Download Pickup ions near Mars associated with escaping oxygen atoms

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 107, NO. A8, 1170, 10.1029/2001JA000125, 2002
Pickup ions near Mars associated with
escaping oxygen atoms
T. E. Cravens, A. Hoppe, and S. A. Ledvina1
Department of Physics and Astronomy, University of Kansas, Lawrence, Kansas, USA
S. McKenna-Lawlor
Space Technology Ireland, National University of Ireland, Maynooth, Co. Kildare, Ireland
Received 3 May 2001; revised 4 December 2001; accepted 7 December 2001; published 7 August 2002.
[1] Ions produced by ionization of Martian neutral atoms or molecules and picked up by
the solar wind flow are expected to be an important ingredient of the Martian plasma
environment. Significant fluxes of energetic (55–72 keV) oxygen ions were recorded in
the wake of Mars and near the bow shock by the solar low-energy detector (SLED)
charged particle detector onboard the Phobos 2 spacecraft. Also, copious fluxes of oxygen
ions in the ranges 0.5–25 and 0.01–6 keV/q were detected in the Martian wake by the
Automatic Space Plasma Experiment with Rotating Analyzer (ASPERA) instrument on
Phobos 2. This paper provides a quantitative analysis of the SLED energetic ion data using
a test particle model in which one million ion trajectories were numerically calculated.
These trajectories were used to determine the ion flux as a function of energy in the
vicinity of Mars for conditions appropriate for Circular Orbit 42 of Phobos 2. The electric
and magnetic fields required by the test particle model were taken from a threedimensional magnetohydrodynamic (MHD) model of the solar wind interaction with
Mars. The ions were started at rest with a probability proportional to the density
expected for exospheric hot oxygen. The test particle model supports the identification of
the ions observed in channel 1 of the SLED instrument as pick-up oxygen ions that are
created by the ionization of oxygen atoms in the distant part of the exosphere. The flux of
55–72 keV oxygen ions near the orbit of the Phobos 2 should be proportional to the
oxygen density at radial distances from Mars of about 10 Rm (Martian radii) and hence
proportional to the direct oxygen escape rate from Mars that is an important part of the
overall oxygen loss rate at Mars. The modeled energetic oxygen fluxes also exhibit a spin
INDEX TERMS: 2780
modulation as did the SLED fluxes during Circular Orbit 42.
Magnetospheric Physics: Solar wind interactions with unmagnetized bodies; 2152 Interplanetary Physics:
Pickup ions; 5421 Planetology: Solid Surface Planets: Interactions with particles and fields; 6215 Planetology:
Solar System Objects: Extraterrestrial materials; KEYWORDS: solar wind interaction, pickup ions, oxygen
corona, Mars, ion acceleration
1. Introduction
[2] Mars lacks a global magnetic field so that the solar
wind interacts directly with that planet’s ionosphere and
atmosphere [Riedler et al., 1990]. However, the magnetometer aboard the Mars Global Surveyor (MGS) has
observed large (hundreds of nanoteslas), albeit localized,
magnetic fields [Acuna et al., 1998], which clearly must
have dramatic effects on the local plasma environment.
Also, the Electron Reflectometer experiment onboard
MGS has made interesting measurements relevant to the
solar wind interaction with Mars [e.g., Cloutier et al., 1999;
Crider et al., 2000]. Mars like Venus possesses a hot
1
Now at Space Sciences Laboratory, University of California at
Berkeley, Berkeley, California, USA.
Copyright 2002 by the American Geophysical Union.
0148-0227/02/2001JA000125$09.00
SMP
oxygen corona [Nagy and Cravens, 1988; Ip, 1990; Kim
et al., 1998; Hodges, 2000], which strongly affects the solar
wind interaction with those planets [e.g., Luhmann, 1991;
Luhmann and Schwingenschuh, 1990; Brecht, 1997; Kallio
and Koskinen, 1999; Kotova et al., 1997] by allowing heavy
pickup ions to be created within the solar wind flow.
[3] Fluxes of both hydrogen and oxygen ions with energies in the ranges 0.5– 25 and 0.01– 6 keV/q have been
recorded in the wake of Mars by the Automatic Space Plasma
Experiment with Rotating Analyzer (ASPERA) Ion Composition Experiment aboard the Phobos Mission to Mars and its
Moons [Lundin et al., 1989; Verigin et al., 1991; Dubinin et
al., 1993]. Kallio and Koskinen [1999] used a test particle
model to interpret these observations and concluded that the
ions are accelerated by the solar wind convective electric
field in the vicinity of Mars. Barabash et al. [1991] also
studied hydrogen pickup ions measured by the ASPERA
instrument onboard Phobos. The energetic particle telescope
7-1
SMP
7-2
CRAVENS ET AL.: PICKUP IONS NEAR MARS
solar low-energy detector (SLED) [McKenna-Lawlor et al.,
1990], onboard Phobos 2, observed significant fluxes of ions
with energies in excess of 50 keV (if oxygen) or 30 keV (if
protons) in the vicinity of Mars on a large number (albeit not
all) of Phobos 2 orbital passes [Afonin et al., 1989, 1991;
Kirsch et al. 1991; McKenna-Lawlor et al., 1993]. These ion
fluxes were most frequently found in the wake and bow
shock regions. Kirsch et al. [1991] and McKenna-Lawlor et
al. [1993] suggested that the observed ions were pickup ions
associated with the ionization of atoms in the hot oxygen
corona. To quote from the conclusions of Kirsch et al. [1991,
p. 767]: ‘‘Pickup ions, most likely constituting O+ and O2+
ions of E > 55 keV, were observed in channels 1 and 2,
especially during periods when Phobos 2 was spin stabilized
and the solar wind speed was >500 km/s.’’ It was also
suggested in this paper that ‘‘the pickup ions detected by
SLED should have their origin 5 – 6 104 km upstream of
Mars’’ (this is due to the very large gyroradius of a pickup
oxygen ion).
[4] Pickup ions are created by the ionization of neutrals
and are initially almost at rest in comparison with the solar
wind. In the absence of waves these ions follow cycloidal
trajectories. This behavior of pickup ions is well known
from studies of the solar wind interaction with comets [cf.
Galeev et al., 1991; Kecskemety et al., 1989]. The cycloidal
motion is a combination of E B drift and gyration. The
maximum ion speed is vmax = 2u sin VB, where u is the
solar wind speed and VB is the angle between the solar
wind velocity and the magnetic field direction. The maximum energy for O+ ions can easily exceed 100 keV (and
can be detected by SLED) if u and VB are sufficiently large
(also see Kirsch et al. [1991]). The distribution function
associated with pickup ions is a ring beam distribution. As
has been shown for comets (see, for example, Johnstone et
al. [1991]), this distribution is unstable against the growth
of very low frequency waves, which then can modify the
distribution via pitch angle and energy diffusion. However,
in the case of Mars where a pickup O+ gyroradius is several
Martian radii, waves will not have sufficient time to grow
significantly and thus modify the ion distribution function.
Consequently, the distribution will be nongyrotropic. This
scenario is similar to the ion pickup process at Pluto as
described in the test particle model of Kecskemety and
Cravens [1993]. Note that the gyroradius of a pickup O+
ion in the unshocked solar wind is several Martian radii near
Mars [Luhmann, 1991], so that in order to be accelerated to
an energy in excess of tens of keV, as observed by SLED,
such an ion would have to be born at least several Martian
radii upstream of the planet, or the observation point.
[5] In this paper, we present the first quantitative analysis
of the SLED energetic (up to 100 keV) oxygen fluxes using
test particle calculations of ion pick up in the vicinity of
Mars. Ion fluxes were calculated at four locations for Circular
Orbit 42 of the Phobos 2 spacecraft and compared with fluxes
measured by the SLED instrument. Note that Phobos 2
executed 114 circular orbits about Mars with period 8
hours and at a radial distance of 2.8 Rm. The radius of Mars
is denoted 1 Rm = 3398 km. The model predictions for the ion
fluxes observed by SLED are directly proportional to the
neutral oxygen density at distances upstream of the planet of
roughly 5 – 15 Rm. This oxygen population should consist
almost entirely of escaping atoms; hence the ion fluxes
Figure 1. Schematic of Phobos 2 orbit, orientation, and
solar low-energy detector (SLED) look direction and view
cone. The spin axis of the spacecraft points toward the Sun
(to the left). The asterisks indicate the centers of three of the
four sampling bins (really planes): bins at the inbound
shock, the wake, and the upstream region.
measured by SLED should be proportional to the ‘‘direct’’
component of the escape flux of oxygen atoms from Mars.
This direct component derives from the hot oxygen produced
by dissociative recombination of O2+ ions in the ionosphere
[Nagy and Cravens, 1988; Fox and Hac, 1997]. Other
contributions to the overall oxygen loss from the planet can
come from the ionization of exospheric oxygen atoms on
ballistic trajectories closer to the planet or due to escape of
oxygen ions directly from the ionosphere in the tail. Good
agreement can be obtained between our predicted energetic
oxygen ion fluxes and the SLED fluxes for reasonable values
of the neutral oxygen densities.
2. Data From the SLED Instrument on Phobos 2
[6] The Phobos 2 spacecraft was sometimes spinning and
at other times was spin-stabilized. During Circular Orbit 42
the spacecraft spin axis was pointed toward the Sun (see
Figure 1). The SLED particle detector onboard Phobos 2,
described by McKenna-Lawlor et al. [1990], consisted of
solid-state detectors in two different telescopes which were
directed at an angle 55 with respect to the spacecraft-Sun
line (in order to avoid the direct impact of sunlight). This
line-of-sight direction is approximately in the direction of
the nominal interplanetary magnetic field once per spin.
[7] The front detector of telescope 1 (Te1) was covered
with 15 mg/cm2 aluminum, whereas telescope 2 (Te2) used
an additional aluminum foil of 500 mg/cm2, which absorbed
protons of energy <350 keV but allowed the detection of
35 – 350 keV electrons. Thus the count rate difference
between the ‘‘open’’ telescope Te1 and the ‘‘foil’’ telescope
Te2 allows differentiation between proton and electron
CRAVENS ET AL.: PICKUP IONS NEAR MARS
Figure 2. Simultaneous magnetic field and energetic ion
measurements made by the Phobos 2 (Revolution 42) SLED
and MAGMA (magnetometer) instruments during Circular
Orbit 42 on 2 March 1989. (top panels): Sun-spacecraft line
component of the magnetic field Bx and the magnetic field
magnitude B. (bottom) Fluxes of energetic ions recorded by
Te1 in the lowest-energy channel of SLED (34 –51 keV for
protons and 55– 72 keV for oxygen ions). The time period
includes a passage through the Martian wake. The neutral
sheet (i.e., the center of the tail) is located where the x
component of the magnetic field changes sign). The bow
shock crossings are also evident in the magnetic field data,
where the field strength changes abruptly. Adapted from
Afonin et al. [1991].
fluxes. In each case the opening angle of the telescope
apertures was ±20. The geometric factor of both telescopes
was 0.25 cm2 sr, and the time resolution was 240 s. The
lowest two energy channels of Te1 are relevant to our study
and utilized energy ranges of 34 – 51 and 51– 202 keV,
respectively, for protons and, owing to pulse height defect
effects, 55– 72 and 72– 223 keV for O+ ions, respectively
[McKenna-Lawlor et al., 1990; Kirsch et al., 1991]. The
instrumental geometry was such that particles reached the
detectors if they were within a cone with half-angle 20.
The differential flux is obtained for a given channel from
the raw number of counts over the integration period by
dividing by the geometric factor, the integration period, and
the energy channel width. This procedure gives the flux
averaged over the solid angle of the aperture, which underestimates the actual differential flux (or intensity) for highly
anisotropic distributions.
[8] The electronic thresholds of SLED 1 were set by
using a calibrated pulse generator. The energy thresholds for
protons were then calibrated from the energy loss curves in
silicon. McKenna-Lawlor et al. [1993] made a calculation
of the energy thresholds similar to an earlier calculation
reported on in the original instrument description paper
[McKenna-Lawlor et al., 1990]. Two SLED instruments
were, in fact, originally launched on the two spacecraft of
the Phobos mission (SLED 1 on Phobos 2 and SLED 2 on
Phobos 1). Phobos 1 and 2 were launched on 7 and 12 July
SMP
7-3
1988, respectively. SLED 2 was activated on 19 July and
SLED 1 on 23 July. Before ground contact with Phobos 1
was lost at the end of August 1988, both SLED instruments
were operating and obtained ‘‘cruise’’ data. The fluxes both
instruments recorded were closely related, indicating a
similar performance.
[9] Figure 2 shows SLED data for Phobos 2 Circular Orbit
42 for channel 1 [from Afonin et al., 1991]. The fluxes in
channel 2 were about a factor of 10 less than the channel 1
fluxes near the outbound bow shock crossing. Ion fluxes are
observed both near the bow shock crossings and in the
Martian wake (at the distance of the Phobos 2 orbit, 2.8
Rm). The bow shock crossings are evident in the magnetic
field strength jumps in Figure 2. The SLED ion fluxes show a
modulation at the spacecraft spin rate, which indicates that
the ion distributions were anisotropic. Typical peak ion fluxes
in channel 1 were 1000 cm2 s1 sr1 near the inbound
shock crossing and 3000 cm2 s1 sr1 near the outbound
shock. The statistical uncertainty associated with these fluxes
is less than one percent.
[10] The solar wind conditions appropriate for the test
particle simulation for Circular Orbit 42 were determined
from Phobos 2 data monitored by the TAUS experiment
[Rosenbauer et al., 1989] for a time just prior to the inbound
shock crossing. We used here average values of the measured solar wind parameters since these parameters displayed
some variability. The average solar wind speed was 600 km
s1 and the magnetic field components were Bx = 2 nT,
By = 4.8 nT, and Bz = 0.0 nT. The angle between the solar
wind velocity and the magnetic field was VB = 67 on
average. The coordinate system used in analyzing the
Afonin et al. [1991] data by Kirsch et al. [1991] is Mars
centered with the x axis directed sunward, the z axis directed
normal to the ecliptic plane, and the y axis completing the
right-handed coordinate system. However, for this paper we
used a coordinate system in which the x axis is directed
antisunward. Hence the magnetic field in our model points
180 away from the actual direction. The E B drift
direction is not changed by this transformation. The only
consequence for our calculations is that the vz velocity
component changes sign, which does not affect any
model-data comparisons we carry out in this paper. For
the pertinent solar wind conditions the gyroradius of a
pickup O+ ion is rc = 1.76 104 km = 5.2 Rm and the
maximum pickup ion speed is 1100 km s1, which corresponds to a maximum kinetic energy of 100.3 keV.
3. Model of Pickup Oxygen Ions Near Mars
[11] The model consists of two parts: (1) a test particle
(or Monte Carlo) part in which the trajectories of 106
randomly created ions are numerically determined and (2)
a magnetohydrodynamic (MHD) model part that is needed
to generate the electric and magnetic fields required by the
test particle code. The test particle model requires a distribution of birth points in space. The locations of these birth
points were chosen randomly with a probability proportional
to the ionization rate and hence to the neutral density
(assuming a constant ionization frequency). Because we
are mainly concerned with ions born more than 5 Rm from
Mars (well outside the bow shock), photoionization is the
main ionization source although charge transfer collision of
SMP
7-4
CRAVENS ET AL.: PICKUP IONS NEAR MARS
Figure 3. The trajectory of an O+ ion born a distance of 105 km from Mars is shown. The inner box is
the region within which the results of the MHD calculation are used for the E and B fields. Mars is
depicted to scale and the 4 sampling planes used are shown. The centers of the planes are at a radial
distance from Mars of 2.8 Rm, corresponding to the distance of the Phobos 2 spacecraft in its circular
orbit. The sample ion trajectory shown is cycloidal upstream of the shock, but the trajectory is not a
simple cycloid downstream where the ion encounters a larger field.
solar wind protons with neutral oxygen makes some contribution. The ionization frequencies for both of these
sources should be independent of distance from Mars outside the shock. We adopt a total ionization frequency, per
neutral, of 4.5 107 s1, appropriate for the period of
rising solar activity and for the heliocentric distance of Mars.
3.1 Neutral Density
[12] The ions observed by SLED, with energies in excess
of 55 keV, must be born at distances in excess of several Rm
from Mars. The relevant neutral oxygen atoms belong to the
hot oxygen corona that is created by the dissociative recombination of ionospheric O2+ [cf. Nagy and Cravens, 1988;
Lammer and Bauer, 1991; Kim et al., 1998; Hodges, 2000].
Most of the oxygen reaching the distances we are interested
in will be escaping oxygen. The calculations of Fox and Hac
[1997] demonstrated that about half of the O atoms thus
produced do indeed exceed the escape speed at the exobase.
Unfortunately, numerical models of the Martian exosphere
reported on in the literature do not extend out to distances of
several Martian radii, which we need. We instead extend out
to larger distances the results of exospheric hot oxygen
models for closer distances. We adopt the neutral atomic
oxygen profile presented in Figure 3 (high solar activity
case) of Kim et al. [1998] for the altitude range of 200 to
6200 km. The Hodges [2000] oxygen profile for noon and
the Kim et al. [1998] profile are consistent. We found that the
following expression fits the oxygen density profile rather
well for the 3000- to 6200-km altitude region:
nO ¼ nexo ðRm =rÞ2 ;
3
ð1Þ
where nexo = 4000 cm is a density at the exobase and r is
the radial distance from Mars. We employ equation (1) for
altitudes greater than 6200 km. Our calculated ion fluxes are
directly proportional to nexo. Our calculated results are
shown for this value of nexo, but it will turn out that better
agreement of the calculated ion fluxes with the SLED data
is obtained if nexo 800 cm3 rather than 4000 cm3.
[13] Exospheric densities at large distances from a planet
vary with radial distance r differently depending whether
the responsible particles are on ballistic, satellite, or escape
trajectories [Chamberlin, 1978]. Neglecting satellite orbits,
the escape component should dominate at radial distances of
many planetary radii. In this case the neutral density for
large radial distances (r Rm) should be roughly given by
an expression like equation (1) with nexo Fesc / huni,
where huni is a typical neutral speed. Hot oxygen produced
by the more energetic reaction channels of dissociative
recombination have speeds of 6– 7 km/s [Fox and Hac,
1997]. Using these speeds and the Hodges escape flux gives
nexo 300 cm3, which is about a factor of 10– 15 less than
the nexo value we used in equation (1). This discrepancy
suggests that a significant fraction of the oxygen in the
3000- to 6000-km altitude region (shown by Kim et al. or
by Hodges) might consist of oxygen atoms on bound
ballistic trajectories. If both satellite and ballistic orbits
were also included, the asymptotic radial dependence of
the exospheric density would be r3/2 rather than r2
[Chamberlin, 1978]), and the oxygen density near 10 Rm
would be about a factor of 2 greater than the values given by
this expression. However, it is not obvious that satellite
orbits are able to be populated at Mars. This aspect of the
Martian exosphere needs further study.
[14] A random number generator was used to create 106
ions whose initial radial distance r is greater than 1.5 Rm
and with a probability proportional to the neutral density
CRAVENS ET AL.: PICKUP IONS NEAR MARS
Figure 4. (top) Scatterplot in vx – vz space for the upstream
bin 12. The straight lines are projections into this plane of
the directions opposite to the modeled SLED look directions
indicated. The mark on the line indicates the speed for a 55keV oxygen ion. (bottom) Same as top panel but scatterplot
in vx – vy space.
profile mentioned above. The ions are created within a
cubical box centered at Mars with 4 105 km for each
side. Also, these initial ion positions are created with
spherical symmetry. This is certainly not entirely accurate
due to day-night differences in the hot oxygen distribution,
but this assumption should not make much difference for
this paper because we are interested in ions born upstream
of the planet.
3.2. Test Particle Model
[15] The test particle code we use for this paper is
essentially the same code used to investigate ion distribution
functions near comets [Cravens, 1989; McKenzie et al.,
1994], near Titan [Ledvina et al., 2000] and near Pluto
[Kecskemety and Cravens, 1993]. The equation of motion is
numerically solved for each ion that is created. The only
force included is the Lorentz force: F = q(E + v B),
where q = e is the charge, v is the velocity vector, and E and
B are the electric and magnetic fields, respectively, at the
location of the particle. The electric field is assumed to be
the motional electric field, E = u B, where u is the bulk
plasma (i.e., solar wind) flow velocity.
[16] Two types of fields (cases) were investigated: (1)
uniform fields everywhere (i.e., the upstream solar wind
field values) and (2) fields taken from a three-dimensional
global MHD model of the solar wind interaction with Mars
SMP
7-5
(which is briefly described later). The results shown in this
paper all come from the second (MHD) case, although
results from the first case are mentioned. The MHD fields
are used only inside a relatively small ‘‘inner’’ region (i.e., a
cube centered at Mars and with dimensions 5.4 104 km ).
Uniform fields are used outside this inner region. Figure 3
shows the trajectory of a typical pickup O+ ion born
upstream of Mars. Note the large size of the ion’s orbit in
comparison with the size of Mars. Ions are removed from
the simulation if they penetrate past the exobase (at an
altitude of 200 km).
[17] The ion trajectories are sampled as they cross four
planes (or spatial ‘‘bins’’), each with dimensions 4 Rm 4
Rm and each oriented perpendicular to the x axis (that is,
the bin surface is oriented perpendicular to the Sun-Mars
direction). Two of the bins are centered on the x axis at the
distance of the Phobos 2 spacecraft: one upstream of Mars
and the other in the wake. The other two bins are located
in the flanks and straddle the bow shock. The centers of
three of these bins (or planes) are indicated in Figure 1.
The sampling bins were made large in order to increase
their collection efficiency, but, even so, they collect less
than 1% of the total ions created in the overall simulation
box. This box was made very large in order not to
accidentally introduce any bias into the counting statistics.
Particles crossing the planes/bins are sorted in a number of
ways such as by their energy and direction or according to
their velocity components. In particular, ion fluxes were
determined in a way that mimics the Phobos/SLED viewing geometry. We arbitrarily chose 16 angles about the
spacecraft spin axis at which we oriented the ‘‘instrument
point direction.’’ For each of these look directions, only
ions traveling at an angle (with respect to the pointing
direction) of less than the SLED semi-acceptance angle of
20 were recorded and included in the flux for this look
direction. We were able to calculate ‘‘absolute’’ instrumental ion flux values by estimating the ratio of the number of
‘‘real’’ ions that would be created per second in the
simulation box to the number of ions (106) that we
randomly generated in the box.
3.3. Electric and Magnetic Fields From the
MHD Model
[18] We used a global 3-D single-fluid MHD model to
specify the electric and magnetic fields in the inner region
of our simulation box. The model included mass loading
(source) terms with both ion production (with the oxygen
densities discussed above) and dissociative recombination.
The upstream boundary conditions for this MHD model
were the solar wind conditions appropriate for orbit 42 and
discussed earlier. The MHD code used was the Zeus-3-D
code developed by the Laboratory for Computational Astrophysics at the National Center for Supercomputer Applications [cf. Hawley and Stone, 1995; Stone, 1999]. A variable
Cartesian spatial grid was employed with a spatial grid size
ranging from 325 km near Mars up to 805 km far from
Mars. This grid size was sufficiently small for a magnetic
barrier, a draped field magnetotail (or wake), and a bow
shock to all appear in the MHD model results. Note that
pickup O+ ions born upstream of Mars are insensitive to the
finer details of the field pattern due to their very large
gyroradii. The field strength in the subsolar magnetic barrier
SMP
7-6
CRAVENS ET AL.: PICKUP IONS NEAR MARS
produced for uniform fields (we performed this calculation
but do not show the results). The distribution is nongyrotropic (more upgoing ions than downgoing ions)
because the upgoing ions are born closer to Mars where
the neutral density is greater. The outlying scattered ions in
Figures 4 and 5, outside the pickup ion ring, result from
reflection from the shock and/or magnetic barrier downstream of the shock.
[20] In the wake the ion distribution is no longer a
simple ring beam distribution. The ion distribution is more
spread out and diffuse. Ions, which are moving upward,
have their velocity vectors shifted downward as the ions
move past the shock and into the wake. This is due to the
larger Lorentz force associated with the increase in the
magnetic field experienced by the ions as they approach
Mars and pass through the magnetic barrier. This change
of velocity has the effect of removing ions from the top
part of the ring distribution and shifting them to lower vz
values. Ions are also deflected away from the E B drift
Figure 5. (top) Same as Figure 4 (vx –vz scatterplot) but
for a bin in the flanks near the outbound bow shock (bin 2).
The lines indicate projections of the directions opposite to
the modeled instrument pointing direction for the spin
number indicated. The mark on the line indicates the speed
corresponding to a 55-keV oxygen ion. (bottom) Same as
top panel but scatterplot in vx – vy space.
is 21 nT, which is a factor of 4 greater than the measured
upstream IMF value of 5.2 nT. The modeled bow shock in
the subsolar region is located at a radial distance of 1.9 Rm
and in the flanks it is located at 3.1 Rm. These distances are
15– 20% greater than the average subsolar and flank bow
shock positions found in Phobos 2 data [Trotignon et al.,
1993] but considerable variability exists in measured bow
shock positions [Schwingenschuh et al., 1990]. In fact, the
distance to the flank bow shock estimated using the location
of the shock crossings for Orbit 42 (i.e., from Figure 2) is
roughly 2.8 Rm, which is not that different than the distance
obtained with the MHD model.
4. Results
4.1. Velocity Space Scatterplots
[19] The velocity vector of each ion was recorded only
if it crossed one of the four bins/planes described in the
model section of the paper. Scatterplots are shown in
Figures 4 – 6. Ions are accepted from a full 4p solid angle
for these scatterplots. The upstream distribution is a
slightly modified nongyrotropic ring beam distribution
and looks very similar to the distribution that would be
Figure 6. (top) Same as Figure 4 (vx – vz scatterplot) but
for the wake bin 15. The lines indicate projections of the
directions opposite to the modeled instrument pointing
direction for the spin number indicated. The mark on the
line indicates the speed corresponding to a 55-keV oxygen
ion. (bottom) Same as top panel but scatterplot in vx – vy
space.
CRAVENS ET AL.: PICKUP IONS NEAR MARS
SMP
7-7
Figure 7. O+ fluxes versus energy for the bin located in the flank and straddling the inbound bow shock
for a SLED geometry with an acceptance cone with half-angle 20. Fluxes were calculated for 16 look
angles in the spin cycle, but for this spatial bin particles were registered only for 8 of these directions, as
shown. Peak SLED channel 1 and 2 count rates for the shock (scaled as indicated, see text) are indicated
in two of the panels.
direction (in the vx – vy plane) and toward the solar wind
direction.
4.2. Calculated Ion Fluxes
[21] We simulated the ion fluxes that SLED should see at
16 angles around its spin cycle at our four different bin
locations. Only ions whose velocity vectors put them within
the 40 instrument (full-width) acceptance cone were
recorded for each of the 16 look directions. The ions were
further sorted according to their energy in 5-keV bins.
Figure 7 shows results for the inbound bow shock crossing
for all instrument look directions for which any ions were
recorded (in this case, 8 of the directions recorded fluxes
while the other 8 directions did not record any ions). The
largest fluxes of O+ ions with energies near 55 keV (the SLED
lower energy limit) were recorded for directions 10 and 11,
whereas smaller fluxes were recorded for direction 5 and
very little or no flux for the other directions. These directions
(i.e., 5, 10, and 11) are indicated by the straight lines in
Figures 4 – 6. For the two shock crossings or for the upstream
case, directions 10 and 11 intersect the ring distribution at
positive vz where the ion flux is large, and direction 5
intersects the ring at negative vz where the ion flux is much
smaller. This asymmetry is due to the nongyrotropy of the ion
distribution which, in turn, is due to increasing production
rate with decreasing distance from the planet.
[22] Figure 8 shows flux versus energy for the other three
spatial locations but only for the look direction (or spin
angle number) at which the flux was a maximum. Some
SLED data from Afonin et al. [1991] are also indicated in
Figure 8. The Afonin fluxes are multiplied by the solid
angle of the instrument (0.39 sr). The SLED fluxes were
spin modulated, and we also chose the peak fluxes for these
comparisons; however, SLED had lower time/angle resolution than our model did, which introduces a factor of 3
difference in the peak flux. The model ion fluxes are about
a factor of 10– 20 greater than the measured fluxes and a
scaling factor is included with the data shown in Figure 8
in order to facilitate the comparison. The channel 2 (72 –
223 keV) SLED fluxes were also placed entirely into a
single 5-keV interval for comparison purposes, which
introduces another scaling factor of 29.6 for this channel.
[23] We also integrated the model ion fluxes over energy
for the SLED channels 1 and 2 energy ranges (55 –72 and
72– 223 keV, respectively) in order to obtain a ‘‘SLED’’
count rate for each of the 16 look directions (0 through 15)
and for each of the four locations around Mars. However,
we still used the better angle resolution of the model in
order to see how sharp the peaks are for the angular
dependence (i.e., time dependence). The results of this
procedure are shown in Figure 9. The spin variations look
very similar for the two shock locations and for the
upstream location; the maximum flux appears near spin
numbers 10 and 11. On the other hand, for the wake the
maximum flux appears at spin look direction number 5. The
model ion fluxes are confined to a rather narrow angular
range in Figure 9 (2 spin numbers or 45). The SLED
integration period is about a third of a spacecraft spin period,
which is 5 to 6 of our spin angular intervals. Consequently, the peak fluxes in a spin cycle from Figure 9 should
be divided by a factor of 3 before being compared with the
SLED maximum flux (in a spin cycle) (from Figure 2).
Table 1 shows these comparisons for the two shock locations and for the wake. To reiterate, both the instrument and
SMP
7-8
CRAVENS ET AL.: PICKUP IONS NEAR MARS
wake of Mars that are about a factor of 6 times greater (after
adjustment for the fraction of spin cycle covered by the
measurements versus the model) than the fluxes observed
by the SLED energetic particle instrument onboard the
Phobos 2 spacecraft during Circular Orbit 42. More accu-
Figure 8. Same as Figure 7, but results are only shown for
the look direction that has the largest flux near an ion energy
of 60 keV. (top) Fluxes for the bin upstream of the bow
shock. (middle) Fluxes for the bin straddling the outbound
shock. (bottom) Fluxes for the bin in the wake of Mars and
peak SLED channel 1 fluxes for the wake are shown.
the simulated instrument had 40 wide acceptance cones,
which results in lower peak particle intensities (flux per unit
solid angle) than the real pickup ion distribution with its
narrow ring. The factor of 3 in Table 1 is just the difference
in integration period between the SLED data and our
simulated SLED ‘‘data.’’
5. Discussion and Summary
[24] Our combined test particle and MHD model (with
nexo = 4000 cm3 in equation (1)) predicts ion fluxes in the
Figure 9. Model flux for SLED channels 1 and 2 energy
bins versus spin orientation for both shock bins, the
upstream bin, and the wake bin. Shock 1 is inbound, and
shock 2 is outbound. The model fluxes are treated in the
same way as the SLED fluxes except that higher time
resolution (i.e., angular resolution, 16 spin positions) is
retained for the model.
SMP
CRAVENS ET AL.: PICKUP IONS NEAR MARS
7-9
Table 1. Model-SLED Comparisonsa
Data (maximum in spin cycle)
Model (maximum in spin cycle)
Model times 1/3 (modified maximum)
Model times 1/3 1/6 (modified maximum and for lower O densities)
a
Shock 1
Shock 2
Wake
1200
33,000
11,000
1830
3000
39,000
13,000
2170
300
3300
1100
180
Oxygen flux 55 – 72 keV (channel 1) (cm2 s1 sr1).
rately, this is true only for those time periods when SLED
actually did observe any ion flux. An examination of the
Orbit 42 data in Figure 2 indicates that for many time
periods SLED did not record any oxygen flux. First, let us
consider this issue and then later we will discuss the factor
of 6 model/data ratio.
[25] The model predicts the existence of significant
pickup ion fluxes outside the bow shock, whereas SLED
only occasionally measured significant ion fluxes outside
the shock. As suggested by Afonin et al. [1991], it is
difficult for pickup ions to enter the narrow SLED entrance
cone at the energies measurable by SLED. Pickup ion
distributions, especially upstream of the shock, depend very
sensitively on the solar wind speed and on the IMF
direction. For the interplanetary conditions pertaining during Circular Orbit 42 (and adopted in the model) the ring
beam distribution was intersected by look directions 10 and
11 (within the half-angle of 20), but if the IMF angle with
respect to the solar wind direction had been 15 less than
the value used (VB = 67) no intersection would have been
present for energies in excess of the SLED channel 1 lower
limit of 55 keV. The magnetic field values we adopted for
Orbit 42 were ‘‘average’’ values estimated from upstream
conditions, but an examination of Figure 2 indicates that
even during this one orbital pass the IMF direction varied
considerably. These variations were probably sufficient to
move the ring beam distribution in and out of the SLED
acceptance cone at different times.
[26] Now we discuss the factor of 6 difference between
the model ion fluxes and the measured fluxes near the
shocks and in the wake. The model ion fluxes are directly
proportional to the hot oxygen density near the birth points
of those ions that reach energies of 55 keV near the
Phobos 2 orbit. These birth points must be located further
from Mars than 10 Rm due to the several Rm gyroradii of
pickup O+ ions. Hence the simplest explanation for the
factor of 6 is that the neutral oxygen density used in the
model (see equation (1)) at such large distances was too
large by this factor. That is, closer agreement would have
been obtained if our O density at 10 Rm was not 40 cm3
but was only 7 cm3. The oxygen population at this
distance primarily consists of escaping oxygen. This is
equivalent to using nexo 800 cm3 rather than nexo =
4000 cm3 in equation (1). As discussed earlier, this smaller
density is not too different from the 300 cm3 we crudely
estimated using the Hodges [2000] escape rate. As mentioned in the model description our O density was based on
the Kim et al. [1998] densities near 2 – 3 Rm; it seems that
the hot oxygen in this near-Mars region must include a
significant fraction of gravitationally bound O atoms as well
as escaping atoms. In any case the flux of 55– 72 keV
oxygen ions seen at the orbit of the Phobos 2 by SLED
should be proportional to the oxygen density in the vicinity
of a radial distance from Mars of 10 Rm, and, hence,
proportional to the direct escape rate of oxygen produced
by the dissociative recombination of molecular oxygen ions
in the ionosphere. Oxygen can also be lost from the planet
due to ionization of bound oxygen atoms located closer to
the planet than 10 Rm yet still in the solar wind or due to
direct outflow of ionospheric ions (mainly on the nightside).
[27] In summary, our quantitative model supports the
identification of the ions observed in channel 1 of the SLED
instrument as pickup oxygen ions that are created by the
ionization of oxygen atoms in the distant part of the exosphere, as suggested by Kirsch et al. [1991]. Our model also
explains the spin modulation exhibited by the SLED data.
Furthermore, the flux of energetic oxygen ions observed
near the orbit of the Phobos 2 provides new information on
the distant part of the hot oxygen corona.
[28] Acknowledgments. Support from NASA Planetary Atmospheres
grant NAG5-4358 and NSF grant ATM-9815574 as well as financial
support for the SLED instrument from the Irish National Board for Science
and Technology is gratefully acknowledged. Computational support from
the Kansas Center for Advanced Scientific Computing and NSF MRI grant
is also acknowledged.
[29] Janet G. Luhmann thanks Stas Barabash and Hannu E. J. Koskinen
for their assistance in evaluating this paper.
References
Acuna, M. H., et al., Magnetic field and plasma observations at Mars: Initial
results of the Mars Global Surveyor mission, Science, 279, 1676, 1998.
Afonin, V., et al., Energetic ions in the close environment of Mars and
particle shadowing by the planet, Nature, 341, 616, 1989.
Afonin, V. V., et al., Low energy charged particles in the near Martian space
from the SLED and LET experiments aboard the Phobos-2 spacecraft,
Planet. Space Sci., 39, 153, 1991.
Barabash, S., E. Dubinin, N. Pissarenko, R. Lundin, and C. T. Russell,
Picked-up protons near Mars: Phobos observations, Geophys. Res. Lett.,
18, 1805, 1991.
Brecht, S. H., Hybrid simulations of the magnetic topology of Mars,
J. Geophys. Res., 102, 4743, 1997.
Chamberlin, J. W., Theory of Planetary Atmospheres, Academic, San Diego, Calif., 1978.
Cloutier, P. A., et al., Venus-like interaction of the solar wind with Mars,
Geophys. Res. Lett., 26, 2685, 1999.
Cravens, T. E., Test particle calculations of pick-up ions in the vicinity of
comet Giacobini-Zinner, Planet. Space Sci., 37, 1169, 1989.
Crider, D. H., et al., Evidence for electron impact ionization in the magnetic
pile-up boundary of Mars, Geophys. Res. Lett., 27, 45, 2000.
Dubinin, E., R. Lundin, O. Norberg, and N. Pissorenko, Ion acceleration in
the Martian tail: Phobos observations, J. Geophys. Res., 98, 3991, 1993.
Fox, J. L., and A. Hac, Spectrum of hot O at the exobases of the terrestrial
planets, J. Geophys. Res., 102, 24,005, 1997.
Galeev, A. A., R. Z. Sagdeev, D. Shapiro, V. I. Shevchenko, and K. Szego,
Quasi-linear theory of the ion cyclotron instability and its application to
the cometary plasma, in Cometary Plasma Processes, Geophys. Monogr.
Ser., vol. 61, edited by A. D. Johnstone, p. 223, AGU, Washington, D. C.,
1991.
Hawley, J. F., and J. M. Stone, A numerical technique for astrophysical
MHD, Comput. Phys. Commun., 89, 127, 1995.
Hodges, R. R., Distributions of hot oxygen for Venus and Mars, J. Geophys. Res., 105, 6971, 2000.
Ip, W.-H., The fast atomic oxygen extension of Mars, Geophys. Res. Lett.,
17, 2289, 1990.
SMP
7 - 10
CRAVENS ET AL.: PICKUP IONS NEAR MARS
Johnstone, A. D., D. E. Huddleston, and A. J. Coates, The spectrum and
energy density of solar wind turbulence of cometary origin, in Cometary
Plasma Processes, Geophys. Monogr. Ser., vol. 61, edited by A. D.
Johnstone, p. 259, AGU, Washington, D. C., 1991.
Kallio, E., and H. Koskinen, A test particle simulation of the motion of
oxygen ions and solar wind protons near Mars, J. Geophys. Res., 104,
557, 1999.
Kecskemety, K., and T. E. Cravens, Pickup ions at Pluto, Geophys. Res.
Lett., 20, 543, 1993.
Kecskemety, K., et al., Pickup ions in the unshocked solar wind at comet
Halley, J. Geophys. Res., 94, 185, 1989.
Kim, J., A. F. Nagy, J. L. Fox, and T. E. Cravens, Solar cycle variability of
hot oxygen atoms at Mars, J. Geophys. Res., 103, 29,339, 1998.
Kirsch, E., et al., Pickup ions (E O+ > 55 keV) measured near Mars by
Phobos-2 in February/March 1989, Ann. Geophys., 9, 761, 1991.
Kotova, G., et al., Study of the solar wind deceleration upstream of the
Martian terminator bow shock, J. Geophys. Res., 102, 2165, 1997.
Lammer, H., and S. J. Bauer, Nonthermal atmospheric escape from Mars
and Titan, J. Geophys. Res., 96, 1819, 1991.
Ledvina, S., T. E. Cravens, A. Salman, and K. Kecskemety, Ion Trajectories
in Saturn’s magnetosphere near Titan, Adv. Space Res., 26(10), 1691,
2000.
Luhmann, J., The solar wind interaction with Venus and Mars: Cometary
analogies and contrasts, in Cometary Plasma Processes, Geophys.
Monogr. Ser., vol. 61, edited by A. D. Johnstone, p. 5, AGU, Washington,
D. C., 1991.
Luhmann, J., and K. Schwingenschuh, A model of the energetic ion environment of Mars, J. Geophys. Res., 95, 939, 1990.
Lundin, R., A. Zakharov, R. Pellinen, H. Borg, B. Hultqvist, N. Pissarenko,
E. M. Dubinin, S. W. Barabash, I. Liede, and H. Koskinen, First measurements of the ionospheric plasma escape from Mars, Nature, 341, 609,
1989.
McKenna-Lawlor, S., et al., The low energy particle detector SLED
(30 keV – 3.2 MeV) and its performance on the Phobos mission to
Mars and its moons, Nucl. Instrum. Methods. Phys. Res., Sect. A, 290,
217, 1990.
McKenna-Lawlor, S. M. P., V. Afonin, Y. Yeroshenko, E. Keppler, E.
Kirsch, and K. Schwingenschuh, First identification in energetic particles
of characteristic plasma boundaries at Mars and an account of various
energetic particle populations close to the planet, Planet. Space Sci., 41,
373, 1993.
McKenzie, M. L., T. E. Cravens, and G. Ye, Theoretical calculations of ion
acceleration in the vicinity of comet, J. Geophys. Res., 99, 6585, 1994.
Nagy, A. F., and T. E. Cravens, Hot oxygen atoms in the upper atmospheres
of Venus and Mars, Geophys. Res. Lett., 15, 433, 1988.
Nagy, A. F., and T. E. Cravens, Hot hydrogen and oxygen atoms in the
upper atmospheres of Venus and Mars, Ann. Geophys., 8, 251, 1990.
Riedler, W., et al., Magnetic field near Mars: First results, Nature, 341, 604,
1990.
Rosenbauer, H., et al., Ions of Martian origin and plasma sheet in the
Martian magnetosphere: Initial results of the TAUS experiment, Nature,
341, 612, 1989.
Schwingenschuh, K., W. Riedler, H. Lichtenegger, Y. Yeroshenko, K.
Sauer, J. G. Luhmann, M. Ong, and C. T. Russell, Martian bow shock:
Phobos observations, Geophys. Res. Lett., 17, 889, 1990.
Stone, J. M., The ZEUS code for astrophysical magnetohydrodynamics:
New extensions and applications, J. Comput. Appl. Math., 109, 261,
1999.
Trotignon, J. G., R. Grard, and A. Skalsky, Position and shape of the
Martian bow shock: The Phobos 2 plasma wave system observations,
Planet. Space Sci., 41, 189, 1993.
Verigin, M. I., The problem of the Martian atmosphere dissipation: Phobos
2 TAUS spectrometer results, J. Geophys. Res., 96, 19,315, 1991.
T. E. Cravens and A. Hoppe, Department of Physics and Astronomy,
University of Kansas, Lawrence, KS 66045, USA. ([email protected])
S. A. Ledvina, Space Sciences Laboratory, University of California,
Berkeley, CA 94720, USA.
S. McKenna-Lawlor, Space Technology Ireland, National University of
Ireland, Maynooth, Co. Kildare, Ireland.