The origin of mass-loss in cool luminous stars is still obscure; several known mechanisms such as thermally driven wind, radiation-driven wind(via dust), wave-driven wind etc all have serious difficulties, if examined in the light of recent observations. At the same time, recent observations in the infrared and radio spectral domains revealed that outer envelope of red (super)giant stars has highly complicated spatial and velocity structures, while inner envelope may have new component that had not been recognized before. For example, recent high resolution infrared spectroscopy revealed a possible presence of a quasi-static turbulent molecular dissociation zone somewhere in the outer atmosphere. This new component may represent a transition zone between the warm chromosphere and the huge expanding molecular envelope, and may be a cool component of chromospheric inhomogeneity or a moleclar condensation in a cool corona extended by turbulent pressure. Such a result can be regarded as observational evidence in support of a recent theory of autocatalytic molecular formation by thermal instability due to molecular cooling. Thus, observation and theory consistently show the presence of a new component - quasi-static turbulent molecular formation zone - in outer atmosphere of cool luminous stars, and a possibility of a unified understanding of outer atmospheric structure and mass-loss, in which turbulence may play important role, can be proposed.

The oool extended shells of giants and supergiants are well known to be places of copious dust formation as indicated by the occurrence of pronounced extinction, reddening, and polarization of the continuous star light and by the appearance of particular absorption features, both manifesting the interaction of photons with particles considerably larger than atoms or molecules.

The accepted explanation of these phenomena is the formation of circumstellar dust, i.e. small solid particles with a typical size of the order of 0.1 micron. However, the analysis of these effects yields only information on the interaction of photons with certain functional groups within the clusters - e.g. Si-O, C-H, C-C bending and Si-O, C-C,... stretching vibrations - and thus allows no definite determination of the “real” physical structure and chemical composition of the grain particles. Therefore, observational conclusions concerning the properties of circumstellar dust can provide only some “mean” information which allows no definitive conclusions regarding the true nature of the observed grains (geometrical shape, crystalline structure,...?).

Two principal mechanisms that may be responsible for mass loss from red giants are considered: shock wave-driven winds and radiatively (dust)-driven winds. Effect of the periodic shocks accompanying nonlinear oscillations of red giants is most prominent in the outer layers of the stellar atmosphere where shocks are able not only to expel gas but also increase gas density so that some molecular components become supersaturated. In O-rich stars the most abundant condensible species are silicon monoxide and iron, whereas in C-rich stars these are carbon, silicon carbide and iron.

Infared polarimetric and photometric mapping observations at K(2.2 μm) and H(1.65 μm) have revealed an extended dust envelope around the late-type star IRC+10216. The observations were made on the 3.8-m United Kingdom Infrared Telescope on Mauna Kea, Hawaii, in 1985 December and 1987 January and February. The polarization observations were made by emplying the Kyoto polarimeter (Sato et al. 1987). Great care was taken to check the contamination by stray light in the telescope and instruments as the source on peak was extremely bright (K~0 mag). From the observations of normal stars, we found that the polarized intensity (degree of polarization times the intensity) was a good measure of the envelope, free from contamination by stray light, although the intensity and the degree of polarization suffered from the contamination separately.

A method of distance determination of evolved carbon stars is proposed and applied to six stars for which the terminal velocity of expansion and the angular size are obtained from the CO and HCN radio observations made at the Nobeyama Radio Observatory. The method assumes the radiation driven wind in spherical geometry. Within those small samples, however, two types of mass losing stars that are systematically different in the antenna temperature ratio TA (CO)/TA(HCN) ≷1, in the velocity ratio υ(CO)/υ(HCN)≶1, and in the angular size ratio θ(CO)/θ(HCN)≷1 seem to be present, suggesting nonspherical geometry of the outflow.

A molecular line search in the range between 85 and 89 GHz has been performed in the circumstellar envelopes of 11 evolved stars. Emissions of 29SiO J=2−1,28SiO J=2−1, HCN J=1−0, H13CN J=1−0, HC5 N J=33−32, HCO+ J=1−0 transitions and other transitions of C2 H, C4 H, and C3 N have been observed in 11 stars. We have detected the ground state 29SiO J=2−1 maser in several stars. We have also detected HCN emission in VY CMa. A narrow H13CN spike feature near the central velocity has been found in the spectrum of CRL 2688.

Mira variables form an important subgroup of the red giant stars which are typical representatives of stars showing burnt material at their surfaces. Since the photospheres of Miras are not in hydrostatic equilibrium but are characterized by spherically very extended density stratifications, their properties and emitted spectra differ substantially from those of non-Miras, and any attempts to analyse Mira spectra by means of conventional techniques must fail. Non-hydrostatic models are needed for analysis work.

We computed a small set of exploratory M type (solar abundances) Mira model photospheres whose density distributions were taken from an improved modification of a pulsation model proposed by wood (1979). Besides the fundamental model parameters, i.e. mass, luminosity and radius (or effective temperature), the parameters of the density stratification (e.g. the heights of the density discontinuities at the shock front positions, or the effective gravity acting between two successive shock fronts) must be treated as free parameters within certain limits, owing to insufficient knowledge of the velocity stratification entering the pulsation model. Typical density and temperature stratifications are shown in Fig. 1 of Scholz (1987). Both CO lines, measuring the outflow and infall velocities of matter between the shock fronts, and selected moleculer band features (colors) were found to react sensitively to adjustments of the model parameters. As an example, Fig.1 shows the differences between CO line profiles computed from the near-maximum model of Fig.1 of scholz (1987) (Fig.1a) and those computed from a model in which the effective gravity between the shock fronts was increased by a factor of 1.5 (Fig. 1b). The same velocity stratification was adopted in of monochromatic radii proves to be a powerful tool of diagnostics of Mira photospheres (Scholz and Takeda 1987; Bessell et al. 1987).

At Pennsylvania’s Flower and Cook Observatory, instrumentation has been developed to measure simultaneously the four Stokes parameters of the filtered radiation field from a celestial source. The instrumental Q/U/V-parameters have been found to be very small and well-behaved. Thus far, the program has concentrated on cool bright giants and supergiants and on hot, evolving close binaries. A single season’s investigation of Alp Ori has already been reported (Holenstein 1987) and the present paper is a summary of current results for the cool, evolved program stars.

For Psi1 Aur, V CVn, 6 Gem, 72 Leo and 119 Tau no V-signal at the level of 3σhas been detected from data from the 1986-1987 season. At the level of 0.0n%, unambiguous and variable V-signals have been detected for VV Cep, Mu Cep, Alp Her, Alp Ori, Bet Peg, and Alp Sco.

Many processes in the convection zone of a star affect the evolution and the atmospheric diagnostics. Here, a progress report is given on our numerical study of some convection related phenomena. The numerical results are obtained by solving the Navier Stokes equations in a three dimensional rectangular domain. The units are chosen such that the initial temperature, pressure, density, and the depth of the domain are all normalized to 1.

The mixing length theory relates the convective (enthalpy) flux FC to the envelope structure quite well (Chan and Sofia 1987). For efficient convection that occurs in deep convective regions, the numerical results are compatible with a mixing length ratio of 2.1. The mixing length theory fails to address the significance of the flux of kinetic energy FKE (see Figure 1). FKE is negative and has a magnitude comparable to the total flux. These results are qualitatively similar to those of two dimensional computations (Hurlburt et. al. 1984).

Hydrodynamic simulations of non-linear pulsation for less-massive cooler supergiants have been perfoemed by several authors (Tuchman, Sack and Barkat, 1979; Fadeyev and Tutukov, 1981; Fadeyev, 1982, 1984; Nakata, 1987; Buchler et al., 1987). The outburst of large amplitude pulsation at times is one of common features of these models, and renders mass-loss from the atmosphere of pulsating stars by generating strong shock waves.

To find out routes of the transition from limit cycles to this type of irregular pulsation, we performed hydrodynamic simulations for a series of models the luminosity log(L/L⊙ ) = 3.505, and Te = 5300 K with a narrow range of the mass, 1.4 M⊙<M<1.5 M⊙ by using the hydrodynamic code, TGRID (Simon and Aikawa, 1986).

Stellar pulsation is one of the condidates for strong mass loss from red giant stars. Recent investigations have shown sporadic outbursts of the pulsation can eject a considerable mass from the stars. Such a sporadic increase of the amplitudes seems to have a connection with the irregularity found in the pulsations. Recently the iregular propreties in stellar pulsation are investigated in the relation to nonlinear dynamics (see for instance Perdang, 1985). Unfortunately no single model oscillator of a star of which the equilibrium state is dynamically stable had been found. In the present paper, we shall discuss a simple oscillator which shows period-doubling and chaotic, that is, irregular oscillations.

There are several methods for solving a structure of a stellar atmosphere in radiative equilibrium. Among them, the matrix method (Nariai and Shigeyama 1984, Nariai and Ito 1985) uses only the energy equation at the nodes; therefore, it is easy to extend this method to an atmosphere that is not in radiative equilibrium. The matrix method gives direct solutions for the cases of a gray model or a picket–fence model, and it is so effective that the calculation converges in 3 or 4 iterations in the case of a non-gray atmosphere with scattering even if the starting model is isothermal. In order to simplify the problem, in this paper, however, we will limit ourselves to the case of a gray atmosphere with pure absorption.

Through the analysis and interpretation of observational material, particularly on the part of Otto Struve and his collaborators, the structure of an interacting binary is depicted (cf. Sahade and Wood 1978) as formed by a) the two stellar components; b) a gaseous stream from the less massive and more evolved component of the system towards the companion; c) a circumstellar gaseous envelope –designated as “ring” or “disk” depending on the density of the material; d) a circumbinary gaseous envelope that surrounds the whole system and is normally in expansion, as suggested by the conventional, ground–based observations.

The evolution of interacting binaries involves the effect of matter outflow and transfer in at least a stage of rapid mass loss and a stage of slow mass loss. As a result, the evolution of the components appreciably departs from the evolution of single stars and produces very bizarre objects which find no counterparts among non-binaries. Both the observational and the computational results suggest that the amount of mass involved in the process of mass outflow must be a large percentage of the total mass of the evolving component. It seems, therefore, reasonable to expect to find evidence, in evolved systems, for processed material from the interior of the mass-losing component.

The preceding invited paper by Sahade (1987) gave a balanced overall picture for the subject matter. I shall attempt to complement his talk in discussing two classes of atmospheric diagnostics of the evolutionary processes in interacting binary systems. These are: (I) Abnormal abundances of elements in the atmosphere resulting from the nuclear processes in the stellar interior. (II) Mass flow as a consequence of the evolution of either of the components and the Algol type binaries.

There are several classes of objects with abundance anomalies, such as Am stars. At least some of these objects are known close binaries. In particular, most Am stars appear to be close binary stars. However, the physical causes of the metallic overabundances in those stars are still unclear. We shall in the following examine Algol type binaries for abundance anomalies.

W Ursae Majoris stars can be understood as contact binary stars with a common envelope (Lucy 1968). They subdivide into two types: The A-type are earlier inspectral class than about F5, are believed to have radiative envelopes, and associate primary (deeper) eclipse minimum with transit eclipse. The W-type have spectral classes later than F5, are believed to have convectlve envelopes, and associate primary minimum with occultation eclipse. Controversy has surrounded the explanation of W-type light curves.

Four distinct models have been introduced to describe the envelopes or photospheres of W UMa stars. (1) The Rucinski hot secondary model directly explains W-type light curves on a postulational basis. Since 70%-90% of the emitted radiation from the secondary (less massive) component is believed to reach the secondary via circulation currents from the primary, there is an apparent thermodynamic mystery why the secondary should be hotter. (2) The Lucy Thermal Relaxation Oscillation (TRO) model argues that the secondary component is perpetually out of thermal equilibrium and that the components are in contact only during part of a given TRO cycle. During contact the photosphere is supposed to be barotropic. In this case primary minimum always associates with transit eclipse, in disagreement with observation for W-type systems. (3) The Shu et al. thermal discontinuity (DSC) model also argues for a barotropic photosphere but differs from Lucy on the gravity brightening exponent. The changes are insufficient to produce W-type light curves, (4) Webbink (1977), and, separately, Nariai (1976), argue for a baroclinic envelope. If the baroclinicity extends to the photosphere there is a possibility that W-type l i g h t curves could be explained. In particular, the Webbink scenario produces a hot secondary.

Recently, Kitamura and Nakamura have found that the anomalous gravity darkening occurs in semi-detached binary systems. The exponent of gravity darkening for the secondary components, which is defined by where F is the radiative flux and g is the gravitational accerelation, is significantly greater than the unity as shown in Table 1. We interpret this in terms of the energy transport by the mass outflow from the secondary component filling the Roche-lobe.

Table 1.

Empirical αc-values and physical quantities of five semi-detached binary systems.

From the analysis based on this interpretation, we show that 1) the mass out-flow carries energy if ∇ < ∇ad and 2) the anomalous gravity darkening requires that the flow must originate from deep interior. The mass loss rate can be estimated from the anomalous gravity darkening. Thus, the anomalous gravity darkening can be used not only to estimate the mass transfer rate but also to probe the interior structure and the evolution of the secondary star of the semi-detached binary. This model links Morton’s instability[2] of the secondary with the accretion disk of the primary.

The ordinary semi-detached close binary system consists of a main-sequence primary and subgiant (or giant) secondary component where the latter fills the Roche lobe. From a quantitative analysis of the observed ellipticity effect, Kitamura and Nakamura (1986) have deduced empirical values of the exponent of gravity-darkening for distorted main-sequence stars in detached systems and found that the empirical values of the exponent for these stars with early-type spectra are close to the unity, indicating that the subsurface layers of early-main sequence stars in close binaries are actually in radiative equilibrium. The exponent of gravity-darkening can be defined by H ∝ gα with H as the bolonetric surface brightness and g as the local gravity on the stellar surface.

U Cephei (V = 6.8–9.0, P = 2.493 d) is an eclipsing binary consisting of a B7V primary and a G8III-IV secondary component. This binary is one of the semidetached Algol systems showing soft X-ray emission which is probably associated with a hot corona surrounding the secondary component (White and Marshall 1983).

We made spectroscopic observations of U Cep with the coudé image-tube spectrograph of the 1.9-m telescope at Okayama Astrophysical Observatory on October 14, 1986. We obtained four spectrograms with a dispersion of 16 Å mm-1 covering λ λ3700—4300 Å during the primary eclipse. The first two exposures were made in a total eclipse, while the last two were slightly after the third contact. The CaII H and K emission lines appear clearly in all the spectrograms. Figure 1 represents an intensity tracing of one of these spectrograms.

The evolution of the carbon abundance at the surface of both components of a mass-exchanging (Algol-type) binary is examined (fig. 1). Distinction is made between case B and case AB (fig. 2) of mass transfer, in view of the different timescales involved. In the mass accreting component thermohaline mixing is adopted when matter with decreasing hydrogen abundance is deposited on the surface.

It is shown that at the surface of the loser a very low C-abundance is present, while at the surface of the gainer different regimes occur. On the average the expected C-abundance on the gainer is clearly lower than the observed solar value, but far above the value at the surface of the loser. The variation in time during the mass-exchange process is compared to the values, derived from observation of several Algol-type systems.

RV Tauri stars are variable yellow supergiants and they display anomalous excesses in infrared radiation indicating the presence of cool and extended dust thermospheres, presumbly formed from the matter ejected by them (Gehrz 1972). AR Pup, recognised as a member of RV Tauri stars, shows strong CH and CN bands in the optical spectrum and is considered to be carbon-rich (Lloyd Evans 1974). During a programme of broad band polarimetric observations of carbon variables, AR Pup was found to exhibit a large and highly time-dependent linear polarisation (Raveendran & Kameswara Rao 1987).

During the period from 1987 January 20 to April 6, AR Pup was observed on 33 nights through standard B and V filters with the 34-cm reflector of the Vainu Bappu Observatory, Kavalur. We also obtained UBVRI polarimetry of AR Pup with the PRL - Polarimeter (Deshpande et al. 1985) attached to the 102-cm reflector on six nights during the same period. Present observations clearly indicate that the large time dependent variations in polarisation, both in the amount and wavelength dependence, observed at earlier epochs by us were not isolated events but were segments of a continuous variation which is cyclic and related to the light curve.