Mercury's magnetic field in the MESSENGER era

5.7 MB·

by  Johannes Wicht

Mercury’s magnetic field in the MESSENGER era
J. Wicht1 and D. Heyner2
1Max-Planck Institut fu¨r Sonnensystemforschung, Go¨ttingen, Germany,
[email protected]
2Institut fu¨r Geophysik und extraterrestrische Physik, TU Braunschweig, Braunschweig,
7
1 Germany
0
2

n January 19, 2017
a
J

8
1 Abstract

] MESSENGER magnetometer data show that Mercury’s magnetic field is not only excep-
P
tionallyweakbutalsohasauniquegeometry. Theinternalfieldresemblesanaxialdipolethat
E
is offset to the North by 20% of the planetary radius. This implies that the axial quadrupol is
h. particularly strong while the dipole tilt is likely below 0.8◦. The close proximity to the sun in
p combination with the weak internal field results in a very small and highly dynamic Hermean
- magnetosphere. WereviewthecurrentunderstandingofMercury’sinternalandexternalmag-
o
r netic field and discuss possible explanations. Classical convection driven core dynamos have
t
s a hard time to reproduce the observations. Strong quadrupol contributions can be promoted
a by different measures, but they always go along with a large dipole tilt and generally rather
[
small scale fields. A stably stratified outer core region seems required to explain not only

1 the particular geometry but also the weakness of the Hermean magnetic field. New interior
v modelssuggestthatMercury’scorelikelyhostsanironsnowzoneunderneaththecore-mantle
0
boundary. The positive radial sulfur gradient likely to develop in such a zone would indeed
6
promote stable stratification. However, even dynamo models that include the stable layer
0
5 show Mercury-like magnetic fields only for a fraction of the total simulation time. Large scale
0 variations in the core-mantle boundary heat flux promise to yield more persistent results but
. are not compatible with the current understanding of Mercury’s lower mantle.
1
0
7 1 Introduction would be too small to support dynamo action.
1 The Mariner 10 measurements also indicated
:
v that Mercury’s magnetic field is special [Ness
i In 1974 the three flybys of the Mariner 10
X et al., 1974]. Being 100 times smaller than
spacecraft revealed that Mercury has a global
the geomagnetic field, it seems too weak to
r
a magnetic field. This was a surprise for many
be supported by an Earth-like core dynamo.
scientistssinceaninternaldynamoprocesswas
And though the data were scarce, they never-
deemed unlikely because of the planet’s rel-
theless allowed to constrain that the internal
ative small size and its old inactive surface
field is generally large scale and dominated by
[Solomon, 1976]. Either the iron core would
a dipole but possibly also a sizable quadrupole
have already solidified completely or the heat
contribution. Both the Hermean field ampli-
fluxthroughthecore-mantleboundary(CMB)
1
tude and its geometry are unique in our solar Mariner 10 data, therefore remains an issue in
system. the MESSENGER era. The situation is fur-
Mercury is the closest planet to the Sun and ther complicated by the fact that the classical
therefore subject to a particular strong and separation of external and internal field contri-
dynamic solar wind. Since Mercury’s mag- butionsdevelopedbyGauss[Olsenetal.,2010]
netic field is so weak, the solar wind plasma doesnotdirectlyapplyatMercury. Itassumes
can come extremely close to the planet and that the measurements are taken in a source
may even reach the surface. Mariner 10 data free region with negligible electric currents, an
showed that Mercury’s magnetosphere is not assumption not necessarily fulfilled in such a
only much smaller than its terrestrial counter- small and dynamic magnetosphere.
part but also much more dynamic. Adapted In order to nevertheless extract information
models originally developed for Earth failed on the internal magnetic field, the MESSEN-
to adequately describe the Hermean magneto- GER team analysed the location of the mag-
sphere which therefore remained little under- netic equator where B , the magnetic field
ρ
stoodintheMariner10era[Slavinetal.,2007]. component perpendicular to the planet’s rota-
Knowing a planet’s internal structure is cru- tion axis, passes through zero [Anderson et al.,
cial for understanding the core dynamo pro- 2011,Andersonetal.,2012]. Sincetheinternal
cess. Mercury’s large mean density pointed field changes on a much slower time scale than
towards an extraordinary huge iron core and themagnetosphere, thetime-averagedlocation
a relatively thin silicate mantle covering only should basically not be affected by the magne-
about the outer 25% in radius. Since little tosphericdynamics. Theanalysisnotonlycon-
more data were available in the Mariner era, firmed that the Hermean field is exceptionally
the planet’s interior properties and dynamics weak with an axial dipole of only 190nT but
remained poorly constrained. also suggested that the internal field is best
Solving the enigmas about Mercury’s mag- described by an axial dipole that is offset by
netic field and interior where major incentives 480km to the north of the planet’s equator
for NASA’s MESSENGER mission [Solomon [Anderson et al., 2012]. This configuration,
et al., 2007]. After launch in August 2004 that we will refer to as the MESSENGER off-
and a first Mercury flyby in January 2008, the set dipole model (MODM) in the following, re-
spacecraft went into orbit around the planet in quiresastrongaxialquadrupoleandaverylow
March 2011. At the date of writing, more than dipole tilt, a combination that is unique in our
2800 orbits have been completed. MESSEN- solar system.
GER’s orbit is highly eccentric with a periap- This article tries to summarize the new
sisbetween200to600kmat60to70◦northern understanding of Mercury’s magnetic field in
latitude and an apoapsis of about 15,000km the MESSENGER era at the date of writing.
altitude. This has the advantage that the MESSENGER is still orbiting it’s target and
spacecraft passes through the magnetosphere continues to deliver outstanding data that will
on each orbit but complicates the extraction further improve our knowledge of this unique
of the internal field component because of a planet. Section 3 briefly reviews the current
strong covariance of equatorially symmetric knowledge of Mercury’s magnetosphere. Sec-
and anti-symmetric contributions [Anderson tion 4 describes recent models for the planet’s
et al., 2012, Johnson et al., 2012]. The trade- interior, focussing in particular on the possible
off between the dipole and quadrupole field core dynamics. The magnetic equator analysis
harmonics, that was already a problem with and the offset dipole model MODM are then
2
discussed in section 4. Explaining the weak- C/(MR2 ) and C /C provide the main con-
M m
ness of Mercury’s magnetic field already chal- straintsformodelsofMercury’sinterior[Smith
lenged classical dynamo theory and the pecu- et al., 2012, Hauck et al., 2013]. Note that
liar field geometry further raises the bar. Sec- Rivoldini and Van Hoolst [2013] follow at
tion 5 reanalysis several dynamo model can- somewhat different approach, taking into ac-
didates in the light of the new MESSENGER count the possible coupling between the core
data. Some concluding remarks in section 6 and the silicate shell. The coupling has the
close the paper. effect that C cannot be determined indepen-
m
dently of the interior model and Rivoldini and
Van Hoolst [2013] therefore directly use g
88
2 Mercury’s internal
rather than C as a constraint. The updated
m
structure interior modelling indicates that the core ra-
dius is relatively well constrained at 2020 ±
MESSENGERobservationsofMercury’sgrav- 30km [Hauck et al., 2013] or 2004 ± 39km
ity field [Smith et al., 2012] and Earth-based [Rivoldini and Van Hoolst, 2013]. This leaves
observations of the planet’s spin state [Margot onlytheouter16to19%ofthemeanplanetary
et al., 2012] provide valuable information on radius R = 2440km to the mantle.
M
the interior structure. That fact that Mercury Hauck et al. [2013] find a mean mantle den-
is in a special rotational state (Cassini state 1) sity (including the crust) of 3380±200kg/m3.
allows to deduce the polar moment of inertia MeasurementsofMESSENGER’sX-RaySpec-
C from the degree two gravity moments and trometer (XRS) show that the volcanic surface
the planet’s obliquity, the tilt of the spin axis rockshavealowcontentofironandotherheav-
to the orbital normal [Peale, 1969]. The mo- ier elements [Nittler et al., 2011]. Smith et al.
ment of inertia factor C/(MR2 ), where M is [2012] and Hauck et al. [2013] therefore spec-
M
the planet’s total mass and R its mean ra- ulate that a solid FeS outer core layer may
M
dius, constrains the interior mass distribution. be required to explain the mean mantle den-
The factor is 0.4 for uniform density and de- sity. RivoldiniandVanHoolst[2013],however,
creases when the mass is increasingly concen- argue that the mantle density is not particu-
trated towards the center. The Hermean value larly well contrained. Compositions compat-
of C/(MR2 ) = 0.346±0.014 [Margot et al., ible with XRS measurements are well within
M
2012] indicates a significant degree of differen- the allowed solutions and a denser lower man-
tiation. tle layer is not required by the data.
The observation of the planet’s 88day libra- Naturally, information about the core is
tion amplitude g , a periodic spin variation of particular interest for the planetary dy-
88
in response to the solar gravitational torques namo. There is a rough consensus on the
ontheasymmetricallyshapedplanet,allowsto mean core density with Hauck et al. [2013]
also deduce the moment of inertia of the rigid andRivoldiniandVanHoolst[2013]suggesting
outer part C . If the iron core is at least par- 6980±280km/m3 and 7233±267km/m3, re-
m
tially liquid, C is the moment of the silicate spectively. However, the core composition and
m
shell and thus smaller than C. The Herman theradiusofapotentialinnercorearenotwell
value of C /C = 0.431±0.025 [Margot et al., constrained. Admissable interior models cover
m
2012] confirms that the core remains at least all inner core radii from zero to very large val-
partially liquid. ues with an aspect ratio of about a = r /r =
i o
In addition to M and R the ratios 0.9 [Rivoldini and Van Hoolst, 2013] where r
M i
3
and r are the inner and outer core radii, re- ing curve in the planetary center. Since the
o
spectively. solid iron phase can incorporate only a rela-
An additional constraint on the inner core tively small sulfur fraction, most of the sulfur
size relies on the observations of so-called lo- is expelled at the inner core front and drives
bate scarps on the planet’s surface which are compositional convection. The latent heat re-
likely caused by global contraction. MESSEN- leased upon iron solidification provides addi-
GER data based on 21% of the surface sug- tional thermal driving power. Contrary to the
gested a contraction between 1 and 3km [Di situation for Earth, freezing could also start
Achille et al., 2012]. This sets severe bounds at the core-mantle boundary (CMB) because
on the amount of solid iron in Mercury’s core of the lower pressures in Mercury’s core. The
becauseofthedensitydecreaseassociatedwith iron crystals would then precipitate or snow
the phase transition of the liquide core al- into the center and remelt when encounter-
loy. Severalthermalevolutionmodelstherefore ing temperatures above the melting point at
favour a completely liquide core or only a very a depth r . This process leaves a sulfur en-
m
small inner core [Grott et al., 2011, Tosi et al., richedlighterresiduuminthelayerr > r . As
m
2014]. Recent more comprehensive MESSEN- theplanetcools, r decreasesandastabilizing
m
GER observations, however, allow for a con- sulfur gradient is established that follows the
tractionofupto7km. Thissomewhatreleases liquidus curve and covers the whole snow zone
the contraints [Solomon et al., 2014] though r > r [Hauck et al., 2006]. Since the heat
m
very large inner cores may still be unlikely. flux through the CMB is likely subadiabatic
Sulfur has been found in many iron-nickel today, thermal effects will also suppress rather
meteorites and is therefore a prime candidate than promote convection in the outer part of
for the light constituent in Mercury’s core. Mercury’score. Astablystratifiedlayerunder-
Rivoldini and Van Hoolst [2013] consider iron- neath the planet’s core mantle boundary and
sulfur core alloys and find a likely bulk sulfur probably extending over the whole iron snow
concentrationof 4.5±1.8wt%. Since thiscom- region therefore seems likely. The liquid iron
position lies on the iron rich side of the eutec- entering the layer below r serves as a com-
m
tic, iron crystalizes out of the liquid when the positional buoyancy source. The latent heat
temperature drops below the melting point. being released in the iron snow zone diffuses to
Where this happens first depends on the form the core mantle boundary. Today’s low CMB
ofthemeltingcurveandtheadiabatdescribing heat flux implies that this can be acchieved by
core conditions. a relatively mild temperature gradient.
SinceMercury’smantleissothinithaslikely The possible core scenarios are illustrated
cooled to a point where mantle convection is in fig. (1) with melting curves for different
very sluggish or may have stopped altogether sulfur concentrations and core adiabats with
[Grott et al., 2011, Michel et al., 2013, Tosi CMB temperatures in the range between 1600
et al., 2014]. The heat flux through the core- and2000Ksuggestedbyinterior[Rivoldiniand
mantle boundary is thus likely subadiabatic Van Hoolst, 2013] and thermal evolution mod-
and therefore too low to support a core dy- els [Grott et al., 2011, Michel et al., 2013, Tosi
namodrivenbythermalconvectionalone. The et al., 2014]. Data on the melting behaviour
required additional driving power may then ei- of iron-sulfur alloys are few and the melting
ther be provided by a growing inner core or by curvesshowninfig.(1)thereforerelyonsimple
an iron snow zone. The solid inner core starts parametrizations [Rivoldini et al., 2011]. The
togrowassoonastheadiabatcrossesthemelt- adiabatshavebeencalculatedbyRivoldiniand
4
Van Hoolst [2013]. Mercury’s core pressure is thin red line from the top). At T = 1750K
cmb
only grossly constrained, with CMB pressures (grey)thereremainsonlyarelativelythincon-
in the range 4 − 7GPa and central pressures vective layer between the inner core bound-
in the range 30−45GPa [Hauck et al., 2013]. ary at r = 1440km and the lower bound-
i
We adopt a central pressure of 40GPa here. ary of the outer snow layer at r = 1650km.
m
Fig. (1) suggests that iron starts to solidify in For the coldest adiabat shown in fig. (1) with
the center for an inital sulfur concentrations T = 1890K (blue) only the outer 300km of
cmb
below about 4wt%. Sulfur released from the the core remain liquid but belong to the iron
inner core boundary increases the concentra- snow zone so that no dynamo seems possible.
tion in the liquid core over time and thereby Additional sometimes complex scenarios
slows down the inner core growth and delays have been discussed in the context of
the onset of iron snow. For an initial sulfur Ganymede by Hauck et al.[2006] and may also
concentration beyond 4wt% iron solidification apply at Mercury since the iron cores of both
starts with the CMB snow regime. A convec- bodies cover similar pressure ranges. For ex-
tive layer that is enclosed by a solid inner core ample, fig. (1) illustrates a kink in the melting
and a stably stratified outer iron snow layer curve for pressures around 21GPa and com-
seems possible for sulfur concentrations be- positions larger than 5wt% sulfur. This could
tweenabout2.5 and7wt%. Forsulfurconcen- lead to a double snow regime where not only
tration beyond 7wt% an inner core would only theveryouterpartofthecoreprecipitatesiron
growwhenthesnowzonesextendsthroughthe but also an intermediate layer around 21GPa.
whole core and the snow starts to accumulate Thispossibilityhasbeenexploredinadynamo
in the center. model by Vilim et al. [2010] that we will dis-
The adiabats and thin red lines in fig. (1) il- cuss in section 2. Since the kink is not very
lustrate the evolution for an initial sulfur con- pronounced, however, such a double snow dy-
centration of 3wt%. For the hot (red) adiabat namo would not be very long lived.
with T = 2000K neither inner core growth Another interesting scenario unfolds when
cmb
not iron snow would have started and there the light element concentration lies on the S-
would be no dynamo. When the temperature rich side of the eutectic. Under these condi-
dops, iron starts to solidify first at the cen- tions, FeS rather than Fe would crystalize out
ter. For a CMB temperature of T = 1910K when the temperature drops below the FeS
cmb
(solid green adiabat), the inner core has al- melting curve. Since FeS is lighter then the
ready grown to a radius of about 600km while residuum fluid, the crystals would rise towards
the outer snow layer is only about 160km the core-mantle boundary. However, eutec-
thick. The sulfur released upon inner core tic or even higher sulfur concentrations can-
growth has increased the bulk concentration not represent bulk conditions since it would be
in the liquide part of the core to 3.4wt% (first difficult to match Mercury’s total mass [Rivol-
thin red line from the top). The decrease in dini et al., 2011]. Inner core growth would in-
the sulfur abundance due to the remelting of crease the sulfur concentration in the remain-
iron snow has not been taken into account in ingfluidovertimebutneverbeyondtheeutec-
this model. When the CMB temperature has tic point. This has likely not been reached in
dropped to T = 1890K (dashed green adia- Mercury because the eutectic temperature of
cmb
bat) the inner core and snow layer have grown 1200−1300K [Rivoldini et al., 2011] is signif-
by a comparable amount while the sulfur con- icantly lower than today’s CMB temperature
centration has increased to 4.4wt% (second suggested by thermal evolution [Grott et al.,
5
2011, Tosi et al., 2014] and interior models magnetizationismuchstrongerinthesouthern
[Rivoldini and Van Hoolst, 2013]. thaninthenorthernhemispherecouldreflecta
An alternative explanation for a locally high specialconfigurationoftheplanet’sancientdy-
sulfurconcentrationwassuggestedbytheXRS namo. Impacts or large degree mantle convec-
observations. The low Fe but large S abun- tion may have significantly decreased the heat
danceinsurfacerocksindicatesthatMercury’s flux through the northern CMB and therefore
core could have formed at strongly reducing weakened dynamo action in this hemisphere
conditions. This promotes a stronger parti- [Stanleyetal.,2008,Amitetal.,2011,Dietrich
tioning of Si into the liquid iron phase lead- and Wicht, 2013]. Mercury’s magnetic field is
ing to a ternary Fe-Si-S core alloy [Malavergne distinctively stronger in the northern than in
et al., 2010]. Experiments have shown that the southern hemisphere and it seems attrac-
Si and S are immiscible for pressures below tivetoinvokeanincreasednorthernCMBheat
15GPa [Morard and Katsura, 2010] which is flux as a possible explanation.
the pressure range in the outer part of Mer- Clues about the possible pattern may once
cury’s core. However, the immiscibility only more come from MESSENGER observations.
happens for sizable Si and S concentrations. A combination of gravity and altimeter data
Experiments by Morard and Katsura [2010], allowed to estimate the crustal thickness in
for example, demonstrate that at 4GPa and the northern hemisphere. On average, the
1900Kabundancesof6wt%Sand6wt%Siare crust is about 50km thicker around the equa-
required to trigger the immiscibility and lead tor than around the pole [Smith et al., 2012]
to the formation of a sulfur rich phase with a which points towards more lava production
composition of about 25wt% S. For FeS crys- and thus a hotter mantle at lower latitudes.
tallization to play a role at today’s CMB tem- This is consistent with the fact that the north-
peratures, the sulfur rich phase should lie sig- ernlowlandsarefilledbyyoungerfloodbasalts
nificantly to the right of the eutectic where the since melts more easily penetrate a thinner
FeS melting temperature increases with light crust [Denevi et al., 2013]. Missing altime-
element abundance. Thus even higher S and ter data and the degraded precision of gravity
Si contributions are required but seem once measurements does not allow to constrain the
more difficult to reconcile with the planet’s crustal thickness in the southern hemisphere.
total mass [Rivoldini and Van Hoolst, 2013]. The lack of younger flood basalts, however,
Since Si partitions much more easily into the could indicate a thicker crust and hotter man-
solidironphasethansulfur,it’scontributionto tle. Since a hotter mantle would reduce the
compositional convection and the stabilization CMB heat flux, these ideas indeed translate
of the snow zone is significantly weaker. into a pattern with increased flux at higher
Several numerical studies in the context of northern latitudes. However, Mercury’s vol-
Earth and Mars have shown that the CMB canism ceased more than 3.5Gyr ago and to-
heat flux pattern can have a strong effect day’s thermal mantle structure may look com-
on the dynamo mechanism (see e.g. Wicht pletelydifferent. Evensimplethermaldiffusion
et al. [2011a] and Dietrich and Wicht [2013] shouldhaveerodedanyasymmetryoversucha
for overviews). Like the mean heat flux out of longtimespan. Thermalevolutionsimulations
thecore, thispatterniscontrolledbythelower show that at least the lower part of the mantle
mantlestructure. TheMartiandynamoceased maystillconvecttoday[Smithetal.,2012,Tosi
about4Gyragobuthasleftitstraceinformof et al., 2014] which would change the structure
a strongly magnetized crust. The fact that the on much shorter time scales. Since the active
6
γ Fe
eutectic
T =2000 K 1
cmb
Tcmb=1910 K 2
T =1890 K
cmb
2500 T =1750 K 3
cmb
T =1600 K 4
cmb
%
5t
w
6r
u
7f
l
] u
K 8s
2000
[
9
T
10
11
1500
10 20 30 40
P [GPa]
Figure1: MeltingcurvesfordifferentinitialsulfurconcentrationsandpossibleMercuryadiabats
fordifferenttemperaturesshownasthickred,green,turqoise,andbluelines. Thinredlinesfrom
top to bottom show the melting curves for the convecting part of the core for an initial sulfur
concentration of 3wt% and a core state described by the solid green, dashed green, gray, and
blue adiabats. The thick solid black line shows the melting curve for pure iron while the thick
dashedblacklineshowstheeutectictemperature. Thefigure,providedbyAttilioRivoldini,and
has been adapted from Rivoldini et al. [2011] to include the Mercury core adiabats calculated
in Rivoldini and Van Hoolst [2013]. A central pressure of 40GPa is assumed for Mercury but
the adiabats are only drawn in the liquid part of the core.
7
shell is so thin, the pattern would be rather rather close to the planet at an average po-
small scale without any distinct north/south sition of only 1.96 planetary radii [Winslow
asymmetry. et al., 2013] compared to 14 planetary radii for
Because of Mercury’s 3:2 spin-orbit reso- Earth.
nance, the high eccentricity of the orbit, and Behind the bow shock, the cold solar wind
theverysmallobliquitythetimeaveragedinso- plasma is heated up and interacts with the
lationpatternshowsstronglatitudinalandlon- planetary magnetic field, thereby creating the
gitudinal variations. Williams et al. [2011] cal- magnetosphere. To first order, the planetary
culates that the mean polar temperature can fieldlines form closed loops within the dayside
be 200K lower than the equatorial. Longitudi- magnetosphereandalongtailonthenightside.
nalvariationsshowtwomaximathatareabout Theouterboundaryofthemagnetosphere, the
100K hotter than the minima at the equator. magnetopause, is located where the pressure
IfMercury’smantleconvectionhasceasedlong of the shocked solar wind and the pressure
ago, the respective pattern may have diffused of the planetary magnetic field balance. The
into the mantle and could determine the CMB solar wind ram pressure, on average 14.3nPa
heatfluxvariation. Higherthanaveragefluxat at Mercury [Winslow et al., 2013], is an or-
the poles and a somewhat weaker longitudinal der of magnitude higher than at Earth while
variation would be the consequence. We dis- the magnetic field is two orders of magnitude
cuss the impact of the CMB heat flux pattern weaker. Likethebowshock,themagnetopause
on the dynamo process in section 2. is therefore located much closer to the planet
at Mercury than at Earth with mean standoff
distances of about 1.45 [Winslow et al., 2013]
3 Mercury’s external
and 10 planetary radii, respectively. Both the
magnetic field Hermean magnetosphere and magnetosheath,
the region between bow shock and magne-
Planetarymagnetospheresaretheresultofthe topause,arethusmuchsmallerthantheterres-
interaction between the planetary magnetic trialequivalentsinrelativeandabsoluteterms.
fieldandtheimpingingsolarwindplasma. Be- Fig. (2) shows the current density in a nu-
cause of Mercury’s weak and asymmetric mag- merical hybrid simulation that models the so-
netic field and the position close to the Sun, lar wind interaction with the planet [Mu¨ller
the Hermean and terrestrial magnetospheres et al., 2012]. The location of the bow shock
differer fundamentally. Mercury experiences and the magnetosphere can be identified via
the most intense solar wind of all solar-system therelatedcurrentsystems. Alongaspacecraft
planets. Under average conditions, the ratio of trajectory these boundaries can be identified
the solar wind speed and the Alfv´en velocity, by the related magnetic field changes. Fig. (3)
called the Alfv´enic Mach-number, is compara- shows MESSENGER magnetic field measure-
bletotheterrestrialone. Withvaluesof6.6for mentsforarelativelyquietorbit(orbitnumber
Mercury[Winslowetal.,2013]and8forEarth, 14) where both the bow shock and the magne-
thesolarwindplasmaissuper-magnetosonicat topause can be clearly classified on both sides
bothplanets,i.e.themediumpropagatesfaster of the planet.
than magnetic disturbances and a bow shock Another important element of the magne-
therefore forms in front of the magnetosphere. tosphere is the neutral current sheet which is
Because of the weak Hermean magnetic field, responsible for the elongated nightside magne-
the sub-solar point of the bow shock is located totail and separates the northern and south-
8
Figure 2: Electrical currents in a numerical simulation of the Hermean magnetosphere. The
amplitude of the current density j is color-coded. An equatorial cross section is shown in an
coordinate system where X points towards the Sun (negative solar wind direction) and the
Y-axis lies in the Hermean ecliptic. The bow shock standing in front of the planet slows down
the solar wind. The magnetopause is the outer boundary of the magnetosphere. The neutral
current sheet is located in the nightside of the planet. An arc of electrical current visible close
to the flyby trajectory (January 14, 2008) could be interpreted as a partial ring current. This
figure is a snapshot from a solar wind hybrid simulation and is adapted from Mu¨ller et al.
[2012].
9
Figure 3: Magnetic field data recorded by the MESSENGER magnetometer (10s average)
during orbit 14 on the DOY 84 in 2011. The upper panel, shows the time series of the absolute
magnetic field |B| (black), the negative absolute field (grey), the radial component B (green),
r
and the component B perpendicular to the rotation axis. Time is measured in hours since
ρ
the last apocenter passage. The plasma boundaries are marked with vertical dashed lines (BS:
bow shock, MP: magnetopause). The location where B vanishes defines the magnetic equator
ρ
(MEQ). The lower panel shows the planetocentric distance r and the co-latitude θ. The data
are taken from the Planetary Data System / Planetary Plasma Interactions Node.
10