 COMPLEX DISTRIBUTION OF PERCHLORATE AT THE MARS PHOENIX LANDING SITE.  S.C. Cull 1 E. Kennedy, A. Clark 2, 1 Bryn Mawr College, 101 N. Merion Ave, Bryn Mawr, PA 19010, scull@brynmawr.edu, 2 Haverford College, Haverford, PA 19010   Introduction In 2008, the Phoenix lander returned evidence of perchlorate in the soils at its landing site on the northern plains of Mars (68.2188N, 234.2508E, IAU 2000 areocentric, Kounaves et al. 2009; Hecht et al. 2009). Perchlorate likely forms atmospherically, a process that should result in a layer of perchlorate on the surface of the soil, with gradually decreasing concentrations further down the soil column.  However, because perchlorate forms highly hygroscopic salts, any contact with water will serve to redistribute it through the soil column.  We can thus use the perchlorate distribution at the landing site as a marker for the aqueous and nonaqueous processes that have affected the soil since perchlorate deposition.  Here, we use spectral data and images from the Surface Stereo Imager (SSI) to map the distribution of perchlorate through the soil column at the Phoenix landing site.    Data Set & Methods Multispectral data from the Phoenix Surface Stereo Imager (SSI, Smith et al. 2009) was used to map the distribution of perchlorate across exposed surfaces at the Phoenix landing site.  SSI was a stereographic imager with a 12-position filter wheel that produced 13 wavelengths between 0.4 to 1.001 micrometers. Perchlorate salts have an absorption at 0.967 micrometers, which corresponds to an SSI filter position (Cull et al. 2010).  This absorption was used to map the distribution of perchlorate across the landing site, including all 12 trenches dug by Phoenix during operations, and undisturbed surface areas.    Results We find that perchlorate occurs almost exclusively in small (<20 mm) patches that appear to be coatings on the surfaces of rocks and soil clods.   These rocks and clods occur both on undisturbed surface soil and in subsurface soil exposed by Phoenix trench digging.  Within trenches, the patches occur on small clods (Figure 1A), on flattened trench surfaces (Figure 1B), and on trench walls (Figure 1C).  On undisturbed surfaces, the patches appear almost exclusively as coatings on rocks and clods.  Few occur on unconsolidated soil.  No perchlorate was observed to be in contact with the subsurface ices exposed by Phoenix trenching, and there appeared to be no correlation between patch position and geomorphology (e.g., polygon trough versus center).     Figure 1 - SSI images of trenched surfaces at the Phoenix landing site.  Areas with a 0.967 micrometer absorption are marked in black.  The black arrows point to the areas used to produce the solid-line spectra at right.  Dotted-line spectra are taken from adjacent to the black pixels.  A) A subsurface perchlorate patch on a soil clod or rock in the subsurface, exposed during Dodo-Goldilocks trenching.  The perchloratebearing portion appears as a coating on the surface of the light-toned clod.  B) A subsurface perchlorate patch on an area of light-toned, flattened soil exposed during Dodo-Goldilocks trenching.  C) A portion of the Dodo-Goldilocks trench wall, showing a perchlorate patch.   The patch does not appear to follow any soil horizons in the subsurface, but instead appears as a rounded clod, similar to those exposed on the surface and in the rest of the trench.   Discussion If perchlorate forms atmospherically (e.g., Catling et al. 2010), this should lead to a homogeneous surface distribution at the Phoenix landing site.  However, because it is also highly soluble in water, its distribution in the soil column is indicative of postformation processes at work in Phoenix soils.   Our SSI mapping shows that perchlorate is neither homogenously distributed nor restricted to the surface.  Instead, it appears to be concentrated as coatings of rock and clod surfaces, both on the surface and in the subsurface.  This distribution points to an aqueous-redistribution of perchlorate; non-aqueous processes are unlikely to produce the observed coatings.  A variety of processes could be responsible for producing the observed coatings, including thin film migration, interaction with seasonal ice meltwater, or episodic wetting events due to orbital variations.  Perchlorate brines, once formed, do appear to be stable under Phoenix surface conditions: Chevrier et al. (2009) found eutectic temperatures of 206 K for Mg(ClO4)2 and 236 K for NaClO4, both within conditions experienced by the Phoenix lander during Martian summer.   Once dissolved and redistributed, the perchlorate-coated rocks could be redistributed through the soil column by a mechanical process, such as cryoturbation or aeolian mixing.  Cryoturbation is known to be a signficiant process at the Phoenix landing site, where annual freeze-thaw cycles in the subsurface ice layer cause cracking that leads to polygonal terrain (Mellon et al. 2010).  Arvidson et al. (2010) also noted that the freeze-thaw cycles appeared to concentrate rocks in the troughs of polygons, illustrating the importance of the process in soil processing.   The distribution of perchlorate at the Phoenix landing site is unlike perchlorate distributions found on Earth.  In the Antarctic Dry Valleys (ADV), Kounaves et al. (2010) report a perchlorate distribution through the soil column similar to what would be expected for atmospheric deposition: a highly concentrated surface layer, followed by a steady decrease downward toward the ice table.  They do report additional concentrated horizons within the soil column that appear to be caused by epidosic wetting events; however, these are distinctly different from the perchlorate coatings we map at Phoenix.   Likewise, the Atacama nitrate/perchlorate deposit appears to be a poor analog for the distribution of perchlorate through the Phoenix soil column.  The Atacama deposits have undergone extensive reworking by liquid water, including from episodic rainfall events (Eriksen 1981), groundwater discharge (Pueyo et al. 1998), and possibly even hydrothermal processes (Erisken 1981).   Our mapping points to a complex set of both aqueous and non-aqueous processes operating at the Phoenix landing site in the geologically recent past.    References: [1] Arvidson, R et al. (2010) JGR 115;   [2] Chevrier, V. F eta l. (2009), GRL 36, L10202, doi:10.1029/2009GL037497.  [3] Cull, S.C. et al. (2010) GRL 37: doi:10.1029/2010GL045269;  [4] Ericksen, G. E. (1981), U.S. Geol. Surv. Prof. Paper, 1188. [5] Hecht, M. H., et al. (2009), Science, 325, 64-67, doi:10.1126/science.1172466; [6] Kounaves, S. P., et al. (2009), J. Geophys. Res., 114, E00A19, doi:10.1029/2008JE003084; [7] Kounaves, S. P., et al. (2010), Environ. Sci. Technol., 44, 2360-2364, doi:10.1021/es9033606; [8] Mellon, M. T., et al. (2010a), J. Geophys. Res. 114: DOI: 10.1029/2009JE003417. [9] Pueyo, J.J. et al. (1998) Andean Geology 25 (1). [10] Smith, P. H., et al. (2009), Science, 325, 58-61, doi:10.1126/science.1172339. [11] Catling, D. C., et al. (2010), J. Geophys. Res., 115, E00E11, doi:10.1029/2009JE003425.   
