R. Lallement,
J.T. Clarke,
Y. Malama,
E. Quémerais,
V.B. Baranov,
J.L. Bertaux,
H.J. Fahr,
F. Paresce
Service d' Aéronomie du CNRS, BP3, Verrieres le Buisson, 91371, France AOSS Department, University of Michigan, Ann Arbor, MI 48109 USA Institute for Problems in Mechanics, Prospect Vernadskogo 101, Moscow 199899, Russia IAEF, Bonn University, Auf dem Hugel 71, D5300 Bonn, Germany STScI, 3700 San Martin Drive, Baltimore, MD 21218 USA
(or
the LIC velocity), the bulk speed of the interstellar neutral hydrogen
appears to have decreased by about 6
during its approach to the
heliospheric interface. This is in qualitative agreement with models
including charge-exchange coupling with the decelerated plasma, itself
excluded from the heliosphere. According to a preliminary analysis the
observed line profiles suggest a significant velocity dispersion when
the viewing direction is parallel to the bulk flow, in agreement with
the models. These observations bring constraints on the interstellar
proton density (estimated to be of the order of 0.1 cm
) and, in
turn, on the size of the heliosphere (about 100 A.U.), of prime importance
for the future crossing of the interface by the two Voyager spacecraft.
A self-consistency check with the electron density deduced from GHRS
interstellar lines ratios towards the nearby stars is then obtained.
Keywords: heliosphere, solar wind, local interstellar medium
Thanks to its relative motion with respect to the Sun, neutral gas
from the Local Interstellar Cloud is invading the interplanetary space.
Neutral hydrogen atoms having succeeded in penetrating close to the
Sun scatter the solar Ly-alpha (121.6 nm) radiation, creating the
interplanetary-interstellar Lyman-alpha glow. Depending on their closest
distance to the Sun, atoms are or are not ionized by the solar EUV flux,
and by the solar wind through charge-exchange reactions with the protons.
As a result, a region void of visible neutral H (the ionization cavity) is
created in the so-called ``downwind'' region. In the supersonic solar wind,
the charge-exchange reaction (H + H
+ H) between
an incident (slow) interstellar H atom and a fast solar wind proton acts
as a loss process for the glow because it replaces the slow atom by a
fast neutral atom which, moving radially with the solar wind velocity,
is shifted out from the solar line. The Ly-alpha emission, when observed
from the earth's orbit, is almost entirely produced within the first 50
astronomical units, essentially due to the solar flux decrease with R,
although multiple scattering effects are present.
Classical models of this emission do assume a homogeneous flow ``at infinity'', characterized by a velocity V and a temperature T . For these models, infinity means at about 50 A.U., i.e., outside the ionization cavity, before any perturbation due to the Sun and the supersonic solar wind.
Spectra of the Ly-alpha emission (Figure 1a) recorded by the GHRS (Echelle A gratings and L S A), provide the best quality spectra ever obtained. Copernicus and IUE had already recorded a spectrum of the Ly-alpha glow, which have been analyzed by Adams & Frisch (1977), and Clarke et al. (1984,1995). Here, however, the observations combine high sensitivity, high resolution, and a good separation of the earth and interplanetary emissions. For each direction the date of observation is calculated in order to maximize the Doppler shift between the earth emission (the geocorona), and the interplanetary background. The geocorona has been used to calculate the point-spread function, and is then modeled and removed.
Figure: a: A GHRS spectrum of the diffuse Ly-alpha background: the strong
and narrow geocoronal emission is very well separated from the Ly-alpha
background. b: Schematic view of the solar wind-interstellar medium
interface
Such a simple view ignores the existence of the interstellar ionized fraction. Charged species of the inflowing interstellar medium are excluded from the heliosphere (the solar wind domain), and forced to flow around the heliopause, the discontinuity surface between the two media (Figure 1b). On the inner side of the interface, the supersonic solar wind, before reaching a pressure equilibrium with the ISM at the heliopause, is expected to be decelerated at the ``termination shock'', while on the outer side, the ISM plasma itself may (or not) be decelerated by a weak shock, depending on the actual values of the ambient ISM plasma density and magnetic field. The clue parameters for the interface are the interstellar plasma density and magnetic field. The magnetic field is known within a factor of 5 only. The interstellar plasma density has been estimated from the magnesium ionization equilibrium in the Local Cloud, itself measured from nearby stars absorption lines. It has been found to be of the same order or larger than the neutral density (Lallement et al. 1995, and this issue).
Interstellar neutral helium atoms are extremely weakly interacting with
the plasma, and are supposed to flow through the ISM/Solar wind interface
without perturbations. Indeed, He atoms have been detected inside the
heliosphere by the Ulysses probe, and found to be characterized by V=
25.7
and T =6700 K (Witte et al. 1993).
These values are in excellent agreement with the Local Cloud heliocentric velocity and temperature found from absorption lines in nearby stars spectra (Lallement et al. 1993, Linsky et al. 1993), providing a link between the heliosphere and the Local Cloud. Thanks to this agreement, the temperature and the relative velocity of all interstellar species before the entrance within the heliosphere are known accurately, allowing inside-outside comparisons for species interacting with the plasma.
Neutral H atoms suffer charge exchange reactions (H + H
+ H) with the protons, and for this reason, they are significantly
coupled to the plasma. This means that in the region where the
interstellar plasma is decelerated and deviated, to flow around the
heliosphere, a fraction of these perturbations is transferred to the
neutrals. Sophisticated self-consistent neutrals/plasma models, taking
into account the charge-exchange processes, allow to estimate the transfer
of plasma momentum to the neutral flow (Baranov & Malama 1993) .
The neutral population approaching the Sun is made from two components:
- atoms having not charge-exchanged : the primary component
- atoms resulting from charge-exchange reactions : the secondary component The proportions of these populations strongly depend on the plasma density in the ISM. The higher the plasma density, the larger the proportions of secondary atoms.
Figure: Secondary interstellar H atoms distribution in the heliosphere
from a self-consistent neutral-plasma model. The interstellar shock, the
heliopause, and the solar wind shock are the white lines. The interstellar
plasma density is 0.1 cm
Figure 2 illustrates the effect of the interface by showing secondary
atoms iso-density contours for an interstellar proton density of 0.1
cm
and a vanishing magnetic field. White lines are the terminations
hock, heliopause and interstellar shock surfaces. There is a marked
concentration of secondaries between the bow shock and the heliopause (the
``hydrogen wall''). Figure 3a shows the neutral H primary and secondary
densities along the sun/wind axis for the same model parameters, as
well as the total density, and Figure 3b displays the mean velocity when
combining the two populations. It is clear from these two figures that for
an observer located close to the Sun, the apparent density ``at infinity''
(at, say, 50--70 AU), is smaller than the true interstellar density,
due to the exclusion from the heliosphere of a fraction of the H atoms.
At the same time the apparent bulk velocity ``at infinity'', is also
smaller than the initial ISM velocity, because the newly created H atoms
are moving less rapidly.
Figure: a) Primary and Secondary H atoms densities along the Sun-Wind
axis on the upwind side. b) Combined primary and secondary H atoms bulk
velocity along the Sun-Wind axis on the upwind side. c) Primary and
secondary H atoms bulk velocities along the Sun-Wind axis on the up wind
side. The difference between the two velocities implies a large dispersion
in addition to the thermal dispersion, and a subsequent broadening of the
emission line. d) Same as c) for a crosswind direction (perpendicular to
the mean flow). The dispersion is very small in this direction, and the
emission line is not expected to be broadened
Figures 3c and 3d display the radial velocities of the two populations
along the upwind (from where the wind blows), and the crosswind (or side
wind= at 90
from the wind axis) directions. Evidently, there is a
strong velocity dispersion for the upwind L-O-S, because the secondary
atoms have been decelerated, while the surviving primary atoms have on
the contrary a velocity larger than the initial one, due to selection
effects (fast particles spend less time in the interface and suffer less
charge-exchange). Along the crosswind direction, both components have
velocities around zero, and there is little dispersion. The velocity
distributions are reflected by the Doppler shifts and the widths of
the Ly-alpha emission. This is why these effects can be searched for in
UV spectra.
Spectra were recorded for the upwind and crosswind (at right angle with the flow) directions. In the case of the crosswind spectrum, it has been possible to find a reasonable fit to the data with a classical model of an initial homogeneous flow (i.e., without any coupling with the plasma) at a temperature of about 8000K. This is in agreement with Figure 3d and the absence of velocity dispersion in this direction.
Figure: An upwind spectrum corrected for the geocorona. A classical model for an
initial velocity of 26
has been superimposed. The observed spectrum
is shifted towards lower heliocentric velocities. The adjustment requires
an initial bulk velocity of 19--20
. The observed line width
is also apparently broader than the model predicts.
The results were different for the upwind spectra. An upwind spectrum
is shown in Figure 4. It appears impossible to fit the data with
an homogeneous flow at an initial velocity of 26
(identical to
the LIC and helium velocity). The data implies a lower velocity of
about 19--20
. This is in agreement with Figure 3b, and suggests
that here has been a significant deceleration of the neutral flow.
The interstellar plasma density is then estimated to be of the order of
0.1 cm
. The modeled line profile is also found to be narrower than the
observed line (Fig. 4b), suggesting a significant velocity dispersion,
in agreement with the model velocities shown in Fig. 3c.
The GHRS Ly-alpha glow data show a deceleration and a dispersion of the
interstellar neutral H flow in the inner heliosphere, as theoretically
expected from coupling with the plasma at the heliospheric interface.
An ambient interstellar proton density of the order of 0.1 cm
is derived from a first rough comparison with the neutral/plasma models.
This density is of the same order than the electron density derived for
the Local Cloud, also with the GHRS, from nearby stars spectra showing
absorptions. This value implies a distance to the
heliopause of the order of 150 A.U., and to the solar wind termination
shock of the order of 90 A.U. This is close enough to the Sun for the
Voyager spacecraft to be still ``alive'' when crossing the shock.
Adams, T.F. & Frisch, P.C. 1977, ApJ, 212, 3000
Baranov, V.B. & Malama, Y. 1993, JGR, 98, 15157
Clarke, J.T., Bowyer, S., Fahr, H., & Lay, G. 1984, A&A, 139, 389
Clarke, J.T., Lallement, R., Bertaux, J.L., & Quemerais, E. 1995, ApJ, 448, 893
Lallement, R. et al. 1995, A&A, 304 (2), 461
Linsky, J.L., et al. 1993, ApJ,402, 694
Witte, M. et al. 1993, Adv. Space Res., 6, (13), 121