Looking bok globules


Published on

Published in: Technology
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Looking bok globules

  1. 1. Looking into the hearts of Bok globules: MM and submm continuum images of isolated star-forming cores R. Launhardt Max Planck Institute for Astronomy, K¨ nigstuhl 17, D-69117 Heidelberg, Germany oarXiv:1004.0869v1 [astro-ph.GA] 6 Apr 2010 rl@mpia.de and D. Nutter, D. Ward-Thompson School of Physics and Astronomy, Cardiff University, Queens Buildings, The Parade, Cardiff, CF24 3AA, UK and T. L. Bourke Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA 02138, USA and Th. Henning, T. Khanzadyan1 , M. Schmalzl, S. Wolf2 Max Planck Institute for Astronomy, K¨ nigstuhl 17, D-69117 Heidelberg, Germany o and R. Zylka Institut de Radio Astronomie Millim´ trique, 300 Rue de la Piscine, Domaine Universitaire, 38406 Saint Martin e dH` res, France e ABSTRACT We present the results of a comprehensive infrared, submillimetre, and millimetre continuum emission study of isolated low-mass star-forming cores in 32 Bok globules, with the aim to investigate the process of star formation in these regions. The submillimetre and millimetre dust continuum emission maps together with the spectral energy distributions are used to model and derive the physical properties of the star-forming cores, such as luminosities, sizes, masses, densities, etc. Comparisons with ground-based near-infrared and space-based mid and far-infrared images from Spitzer are used to reveal the stellar content of the Bok globules, association of embedded young stellar objects with the submm dust cores, and the evolutionary stages of the individual sources. Submm dust continuum emission was detected in 26 out of the 32 globule cores observed. For 18 globules with detected (sub)mm cores we derive evolutionary stages and physical parameters of the embedded sources. We identify nine starless cores, most of which are presumably prestellar, nine Class 0 protostars, and twelve Class I YSOs. Specific source properties like bolometric temperature, core size, and central densities are discussed as function of evolutionary stage. We find that at least two thirds (16 out of 24) of the star-forming globules studied here show evidence of forming multiple stars on scales between 1,000 and 50,000 AU. However, we also find that most of these small prototstar and star groups are comprised of sources with different evolutionary stages, suggesting a picture of slow and sequential star formation in isolated globules. Subject headings: dust, extinction — infrared: stars — ISM: clouds — stars: formation — submillimeter: ISM 1
  2. 2. 1. Introduction to the sharper rim at the “head” side (e.g., Launhardt 1996). Similarly, pre-stellar cores, which are the ear- Different aspects of star formation can be stud- liest stage of star formation (Ward-Thompson et al.ied on different size scales and in different environ- 1994, 2007) often appear to be fragmentary and fil-ments. The large-scale distribution of star-forming re- amentary. However, the protostellar cores and en-gions and the relation between molecular cloud life velopes of the more evolved Class 0 & I YSOs (Ladacycles, galactic spiral density waves, and star forma- 1987; Andr´ et al. 1993) are more spherically symmet- etion can be studied by observing nearby galaxies (e.g. ric, which can be understood as a result of the gravi-Brouillet et al. 1998; Scheepmaker et al. 2009). The tational collapse of the inner dense R ∼ 5000 AU re-stellar initial mass function (IMF), which is needed gion. Many of these isolated cores were found to be theto interpret these data, is usually derived from rich sources of bipolar molecular outflows, indicating theyoung stellar clusters in our own Galaxy (e.g. Kroupa presence of embedded protostars (e.g., Yun & Clemens2002; Chabrier 2003). Dense star-forming dark cloud 1994b; Ward-Thompson et al. 1995; Bontemps et al.complexes such as the Orion and Ophiuchus molecu- 1996).lar clouds are the places to study the relation between In order to investigate the star-forming potentialthe molecular core mass spectrum (CMF) and the in- and evolutionary stages of Bok globules in the solarterstellar IMF (e.g. Motte et al. 1998; Nutter & Ward- neighbourhood, we had surveyed a large number ofThompson 2007; Simpson et al. 2008; Goodwin et al. globules for signs of star-forming cores, using as trac-2008). ers, e.g., the mm dust continuum emission (Launhardt Nearby and more isolated star-forming cores, such 1996; Launhardt & Henning 1997; Henning & Laun-as Bok globules, are the best places to study in de- hardt 1998; Launhardt et al. 1998), NH3 (Bourke et al.tail the initial properties of individual star-forming 1995b) or CS line emission (Launhardt et al. 1998).cores, their chemical evolution, kinematic structure, In this paper we present a submillimetre and mil-and the physics of their collapse and fragmentation limetre – hereafter (sub)mm – continuum study of 32(e.g., Clemens & Barvainis 1988; Bourke et al. 1995a; Bok globules, which were identified from these previ-Launhardt & Henning 1997; Launhardt et al. 1997; ous surveys as promising candidates for globules withHenning & Launhardt 1998; Ward-Thompson et al. currently ongoing star formation. The (sub)mm maps2007; Stutz et al. 2008). Bok globules are small, are complemented by deep near-infrared (NIR) imagessimply-structured, relatively isolated, opaque molec- and NIR to mm spectral energy distributions (SEDs).ular clouds that often contain only one or two star- In Sect. 2, we describe the observations and data re-forming core. They are often not completely iso- duction. In Sect. 3, we present the NIR and (sub)mmlated, but reside in the filamentary outskirts of larger images, SEDs, and results on multiplicity, physical pa-dark cloud complexes (Launhardt & Henning 1997), a rameters, and evolutionary stages. In Sect. 4, we dis-fact that may tell something about their origin. With cuss in particular the source properties as function oftheir size, mass, densities, etc., Bok globules resemble evolutionary stage and the results on multiplicity. Insmall clumps in larger molecular clouds (cf. Bergin Sect. 5, we describe and discuss the individual glob-& Tafalla 2007), only that they lack the surrounding ules, and Sect. 6 summarizes the main results of thiscloud. Table 1 summarizes the average general prop- study.erties of typical Bok globules and their star-formingcores. 2. Observations and data reduction Although they are the most simple star-formingmolecular clouds, many globules deviate considerably A total of 32 Bok globules were studied using afrom spherical geometry. They are often cometary or combination of NIR and (sub)mm continuum obser-irregularly shaped. The dense star-forming cores are vations. Coordinates, distances, and observation sum-not always located at the center of the globule, but mary of these sources are given in Table 2. Althoughin cometary-shaped globules are often located closer many of the globules appear in different catalogues and thus often have several names assigned, we use 1 Dept. Experimental Physics, National University of Ireland throughout this paper the CB name and number in-Galway, Galway, Ireland troduced by Clemens & Barvainis (1988) for northern 2 University of Kiel, Institute of Theoretical Physics and Astro- sources, and the BHR name and number introduced byphysics, Leibnizstrasse 15, 24098 Kiel, Germany 2
  3. 3. Bourke et al. (1995a) for southern sources. and 450 µm using a 64-point jiggle pattern to create fully sampled maps of 2.3 diameter fields around the2.1. (Sub)millimetre continuum observations coordinate centers (for more details, see Holland et al. 1999). Time-dependent variations in the sky emis- Submillimetre and millimetre dust continuum ob- sion were removed by chopping the secondary mir-servations at 450 µm, 850 µm, and 1.3 mm were per- ror at 7.8 Hz with a chop throw of 120 in azimuth.formed during several observing runs between 1994 During the 2000 and 2001 observations, we also ob-and 2002 at the 15-m James Clerk Maxwell Telescope1 tained 850 µm polarization maps of six globule cores(JCMT) on Mauna Kea, Hawaii, at the 30-m Institut de (see Henning et al. 2001; Wolf et al. 2003). In addi-Radio Astronomie Millimetrique (IRAM) telescope on tion, the JCMT archive, operated by the Canadian As-Pico Veleta, Spain, and at the 15-m Swedish-ESO Sub- tronomy Data Centre (CADC), was searched for addi-millimetre Telescope (SEST) on La Silla, Chile. tional data for the sources observed at 1.3 mm. One Observations at 1.3 mm of the northern sources of the sources (CB224) was observed using SCUBAwere carried out during four observing runs between in the scan-mapping mode. In this observing mode,April 1994 and March 1997 using the MPIfR 7- the array is rastered across the sky to produce a rect-channel (1994–95) and 19-channel (1996–97) bolome- angular map several arcminutes in extent. Variationster arrays (Kreysa et al. 1998) at the IRAM 30-m tele- in sky emission are again removed by chopping thescope. The maps were obtained with the standard secondary mirror, though six chop configurations aredouble-beam technique described by Emerson et al. used, chopping in 30, 44 and 68 in both RA and Dec(1979). The telescope was scanned continuously at (Jenness et al. 1998).4 /sec in azimuth while the wobbling secondary, op- During all observing runs, the atmospheric trans-erating at 2 Hz with a chopper throw of 30 along the mission was determined using the ‘skydip’ method,scan direction, generated the dual beams. Scan lengths which were performed every 2–3 hours. At the JCMT,and map sizes were adapted to source sizes to pro- we also used polynomial fits to the 225 GHz atmo-vide sufficient basline for background determination. spheric opacity data which was measured by the Cal-Individual scan maps of a source, which were later tech Submillimeter Observatory (CSO) radiometer.combined, were obtained at different hour angles in Average zenith optical depths for each run are listed inorder to reduce scanning effects. Table 3. Observations at 1.3 mm of southern sources were Telescope pointing and focus were checked regu-carried out in September and November 2002 at the larly by observing strong standard continuum pointSwedish/ESO Submillimetre Telescope (SEST) at La sources. Pointing was found to be repeatable withinSilla in Chile, using the 37-channel bolometer array ≈3 at the IRAM 30-m telescope and the JCMT, andSIMBA (SEST Imaging Bolometer Array). Since ≈5 at SEST. For flux calibration, maps of Uranus,SIMBA was used without a wobbling secondary mir- Mars, Neptune, or different secondary calibratorsror, several fast-scanning maps (at 80 /sec) were ob- (Sandell 1994) were obtained during each shift andtained for each field at different hour angles to re- at different elevations. Observing parameters for allduce scanning effects. Source BHR 71 was already runs are summarized in Table 3.observed in 1995 with the MPIFR single-channelbolometer at SEST (Bourke et al. 1997). Individual 2.2. Reduction and calibration of the (sub)millimetremaps at different hour angles were obtained with the datasame double-beam mapping technique used at the 30-m telescope, but by employing the SEST focal plane 1.3 mm bolometer maps observed using the IRAMchopper. 30-m and SEST telescopes were reduced and an- The submillimetre observations were carried out alyzed with the MOPSI2 software package. Theduring three observing runs between August 1997 and chopped dual-beam scan maps were flat-fielded, cor-September 2001 using the Submillimetre Common rected for atmospheric extinction, de-spiked, base-User Bolometer Array (SCUBA) on the JCMT. The line subtracted, and processed with an advanced skyobservations were conducted simultaneously at 850 2 MOPSI (Mapping On-Off Pointing Skydip Infrared) is a software1 JCMT package for infrared, millimetre, and radio data reduction that has is operated by the JAC, Hawaii, on behalf of the UK STFC, been developed and is updated by R. Zylka.The Netherlands OSR, and the Canadian NRC. 3
  4. 4. noise algorithm. The double-beam maps were then negligible, but at 450 µm a significant error beam wasrestored into single-beam maps using the Emerson- found to contribute up to ∼50% of the flux density.Klein-Haslam algorithm (Emerson et al. 1979). The Note that the main beam flux calibration accounts forindividual maps were averaged with corresponding this (to first order). Flux density measurements in theweights and converted into the equatorial coordinate final maps were done using the GILDAS3 software andsystem. The SIMBA fastscan maps were reduced ac- are are described in Sect. 3.1.cording to the instructions in the SIMBA Observer’s For weak and extended sources, the integrated fluxHandbook (2002; see also Chini et al. 2003, Appendix density and the derived source size also depend onA). the map size, the curvature of the extended flux den- The SCUBA 450 µm and 850 µm jiggle maps were sity distribution, and the method of baseline subtrac-reduced with the SURF package (SCUBA User re- tion, since the sources are always observed against aduction Facility, Jenness & Lightfoot 1998). The in- more extended emission background originating fromdividual maps were flat-fielded, corrected for extinc- the parental molecular cloud.tion, processed with a sky-noise routine (Jenness et al.1998), and combined. The SCUBA scan maps were 2.3. Near-infrared observationsprocessed in a similar manner, before removing the We use the NIR images only for overlays with thedual-beam function. Finally, the maps from the six dif- (sub)mm maps, to identify which core is associatedferent chop configurations were combined in Fourier with a star or NIR nebula, and for discussion of evo-space, weighting each to minimize noise on spatial lutionary stages. NIR photometry is not be describedscales corresponding to the chop throw (Jenness et al. in this paper, but we use the fluxes for the SEDs where1998; Holland et al. 1999). relevant. To this end, NIR observations of the globules The data were calibrated using nightly observations listed in Table 2 were performed during five observ-of either primary planetary calibrators, or secondary ing runs between December 1993 and October 2003calibrators when a planet was unavailable. Main beam at the 3.5m telescope on Calar Alto, Spain and at theflux calibration factors for all maps were computed MPG/ESO 2.2-m telescope on La Silla, Chile. Observ-by integrating over the main antenna beam in the ing parameters for all runs are compiled in Table 4.planet maps, assuming planet brightness temperatures Images were obtained in standard dithering mode withlisted by Orton et al. (1986) and Griffin & Orton 4 to 5 positions per field. Flat field calibrations were(1993) and correcting for the error beams (see below). obtained each night via dome flats using differenceThe flux densities of the secondary calibrators were frames of integrations with a halogen lamp on and off.taken from Sandell (1994). The statistical scatter of Sets of photometric standard stars from lists by Eliasthe 1.3 mm flux calibration factors derived from all et al. (1982) and Hawarden et al. (2001) were observedIRAM 30-m and SEST data was found to be ≈6%. during each night, but the calibration for most sourcesDue to the uncertainty of the adopted planet temper- was done by using 2MASS catalogue fluxes for starsatures and uncertainties in the flux integration proce- in the fields.dure, we estimate that the total calibration uncertainty The data were reduced using a combination of theis ±15% at 1.3 mm, ±20% at 850 µm, and ±30% at IRAF4 “Experimental Deep Infrared Mosaicking Soft-450 µm. Typical 1σ noise levels (off-source) in the fi- ware” (XDIMSUM5 ) and Starlink Project packages.nal maps are 5.5 ± 1.3 mJy/beam (IRAM 30-m) and For photometric calibration, zero-points were derived20 ± 5 mJy/beam (SEST) at 1.3 mm, ∼20 mJy/beam at from observed standard stars of each night. Then,850 µm, and ∼200 mJy/beam at 450 µm, respectively. magnitudes for individual objects were derived using For each instrument and observing run, the aver-age beam size full-width at half maximum (FWHM) 3 http://www.iram.fr/IRAMFR/GILDASwas determined from the flux calibration maps, mostly 4 IRAF is distributed by the NOAO, which are operated by the Asso-on Uranus, taking into account the intrinsic size of the ciation of Universities for Research in Astronomy, Inc., under coop-planet disk (see Table 3). The secondary calibrators are erative agreement with the National Science Foundation. 5 XDIMSUM is a variant of the DIMSUM package (V3: August 19,less well characterized, and so on nights when plane- 2002) developed firstly by Eisenhardt et al. (1996)tary beam-maps were unavailable, the average beamFWHM for the entire data-set is quoted in Table 3.Error beams at 1.3 mm and 850 µm were found to be 4
  5. 5. GAIA6 with the embedded PHOTOM7 package. The optical images of the globules are also shown, together average 3 σ point source detection limit in all three with the submm contours (850 µm or 1.3 mm) and the NIR bands is ∼ 19 mag. Only sources that are associ- boundaries of the submm maps, to illustrate their over- ated with a dense globule core (i.e. a (sub)mm source) all morphology and the relative location of the dense are considered in this paper. cores within the globules. Note that in the case of For astrometric calibration of mosaicked frames we SCUBA jiggle maps the entire mapped region is indi- used the Guide Star Catalogue II8 and the SIMBAD9 . cated, while in the case of IRAM (dual-beam) or SEST The resulting typical positional accuracy is better 0. 8 scan maps only the area from which the mm emission (1 σ rms). was restored is marked. In addition, SEDs, includ- ing fluxes from optical wavelengths up to 3 mm, are 2.4. Spitzer data shown for those sources were we could compile suf- ficient data and associate them unambiguously to one All sources listed in Table 2 were observed by source. Spitzer within different programs, most with IRAC (Sub)mm flux densities are summarized in Table 5. and MIPS, except CB 17 SW, BHR 58, BHR 137, and The center coordinates for each dust core were derived CB 224. The relevant data were obtained from the from gaussian fits to the 1.3 mm intensity maps af- Spitzer Science Center (SCC) Archive. In particular, ter decomposition into extended envelope and individ- we retrieved, where available, IRAC maps at 3.6, 4.5, ual compact components. For some sources (marked 5.8, and 8 µm and MIPS scan maps at 24, 70, and in Table 5), the peak position was derived from the 160 µm. The data were processed by the SCC using peak 850 µm maps. The peak intensity Iν of each com- their standard pipeline to produce Post-Basic Cali- pact component is given in Columns 3, 5 and 7 of Ta- bration Data (P-BCD) images. Photometry was done ble 5 in units of Jy/beam. The beam size is also quoted in IDL by applying aperture photometry, where the here in parentheses. Note that the maps of some ex- aperture corrections available at the instruments web- tended sources (e.g., CB 246) were convolved with a page were used. The conversion from P-BCD units larger gaussian beam to increase the SNR. We also (MJy sr−1 ) to flux density was achieved by applying tot quote the total integrated flux density, S ν , for each Quick Point Source Photometry as described online. source. This was derived by integrating within a poly- The overall flux calibration and flux measurement ac- gon enclosing the (closed) 2 σ contour. If there is one curacy is estimated by the SSC to be accurate to within or more compact sources within a larger envelope, we 20%. also give the integrated flux densities of the compact sources, derived after fitting and subtracting the ex- 3. Results tended envelope from the maps. We do not list indi- 3.1. (Sub)millimetre continuum maps vidual error bars for the flux densities, because uncer- tainties are mostly dominated by the calibration uncer- In Figs. 1 through 24, we present the (sub)mm tainties (see Sect. 2.2). However, for extended sources, maps, together with NIR K-band images for each of the uncertainty is dominated by the size of the area the sources listed in Table 2 (sources with only upper from which emission could be recovered above the 2- limits are not shown). Digital Sky-Survey (DSS2)10 σ-level, which in turn depends on the curvature of the emission background, the ratios of the angular sizes of6 GAIA is provided by Starlink, and is a derivative of the Skycat cata- the beam, the source, and the chop throw, as well as on logue and image display tool, developed as part of the VLT project at ESO. Skycat is free software under the terms of the GNU copyright. the SNR in the map. In general, the relative fraction of 7 The PHOTOM photometry package is provided by Starlink recovered to total actual flux increases from faint ex- 8 The Guide Star Catalogue II is a joint project of the STSI and the tended to strong compact, and from nearby to distant Osservatorio Astronomico di Torino. STSI is operated by the As- sources. sociation of Universities for Research in Astronomy, for the NASA under contract NAS5-26555. 3.2. Association of sub(mm) and infrared sources 9 SIMBAD is an online database, operated at CDS, Strasbourg, France10 The Digitized Sky Survey was produced at the Space Telescope Sci- When compiling SEDs and attempting to model ence Institute under US Government grant NAG W-2166. The im- and interpret the observational results, it is important ages of these surveys are based on the photographic data obtained using the Oschin Schmidt Telescope on Palomar Mountain and the UK Schmidt Telescope. 5
  6. 6. to reveal which of the sources and flux densities de- main mm peak.tected at different wavelengths, with different angularresolution, apertures, and astrometric precision, are ac- 3.3. Multiplicitytually physically associated and which ones have dif- Although the (sub)mm maps are not sensitive toferent origins. For this reason, we overlay the (sub)mm spatial scales that cover the peak of the binary starmaps on the optical (DSS) and NIR images, show the separation distribution (20 – 100 AU; Duquennoy &IRAS point source position error ellipses where appli- Mayor 1991), we find that at least 16 out of 24 glob-cable, and also inspect the Spitzer maps where avail- ules with detected (sub)mm emission show signaturesable (see 2.4). of multiplicity at scales larger than ≈1,000 AU, out We found the following associations between to 80,000 AU (0.4 pc; see Table 6). This includesNIR/MIR sources and mm cores: No NIR sources prestellar and protostellar double cores resolved in thebrighter than 0.02 mJy (≈19 mag) at 2.2 µm or ≈1 mJy mm dust continuum (e.g., CB 17, CB 130, CB 246,at 4.5 µm are detected in CB 17-SMM1 and 2, CB 130- BHR 12), unresolved protostellar cores where NIR orSMM2, CB 232-SMM, CB 243-SMM1 and 2, and MIR observations have revealed a secondary embed-CB 246. However, since local dust extinction, source ded source (e.g., CB 230, BHR 71), prestellar or pro-morphology, and projection effects (e.g., the alignment tostellar cores where NIR/MIR observations have re-between the line of sight and outflow cavities) play a vealed sources that are significantly offset from themajor role at these early evolutionary stages, this de- (sub)mm peak and that are most likley not embeddedtection limit cannot be directly converted into an upper in these cores (e.g., CB 17, CB 130, CB 232, CB 243;limit for the mass or luminosity of possible embedded see Fig. 28), as well as large dust cores that are associ-sources. No K-band emission, but faint NIR sources ated with clusters of NIR sources, presumably youngwith steeply rising NIR SEDs peaking at ≈ 4 . . . 6 µm stars that formed from the same globule (e.g., CB34,have been found in BHR 55, BHR 86, CB 68, and CB 54, CB 205). Nine of these 24 globules were re-CB 199. Extended NIR nebulosities, in several cases solved to have two or more cores in the (sub)mmNIR jets, but no star-like objects were detected in dust emission maps (see Table 6). However, onlyCB 26, BHR 12-SMM1 and SMM2, BHR 36, BHR 71, three of these double cores have projected separationsCB 130-SMM1, CB 224-SMM1, CB 230, and CB 244- <10,000 AU (CB 17, BHR 12, and CB 130). Separa-SMM1. A single red star or compact star-like NIR tions >10,000 AU are probably not relevant for the for-source, in some cases with extended nebulosity, was mation of bound stellar systems.detected at the mm peak position in CB 6, CB 188, and Altogether, we find that at least two thirds (16 outCB 224-SMM2. Several NIR sources or clusters, re- of 24) of the star-forming globules studied here showspectively, are associated with the (sub)mm cores in evidence of forming multiple stars, either in multipleCB 34, CB 54, CB 205, and CB 240. The situation is star-forming cores, wide embedded binaries, or smallsomewhat inconclusive in CB 58 and BHR 137, where star clusters. The fraction of closer binaries formedit is not clear if the NIR sources are actually associated from unresolved (sub)mm cores might be higher, butwith the mm core or not. remains unknown from this study (cf. Maury et al. The following (sub)mm cores have very nearby, but 2010).most likely not directly associated NIR/MIR sourcesand the flux measurements could not always be dis- 3.4. Spectral Energy Distributionsentangled: CB 17 has an IRAS point source and aSpitzer-detected low-luminosity IR point source that Spectral Energy Distributions (SEDs) were com-is clearly not associated with the mm core (Fig. 2). piled from aperture photometry in the (sub)mm mapsCB 130-SMM1 is a strong and slightly extended (Table 5) and NIR images (Figures 1–24) presentedsubmm source with a small NIR nebula at the mm peak in this work (Sects. 3.1 and 3.2), the 2MASS cat-position. A very nearby (14 ) strong and pointlike alogue, the IRAS point source catalogue, pipeline-IR source (IRS) could not be completely disentagled calibrated Spitzer IRAC and MIPS maps, and otherfrom SMM1, but does not seem to exhibit detectable publicly available databases and literature papers. Wemm emission on its own. CB 232 and CB 243 both do not explicitly list all flux values which we havehave bright reddened stars within the area of the mm compiled, but show in Figs. 1 through 23 the SEDs foremission, but clearly offset and not associated with the those sources where we are confident that the measure- 6
  7. 7. ments represent one source or, in some cases, physi- heating sources or externally heated very small grainscally related, mostly unresolved double sources (e.g., in the envelopes. Thus, they usually do not repre-CB 230). Note that we actually show the energy den- sent the mass-averaged dust temperature. Instead, wesity (ν S ν ) and not the flux density (S ν ). We do not use fixed mass-averaged dust temperatures as functionshow SEDs for sources where we either do not have of evolutionary stage (see Sect. 3.6) and derive sourceenough data points (e.g., BHR 137, CB 246) or where sizes directly from the (sub)mm maps.the (sub)mm emission is associated with multiple NIRsources (e.g., CB 34, CB 54, CB 58, CB 205, CB 240). 3.5. Evolutionary stagesFor the other sources, we indicate in the figure captions We use a combination of different diagnostics toand in Table 7 whether we use the total (sub)mm fluxes derive the evolutionary stages of the globule cores.or only one compact component. Non-detections are We do not attempt to classify those cores of dis-indicated by down-arrows as 3 σ upper limits. Flux tant globules that are associated with clusters of NIRvalues that derive from apertures smaller than the ex- sources and for which the NIR, FIR (IRAS), andpected emission region (e.g., single-point measure- (sub)mm emission could not be uniquely associatedments with the JCMT-UKT14 bolometer) or interfer- (e.g., CB 34, CB 54, CB 58, CB 205, CB 240). Sourcesometric measurements, which are not sensitive for ex- for which we have barely more information than a mmtended emission, are indicated by up-arrows as lower flux (e.g., BHR 58, BHR 137) are also not consideredlimits. A comprehensive overview of all SEDs, or- here. For the other sources, we analyze the followingdered by evolutionary stage (see Sect. 3.5) is shown in tracers of evolutionary stage:Fig. 25. In order to derive luminosities and bolometric tem- 1. The ratio of submm to bolometric luminos-peratures, we first interpolated and then integrated the ity, Lsmm / Lbol , where Lsmm is measured long-SEDs, always assuming spherical symmetry. In the ward of 350 µm, which roughly reflects the ratiocase of, e.g., bipolar NIR reflection nebulae, this may of envelope to stellar mass, Menv / M∗ (Andr´ elead to either an under-estimate of the NIR luminos- et al. 1993). Andr´ et al. (2000) propose to use eity (e.g., CB 26, Fig. 3) or an overestimate (e.g., Lsmm / Lbol > 0.5% as a threshold for Class 0CB 230, Fig. 19), depending on the inclination with protostars, reflecting Menv / M∗ > 1.respect to the line-of-sight. However, all sources stud-ied in this paper emit the majority of their total en- 2. The bolometric temperature, T bol (Myers &ergy at much longer wavelengths, where the radiation Ladd 1993), which is defined by the tempera-is reprocessed by dust, so that the general results de- ture of a blackbody having the same mean fre-rived here are not significantly affected by the sim- quency (ratio of the first and the zeroth momentplifying assumption of spherical symmetry. Interpo- of the SED) as the observed continuum spec-lation between the measured flux densities was done trum. Chen et al. (1995) and Myers et al. (1998)by a χ2 grey-body fit to all points at λ ≥ 100 µm and derive the boundary between Class 0 and Class Iby simple logarithmic interpolation between all points YSOs to be at T bol ∼ 70 K. Note that we use theat λ ≤ 100 µm. Free parameters in the grey-body interpolated SEDs (see Sect. 3.4) with flux val-fits were the dust temperature, source size, and op- ues at all wavelengths to compute T bol and lumi-tical depth. The (sub)mm spectral index of the dust nosities, rather than just the actual data points.emissivity was fixed to β = 1.8 (see Sect. 3.6). Fit- The latter approach could lead to significant bi-ted dust temperatures were in the range 15 K (CB 17- ases if the SED is not well-sampled (cf. EnochSMM) to 53 K (CB 224-SMM2), with mean values of et al. 2009).33 K for Class 0 sources and 42 K for Class I sources 3. The IR spectral index, αIR = −d log(ν Fν ) / d log ν(see Sect. 3.5). (Lada 1987), which characterizes the IR excess Note, however, that we use these fits only to de- with respect to stellar blackbodies. Where suf-rive an approximate interpolation for integrating the ficient data are available, we calculate αIR be-SEDs. In particular, we do not use the dust temper- tween λ = 2.2 and 24 µm, thus avoiding to car-atures or source size from these grey-body fits because acterize only the NIR/MIR nebulosity “bump”the derived effective temperatures are usually dom- in the Class 0 SEDs (see below), the possibleinated by small contributions from either embedded 7
  8. 8. influence of the silicon absorption, and the low- ilar properties of the SEDs, appears to be a more resolution IRAS measurements at 12 and 25 µm, reliable indicator of the evolutionary stage than the which are prone to confusion problems. Quanti- Lsmm / Lbol ratio, as it is less sensitive to the amount of tatively, we follow Andr´ & Montmerle (1994) e extended envelope emission recovered in the submm and use αIR = 0 as boundary between Class I maps (see Sect. 4.1). We therefore use the bolometric and II. No well-defined boundary exists for temperature as the main discriminator between Class 0 Class 0 sources since only Spitzer lead to the protostars and Class I YSOs, along with the other sup- detection of the NIR/MIR emission in many of porting indicators mentioned above (cf. Enoch et al. them. 2009). The derived evolutionary stages, along with other Along with these three quantitative classification source properties, for 18 globules are summarized incriteria, we evaluate the morphology of the NIR emis- Table 7. Figure 26 shows the Lsmm / Lbol ratio vs. T bolsion and the presence of molecular outflows. The first for these sources. In total, we identified nine star-one can be either direct star light, indicating that the less cores, presumably prestellar, in six globules, ninecircumstellar emission originates either from an op- Class 0 protostars in nine globules, and twelve Class Itically thin envelope or a phase-on disk, or diffuse YSOs in eleven globules. Interestingly, seven glob-nebulosity, indicating, e.g., scattered light from out- ules contain co-existing sources at different evolution-flow cavities in an otherwise optically thick envelope. ary stages (see discussion in Sect. 4.2). Six globulesThe presence of a molecular outflow is an indicator were not detected at (sub)mm wavelengths, and for theof an embedded source that distinguishes protostars submm cores in eight globules we could not or did notfrom prestellar cores. Two notes of caution are: (i) the attempt to derive an evolutionary stage because we ei-non-detection of an outflow does not prove the absence ther had insufficient data, or the available data indi-of an embedded protostar, in particular as we have no cated that several sources with possibly different evo-homogeneous high-sensitivity, high-angular resolution lutionary stages contribute to the partially unresolvedoutflow survey at hand, and (ii) the driving source of an flux measurements.outflow is not necessarily identical with the (sub)mmsource - this must be verified in each individual case. 3.6. Physical parameters of the star-forming cores The applicabilty and drawbacks of the classificationcriteria mentioned above has been discussed in detail The following physical source quantities are de-by Enoch et al. (2009). An analysis of our data con- rived from the (sub)mm continuum maps and thefirms the finding of these authors in that αIR does not SEDs: bolometric and submm luminosities (Sect. 3.5),correlate well with T bol and embeddedness for Class 0 source sizes, masses, beam-averaged peak columnand embedded Class I sources. The deep NIR images, densities, and source-averaged volume densities. √in particular those from Spitzer, show that most of even FWHM mean source sizes ( a b) are simply de-the youngest Class 0 sources already have a small NIR rived by measuring the angular diameters of the 50%nebulosity associated. This can also be clearly seen contours in the 850 µm or 1.3 mm maps (after decom-in the compilation of SEDs shown in Fig. 25. We at- position into envelope and compact core, where appli-tribute this to the fact that the amount of NIR/MIR flux cable), deconvolving them with beam size, and apply-seen from such an embedded object largely depends on ing the distances listed in Table 2. The results are listedthe alignment between the line of sight and the outflow in Table 7.cavity, allowing or not a direct view onto the hot dust To calculate masses, densities, and column densi-in the inner protostellar core and/or forward-scattered ties, we adopt mass-averaged mean dust temperatureslight in the outflow cavity. Therefore, we only list the that depend on the evolutionary stage of the sourceαIR values in Table 7, but do not use them for the clas- (Sect. 3.5). Quantitatively, we follow Motte & Andr´ esification of evolutionary stages. (2001), Shirley et al. (2002), and Young et al. (2003), We find that all sources, for which we could com- who have derived mean mass-averaged isothermal dustpile meaningful SEDs, have Lsmm / Lbol ≥ 0.8% (Ta- temperatures from radiative transfer models for sam-ble 7), i.e., according to the criterion proposed by ples of starless cores, Class 0 protostars, and envelopesAndr´ et al. (2000) they would all be Class 0 sources. e of Class I YSOs. We use here T d = 10 K for star-The bolometric temperature, although reflecting sim- less (prestellar) cores, T d = 15 K for Class 0 pro- 8
  9. 9. tostars, and T d = 20 K for the envelopes of Class I of hydrogen atoms, nH = n(H) + 2 n(H2 ), was calcu-YSOs. The first dust temperature map directly derived lated byfrom new Herschel data of the Bok globule CB 244 MH nH = , (3)with its two embedded cores confirms the validity of mH Vthese temperature assumptions (Stutz et al. 2010). The with V ∼ π/6 dS being the volume of the source, and 3adopted temperatures and resulting masses for each dS the FWHM size listed in Table 7. The resultingsource are listed in Table 7. densities are listed in Table 7. Note that these are Assuming optically thin isothermal dust emission, source-averaged densities and the local peak densitiesthe hydrogen gas mass MH = M(H) + 2 M(H2 ) is cal- in sources with density gradients can be orders of mag-culated from the (sub)mm flux density by nitude higher. In an earlier paper, we had also derived radial den- S ν D2 MH MH = . (1) sity profiles for most of the globule cores by fitting ray- κd (ν) Bν (ν, T d ) Md traced and beam-convolved model cores to the radiallyHere, ν is the observing frequency, S ν the observed averaged (sub)mm data at all three wavelegths (Laun-flux density from the region of interest, D is the hardt 2005). The extended dataset presented in this pa-distance to the source, κd (ν) is the dust opacity, per would certainly permit to derive better-constrainedBν (ν, T d ) the Planck function, T d the dust tempera- density profiles for many globules cores. However, asture, and MH /Md the hydrogen-to-dust mass ratio. in this previous paper, such an analysis would be ham-Throughout this paper, we use a hydrogen-to-dust pered by the fact that we have to assume temperaturemass ratio of MH /Md = 110 and a dust opacity of profiles, since the (sub)mm data provide only weakκd (λ) = κd (1.3mm) × (λ/1.3mm)−1.8 with κd (1.3mm) = constraints. In particular for the prestellar cores, which0.5 cm2 g−1 (Ossenkopf & Henning 1994, uncoagu- do not have a central heating source and may have alated MRN dust grains with thin ice mantles), a that positive radial temperature gradient, this can introduceis value typical for dense and cold molecular cloud significant biases and may lead to an underestimationcores and intermediate between the opacities usually of the density gradient. Since we expect to obtain soonadopted for unprocessed interstellar grains in the dif- Herschel data for many of the sources presented here,fuse gas (e.g., Draine & Lee 1984) and for processed which will provide direct constraints on the tempera-and coagulated grains in protoplanetary disks (e.g., ture profiles, we postpone the derivation and analysisBeckwith et al. 2000). of the radial density profiles to a succeeding paper. We derive total Hydrogen masses of 1 – 20 M for 4. Discussionprestellar cores, 0.5 – 10 M for Class 0 protostars, and0.11 – 1 M for Class I YSOs (not counting the ex- 4.1. Source properties as function of evolutionarytended envelopes; see Table 7). The total gas mass, stageaccounting for helium and heavy elements, amountsthen to Mg ≈ 1.36 MH . Being relatively nearby and isolated, the cores of Bok globules are the best-suited places to observe the The beam-averaged peak Hydrogen column density properties and evolution of individual low-mass star-NH = N(H) + 2 N(H2 ) was derived from the observed forming cores. However, this isolation, which makespeak flux densities by them such good targets, may also alter the core proper- Iν 1 MH ties and evolutionary process as compared to the more NH = , (2) representative (for general star formation), but harder κd (ν) ΩB Bν (ν, T d ) mH Md to observe, star-forming cores in clustered regions.where Iν is the peak flux density (Table 5), ΩB = Some of the most obvious differences in environmen-1.133 θB the beam solid angle, θB the beam FWHM 2 tal conditions are the interstellar radiation field, whichsize, and mH is the the proton mass. The resulting is not shielded by the surrounding cloud, the miss-beam-averaged peak column densities are also listed ing external pressure due to the lack of inter-core gas,in Table 7. and the lack of turbulence input from the surrounding Assuming that the sources have the same extent cloud.along the line of sight as the mean size perpendicular The most obvious changes in core properties duringto the line of sight, the source-averaged volume density the evolution from prestellar cores, through collapsing 9
  10. 10. protostars, towards Class I YSOs are expected to be sources as “Class I” in Table 7, despite the fact thatseen in the core size, central core density, ratio of core they have αIR < 0.mass to envelope mass, and contribution of the central To verify these relations between T bol , Lsmm /Lbol ,(proto)star to the source luminosity at shorter wave- and evolutionary stages, we have compared our sam-lengths. The latter two effects have been used in the ple of globule cores to the sample of Class 0 proto-form of parameters Lsmm / Lbol and T bol as main cri- stars from Andr´ et al. (2000) . Plotting these sources eteria to identify the evolutionary stage of the sources into the same diagramme (not shown here) shows a(Sect. 3.5). different picture. While none of the globules sources In Figure 26, we show a plot of Lsmm /Lbol vs. T bol have both T bol < 70 K and Lsmm /Lbol < 2%, numer-for the sources compiled in Table 7. We find that all ous protostars from the Andr´ sample do populate this esources with T bol < 70 K, the empirical boundary be- region. The two different source samples have similartween Class 0 protostars and Class I YSOs proposed distance distributions, and in each case, both nearbyby Myers et al. (1998), have Lsmm / Lbol > 2%. Within and more distant sources are distributed evenly aroundthe Class 0 phase, there seems to be no significant cor- the Lsmm /Lbol vs. T bol plot. Therefore there is no ev-relation between the two parameters and no obvious idence to suggest that limited resolution or sensitivityevolution from lower to higher values of T bol . This for either of the samples is causing the discrepancy.may reflect the fact that during this early phase, the Similarly, there is no evidence that any particular as-optically thick envelope dominates over the still very pect of the morphology of a Bok globule is causinglow-mass and low-luminosity central protostar. How- this discrepancy between the two samples. As noneever, there is a relatively large spread in Lsmm / Lbol ra- of the above characteristics can possibly cause a sys-tios, probably reflecting the fact that during this phase tematic shift of the Bok globules in the Lsmm /Lbol vs.the mass of the infalling envelope is relatively quickly T bol diagramme, we conclude that the absence of glob-reduced. ule cores with both low T bol and low Lsmm /Lbol -ratios The more evolved (Class I) YSOs in our sam- is a real effect, caused by the systematic environmentalple have bolometric temperatures between 100 K and differences between the cores of Bok globules and pro-350 K and Lsmm /Lbol ∼ 0.8 . . . 3.5% (see Fig. 26). All tostars located in larger molecular clouds mentioned atsources with NIR nebulosities but no visible stars the beginning of this section.have Lsmm /Lbol ≥ 1.6% and extended (resolved) The outer boundary of a protostellar core that is em-(sub)mm emission. These sources are probably less- bedded in a larger molecular cloud is generally judgedevolved, envelope-dominated Class I YSOs. The three to be where it’s envelope merges with the surroundingsources with Lsmm /Lbol ≤ 1.3% (CB 188, CB 224- cloud material, or the envelope of an adjacent object.SMM2, CB232-IRS1) all have visible stars and com- In contrast, the column density profile of a globule corepact (sub)mm emission arising presumably from cir- continuously decreases towards the ”edge” of the smallcumstellar disks. These sources probably represent cloud, making it more difficult to distinguish betweenthe more evolved, disk-dominated Class I YSOs. We the core and the surrounding cloud, both physicallytake these findings as tentative indication that the and observationally, as the envelope (cloud) emissionLsmm / Lbol ratio, as indicator of envelope dispersal, is usually not chopped away in the small globules. Thismay better trace the evolution within the Class 0 and has the effect of making the detected envelopes largerClass I phases than T bol . and elevating the submm luminosity as compared to Two of the Class I sources in our sample have dense cores embedded in larger clouds. It is thereforedouble-peaked SEDs with NIR nebulosities and αIR < possible that the cutoff in Lsmm /Lbol , used to differ-0 (CB 26, CB 230), which would indicate Class II. entiate between Class 0 protostars and Class I YSOsHowever, the IR emission seems completely originat- should be raised from 0.5% in the case of protostarsing from scattered light and the IR fluxes and morphol- in larger molecular clouds, to ≈ 2% in the case of iso-ogy seem dictated by projection effects. For CB 26, lated Bok globules, to account for their larger enve-a massive edge-on protoplanetary disk with a rem- lope of cold dust. The value of 2% is chosen as thisnant envelope, Launhardt & Sargent (2001) derive an is the lowest value of Lsmm /Lbol in the observed sam-age of ≈ 1 Myr. CB 230, with a more massive enve- ple of globule cores with T bol < 70 K. Although welope and less-developed disk(s) (Launhardt 2001), is cannot present a modified quantitative physical modelmost likely younger. We therefore classify these two (cf. Andr´ et al. 1993), we think that this different cut- e 10
  11. 11. off value for Lsmm /Lbol still represents the point where sequential star formation. This includes neighbour-the mass of the infalling protostellar envelope is equal ing mm sources with obviously different evolutionaryto that of the accreting protostar, but it accounts for stages (e.g., BHR 12, CB 130, CB 224), prestellar orthe contribution from the extended envelope which can protostellar cores with nearby IR sources, presumablyno longer be clearly distinguished from the protostellar more evolved protostars or Class I YSOs (e.g., CB 17,core. CB 130, CB 232, CB 243), as well as NIR star clusters We conclude that T bol appears to be the most re- next to large (sub)mm cores with the potential to formliable parameter to discriminate between Class 0 pro- more stars (e.g., CB 34, CB 54, BHR 137) (see Ta-tostars and Class I YSOs, with no hint for a system- ble 7). In only three globules we find (approximately)atic difference between sources in isolated globules coeval pairs, ranging from multiple pre-stellar cores inand those embedded in larger molecular clouds. On CB 246, embedded Class 0 protostars in BHR 71, to athe other hand, the evolution during the Class 0 phase pair of embedded Class I YSOs in CB 230.and during the Class I phase may probably be better The most interesting sources in this respect aretraced by the Lsmm / Lbol ratio. Despite the fact that the shown in Fig. 28. CB 17 contains two prestellar corestwo parameters are not completely independent, this and one low-luminosity Class I YSO within 5 000 AUcan be understood as a consequence of T bol being more (projected separation). CB 130 contains three sourcessensitive to a flux contribution at shorter (NIR) wave- within 9 000 AU, all with different evolutionary stages.length from a central heating source, while Lsmm / Lbol A Class 0 protostellar core is flanked by a prestellaris more sensitive to the size and mass of the cold enve- core at 6,000 AU to the west, and by a Class I or IIlope. YSO at 3,000 AU to the east. CB 232 and CB 243 both We have also determined the mean FWHM core contain a prestellar core with a Class I YSO withinsizes and the mean source-averaged volume densities 7 000 AU.(see Table 7) of the three evolutionary subgroups. We This widespread non-coevality suggests that thefind that the mean core sizes decrease from ≈8,000 AU multiple sources observed here have in general notfor prestellar cores, to ≈3,000 AU for Class 0 proto- formed by gravitational fragmentation of collapsingstars, to <2,000 AU for Class I YSOs (Fig. 27). We protostellar cores (e.g., Bonnell & Bate 2006), butalso find that the Class I YSOs usually have extended rather by initial turbulent fragmentation prior to theenvelopes with a mean diameter of 20,000 AU. The collapse phase (e.g., Padoan & Nordlund 2002), fol-source sizes of the Class 0 protostars might be some- lowed by a relatively independent evolution of the in-what underestimated since we did not decompose the dividual fragments. Whether or not the first formingsources into central point source (the warm protostar) star can trigger collapse of the other fragment(s) andand envelope before measuring the FWHM size. The thus lead to small-scale sequential star formation can-source-averaged volume densities, nH , increase from not be directly answered by this study. The lack of1 × 107 cm−3 for prestellar cores, to 7 × 107 cm−3 for close coeval pairs can be understood as an observa-Class 0 protostars. At these high densities, one can tional bias as our study is in general not sensitve toin general assume that gas and dust temperatures are size scales <1000 AU, which would be more relevantcoupled. For the cores of the Class I YSOs, which are to the first mechanism. Consequently, we cannot drawmostly unresolved and presumably contain significant direct conclusions on the frequency of bound binaryflux contrinutions from circumstellar disks, we derive stars formed in isolated Bok globues (cf. Maury et al.a lower limit to the mean density of > 8 × 107 cm−3 . 2010).The extended remnant envelopes of the Class I YSOs These findings also call for special attention whenhave a mean density of 7 × 104 cm−3 . compiling SEDs and attempting to derive source prop- erties from flux measurements with insufficient angu-4.2. Multiple and sequential star formation in lar resolution. One may easily end up classifying the globules combined SED of a prestellar core and a nearby, more One interesting aspect of the combined results for evolved YSO as a Class 0 protostar (see, e.g., CB243,multiplicity and evolutionary stages is that the large Fig. 22f).majority of the globules with signs of multiple starformation shows evidence of non-coeval, possibly 11
  12. 12. 5. Description of individual globules controversial: while Clemens & Barvainis (1988) de- scribed it as Class III YSO, Launhardt et al. (1997) de- In this section, we describe in somewhat more detail rived a Lbol / Lsmm ratio indicative of a Class 0 protostarthe individual globules, their morphology and physical (Andr´ et al. 1993). Our (sub)mm continuum maps eproperties, and discuss the evolutionary stages of the (Fig 1) show that the thermal dust emission is domi-embedded sources. Distance references are not given nated by a compact source centered at the origin of thein the text, but are compiled in Table 2. Globule sizes NIR nebula, indicating that the bulk of the mm dustand morphology are summarized in Table 6 and prop- emission does not originate from an optically thick en-erties and evolutionary stages of the embedded sources velope, but rather from a disk. The disk and the moreare summarized in Table 7. extended envelope probably obscure the southern out- CB 4 is a small, roundish, and relatively isolated flow lobe which is presumably pointed away from us.dark globule at a somewhat uncertain distance. Al- This morphology, combined with the bolometric tem-though Wood et al. (1994) assumed a distance of perature of T bol ∼ 180 K let us conclude that this is a200 pc to the Cas A dark clouds in this region of the Class I YSO.sky, we follow the analysis of Dickman & Clemens CB 6 N is a small globule located ≈10 north of(1983) who derived a distance of 600 pc. This latter CB 6. It has no IR source associated and we did notvalue is also close to the distance of the “-12 km s−1 ” detect it at mm wavelengths, indicating that it does notclouds (800 pc; Dame et al. 1987) with which Laun- contain a star-forming core.hardt & Henning (1997) associated the globules CB 4 CB 17 (L 1389; Fig. 2) is a small and slightlyand CB 6. CB 4 is associated with an IRAS point cometary-shaped globule, located near Perseus andsource, located at its southern rim, which is detected associated with the Lindblad ring (Lindblad et al.at 100 µm only. There is no further evidence of an 1973). It has a long diffuse tail, pointing north-eastembedded source, including the Spitzer maps up to and visible as cloudshine reflection. The distance of160 µm. Deep IRAS aperture photometry and molecu- CB 17 was derived by Launhardt & Henning (1997)lar line studies showed that the globule is cold (∼ 7 K) via association in projected space and radial veloc-and has a relatively low mean density (Clemens et al. ity with other Lindblad Ring clouds. In the direc-1991; Kane et al. 1994; Turner et al. 1997). Despite its tion of CB 17, the Lindblad Ring structures have acompact and dark appearance it thus rather resembles mean distance from the Sun of ≈ 300 pc (Dame et al.a cirrus cloud. CCD polarimetry of background stars 1987). HD 25347, a bright G5 III star at a distancerevealed a faint and well-ordered magnetic field with of 210 ± 40 pc (van Leeuwen 2007) is located abouta position angle (P.A.) of 71◦ (Sen et al. 2005), which ≈11 (0.65 pc at 200 pc) south of CB 17 and could beis apparently coupled to the surrounding Galactic field responsible for the cometary shape and diffuse cloud-(Kane et al. 1995). Our non-detection of compact ther- shine from the rim and tail of CB 17 if it is locatedmal dust emission confirms the assumption that this at the same distance as the globule. Combining theglobule does not seem to form stars. possible associations of CB 17 with both the Lindblad CB 6 (LBN 613; Fig. 1) is a small, cometary-shaped Ring and HD 25347, and the distances and uncertain-globule with a long diffuse tail, visible as scattered ties involved, we adopt a distance of 250±50 pc forlight cloudshine. It is located only a few degrees from CB 17.CB 4 and has the same radial velocity. We therefore The roundish cloud core of CB 17 is associatedassign CB 6 the same distance (600 pc) as CB 4. At its with a dense submm core and a faint and cold IRAScentre, the globule harbours an embedded YSO with point source (IRAS 04005+5647) that is detected onlycircumstellar dust emission and a conical reflection at 60 and 100 µm. CB 17 has been studied extensivelynebula (RNO 3; Cohen 1980; Eiroa et al. 1994), likely by various authors using different mm molecular linerepresenting scattered light in an outflow cone from the transitions. The core was found to have a mean kineticembedded YSO. Indeed. broad line wings in CO (2-1) gas temperature of T kin ∼ 10 K and a virial mass of(Clemens et al. 1991) and CS (2-1) (Launhardt et al. ≈3 M (e.g., Lemme et al. 1996; Caselli et al. 2002,1998) indicate the presence of a molecular outflow scaled to D = 250 pc). Pavlyuchenkov et al. (2006)from the YSO, but no outflow maps exist yet. Scap- compared a number of different molecular line mapspini & Codella (1996) searched for, but did not detect of CB 17 with chemodynamical models. They deriveNH3 and C2 S. The evolutionary stage of the YSO is a chemical age of ≈2 Myr, as well as the relative rota- 12
  13. 13. tional, thermal and turbulent energies, concluding that of similar size as CB 17, located ≈12 south-west ofthe core will most likely fragment. CB 17. It has no IR source associated and we did not Indeed, our deep 1.3 mm continuum map (Fig. 2) detect it at mm wavelengths, indicating that it does notshows a double core with 14 separation and a com- contain a star-forming core.mon envelope, located at the south-western rim of the CB 26 (L 1439: Fig. 3) is a small, cometary-shaped,globule head. The low-SNR 850 µm SCUBA map double-core Bok globule located ∼ 10◦ north of thecan, however, not confirm the double-core nature, Taurus-Auriga dark cloud, at a distance of ∼ 140 pc.and the source remains undetected at 450 µm. The An IRAS point source and dense core with signa-Spitzer IRAC maps show a 8 µm shadow that agrees tures of star formation is located at the south-west tipin position and shape with the double core (Fig. 28). of the western globule core (Launhardt & HenningNo emission is detected at wavelengths shorter than 1997). While the eastern globule core (at 4:56:20,160 µm, indicating these cores have no, or only very 52:01:30, B1950, in Fig. 3) is visible as 8 µm shadowlow-luminosity internal heating sources. Since the in the Spitzer IRAC maps, the IR source is locateddouble core is resolved only in the 1.3 mm map, we ≈4 further west. The eastern globule core was notcompile the SED only for the combined core, which searched for mm emission, but the Spitzer maps alsowe call SMM. However, a faint point-like NIR/MIR do not indicate the presence of an embedded source.source (IRS), that is clearly not directly detected in OVRO observations at the IRAS position of the mmthe (sub)mm continuum, is located ≈10 northwest dust continuum emission and of the 13 CO (1-0) lineof SMM and just outside the 8 µm shadow, indi- have revealed a nearly edge-on circumstellar disk ofcates the presence of a third, more evolved source radius 200 AU with Keplerian rotation (Launhardt &in CB 17 (Fig. 28). This source dominates the emis- Sargent 2001), surrounding a very young (obscured)sion at wavelengths shorter 160 µm, while SMM dom- low-mass T Tauri star. It is associated with a smallinates at longer wavelengths. Although the MIPS3 bipolar near-infrared (NIR) nebula bisected by a darkmap at 160 µm does not resolve the two positions, extinction lane at the position and orientation of thea slight shift of the peak position from IRS towards edge-on disk (Stecklum et al. 2004). The star andSMM indicates detectable emission from SMM at this disk are surrounded by an optically thin asymmetricwavelength. For compiling the SEDs, we assigned envelope with a well-ordered magnetic field directed15% of the total 160 µm flux to SMM, but our re- along P.A. ∼ 25◦ (Henning et al. 2001). Furthermore,sults depend only weakly on the adopted flux splitting. an Herbig-Haro object (HH 494) was identified by HαThe morphology and SED of SMM (Tbol ≈ 15 K, and S[II] narrow-band imaging, 6.15 arcmin northwestLsmm / Lbol > 30%) suggest this is a prestellar (dou- of CB 26 at P.A. = 145◦ (Stecklum et al. 2004). Theble) core (cf. Caselli et al. 2002; Pavlyuchenkov HH object is thus perfectly aligned with the symmetryet al. 2006). Source IRS, which is located at a pro- axis of the disk and the bipolar nebula. Recently, Laun-jected separation of ∼ 2000 AU from SMM2 within hardt et al. (2009) reported the discovery of a smallthe same globule core, is tentatively classified as a collimated molecular outflow along the same direc-low-luminosity Class I YSO protostar (Tbol > 55 K, tion, that appears to rotate about its polar axis. It isLsmm / Lbol < 2.6%), although an outflow remains to speculated that this rotation is related to the possiblebe detected (Chen et al., in prep.). A faint and diffuse binarity of the central star. We are currently analyzingNIR nebula at the position of SMM2 is speculated to higher resolution interferometric CO maps, attemptingbe related to IRS rather than to a source embedded in to reveal the mechanism that drives this outflow rota-SMM2, although the nature of a faint compact NIR tion (Launhardt et al., in prep.). Due to the edge-onknot at the peak position of SMM2 remains unknown. configuration, the bolometric luminosity listed in Ta-We are currently analyzing interferometric molecular ble 7 is only a lower limit, while the Lsmm / Lbol ratioline maps, attempting to better constrain evolutionary is an upper limit. Although the IR spectral index in-stages and the kinematic relation of the three sources dicates a Class II source, the other evolutionary tracers(Schmalzl et al., in prep.). CB 17 is thus one of those and the estimated age of 1 Myr (Launhardt et al. 2009)globules that contain two or more sources of differ- suggest it is a Class I YSO.ent evolutionary stage with a few thousand AU (see CB 34 (Fig. 4) has the appearance of a Bok glob-Sect. 4.2 and Table 6). ule, but, with a distance of ∼ 1.5 kpc, it is rather atyp- CB 17 SW (L 1388) is a small elliptical globule ical and more massive than other Bok globules. It 13
  14. 14. has three dense cores and is associated with numer- (2006) detected a water maser. Polarimetry of back-ous young stars that seem to have formed from this ground starlight revealed a magnetic field with an av-cloud (Launhardt & Henning 1997; Alves et al. 1997; erage P.A. of 116◦ (Sen et al. 2005). Because of thisHuard et al. 2000). A water maser (G´ mez et al. 2006) o complexity, we do not further investigate the individ-as well as several collimated outflows testify ongoing ual sources in this globule, but only present our data.star formation activity. A bipolar CO outflow from CB 58 (Fig. 6) is a relatively large, slightly elon-cool gas (Yun & Clemens 1994b), shocked optical gated and cometary-shaped globule at a distance ofknots of atomic emission lines from radiative shocks 1.5 kpc. It is associated with IRAS point source 07161-(Herbig-Haro objects), as well as H2 infrared jets (Yun 2336 and the NIR images indicate the presence of& Clemens 1994a; Khanzadyan et al. 2002) have been several embedded YSOs. Afonso et al. (1998) con-reported. The chemical age of the globule was esti- cluded from HCN observations that the core repre-mated to be > 105 yr (Codella & Scappini 1998). The sents a Class I evolutionary stage. Our (low-quality)presence of a pre-main-sequence star, CB34FU, with mm dust continuum map suggests the presence of twoan age of ∼ 106 yr supports this young age (Alves et al. dense cores in the head of the globule, where also the1997). Codella & Scappini (2003) studied the molecu- IRAS source and a number of reddened stars with as-lar gas of this globule and showed that the current star- sociated nebulosities are located. However, none offormation activity is concentrated in the three main the IR sources can be directly associated with one ofclumps which have sizes of ∼ 0.25 pc. Because of this the two dust clumps. Because of this complexity, wecomplexity, we do not further investigate the individ- do not further investigate the individual sources in thisual sources in this globule, but only present our data. globule, but only present our data. CB 39 is a small and not very opaque globule, lo- BHR 12 (CG 30, DC 253.3-1.6; Fig. 7) is a small,cated in direction of the Gemini OB1 cloud. It is opaque, and very pronounced cometary globule withassociated with the well-known Herbig Ae/Be star a bright rim, located in the Gum nebula region. TheHD 250550, the distance of which was derived with distance towards the Gum nebula clouds is somewhat700 pc. No mm dust continuum emission has been de- uncertain, with estimates ranging from 200 pc (Knudetected from this star, indicating that its circumstellar et al. 1999) to 450 pc (Reipurth 1983). For consistencydisk, if still present, cannot be very massive. Yun & with earlier papers (e.g., Henning & Launhardt 1998),Clemens (1994b) found a small, bipolar, low-velocity we use here 400 pc. The globule has an angular diam-(∼ 3 km s−1 ) CO outflow with not well-separated eter of ∼ 2.3 and its tail (which merges with the taillobes. CB 39 has been observed in different molec- of CG 31) has a length of 25 (Reipurth 1983). At itsular lines. It was detected in SO by Codella & Mud- head side, CG 30 harbors two protostellar cores with aers (1997) but not detected in CS by Launhardt et al. projected separation of ∼20 (∼8000 AU), embedded(1998). Sen et al. (2005) measured a magnetic field in a common envelope (Launhardt et al. 2000). TheP.A. of 150◦ , based on the polarization of a number of northern source (SMM 1, CG 30N) is associated withbackground stars along the line of sight of the globule. the IRAS point source 08076-3556 and a NIR source, CB 54 (LBN 1042; Fig. 5) is a relatively compact which drives the Herbig-Haro flow HH 120 (see Ho-molecular cloud with the appearance of a Bok globule, dapp & Ladd 1995, and references therein). Thebut seems to be associated with the Vela OB1 cloud southern core (SMM 2, CG 30S; Launhardt et al. 2001)complex at a distance of 1.1 kpc. It contains a clus- is the origin of a protostellar jet with position angleter of embedded young stars, visible on NIR images 44◦ (Hodapp & Ladd 1995), but no NIR counterpart(Yun & Clemens 1994a; Yun 1996), as well as a cold is seen at this position. Both cores are detected in allIRAS point source (07020-1618) at the position of the IRAC bands (Chen et al. 2008b). The IRAC 4.5 µmmm peak and the embedded star cluster. CO obser- image also shows bipolar jets, which have been tracedvations revealed the presence of a collimated bipo- in CO(2–1) using the SMA (Chen et al. 2008a). Fig. 7lar molecular outflow with well-separated lobes and shows that the two cores are well resolved at 450 andcentered on the IRAS point source (Yun & Clemens 850 µm, though they are surrounded by a common en-1994b). This source is also associated with cm contin- velope. The cores are unresolved in our 1.3 mm map.uum emission (VLA AY0073; Yun et al. 1996). C18 O The SED is well sampled for both cores, and showsobservations suggests the presence of multiple cores that SMM 2 is colder or more heavily extincted thanwith infall signatures (Wang et al. 1995). G´ mez et al. o SMM 1. We classify SMM 1 as a Class I YSO, based 14
  15. 15. on its bolometric temperature (see Table 7). The low globule contains a dense protostellar core, which emitsbolometric temperature of 50 K suggests that SMM 2 strong mm dust continuum emission and is detected byis a Class 0 protostar and thus younger than SMM 2. various instruments down to MIR wavelengths. Ob-BHR 12 is thus one of those globules that contain two servations with higher angular resolution revealed thator more sources of different evolutionary stage within this is actually a double core (Myers & Mardonesa few thousand AU (see Sect. 4.2 and Table 6). 1998) with two embedded IR sources, IRS 1 and IRS 2, BHR 36 (DC 267.4-7.5; Fig. 8) is a cometary- separated by 17 (3400 AU – Bourke 2001). Out-shaped globule in the Vela sheet at a distance of 400 pc. flows are seen in the H2 v = 1 − 2 S(1) line emanat-A dense core, mapped at 1.3 mm continuum, is located ing from both sources. A highly collimated molecularat the head side of the globule. The mm emission peaks outflow has been detected (Bourke et al. 1997) whichat the position of the IR source detected both by IRAS is associated with two Herbig-Haro objects (HH 320and Spitzer and at the center of a bipolar jet seen on and HH 321; Corporon & Reipurth 1997). BourkeNIR images. Two Herbig-Haro objects, HH46/47, are (2001) and Parise et al. (2006) resolved it into twoassociated with a collimated bipolar outflow centered outflows, mapping CO and methanol, and found thatat this position (Dopita et al. 1982). From their ob- the IRS 2 outflow is significantly hotter than the IRS 1servations of 13 CO(2–1), L¨ hr et al. (2007) derive a o outflow. From 13 CO(2–1) measurements, L¨ hr et al. ototal mass of 255 ± 60 M and a size of ∼ 1.3 pc for (2007) derived a mass of 12 ± 4 M for the entirethe entire globule. In their ammonia survey, Bourke cloud. From their ammonia measurements, Bourkeet al. (1995b) obtain a mass of ∼ 11 M within the et al. (1995b) derived a mass of 2.2 M for the densecentral 0.3 pc for the dense core. They also derive a core and Bourke et al. (1997) derive 2.4 M from thekinetic temperature of T k = 18 K in the centre. The 1.3 mm dust continuum map. The 1.3 mm contin-chemical age is of the globule was estimated to be uum image does not resolve the two sources, but the∼ 0.1 Myr (Cecchi-Pestellini et al. 2001). BHR 36 is low Tbol of the combined SED indicates that at leasttoo far south for the JCMT to observe. Like for the IRS 1 (the brighter source at the mm peak position) isother southern globules, we are therefore limited to a Class 0 protostar. This is in agreement with earlierthe 1.3 mm maps. The fit to the SED of this source work (Bourke et al. 1997).indicates Tbol of ≈150 K (see Table 7). We therefore BHR 86 (DC 303.8-14.2; Fig. 11) is a cometary-classify the embedded YSO as a Class I source. shaped globule located in the Chameleon dark cloud BHR 55 (DC 275+1.9; Fig. 9) is an average-sized, complex at a distance of 150 to 180 pc. L¨ hr et al. olow-extinction globule, associated with the annulus of (2007) carried out 12 CO(2–1) observations and mea-cometary globules in the Vela-Gum complex (Bourke sured a size of 15 × 15 and a mass of 19 ± 4 M .et al. 1995a), though it has not a very pronounced From ISO observations, Lehtinen et al. (2005) derivedcometary appearance. Bourke et al. (1995b) estimate central dust temperature of 14 ± 1 K and a mass ofthe distance to the globule to be 300 pc by looking 2.7 M for the dense core and concluded that the em-at the reddening of background stars. From ammo- bedded source is a Class 0/I transition object. Theynia observations, Bourke et al. (1995b) derive kinetic also showed that the main heating source of the denseand rotational temperatures of 13 K and 12 K respec- core is not the embedded source, but the interstellar ra-tively, a central volume density of 8.1 × 103 cm−3 , and diation field. From CS(2–1), HCO+ (1–0) and HCN(1–a core mass of 3.7 M . The globule is too far south 0) line profiles Lehtinen (1997) could classify BHR 86to observe with JCMT, therefore only the 1.3 mm data as a core with simultaneous infall of dense gas in theconstrain the long-wavelength end of the SED. The centre and a non-collapsing envelope. They also de-very compact (unresolved) mm source is associated tected a bipolar molecular outflow in 12 CO(1–0). NIRwith an IRAS point source (09449-5052), but has no observations in JHK s showed an exponential extinc-near-infrared counterpart. Due to the lack of data at tion profile with p = 2.29 ± 0.08 (Kainulainen et al.submm and MIR wavelengths, the SED is not well- 2007). From the low bolometric temperature of ≈60 K,constrained. However, the low Tbol of ≈40 K suggests the high Lsmm / Lbol ratio of ≈6%, and the absence of athis is a Class 0 protostar. NIR source we conclude this this is a Class 0 protostar, BHR 71 (DC 297.7-2.8; Fig. 10) is a very opaque, which is in agreement with the ISO observations.elongated globule located near the southern Coalsack CB 68 (L 146; Fig. 12) is a small, nearby, slightlyat an estimated distance of ∼ 200 pc. At its center, the cometary-shaped opaque Bok globule located in the 15