Our solar system is moving through a partially ionized interstellar medium. The ionized fraction of this local interstellar medium (LISM), interacts with the expanding solar wind and forms the LISM - Solar Wind interface. The characterization of this interface is one of the major objectives for the near future in astrophysics and space plasmas physics. For recent reviews see Holzer (1989), Axford (1989), Suess (1990), Baranov (1990), Lee (1993), Jokipii and McDonald (1995). The choice of an adequate model of the interface depends on the parameters of the parcel of interstellar gas surrounding the Sun, also called the Local Interstellar Cloud (LIC). Some of these parameters, as the Sun/LISM relative velocity, or the LISM temperature are now well constrained (Witte et al, 1993, Lallement & Bertin, 1992, Linsky et al, 1993, Lallement et al, 1995), but unfortunately there are no direct ways to measure the local interstellar electron (or proton) density, nor the local interstellar magnetic field, while these two parameters govern the structure and the size of our heliosphere. Therefore, there is a need for indirect observations which can bring stringent constraints on the plasma density and the shape and size of the interface to allow for the choice of an adequate theoretical model. Such constraints should help to predict when the two Voyager spacecraft will cross the interface, and whether or not they will be able to perform and transmit direct observations.
Among the various types of diagnostics are the observations of species which can penetrate deeply in the heliosphere, as the interstellar neutral atoms. These neutrals suffer significant modifications during the crossing of the interface (Wallis, 1975; Holzer, 1989; Osterbart and Fahr, 1992; Malama, 1991; Gangopadhyay and Judge, 1989). These modifications depend on the structure of the interface and the penetration factor of the atoms through the interface. Models show that the changes in the abundances and velocity distributions due to the interface filtering do vary strongly from one species to the other, because they depend on the strength of the coupling between the atom and the ionized species. As a result, the relative abundances and the velocity distributions of the different species inside the heliosphere are different from the original interstellar abundances and velocity distributions. In fact there are two types of diagnostics in connection with the perturbations of the neutrals. First, species-to-species comparisons within the heliosphere are in some cases sufficient to quantify the amount of perturbations at the interface. In particular, a long-standing debate has been the comparison between the properties of the heliospheric neutral helium and those of the heliospheric neutral hydrogen. In principle neutral helium velocity and temperature are identical to those in the original LISM due to the extremely weak neutral/plasma coupling. In this respect, analyses of the backscattered radiation in the resonance line of H I and He I , together with models of the penetration of interstellar neutrals into interplanetary space, have brought useful information on the neutral component of the LISM and have allowed a indirect detection of the terminal shock by showing evidence for a neutral H deceleration (Lallement et al, 1993).
The second method is the comparison between the local interstellar medium properties, which can be derived e.g. from nearby stars spectroscopy, and the properties of the interstellar species inside the heliosphere. Not only the neutrals themselves but also the two types of derivatives of the neutrals, the pick-up ions and the anomalous component of cosmic rays (ACR), can be used to constrain the perturbations at the interface. As a matter of fact, according to the now well established ACR theory (Jokipii J.R., 1971; Fisk L.A., Kozlovsky B., Ramaty R., 1974,Pesses et al, 1981, Cummings et al, 1984) the sources of the ACR are the LISM atoms which penetrate in the heliosphere where they get ionized. The resulting ions are convected with the plasma as pick-up ions towards the termination shock where they are accelerated to high energies by Fermi acceleration and the ACR are this fraction of the pick-up ions which, after outwards convection by the solar wind, has been accelerated. Pick-up ions have been observed close to the Sun (e.g. Mobius et al, 1988, Gloeckler et al, 1995) and ACR in the whole heliosphere with in particular the two Voyager (e.g. Cummings and Stone, 1990). Observations of both types of ions can potentially bring extremely important constraints on the relative abundances of the heliospheric neutral atoms (Fahr, 1991, Fahr et al, 1995, Frisch, 1995) just after their entrance in the heliosphere. These results can then be compared with LISM abundances, and, through reliable models predicting the filtration of the different species as a function of the interface parameters, a self-consistent picture should be buildt.
The major difficulty when modelling the coupling between heliospheric interstellar neutrals and the plasmas is related to the fact that, at variance with charged particles, neutral atoms have an interaction length as large as the characteristic length of the interface, and penetrate deeply into the heliosphere. A significant amount of work has been done on the heliospheric interface filtering of hydrogen (Wallis, 1975; Baranov et al., 1979; Ripken and Fahr,1983; Fahr and Ripken, 1984; Bleszynski, 1987; Gangopadhyay and Judge 1989, Baranov et. al., 1991, Osterbart & Fahr, 1992; Baranov & Malama 1993), and both hydrogen and helium (Bleszynski, 1987), primarily motivated by their large cosmic abundances and the availability of observations of resonantly scattered solar radiation by H and He atoms having penetrated the heliosphere.
There has been more recently an increasing interest in heavier elements of the LISM, in particular O, N, Ne, C and others. This was initially due to experimental projects devoted to the detection of the backscattered radiation in the resonance lines of these elements. In particular, solar radiation scattered by the oxygen particles could be measured in the far UV at the 1304 primary resonance line (Bowyer & Fahr, 1989; Fahr, 1989; Fahr et. al. 1995 and references therein). However, this interest in minor species is now growing due to the recent successfull detections of pick- up (Geiss et al, 1994) and ACR (Cummings et al, 1990) ions, with the Ulysses and Voyager spacecraft respectively. Oxygen is of particular interest, because it is one of the most perturbed elements due to its large charge exchange cross-section with the protons. Its entrance in the heliosphere has thus been thought to be a very good diagnostic of the interstellar plasma density (e.g. Fahr 1991). Initial calculations suggested a very small filtration factor, even for a very low interstellar plasma density. Here we call the filtration factor the ratio between the neutral oxygen density inside the heliosphere, at a large enough distance from the Sun for the direct solar wind and EUV ionization to be insignificant, and the initial interstellar neutral oxygen density outside the heliosphere, in the unperturbed interstellar medium. The filtration factor is a measurement of the coupling with the plasma, and then of the interstellar plasma density. According to these first results, the quasi-normal heliospheric oxygen abundance inferred from the ACR, implying a very large filtration factor, was incompatible with a non negligible LISM plasma density. Contrary to these previous calculations (Fahr, 1991), an improved modelling has shown that the degree of penetration is not so low, at least in the case of a Parker's type interface and a modified twin-shocks model with incompressible plasma and no gravitation (Fahr et al, 1995). Recently Frisch (1994) has demonstrated how oxygen measurements could be used in conjunction with carbon data to infer the interstellar electron density in the LISM. However, filtration of neutral oxygen and neutral carbon at the interface should be taken into account to obtain a more precise result.
In the present paper, we discuss the penetration of the neutral oxygen, whose cosmic LISM abundance is third, inside the heliosphere, through a realistic two shocks interface. The model allows for plasma compressibility, and includes gravitation. The large cross-section for charge-exchange with the protons, comparable to the neutral H cross-section, implies that the structure of the plasma interface has a strong influence on the characteristics of the oxygen atoms into heliosphere. However, there is an important difference between the two species. In the case of neutral hydrogen (charge-exchange H + H H + H), its abundance makes the quantity of newly created H and H large enough to influence the overall structure of the interface (Baranov and Malama, 1993). For oxygen, its small cosmic abundance makes the newly created species unimportant, and it is possible to neglect the effect of charge-exchange on the H and H parameters. It has to be noticed that neglecting the counter-back reaction onto the hydrogen and protons does not mean that the charge-exchange inverse reaction (O with H) is negligible in comparison with the direct charge-exchange (O with H ), something assumed in the past (Fahr and Ripken, 1984). As far as we are concerned with oxygen distribution, both reactions are equally important. The inverse process has been taken into account recently for the first time by Fahr & al (1995).
The basic principles for the present theoretical description (gas dynamics for the plasma, kinetic approach for the neutrals) can be found in Malama (1991). Instead of the solution of the Boltzmann integro- differential equation calculated by Osterbart & Fahr (1992) and Fahr & al (1995), the present calculation is based on a Monte - Carlo simulation of the flow of O and O through the H and H interface (Bleszynski, 1987; Gangopadhyay & Judge 1989; Malama, 1991; Baranov & Malama, 1993). It uses the Baranov two-shock interface model which is a self-consistent gasdynamic model of the solar wind interaction with the local interstellar medium, which takes into account the mutual influence of the plasma component of the LISM and the LISM H atoms in the approximation of axial symmetry (Baranov and Malama, 1993). The oxygen and oxygen ions simulation is started in such an already self-consistent H atoms and protons background.
Such a modelling of the neutral oxygen flow through the interface provides the production term for the pick-up ions, of direct relevance to pick-up and anomalous cosmic rays measurements. More precisely, it provides the number and the location of newly created oxygen ions in heliosphere due to ionization and charge-exchange processes, the density distribution of the pick-up ions, assuming new born ions are picked up by the plasma instantaneously, and subsequently the source term for the acceleration processes. Ideally, such a model applied to pick-up and ACR data should provide heliospheric interstellar abundances on one hand, which can be compared with original interstellar abundances on the other. Because, independently, the ACR spectral and spatial gradients observations by the outer probes constrain the shock location (and then the structure of the interface) through diffusion- convection-acceleration models (Cummings et al, 1993), then, if the interstellar plasma pressure is the dominant confining agent for the heliosphere, a self consistency check of this pressure (or of the plasma density) should be obtained in the future from the two types of modelling. Simultaneously, the heliospheric radio emissions detected by the Voyager, which depend on the interstellar plasma density too (Kurth and Gurnett, 1993), have to be found compatible with the other descriptions.
The core of this paper is divided into five sections. In Sect. 2 is recalled the basic structure of the Baranov and Malama (1993) LISM vs solar wind interaction region, which is the starting point for the oxygen flow simulation. Sect. 3 describes the mathematical formulation of the problem, while Sect. 4 describes the computational method. In the fifth section results are presented for both neutral and pick-up ions. In Sect. 6 we discuss further the particular importance of the oxygen penetration in relation with the observations.
© European Southern Observatory (ESO) 1997