Bipolar outflow sources in the Serpens core: SVS 2 and SVS 20

1997MNRAS.290..598H

Mon. Not. R. Astron. Soc. 290, 598-606 (1997)

Bipolar outflow sources in the Serpens core: SVS 2 and SVS 20

T. L. Huard,l* D. A. Weintraub1* and J. H. Kastner2* 1 Department of Physics and Astronomy, Vanderbilt University, Box 1807 Station B, Nashville, TN 37235, USA 2M1T Center for Space Research, 37-667a, Cambridge, MA 02139, USA

Accepted 1997 April3. Received 1997 March 13; in original form 1996 October 30

1 INTRODUCTION

ABSTRACT Polarimetric maps of the Serpens Reflection Nebula (SRN) obtained at H (1.65 J.1m) and K (2.20 J.1m) show centro symmetric patterns of polarization vectors centred about each of the proto stars SVS 2 and SVS 20. These patterns indicate that evacuated, bipolar cavities surround each source. Such bipolar structures suggest the presence of outflows. While the polarization discs are not apparent in these maps, the position angles of the discs can be inferred from the elliptically symmetric morphology of the polarization pattern. We determine these position angles to be 400 ± 10° for SVS 2 and 1400 ± 10° for SVS 20. Three 'knots' of emission are evident in a narrow-band image of the SRN taken at the /I = 1-+0 S(I) transition of molecular hydrogen. The association of these knots with the projected directions of the polar axes of the SVS 2 and SVS 20 outflow cavities suggest the knots are collisionally excited by outflows from these two protostars.

We obtained high spatial resolution images of the SVS 20 binary system through four narrow-band filters at and near the 3.1-J.1m H20 ice absorption feature. The northern (SVS 20N) and southern (SVS 20S) components of the 1.6-arcsec separation binary are well separated in these images. We measured the intensity of SVS 20S relative to that of SVS 20N using a small 1.1 x 1.1 arcsec2 synthetic aperture and a larger 1.1 x 8.1 arcsec2

synthetic aperture oriented east-west. We found that the relative intensity was smaller through the larger aperture for all of the narrow-band filters except the 3.1-J.1m filter. These data imply that ice must exist within the local circumstellar environment of this binary.

Images of the SRN reveal that the cometary nebula EC 81 has faded by at least two magnitudes at H and K between 1989 and 1994. The rapid and dramatic drop in the nearinfrared brightness of EC 81 suggests that this protostar may be in a post-FUor or post-EXor outburst phase. Based on recently presented evidence for several other flaring or fading episodes over the last two decades for other protostars in or near the SRN, we conclude that such FUor or EXor activity may be fairly common in the Serpens star-forming regions.

Key words: techniques: polarimetric - circumstellar matter - stars: pre-main-sequence -ISM: individual: Serpens Reflection Nebula - ISM: jets and outflows - infrared: stars.

The Serpens molecular cloud is a very active site of star formation. Recent work at both near-infrared (Eiroa & Casali 1992) and subrnillimetre wavelengths (Casali, Eiroa & Duncan 1993) reveals that the Serpens core contains dozens of young stars and multiple molecular outflows (White, Casali & Eiroa 1995). The first measurements of the CO J = 1--0 transition toward the Serpens nebula, made at low angular resolution, showed a high-velocity molecular outflow from the core (Bally & Lada 1983). More

recently, White et al. obtained higher resolution observations of the CO, l3CO, Cl80 and C170 J = 2--1 and the CO J = 4-+3 transitions to identify several outflows in the Serpens nebula. They associate several subrnillimetre sources (SMMI = FIRS 1, SMM2, SMM3, SMM4, SMM8 and possibly SMM9) within the Serpens core with outflows. However, neither of the two brightest nearinfrared protostars in the Serpens core, SVS 2 (Strom, Vrba & Strom 1976) (located near the centre of the Serpens Reflection Nebula, or SRN) and SVS 20 (located at the southern edge of the SRN), is unambiguously associated with any of the identified outflows.

*E-mail: [email protected] (TLH); [email protected] (DAW); [email protected] (JHK)

In a 0.95-J..Lm polarization map of the SRN (Gomez de Castro, Eiroa & Lenzen 1988), a centrosymmetric pattern of polarization

© 1997RAS

© Royal Astronomical Society • Provided by the NASA Astrophysics Data System

Dow

nloaded from https://academ

ic.oup.com/m

nras/article/290/4/598/981907 by guest on 24 June 2022

http://adsabs.harvard.edu/abs/1997MNRAS.290..598H

1997MNRAS.290..598H

vectors indicates that this nebula is illuminated by SVS 2. Furthermore, a comparison of polarization models (Whitney & Hartmann 1993) to the relatively high degree of polarization and morphology in the map suggests that SVS 2 is surrounded by an evacuated, bipolar cavity. This, in turn, suggests that SVS 2 may be associated with an outflow.

At 2.2 fLm (Kband), the extended emission near the binary source SVS 20 appears as an arc-shaped tail trailing toward the south-west (Eiroa & Casali 1992). Eiroa & Casali suggest that the tail could be an outflow from the source or it could be gas swept up by the orbiting binary. Similarly shaped nebulae are observed around a few other sources including SSV 19 in NGC 2068 and IRAS 04248+2612 in Taurus. Polarimetric measurements of SSV 19 and IRAS 04248+2612 show centrosymmetric patterns with polarization discs suggesting the presence of outflows (Weintraub, Kastner & Lowrance 1994).

The SVS 20 binary, having a separation of only 1.6 arcsec, is a good place to probe for H20 ice within the circumstellar environments of pre-main-sequence binaries. Eiroa & Leinert (1987) presented one-dimensional, near-infrared, speckle observations in the H20 ice feature and found that the intensity ratios of SVS 20S relative to SVS 20N were 3.4::!::0.1 and 3.3::!::0.1 at 3.1 and 3.5 fLm, respectively. Because these ratios are statistically indistinguishable, Eiroa & Leinert argued that the 3.1-fLm absorption was due to icecarrying grains in the foreground, within the Serpens cloud. However, their measurements were taken with an 8 x 0.6 arcsec2 slit oriented east-west. Such measurements could be affected by reflected light and, if the dust properties around SVS 20N and SVS 20S are different, the ratios would depend upon the amount of reflected light as well as the levels of ice absorption. For example, Kastner & Weintraub (1996) demonstrated that the H20 ice optical depths along the line of sight to the protostar AFGL 2136 IRS 1 were greater than those along the line of sight to each of the lobes of the surrounding Juggler reflection nebula.

In order to investigate SVS 2 and SVS 20 in more detail and search for additional evidence of outflow activity, we have obtained polarimetric images atH(1.65 fLm) and Kbands supplemented with images in the v = 1- 0 S(l) transition of molecular hydrogen at 2.122 fLm. To revisit the question as to whether there is circumstellar H20 ice in the SVS 20 system in a way that avoids possible scattered light problems, we also have obtained high-resolution shift-and-add images of SVS 20 in four narrow-band filters with central wavelengths of 3.1,3.3,3.6 and 4.0 fLm.

2 OBSERVATIONS

We used the National Optical Astronomy Observatories 1 nearinfrared Cryogenic Optical Bench (COB) mounted at the fl15 Cassegrain focus of the Kitt Peak National Observatory (KPNO) l.3-m telescope to obtain images of the Serpens core on 1994 May 27 UT. The entire COB filter assembly and associated baffling is cooled to 77 K and the light-tight detector assembly is cooled to 40 K. The out of band signal and internal background have been reduced to negligible levels. Detector dark current is typically limited by self-luminosity within the focal plane readout itself. For these observations, the detector was an SBRC 256 x 256 InSb hybrid focal plane array with 30 x 30 p.m2 square pixels, 25 electrons read noise, and dark current of < 1 electron s -I (Fowler et al.

1 National Optical Astronomy Observatories is operated by Associated Universities for Research in Astronomy, Inc., for the National Science Foundation.

© 1997 RAS, MNRAS 290, 598-606

Bipolar outflow sources in the Serpens core 599

1994). The pixel scale was 0.93 arc sec, as determined astrometrically from images of the OMC 1/0MC2 cloud. The seeing was ~2.5 arcsec full width half maximum at both H and K.

For measurements of the near-infrared polarization, we used broad-band H and K filters in combination with a cooled analyser and warm half-wave plate. An image suite at each waveband, from which we obtained polarimetry, consists of four images obtained at 22~5 intervals of the half-wave plate. This sequence modulates the position angle of polarization incident on the detector at intervals of 45°, allowing for straightforward reconstruction of the Stokes parameters (Weintraub et al. 1992). We obtained three suites of data at K, where each image included in the suites represents an exposure time of 1 min. For the images of one of these suites, this I-min exposure time was obtained by co-adding 60 1-s exposures. For the images of the other two suites, three 20-s exposures were co-added. These three suites were aligned, weighted by the exposure times, and combined to produce the working K-band suite. The corresponding images of a suite of lOco-added 2-s exposures and a suite of six co-added 1O-s exposures were aligned, weighted by the exposure times, and combined to produce the working H-band suite. Most of the pixels within 2 arcsec of SVS 2 and SVS 20 were intentionally saturated in the longer exposure suites at K and H. The saturated pixel values were replaced with the unsaturated values of the corresponding pixels in the shorter exposure suites. Polarimetric maps were constructed from the working K- and H-band suites via a formalism described in Weintraub et al. and by binning the data into 3 x 3 pixel bins. In order to determine the absolute polarization position angles, we observed the highly polarized source IRC+10216 (Kastner & Weintraub 1994).

The total intensity (i.e., intensity measured without an analyser) at a given location can be reconstructed from the intensities measured with an analyser positioned at the four Stokes angles (0°,45°,90°, 135°) according to

1 _ 10 +145 +/90 +/135 total - 2

Photometry of sources in the SRN was obtained from these total intensity H and K images. In order to transform to the standard magnitude system, the standard stars Gl 748 and HD 106965 were observed. Across most of the field, the limiting magnitudes (i.e., magnitudes corresponding to the 30" level above the background) are ~ 15.5 and ~ 15.1 at H and K, respectively. However, the limiting magnitudes may be as much as a magnitude smaller in regions with bright nebulosity.

We obtained images using a I per cent interference filter centred at 2.122 fLm, the wavelength of the v = 1-0 S(I) line of H2, to search for evidence of shocked gas. In an environment like the SRN, any such gas probably would have been shock heated by an outflow from a nearby protostar. The level of continuum emission in this filter was determined from images obtained using a 1 per cent interference filter centred at 2.14 fLm. Because these narrow-band filters have slightly different effective bandwidths, we rescaled the 2. 14-fLm images to the same effective transmission efficiency as the 2.122-fLm images using field stars to calibrate the relative filter bandwidths.

On 1996 March 31 UT, using the NOAO 4-m telescope on Kitt Peak, we again used COB. This time we made observations in the new diffraction-limited infrared imaging (DLIRIM) configuration of COB (see Weintraub et al. 1996). During this experiment, we obtained near-infrared images of SVS 20 through four narrow-band filters. The wavelengths of peak transmission (and half-power


Dow


ic.oup.com/m



1997MNRAS.290..598H

600 T. L. Huard and D. A. Weintraub

SRN : H Band SRN : K Band

40 40

r--. 20

r--. 20 u u

Q) Q) (fl (fl

u U l- I-

0 0 0 0 '-'

a:; a:; !' -20 !' -20 0 0 U -40 U -40 Q) Q)

0 0

-60 -60

-80 - 80 80 60 40 20 0 -20 -40 80 60 40 20 0 -20 -40

RA Offset (arcsec) RA Offset (arcsec)

SRN: 2.122 J-Lm SRN: 2.1 4 J-L m

40 40

r--. 20

r--. 20 u u

Q) Q) (fl (fl

~ U I-

0 0 0 0 '-' '-'

a:; a:; !' -20 !' -20 0 0 U -40 U -40 Q) Q)

0 0

-60 -60

-80 -80

80 60 40 20 0 -20 -40 80 60 40 20 0 -20 -40 RA Offset (arcsec) RA Offset (arcsec)

Figure 1. Images of the SRN at H, K, 2.122 f1.m and 2.14 f1.m. The origin in allimages has been placed at the position of SVS 2 (Epoch 1950: RA = 18h27m24~5, Dec. = 1°12'41"). The near-infrared peak associated with SVS 20 is located at position (12", -41"). The H and Kimages are displayed on a logarithmic scale, while the 2. 122-f1.m and 2.14-f1.m images are scaled to the same effective bandwidth and displayed on a linear scale. In the Hand K images, the hourglass-shaped nebula around SVS 2 is visible with dark lanes separating the north-western and south-eastern regions. A comparison of the 2. 122-f1.m and 2. 14-f1.m images reveals three regions of emission from shocked molecular hydrogen (identified by arrows).

bandwidths) for our four filters were 3.072lLm (0.096ILm), 3.305 ILm (0.074ILm), 3.592 ILm (0.078ILm) and 3.990 ILm (0.052ILm). Images are shifted and added, using the peak pixel within a preselected 5.0 x 5.0 arcsec2 field on the array. No image quality criteria are applied before an image is co-added into the final frame. A bad-pixel mask is applied to each raw image before implementing the shift-and-add step. No fiat-field is applied to any raw or final images; however, the shift-and-add process effectively flat-fields the final image. All of these processing steps are done in real time. Each DLIRIM image collected for SVS 20 is the shift-and-add summation of 256 such 50-ms images. Three DLIRIM images were obtained, aligned, co-added and normalized for the final image at 3.1 ILm. Two DLIRIM images were similarly processed at each of 3.3, 3.6 and 4.0 ILm. From observations of the binary system 'Y And AB, we established that the plate scale was 0.100::!:: 0.001 arc sec pixel-I .

3 RESULTS AND DISCUSSION

In Fig. 1, we present H and K images of the -2 x 2 arcmin2 region approximately centred on the SRN. The morphology of the faint nebulosity at both H and K resembles a butterfly shape extending as far north as EC 93 [object 93 from Eiroa & Casali 1992; located at position (15/1, 45/1) in our maps] and as far south as EC 117 (55/1, -67"). A brighter, hourglass-shaped nebula is centred on SVS 2 (0", 0"). The dark lanes, located on either side of the neck of the hourglass-shaped nebula and oriented at a position angle of -40°, are barely visible at H, but are clearly apparent at K. The double star SVS 20 is unresolved in the near-infrared and is surrounded by an arc-shaped nebula located at position (12", -41"). At 0.911Lm this arc-shaped nebula is invisible and the SRN is illuminated predominantly by SVS 2 (see the 0.91-lLm photograph in fig. 1 of Gomez de Castro et al. 1988). The nebulosity surrounding SVS 2

© 1997 RAS, MNRAS 290, 598-606 © Royal Astronomical Society • Provided by the NASA Astrophysics Data System

Dow


ic.oup.com/m



1997MNRAS.290..598H

still dominates the SRN at H, but the arc-shaped nebula is clearly a major component of the SRN. At K, however, the hourglass-shaped and arc-shaped nebulae are comparably bright features of the SRN. Furthermore, in both the H and K images, two narrow arms can be observed arching away from the arc-shaped nebula. From the southwestern edge of this nebula, an arm extends as far as 15 arcsec south. From the north-eastern edge, a fainter arm extends up to the region directly between EC 105 (34", -39") and EC 103 (28", - 21"). In the K image, the southern arm is brighter and broader than at H. The north-eastern arm also is brighter at K, while a third arm is visible lying just north of and more-or-Iess parallel to this northeastern arm.

Eiroa & Casali (1992) discovered a cometary nebula located 75 arcsec directly south of SVS 2 in their K image (BC 81: see their fig. 1). The most striking difference between our near-infrared observations and those published by Eiroa & Casali is the absence in our images ofEC 81, which should be located at (0", -75''). They determined an H magnitude of 12.8 and K magnitude of 12.2 for this object. In this region of our image, we find a limiting magnitude of 15.3 at H and 14.4 at K. These data imply that EC 81 faded by more than two magnitudes between the years 1989 and 1994. Other sources exhibiting large-amplitude, short-time-scale, near-infrared variability typical of very young, embedded stars have been observed in the Serpens north-west star-forming region located just north-west of the field imaged by us. Hodapp et al. (1996) have discovered an outburst from a deeply embedded star. This star has brightened by over 4.5 mag between 1994 August and 1995 July. In addition, the cometary-shaped nebula EC 53, located just 18 arcsec south and 29 arcsec west of the flaring star, faded by a couple of magnitudes at Kbetween 1994 August and 1995 December, and has since brightened by the same amount. Evidence for such variability in the SRN also exists. Strom et al. (1976) noted that SVS 20 (it was not known to be a binary until 11 years later: Eiroa et al. 1987) faded by 1.5 mag at K between 1974 and 1975. The variability of these sources is attributed to instabilities in disc accretion observed in other very young, embedded pre-main-sequence stars and is used to explain the flaring ofFU Orionis and EX Lupi stars. Therefore, it is possible that the fading of EC 81 may be a result of a reduction in disc accretion, perhaps following an FUor- or EXor-type outburst. Alternatively, it is also possible that this fading results from temporally variable obscuration near EC 81, rather than an inherent change in the luminosity of the embedded source itself.

For comparison with EC 81, we examined the published nearinfrared photometry ofV1057 Cyg (Gezari, Schmitz & Mead 1987; Kenyon & Hartmann 1991), an FU or that brightened in 1969-1970 and since then has faded by -3.5 mag at B (see light curve in fig. 1 of Bell et al. 1995). Since its outburst, V 1057 Cyg has faded by just over a magnitude at K. While V1057 Cyg is one of the more rapidly fading FUors, it faded much more slowly than EC 81. However, the post-outburst fading of FUors is not a good defining characteristic of this group of stars since it varies considerably from protostar to protostar.

A close comparison of our K image with that of Eiroa & Casali (1992) reveals no other sources that have notably faded or brightened. Our photometry agrees to within the measurement error for all sources for which Eiroa & Casali report magnitudes. A bright source can be seen 55 arcsec east and 47 arcsec south of SVS 2 in our K image as well as in the image of Eiroa & Casali. However, this source did not appear in their table of near-infrared magnitudes. From our observations, we determine an H magnitude of 16.5 ± 0.2 and a K magnitude of 13.7 ± 0.2. Thus, this source represents one of the reddest objects in this region.

© 1997 RAS, MNRAS 290, 598-606


We also present the 2.122- and the 2. 14-lLm images of the SRN in Fig. 1. The arrows in the 2. 122-lLm image indicate regions where we have unambiguously detected H2 emission. The regions of H2 emission ('knots') are located at (31", -33"), (37", -50") and (2", -53"). Note that all three knots are clearly evident in comparing the 2.122- and 2. 14-lLm images before scaling these images to the same effective transmission efficiency as described in Section 2. We will discuss the significance of these knots after presenting the polarimetry results.

We present H and K polarization maps of the SRN in Figs 2 and 3, respectively. The centrosymmetric pattern associated with SVS 2 at Hand K is similar to that observed at 0.95 ILm (Gomez de Castro et al. 1988). The polarization amplitudes around SVS 2 are generally between 15 and 30 per cent at H, but reach as high as 45 per cent. These amplitudes are typically between 25 and 45 per cent at K, with some as high as 65 per cent. The fact that these amplitudes are so large and comparable to the values of 30-40 per cent observed at 0.95 ILffi implies that the scattering efficiency is effectively independent of wavelength for A < 2.2ILm. This is a characteristic of dust grains comparable in size to (or larger than) the wavelength (i.e., with a ~ Al2-rr where a is the grain size). We therefore conclude that the dust population within the SRN must include a population of large grains. Furthermore, the morphology of the polarization map and the high polarization amplitudes indicative of single scattering suggest that SVS 2 is illuminating an evacuated, bipolar cavity. This, in tum, suggests the presence of a bipolar outflow from SVS 2.

A comparison of our maps with the models of Whitney & Hartmann (1993) reveals that the polarization pattern and intensity contours surrounding SVS 2 show a remarkable similarity to Models 7 and 8 in their fig. 3(a). These two models exhibit the broad hourglass-shaped morphology for the intensity contours and the same elliptically symmetric polarization pattern. Models 7 and 8 assume an edge-on disc and streamline outflow with an opening angle of 30°. The disc axis, inferred from the elliptical morphology of the polarization pattern, is determined to be at a position angle of 400 ± 10°. Because the models are calculated for a much smaller spatial scale than our observations as well as the large number of uncertain parameters involved in such a calculation, a comparison of our results with these models is difficult and should only be used to provide a qualitative understanding of the SVS 2 environment.

At wavelengths at which a circumstellar disc is optically thick, a band oflow polarization (the classical 'polarization disc') is usually observed and attributed to multiple scattering (Whitney & Hartmann 1993). For the case of SVS 2, the band of low polarization surrounding SVS 2 in the 0.95-lLm map is not visible in our H and K polarimetry. We offer two possible explanations for this observation. In addition to the large grain popUlation within the SRN, the neck of the SRN may contain a population of small grains. If this is the case, the neck region would be optically thick for the 0.95-lLm photons due to both grain populations. However, the smaller grain population would not be optically thick for the near-infrared H and K photons. Thus, any H and K polarization seen in this region would be due to single scattering from the large grain population as is the case in the other regions of the SRN. A second explanation is that the neck region may be optically thick for the Hand K photons, just as it is for the 0.95-lLm photons. Yet, because our maps have a resolution three times lower than the 0.95-lLm map, in any given pixel along the disc axis the effect of the multiple scattering in the thin disc is completely dominated by the single scattering from the envelope boundary above and below the polarization disc.


Dow


ic.oup.com/m



1997MNRAS.290..598H


40

(0

20

0

U 0 (I) VI

~ ~ Q) ~ (5

-20 U (I)

0

-40

-60

60

Band

40 20 0 RA Offset (orcsec)

- 30% Polarization • Knot 1 '" Knot 2 • Knat 3

-20 -40

Figure 2. Polarization map with intensity contours of the SRN at H. The origin has been placed at the position of SVS 2. The near-infrared peak associated with SVS 20 is located at position (12", -41"). The first contour is at the 30- level (0.015 m1y arcsec-2), and subsequent contours are plotted at I-mag increments above this level. Overlaid on this map are the positions of the three knots of H2 emission. Note the positions of these three knots relative to the polar axes of SVS 2 (at position angle 130°) and SVS 20 (at position angle 50°).

Entirely invisible at A < 1 f.lm the arc-shaped nebula surrounding SVS 20 is very prominent at 1.6-2.2 f.lm. Thus, the H and K polarimetry provides information about this region that was absent in the 0.95-f.lm polarimetry. In the nebulosity around SVS 20, we find typical polarization amplitudes of 5-25 per cent at H, and peaking around 35 per cent. Typical amplitudes at K range from 10-30 per cent, with amplitudes as high as 40-50 per cent in the outer portion of the north-eastern lobe and within the southwestern tail. Although the polarization is weaker than that around SVS 2, a similar elliptically symmetric morphology is observed around SVS 20 suggesting that it, too, is an outflow candidate. Again, no low-polarization band is obviously visible. However, the position angle of the disc, 1400 ± 10°, can be inferred from the major axis of the elliptical pattern. The central axis of the reflection lobes is perpendicular to the disc axis except in the most remote areas of the south-western lobe. In this region, the polarization position angles tend to stay more-or-Iess perpendicular to the arcing tail.

It is especially noteworthy that the H2 emission region designated as Knot 1 is found almost exactly along the south-western line of polar symmetry of the polarization map associated with SVS 20 and the arc-shaped nebula. If, in fact, the polarization map indicates scattering from an evacuated, polar lobe generated by an outflow, it is likely that this H2 knot is collisionally excited by material ejected in that outflow.

The proximity of Knots 2 and 3 to the bright source EC 105 suggest that perhaps this source is the origin of the outflow exciting these knots.

However, H20 ice measurements (Eiroa & Casali 1992), polarimetric measurements at 0.95 f.lm (Gomez de Castro et al. 1988), and our nearinfrared polarimetry all demonstrate that EC 105 is not associated with the Serpens nebula. Instead, as with Knot 1, the region of H2 emission designated as Knot 2 lies along the line of polar symmetry (oriented perpendicular to the disc axis) of the polarization map associated with SVS 2. The close association between the position of this knot and the polar axis of SVS 2 suggests that Knot 2 is collisionally excited by an outflow from SVS 2. The particular outflow source exciting Knot 3 is ambiguous since it lies along both the south-eastern polar axis ofSVS 2 and the north-eastern polar axis of SVS 20. Furthermore, it is found near the interface between the south-eastern lobe of the bipolar cavity associated with SVS 2 (as outlined by the outer edges of the polarization pattern associated with SVS 2) and the north-eastern lobe of the SVS 20 reflection nebula (as outlined by the outer edges of the polarization pattern associated with SVS 20). However, the arm extending from SVS 20 to the position of Knot 3 may provide a slightly stronger association of this knot with SVS 20. Note that a similar arm extends towards the south-west to Knot 1 from this same source.

To learn more about the mass distribution surrounding SVS 2, especially in the nebula, we used the H and K total intensity images to make a map of the optical extinction A" for the SRN. This map was generated by assuming

A" = 15.5 x [(H - K)observed - (H - K)o]

© 1997 RAS, MNRAS 290, 598-606


Dow


ic.oup.com/m



1997MNRAS.290..598H

40

20

,-.,. u 0 Q) III

~ ~

w 2 (5

-20 U Q)

0

-40

-60

60 40


SRN: K Bond

o

20 0

0

v

- 30% Polarization • Knot 1 '" Knot 2

.• Knat 3

-20 -40 RA Offset (arcsec)

Figure 3. Polarization map with intensity contours of the SRN at K. The origin has been placed at the position of SVS 2. The first contour is at the 311 level (0.0l3 mly arcsec -2), and subsequent contours are plotted at I-mag increments above this level. Overlaid on this map are the positions of the three knots ofR2 emission. Note the positions of these three knots relative to the polar axes of SVS 2 and 20.

where an intrinsic colour index of(H - K)o = 0.76 (corresponding to a blackbody temperature of 2100 K) was assumed for SVS 2 (Churchwell & Koornneef 1986). This calculation assumes that the scattering nebula sees unreddened starlight. This is reasonable since the very high polarization levels indicate that most of the light we see is dominated by single scattering. In other words, starlight is unimpeded in the evacuated bipolar cavity until it scatters into our line of sight. Thereafter, it suffers some extinction travelling toward us through the local nebula. We present the map of optical extinction in Fig. 4. That portion of the SRN illuminated by SVS 2 shows a fairly constant and low extinction (A.,= 1) as would be expected from a bipolar envelope in which single scattering dominates. Furthermore, an elongated region of higher extinction is observed in the neck of the bipolar structure surrounding SVS 2 presumably due to higher line-of-sight extinction in a circumstellar disc. It is important to note that this elongated, disc-like morphology is still evident even after considering that the regions of apparent higher extinction just 10 arcsec north-east and 10 arcsec south-west of SVS 2 are due toa point source at position (8",4") (see Fig. 1) and an artefact (three adjacent, black pixels) of the reduction process, respectively.

In general, one should be careful in interpreting the extinction map since we assumed an intrinsic colour index for SVS 2 in order to produce it. Those regions in the map illuminated by other sources, most notably the nebulous region around SVS 20, may contain an offset in Av depending upon the difference between the intrinsic colour index of the local source and that assumed for

© 1997 RAS, MNRAS 290,598-606

SVS 2. However, Churchwell & Koornneef(1986) also suggest that SVS 20 has an intrinsic colour index of (H - K)o = 0.76. Thus, the map may be accurate in the vicinity of SVS 20. The arc-shaped nebula surrounding SVS 20 has a fairly large extinction of A., >10. Also, note that the two north-eastern arms extending from SVS 20 are visible, although barely above the local background Av level, in the extinction map. EC 105, the object with the largest H20 ice optical depth in the SRN region (Eiroa & Casali 1992), is located at position (57", 32") and has an optical extinction A., >10 in our extinction map.

To gain further insight into the distribution of the dust around the SVS 20 binary, we present narrow-band DLIRIM images at 3.1,3.3,3.6 and 4.0 fLm in Fig. 5. These images are centred on the brighter component SVS 20S, while SVS 20N can be seen just 1.6 arcsec north-north-east of SVS 20S. Using 1.1 x 1.1 arcsec2

synthetic apertures centred on each of SVS 20S and SVS 20N, we find the intensity of SVS 20S relative to SVS 20N to be 2.51±0.05, 2.89±0.05, 2.99±0.03 and 2.87±0.03 at 3.1, 3.3, 3.6 and 4.0 iJ.m, respectively. Because of the small, apparent separation of SVS 20S and SVS 20N, the two components are observed through columns of molecular cloud material that are almost certainly nearly identical in column density and composition. Thus, the measurements should depend primarily upon three factors: the column density of interstellar ice present along the line of sight to the binary, the relative abundances of ice within the local circumstellar environments of SVS 20N and SVS 20S, and the relative temperatures of the two protostars.


Dow


ic.oup.com/m



1997MNRAS.290..598H


SRN : Extinction Ma p

.... • I ... ...... ..., 40 ..... .:t. 1'. 1

.. "r:.~·i( .~ =- -.:. . . .. ,=-

20 . ~ ..... .r , 0 -----.

-J t~'-u r • .. .1 •• I' - -:. (j) •• -1 .........

5 (f) -. "" ." u : ,-1 L 0 0 I. •

'I". ... -' - . .. ... +-' .. ' .. ...

10 > Q)

-.~ ; 4:

~ .. 0 -20 U 15 Q)

0

-40 '. >20

It • I.

-60 r

r-'

80 60 40 20 0 -20 -40 RA Offset (arcsec)

Figure 4. A map of the optical extinction in the SRN. The origin has been placed at the position of SVS 2. Note the elongated, disc-like region of high Av centred on SVS 2 in the neck of the hourglass-shaped nebula.

Eiroa & Leinert (1987) argued from near-infrared, speckle observations that the H20 ice absorption seen towards SVS 20 was due to ice-carrying grains within the Serpens cloud rather than the circumstellar environments of SVS 20S or SVS 20N. As explained in Section 1, these measurements may be affected by reflected light and, thus, may not offer a clear interpretation of relative abundances of ice near SVS 20S and SVS 20N.

In order to investigate this assertion further, we repeated our measurements above with 8.1 x 1.1 arcsec2 synthetic apertures, oriented east-west, similar to the measurements ofEiroa & Leinert. We found the intensity ratios to be 2.40±0.04, 2.08±0.03, 2.16±0.02 and 2.17±0.02 at 3.1,3.3,3.6 and 4.0 j.Lm respectively. The change in the intensity ratios from -3.4 in 1986 (determined by Eiroa & Leinert) to -2.2 in 1996 (from our large-aperture measurements) can be attributed to a decrease in brightness of SVS 20S relative to SVS 20N by -0.5 mag over 10 years. Such a change is consistent with our knowledge that one (or both) of the components of SVS 20 is variable (Strom et al. 1976). The large-aperture intensity ratios we found are quite different from the ratios found with the smaller aperture. However, they are consistent with the sense of the ratio being nearly unchanged, or perhaps slightly greater at 3.1 j.Lm than at 3.5 j.Lm, as found by Eiroa & Leinert. Because the ratios at 3.3,3.6 and 4.0 j.Lm are all consistently smaller through the larger aperture than through the smaller aperture, the environment surrounding SVS 20N must contain more reflected light than the environment surrounding SVS 20S. The largeaperture 3.1- j.Lm measurement relative to the small-aperture measurement depends upon any interstellar and circumstellar ice that may exist along the lines of sight. Consider, for example, the case in which there is no ice in the immediate circumstellar environments;

however, ice exists along the line of sight through the molecular cloud to the binary. In such a case, we would expect the 3.1-j.Lm ratio to decrease when we switch from the small to the large aperture in a similar way to that in which we observe a decrease at 3.3, 3.6 and 4.0 j.Lm. Instead, the large-aperture 3.1-j.Lm measurement is nearly the same as the small-aperture measurement. This implies that there must be some ice present within the circumstellar environment of SVS 20, around one or both protostars.

4 CONCLUSIONS

The similarity of polarimetric maps of the SRN at H and K to existing polarimetry models suggest that evacuated, bipolar cavities surround both of the protostars SVS 2 and SVS 20. Such bipolar cavities are typical around young, pre-main-sequence stars and are signatures of outflow activity characteristic of the star formation process. The polar angles of the discs surrounding SVS 2 and SVS 20 can be inferred from the symmetry of the polarization angles around each source. A search for emission from molecular hydrogen in the SRN reveals one knot lying along the south-western polar axis of SVS 20, a second knot lying along the south-eastern polar axis of SVS 2, and a third knot lying along both the southeastern polar axis of SVS 2 and north-eastern polar axis of SVS 20. Such emission is often observed in the environments around young stars and is attributed to ambient molecular hydrogen in the local cloud that is collisionally excited by outflowing material from a protostar. The association of these knots with the projected directions of the polar axes of SVS 2 and SVS 20 strongly suggests that both SVS 2 and SVS 20 are outflow sources. In previous millimetre and submillimetre studies, multiple molecular outflows have been

© 1997 RAS, MNRAS 290, 598-606


Dow


ic.oup.com/m



1997MNRAS.290..598H

4

2 .---.. (J (\) en ~ ~

v 0 2 0 ,; (\)

a -2

-4 4 2 0 -2 -4

RA Offset (orcsec)

SVS 20: 3.6 J.Lm 4

2 .---.. (J (\) en ~ ~

v 0 2 0 ,; (\)

a -2

-4 4 2 0 -2 -4

RA Offset (orcsec)


.---.. (J (\) en (J '-~

v 2 0 ,; (\)

a

.---.. (J (\) en ~ ~ ..., (\)

2 0 ,; (\)

a

4

2

0

-2

-4 4

4

2

0

-2

-4 4

2 0 -2 RA Offset (orcsec)

SVS 20: 4.0 J.Lm

2 0 -2 RA Offset (orcsec)

Figure 5. Diffration-limited, high-resolution images of the SVS 20 binary at 3.1, 3.3, 3.6 and 4.0 j.Lm. It is clear from these images that both SVS 20N and SVS 20S fade in the 3.1-j.Lm ice absorption band suggesting the presence of ice along the lines of sight. A comparison of the relative intensities of the two components through large and small synthetic apertures reveals the presence of circumstellar ice.

detected in the Serpens core, but none has been unambiguously identified with SVS 2 or SVS 20. This paper demonstrates the power of near-infrared polarimetry combined with narrow-band imaging of molecular hydrogen emission to detect and identify outflow sources.

From high-resolution images of the SVS 20 binary in the 3.1-fLm H20 ice feature, we conclude that ice-carrying dust grains must exist in the circumstellar environment of this binary. These images, however, do not preclude the existence of interstellar ice within the Serpens cloud.

Comparison of our images with previous images of the Serpens core reveals that the cometary nebula EC 81 has faded by at least two magnitudes at H and K between 1989 and 1994. Given the large amplitude and short time-scale of variation, as well as the number of other young stars exhibiting FUor- or EXor-type activity in the Serpens core, we suggest that the fading of EC 81 may be yet

© 1997 RAS, MNRAS 290, 598-606

another example ofFUor- or EXor-type activity in the Serpens core. It will be worthwhile to continue monitoring the Serpens core to gain a longer baseline for these young variable sources and to search for additional FU ors and EXors. Such data are relatively scarce, and a study of these sources in Serpens could significantly increase our understanding of the accretion process.

REFERENCES

Bally 1., Lada C. J., 1983, ApJ, 265, 824 Bell K. R., Lin D. N. C., Hartmann L. w., Kenyon S. J. , 1995, ApJ, 444,376 Casali M. M., Eiroa C., Duncan W. D., 1993, A&A, 275, 195 Churchwell E., Koomneef J ., 1986, ApJ, 300, 729 Eiroa C., Casali M. M., 1992, A&A, 262, 468 Eiroa C., Leinert C., 1987, A&A, 188, 46 Fowler A. M., Gatley I., Vrba F. J., Ables H. D., Hoffman A., Woolaway J.,

1994, Exp. Astron., 3, 243


Dow


ic.oup.com/m



1997MNRAS.290..598H


Gezari D. Y., Schmitz M., Mead I. M., 1987, Catalog of Infrared Observations, NASA RP 1294

Gomez de Castro A.I., Eiroa C., Lenzen R., 1988, A&A, 201, 299 Hodapp K.-w., HoraJ. L., Rayner I. T., Pickles A. I., LaddE. F., 1996, ApI,

468,861 Kastner I. H., Weintraub D. A., 1994, ApI, 434, 719 Kastner I. H., Weintraub D. A., 1996, ApI, 466, L103 Kenyon S. J., Hartmann L. W., 1991, ApI, 383, 664 Strom S. E., Vrba F. I., Strom K. M., 1976, AI, 81, 638 Weintraub D. A., Kastner J. H., Zuckerman B., Gatley I., 1992, ApI, 391,

784

Weintraub D. A., Kastner I. H., Lowrance P., 1994, in Clemens D., Barvainis R., ASP Conf. Ser. 65, Clouds, Cores and Low Mass Stars. Astron. Soc. Pac., San Francisco, p. 266

Weintraub D. A., Kastner I. H., Gatley I., Merrill K. M., 1996, ApI, 468, L45

White G. I., Casali M. M., Eiroa C., 1995, A&A, 298, 594 Whitney B. A., Hartmann L., 1993, ApI, 402,605

This paper has been typeset from aTE XIL ATE X file prepared by the author.

© 1997 RAS, MNRAS 290, 598-606


Dow


ic.oup.com/m



Bipolar outflow sources in the Serpens core: SVS 2 and SVS 20

Documents