSE image of the Tadpole globule and the HH 900 jet+outflow system ([S ii ]=red, H α =green, [O iii ]=blue). Asecond YSO (PCYC 838) and possible microjet that lie projected onto the western limn of HH 900 are labeled. Right:
Three-color imagehighlighting different components of the jet+outflow system: the molecular outflow ([C i ]=red), the atomic jet ([Fe ii ]=green), and theionized outflow ([O i ]=blue). posure to reduce instrument artifacts. The integration timefor each exposure was 240 s, for a total on-source integra-tion time of s. Data were reduced in the ESO esorex environment of the MUSE pipeline (Weilbacher et al. 2012)using standard calibrations.Due to lunar contamination, the MUSE spectra showedincreased fluxes towards the blue regime of the wavelengthcoverage with respect to the red. This would result infalsely high fluxes of integrated maps of emission linesin that portion of the spectrum (e.g., H β , He I λ iii ] λλ We detect ∼ emission lines with MUSE, with more thana dozen hydrogen recombination lines and forbidden emis-sion lines from at least seven different atoms. A full list oflines detected with MUSE is presented in Table A1. Detectedemission lines represent multiple ionization states in the gas,although the majority have ionization potentials (cid:46) . eV(see Table 1 for a list of first and second ionization poten-tials). The tadpole-shaped globule itself is generally seen insilhouette against the bright nebular background. Emissionlines like H α , [N ii ], and [S ii ] trace the ionization fronton the globule surface. With these spatially resolved MUSEobservations of diagnostic emission lines, we measure thevariation in the physical conditions in the ionization frontboth in the globule and the externally irradiated HH 900jet+outflow system.Many of the emission lines that are bright in the globuleionization front also trace the wide-angle outflow originallydiscovered in narrowband H α images from HST . Forbiddenemission lines that are collisionally deexcited in high-density
Table 1.
Ionization potentials for detected elementsName first second(eV) (eV)H 13.6 —He 24.6 54.4C 11.3 24.4N 14.5 29.6O 13.6 35.1S 10.4 23.3Cl 13.0 23.8Ar 15.8 27.6Ca 6.1 11.9Fe 7.9 16.2Ni 7.6 18.2 gas like [Fe ii ], [Ca ii ], and [Ni ii ] trace the collimated bipo-lar jet that threads through the center of the wide-angleoutflow (see Figure 1 and Reiter et al. 2015a). The spectralresolution of MUSE allows us to resolve velocity differences (cid:38) km s − , enabling a kinematic analysis of the faster por-tions of the jet+outflow system. In the following sections, wederive the physical properties of the tadpole-shaped globuleand the HH 900 jet+outflow system. We use the Balmer decrement to estimate the line-of-sightextinction as E ( B − V ) = E ( H β − H α ) κ ( H β ) − κ ( H α ) = − . κ ( H β ) − κ ( H α ) × log (cid:20) ( H α / H β ) int ( H α / H β ) obs (cid:21) (1)where κ ( λ ) is the value of the extinction curve at a givenwavelength (see the Appendix of Momcheva et al. 2013).Assuming typical Milky Way extinction, κ ( H β ) − κ ( H α ) = MNRAS , 1–18 (2019)
M. Reiter et al. . (Fitzpatrick 1999), this becomes E ( B − V ) = − . × log (cid:20) . ( H α / H β ) obs (cid:21) (2)where we have assumed Case B recombination for gas with T = K and n e = cm − for which we expect an in-trinsic flux ratio ( H α / H β ) int = . (Osterbrock & Ferland2006). The observed flux ratio in and around the Tadpoleglobule and outflow are ( H α / H β ) obs ≈ . − . , correspond-ing to E ( B − V ) ≈ . − . . To convert the color excess to anextinction, we adopt R V = . ± . (Hur et al. 2012), whichgives . (cid:46) A V (cid:46) . mag, with a median A V ≈ . mag. Theestimated extinction is slightly higher if we use R V = . , asother authors have found for Carina (e.g., Smith 1987, 2002).The extinction we estimate using the Balmer decrementis ∼ mag lower than that estimated by Reiter et al. (2015a)using the ratio of near-IR [Fe ii ] lines ( A V ≈ . mag). Hy-drogen recombination lines like H α trace a notably differentmorphological and kinematic component than the forbiddennear-IR [Fe ii ] lines in the HH 900 jet+outflow system. Dif-ferences in the estimated extinction may therefore reflectreal differences in the extinction to different components ofthe system. More material may obscure the jet embeddedwithin the outflow, especially if the outflow is not purely ion-ized gas (see Section 4.2), leading to a higher extinction tothe collimated jet. Raga et al. (2015) also find that E ( B − V ) estimates from the Balmer decrement are lower than thoseobtained with other lines that do not require assuming arecombination cascade. In the following analysis, we adoptthe extinction estimated from the Balmer decrement anduse Pyneb (Luridiana et al. 2015) to correct the observedfluxes.
With AO-assisted IFU data, we can make spatially-resolvedestimates of the density and temperature to determine howthey vary as a function of position in the system. Many tra-ditional nebular emission lines are detected from the wide-angle ionized outflow (both [S ii ] and [N ii ] have the samemorphology as H α , see Figure B1).Density-sensitive line ratios compare the flux of emis-sion lines with significantly different radiative transitionprobabilities. We estimate the electron density, n e , from theratio [S ii ] λ / λ which is sensitive to densities in therange − cm − . Emission lines that originate from sig-nificantly different excitation energy levels from the sameatom provide an estimate of the temperature in the gas. Toestimate the electron temperature, T e , we use the ratio [N ii ] λ + λ / λ . We use pyneb to solve for both quan-tities simultaneously at each pixel in the MUSE map. Varia-tions in each quantity as a function of position are shown inFigure 2. Electron densities vary from (cid:46) n e (cid:46) cm − with the highest densities found in the edge of the globuleand inner outflow. The derived T e shows a smaller range ofvariation along the outflow / globule ( T e ≈ K, see Fig-ure 2), with the lowest temperatures near the globule.Both [S ii ] and [N ii ] trace the wide-angle outflow, fol-lowing the morphology of H α . Thus the temperature anddensity derived from these ionized gas tracers may not re-flect the physical conditions in the core of the jet seen in near-IR [Fe ii ] emission. Reiter et al. (2015a) argue thathigh densities in the jet core ( (cid:38) cm − ) are required toshield the Fe + from further ionization. Nisini et al. (2002,2005) also derived smaller electron densities from the [S ii ]lines compared to those determined from the near-IR [Fe ii ]lines. Bautista et al. (1994) argue that the optical [Fe ii ]lines reveal even higher densities, originating from regionswith n e ≈ − cm − . We detect multiple [Fe ii ] emissionlines in the MUSE data (see Table A1); all trace the samecollimated morphology seen in the near-IR. Using diagnosticdiagrams from Bautista & Pradhan (1998), we find that theratio [Fe ii ] λ / λ suggests n e > cm − , providedthat photoexcitation also contributes to the emission. Lineswith lower critical densities, like the classic nebular tracersused to estimate the temperature and density of the ionizedoutflow, will be de-excited in regions with such high densi-ties and are therefore a poor tracer of the physical conditionsin the [Fe ii ]-emitting jet. Thus, we allow that temperaturesmay also be lower in the high density, lower ionization jetcomponent (see Section 3.3). The intensity and spatial morphology of the emission lineswe detect with MUSE depend on the dominant excitationmechanism(s). Studies of ionized interfaces often find thatphotoexcitation dominates over shocks (e.g. Yeh et al. 2015;McLeod et al. 2015). Reiter et al. (2015a) argued that this isalso true in the HH 900 outflow based on the H α /[Fe ii ] ra-tio, as well as the morphology and kinematics of both lines.The spectral coverage of MUSE includes many lines thatmay be used in diagnostic ratios to map the excitation overthe Tadpole globule and HH 900 jet+outflow system to testwhether – and where – shock excitation dominates over pho-toexcitation.Tracings of the intensity profile of several key emis-sion lines in slices parallel and perpendicular to the outflowaxis are shown in Figure 3. For shock excitation, we expect[N ii ]/H α > . and [S ii ]/H α > . (Allen et al. 2008).However, the line ratios we measure in the Tadpole glob-ule and the HH 900 jet+outflow system are typically muchless than these values (median ratios are [N ii ]/H α = . and [S ii ]/H α = . ). No point in the contiguous systemmeets or exceeds the shock-excitation threshold. Values ofthe [N ii ]/H α ratio are similar in the bow shocks ( ∼ . and ∼ . for the eastern and western bow shocks, respectively),and slightly higher in the [S ii ]/H α ratio ( ∼ . and ∼ . ,respectively). Overall the line ratios indicate that photoex-citation dominates over shocks.Multiple line ratios in this wavelength range have beenused to diagnose the degree of ionization in the gas. Wetake two approaches to estimate the degree of ionization:(1) the ratio of lines of different ionization from the sameelement, in this case [S iii ] 9069˚A / [S ii ] 6717˚A; and (2)the ratio of lines from elements with similar ionization po-tentials (see Table 1), in this case I(He i λ ) / I([S ii ]( λ + λ )/2)). As the ionization increases, so too doesthe ratio of these lines since the line with the slightly lowerionization potential (the denominator) decreases as more ofit gets ionized. Glushkov (1998) used the latter ratio to ar-gue that M17 has the highest degree of ionization of galacticstar forming regions ( > in M17 compared with ≈ . in MNRAS000
With AO-assisted IFU data, we can make spatially-resolvedestimates of the density and temperature to determine howthey vary as a function of position in the system. Many tra-ditional nebular emission lines are detected from the wide-angle ionized outflow (both [S ii ] and [N ii ] have the samemorphology as H α , see Figure B1).Density-sensitive line ratios compare the flux of emis-sion lines with significantly different radiative transitionprobabilities. We estimate the electron density, n e , from theratio [S ii ] λ / λ which is sensitive to densities in therange − cm − . Emission lines that originate from sig-nificantly different excitation energy levels from the sameatom provide an estimate of the temperature in the gas. Toestimate the electron temperature, T e , we use the ratio [N ii ] λ + λ / λ . We use pyneb to solve for both quan-tities simultaneously at each pixel in the MUSE map. Varia-tions in each quantity as a function of position are shown inFigure 2. Electron densities vary from (cid:46) n e (cid:46) cm − with the highest densities found in the edge of the globuleand inner outflow. The derived T e shows a smaller range ofvariation along the outflow / globule ( T e ≈ K, see Fig-ure 2), with the lowest temperatures near the globule.Both [S ii ] and [N ii ] trace the wide-angle outflow, fol-lowing the morphology of H α . Thus the temperature anddensity derived from these ionized gas tracers may not re-flect the physical conditions in the core of the jet seen in near-IR [Fe ii ] emission. Reiter et al. (2015a) argue thathigh densities in the jet core ( (cid:38) cm − ) are required toshield the Fe + from further ionization. Nisini et al. (2002,2005) also derived smaller electron densities from the [S ii ]lines compared to those determined from the near-IR [Fe ii ]lines. Bautista et al. (1994) argue that the optical [Fe ii ]lines reveal even higher densities, originating from regionswith n e ≈ − cm − . We detect multiple [Fe ii ] emissionlines in the MUSE data (see Table A1); all trace the samecollimated morphology seen in the near-IR. Using diagnosticdiagrams from Bautista & Pradhan (1998), we find that theratio [Fe ii ] λ / λ suggests n e > cm − , providedthat photoexcitation also contributes to the emission. Lineswith lower critical densities, like the classic nebular tracersused to estimate the temperature and density of the ionizedoutflow, will be de-excited in regions with such high densi-ties and are therefore a poor tracer of the physical conditionsin the [Fe ii ]-emitting jet. Thus, we allow that temperaturesmay also be lower in the high density, lower ionization jetcomponent (see Section 3.3). The intensity and spatial morphology of the emission lineswe detect with MUSE depend on the dominant excitationmechanism(s). Studies of ionized interfaces often find thatphotoexcitation dominates over shocks (e.g. Yeh et al. 2015;McLeod et al. 2015). Reiter et al. (2015a) argued that this isalso true in the HH 900 outflow based on the H α /[Fe ii ] ra-tio, as well as the morphology and kinematics of both lines.The spectral coverage of MUSE includes many lines thatmay be used in diagnostic ratios to map the excitation overthe Tadpole globule and HH 900 jet+outflow system to testwhether – and where – shock excitation dominates over pho-toexcitation.Tracings of the intensity profile of several key emis-sion lines in slices parallel and perpendicular to the outflowaxis are shown in Figure 3. For shock excitation, we expect[N ii ]/H α > . and [S ii ]/H α > . (Allen et al. 2008).However, the line ratios we measure in the Tadpole glob-ule and the HH 900 jet+outflow system are typically muchless than these values (median ratios are [N ii ]/H α = . and [S ii ]/H α = . ). No point in the contiguous systemmeets or exceeds the shock-excitation threshold. Values ofthe [N ii ]/H α ratio are similar in the bow shocks ( ∼ . and ∼ . for the eastern and western bow shocks, respectively),and slightly higher in the [S ii ]/H α ratio ( ∼ . and ∼ . ,respectively). Overall the line ratios indicate that photoex-citation dominates over shocks.Multiple line ratios in this wavelength range have beenused to diagnose the degree of ionization in the gas. Wetake two approaches to estimate the degree of ionization:(1) the ratio of lines of different ionization from the sameelement, in this case [S iii ] 9069˚A / [S ii ] 6717˚A; and (2)the ratio of lines from elements with similar ionization po-tentials (see Table 1), in this case I(He i λ ) / I([S ii ]( λ + λ )/2)). As the ionization increases, so too doesthe ratio of these lines since the line with the slightly lowerionization potential (the denominator) decreases as more ofit gets ionized. Glushkov (1998) used the latter ratio to ar-gue that M17 has the highest degree of ionization of galacticstar forming regions ( > in M17 compared with ≈ . in MNRAS000 , 1–18 (2019) adpole metamorphasis Figure 2.
A [S ii ] image showing the slices through the globule parallel and perpendicular to the jet+outflow. We plot the variationalong the jet+outflow (red line; plots shown on the left ) and perpendicular to it (blue dotted line; plots shown on the right ) of thefollowing physical parameters (see Sections 3.1 and 3.2): (a) A V ; (b) n e ; (c) T e ; (d) the degree of ionization as traced by the He i ii ] λ + λ / ) ratio (black line) and the [S iii ] λ / [S ii ] λ ratio (gray dotted line). Vertical dashed gray linesshow the approximate position of the globule edges.MNRAS , 1–18 (2019) M. Reiter et al.
Figure 3.
Emission line intensity tracings parallel (left) and perpendicular (right) to the HH 900 jet+outflow system. Cuts are made atthe same locations shown in Figure 2. We have added a small offset to the normalized intensity; horizontal gray dotted lines indicate themean flux level. Vertical dashed gray lines shown the approximate edges of the globule.
NGC 1976, near the Trapezium). Typical ratios are lower inthe nebular gas surrounding the Tadpole ( ∼ . ) and lowerstill in the globule and outflow ( ∼ . ). There is a cleartrend of increasing ionization with distance from the centerof the globule. In both tracers, the ratio at the terminus ofthe contiguous outflow is ∼ . × greater than the ratio mea-sured in the least ionized gas in and immediately around theglobule (see Figure 2).Most of the lines that we detect in the outflow haveionization potentials < . eV; emission lines from morehighly ionized elements requiring more energetic photons aresignificantly fainter. Notable exceptions to this trend are: (1)Neutral carbon, [C i ] λ , which is significantly brighterthan the possible C ii λ line, despite the fact that theionization potential of C is 11.3 eV. Both lines are seen onlyin the immediate vicinity of the globule. (2) Highly ionizedlines like [S iii ] and [Ar iii ], are seen in the entire outflow and both bow shocks. Both require photon energies > eVto achieve the second ionization, although we do not detectlines like He + that also have ionization potentials in thisrange. (3) The detection of [N ii ] but not [N i ], even thoughthe ionization potential of N is 14.534 eV.We can simultaneously compare ionization and shockexcitation in the gas using the diagnostic diagram shownin Figure 4. More highly ionized gas will have smaller[O ii ]/[O iii ] ratios; shock excitation will enhance the ra-tio S = ([S ii ] λ + [S ii ] λ + [S iii ] λ ) / H β . Weplot the line ratio measured in individual pixels in Figure 4using separate colors to distinguish between different com-ponents of the system. Points in the Tadpole system occupya smaller portion of parameter space than sampled in theOrion Nebula Cluster (the ONC, see Mc Leod et al. 2016).The most highly ionized gas in this study overlaps with thelower excitation proplyds in Mc Leod et al. (2016). At the MNRAS000
NGC 1976, near the Trapezium). Typical ratios are lower inthe nebular gas surrounding the Tadpole ( ∼ . ) and lowerstill in the globule and outflow ( ∼ . ). There is a cleartrend of increasing ionization with distance from the centerof the globule. In both tracers, the ratio at the terminus ofthe contiguous outflow is ∼ . × greater than the ratio mea-sured in the least ionized gas in and immediately around theglobule (see Figure 2).Most of the lines that we detect in the outflow haveionization potentials < . eV; emission lines from morehighly ionized elements requiring more energetic photons aresignificantly fainter. Notable exceptions to this trend are: (1)Neutral carbon, [C i ] λ , which is significantly brighterthan the possible C ii λ line, despite the fact that theionization potential of C is 11.3 eV. Both lines are seen onlyin the immediate vicinity of the globule. (2) Highly ionizedlines like [S iii ] and [Ar iii ], are seen in the entire outflow and both bow shocks. Both require photon energies > eVto achieve the second ionization, although we do not detectlines like He + that also have ionization potentials in thisrange. (3) The detection of [N ii ] but not [N i ], even thoughthe ionization potential of N is 14.534 eV.We can simultaneously compare ionization and shockexcitation in the gas using the diagnostic diagram shownin Figure 4. More highly ionized gas will have smaller[O ii ]/[O iii ] ratios; shock excitation will enhance the ra-tio S = ([S ii ] λ + [S ii ] λ + [S iii ] λ ) / H β . Weplot the line ratio measured in individual pixels in Figure 4using separate colors to distinguish between different com-ponents of the system. Points in the Tadpole system occupya smaller portion of parameter space than sampled in theOrion Nebula Cluster (the ONC, see Mc Leod et al. 2016).The most highly ionized gas in this study overlaps with thelower excitation proplyds in Mc Leod et al. (2016). At the MNRAS000 , 1–18 (2019) adpole metamorphasis Figure 4.
Same excitation / ionization parameters as exploredfor the nebular emission in the ONC by Mc Leod et al. (2016).All points are shown in purple, with a few select portions of theoutflow (east=green, west=blue) and the globule face (gold) high-lighted. Ambient gas (purple points in lower left portion of theplot) tends to have higher ionization than the globule and out-flow. The S parameter shows a large spread in all components,although there is a hint that the value of S is slightly larger inthe western limb of the jet. same time, the S parameter is higher overall, with mostTadpole points lying above the values Mc Leod et al. (2016)measure in the Bright Bar, overlapping with the lower end ofthe shock-excited emission measured in the HH 201 proto-stellar jet. Overall, shock-excitation does not dominate andthe globule and outflow are less ionized than the ambientgas surrounding the system. Multiple emission lines in the MUSE data trace the irradi-ated HH 900 jet+outflow system. H ii region emission lineslike H α , [N ii ], [S ii ], and [O ii ] trace the wide-angle, ex-ternally irradiated outflow. Line ratios in the irradiated out-flow suggest that photoexcitation dominates over shocks (seeSection 3.3). In the following sections, we examine the jetand outflow components separately, and finally examine thekinematics of the jet+outflow system. Forbidden emission lines such as [Ca ii ], [Ni ii ], and [Fe ii ]trace the fast, collimated jet seen in near-IR [Fe ii ] emis-sion (Reiter et al. 2015a). The detection of the optical [Fe ii ] lines suggests n e (cid:38) cm − (Bautista et al. 1994), indicat-ing that densities in the jet may be an order of magnitudehigher than estimated by Reiter et al. (2015a). [Ni ii ] tracesthe same jet components that are bright in [Fe ii ], as is oftenobserved in protostellar jets given their similar ionizationpotentials (7.9 and 7.6 eV, respectively, see e.g., Bautistaet al. 1996; Nisini et al. 2005). For n e ∼ cm − , the [Ni ii ] λ / λ flux ratio in the jet is not consistent with purecollisional excitation (Bautista et al. 1996), suggesting sign-ficant photoexcitation from the strong UV environment ofCarina (e.g., Reiter & Smith 2013). The [Ca ii ] λ / λ flux ratio is also not consistent with pure collisional excita-tion, with values < . throughout the jet (Hartigan et al.2004). Together, this suggests that the density in the fast,collimated jet-like component of HH 900 may be an orderof magnitude higher than that estimated by Reiter et al.(2015a), leading to a corresponding increase in the mass-loss rate and momentum estimates. Nebular emission lines like H α and [S ii ] both trace the wide-angle irradiated outflow, as also seen in HH 666 (Reiter et al.2015b). The MUSE data also provide multiple forbiddenoxygen emission lines of different degrees of ionization. All ofthe detected oxygen lines trace the wide-angle outflow (seeFigure B1). [O i ] 6300 ˚A has the same ionization potential ashydrogen (see Table 1) and follows the same tapered shapedelineated by H α emission from the ionized inner outflow.[O i ] 6300 ˚A is not detected in the terminal bow shocks.[O ii ] 7319,7330 ˚A emission is bright throughout the sys-tem, including both bow shocks. In contrast, [O iii ] 5007 ˚Ais not seen in emission. Instead, the silhouette of the globulestands out against the bright H ii region and we detect weakabsorption at the position of the outflow (see Figure 5). Wesee similar weak absorption of the high-excitation [S iii ] line.Reiter et al. (2017a) saw a similar trend in [O iii ] inthe knots of the dusty jet HH 1019. Absorption in the highexcitation lines in HH 900 suggest that there may be somedust entrained in the outflow (also suggested by morpholog-ical features seen in the HST images; Smith et al. 2010).Finally, we note that [C i ] 8727 ˚A emission increases at thesame location where the high excitation lines are seen inaborption (see Figures 3 and 5; we discuss [C i ] further inthe next paragraph). Strong [O i ] emission and evidence fordust in the outflow points to the presence of low-ionizationand neutral material in the outflow itself (in addition to thejet). We detect [C i ] 8727 ˚A emission from the innermost re-gions of the outflow (see Figures 1 and 6). The ionizationpotential of carbon is 11.26 eV, so the [C i ] recombinationline traces partially ionized gas and is often observed to co-exist with H . Indeed, the [C i ] detected with MUSE appearsto have the same morphology as the near-IR H emission re-ported by Reiter et al. (2015a). The critical density of [C i ]is ∼ . − . × cm − for temperatures T = − K.We estimate densities ∼ × cm − from the [S ii ] linesratio (see Section 3.2) from gas with similar morphology asthe [C i ] emission, suggesting that densities in the outfloware too low for significant collisional excitation. Escalanteet al. (1991) argue that collisional excitation of the [C i ] lineis negligible where the fractional ionization is low, which MNRAS , 1–18 (2019)
M. Reiter et al.
Figure 5.
Left:
Intensity tracing through the eastern limb of the irradiated outflow.
Middle: [O i ] image showing the location of theslices through the eastern and western limbs of the jet. Right:
Intensity tracing through the western limb of the irradiated outflow.
Figure 6.
Emission line tracing through the HH 900 jet+outflowsystem showing the ionized outflow (H α =blue), partially ionizedgas ([C i ]=orange), and predominantly neutral gas ([Fe ii ]=green)using the same jet tracing position as shown in Figure 2. appears to be the case in the HH 900 outflow close to theglobule, where [C i ] emission is bright (see Figure 2). We measure the jet velocity using [Fe ii ] 7155 ˚A and thekinematics of the irradiated outflow using [O i ] 6300 ˚A. Bothlines are in the redder portion of MUSE’s wavelength cov-erage where the velocity resolution is (cid:38) km s − . We fita Gaussian to the line profiles as a function of position inthe jet/outflow, allowing us to marginally resolve the ve-locity difference between the two lobes. The western limbof the jet is blueshifted compared to the eastern limb ofthe jet, with a velocity difference between the two limbsof ∼ km s − (see Figure 7). The kinematic structure ofthe irradiated outflow is spectrally unresolved with MUSE,consistent with the small line-of-sight velocities reported byReiter et al. (2015a). The HH 900 jet-driving source is embedded in the opaqueglobule and cannot be seen with MUSE. However, the less- obscured protostar to the west of the globule, identified asPCYC 838 by Povich et al. (2011), is readily detected withMUSE. This YSO is spatially coincident with the westernlimb of the HH 900 outflow and may drive a microjet of itsown (Smith et al. 2010). We perform aperture photometry toextract the spectrum of the star, although given its location,the stellar spectrum could not be cleanly separated from theoutflow. We therefore focus our analysis on lines that areunlikely to be excited in the nebular gas.The spectrum of PCYC 838 is shown in Figure 8.Continuum emission rises slightly toward redder wave-lengths indicating that the source is young. Lines like Ca i λ , , ˚A seen in absorption suggest a G or F spec-tral type. Determining a more robust spectral type of thesource is difficult given the complex structure of the neb-ular emission surrounding the source that prevents a cleansubtraction from the stellar spectrum. Model fits to the IRspectral energy distribution from Povich et al. (2011) giveM (cid:63) = . ± . M (cid:12) (L (cid:63) = . ± . L (cid:12) ), suggesting that theforming YSO will be an intermediate-mass star. In addition,excess infrared emission from the source indicates circum-stellar material (Povich et al. 2011) that may feed on-goingaccretion which will veil the photospheric absorption lines.Indeed, the near-IR Ca ii triplet is seen in emission, whichHillenbrand & Hartmann (1998) showed is tightly correlatedwith strong accretion. The Ca ii λ ˚A line is stronger thanthe pure H line at λ ˚A (see Figure 8 and Hillenbrandet al. 2013), providing additional evidence for accretion.The spectral resolution of MUSE is inadequate to tellif emission lines likely to be excited in the putative micro-jet are at a different velocity than the HH 900 jet+outflowsystem. Given the clear relationship between accretion andoutflow activity (e.g., Hartigan et al. 1995), there may bea jet and/or outflow associated with PCYC 838 that liesclose to the plane of the sky that cannot be detected at thisspectral resolution. We present new MUSE+GALACSI-AO observations of theTadpole globule and HH 900 jet+outflow system in Carina,providing an unprecedented view of a small globule subjectto intense stellar feedback. Together with spatially-resolvedmaps of the cold, molecular gas (Paper II), this multiwave-length view reveals the anatomy of the system and allows us
MNRAS000
MNRAS000 , 1–18 (2019) adpole metamorphasis Figure 7.
Top:
Position-velocity diagram of the wide-angle outflow ([O i ] 6300 ˚A; colorscale) and collimated jet ([Fe ii ] 7155 ˚A; contours)taken through the same tracing shown in Figure 2. Bottom:
The peak velocity at each position of the jet/outflow determined from aGaussian fit to the line profile.
Figure 8.
Spectrum of the YSO, PCYC 838 (see Figure 1), that lies in the western limb of the outflow ( top ). Ca i absorption suggestsa G or F spectral type, although model fits suggest M (cid:63) ∼ . M (cid:12) ( left ; see Section 5). Strong Ca ii emission suggests ongoing accretion( right ).MNRAS , 1–18 (2019) M. Reiter et al. to measure the external influence of the nearby high-massstars on both the globule and the jet+outflow system. As anisolated globule adrift in the H ii region, the Tadpole rep-resents an evolutionary stage between star-forming pillars(e.g., Klaassen et al. 2014, Klaassen et al. 2019, subm.) andrevealed photoevaporating protoplanetary disks (e.g., O’dell& Wen 1994). Understanding this intermediate stage is im-portant for developing a more complete view of star forma-tion in feedback-dominated environments. In the followingdiscussion, we consider the fate of the Tadpole globule inthe context of its environment. Physical properties in the ionization front depend on thestrength and shape of the ionizing spectrum (Sankrit &Hester 2000). The Tadpole resides near Tr16 which has awell-studied population of OB stars with tabulated spectraltypes and ionizing photon luminosities (e.g., Smith 2006a;Alexander et al. 2016). Several morphological clues suggestthat the Tadpole lies in front of Tr16: (1) the globule hasa v LSR ≈ − km s − compared to v LSR ≈ − km s − forCarina (see Paper II and Rebolledo et al. 2016); (2) the Tad-pole tail lies in front of the eastern outflow limb (Smith et al.2010) and is blueshifted with respect to the head, as thoughtracing the remnants of a pillar that points away from theobserver, toward the star cluster (Paper II); (3) bright emis-sion lines from the ionization front are detected on both theside of the Tadpole closest to Tr16 and on the far side ofthe globule (see Figure 1); and (4) the center of the globuleappears to be opaque, suggesting little illumination of thenear side of the globule.Excitation diagnostics from the MUSE data allow us tofurther constrain the local irradiation of the Tadpole. Theratio [O iii ]/H α is a proxy for the spectral type of the dom-inant ionizing source as it provides an estimate of the ratioof photons with energy to ionize H (13.6 eV) and those ade-quately energetic to further ionize O + (35.1 eV, see Table 1).In the Tadpole globule and HH 900 jet+outflow system, theratio varies between ∼ . − . , suggesting an ionizingsource with spectral type ∼ O8 or later (Sankrit & Hester2000). This is somewhat later than the spectral type of thenearest O-type star seen in projection, the O5 V star CPD-59 2641 (see Reiter et al. 2015a).In photoionization equilibrium, the density in the ion-ization front will increase in proportion to the overall ioniz-ing flux (Sankrit & Hester 2000). Reiter et al. (2015a) usedthis to estimate n IF ≈ cm − by assuming that the near-est O-type star dominates the excitation. This density maybe computed as n IF = (cid:115) Q H π D r α B (3)where Q H is the ionizing photon luminosity, r is the radiusof curvature of the globule, and D is the projected distancebetween the ionizing source and the globule. Ambiguitiesin the projected separations contribute as ∼ D − and thusare a larger source of uncertainty than the ionizing flux ofa given star (which goes as ∼ √ Q H ). This uncertainty willaffect the estimated photoevaporation rate (See Section 6.3), particularly when no independent density measurement isavailable.Fortunately, the MUSE data allow us to make an inde-pendent measurement of the density (see Section 3.2). Thisallows us to invert the computation and estimate the inci-dent ionizing flux required to create the observed density.Assuming ionizations are balanced by recombinations in apure hydrogen nebula, Q H π D = an e n p L α H α ef f (4)where Q H is the luminosity of ionizing photons in s − , D is the distance in pc, a is a dimensionless factor of order 1that accounts for geometric effects, n e = n p are the electronand proton (H + ) density, respectively, L is the emitting pathlength, and α H α ef f is the recombination coefficient. Assumingthe median distance to the OB stars in Tr16 from Alexan-der et al. (2016) and taking the emitting path length to be L ≈ . pc (Smith et al. 2010), we find lo g ( Q H ) ∼ . , twoorders of magnitude less than the cumulative ionizing photonluminosity from Tr16 ( lo g ( Q H ) = . , see Smith 2006a).This suggests that the Tadpole/HH 900 are not subject toor shielded from the most extreme feedback from Tr16, de-spite its close projected separation. Nevertheless, feedbackis clearly affecting the jet and globule and may be rapidlyclearing the remaining reservoir of gas and dust. The prototypical cloud model used in many theoretical in-vestigations of the impact of ionizing radiation is the isother-mal Bonnor-Ebert (BE) sphere (Bonnor 1956; Ebert 1955).While the model assumes equilibrium, it is unstable to col-lapse if sufficiently compressed. Haworth et al. (2015) usedpressure-bounded BE spheres to model the stability and evo-lution of small globulettes in Carina cataloged by Grenman& Gahm (2014), including the Tadpole.We revisit the parameters of the Tadpole assuming it iswell described by a BE sphere but using the physical param-eters we measure with the MUSE data. The key differencein our approach compared to that of Haworth et al. (2015)is that we consider the pressure provided by ongoing ioniza-tion on the globule surface, rather than assuming that thehot, diffuse gas of the H ii region confines the globule. In thisway, we examine the stability of the globule against collapseunder the influence of external ionizing radiation, in roughanalogy to the approach of Bertoldi (1989). For sufficientlylow incident ionization and/or adequately high column den-sities, clouds will be relatively unaffected by the influence ofthe ionizing radiation.Following Hartmann (2009), we compute the BE massas M BE = . c s G / √ p (5)where c s = (cid:112) k B T / µ m H is the sound speed in the glob-ule, and p is the gas pressure of the confining medium, p = n e k B T e . Assuming the temperature in the globule is T ≈ K (Roccatagliata et al. 2013) and a mean molecu-lar weight of µ = . , we find c s ≈ . km s − in the coldgas. Together with the density and temperature we measure MNRAS000
Spectrum of the YSO, PCYC 838 (see Figure 1), that lies in the western limb of the outflow ( top ). Ca i absorption suggestsa G or F spectral type, although model fits suggest M (cid:63) ∼ . M (cid:12) ( left ; see Section 5). Strong Ca ii emission suggests ongoing accretion( right ).MNRAS , 1–18 (2019) M. Reiter et al. to measure the external influence of the nearby high-massstars on both the globule and the jet+outflow system. As anisolated globule adrift in the H ii region, the Tadpole rep-resents an evolutionary stage between star-forming pillars(e.g., Klaassen et al. 2014, Klaassen et al. 2019, subm.) andrevealed photoevaporating protoplanetary disks (e.g., O’dell& Wen 1994). Understanding this intermediate stage is im-portant for developing a more complete view of star forma-tion in feedback-dominated environments. In the followingdiscussion, we consider the fate of the Tadpole globule inthe context of its environment. Physical properties in the ionization front depend on thestrength and shape of the ionizing spectrum (Sankrit &Hester 2000). The Tadpole resides near Tr16 which has awell-studied population of OB stars with tabulated spectraltypes and ionizing photon luminosities (e.g., Smith 2006a;Alexander et al. 2016). Several morphological clues suggestthat the Tadpole lies in front of Tr16: (1) the globule hasa v LSR ≈ − km s − compared to v LSR ≈ − km s − forCarina (see Paper II and Rebolledo et al. 2016); (2) the Tad-pole tail lies in front of the eastern outflow limb (Smith et al.2010) and is blueshifted with respect to the head, as thoughtracing the remnants of a pillar that points away from theobserver, toward the star cluster (Paper II); (3) bright emis-sion lines from the ionization front are detected on both theside of the Tadpole closest to Tr16 and on the far side ofthe globule (see Figure 1); and (4) the center of the globuleappears to be opaque, suggesting little illumination of thenear side of the globule.Excitation diagnostics from the MUSE data allow us tofurther constrain the local irradiation of the Tadpole. Theratio [O iii ]/H α is a proxy for the spectral type of the dom-inant ionizing source as it provides an estimate of the ratioof photons with energy to ionize H (13.6 eV) and those ade-quately energetic to further ionize O + (35.1 eV, see Table 1).In the Tadpole globule and HH 900 jet+outflow system, theratio varies between ∼ . − . , suggesting an ionizingsource with spectral type ∼ O8 or later (Sankrit & Hester2000). This is somewhat later than the spectral type of thenearest O-type star seen in projection, the O5 V star CPD-59 2641 (see Reiter et al. 2015a).In photoionization equilibrium, the density in the ion-ization front will increase in proportion to the overall ioniz-ing flux (Sankrit & Hester 2000). Reiter et al. (2015a) usedthis to estimate n IF ≈ cm − by assuming that the near-est O-type star dominates the excitation. This density maybe computed as n IF = (cid:115) Q H π D r α B (3)where Q H is the ionizing photon luminosity, r is the radiusof curvature of the globule, and D is the projected distancebetween the ionizing source and the globule. Ambiguitiesin the projected separations contribute as ∼ D − and thusare a larger source of uncertainty than the ionizing flux ofa given star (which goes as ∼ √ Q H ). This uncertainty willaffect the estimated photoevaporation rate (See Section 6.3), particularly when no independent density measurement isavailable.Fortunately, the MUSE data allow us to make an inde-pendent measurement of the density (see Section 3.2). Thisallows us to invert the computation and estimate the inci-dent ionizing flux required to create the observed density.Assuming ionizations are balanced by recombinations in apure hydrogen nebula, Q H π D = an e n p L α H α ef f (4)where Q H is the luminosity of ionizing photons in s − , D is the distance in pc, a is a dimensionless factor of order 1that accounts for geometric effects, n e = n p are the electronand proton (H + ) density, respectively, L is the emitting pathlength, and α H α ef f is the recombination coefficient. Assumingthe median distance to the OB stars in Tr16 from Alexan-der et al. (2016) and taking the emitting path length to be L ≈ . pc (Smith et al. 2010), we find lo g ( Q H ) ∼ . , twoorders of magnitude less than the cumulative ionizing photonluminosity from Tr16 ( lo g ( Q H ) = . , see Smith 2006a).This suggests that the Tadpole/HH 900 are not subject toor shielded from the most extreme feedback from Tr16, de-spite its close projected separation. Nevertheless, feedbackis clearly affecting the jet and globule and may be rapidlyclearing the remaining reservoir of gas and dust. The prototypical cloud model used in many theoretical in-vestigations of the impact of ionizing radiation is the isother-mal Bonnor-Ebert (BE) sphere (Bonnor 1956; Ebert 1955).While the model assumes equilibrium, it is unstable to col-lapse if sufficiently compressed. Haworth et al. (2015) usedpressure-bounded BE spheres to model the stability and evo-lution of small globulettes in Carina cataloged by Grenman& Gahm (2014), including the Tadpole.We revisit the parameters of the Tadpole assuming it iswell described by a BE sphere but using the physical param-eters we measure with the MUSE data. The key differencein our approach compared to that of Haworth et al. (2015)is that we consider the pressure provided by ongoing ioniza-tion on the globule surface, rather than assuming that thehot, diffuse gas of the H ii region confines the globule. In thisway, we examine the stability of the globule against collapseunder the influence of external ionizing radiation, in roughanalogy to the approach of Bertoldi (1989). For sufficientlylow incident ionization and/or adequately high column den-sities, clouds will be relatively unaffected by the influence ofthe ionizing radiation.Following Hartmann (2009), we compute the BE massas M BE = . c s G / √ p (5)where c s = (cid:112) k B T / µ m H is the sound speed in the glob-ule, and p is the gas pressure of the confining medium, p = n e k B T e . Assuming the temperature in the globule is T ≈ K (Roccatagliata et al. 2013) and a mean molecu-lar weight of µ = . , we find c s ≈ . km s − in the coldgas. Together with the density and temperature we measure MNRAS000 , 1–18 (2019) adpole metamorphasis from the MUSE data ( n e ∼ cm − and T ∼ K, seeFigure 2 and Section 3.2) to compute the confining pres-sure, we estimate a globule mass M BE ∼ . M (cid:12) . This istwo orders of magnitude higher than the ∼ M Jupiter (or0.01 M (cid:12) ) estimated by Grenman & Gahm (2014) and re-garded as a lower limit due to superimposed bright nebu-losity. Indeed, subsequent single-dish observations of someof the Carina globulettes (not including the Tadpole) foundsystematically higher masses than those estimated from theextinction (Haikala et al. 2017).Using M BE , we compute the corresponding radius of theBE sphere, R BE = . Gc S M BE ≈ , AU, or ∼ . (cid:48)(cid:48) at adistance of 2.3 kpc, roughly twice the observed major axisof the globule ( ∼ . (cid:48)(cid:48) ). The BE radius corresponding to themass estimated by Grenman & Gahm (2014), ∼ . (cid:48)(cid:48) wouldbe unresolved, even with HST .Assuming a uniform density, we can use the BE massto estimate the column density. This allows us to comparemore directly with the approach taken by Grenman & Gahm(2014) who used A V estimates to infer the column den-sity, and thus mass of small globulettes. Given our highermass, we estimate a column density and extinction ∼ × higher than that obtained by Grenman & Gahm (2014). A14 M Jupiter uniform-density globule with the size of the Tad-pole corresponds to a column density N ( H ) ≈ × cm − ,or an A V ≈ . mag. This is within ∼ mag of the A V weestimate in the ionized gas using the ratio of hydrogen re-combination lines (see Section 3.1 and Figure 2). In contrast,from M BE , we estimate N ( H ) ≈ × cm − , correspond-ing to A V (cid:38) mag. We find the latter estimate morephysically realistic given that the jet morphology and kine-matics demand a protostar embedded in the opaque globulethat remains unseen even in the near-IR, requiring a highextinction to obscure the driving source of one of the mostpowerful jets in Carina.Bertoldi (1989) explored many of these same parame-ters to determine the fate of initially neutral clouds subjectto external ionizing radiation. The stability of clouds de-pends on the column density ( A V ), and the strength of theincident radiation. Bertoldi (1989) defined regions of this pa-rameter space more or less stable against the compressionfrom external ionization. For both the BE mass estimate andthat produced by Grenman & Gahm (2014), the Tadpoleparameters suggest that the cloud will be compressed bythe ionization front (region II in Figure 1 of Bertoldi 1989).Within the Bertoldi (1989) framework, we expect that moreionizing photons will be absorbed before reaching the ioniza-tion front at higher column densities (the BE estimate) andthe fractional size of the ionization front, compared to thesize of the cloud itself, will be smaller. Using the emittingpath length of ionizing photons L ≈ . pc, we estimatethe fractional size of the ionization front compared to theglobule is ∼ . , larger than Bertoldi estimate even for theGrenman & Gahm (2014) mass. Since the H α emission willalso trace the photoevaporative flow, a more representativemeasure of the size of the ionization front may be the off-set between emission lines of different ionization potentials(e.g., Carlsten & Hartigan 2018), but we cannot resolve theseparation between the characteristic emission lines in theMUSE data even with AO (see Figure 3).This simplistic analysis neglects some key observational features of the system, most notably that there is unambigu-ous evidence that the globule has already undergone col-lapse and formed a protostar. The free fall time for a cloudwith the Tadpole’s estimated mass and radius is < . Myr( < . Myr for the Grenman & Gahm 2014 mass estimate),suggesting that the globule has already contracted from anequilibrium sphere. In addition, the existence of the embed-ded jet-driving source suggests that the globule is likely notisothermal. More subtly, most of the assumptions that wehave made presume the present-day distribution and fluxof ionizing sources. However, until recently, η Carinae, lo-cated ∼ . pc in projection from the Tadpole, dominatedthe ionizing flux of Tr16 with an estimated main-sequenceionizing photon luminosity log ( Q H ) ≈ . s − . On the mainsequence, η Car alone would increase the incident ionizingphoton luminosity by an order of magnitude.
Assuming the Tadpole is a sphere, Reiter et al. (2015a) es-timated the photoevaporation rate as (cid:219) M phot (cid:39) π r µ m H n H v (6)where r ≈ . pc is the radius of curvature of the globule, m H is the mass of hydrogen, n H ≈ . n e ∼ cm − is thedensity of neutral hydrogen (Sankrit & Hester 2000), and v ≈ c I I is the speed of the evaporative flow, assumed to be thesound speed in ionized gas. The resulting photoevaporationrate of (cid:219) M ∼ × − M (cid:12) yr − suggests that a 1 M (cid:12) globulewill survive an additional (cid:46) yr.Our new MUSE data allow for better constraints on thephysical properties of the globule. Using the density deter-mined from the [S ii ] line ratio (see Section 3.2), we find avalue that is an order of magnitude lower than Reiter et al.(2015a) infer from the ionizing photon flux of the nearestO-type star ( ∼ × cm − compared to ∼ × cm − ,respectively). This leads to a corresponding reduction in theestimated photoevaporation rate, ∼ × − M (cid:12) yr − , andincrease in the estimated globule lifetime.At the same time, the structure of the wide-angle out-flow emerging from the Tadpole demonstrates that signifi-cant mass may be lost to jet entrainment. The outflow isnearly the same width as the globule itself with dusty fila-ments extending into the outflow from the western edge ofthe globule (Smith et al. 2010). The mass-loss rate in theionized outflow is comparable to the photoevaporation rate, (cid:219) M out f low ∼ × − M (cid:12) yr − (from Smith et al. 2010; wenote that the lower outflow velocity measured by Reiter et al.2017b is offset by the higher density we measure with theMUSE data in Section 3.2). However, this mass-loss rate isa lower limit as it only accounts for the ionized gas and doesnot include the partially ionized gas traced by [C i ] and H or any molecules or dust in the outflow (see Section 4 andReiter et al. 2015a, Paper II).It is difficult to constrain how much mass-loss from jetentrainment contributes to destruction of the globule, giventhe time-variable nature of accretion and outflow as well asthe uncertainties in the entrainment efficiency. As a roughestimate, we compare the kinetic energy of the jet to thegravitational binding energy of the globule. HH 900 is one ofthe most energetic jets in Carina, with E (cid:38) . × erg (Re-iter et al. 2017b), and likely higher given evidence for higher MNRAS , 1–18 (2019) M. Reiter et al. densities in the jet (see Section 4.1). Assuming the Tadpoleis a sphere, the gravitational binding energy U = GM R is U ≈ × erg using M BE estimated in Section 6.2. Thekinetic energy of the jet alone is within a factor of 3 of thegravitational binding energy of the globule. The similarityof these numbers suggests that the jet may play a signifi-cant role in the ultimate destruction of the globule. At thesame time, the survival of the globule despite the power-ful jet+outflow illustrates the inefficient transfer of momen-tum and energy to the natal cloud. Observations of the coldmolecular gas in the globule (Paper II) and more detailedmodels are needed to constrain the role of the jet in theultimate destruction of the globule. Gahm et al. (2007); Grenman & Gahm (2014) argue thatthe small globulettes seen in H ii regions may contributesignificant numbers of planetary mass objects even if only asmall number of the existing clouds collapse. However, Ha-worth et al. (2015) conclude that the globulettes are stableand unlikely to collapse without an external perturbation.Results for the Tadpole are significantly at odds with theconclusions of Grenman & Gahm (2014). We argue that theTadpole and other globules must be of order stellar masses(rather than planetary masses) in order to prevent rapidphotoablation. For either mass estimate, we find the exter-nal force of the ionization front on the globule surface willcompress the globule.While the Tadpole globule itself is unambiguously star-forming, its evolution and interaction with its environmentare relevant for any planets that may form around the em-bedded jet-driving source. The Tadpole shares a few char-acteristics with some models for the formation of the SolarSystem. Fossil evidence in Solar System meteorites indicatesthat the Sun formed near a source of short-lived radioactiveisotopes that are synthesized in the death of high-mass stars(e.g., Adams 2010). Various authors have considered the roleof a nearby supernova explosion in enriching a prestellarcloud core and triggering its collapse (e.g., Cameron & Tru-ran 1977; Boss 2018). The timescales required to trigger anew star formation event are at odds with the short half-lives of the essential radioactive elements (e.g., Parker &Dale 2016), suggesting that direct disk enrichment is stillthe most likely pathway. However, disks may be rapidly de-stroyed under the influence of external ionization (e.g., Win-ter et al. 2018; Nicholson et al. 2019, although see Richertet al. 2015), leaving little to enrich by the time nearby starsexplode as supernovae.Young objects like the Tadpole may be a prototype for athird option that is intermediate between these well-studiedpossibilities. The young dynamical age of the HH 900 jet( ∼ yr, see Reiter et al. 2015a) suggests that its driv-ing source formed after most of Tr16, where multiple post-main-sequence stars suggest an age (cid:38) Myr (Walborn 1995;Getman et al. 2014). For the estimated Tadpole photoevap-oration rate of ∼ × − M (cid:12) yr − , a ∼ . M (cid:12) globule willbe completely photoevaporated in ∼ Myr. Including theimpact of the HH 900 jet+outflow system on the globule,this remaining lifetime may be even shorter. This raises thepossibility that the disk-bearing YSO will emerge into the H ii region (similar to the proplyds in Orion) during thesupernova era in Carina.The Tadpole is one of the largest objects in the Gren-man & Gahm (2014) catalog in both its physical extent andestimated mass. It is unclear what fraction of the other glob-ulettes may also be of order stellar mass. If they have indeedbeen compressed in the H ii , then high optical depths mayobscure embedded protostars, as is the case in the Tadpole.Nevertheless, the Tadpole provides a cautionary tale thatdemonstrates the limits of indirect mass estimates. Resolvedsubmillimeter observations provide a more direct mass esti-mate as well as gas kinematics to measure the dynamicalimpact of feedback on the gas (Paper II). In the absense ofsuch a survey of globulettes, we urge caution in consideringtheir contribution and influence on the bottom of the IMF. We present new AO-assisted MUSE observations of theTadpole globule and HH 900 jet+outflow system in theCarina Nebula. The broad spectral coverage of MUSE al-lows us to measure several diagnostic ratios and determinethe variation in physical parameters as a function of posi-tion throughout the system. We measure modest extinctions( A V ≈ . mag) in the system using a tracer only sensitiveto the ionized gas. Densities ( n e ≈ cm − ) and temper-atures ( T e ≈ K) are high in both the globule and thejet+outflow system. There is a clear increase in the excita-tion and ionization of the outflow with increasing distancefrom the opaque Tadpole globule.Atomic emission lines also reveal the change in excita-tion in the outflow with increasing distance from the globule.We detect [C i ] 8727 ˚A, which traces partially ionized gas,in the inner outflow close to the globule. Evidence for ab-sorption in high excitation lines like [O iii ] and [S iii ] in theoutflow suggest the presence of dust. Together, this demon-strates that the HH 900 outflow is not fully ionized. The col-limated jet that threads the ionized HH 900 outflow is seenin high-density tracers, suggesting that the density, and thusthe mass-loss rate, in the jet may be as much as an order ofmagnitude higher than estimated by Reiter et al. (2015a).As a result, the kinetic energy of the jet alone (not includingthe outflow) is within a factor of three of the gravitationalbinding energy of the Tadpole.We extract the spectrum of the YSO that lies projectedonto the western limb of the HH 900 outflow. While con-tamination from the environment prevents robust spectraltyping, bright emission from the near-IR Ca ii triplet indi-cates strong on-going accretion, consistent with evidence fora circumstellar disk presented by Povich et al. (2011). Theprotostar that drives the HH 900 jet+outflow system is em-bedded within the opaque Tadpole globule and not detectedwith MUSE.We estimate a Bonnor-Ebert mass of ∼ . M (cid:12) for theglobule, ∼ × larger than previous estimates. Assuminga uniform density globule, we estimate a column density N ( H ) ∼ × cm − , corresponding to A V (cid:38) mag, inagreement with a globule so opaque that the HH 900 driv-ing source remains unseen even in the infrared. The electrondensity in the ionization front is lower than previous esti-mates, leading to a smaller photoevaporation rate and longer MNRAS , 1–18 (2019) adpole metamorphasis survival time of the globule assuming that photoevaporationdominates the destruction of the globule. Observations of thecold, molecular gas in the globule (Paper II) will allow us totest these estimates.The results from this MUSE study suggest that smallglobules may play an interesting role in the H ii regionecosystem. Stars that form in small globules will be exposedto the most intense feedback later in the life of the nearbyhigh-mass stars. This may provide a pathway for planet-forming disks to survive destructive ionizing radiation andthus be enriched with essential elements synthesized in thedeath of high-mass stars. While we present strong evidencethat the Tadpole globule is orders of magnitude more mas-sive than previous estimates, it is not clear what fractionof the globulettes may be stellar-mass objects (rather thanplanetary-mass). Nevertheless, the discrepancy in the Tad-pole mass estimates point to the need for caution when con-sidering the contribution that globulettes might make to thebottom of the IMF. ACKNOWLEDGEMENTS
We thank the referee, Dr G¨osta Gahm, for a timely andthoughtful report. MR would like to thank Libby Jones forhelpful discussions and Tom Haworth for a careful readingof the manuscript. In loving memory of John Causland. Thisproject has received funding from the European Union’sHorizon 2020 research and innovation programme under theMarie Sk´lodoska-Curie grant agreement No. 665593 awardedto the Science and Technology Facilities Council. JCM ac-knowledges support from the European Research Councilunder the European Communityˆa ˘A´Zs Horizon 2020 frame-work program (2014-2020) via the ERC Consolidator grantˆa ˘A¨YFrom Cloud to Star Formation (CSF)ˆa ˘A´Z (project num-ber 648505). This paper is based on data obtained with ESOtelescopes at the Paranal Observatory under programmeID 0101.C-0391(A). This research made use of Astropy, a community-developed core Python package for Astron-omy (Astropy Collaboration et al. 2013; Price-Whelan et al.2018); APLpy, an open-source plotting package for Python(Robitaille & Bressert 2012); and PySpecKit (Ginsburg &Mirocha 2011). REFERENCES
Adams F. C., 2010, ARA&A, 48, 47Alexander M. J., Hanes R. J., Povich M. S., McSwain M. V.,2016, AJ, 152, 190Allen M. G., Groves B. A., Dopita M. A., Sutherland R. S., Kew-ley L. J., 2008, ApJS, 178, 20Astropy Collaboration et al., 2013, aap, 558, A33Bacon R., et al., 2010, in Ground-based and Airborne Instrumen-tation for Astronomy III. p. 773508, doi:10.1117/12.856027Bally J., Licht D., Smith N., Walawender J., 2006, AJ, 131, 473Bastian N., Covey K. R., Meyer M. R., 2010, ARA&A, 48, 339Bautista M. A., Pradhan A. K., 1998, ApJ, 492, 650Bautista M. A., Pradhan A. K., Osterbrock D. E., 1994, ApJ,432, L135Bautista M. A., Peng J., Pradhan A. K., 1996, ApJ, 460, 372 arXiv:1608.00005 )Miao J., White G. J., Thompson M. A., Nelson R. P., 2009, ApJ,692, 382Momcheva I. G., Lee J. C., Ly C., Salim S., Dale D. A., OuchiM., Finn R., Ono Y., 2013, AJ, 145, 47MNRAS , 1–18 (2019) M. Reiter et al.
Nicholson R. B., Parker R. J., Church R. P., Davies M. B., FearonN. M., Walton S. R. J., 2019, arXiv e-prints,Nisini B., Caratti o Garatti A., Giannini T., Lorenzetti D., 2002,A&A, 393, 1035Nisini B., Bacciotti F., Giannini T., Massi F., Eisl¨offel J., PodioL., Ray T. P., 2005, A&A, 441, 159O’dell C. R., Wen Z., 1994, ApJ, 436, 194Ohlendorf H., Preibisch T., Gaczkowski B., Ratzka T., GrellmannR., McLeod A. F., 2012, A&A, 540, A81Osterbrock D. E., Ferland G. J., 2006, Astrophysics of gaseousnebulae and active galactic nucleiParker R. J., Dale J. E., 2016, MNRAS, 456, 1066Pattle K., et al., 2018, ApJ, 860, L6Pottasch S. R., 1956, Bull. Astron. Inst. Netherlands, 13, 77Pottasch S. R., 1958, Bull. Astron. Inst. Netherlands, 14, 29Povich M. S., et al., 2011, ApJS, 194, 14Price-Whelan A. M., et al., 2018, aj, 156, 123Raga A. C., Reipurth B., Castellanos-Ram´ırez A., Chiang H.-F.,Bally J., 2015, ApJ, 798, L1Rebolledo D., et al., 2016, MNRAS, 456, 2406Reipurth B., 1983, A&A, 117, 183Reiter M., Smith N., 2013, MNRAS, 433, 2226Reiter M., Smith N., Kiminki M. M., Bally J., Anderson J., 2015a,MNRAS, 448, 3429Reiter M., Smith N., Kiminki M. M., Bally J., 2015b, MNRAS,450, 564Reiter M., Kiminki M. M., Smith N., Bally J., 2017a, MNRAS,467, 4441Reiter M., Kiminki M. M., Smith N., Bally J., 2017b, MNRAS,470, 4671Richert A. J. W., Feigelson E. D., Getman K. V., Kuhn M. A.,2015, ApJ, 811, 10Robitaille T., Bressert E., 2012, APLpy: Astronomical Plot-ting Library in Python, Astrophysics Source Code Library(ascl:1208.017)Roccatagliata V., Preibisch T., Ratzka T., Gaczkowski B., 2013,A&A, 554, A6Sandford II M. T., Whitaker R. W., Klein R. I., 1982, ApJ, 260,183Sankrit R., Hester J. J., 2000, ApJ, 535, 847Schneps M. H., Ho P. T. P., Barrett A. H., 1980, ApJ, 240, 84Smith R. G., 1987, MNRAS, 227, 943Smith N., 2002, MNRAS, 331, 7Smith N., 2006a, MNRAS, 367, 763Smith N., 2006b, ApJ, 644, 1151Smith N., Barb´a R. H., Walborn N. R., 2004, MNRAS, 351, 1457Smith N., Bally J., Walborn N. R., 2010, MNRAS, 405, 1153Thackeray A. D., 1950, MNRAS, 110, 524Tremblin P., et al., 2013, A&A, 560, A19Walborn N. R., 1995, in Niemela V., Morrell N., Feinstein A., eds,Revista Mexicana de Astronomia y Astrofisica Conference Se-ries Vol. 2, Revista Mexicana de Astronomia y AstrofisicaConference Series. p. 51Weilbacher P. M., Streicher O., Urrutia T., Jarno A., P´econtal-Rousset A., Bacon R., B¨ohm P., 2012, in Softwareand Cyberinfrastructure for Astronomy II. p. 84510B,doi:10.1117/12.925114Winter A. J., Clarke C. J., Rosotti G., Ih J., Facchini S., HaworthT. J., 2018, MNRAS, 478, 2700Yeh S. C. C., Seaquist E. R., Matzner C. D., Pellegrini E. W.,2015, ApJ, 807, 117Zhang Z.-Y., Romano D., Ivison R. J., Papadopoulos P. P., Mat-teucci F., 2018, Nature, 558, 260
APPENDIX A: LINES DETECTED IN THETADPOLE
MNRAS000
MNRAS000 , 1–18 (2019) adpole metamorphasis Table A1: Lines detected in MUSE dataLine λ componentname [˚A]H β [ O iii ] [ O iii ] [ N ii ] [ O i ] [ S iii ] [ O i ] [ N ii ] α ii ? 6578.03 haze around globule [ N ii ] i [ S ii ] [ S ii ] i [ Ar iii ] [ Fe ii ] i [ Ca ii ] ∗ [ O ii ] [ Ca ii ] ∗ [ O ii ] [ Ni ii ] [ Fe ii ] [ Fe ii ] ∗ [ Ar iii ] i [ Cl ii ] [ Fe ii ] [ C i ] ∗∗ -bright inner outflowPaschen 12-3 8750.48 outflow, globule edgePaschen 11-3 8862.79 outflow, globule edge, bow shocks [ Fe ii ] [ Fe ii ] ∗ [ S iii ] ∗ ID from Ellerbroek et al. (2013); ∗∗ ID from Liu et al. (1995)
APPENDIX B: INTENSITY MAPS
MNRAS , 1–18 (2019) M. Reiter et al.
Figure B1.
Continuum-subtracted integrated intensity maps of the Tadpole globule and the HH 900 jet+outflow system.MNRAS000
Continuum-subtracted integrated intensity maps of the Tadpole globule and the HH 900 jet+outflow system.MNRAS000 , 1–18 (2019) adpole metamorphasis Figure B2. (continued) Continuum-subtracted integrated intensity maps of the Tadpole globule and the HH 900 jet+outflow system.MNRAS , 1–18 (2019) M. Reiter et al.
Figure C1.
Line ratio maps of the Tadpole globule and the HH 900 jet+outflow system.
APPENDIX C: LINE RATIO MAPS
This paper has been typeset from a TEX/L A TEX file prepared by the author. MNRAS000