Please Note: the e-mail address(es) and any external links in this paper were correct when it was written in 1995, but may no longer be valid.
Mullard Radio Astronomy Observatory, Cavendish Laboratory, Madingley Road, Cambridge CB3 0HE, UNITED KINGDOM
Recent observations at the James Clerk Maxwell Telscope (JCMT) and elsewhere have identified a class of very deeply embedded, possibly protostellar, sources which are not associated with any of the traditional signposts of star formation - HII regions and near infrared emission for example -but which do lie close to otherwise isolated masers. This raises the possibility that it may be possible to identify many more of these `class 0' sources from an appropriate sample of maser sources. We have defined such a sample using the Arcetri catalogues of Comoretto et al. (1990) and Brand et al. (1994), and have observed 44 of these sources in continuum (and in some cases at also), using the 15-m JCMT. Most (95 per cent) of the regions showed significant far infra-red (FIR) emission.
Figure 1: From Hobson et al. (1993) showing the three water masers and their associations in the Northern Condensation of M17 SW.
The association of powerful masers, bright ultra-compact HII regions and high-velocity flows in regions of OB star formation is well established, though not well understood in detail (e.g. W51, Genzel et al. (1981)). For these regions, theoretical models involving dissociative shocks have been reasonably successful in explaining the maser emission Elitzur et al. (1989). The precise mechanism (stellar wind, molecular outflows, cloud collisions etc.), and the conditions required to produce strong maser emission have not yet been clearly identified however. masers have also been found in a number of less active star-forming regions e.g. associated with Sharpless or radio HII regions, and studies of such regions have recently been undertaken by Tofani et al. (1994) and Ladd (1994). In this paper we consider maser sources which have no obvious stellar association but may be associated with an earlier stage of star formation. An example of maser association with sub-mm cloud cores can be found in M17 SW where each of the three water masers found in the Northern Condensation are associated with a separate FIR peak as can be seen in Figure 1 of Hobson et al. (1993).
The first step in the project was to find a list of all the known water maser sources. This task was made much easier by the publication of the Arcetri water maser atlas Comoretto et al. (1990) and more recently the release of an updated version by Brand et al. (1994). These lists contain a total of 718 water maser sources and they had been classified according to their IRAS fluxes as belonging to either evolved stars or young star forming regions following the method explained in Palagi et al. (1993). They used a method of principal component(PC) analysis to identify the maser type. They took the Arcetri Atlas and used 246 masers of known type to construct three linear combinations of the logarithms of the IRAS fluxes, with 12, 25, 60, 100, to determine ways of distinguishing future IRAS sources.
With these definitions a maser can be associated with a star or star-forming region using the boundaries specified in Table 1. This method reclassifies the first Arcetri Atlas as 229 masers from evolved stars and 262 from star forming regions. The complete Atlas has 409 sources satisfying the star forming region (SFR) criteria.
Table 1: Classification of probable maser types from the PC analysis of Palagi et al. (1993).
A further constraint on our sample was caused by the timing of our telescope runs, which restricted observations to the range to in right ascension. This did not affect the sample too much since few water masers are known outside this range. Well known sources such as M17SW (e.g. Hobson et al. (1993)), S106 (e.g. Richer et al. (1993)), Taurus and Oph, for which considerable data already exist, were also removed from the sample.
The sample was further reduced by limiting it to those sources which had a distance (actual or kinematic) less then 5 kpc. This had the advantage of retaining a reasonably good spatial resolution and reduced the sample size to a more manageable 119 sources.
Limited observing time on the James Clerk Maxwell telescope (JCMT) meant that only 44 sources could be mapped in the sub-mm continuum.
The sub-mm continuum data presented here were observed at the JCMT in 1993 September and 1994 August. All the data were taken using receiver UKT14, a single element -cooled Ge:In:Sb bolometer (Duncan et al. (1990)), and the continuum backend (CBE). The source was mapped in an `on the fly' observing mode, where the telescope was continuously scanned in azimuth whilst using the secondary mirror to chop in the azimuth directions; several scans at different elevations are used to make the final map. (18 arcsec FWHM) maps and, where possible, high resolution maps (7 arcsec beam FWHM), were made of 44 sources.
The mm spectral line observations were made with the JCMT during May 1994 and August 1994 using the common-user SIS Receiver A2 (Davies et al. (1992)) and the Dutch Autocorrelation Spectrometer (DAS) with a bandwidth of 250 MHz over 2048 channels. This gave a velocity resolution of 0.156 MHz, equivalent to a velocity resolution of approximately . We observed the and rotational transitions which have rest frequencies of 230.5380 GHz and 219.5603 GHz respectively and also at 244.9357 GHz (which is approximately 1.3 mm wavelength). All the maps were done as grids in position switch observing mode where the telescope integrates on source and then moves to a fixed off position - the difference between the two spectra is then assumed to be from the source.
Observations were made using the VLA in `B' array in July 1994 to find accurate positions of some of the water masers and their relation to any ultra-compact HII regions which might be present. Single snapshots of 7 minutes were done at 8 GHz continuum and in the maser transition at 22.23508 GHz.
A sample from the complete data set can be found in the following pages. Most of the spectral energy distributions (SEDs) lack information at near-infra-red wavelengths. For more information on similar work of this kind but in the near-infra red see the paper by Leonardo Testi in these proceedings.
Figure 2: and 8 GHz continuum, and the SED of IRAS 20188+3928
Figure 3: and continuum, and the SED of IRAS 22198+6336
Figure 4: and 8 GHz continuum, and the SED of IRAS 18273+0113
Figure 5: and continuum, and the SED of IRAS 23314+6033
Figure 2 shows some data for IRAS 20188+3928. This source is interesting in that it is one of only two sources in the sample where the water maser is not associated with the sub-mm dust emission. The VLA map does, however, show that the masers (crosses on the figures) are near a weak ultra-compact HII region.
IRAS 22198+6336 and IRAS 23314+6033 and (Figures 3 and 5) show evidence of velocity components in the maps of a few separation which may be due to outflow activity.
IRAS 18273+0113 (Figure 4) shows an example of a possible `class 0' object, since the infra-red fluxes are only upper limits. For this source the maser is centred on radio continuum and sub-mm continuum.
The following conclusions can be drawn from the initial examination of the data. A complete analysis can be found in Jenness et al. (1995b).