Astron. Astrophys. 361, 303-320 (2000)

1. Introduction

The solar system is embedded in a partially ionized cloud, conventionally called the Local Cloud, Local Fluff or Local Interstellar Cloud (LIC), which is the most local patch of interstellar medium (ISM). Absorption line studies show, that in the neighbourhood of the LIC there are also other cloudlets of comparable size like e.g. the G-cloud (cf. Lallement 1998, and references therein). The state of the LIC plasma is known within some uncertainty. From GHRS measurements onboard HST its temperature is inferred to be about (Linsky 1996). For the G cloud it could be even lower. The density inferred from line ratios, such as MgII /MgI , ranges from (Lallement et al. 1994) to , for NaI (Lallement & Ferlet 1997), along different lines of sight. Using a Doppler triangulation method, the LIC velocity vector in space could be determined, having a magnitude of about 26 km s^-1 (s. Lallement 1998), in agreement with in situ measurements of neutral helium from Ulysses (Witte et al. 1996). It is noteworthy, that for the G-cloud a value of 29 km s^-1 (s. Lallement 1998) has been inferred, and that for both clouds the velocity vector points away from the wall or Sco-Cen direction.

As we will show in some detail in this paper, the origin of the local clouds is intimately related to the origin of the local X-ray emitting cavity, called the Local Bubble, itself. There are currently several models concerning this issue. In this respect it should be emphasized that the pressure of the LIC is , where is Boltzmann's constant, and hence at least a factor of 4-5 lower than the Local Hot Bubble model predicts (e.g. Snowden et al. 1990) for the local soft X-ray emitting plasma on the basis of collisional ionization equilibrium (CIE). However, the pressure of the LIC is in agreement with the pressure of the Local Bubble as derived from non-CIE models (Breitschwerdt & Schmutzler 1994).

It would be straightforward to assume that the local X-ray emitting plasma represents a normal patch of hot interstellar medium (HIM). Our range of sight in X-rays beyond 100 pc would then be blocked by intervening HI clouds (Jakobsen & Kahn 1986). However, both the overall sphericity of the cavity (cf. Snowden 1998) and the canonical temperature of of the HIM (McKee & Ostriker 1977), which is on the low side for Raymond & Smith (1977) CIE model fits to the soft X-ray background (SXRB), argue against it. Moreover UV absorption line studies (Fruscione et al. 1994) could not find any significant amount of interspersed HI within the Local Bubble. At an average radius of 100 pc, however, there is a significant increase in the HI column density indicating the presence of a swept-up shell.

The most plausible model in our view is that the Local Bubble has been created by a number of successive supernovae, presumably or more years ago. Although there is some disagreement whether the X-ray emission is due to the reheating of the plasma by a recent supernova (e.g. Cox & Anderson 1982) or whether it is due to the slowly recombining gas of an extinct superbubble (Innes & Hartquist 1984; Smith & Cox 1998), it is suggestive that the Local Bubble is an individual object.

This is challenged in an alternative scenario by Frisch (1995, 1996), who has pictured the Local Bubble to be an appendix of the Loop I superbubble which expanded into a low density interarm region, an idea that had been proposed earlier by Bochkarev (1987). Such an expansion may have been caused by an epoch of star formation in the Sco-Cen association some years ago, and the local clouds surrounding the solar system would just be fragments of the expanding shell. Based on absorption line studies within 50 pc of the Sun, it is argued that the local ISM flow has velocities in the range of  km s^-1 and that the velocity vector is pointing away from Sco-Cen. It is however difficult to explain the existence of a neutral HI wall ¹ between the Local Bubble and Loop I at a distance of  pc in the Sco-Cen direction, which was inferred from ROSAT WFC star counts (Warwick et al. 1993) and optical and UV spectral analysis of stars in Loop I (Centurion & Vladilo 1991). The latter is still consistent with a later analysis of ROSAT EUV observations by Diamond et al. (1995), who place the rise in column density () towards the Galactic centre direction at a distance of 25-30 pc. As we will show in Sect. 4 the rise in column density to occurs at a distance of  pc using recent HIPPARCOS distances (see Fig. 9). Thus a third epoch of star formation has been postulated and the wall has been identified with the swept-up shell resulting from these explosions. The problem is then somehow shifted towards explaining how supernova remnants (SNRs) expanding into a hot low density medium can produce a neutral shell. Contrary to this view one would expect in this case that the shock wave weakens rapidly due to the high speed of sound in the medium it is propagating into, and the compression of a swept-up, expanding shell is weak. Therefore the cooling time is very large and the formation of a dense neutral shell is severely inhibited.

The North Polar Spur and its associated Loop I bubble with an huge apparent size covering a solid angle of on the sky have been studied in detail with the ROSAT PSPC instrument by Egger (1993) and Egger & Aschenbach (1995, henceforth EA95), using diffuse X-ray background maps of the ROSAT All-Sky Survey. Owing to its low instrumental background and high spatial resolution the PSPC is ideal for investigating the hot gas produced by supernova (SN) explosions in the nearby Sco-Cen association. It was shown (i) that the Loop I is still an active superbubble with SNe having already occurred and another to come, (ii) that the dense neutral HI shell found to be surrounding Loop I casts deep X-ray shadows in the  keV band, (iii) that a ring-like structure ("ring") with an even higher HI column density is located between Loop I and the Local Bubble, indicated by its strong 21 cm line emission and a strong anticorrelation with the ROSAT  keV band. By comparing these findings to numerical simulations of colliding bubbles (Yoshioka & Ikeuchi 1990), Egger (1993) concluded that the shell and ring structures were the result of an ongoing interaction between Loop I and the Local Bubble, with at least one bubble having already formed a dense and cool shell prior to the collision. In this picture the "wall" is just part of the interaction zone, held under pressure by the two colliding bubbles, and bounded by the ring. In the following we will adopt this model, because it can explain present observations in a consistent and physical way.

The purpose of this paper is to show that the existence of neutral clouds and their predominant flow away from the Sco-Cen association, is a natural consequence of the interaction between the bubbles and subsequent local fragmentation of the interaction zone (Fig. 1). The structure of the paper is as follows: Sect. 2 gives a detailed discussion of the interaction process and a thorough treatment of the resulting hydromagnetic Rayleigh-Taylor instability, in Sect. 3 the formation and dynamics of clouds is investigated and in Sect. 4 we compare the results of our calculations to ROSAT observations. Sect. 5 contains the discussion and our conclusions.

Fig. 1. Schematic representation (not drawn to scale) of the interaction between the Local Bubble and the neighbouring Loop I superbubble (longitudinal cut perpendicular to the galactic plane). The dark slab represents the dense, compressed interaction zone ("wall"), bounded by the "ring", as seen in absorption in the ROSAT images. Also shown are HI cloudlets that seem to be associated with the wall and have systematic velocities pointing towards us.

© European Southern Observatory (ESO) 2000

Online publication: January 29, 2001
helpdesk.link@springer.de