Star formation in Cometary globule 1: the second generation
aa r X i v : . [ a s t r o - ph . GA ] J u l Astronomy&Astrophysicsmanuscript no. 14524 c (cid:13)
ESO 2018November 7, 2018
Star formation in cometary globule 1: the second generation ⋆,⋆⋆
L. K. Haikala , , M. M. M¨akel¨a , , and P. V¨ais¨anen , Observatory, University of Helsinki, Finland Department of Physics, Division of Geophysics and Astronomy,P.O. Box 64, FI-00014 University of Helsinki, Finlande-mail: [email protected] South African Astronomical Observatory,P.O. Box 9, Observatory 7935, Cape Town, South Africa Southern African Large Telescope Foundation,P.O.Box 9, Observatory, 7935, Cape Town, South Africa
ABSTRACT
Context.
Cometary globule 1 (CG 1) is the archetype cometary globule in the Gum nebula.
Aims.
We attempt to discover stars possibly embedded in the globule head and to map the distribution of ISM in it.
Methods.
We analyse C O spectral line observations, NIR spectrosopy, narrow and broad band NIR imaging, and stellar photometryto determine the structure of the CG 1 head and the extinction of stars in its direction.
Results.
A young stellar object (YSO) associated with a bright NIR nebulosity and a molecular hydrogen object (MHO 1411, aprobable obscured HH-object) were discovered in the globule. Molecular hydrogen and Br γ line emission is seen in the direction ofthe YSO. The observed maximum optical extinction in the globule head is 9 . m
2. The peak N(H ) column density and the total massderived from the extinction are 9 . × cm − and 16.7 M ⊙ (d / . The C O emission in the globule head is detected in a 4 ′ by1 . ′ O line velocity and excitation temperature.Because of variations in the C O excitation temperature the integrated C O line emission does not follow the optical extinction. Itis argued that the changes in the C O excitation temperatures are caused by radiative heating by NX Pup and interaction of the YSOwith the parent cloud. No indication of a strong molecular outflow from the YSO is evident in the molecular line data. The IRASpoint source 07178–4429 located in the CG 1 head resolves into two sources in the HIRES enhanced IRAS images. The 12 and 25 µ memission originates mainly in the star NX Puppis and the 60 and 100 µ m emission in the YSO. Key words.
Stars: formation – Stars:pre-main-sequence – ISM:individual (CG 1) – Infrared: stars
1. Introduction
Sanduleak (1974) noted a 5 ′ by 10 ′ dark cloud west of theemission-line star Hen 32 (CoD -44 3318, NX Pup). This darkcloud was later included in a list of cometary globules byHawarden & Brand (1976). Cometary globules are elongateddark clouds with compact, dusty “heads” and long faintly lu-minous “tails”. The first list contained 12 sources of which tenwere in the outskirts of the Gum nebula. The sizes vary fromfew arcminutes (e.g. CG 5) to tens of arcminutes (CG 12). Sincetheir discovery similar cometary structures have been found tobe common among interstellar clouds, the scales ranging fromisolated tiny globules to complete cometary shaped star-formingregions such as the Corona Australis cloud. It has been suggestedthat the formation of the classical CGs is caused by radiation-driven implosion (Reipurth 1983) of an isolated, extended glob-ule or a passage of a supernova plane blast wave colliding withan extended globule (Brand et al. 1983).CG 1 (Hawarden & Brand 1976) is a prototypical cometaryglobule. It has a few arcminute sized compact bright rimmedhead and nearly a half-a-degree-long tail. The bright pre-main-sequence binary star NX Pup lies just outside the opaque glob- ⋆ Based partly on observations collected at the European SouthernObservatory, La Silla, Chile and partly obtained from the ESO / ST-ECFScience Archive Facility ⋆⋆ Appendices A, B and C are only available in electronic form viahttp: // ule head. The IRAS point source 07178–4429 lies between NXPup and the opaque globule head. It has far infrared (FIR)colours similar to young stellar objects. Extensive molecu-lar line observations of CG 1 are presented in Harju et al.(1990) (hereafter HSHWSP). Even though HSHWSP reportrather strong CO lines (SEST antenna temperatures up to10K), the C O (1–0) emission is weak. In the C O maximumlocated in the globule head, the integrated C O (1–0) main-beam brightness temperature (T MB ) was less than 0.55 K km s − .Two CO velocity components are seen in the globule head. Thestronger one follows the globule optical image from head to tail.The weaker velocity component is distributed in an elongatedstructure and coincides with the bright “nose”-like extension inoptical surface emission, which extends to below the star NXPup. The Bourke et al. (1995) single point ammonia observationin the CG 1 head was a non-detection. Single point observationsof IRAS 07178–4429 in CO(3–2), C O (3–2), HCO + (4–3), andH CO + (4–3) are presented in van Kempen et al. (2009). Theauthors conclude that the point source has no associated HCO + core and that it very likely does not have a circumstellar shell.The distance to CG 1 is uncertain and the estimates rangefrom 300pc (Franco 1990) to 500pc (Brand et al. 1983). Franco(1990) analysed the distribution of stellar E(b–y) colour excessesin the direction of the Gum nebula and concluded that the dis-tance to the centre of the nebula is 290 ± L. K. Haikala et al.: Star formation in cometary globule 1: the second generation
Fig. 1.
Colour-coded SIRIUS image of the CG 1 head. The J , H , and K s bands are coded in blue, green, and red, respectively.Square root scaling has been used to enhance the faint surfacebrightness structures. NX Pup is the brightest star in the imageat 7 h m . s − o ′ . ′′ ′′ away was discovered byBrandner et al. (1995). The globule head is strongly obscuredin the optical and in the 2MASS survey the number of starsin the globule head is also smaller than in the surroundings.A deeper-than-2MASS near-infrared (NIR) study is thereforeneeded to study the possible stellar population still embeddedin the globule and the distribution of dust and gas in the globulehead. We present J , H , and K s NIR imaging of the CG 1 headwith the InfraRed Survey Facility (IRSF) 1.4m telescope at theSouth African Astronomical Observatory SAAO and the 3.5mNew Technology Telescope (NTT) at La Silla. Low resolutionNIR spectroscopy and narrow band imaging in the H µ m line were also conducted with the NTT. The C O (1–0)and (2–1) molecular line observations of the globule head wereobtained at the Swedish ESO Submillimetre Telescope (SEST)at La Silla. Observations, data reduction, and calibration proce-dures are described in Sect. 2 and the observational results inSect. 3. The new results are discussed and compared with avail-able data at other wavelengths in Sect. 4. The results are sum-marised and the conclusions drawn in Sect. 5.
2. Observations and data reduction
The head of CG 1 was imaged in J , H , and K s with theSimultaneous InfraRed Imager for Unbiased Survey (SIRIUS,Nagayama et al. 2003) on the IRSF 1.4m telescope on Jan. 2007.The three colours were observed simultaneously. The SIRIUSfield of view is 7 . ′ × . ′ . ′′ ′′ . The observations were carried out in the on-o ff mode instead of the standard jitter mode to preserve the sur-face brightness. After each on-integration of 10s, an o ff -position outside the globule was observed with the same integration time.Jittering was performed after each on-o ff pair. Sky flats were ob-served every evening and morning.The Son of Isaac (SOFI) imaging in J s, H , and K s at NTTwas performed in Feb 2007. The SOFI field of view is 4 . ′ . ′′ ∼ . ′′
7. The bright pre-main-sequence star NX Pup, which is located to the east of the glob-ule head, was not within the field of view. A reference positionat the same Galactic latitude but outside the globule (7 h m -44 o ′ , J2000) was observed in all colours. Standard stars fromthe Persson et al. (1998) faint NIR standard list were observedfrequently during the nights.Extended molecular hydrogen emission was searched for byimaging the CG 1 head with the narrow SOFI filters NB H2 S1(H µ m) and NB 2.090 (continuum). The narrow-band imaging was done in the standard jittering mode.The IRAF external XDIMSUM package was used to reducethe SIRIUS and SOFI imaging data. The images were examinedfor cosmic-ray hits and then sky-subtracted. The two nearest im-ages in time to each image were used in the sky subtraction.An object mask was constructed for each image. Applying thesemasks in the sky subtraction produced hole masks for each sky-subtracted image. Special dome flats and illumination correctionframes provided by the NTT team were used to flat field andillumination-correct the sky-subtracted SOFI images. Sky flatswere used to flat-field the SIRIUS images. Rejection masks com-bined from a bad pixel mask and individual cosmic ray and holemasks were used when averaging the registered images. The re-duced SOFI J s, H , and K s images are shown in Figs. A.1, A.2,and A.3.NX Pup is very bright and produces an artifact (inter-quadrant row cross talk) which appears as a stripe in the samecolumn as the star and symmetrically on the other half of thedetector in the SIRIUS images. Reflections in the SIRIUS filterscause ring-like ghosts around NX Pup. We attempted to correctfor neither the inter-quadrant row cross talk nor the reflections.The SExtractor software v 2.5.0 (Bertin & Arnouts 1996)package was used to obtain stellar photometry of the reducedSOFI images. The galaxies were excluded using the SExtractorCLASS keyword and by visually inspecting the images. Afterelimination of the non-stellar sources, 262 stars remained. Themagnitude zero points of the summed data were fixed usingthe standard star measurements. The instrumental magnitudeswere first converted to the Persson et al. (1998) photometric sys-tem and then to the 2MASS photometric system as describedin Ascenso et al. (2007). The magnitude scale was checkedby comparing the SOFI photometry of stars in common with2MASS that have high quality 2MASS photometry. The limit-ing magnitudes are approximately 21 . m
5, 20 . m
5, and 20 . m J , H , and K s, respectively. The limiting magnitude however variesover the observed area and is brighter in the regions where thebackground surface brightness is strong. IRAF is distributed by the National Optical AstronomyObservatories, which are operated by the Association of Universitiesfor Research in Astronomy, Inc., under cooperative agreement with theNational Science Foundations. K. Haikala et al.: Star formation in cometary globule 1: the second generation 3
Fig. 2.
Colour-coded SOFI image of CG 1 head. The J s, H , and K s bands are coded in blue, green, and red, respectively. Squareroot scaling has been used to enhance the faint surface brightness structures. A low resolution SOFI spectrum was acquired covering thewavelengths from 1.53 µ m to 2.52 µ m (resolving power 980)using a slit oriented along a bright surface brightness feature (seeFigs. 3 and 4). The SOFI standard nodding observing templatewas used. La Silla observatory IR standard F7V star Hip37123was observed as a telluric standard.The spectroscopic science frames were flatfielded using ON-OFF flats, illumination and cosmic-ray corrected, and then sub-tracted pairwise from each other. Xe-arcs were used to fit thewavelength solution and correct for the two dimensional shapedistortion. Frames were then shifted, coadded, and traced. Thestandard star was reduced in exactly the same way, and its onedimensional extracted spectrum was divided into the two di-mensional target per row, and the result multiplied by a smoothblack-body model of the star, thereby removing telluric featuresand also performing a relative flux calibration. C O mapping The C O (1–0) and (2–1) molecular line observations weremade in Sep. 2000 at SEST. The observations were con-ducted with the SEST 3 and 1 mm (IRAM) dual SiS SSBreceiver using the frequency-switching observing mode. A 6MHz and 18 MHz frequency switch was used at the C O (1–0) and (2–1) line frequencies, respectively. The CG1 glob-ule head was mapped simultaneously in the C O (1–0) andthe C O (2–1) transitions in a regular grid with a spacingof 20 ′′ . Altogether 137 positions were observed using one-minute integration times. Selected positions were integratedlonger. Typical values of the e ff ective SSB system tempera-tures outside the atmosphere ranged from 170 K to 250 K. TheSEST high-resolution 2000-channel acousto-optical spectrom-eter (bandwidth 86 MHz, channel width 43 kHz) was dividedinto two to measure the two receivers simultaneously. At theobserved frequencies, 109.782173 GHz and 219.560353 GHz, L. K. Haikala et al.: Star formation in cometary globule 1: the second generation the spectrometer channel width corresponds to ∼ − and ∼ − , respectively. The SEST HPBW and main-beam e ffi ciency were 47 ′′ and 0.7,respectively, for the C O (1-0) frequency, and 23 ′′ and 0.5 for the C O (2–1) frequency.Calibration was performed using the chopper wheel method.Pointing was checked regularly towards SiO maser sources andthe pointing accuracy is estimated to be better than 5 ′′ .The frequency-switched spectra were first folded and a sec-ond order baseline was then subtracted. The observed spectrumrms was typically ∼ ∼ MB scale for theC O (1–0) and (2–1) transitions, respectively. All the SEST linetemperatures in this paper are in the T MB units, i.e, corrected tooutside the atmosphere, assuming that the source fills the mainbeam.
3. Results
JHK s Imaging
False colour SIRIUS and SOFI images of the CG 1 head areshown in Figs. 1 and 2, respectively. The J , H , and K s bandsare coded in blue, green, and red, respectively. The SOFI imagecontains only the obscured globule head, whereas the SIRIUSimage also contains the bright pre-main-sequence star NX Pupand a bright reflection nebulosity south of it. The SIRIUS obser-vations were made in the on-o ff mode, which conserves the ex-tended surface brightness. The SOFI observations were obtainedin the jitter mode where the surface brightness with scale largerthan the jitter box (30 ′′ in this case) is smeared and / or cancelledin the data reduction. Only small-size features and gradients inthe original surface brightness structure are retained. Point-likesources and galaxies are una ff ected.The edges of the globule are well defined in both the SIRIUSand the SOFI images. In the SIRIUS image, the surface bright-ness is extended NW of NX Pup and a narrow filament followsthe globule southern edge. The surface of the CG 1 head is cov-ered with narrow, faint filaments in the SOFI image. Because ofthe jittering observing mode, the extended surface brightness istransformed into filaments that trace small-scale gradients in theactual surface brightness.Besides the stars and numerous galaxies, a semistellar sourcesurrounded by a bright nebulosity is seen in both the SIRIUS andSOFI images. The nebulous source has the appearance of a typ-ical young stellar source (see e.g. Zinnecker et al. 1999) and isreferred to as YSO in the following. A bright and elongated sur-face brightness feature protrudes SE of the YSO in the K s image.This feature is referred as the YSO SE filament in the following.Thin, wispy filaments extend both west and east of the YSO inFig. 2. The nebulosity associated with the YSO is extended andvery bright. Therefore, it had to be masked heavily in the jittersky frames during the data reduction. As a consequence, most ofthe structure seen in the reduced images around the YSO is realand not smeared like the other extended structures in the SOFIimages. A faint surface brightness patch is seen in the SOFI im-ages ∼ ′′ west of the YSO. Both the thin filaments and thesurface brightness patch west of the YSO are seen in all threecolours. Except for the YSO and the regions NW and SE of itthe globule NIR surface brightness is predominantly blue in Fig.2, i.e. more pronounced in the J s band than in the two othercolours. The ring-segment-like structure facing the bright starNX Pup seen in Figs. 1 and 2 as well as the bright patch south ofthis star (Fig. 1) have counterparts in the optical images of CG 1(see e.g. Brand et al. 1983, Plate 1). The YSO and these featuresare identified in Fig. 3. The ring-segment-like structure and the Fig. 3.
The notable surface brightness features discussed in thetext are marked on an extract of the SOFI K s image. The positionof the semi-stellar source is marked in the centre of the brightnebulosity (white in the figure) surrounding it. The YSO SE fil-ament is shown with a line in the insert. The contours in the in-sert delineate the central source. The continuous curve marks thering-segment-like structure facing the NX Pup and the dashedcurves the brightest filaments emanating from the YSO. Thefaint surface brightness patch is inside the circle in the right. Fig. 4.
SOFI
JHK s images of the CG 1 YSO. The lower panelsshow the original image. The filter is indicated in the upper leftcorner of each panel. The size of the lower panels is 46 ′′ by 46 ′′ .The upper panels (23 ′′ by 23 ′′ ) show surface brightness contoursoverlayed on Gaussian smoothed images. The line in the upper K s panel shows the orientation of the SOFI slit on the nebulosity(see. Sect.3.2). The cross in all the panels indicates the positionof the semistellar YSO. The contour levels in SOFI counts are: J s 2 to 5 by 1 and (in white) 6; H K s 3 to 23 by 5, (in white) 28 to 112 by 16, and 200 to600 by 80brightest thin filaments are indicated by a continuous line anddashed lines, respectively. The YSO SE filament is marked witha line in the insert and the nebulous patch west of the YSO witha circleA zoomed image of the YSO in the SOFI image in threecolours is shown in the three lower panels of Fig. 4. The filter isindicated in the upper left corner of each panel. The size of thelower panels is 46 ′′ by 46 ′′ , which approximately correspondsto the SEST HPBW at the C O (1–0) line frequency, 109 GHz.The upper 23 ′′ by 23 ′′ panels (SEST HPBW at the C O (2–1)frequency, 219 GHz) show surface brightness contours superim- . K. Haikala et al.: Star formation in cometary globule 1: the second generation 5 posed on Gaussian smoothed images. The contour levels and thegrey scale were chosen to enhance the contrast in surface bright-ness for each filter. The line in the K s image shows the orienta-tion of the spectrometer slit. The cross in all the panels indicatesthe position of the semi-stellar source in the K s image. The H and K s surface brightness contours are overlaid on a 23” by 23” J s grey scale image in Fig. A.4. The absolute surface brightnessin the J s image is very low compared with that in the K s im-age. The K s semi-stellar source lies behind a narrow obscuringlane in the J s image. In the H filter, the maximum emission lieseast of the K s maximum and is not stellar like. The half widthof the K s semi-stellar source intensity profile in the EW direc-tion is nearly twice that of isolated stars elsewhere in the image.This indicates that even at the K s maximum, a large part of theobserved emission is reflected light and not direct light from thecentral source.The YSO is visible in the 2MASS survey (2MASS07192185-4434551), which is a K s band detection (13 . m
42) only.The 2MASS extended source 2MASX J07192176-4434591 cor-responds to the extended YSO nebulosity. The YSO is also visi-ble in the K s image presented in Santos et al. (1998), who iden-tify it as a red object, possibly a YSO in an early evolutionarystage embedded in the globule. Part of the SOFI low-resolution long-slit spectrum, in both twodimensional and extracted format, is shown in Fig. 5. The 12 ′′ centre part of the spectrum through the YSO nebulosity and theYSO SE filament is shown in Fig. 5, lower panel. Extractedcontinuum- subtracted spectrum convolved with a 3-pixel box-car is shown in the upper panel. The wavelengths of the H lineswithin the displayed spectral region and of the Br γ line are in-dicated. The spectrum has been normalized for an (1–0) S(1)intensity of unity.At least three H lines and the Br γ line are evident in thespectrum. The H lines are detected in the centre of the nebulos-ity and also at the base of the bright YSO SE filament. The fila-ment is most clearly seen in the H γ emission is strictly constrained to the spatial area of the contin-uum peak. A relatively strong continuum is seen in the directionof the semi-stellar YSO. H imaging The di ff erence between the images obtained using the narrowfilter NB H2 S1 covering the H µ m line and theadjacent continuum obtained with the filter NB 2.090 reveals twoobjects. The stronger one lies at the centre of the YSO nebulosityand another 90 ′′ west of it. The latter coincides with the faintsurface brightness patch, which is seen in the SOFI images in allthree colours and noted in Sect. 3.1. An extract of the NB H2 S1image and the NB H2 S1 - NB 2.090 di ff erence image coveringthe two objects is shown in the two upper panels in Fig. 6. Thecross in the middle panel shows the position of the YSO. Thestars and the strong surface brightness surrounding the YSO arecancelled out and only a small size source in the direction ofthe YSO and a faint patch west of it remain in the di ff erenceimage. The lowest panel shows the di ff erence image smoothedby a three-pixel half-width Gaussian. Fig. 5.
Part of the SOFI spectrum over the YSO and its asso-ciated nebulosity in both two dimensional and extracted for-mat. Lower panel: 12 ′′ centre part of the low-resolution long-slit spectrum through the YSO and the YSO SE filament. Upperpanel: extracted continuum-subtracted spectrum convolved witha 3-pixel box-car. The wavelengths of the H lines within thedisplayed spectral region and of the Br γ line are indicated.The spectrum has been normalized for an (1–0) S(1) intensityof unity. The horizontal bar at 2.01 µ m indicates the extent ofthe strong atmospheric absorption between the H and K s bands. Fig. 6. H emission in the CG1 head. The upper panel shows anextract of the image obtained with the SOFI NB H2 S1 filter.The centre panel shows the di ff erence NB H2 S1 - NB 2.090.The cross shows the position of the YSO in the K s image. Thelowest panel shows the di ff erence image smoothed with a three-pixel half-width Gaussian. The
JHK s colour-colour diagram for stars in the CG 1 head isshown in Fig. 7, left panel. The main-sequence (Bessell & Brett1988) transformed to the 2MASS photometric system is alsoplotted. The arrow shows the e ff ect of 5 m of visual extinctionin the diagram according to the Bessell & Brett (1988) red-dening law. A small symbol is used if the ( J - H ) and / or the( H - K s) formal error is larger than 0 . m
1. The right panel showsthe colour-colour diagram for stars in the reference position.Normal reddened main-sequence stars should lie near or be-tween the reddening lines indicated with dashed lines in Fig. 7.Stars with circumstellar shells and disks have infrared excess and
L. K. Haikala et al.: Star formation in cometary globule 1: the second generation
Fig. 7.
Left:
JHK s colour-colour diagram for stars in the CG 1head. The stars with ( J - H ) and / or ( H - K s) error larger than0 . m ff ect of 5 m of vi-sual extinction in the diagram according to the Bessell & Brett(1988) reddening law. Right: As on the left, but for stars ob-served at the reference position.lie to the right of the reddening line. Carbon stars, long periodvariables, and extragalactic sources may mimic infrared excesssources (Foster et al. 2008). Even though the obvious extragalac-tic sources were filtered out from the original data set it is highlylikely that some of these sources remain in the final data set. Thetwo stars in CG 1 that lie to the right of the reddening line inFig. 7 are faint ( J s ∼ . m
7) and were possibly not recognisedas extended sources. If these objects were stars associated withCG 1, they would have to be extremely faint (substellar). Anunreddened M5 star at a distance of 300pc has an apparent J magnitude of 14.6. C O mapping The observed C O (1–0) (in black) and (2–1) (in red) spectraare shown in Fig. 8. The spectra are plotted in the main-beambrightness scale from 1.0 km s − to 6.0 km s − in velocity andfrom -0.5 K to 2.8 K in temperature. The positional o ff sets arearcseconds from 7 h m . s o ′ . ′′ . ′′ O (1–0)and (2–1) emission are shown in Figs. B.1 and B.2. The distribu-tion of the C O (1–0) emission in Fig. 8 and Fig. B.1 is consis-tent with that shown in HSHWSP, Fig. 1. The di ff erences can beattributed to the coarse sampling used by HSHWSP (40 ′′ com-pared to the 20 ′′ spacing in this paper). In addition the low valueof the HSHWSP maximum integrated C O (1–0) line integral,which is less than 50% of the present data, may be explained ina similar way.A contour map of the integrated C O (2–1) emission in themain-beam brightness scale superimposed on the SOFI K s im-age is shown in Fig. 9. The contours are from 0.4 K km s − to2.4 K km s − in steps of 0.4 K km s − . The distribution of theemission is elongated with a sharp maximum SW of the YSO.The general C O distribution correlates well with the opticalextinction evident in the SOFI J s image (Fig. A.1). Fig. 8.
The observed C O (1–0) (in black) and (2–1) (in red)spectra. The spectra are plotted in the main-beam brightnessscale from 1.0 km s − to 6.0 km s − in velocity and from -0.5K to 2.8 K in temperature. The positional o ff sets are in arcsec-onds from 7 h m . s o ′ . ′′ . ′′ Fig. 9.
Contour map of the C O (2–1) integrated emission(main-beam brightness scale) overlaid on the K s SOFI image.Contour levels are from 0.4 K km s − to 2.4 K km s − in steps of0.4 K km s − . SEST HPBW at the C O (2–1) frequency is indi-cated in the upper right. The positional o ff sets are in arcminutesfrom NX Pup indicated with a filled circle. . K. Haikala et al.: Star formation in cometary globule 1: the second generation 7 Fig. 10. C O (2–1) (histogram) and (1–0) (line) spectra in threepositions. The o ff set from the map centre is indicated in the up-per left corner of each plot. The velocity marker at the top ofeach panel is at velocity 3.4 km s − .The C O (1–0) and (2–1) spectra at three positions areshown in Fig. 10. The o ff set from the map centre position isshown in arcseconds at the upper left corner of each panel. Thevelocity marker at the top of each panel is at velocity 3.4 km s − .The o ff set positions (-20 ′′ ,-40 ′′ ), (-80 ′′ ,20 ′′ ) and (-140 ′′ ,80 ′′ ) arereferred to as SE, centre and NW positions, respectively. In Fig.10, the C O (2–1) line intensity is significantly stronger than theC O (1–0) intensity at the SE and centre positions. However,the centre of line velocities are di ff erent, 3.69 ± − at SEand 3.33 ± − at the centre position. Also the line widthsdi ff er being 0.46 km s − and 0.70 km s − at the two positions,respectively. The velocity di ff erence was seen by HSHWSP in CO and here it is confirmed in the C O emission. At the NWposition, the intensities in the two transitions are equal and theline velocity is similar to that at the centre position. The lines areasymmetric (blue wing at centre position, and red at NW posi-tion).
4. Discussion C O mapping In the following discussion, the C O map is divided into threeregions: 1) C O SE (the 3.7 km s − velocity component is dom-inant); 2) C O maximum (the integrated C O emission maxi-mum); and 3) C O NW (region in the NW where the C O (1-0) and (2–1) T MB intensities are similar).In the LTE approximation, the C O line intensity dependson both the C O column density and the C O excitation tem-perature, whereas the C O (1–0) to C O (2–1) intensity ratiodepends on the excitation temperature. If the C O excitationtemperature were constant in the mapped area, the integrated linearea of the optically thin C O emission would trace the cloudgas column-density linearly. However, the C O peak T MB (2–1) / T MB (1–0) ratio varies over the map. In Figs. 8 and 10, theC O (2–1) emission is significantly stronger than the C O (1–0) emission in the direction of the C O maximum and inC O SE. In C O NW, the intensity of the two transitions isthe same. The varying line ratio indicates that the excitation tem-perature is higher in both C O maximum and C O SE than inC O NW.The interaction of the newly born star with the surroundingcloud (see the discussion below) would be a likely reason fora warmer spot in the cloud (C O maximum). The C O linewidth is also the broadest in C O maximum, which is indica-tive of extra turbulence at this position. The C O line veloc-ity in C O SE di ff ers from that in C O maximum, implyingthat these are separate structures in the cloud. The high NIRsurface brightness around C O SE in Figs. 1 and 2 indicates that the radiation from the star NX Pup is the reason for theelevated temperature. C O NW lies NW of the YSO and theC O maximum and is most probably well shielded against theradiation from NX Pup. This region of the cloud would corre-spond to a quiescent, low temperature part of the cloud.The C O (2–1) emission is distributed in a north-south ori-ented ridge, which is inconsistent with a point-like source. It canbe seen in Figs. 8 and 9 that the C O (2–1) maximum doesnot fall on the measuring grid but lies slightly NE of the mea-sured maximum line integral and west of the location of theYSO. If the C O maximum were point-like, it would be jus-tifiable to argue that the positional o ff set between the maximumand the YSO would be caused by a larger than expected tele-scope pointing error during the observations. However, the max-imum is elongated and there is no obvious reason why the north-south oriented C O ridge should be centred on the YSO posi-tion. The C O mapping grid spacing is only marginally finerthan the SEST beam HPBW at the C O (2–1) frequency (grid20 ′′ , HPBW 23 ′′ ) and therefore a fully sampled map is neededto pinpoint the exact location of the maximum.The HCO + (4–3) line observed by van Kempen et al. (2009)in the direction of the YSO is weak (T MB CO + (4–3) was a upper limit of 0.08K. This rules out a dense core inthis direction (van Kempen et al. 2009). Even though the posi-tion where van Kempen et al. (2009) observed the C O (3–2)and the HCO + (4–3) lines, the YSO, lies between the mappingpositions in this paper, their measurement can be used to evaluatethe nature of the C O maximum. The APEX beam size at 319GHz is similar to the SEST beam size at 220 GHz, which makescomparing the line intensities observed with APEX and SESTrealistic. The APEX C O (3–2) T MB line intensity, 2.3K, is sim-ilar to the SEST C O (2–1) line intensities observed aroundthe YSO. Because the C O (3–2) line temperature is low, theC O maximum position cannot be a C O hotspot similar tothat discovered in CG 12 (NGC 5367) (Haikala et al. 2006).Even though the emission from high density tracers is weak inthe CG 12 C O hotspot, the C O (3–2) T MB of 10K is highcompared to the T MB temperatures in the (2–1) and (1–0) tran-sitions (5 K and 2.2 K, respectively). Observations of the highdensity tracers (van Kempen et al. 2009) and C O thus indicatethat C O maximum is a moderately dense elongated structurewest of the YSO position.
The NICER method presented in Lombardi & Alves (2001) andthe SOFI NIR photometry can be used to estimate the visual ex-tinction within the imaged area in CG 1. However, the positionof CG 1 14 o below the Galactic plane is unfavourable for apply-ing the NICER method because the expected number of brightbackground stars, especially of giant stars, is low. Therefore thespatial resolution of the resulting extinction map is only 20 ′′ .The J - H / H - K s colour-colour diagram for the observed o ff -fieldis shown in Fig. 7, right panel. The diagram shows that the in-terstellar extinction is small in the o ff -field direction. The fieldlies at the same Galactic latitude as the on-field and therefore thestellar population should be statistically similar to that of the on-field. Thus the o ff -field is well suited to being used in the NICERmethod. The contour map (thick lines) of the extinction derivedfor the CG 1 head superimposed on the SOFI K s image is shownin Fig. 11. The contours are from 2 . m . m . m
0. hecontours of the C O (2–1) integrated emission (0.8 K km s − to2.4 K km s − in steps of 0.4 K km s − ) are shown with thin lines.The minimum visual extinction, 1 . m
3, in the map is in the NE cor-
L. K. Haikala et al.: Star formation in cometary globule 1: the second generation
Fig. 11.
Contour map of the optical extinction (thick lines) ob-tained with the NICER method overlaid on the K s SOFI image.Contour levels are from 2 magnitude to 9 magnitudes in stepsof 1 magnitude. Contours of the C O (2–1) integrated emis-sion (0.8 K km s − to 2.4 K km s − in steps of 0.4 K km s − ) areshown with thin lines. The positional o ff sets are as in Fig. 9ner of the image. This non-zero minimum extinction agrees withFig. 7, where practically all late main sequence stars are abovethe main-sequence also plotted in the figure. The maximum ex-tinction, 9 . m
2, is located NW from the YSO. However, a word ofcaution is required because the NICER method assumes that theextinction is constant in each NICER cell (in this case 20 ′′ by20 ′′ ). If a cell contains a compact (with respect to the cell size)high extinction structure, the measured extinction tends to be bi-ased towards lower extinction (Lombardi 2005). Thus, localisedsmall-size extinction maxima may go unnoticed.The position of the maximum extinction does not coin-cide with the maximum C O integrated emission (Fig. 9). Thisshows, as discussed in Sect. 4.1, that because of the varyingC O excitation temperature the distribution of the observedC O integrated emission does not linearly trace the gas / dustcolumn density distribution. In addition, the optical extinctionalso traces the low density cloud envelope where C O moleculeis not excited.
A detailed three-dimensional non-LTE model that accounts forboth the varying density and excitation conditions would be nec-essary to derive the cloud physical properties. The constructionof this model is not possible because only C O (1–0) and (2–1) mapping data can be used and only a few pointed observa-tions are available for other molecules. The LTE-approximationapproach is used instead of a sophisticated model to ob-tain a zeroth-order estimate of the H column densities andmasses. An average C O excitation temperature for C O SE,C O maximum, and C O NW is estimated assuming opticallythin C O emission and using the observed relative C O (2–1)and (1–0) line intensities. This assumes that the observed emis-sion in both transitions originates in the same volume of gasat a constant excitation temperature. The observed C O line T MB (2–1) / T MB (1–0) ratio is 2 in C O SE and C O maximum,which is compatible with an excitation temperature near 15 K. InC O NW, the ratio is one indicating an excitation temperatureof the order of 10 K. The low C O (2–1) T MB temperature inC O NW may also be due to subthermal excitation. However,according to model calculations presented in Warin et al. (1996),the deviation of the populations in the lowest C O energy levelsfrom their LTE values is small in typical conditions prevailing indark clouds. Considering the high H column density derived forC O NW from optical extinction (see below), subthermal exci-tation is unlikely.The N(H ) column density can be estimated using boththe C O data and the optical extinction. The relation betweenthe molecular hydrogen and C O column density is N(H ) = [N(C O ) / × + .
3] 10 (Frerking et al. 1982). AssumingLTE and C O excitation temperature of 15 K in C O SE andC O maximum and 10 K in C O NW, the peak N(H ) columndensities in these regions are 5 . × cm − , 7 . × cm − ,and 4 . × cm − , respectively. The relation between the op-tical extinction and hydrogen column density is N(H) = × cm − mag − (Bohlin et al. 1978). The peak N(H ) columndensity estimated from the extinction data is 9 . × cm − . Themaximum is in the direction of C O NW and is nearly twicethat estimated from the C O data.The LTE column densities can be compared with thosecalculated with a non-LTE radiative transfer code such asRADEX (van der Tak et al. 2007), which is available on-line.The C O (1–0) and (2–1) line intensities are calculated byproviding the kinetic temperature, H number density, theC O column density, and the C O line width as input parame-ters. The line width is known in CG 1 and a H number densityof 10 cm − is a reasonable guess. The observed C O (2–1) toC O (1–0) line intensities in C O NW can be reproduced withRADEX assuming a C O 10K excitation temperature and theLTE column density calculated above. However, in C O SE andC O maximum the C O excitation temperature has to be in-creased from 15K to 30K to make the line intensities agree. Thisis because of subthermal excitation of the C O (2–1) transition.Increasing the N(H ) column density makes the C O (1–0) linestronger in respect to the C O (2–1) line which does not agreewith the observations. The comparison of LTE and non-LTE re-sults shows that the two methods produce similar N(H ) columndensities if the input kinetic temperature is adjusted. The conver-sion factor from the line integral to column density in the LTEmethod is temperature dependent but the variation is only 20%in the 10K to 30K range.The cloud mass can be calculated from the C O data bysumming up the calculated N(H ) column densities point bypoint and using the mean molecular weight per H molecule 2.8.A major uncertainty in the mass calculation is the distance toCG 1, as the mass scales as the square of the distance. If thedistance d to CG 1 is 450pc instead of the assumed 300 pc, thecalculated mass more than doubles. Assuming C O excitationtemperatures as above, the masses of C O SE, C O maximum,and C O NW are 0.75M ⊙ ( d / pc ) , 1.35M ⊙ ( d / pc ) ,and 1.6M ⊙ ( d / pc ) , respectively. These numbers shouldbe considered very rough estimates because the regions over-lap and the division of data between them is subjective. Themasses are lower limits because they refer to gas traced by theC O emission. The true masses are higher as the clumps havelow density envelopes that are not detected in C O. The mostmassive of the three regions is C O NW. The C O line inten-sity in its direction is lower than in the other two regions but thisis compensated by its larger size. . K. Haikala et al.: Star formation in cometary globule 1: the second generation 9
Summing up the extinctions point by point, using theBohlin et al. (1978) relation between the extinction and hydro-gen column density and mean molecular weight per H molecule2.8, the total mass from the extinction data in the imaged area is16.7M ⊙ ( d / pc ) . The A v = O . The mass within thiscontour is 9.2M ⊙ ( d / pc ) , which is more than twice thesummed up C O mass. The discrepancy between the massescalculated from the extinction data and the C O data is expectedbecause the optical extinction also traces the low density cloudenvelope, which is not detected in C O. The SOFI spectrum was taken to investigate the nature of thenebulosity, i.e., ascertain whether it is only reflected light orwhether line emission is also present. The spectrum in Fig. 5shows an underlying continuum, at least three H lines, and theBr γ line. The H line emission does not only originate in thelocation of the YSO but also from the base of the YSO SE fil-ament. The spectral resolution and the spectrum low signal-to-noise ratio do not allow us to extract detailed physical informa-tion, though by comparing with the models of e.g. Smith (1995)we note that the line strengths of the H lines appear more con-sistent with C-type shocks than J-type. As expected, the Br γ lineemission is seen only in the direction of the YSO as this line isexpected to originate very near the YSO (see e.g. Malbet 2007).The underlying continuum emission is direct light from the YSOand light reflected from the surroundings. The spectrum is notsensitive enough to show the continuum in the direction of theSE filament.The H emission region can also be seen in the narrow-bandH / continuum di ff erence image in Fig. 6. It is centered on the K sYSO and is resolved in the E-W direction but not in the N-S di-rection. The extent in the E-W direction is 5 ′′ . The NIR spectrumshows that H emission comes also from the base of the YSO SEfilament but the narrow-band images are not deep enough andemission in Fig. 6 is confined only to the YSO.For the o ff set 90 ′′ west from the YSO, a faint patch of emis-sion is detected. This patch coincides with those seen in the J s and H bands (Sect. 3.1). The di ff erence image in Fig. 6shows that in the K s band this emission is mostly due to H line emission. Though this object is likely to be a HH-object itcannot be confirmed in the optical. Therefore it will be desig-nated as a molecular hydrogen emission-line object MHO 1411,(Davis et al. 2010). CG 1 was mapped by the IRAS satellite in wavelengths 12 to 100 µ m. Besides extended surface emission, IRAS detected a pointsource, IRAS 07178–4429, at the edge of the globule head. Thenon-colour-corrected fluxes of the point source are 6.68, 7.60,13.12, and 33.59 Jy at 12, 25, 60, and 100 µ m, respectively.The positional uncertainty ellipse major and minor axis are 10 ′′ ,3 ′′ and the position angle (East through North) 178 ◦ . The 60 and100 µ m fluxes point to an embedded source. However, consid-ering the 100 µ m flux of 33.59 Jy, the 12 µ m flux of 6.68 Jy istoo strong for such a source.The spatial resolution of original IRAS images can be en-hanced using HIRES processing, which uses the maximum cor-relation technique (Auman et al. 1990). Figure 12 displays con-tour maps of the CG1 HIRES processed (20 iterations) IRAS maps superimposed on the SOFI K s band image. The IRASwavelength in microns is shown in the upper left corner of eachpanel and the position of the star NX Pup is marked with afilled circle. The IRAS 07178–4429 point-source positional un-certainty ellipse is shown in the 100 µ m panel. It lies west ofNX Pup at the edge of the K s image. The C O (2-1) integratedemission contours (1.2 K km s − to 2.0 K km s − in steps of 0.4K km s − ) are superimposed on the 12 µ m panel. The 12 µ mIRAS contours in Fig. 12 are centred on the star NX Pup. At25 µ m, the maximum contours are still centred on NX Pup butat a lower level the contours encircle the YSO. At 60 and 100 µ m, the maximum of the emission has shifted significantly tothe west of NX Pup and lies near the YSO and the maximum ofthe C O (2–1) integrated emission. The nominal point sourceposition lies between the 12-25 µ m and the 60-100 µ m max-ima.The likely explanation of the disagreement between the pointsource fluxes (hot or cold source) and the systematic shift ofthe position of the IRAS emission maximum from 12 µ m to100 µ m is that there are two sources, one warm (NX Pup) andone cold (the YSO). Spitzer 24 and 70 µ m MIPS observationsnow available online confirm the HIRES analysis. Spitzer de-tected two sources, one centered on NX Pup and another at theYSO. Their relative flux ratios (YSO / NX Pup) are ∼ µ m)and ∼
100 (70 µ m). A similar case of two infrared sources ob-served as one IRAS point source is observed in CG 12 (NGC5367). The Herbig AeBe binary star h4636 and an adjacentcold source are merged into a single source, IRAS 13547-3944(Haikala & Reipurth 2010).The IRAS 12, 25, 60, and 100 µ m point source catalog fluxesin janskys can be converted into in band fluxes by multiplyingthem with the synthesized IRAS bandpasses of 2 . × Hz,7 . × Hz, 4 . × Hz, and 1 . × Hz, respec-tively (Emerson 1988) and the source luminosity can be cal-culated. Though the Spitzer 70 µ m and the IRAS 60 µ m band-passes are not the same, it can be assumed that the contributionof NX Pup to the IRAS 60 µ m flux is not significant. Using theIRAS 07178–4429 60 and 100 µ m fluxes the FIR luminosity ofthe YSO is 3.1 L ⊙ × ( d / pc ) .There are two secondary maxima in addition to the IRASpoint source at 60 µ m in Fig. 12. These lie at the globulebright edge seen in Fig. 2. In the three-component dust model(Puget & Leger 1989), interstellar dust consists of large aromaticmolecules (PAHs), very small grains (VSGs), and the classicallarge grains. The VSGs are transiently heated and emit non-thermal emission in the mid- to FIR. In dark clouds (temper-atures < ≥ µ m. The two secondary 60 µ m maxima in Fig.12 are therefore most likely to be produced by VSGs at the sur-face of the globule and heated by the star NX Pup or the UV ra-diation from the central part of the Gum nebula. Consistent withthe HIRES 60 µ m image the globule bright edges can also be seenin the Spitzer 70 µ m image. In addition, the bright reflection neb-ulosity at the SE tip of CG 1 seen in Fig. 1 appears as a bright,extended source in the Spitzer 70 µ m image. This extended emis-sion is likely to be due to VSGs heated by NX Pup. This wouldalso agree with the argument presented in Sect. 4.1 that the ra-diation from NX Pup explains the elevated C O excitation tem-perature in C O SE. The Spitzer 70 µ m integrated flux fromC O SE is similar to that of NX Pup and does not invalidate theYSO IRAS FIR luminosity calculation presented above.
Fig. 12.
Contour maps of the HIRES processed IRAS maps su-perimposed on the SOFI K s band image. The IRAS wavelengthis shown in microns in the upper right corner of each panel.The position of the star NX Pup is marked with a filled cir-cle. The IRAS 07178–4429 point source positional uncertaintyis shown as an filled ellipse west of NX Pup in the 100 µ m panel.The C O (2-1) integrated emission contours (indicated by thicklines, 1.2 K km s − to 2.0 K km s − in steps of 0.4 K km s − ) aresuperimposed on the 12 µ m panel. The o ff set is in arcminutesfrom the centre of the SOFI image. The contour levels in MJysr − are: 12 µ m 5 to 30 by 5; 25 µ m 15 to 115 by 20; 60 µ m 3,5, 15 to 91 by 19; and 100 µ m 7, 16, 31 to 71 by 20. The HAEBE star NX Pup lies north of the bright “nose” nebu-losity and east of the obscured CG 1 globule head. Being close toCG 1, the star NX Pup dominates the local radiation field overthe general interstellar radiation field at least around the glob-ule head. Two surface brightness features in CG 1 indicate thatthe position of the star near the globule is not a projection ef-fect but that the star is physically located near the globule. It islikely that the illuminating source of the very bright reflectionnebulosity at the SE tip of CG 1 seen in Fig. 1 and the corre-sponding Spitzer 70 µ m surface brightness is NX Pup. The starmay also be the reason for the elevated C O excitation temper-ature of C O SE. The second surface brightness feature is thearc reaching NE from the YSO. The arc is symmetric with re-spect to NX Pup and is also seen in the optical images, not onlyin NIR. The arc could indicate the edge of a bubble vacated fromgas and dust by NX Pup.
In the SOFI imaging, the YSO is visible only at K s. In H and J s,it lies behind an obscuring dust lane (Figs. 4 and A.4). Thereforeits ( J - H ) and ( H - K ) colour indices cannot be calculated and itsposition in the JHK s colour-colour diagram is not known. TheYSO 2Mass K s magnitude is 13 . m
42, which is consistent with theSOFI photometry. The SOFI imaging H limiting magnitude is20 . m H - K ) colour index wouldbe & H - K ) colour index forthe SOFI o ff -field in Fig. 7 is 0 . m
15. The YSO visual extinctioncalculated with the NICE method (Lada et al. 1994) would be ∼
93 magnitudes. This is however an upper limit as the objectmost probably has an infrared excess (the extent of the YSOsource profile in the EW direction is nearly twice that of isolatedstars elsewhere in the image (Sect. 3.1). Despite the IR excess,the visual extinction towards the YSO is likely to be more than50 magnitudes. This extinction is significantly higher than themaximum extinction in the CG 1 head (9 . m
2) and the estimatedextinction in the direction of the YSO ( ∼ / core andmolecular outflow. The position observed by van Kempen et al.(2009) was however the YSO and not the nominal IRAS pointsource position. This pointing was possibly selected using theCG 1 K s image in Santos et al. (1998). van Kempen et al. (2009)found evidence of neither protostellar envelope nor a molecularoutflow and concluded that it is most likely not an embeddedYSO (class I source). This indicates that the observed high ex-tinction towards the YSO takes place in a circumstellar disk andnot in a circumstellar shell. This makes the source a likely classII source. Besides mass infall, the early stage of low-mass star formationis also associated with energetic mass outflows. Manifestationsof outflows are observed, e.g., as jets, molecular outflows andHerbig-Haro objects. The outflows inject momentum and energyinto the parent cloud and can have a highly disruptive e ff ect onthe cloud core they are forming in.Molecular hydrogen emission detected in the low-mass star-formation regions is usually associated with shocks due to hotand dense gas flowing out from the newly born stars. The H NIRline emission can also be due to UV absorption and fluorescence(Black & Dalgarno 1976). Fluorescent H line emission hasbeen observed in elephant trunks in HII regions (e.g. Allen et al.1999). The fluorescent H emission appears as a bright, thinlayer on the surface of the trunks on the side facing the ioniz-ing source. However, the ionizing source is missing in CG 1.The FUV spectrum of NX Pup is most consistent with a spec-tral type F2III (Blondel & Djie 2006). Thus the UV flux fromNX Pup cannot ionize the cloud surface. It is also not plausiblethat the early-type stars ionizing the Gum nebula would exciteonly a very localized region in CG 1. This leaves shock excitedH as the only viable option. On similar grounds, MHO 1411,which lies 90 ′′ west of the YSO, is likely to be a highly ob-scured Herbig-Haro object. The higher C O excitation temper-ature around the YSO relative to C O NW also points to anadditional energy source, probably the YSO. An indirect indica-tion of shocks is the [HCN] / [HNC] molecular abundance ratiobeing found to be larger than 1 by HSHWSP. Such a ratio canbe produced by shock chemistry.Even though the typical tracers of an outflow are present(molecular hydrogen and Br γ line emission indicating shocksand a probable heavily obscured HH object), actual outflow canbe seen neither in the HSHWSP CO velocity interval plots northe van Kempen et al. (2009) CO (3–2) spectrum in the direc- . K. Haikala et al.: Star formation in cometary globule 1: the second generation 11 Fig. 13.
APEX CO (3–2) spectra in the direction of the YSO(histogram) and in a ±
10” slice in declination (0,0 position ex-cluded) (line) and a Hanning smoothed C O (3–2) spectrumof the centre position. The temperature unit is T ∗ a . The spectrawere obtained from the ESO data archive (ESO project 077.C-0217A).tion of the YSO. The ESO data archive contains a ∼
1’ by 1 . ′ ±
10” EWslice in declination through the YSO. The 0,0 position is not in-cluded in the sum. A Hanning-smoothed C O (3–2) 0,0 positionspectrum from the same data set is also shown in the figure. Themorphology of the YSO nebulosity in the SOFI images and es-pecially the location of MHO 1411 west of the YSO implies thatthe associated molecular outflow, if present, should be orientedin the EW direction. However, there is no indication of outflowwings in the declination-slice summed-up spectrum. In contrast,it is more narrow than the spectrum in the centre position. TheC O (3–2) spectrum is centred on the CO line.It is puzzling why no molecular outflow from the YSO isdetected. If an outflow were present, it would have to be highlycollimated and exactly in the plane of the sky to remain unno-ticed.
5. Summary and conclusions
We have combined our new NIR imaging, photometry, spec-troscopy, and mm molecular line observations with already ex-isting data at other wavelengths to analyse the structure of theglobule head in greater detail than possible before. The conclu-sions are the following:1. A young stellar object (a likely class II source) with anassociated bright NIR nebulosity and a molecular hydrogen ob-ject, MHO1411, were detected in the globule head. The YSO istotally obscured in J s. The estimated optical extinction towardsthe YSO is 50 magnitudes or more. The H and Br γ line emis-sion is detected in the direction of the YSO. The H emission isalso seen in the YSO SE filament.2. The visual extinction, as estimated with the NICERmethod, in the imaged area excluding the YSO direction, rangesfrom 1 . m . m ′′ NW of the YSO. The peak N(H ) column density estimatedfrom the extinction data is 9 . × cm − and the total mass16.7M ⊙ ( d / pc ) . The mass within the A v = ⊙ ( d / pc ) .3. The C O emission in the globule head is distributed in an4 ′ by 1 . ′ ff set SW of the YSO.The C O T MB (2–1) is equal to T MB (1–0) in the NW part ofthe cloud (the Av maximum) but is significantly stronger thanin the C O (1–0) transition in the C O maximum and in theSE. This indicates that the C O excitation temperature is higherin the latter two positions. The C O centre of line velocity is3.33 ± − in the C O maximum and 3.69 ± − in SE indicating, that these are physically separate entities. It islikely that the elevated C O excitation temperature SW of theYSO is caused by the interaction of the YSO with the surround-ing cloud. The SE “nose” coincides with a bright reflection neb-ulosity south of NX Pup in the SOFI and SIRIUS images. Anelevated C O excitation temperature in the “nose” is likely be-cause of heating by NX Pup.4. The peak N(H ) column densities and masses,as calculated from the C O data, are (5 . × cm − ,0.75M ⊙ ( d / pc ) ), (7 . × cm − , 9.2M ⊙ ( d / pc ) )and (1 . × cm − , 1.6M ⊙ ( d / pc ) ) for C O SE,C O maximum, and C O NW, respectively. The columndensities and the masses are calculated from only twoC O transitions and should be considered as very roughestimates. The C O masses are lower than those calculatedfrom the extinction data. The likely reason for this is that thecloud low density envelope is not visible in C O .5. The peak integrated C O emission does not coincide withthe position of maximum visual extinction. This is primarily dueto the C O excitation temperature, which is lower in the extinc-tion maximum than in the C O emission maximum.6. The IRAS point source 07178–4429 resolves into twosources in the HIRES enhanced IRAS images. The 12 and 25 µ memission originates mainly in the Herbig AeBe star NX Puppisand the 60 and 100 µ m emission in the adjacent YSO.The resultsof the HIRES analysis are confirmed by Spitzer 24 µ m and 70 µ mMIPS mapping, which has become publicly available. The 60and 100 µ m FIR luminosity of the point source is 3.1 L ⊙ .7. Even though typical signs of YSO–parent cloud interac-tion (shocked gas, elevated C O excitation temperature, MHOobject) were detected no strong molecular outflow seems to bepresent.8. If the binary NX Pup and the T Tau star adjacent to it wereformed in CG 1, they have already cleared away the surroundingdust and molecular gas. The YSO detected in NIR in the globulehead is the second generation of stars formed in the cloud.The new NIR photometry and molecular line data presentedin this paper reveal a newly born star and a rather complex dis-tribution of interstellar material in the globule head. However,the derived cloud parameters, e.g., mass, column density, andexcitation temperature, are only rough estimates. Further obser-vations are necessary to refine the cloud model. Mapping of thecloud head in density sensitive molecules and in both FIR and(sub)mm continuum is needed to have su ffi cient input parame-ters to run a radiative transfer code like e.g. the one presentedin Juvela (1997). Additional NIR imaging and spectroscopy areneeded to provide information on the nature of the YSO and thesurrounding nebulosities. Acknowledgements.
It is a pleasure to thank the NTT team for the support dur-ing the observing run. This research has made use of the SIMBAD database,operated at CDS, Strasbourg, France, and of NASA’s Astrophysics Data SystemBibliographic Services.This work is based in part on archival data obtained with the Spitzer Space Telescope, which is operated by the Jet Propulsion Laboratory,California Institute of Technology under a contract with NASA. M.M acknowl-edges the support from the University of Helsinki Senat’s graduate studies grantand from the Vilho, Yrj¨o and Kalle V¨ais¨al¨a Fund.
References
Allen, L. E., Burton, M. G., Ryder, S. D., Ashley, M. C. B., & Storey, J. W. V.1999, MNRAS, 304, 98Ascenso, J., Alves, J., Vicente, S., & Lago, M. 2007, A&A, 199Auman, H., Fowler, J., & Melnyk, M. 1990, AJ, 99, 1674Bernacca, P. L., Lattanizi, M. G., Bucciarelli, B., et al. 1993, A&A, 278, L47Bertin, E. & Arnouts, S. 1996, A&AS, 117, 393Bessell, M. & Brett, J. 1988, PASP, 100, 1134Black, J. H. & Dalgarno, A. 1976, ApJ, 203, 132Blondel, P. F. C. & Djie, H. R. E. T. A. 2006, A&A, 456, 1045Bohlin, R. C., Savage, B. D., & Drake, J. F. 1978, ApJ, 224, 132Bourke, T., Hyland, A., Robinson, G., James, S., & Wright, C. 1995, MNRAS,276, 1067Brand, P. W. J. L., Hawarden, T. G., Longmore, A. J., Williams, P. M., &Caldwell, J. A. R. 1983, MNRAS, 203, 215Brandner, W., Bouvier, J., Grebel, E. K., et al. 1995, A&A, 298, 818Davis, C. J., Gell, R., Khanzadyan, T., Smith, M. D., & Jenness, T. 2010, A&A,511, A24 + Emerson, J. P. 1988, in NATO ASIC Proc. 241: Formation and Evolution of LowMass Stars, ed. A. K. Dupree & M. T. V. T. Lago, 21–Foster, J. B., Rom´an-Z´u˜niga, C. G., Goodman, A. A., Lada, E. A., & Alves, J.2008, ApJ, 674, 831Franco, G. A. P. 1990, A&A, 227, 499Frerking, M. A., Langer, W. D., & Wilson, R. W. 1982, ApJ, 262, 590Haikala, L. K., Juvela, M., Harju, J., et al. 2006, A&A, 454, L71Haikala, L. K. & Reipurth, B. 2010, A&A, 510, A1 + Harju, J., Sahu, M., C., H., et al. 1990, A&A, 233, 197Hawarden, T. & Brand, P. 1976, MNRAS, 175, 19PJuvela, M. 1997, A&A, 322, 943Lada, C., Lada, E., Clemens, D., & Bally, J. 1994, ApJ, 429, 694Lombardi, M. 2005, A&A, 438, 169Lombardi, M. & Alves, J. 2001, A&A, 377, 1023Malbet, F. 2007, in IAU Symposium, Vol. 243, IAU Symposium, ed. J. Bouvier& I. Appenzeller, 123–134Nagayama, T., Nagashima, C., Nakajima, Y., et al. 2003, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, Vol. 4841,Society of Photo-Optical Instrumentation Engineers (SPIE) ConferenceSeries, ed. M. Iye & A. F. M. Moorwood, 459–464Persson, S. E., Murphy, D. C., Krzeminski, W., Roth, M., & Rieke, M. J. 1998,AJ, 116, 2475Puget, J. L. & Leger, A. 1989, ARA&A, 27, 161Reipurth, B. 1983, A&A, 117, 183Sanduleak, N. 1974, Information Bulletin on Variable Stars, 880, 1Santos, N., Yun, J., Santos, S., & Marreiros, R. 1998, A&A, 116, 1376Smith, M. D. 1995, A&A, 296, 789van der Tak, F. F. S., Black, J. H., Sch¨oier, F. L., Jansen, D. J., & van Dishoeck,E. F. 2007, A&A, 468, 627van Kempen, T. A., van Dishoeck, E. F., Hogerheijde, M. R., & G¨usten, R. 2009,A&A, 508, 259Warin, S., Benayoun, J. J., & Viala, Y. P. 1996, A&A, 308, 535Zinnecker, H., Krabbe, A., McCaughrean, M. J., et al. 1999, A&A, 352, L73 . K. Haikala et al.: Star formation in cometary globule 1: the second generation , Online Material p 1
Appendix A: Sofi photometry
Fig. A.4.
SOFI K s (in black) and H (in red) surface brightnesscontours superimposed on a grey scale J s image. The contourlevels in SOFI counts are from 100 to 1100 in steps of 100 in K sand from 10 to 90 in steps of 10 in H . The wedge in the rightshows the J s grey scale levels. The image size is 23” by 23” andSOFI pixel scale is 0 . ′′ Appendix B: C O mapping Appendix C: ESO archive data Fig. C.1.
CO (3–2) map in the direction of the YSO. The spectraare plotted from 1 km s − to 7km s − in velocity and from -1 K to12 K in T ∗ a scale. The o ff sets are in arcseconds in right ascensionand declination from the YSO position. The data is ESO archivedata (ESO project 077.C-0217A) . K. Haikala et al.: Star formation in cometary globule 1: the second generation , Online Material p 2
Fig. A.1.
SOFI J s image of CG 1 head. Square root scaling has been used to enhance the appearance of the faint surface brightnessstructures. . K. Haikala et al.: Star formation in cometary globule 1: the second generation , Online Material p 3
Fig. A.2.
SOFI H image of CG 1 head. Square root scaling has been used to enhance the appearance of the faint surface brightnessstructures. . K. Haikala et al.: Star formation in cometary globule 1: the second generation , Online Material p 4
Fig. A.3.
SOFI K s image of CG 1 head. Square root scaling has been used to enhance the appearance of the faint surface brightnessstructures. . K. Haikala et al.: Star formation in cometary globule 1: the second generation , Online Material p 5
Fig. B.1.
Contour map of the integrated C O (1–0) emission in CG 1 head in T MB scale. The contour levels are from 0.25 K km s − to 1.25 K km s − in steps of 0.25 K km s − . The points indicate the measured points and the square the location of the YSO in thefigure. The SEST HPBW at the C18
Contour map of the integrated C O (1–0) emission in CG 1 head in T MB scale. The contour levels are from 0.25 K km s − to 1.25 K km s − in steps of 0.25 K km s − . The points indicate the measured points and the square the location of the YSO in thefigure. The SEST HPBW at the C18 O (1–0) frequency is shown in the upper left corner. . K. Haikala et al.: Star formation in cometary globule 1: the second generation , Online Material p 6
Fig. B.2.
Contour map of the integrated C O (2–1) emission in CG 1 head in T MB scale. The contour levels are from 0.4 K km s − to 2.0 K km s − in steps of 0.4 K km s − . The points indicate the measured points and the square the location of the YSO in thefigure. The SEST HPBW at the C18