Evolution of post-AGB stars
(Gledhill, Cioni)
Low to intermediate mass stars (0.8 to 8 Mo) evolve between the Asymptotic Giant Branch (AGB), across the top
of the HR diagram, to end their days as cooling white dwarfs. This is a transition that will be undergone by the majority of stars in the Galaxy, probably including the Sun, and
these end phases of a star's life are characterised by huge mass loss, which is the major source of dust and chemically enriched material in the interstellar medium. During
the final AGB stage, mass-loss rates can exceed 10-4 Mo yr-1 and can become highly asymmetric, resulting in the formation of circumstellar discs,
bipolar outflows and jets, such as seen in the protoplanetary nebula AFGL 2688 (left). The most likely mechanism for this asymmetry is the interaction of the mass-losing star with
a binary comapnion, or perhaps with a planetary system. The group studies these objects in the post-AGB phase using a variety of observational and modelling techniques:
Near-infrared Imaging Polarimetry of Post-AGB Stars
In the NIR, the dusty circumstellar envelopes (CSEs) around post-AGB stars are seen by scattered light and are therefore polarized, whereas the point-spread function (PSF) of the bright central star is unpolarized. Imaging polarimetry allows us to separate the two components and to image the extended CSE in compact AGB and post-AGB sources. This technique was first used by Gledhill et al. (2001) to survey 16 post-AGB stars and detect a range of morphologies from shells to bipolars. Two broad classes are identifiable: (i) mildly asymmetric CSEs where the star is usually visible and (ii) strongly asymmetric or bipolar CSEs where the star is usually obscured. Class (i) are identified with a freely expanding and detached CSE wherein binary interaction does not appear to have had a strong role, either because the mass-losing star is single, or because the binary separation is too large to have had an effect. Class (ii) objects appear to have a dense circumstellar disc hiding the star, which may be long-lived, and suggests a binary origin ( Gledhill 2005).
Themulti-waveband imaging polarimetry allows us to construct detailed scattering models of the CSE to investigate the geometry, the distribution of dust and the composition (O- or C-rich) and sizes of the dust grains. We are refining current models of CSE structure to determine the key physical parameters such as the dynamical timescale for outflow on the AGB, the mass of dust in the outflow and the evolution of mass loss rate. These CSE models can be used to predict the mid-infrared structure for comparison with imaging observations on Gemini with MICHELLE, and the spectral energy distribution (SED) out to 1mm. A dust model of the CSE of the young PN 19306+1407, consistent with observations from optical to mm wavelengths, is given by Lowe and Gledhill (2007).
Mid-Infrared Imaging and Modelling
The dusty CSE seen in scattered light in the near-IR is seen directly at longer wavelengths (> 8
microns) due to thermal emission by the warm dust grains. Dust at the inner edge of a detached CSE, or in a circumstellar/binary disc, will radiate strongly in the 10 and 20 micron
windows, and can be imaged using instruments such as MICHELLE and TReCS on the Gemini telescopes. The images and dust emission features can then be modelled using a radiation transport
(RT) code, to determine the degree of axisymmetry in the CSE, the mass of dust, size of dust grains, and chemistry. Models of oxygen- and carbon-rich objects can be found in Gledhill
and Yates (2003) and Clube and Gledhill (2004) respectively. The latter study is extended
in Kim Clube's PhD Thesis "The dusty envelopes of post-AGB stars" (University of Hertfordshire 2007). However, these models do not account for the presence of circumstellar
discs, which are thought to harbour cold large dust grains. We are therefore extending our RT codes to treat binary systems with outflows and circumstellar discs, so that we can
address the role that discs play in dust grain growth (possibly up to macro-objects), processing and storage. An example of an oxygen-rich object which appears to possess a massive
circumstellar disc, is provided by OH231.8+4.2. In the 2 micron HST image (right), the central star is totally obscured by a dense circumstellar disc. In our 11 micron TReCS image,
a bright (19.6 Jy) point source is seen. This is thermal radiation from the disc and an analysis of the fluxes at 11 and 18 microns indicates that there is a lot of cold dust close
to the star. The SED has been modelled as a flared optically thick disc by Jura et al. (2002,ApJ 574 963) and optical spectroscopy indicates that the source is a M10 evolved star
with a A0 V main sequence companion (Sanchez-Contreras et al. 2004 ApJ 616 519).
PN formation and Molecular Hydrogen Imaging using Integral Field Techniques
Planetary Nebulae (PN) are thought
to form when a fast wind begins to blow from a post-AGB star as it evolves to hotter temperatures. However there is little observational information on how the PN formation process
occurs, resulting in a huge parameter space for models to explore. The ro-vibrational transitions of H2 in the infrared K-band are an important tracer of excited molecular
gas in post-AGB outflows. Using Integral Field Spectroscopy (IFS) we can obtain both the spatial and spectral distribution of the emission simulataneously. The image to the right
shows the continuum subtracted H2 1-0 S(1) line for the hot post-AGB star IRAS 19306+1407 taken with the UIST imager on the UK Infrared Telescope. This object has a B1
spectral type and is likely to be a young PN. The H2 emission traces a bipolar structure along PA 30 deg, which is colinear with the axis seen in HST images. In the central
region, the flattened contours provide evidence for a torus structure. This is corroborated by high spatial resolution IFS taken with the SINFONI instrument on VLT (Lowe et al.
2007 in prep). A particular strength of the IFS technique is the ability to obtain high quality spectra over the whole emission region. This enables line-ratio maps to be constructed,
allowing discrimination between radiative and shock excitation. For example, ratios of the 1-0 S(1) and 2-1 S(1) lines of greater than 10 indicate collisional shocks, whereas radiative
excitation is expected to result in ratios of between 1 and 2. For further details see Lowe
and Gledhill (2006).
High Resolution Maser Interferometry Observations of PPN with MERLIN
Our NIR imaging-polarimetry observations of the dust content of the CSEs of AGB and post-AGB sources are complemented by a programme of radio-interferometric observations of maser emission from OH and H2O in the envelopes. These molecules trace gas at radii of between 1015 and 1017cm. The 18cm OH maser lines have a MERLIN spatial resolution of 0.2 arcsec, allowing us to directly compare the gas and dust structure in the envelopes. The excellent kinematic resolution of MERLIN (0.35 km s-1 at the OH lines) allows the outflow velocity to be mapped across the shell in detail. This technique is illustrated in the MERLIN observations of HD 179821, a possible hypergiant (Gledhill, Yates and Richards 2001). Polarimetric observations of the OH maser lines afford the possibility of detecting ordered magnetic fields around post-AGB stars. This is an important topic, since a number of theoretical models have been advanced recently which attempt to explain the development of asymmetric outflows by the presence of a dynamically important mangetic field. MERLIN allows spectral line observations in full polarization, so that both linearly and circularly polarized components of the maser lines, arising due to Zeeman splitting in an ordered magnetic field, can be detected. We have used this technique to detect the first dynamically important magnetic field around a post-AGB star and the figure to the right shows the spatial distribution of linearly polarized maser components in the 1612 MHz line towards the bipolar AGB/post-AGB object OH17.7-2.0. In combination with the circularly polarized features, the results are consistent with a stretched dipole field with strength B= 4.6 mG and comparison with infrared images shows that the masers appear to be contained within a dense dusty circumstellar disc around the star (Bains et al. 2003).
Imaging polarimetry of AGB stars (Cioni)
AGB stars lose mass at a rather high rate. Thus, they are among the major contributors of heavy elements to the interstellar medium for the formation of new stars and planets. They are important sources of elements produced via the s-process and of dust which forms in the outer part of their atmosphere. AGB stars are also the precursors of beautiful, mostly a-spherical PNe, but the transition is poorly understood: is the shape the result of binary interaction or is it due to intrinsic properties of the AGB progenitor? Recent results indicate that circumstellar envelopes already become highly a-spherical during the AGB phase, possibly under the influence of a strong magnetic field. Polarimetry of a large sample of stars can provide important clues about when the AGB star envelopes become spherical, as polarisation is a measure of asymmetry of the scattering particles around the star.
Imaging polarimetry of the Magellanic Cloud, to which distances are sufficiently accurate to measure the degree of linear polarisation, have been obtained from La Silla using the EFOSC2 instrument (optical) while a collaboration has been established to use the SIRPOL instrument (near-infrared) to significantly contribute to the investigation of a large sample of stars. This sample will allow us to study the relation between the degree of polarisation and several key stellar parameters providing crucial constraints for the models that describe the formation of a-spherical PN. Since the new opportunity offered by wide-field imaging polarimetry includes stars of different types (not only evolved giants) other aspects of stellar structure and evolution, such as the formation of massive stars, will be constrained.
Variability and chemistry of AGB stars (Cioni)
The luminosity of AGB stars varies periodically with long periods and large amplitudes. Long term microlensing programs (i.e. EROS, MACHO and OGLE) have provided light-curves for many AGB stars in the Magellanic Clouds and in the Bulge. At least four period-magnitude relations were discovered (i.e. Wood et al. 1999, Cioni et al. 2001, 2003). Each but one relation is associated to a different mode of radial pulsation which changes from overtone to fundamental mode for ageing stars. Mira variables, important distance indicators, are luminous fundamental mode pulsators. This result has been confirmed by comparing interferometric angular diameters with models of large amplitude pulsating atmospheres. The yet unexplained relation is perhaps occupied by binary stars with an AGB companion. Wood and Cioni are investigating this possibility; data from a spectroscopic monitoring campaign with FLAMES at the VLT are being analysed. The goal is to combine photometric and radial velocity variations in order to constrain the nature of those stars with an unexplained type of large amplitude variability. They represent about 25% variable AGB stars in the Large Magellanic Cloud.
AGB stars produce a considerable fraction of the heavy elements present in the interstellar medium which set the basis for the formation of a new generation of stars. High-resoultion spectrographs operating on large telescope allow us to directly measure the abundance of these elements. Although the spectra of AGB stars are dominated by molecular bands, often preventing the analysis of individual atomic lines, the constant developement of model atmospheres and spectral synthesis for these complex and dynamical objects shows a very active field of research likely to bring exiting results in the near future.