Introduction

Most long-period variables (LPVs) are surrounded by extended dusty circumstellar shells accompanied by considerable mass loss, which is often large enough to influence their evolution and to provide substantial amounts of chemically processed material to the ISM (Lafon & Berruyer 1991). Due to the improved instrumental capabilities at infrared and longer wavelengths a large amount of observational data is available for these objects (for a review see Habing 1990). On the other hand based on the progress in theoretical dynamical modelling of LPV atmospheres achieved during the last years (e.g. Bowen 1988, Fleischer et al. 1992) a detailed comparison of theory and observations seems to become accessible now.

The modelling method

Dynamical model calculations

Our approach for the dynamical modelling of circumstellar dust shells of LPVs, which is described in detail in Fleischer et al. 1992, comprises the explicit solution of the hydro-equations in spherical symmetry adopting Lagrangian coordinates, the treatment of radiative transfer by the Eddington approximation for a spherical grey atmosphere according to Lucy (1976), the description of postshock cooling of the gas either by limiting cases (e.g. isothermal shocks) or by cooling functions, and the consistent detailed treatment of the formation, growth and destruction of carbon grains by means of a moment method (cf. Gauger et al. 1990).

Introduction

Strong winds from massive stars can sweep up the ambient gas forming stellar wind bubbles, also called ring nebulae. Classically, ring nebulae around Wolf-Rayet (WR) stars have been modeled assuming a homogeneous interstellar medium (ISM), following Weaver et al. (1977). However, theory and observations have progressed to the point that, this simplification can no longer be justified. The evolution of massive stars has been studied by Maeder (1990). He shows tracks for 15–120 M⊙ in his plots (Figures 1-4). Main sequence (MS) stars between 25–40 M⊙ evolve to WR stars after passing through a red supergiant (RSG) phase. Observations of MS stars (Herrero et al. 1992) and WR stars (Willis 1991) reveal fast winds, as opposed to RSG stars (Humphreys 1991), where the winds are dense and slow (see also Chevalier & Liang 1989, Stencel et al. 1989). The above studies, suggest to us that the ISM initially encountered by a WR wind is far from homogeneous. This is the base of our three-wind model. In order to explain WR ring nebulae, we must take into account the history of the central stars, not just their interstellar environment. We have already presented a brief description of an analytic calculation of the dynamical behavior of the swept-up shell of RSG wind (Garcia-Segura. and Mac Low 1993). In this paper, we present numerical computations of the shell that follow it after instability sets in and it can no longer be modeled analytically.

Introduction

Wolf-Rayet Shell Nebulae (WRSN) provide a “quick look” at an intermediate stage of evolution of massive stars between the main sequence O stage and their ultimate demise as SNII. During this evolutionarily brief epoch, the O star develops a strong wind which affects the surrounding ISM, and can even have significant mass loss which enriches the ISM with H-burning products –specifically He and N (Maeder 1990). Therefore, studies of these objects are both interesting and important regarding the physics of wind-shock effects on the ISM and in the role they have in galactic chemical evolution.

In this short contribution I will present some of the results of two recent students of mine who completed Ph.D. theses studying the morphology and spectra of the WRSN NGC 6888 (Mitra 1990) and NGC 2359 (Jernigan 1988). A more comprehensive review of the literature on WRSN is given by the fine paper by L. Smith in this volume. The theses studies incorporated CCD imagery mapping of the ionization structure of the nebulae in the emission lines of Hβ, [OIII]λ5007, Hα, [NII]λ6583, & [SII]λ6717+30; followed by spectroscopy of parts of the two nebulae that were of special interest from the imagery. Herein I will note some of the spectroscopic results regarding the hot wind-driven gas; the imagery mapping is available in their theses and moreso in a forthcoming Atlas of CCD Imagery of Galactic HII Regions (Hester et al. 1994).

Introduction

A variety of supernova interactions with circumstellar material (CSM) has been observed to date. The best, and most direct, example is the ring of emission around SN 1987A (Jakobsen et al. 1991). This material has been ionized by the UV and soft X-ray flash of the shock breakout at the surface of the supernova (Fransson et al. 1989, Lundqvist & Fransson 1989). The density enhancement in the ring is caused by the interaction of the fast blue supergiant wind colliding with the slow red supergiant wind of a previous epoch (Blondin & Lundqvist 1993). In the case of SN 1993J, the early detection of radio and X-ray emission, in combination with narrow emission lines in the UV and optical, are indicative of interaction with the CSM. Blue optical continua, X-ray detection at early phases, as well as the UV emission have been proposed as characteristics of a shock in the CSM around SN 1979C (Fransson 1984).

Introduction

Mass loss from late type stars is a ubiquitous phenomenon which has important ramifications for the further evolution of the star. Typically, the nuclear burning timescale of a giant is 10−7 M⊙/yr. For comparison, the mass loss rate on the Asymptotic Giant Branch (AGB) varies from 10−6 to a few times 10−4M⊙/yr.

Supernova 1993J in the spiral galaxy M81 is the brightest supernova since SN 1987A and, like the latter, appears to be another peculiar type II supernova. Its early light curve is characterized by a very sharp initial peak (lasting for less than ten days) followed by a less rapid secondary brightening, which was qualitatively similar to the secondary brightening observed in SN 1987A.

Humphreys et al. (1993) have identified a candidate progenitor consistent with the position of the supernova. Combining their UBVR. photometry with the I magnitude obtained by Blakeslee & Tonry (1993), they concluded that the colors of the apparent progenitor require the presence of at least two bright stars. One star is an early-type supergiant (most likely a late-B to early-A supergiant), the other a late-type supergiant (most likely a G to early-K supergiant). The bolometric magnitudes of both stars are in the range of –6 to –8, with best-fit values of –7 to –7.5 (for an assumed distance of 3.3 Mpc). We have performed our own fits to the photometric data and obtained similar results. These best-fit magnitudes imply mainsequence masses of ∼ 15 M⊙, but the masses could be as low as 8 M⊙ or as large as 20 M⊙. The image of the candidate progenitor appears extended on some plates (Blakeslee & Tonry 1993). This suggests that, at the distance of M81, the two stars do not form a close binary (although either star could have an undetected binary companion).

The historic series of Herstmonceux Conferences was started by the new Astronomer Royal, Sir Richard Woolley, in the late 1950's. The Royal Greenwich Observatory had by then recently finished moving its scientific operations from Greenwich to Herstmonceux Castle in East Sussex. Evidently the first few such conferences were relatively small private affairs, and there are few written records of them, but in later years they grew in size. After the moving of the RGO to Cambridge in 1989, they have been organised jointly with the University's Institute of Astronomy.

Our idea for the 34th Conference was to bring together different astronomical communities who study stellar evolution, stellar winds and the physics of circumstellar media, and to bring out the common physics affecting matter around both high and low-mass stars. This volume presents the proceedings. We have included all the invited reviews and the contributed oral talks, and there is a summary listing the titles and authors of all the poster papers.

Thanks are due to many people for helping to put together what was the largest-ever Herstmonceux Conference. The Organising Committee were Robin Catchpole, Robin Clegg, Peter Meikle, Jim Pringle, Anne Reynolds and Ian Stevens; Anne did a huge job as Conference Secretary and deserves special mention. The Advisory Committee were John Dyson, Alex Filippenko, Claes Fransson, Harm Habing, Alain Omont and Guillermo Tenorio-Tagle. The RGO and the IoA gave financial support, and the International Science Foundation funded speakers from Russia.

Introduction

The detection of narrow emission lines from supernova 1993J gives us a rare opportunity to study the circumstellar medium produced in the preexplosion phase. This can give insights into the evolution and mass loss history of the progenitor star, and test theories of the interaction of a supernova with its surroundings.

Observations

Optical spectra of SN 1993J were obtained with the IDS and ISIS spectrographs on, respectively, the INT and WHT on La Palma. For the first two weeks, spectroscopy was carried out nightly, and less frequently thereafter. The resolution and wavelength coverage varied according to the scheduled observing programme being carried out at the time.

Circumstellar matter in symbiotic stars

By definition, symbiotic stars exhibit simultaneously the absorption features of a cool giant and high excitation emission lines, e.g. HeII, [OIII], of an ionised nebula. Thus the presence of circumstellar material is a necessary classification criterion of these objects. It is now generally accepted that symbiotic stars are binary systems consisting of a red giant and a hot radiation source, in most cases a hot white dwarf. The strong emission lines originate in a dense nebula which is thought to be wind material lost by the cool giant and ionised by the hot companion.

Besides the compact nebula other components of circumstellar material are observed. Enhanced IR emission due to circumstellar dust is found in D-type (D for dust) symbiotic systems. Very extended ionised regions have been mapped with radio interferometers or optical imaging techniques (e.g. Taylor 1988, Solf 1988, Corradi & Schwarz 1993, Schwarz 1993). Some features of the extended structures can be associated with bipolar outflow (velocities ∼ 100 km/s) from the central, unresolved binary system.

In this paper we discuss how the geometric structure of the circumstellar environment of symbiotic systems can be clarified from an analysis of light scattering processes.

Scattering processes in the circumstellar environment of symbiotic stars

Polarisation measurements are a well known tool for studying scattering processes. Polarimetric observations of symbiotic stars in broad and narrow band filters have shown that these objects are often intrinsically polarised (e.g. Piirola 1983, Schulte-Ladbeck 1985, Schulte-Ladbeck et al. 1990).

Introduction

The defining characteristic of type II SNe is the presence of very broad Hα emission with a strong P-Cygni profile. A small number of peculiar type II SNe have been found in recent years which are either very bright in the optical continuum with very strong and broad Hα emission without a P-Cygni profile, or are strong radio sources then called radio supernovae.

Introduction

The circumstellar gas (CSG) around supernovae (SNe) provides information on the mass loss history of the dying star. When the SN explodes, the CSG is ionized by the radiation from both the SN and the gas shocked by the expanding ejecta. By looking at spectral signatures from the ionized CSG at increasingly large radii, we may peer deeper and deeper back through time into the mass loss history of the pre-SN. The recent bright and well-studied Type II SNe 1987A and 1993J have given us an unprecedented chance of doing so. Here we briefly discuss the CSG of these two SNe.

SN 1987A

Light curves for the narrow UV emission lines from SN 1987A (Sonneborn et al. 1994) show that the emission starts ∼ 70 days after the outburst with a roughly linear increase in strength until day ∼ 400, followed by a gradual decline up to day ∼ 1000 when the lines start to fall below detectability. In the optical there is very good information on the spatial flux distribution from observations with the NTT (Wampler et al. 1990; Wang & Wampler 1992) and the HST (Jakobsen et al. 1991; Plait et al. 1994); the emission mainly comes from a patchy, elliptically shaped ring with semimajor and semiminor axes of ∼ 0.830 (corresponding to ∼ 6.2 × 1017 cm at 50 kpc) and ∼ 0.605 arcsec, respectively.

Introduction

Supernovae are important for the study of several astrophysical problems – nuclear processing in stellar interiors, distance scale determinations, and the chemical enrichment of the interstellar medium have all been explored. Additionally, they may be used as background probes of interstellar gas by studying, for example, NaI and CaII lines, sampling the gas in the host galaxy, the Milky Way halo gas, and any intervening intergalactic gas. SN 1987A allowed the detailed study of gas towards and within the LMC to a distance D ≃ 0.05 Mpc (Vidal-Madjar et al. 1987, de Boer et al. 1987). With the unusually bright SN 1993J in M81 it is now possible to extend the search for interstellar/intergalactic absorptions beyond the local group of galaxies, out to the distance of the M81 group at a distance of D~3.25 Mpc. In this paper we present a preliminary study of high resolution optical interstellar spectra towards SN 1993J. The observations are described in section 2. The origin of the absorption lines, which fall into three distinct groups are discussed in section 3.

Observations

The observations of SN 1993J were obtained during the nights 1993 April 4–8, using the Utrecht Echelle Spectrograph on the William Herschel Telescope at the Observatorio del Roque de los Muchachos, La Palma.

Introduction

An early result from ultraviolet astronomy was that OB stars with bolometric magnitudes brighter than Mbol ≃ −6 suffer significant mass loss (Snow & Morton 1976). At about the same time the framework was laid down by Castor, Abbott & Klein (1975; hereafter CAK) for what, has proved since to be a very successful theory of mass loss from OB and other similarly high-luminosity stars. In bare outline the physical model is a simple one in which the outward force able to overcome gravity is the pressure exerted by the hot star's radiation field on its own atmosphere. In practical application, complexity arises from the fact that it is overwhelmingly the scattering of radiation in spectral lines that, mediates the force (a point first appreciated by Lucy & Solomon 1970).

Introduction

The standard picture of a Type II SN progenitor star is a red supergiant (RSG) that has evolved from a massive star with an initial main-sequence mass above ∼ 10M⊙. These RSGs are observed to have very massive, slow winds with terminal speeds in the range of 10 − 50 km s−1, and mass loss rates in the range of 10−7 − 10−5M⊙yr−1. These slow winds will gradually blow a stellar wind bubble of RSG wind into the relic main-sequence stellar wind bubble, building up a shell of shocked RSG wind at the edge of the expanding bubble.