Compositional Interpretation of PFS/MEx and
TES/MGS Thermal Infrared Spectra of Phobos
M. Giuranna1, T. L. Roush2, T. Duxbury3, R. C. Hogan4,
C. Carli5, A. Geminale1, V. Formisano2
1.
2.
3.
4.
5.
Istituto di Fisica dello Spazio Interplanetario INAF-IFSI, Via del Fosso del Cavaliere 100, 00133 Roma, Italy
NASA Ames Research Center, MS 245-3, Moffett Field, CA 94035-1000 USA
Jet Propulsion Laboratory, California Institute of Technology, 4800 Oak Grove Drive, Pasadena, 91109 California
Bay Area Environmental Research Institute, C/O MS 245-3, Moffett Field, CA 94035-1000 USA
Isituto di Astrofisica Spaziale e Fisica cosmica INAF-IASF, Via del Fosso del Cavaliere 100, 00133 Roma, Italy
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
FORMATION SCENARIOS AND THEIR
COMPOSITIONAL IMPLICATIONS
The origin of the Martian satellites presents a puzzle of long standing.
Addressing the composition of Phobos will help constrain theories of its formation.
Table 1. Different origin scenarios for Phobos, compositional implications, and importance for Mars science
Capture
Volatile and organicsrich?
Constrains input to any
early biosphere
In-situ
Formation
Silicate material
Ordinary chondrite-like?
Mars “building blocks”
Mars Ejecta
Space-weathered basalt?
Early Martian crust,
preserved from erosion?
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
TES/MGS and PFS/MEx
TES is a Michelson interferometer operating from 1700 to 200 cm-1 (~6 to 50 mm)
with spectral resolutions of 6.25-12.5 cm-1 (Christensen et al., 1992, 1998).
PFS is a Fourier spectrometer working in two different channels. The SWC covers
wavenumbers from 1700 to 8200 cm−1 (1.2–5.9 μm) while the LWC operates in the
range 250–1700 cm−1 (5.9–40 μm). The spectral resolution is 1.3 cm−1 when no
apodization function is applied, and about 2 cm−1 when Hamming function is applied
to the interferograms. In the present analysis we use data from LWC.
At these wavelengths, both instruments can detect the fundamental vibrational
modes of materials. The number, position, intensity, and shape of which are
dependent upon the atomic masses, inter-atomic force fields, and molecular geometry,
providing a sensitive means of determining mineralogy.
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
Available Data: TES/MGS
Table 2. TES Phobos observations.
Orbit #
MGS-Phobos
Distance, km
Phobos-Sun
Distance, AU
Sun-Phobos-MGS
Phase Angle (º)
Approximate
pixel size, km
Number of Spectra*
476
1437
1.58587
105-107
11.9
6
501
1081
1.59916
110-113
9.0
9
526
1152-1269
1.61140
93-149
9.6-10.5
106
551
275-785
1.62260
51-131
2.3-6.5
149
*>50% of pixel filled with Phobos
The MGS mission was designed chiefly to investigate the surface and atmosphere of
Mars (Albee et al., 1998). During the science phasing orbital period there were several
opportunities to observe Phobos (see Table 2).
The first three were essentially hemispherical observations of Phobos’ unilluminated
hemisphere.
However, the final encounter (orbit 551) provided modest spatial resolution of the
surface of Phobos during which TES acquired several spectra of Phobos.
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
Available Data: PFS/MEx
Table 3. PFS Phobos observations.
Orbit #
5851
756
5870
6906
Number of Spectra
1
2
1
2
MEx-Phobos Distance, km
97
155
354
530
Footprint diameter, km
4.7
7.5
17.3
25.9
1.64121
1.66508
1.63740
1.38759
96.9
64.4
52.5
34.6
23.5, 219.7
24, 266.8
23.2, 222.7
-25.5, 229.4
Phobos-Sun Distance, AU
Sun-Phobos-MEx Phase Angle (±1.5º)
Sub-Solar Point (°)
(Lat, ELon)
In observing Phobos, MEx benefits from its highly elliptical orbit allowing
observations of both the planet-facing and the far side of the moon.
PFS had several opportunities to observe Phobos. In particular, between July and
September 2008 the MEx spacecraft made a series of eight flybys of Phobos at
distances ranging from 4500 km down to 93 km from the centre of the moon.
For our analysis we selected orbits with the closest possible approach, in order to
have Phobos always larger than, or at least comparable to, the FOV of PFS.
Four orbits satisfy these criteria, as summarized in Table 3.
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
Available Data: PFS/MEx
In each orbit PFS observes a
different region of the Phobos’
surface.
A simulation of Phobos during
the four PFS observations is
shown in Figure.
The PFS footprint has been
super-imposed on each of the
images.
The simulated images of Phobos
were created by illuminating a
shape model consisting of a
harmonic expansion (Willner et
al.,
2009)
with
craters
superimposed (Duxbury, 1989).
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
Data Reduction: TES
• A linear combination of 3 black bodies is used to perform a least squares fit
• Using these results as inputs, an upper hull is fit to the radiance maxima
• Emissivity values are produced by dividing the measured radiance by the hull.
The derived temperatures can be compared with those of PFS
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
Data Reduction: PFS
1) Hamming
apodization function
is applied to the
interferograms.
2) spectra within the
same
orbit
are
averaged together (4
final spectra).
3) The resulting spectra
are smoothed over
thirteen points.
The actual noise is
shown as error bars, one
every 5 spectral points.
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
Comparison of TES & PFS Surface Temperatures
Table 4 Derived Surface Temperatures
Heliocentric Longitude (Ls, deg)
Avg. T1±sd, K
Avg. T2±sd, K
Avg. T3±sd, K
PFS-5851
102.5
240
160
130
PFS-756
77.3
265
260
250
PFS-5870
104.9
270
265
260
PFS-6906
269.7
(~winter solstice, 1.388 AU)
353
290
260
TES-476
12
218±15
149±30
114±40
TES-501
17
194±15
139±20
102±25
TES-526
23
190±20
146±27
102±25
TES-551
29
271±51
206±44
143±51
Instrument-Orbit #
PFS min and max derived surface temperatures: 130-160 K and 290-353 K.
Good agreement with the minimum (night, 140 K) and maximum (day, 300 K) brightness
temperatures of the satellite surface derived from Viking measurem.s (Lunine et al., 1982)
Compared to the temperatures derived (Earth-based observations) for Phobos by Lynch et
al. (2007, 320-340 K), only those for PFS orbit 6906 approach these values. The
explanation for this systematic difference is that during the observations of Lynch et al.
(2007) Mars’ heliocentric distance was ~1.38-1.39 AU.
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
Comparison of Temperatures with models
We compare our results with
the results of the numerical
modeling of the thermal regime
of Phobos’ surface regolith layer
(on seasonal time scales) in
Kuzmin and Zabalueva (2003).
There is a general good
agreement, as our derived
temperatures are mostly within
the range of predicted diurnal
temperatures on the Phobos
surface, and the variation of
surface temperature as predicted
by the model is also reproduced
by our results.
However, our max derived
temperature is ~20 K higher than
the max diurnal temperature
predicted.
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
TES Spectra
Orbit 551
TES spectra with similar properties are combined
into clusters. Statistical information is associated
with each cluster (e.g., mean, variance, etc.).
Most spectra fell into one of four groups that are
spatially associated with areas located
1) on the rim; 2) on the floor; 3) in the grooves to
the east; and 4) to the northeast, of Stickney crater.
The emissivities of the four regions were averaged
and used as representative spectral measurements
in the compositional efforts.
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
PFS Spectra
TF
By chance, the same region, namely the area located to the northeast of Stickney crater, has been observed by
both TES and PFS. The two spectra look extremely similar, showing essentially the same bands in the same
positions, with similar widths and relative depths.
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
Analyses approaches
Our compositional interpretation of thermal emission data relies upon several digital
spectral libraries.
One contains measured emissivity for a variety of materials (carbonates, nitrates, borates,
halides, oxides, phosphates, arsenates, vanadates, sulphates, chromates, molybdates and
silicates), (TES library).
Additionally, a digital library containing spectra of terrestrial rocks, lunar materials, and
various meteorites reflectivities was used (ASTER spectral library).
The RELAB spectral library is also used for comparison (Pieters and Hiroi, 2004). It
contains spectra of lunar samples.
Additionally, we compared Phobos spectra with a limited number of materials that are
residues of the processing of various organic precursors (Cruikshank et al., 1991). Such
samples are analogs for the processed products of ultraprimative materials found in the
outer solar system.
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
Surface Compositional Interpretations
Results are dominated by Phyllosilicates for all classes
Mostly Micas
(Biotite)
Antigorite
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
Surface Compositional Interpretations
Phyllosilicate, Rocks, Lunar samples
Ijolite: the sample from which the powder was derived is a dark gray to black, medium-to coarsegrained rock composed of Nepheline, Biotite, a black mafic mineral and opaques.
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
Surface Compositional Interpretations
Phyllosilicates (Tectosilicate?)
Phyllo Mixture:
Nepheline:
50% Biotite
+
50% Antigorite
Unusually dark
gray
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
Surface Compositional Interpretations
TES observations allow evaluation
of
the
position
of
the
Christiansen frequency (CF) that
has been related to rock type
(Salisbury and Walters, 1989).
Using the entries in Table 2 of
Salisbury and Walter (1989) that
lists the CF of rocks observed in a
vacuum, we created an average for
several igneous rock types. These
are shown as the points with 1
standard deviations in Figure.
We used the 4 classes of spectra identified from the region in and around Stickney for
comparison to these points. For each class we fit the data in the 7.1-9.8 mm wavelength region
using a quadratic curve. The results are shown as the large letters in Figure. It is clear that all
the classes are consistent with the range of CF values associated with ultramafic rocks.
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
Surface Compositional Interpretations
CF vs SiO2 content (ASTER library) compared to CF calculated for TES spectra observed near Stickney
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
Surface Compositional Interpretations
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
Surface Compositional Interpretations
Feldspars?
TES Library
ASTER Library
This is the only spectrum that can not be matched by phyllosilicates.
Only Tectosilicates (Feldspars) provide reasonable-to-good matches.
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
TES average spectrum for selected
observations during orbit 526
TES footprints cover the PFS footprint for Orbit 5851. The two measured spectra look very similar.
This supports the indication of feldspars on the surface of Phobos.
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
METEORITES: Poor Matches
Meteorites “best” matches
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
Two different portions of a black chondrite
(Paragould, LL5)
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
Conclusions (1/3)
•
The lack of deep absorption features in the VNIR spectra of Phobos have allowed a
number of possible analogs to be ruled out, though they have also made it difficult to
make any positive identifications.
•
Our observations in the thermal infrared show several spectral features that have
been used to investigate the composition of the surface.
•
Our results show that the majority of the spectra are consistent with a dominance of
phyllosilicate, particularly in the areas around Stickney crater (“blue” units).
•
Some spectra also suggest the presence of Feldspars/Feldspathoids (“red” units).
•
The spectral compositional results suggest ultramafic assemblages. This is consistent
with the position of the CF in the TES data that suggests ultramafic rock types.
•
Meteorites spectra provide poor matches. Achondrites?
•
Two independent approaches of compositional analyses yield very similar results.
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
Conclusions (2/3)
•
Our results show that the surface temperature of Phobos varies with solar incidence
angle and heliocentric distance, reconciling different results (Viking, Lunine et al. 1982;
Earth-based, Lynch et al. 2007; models, Kuzmin and Zabalueva 2003).
•
The lack of consistency of the PFS and TES spectra to analogs of ultraprimitive
materials (organic residues) suggests that an origin via capture of a TNO object is not
supported by these observations.
•
Silicates: silicate material on Phobos yield high porosity content (see P. Rosenblatt),
which seem difficult to reconciliate with the capture scenario, but is required in order
to absorb the energy of the large impact that generated Stickney without destroying
the body. High porosity is an expected consequence in some scenarios of in-situ
formation, e.g., re-accretion process.
•
Phobos may represent a case analogous to that of Mercury, where the use of mid-IR
emissivity spectra can improve surface compositional interpretations.
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
Conclusions (3/3)
Table 1. Summary of hints for Phobos origin.
Data/model
Capture scenario
In-situ scenario
Spectral reflectance (VNIR)
(previous works)
Very likely
carbonaceous chrondrites
D-type asteroids
Likely
Silicate material
Spectral emissivity (TIR)
(this work)
Possible
achondrites?
HEDs and Shergotty ?
Likely
Silicate material, Mars-like
Celestial mechanics
Unlikely
Likely
Bulk density (1.867 to 1.885 ± 0.06 g/cm3)
(Willner et al., 2009)
Unlikely
D-class material
(1.64 ± 0.02 g/cm3)
Likely
Silicate material
Bulk porosity
(Rosenblatt et al., 2009)
Unlikely
Anhydrous chondrites
Very likely
In Table 1 we collect and summarize the compositional clues for the origin of Phobos
Currently, the most likely scenario is the in-situ formation of Phobos (re-accretion?),
although a capture of achrondrite-like meteorites is not completely ruled out.
A more definitive answer to the origin of Phobos will benefit from in-situ measurements, or
sample return. The future Russian Phobos-Grunt mission (Phobos Sample Return) will
certainly contribute to our understanding regarding the origin of Phobos.
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
Several features in mid-IR spectra…
…and VIS/NIR spectra almost featureless?
There are several physical and chemical mechanisms that can act to eliminate or reduce the
strength of the 1- and 2-mm bands.
1. Both the 1- and 2-mm bands arise from electronic transitions of ferrous-bearing minerals
such as pyroxene (1- & 2-mm) and/or olivines (1 mm). If the surface materials on Phobos
are Fe-free, or Fe-poor, then these spectral features will be absent, or reduced in their
intensity.
2. Mixtures of minerals will have an influence on the observed band depths. Both
laboratory and theoretical studies of the reflectance behaviour of opaque materials
(metals, amorphous carbon, et cetera) mixed with olivines and pyroxenes show that
relatively low abundances of the opaques reduce the 1 and 2 micrometer bands
drastically.
3. Mineraloids such as glasses produced by quenching of materials from the liquid state, or
by impact processes exhibit much less pronounced spectral features at VNIR wavelengths
because of the lack of well-defined crystalline structure associated with the electronic
transitions seen in pyroxenes and olivines.
4. Many laboratory reflectance spectra of pure fine-grained minerals document the decrease in
band depth as the particle size decreases.
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia
Several features in mid-IR spectra…
At mid-IR wavelengths, both
…and VIS/NIR spectraAllende
almostandfeatureless?
Elenovka become
5.
Space weathering?
brighter and almost all of the
spectral maxima become more
pronounced after laser irradiation.
The effects of space weathering include glass and agglutinate formation, comminution, and
deposition of fine metallic iron, all due to vapor condensation created via sputtering
and/or micrometeorite impacts.
At visible and near-IR wavelengths, documented effects of space weathering on silicates
include darkening, reddening, and reduction of spectral bands.
The effects of space weathering in mid-IR have been
mainly forcomparison
the Moon, Mercury,
Thisstudied
limited
of
and few asteroids. Salisbury et al. (1997) demonstrated
Lunar
space weathering
has
laboratorythat
data
suggests
that mid-IR
little or no effect on the Christiansen feature of lunar samples.
spectra appear to be less influencedWe found two samples in the RELAB spectral library that had been exposed to nano-pulsed
degraded
by Elenovka
space weathering than
laser bursts, simulating micrometeorite impacts:
Allende and
VIS/NIR spectra, and perhaps
spectral features are enhanced.
Space weathering has the potential to reduce the VIS/NIR bands, with possibly little or no
effect in the mid-IR.
The First Moscow Solar System Symposium , 11 – 15 October 2010, Moscow, Russia