Mass loss evolution and the formation of detached shells around TP-AGB stars

Abstract

Context. The origin of the so called "detached shells" around AGB stars is not fully understood, but two common hypotheses state that these shells form either through the interaction of distinct wind phases or an eruptive mass loss associated with a He-shell flash. We present a model of the formation of detached shells around thermal pulse asymptotic giant branch (TP-AGB) stars, based on detailed modelling of mass loss and stellar evolution, leading to a combination of eruptive mass loss and wind interaction. Aims. The purpose of this paper is first of all to connect stellar evolution with wind and mass loss evolution and demonstrate its consistency with observations, but also to show how thin detached shells around TP-AGB stars can be formed. Previous attempts to link mass loss evolution with the formation of detached shells were based on approximate prescriptions for the mass loss and have not included detailed modelling of the wind formation as we do here. Methods. Using stellar parameters sampled from an evolutionary track for a 2 M star, we have computed the time evolution of the atmospheric layers and wind acceleration region during a typical thermal pulse with detailed radiation hydrodynamical models including dust formation. Based on these results, we simulate the subsequent circumstellar envelope (CSE) evolution using a spherical hydrodynamic model. Results. We find that existing simple mass loss prescriptions all suggest different mass loss evolutions and that they differ from our detailed wind modelling. The most important factor for the formation of a detached shell is the wind velocity evolution which has a strong impact on the wind interaction and the resulting pile-up of matter. Our CSE model shows that a thin shell structure may be formed as a consequence of a rather short phase of intense mass loss in combination with a significant variation in the wind velocity, as obtained by our wind models. This situation can only be obtained for a limited range of amplitudes for the piston boundary used in the dynamic atmosphere models. Conclusions. The combined mass loss eruption and wind interaction scenario for the formation of detached shells around AGB stars (suggested by previous work) is confirmed by the present modelling. Changes in mass loss rate and wind velocity due to a He-shell flash are adequate for creating distinct wind phases and a "snow plow effect" that is necessary to form a geometrically thin detached shell. The derived properties of the shell (i.e. radius, thickness and density) are more or less consistent with existing observational constraints.

Unexpectedly large mass loss during the thermal pulse cycle of the red giant star R Sculptoris

“…Although a slight decrease in the expansion velocity of the detached shell is possible, we find no evidence of a significant decrease due to the sweeping-up of material from the pre-pulse wind. This is contrary to the current theory of how detached shells are formed during thermal pulses 3,4 . Also, no material is likely to have piled onto the shell due to the post-pulse mass loss.…”

“…a factor ≈30 lower than during the pulse. This general evolution of the mass-loss rate is consistent with stellar evolution models, however, the ratio between the derived pulse and prepulse mass-loss-rate is significantly higher than found in the models 4 . To further constrain the mass-loss-rate evolution of R Sculptoris, we modeled the system with a modified version of the GADGET-2 SPH code 14 , including detailed radiative cooling 15 .…”

“…3, and supplementary information). The derived evolution of the expansion velocity since the last thermal pulse is consistent with models of thermal pulses 4 . However, the observed emission implies variations in the expansion velocity of ±1.5 km s −1 on timescales of a few hundred years.…”

“…The overall velocity fits the predictions from theoretical models of thermal pulses well. However, velocity variations of ±1.5 km s !1 are apparent throughout the evolution, while theoretical models only predict significant variations in the expansion velocity <200 years after the pulse 4 . Theoretical models predict an increasing widening of the spiral with ± the sound speed, which may explain at least part of the velocity variations 8 .…”

“…Recent images of thermal dust emission show the detached shell, as well as a more distant, and spatially distinct, region of interaction with the interstellar material 12 , showing the presence of a stellar wind from R Sculptoris before the pulse. Collision with a previous, slower wind is required in order to prevent the thin shell from quickly diffusing 4,13 . The average CO line intensity in the area surrounding the shell sets an upper limit to the pre-pulse mass-loss rate, and its ratio to the average CO line intensity in the shell suggests an increase in mass-loss rate during the thermal pulse of a factor of !…”