[1]Physical Research Laboratory, Ahmedabad, India - 380009
[2]Joint Astronomy Centre, 660 N. Aohoku Pl., Hilo, Hawaii-96720, USA
[3]SCFAB-KTH, At. & Mol. Physics, Roslagstullsbacken 21, Stockholm, Sweden

The nova-like outburst of V4332 Sagittarii (V4332 Sgr) in February, 1994 was peculiar. Martini et al. (1999), who studied the object, showed that its outburst did not conform to known categories of cataclysmic variables, like novae or "born-again" AGB stars. In what has become the signature for the evolution of such objects, V4332 Sgr showed a rapid cooling over three months - from 4400 to 2300 K - and evolved into a cool M-type giant.

A similar evolutionary behavior has also been observed in the recent (January 2002) outburst of V838 Monocerotis (Munari et al. 2002), and earlier in a red variable that erupted in the Andromeda galaxy called M31 RV (Rich et al. 1989). These objects appear to be analogues and are thought to constitute a new class of objects (Munari et al. 2002; Banerjee and Ashok, 2002; Bond et al. 2003).

It seems that new mechanisms are necessary to explain the outbursts of these objects. For example, Soker and Tylenda (2003) explain the outburst as a merger event between two stars, where the release of gravitational energy leads to the the nova-like outburst, in distinct contrast to the classical nova scenario.

FIGURE 1: Bottom: observed JHK spectra of V4332Sgr showing the prominent A^2 \Pi_i -- X^2 \Sigma^+ emission bands of AlO. Top: a synthetic spectrum with T(rot)=300 K and T(vib)=3000 K.

Because of the limited studies of V4332 Sgr and the considerable interest in V838 Mon at present, we thought it useful to re-observe V4332 Sgr to see how such an object evolves. We observed this interesting object via the UKIRT service programme with UIST and UFTI. The JHK spectra obtained are shown in Figure 1. The most striking feature of the spectra is the presence of several strong, molecular bands in emission. Based on laboratory spectra by Launila & Jonnson (1994), we have identified the bands as being due to transitions in the A^2 \Pi_i - X^2 \Sigma^+ band system of AlO. A synthetic spectrum of AlO is also shown in Figure 1 - with the prominent bands marked - where there is clearly a good match between the observed and expected features.

FIGURE 2: Model fits to the AlO (2,0) band profile for different values of T(rot), overlaid on to the observed (2,0) band (plotted with a thick black line). T(vib) = 3000 K for all the fits.

At near-IR wavelengths, the only other detection of AlO is very recent and is coincidentally seen in V838 Mon (Geballe et al. 2002, Bernstein et al. 2003). However, only some of the AlO bands detected here were seen in V838 Mon, and several of the other bands seen by us are first detections.

The dereddened spectral energy distribution (SED) of V4332 Sgr, derived from 2MASS data observed in 1998, can be fitted with a 3250 K black body. By contrast, our current JHK photometry (from 2003) shows an SED with a steep rise towards longer wavelengths (with some excess in the J-band). These new data can be fitted with two black bodies with temperatures of 3250 K and 900 K. To infer anything more about the hot component from the present data is difficult. What is clear, however, is the evidence for a cool 900-1000 K component, which we interpret as arising from a dust shell that has formed since the 2MASS observations were taken. If the ejecta had an expansion velocity of 200-300 km/s (from the H-alpha line widths of Martini et al. 1999), then in the absence of any deceleration, a shell formed from the 1994 outburst should now have a radius of ~10 solar radii. At a distance of 300pc, this is equivalent to an angular diameter of 3.0", so it should be resolved in our K-band UKIRT images (taken during <0.7" seeing). But in the K-band data we do not see any extended features. It therefore appears that the dust shell diameter is much smaller than expected, and that it is not associated with the 1994 outburst. Instead, it must be associated with a second episode of mass-loss.

We interpret the AlO as being of non-photospheric origin because it is seen in emission. It may be mixed with the dust in the shell or it may be located outside of the shell. It appears to be pumped by the near-IR photons emitted from the star or from the dust shell. The synthetic spectrum in Figure 1 has been computed for vibrational and rotational temperatures of 3000K and 300K respectively. The unequal values of T(vib) and T(rot) indicate that the system is not in thermal equilibrium. The reasons for this are being investigated. The model calculations, done for different bands, also indicate that a low rotational temperature (in the range of 200-400 K) is required to simulate the observed AlO spectra. This is clearly seen in Figure 2 where model fits at different values of T(rot) are overlaid on to the observed (2,0) band profile.

The present studies of V4332 Sgr give valuable insights into the evolution of this new class of object. Specifically, it is seen here that the mass loss in such objects may continue beyond the main eruption phase. The presence of molecular lines indicates that a cold environment can surround these stars while they are in their post-outburst stage. This is in sharp contrast to the extremely hot coronal gas associated with classical novae. We are looking forward to further multi-wavelength observational studies of this extremely interesting object.

References

Banerjee D.P.K. and Ashok N.M., 2002, A & A, 395, 161
Bernstein, L.S., Rudy, R.J., Lynch et al., 2003, IAU Circ. 8082
Bond H.E., Henden A., Levay Z.G.et al., 2003, Nature, 422, 405
Geballe, T.R., Smalley, B., Evans, A. & Rushton M.T., 2002, IAU Circ. 8016
Martini P., Wagner R., Tomaney A. et al., 1999, AJ 118, 1034
Munari U., Henden A., Kiyota S. et al., 2002, A & A, 389L, 51
Rich R.M., Mould J., Picard A. et al., 1989, ApJ, 341, L51
Soker N., and Tylenda R., 2003, ApJ, 582, L105

[1]Armagh Observatory, Armagh, N. Ireland and
[2]MPIA fur Astronomie, Heidelberg, Germany,
[3]Joint Astronomy Centre, Hilo, Hawaii, USA

When a star is born, some of the gas that tries to accrete onto the protostar is driven into a collimated outflow which disrupts the surrounding environment. The outflow sweeps up, compresses and excites cloud material through shock waves. As part of our ongoing programme to scrutinise these environments, we recently investigated the S233 HII region, which is located in the Perseus Arm at a distance of 1.8kpc. This region contains high mass protostars within dense molecular cores which drive multiple outflows into the ambient medium, disturbing and disrupting it. Besides disordered clumps and knots, S233 contains spectacular examples of bright bow shocks. Here we concentrate on two of these, N1 and N6.

The narrow-band high-resolution UFTI images presented in Figs 1a and 1b are of excited molecular and atomic gas, in the H2 1--0S(1) and 2--1 S(1) lines at 2.122 microns and 2.247 microns, and the [FeII] line at 1.64 microns. Only a few bow shocks have previously been imaged in H2 lines originating from above the first vibrational level (Davis et al. 1996, ApJ 443, L41; Eisloeffel et al. 1996, AJ 112, 2086). Recent studies of bow shocks in HH 7 and HH 240 use steady magneto-hydrodynamic models of planar and bow-shaped shocks (Smith & Brand 1990, MNRAS 245, 108; Smith 1994, MNRAS 266, 238), models which have been newly refined and which match reasonably well the observed parameters for both objects (Smith et al. 2003, MNRAS, 339, 524; O'Connell et al. 2003, in prep.). Physical, chemical and geometrical factors can often be distinguished given a sufficient variety of data. A basic network of oxygen and carbon chemistry is included in the models to account for the formation and cooling through water and CO rotational and vibrational transitions. Given the shape of the shock front, the shock surface is discretised and the deflected and compressed fluid is followed by determining the steady shock solution through numerical integration.

In our present work (Khanzadyan et al., 2003, A&A, submitted) we apply this latest version of the model to the S233IR N1 and N6 bows.

FIGURE 1: Left/Top - the S233IR N1 bow (and part of N2) in H2 1-0 S(1), 2-1 S(1) and [FeII] 1.64 micron emission. Colour-bars show the flux level in units of 10{-19} W/m2/arcsec2. Right/Bottom - result for the N1 bow using a C-type paraboloidal bow shock model. See Khanzadyan et al. (2003) for details of the model parameters.

The UKIRT data are compared to synthetic bow-shock images in Figs. 1 and 2. Both "Jump-type" and "Continuous-type" shocks are considered in our models. In the J-type shock, the conditions in the gas (density and temperature) change discontinuously across the shock front; molecules tend to be dissociated unless the shock velocity is low (<20 km/s). However, if there is a sizable magnetic field in the pre-shock gas and a low ionisation fraction, ions will travel ahead of the shock front and "communicate the arrival of the shock" to the pre-shock medium. The pre-shock gas is slowly (and more "continuously") accelerating to the post-shock speed without dissociation. Essentially, this process of ambipolar diffusion "cushions" the gas against the shock, thereby allowing for higher shock velocities (up to 50 km/s) without dissociation and potentially for more extensive molecular excitation across the shock front.

The model produces synthetic images that are sensitive to the bow shock speed and the ambient (pre-shock) gas density, as well as to the strength and orientation of the magnetic field in the ambient gas. For example, for N1, a bow shock speed of 65 km/s provides the best C-type shock fit to the H2 emission structure. Higher speeds tend to push the emission into two separated wings (i.e. away from the bow head). However, [FeII] emission from a C-type shock is negligible. Strong iron emission therefore only arises from a fast, dissociative, J-type shock around the bow cap.

FIGURE 2: Left/Top - the S233IR N6 bow in H2 1-0 S(1), 2-1 S(1) and [FeII] emission. Colour-bars are same as Fig.1. Right/Bottom - result from a C-type bow shock model for bow N6. See Khanzadyan et al. (2003) for details of the model parameters.

The magnetic field is critical to the C-shock structure, and this strongly dictates the observed flux distribution across the curved, bow shock surface. In N1 the modelling predicts a field in the plane of the sky but at 45 degrees to the bow direction of motion. This produces the observed axial asymmetry in the emission across the bow (particularly evident across N6 in Fig.2).

The model also predicts the observed line fluxes, as well as the gas excitation across the bow shock surface, measured here from 2-D 1-0S(1)/2-1S(1) ratio maps compiled from the data in Figs.1 and 2.

To produce the longer tail of 1-0\,S(1) emission seen in N6 (Fig.2) we simulate a C-type bow with stronger shocks in the wings by taking a more conical bow shape. Again, a C-shock generates strong 1--0 S(1) emission and the observed excitation. The latter is now considerably lower than in N1, due to the more aerodynamic bow-shock shape, but it is still high in comparison to the observed excitation across N6. Given the considerable fine-scale structure in this object, and the possible striations across the bow, some weak turbulence may be dissipating energy in the bow wake, producing the low-excitation emission observed.

An asymmetric magnetic field again improves the comparison between the data and the model. In N6 a compact H2 knot appears at the apex location where the field inhibits dissociation. Enhanced H2 emission is also produced in the bow shock flanks, although the tail length appears somewhat shorter than is observed.

Overall, the models yield convincing fits to the observed data, in S233 but also in a number of other systems. We can be confident that, bar possible refinements associated with turbulence and inhomogeneities in the ambient gas, we are approaching a reasonably complete understanding of molecular bow shocks in outflows from young stars.

University of Leeds, U.K.

Massive stars (>8-10 Msun) have a pivotal impact on the interstellar medium and galactic evolution. Their energy output can be sufficient to dominate the luminosity of their natal galaxy. Mass loss is comparable with that of all the other stars in the galaxy. At the start of their life they power strong outflows and HII regions, ending in a supernova explosion releasing energy and chemically enriched material. It is imperative that we understand their formation and upper initial mass functions.

FIGURE 1: UFTI image of a MYSO candidate. One can easily see the bright centre surrounded by the extended emission.

However, our understanding of massive star evolution is far from complete and much poorer than that of their low mass counterparts. To try and fill in some of the gaps we are conducting a systematic search for massive young stellar objects (MYSOs). MYSOs are the early formation phase of massive OB stars where fusion has started in the core but the surrounding dust cloud has not yet been ionized to form a HII region.

Observationally MYSOs are very bright in the mid-infrared, where most of the bolometric luminosity emerges after reprocessing by the surrounding dust cloud. They invariably power massive molecular bi-polar flows, that often manifest at near-IR wavelengths as reflection nebulae and shrouded emission.

Searches for MYSOs must be conducted in the thermal IR, because they suffer heavy extinction from their natal dust cloud. Known MYSOs have very red colours, but unfortunately so do compact HII regions and compact planetary nebula. An initial sample of 3000 candidate MYSOs was created using colour information from the Mid-IR Galactic Plane survey carried out by the MSX satellite. 2MASS was then utilised to remove objects that are too blue in the near-IR, e.g. planetary nebula. UKIRT is the ideal telescope for carrying out the necessary follow-up observations to remove the remaining contaminating objects and determine MYSO classification.

To identify genuine MYSOs a sub-sample of candidate MYSOs were observed with MICHELLE (at 10 microns) and UFTI (at 2.2 microns). For a genuine candidate we should see a point source with MICHELLE and evidence of youth, e.g. nebulosity, extinction or a cluster with UFTI. An UFTI image of a MYSO candidate can be seen in Figure 1, which clearly shows extended bi-polar emission. The corresponding MICHELLE image shows a point source situated at the bright centre of the UFTI image.

FIGURE 2: Long-slit spectrum taken at the bright centre of Figure 1. The data between 1.8-2 microns have been clipped to remove atmospheric effects.	FIGURE 3: Long-slit spectrum taken along the South-West extended emission of Figure 1. The data between 1.8-2 microns have again been clipped to remove atmospheric effects.

To confirm the objects MYSO status a long-slit spectrum was taken with UIST. Figure 2 shows a spectrum taken at the core of Figure 1. Evidence for a heavily embedded object with hot dust emission can be seen in the rise of the red continuum. There is also evidence of a massive young star from the Brackett Gamma line (at 2.16 microns), which is usually attributed to a stellar wind, and the weak CO bandhead emission (at 2.3-2.4 microns), which may be evidence for a disk. A spectrum taken along the south-west outflow shows evidence for shocks in many of the emission features, e.g. [FeII] at 1.64 microns and the molecular hydrogen lines (Figure 3).

These UKIRT data all point to the identification of this object as a MYSO; the target was only previously known as an infrared and maser source. This is just one example from a survey that we believe will eventually define a sample of MYSOs that is an order of magnitude larger than that currently known.