 DETECTABILITY BY RADAR OF SALTS IN MARTIAN ICE DEPOSITS.  D.P. Winebrenner1,2, D.E. Stillman3, and R.E. Grimm3. 1Applied Physics Laboratory, University of Washington, Seattle, WA 98185 USA, dpw@apl.washington.edu; 2Dept. of Earth and Space Sciences, University of Washington, Seattle, WA 98195 USA, 3Southwest Research Institute, Boulder, CO USA.  Introduction: Dust is common in Martian ice deposits, as a result of atmospheric deposition and concentration by ice-ablation [1,2]. Recent discoveries of salts in Martian surface soil and dust [3,4] therefore suggest that dust impurities could result in salt impurities, at micro- to millimolar concentrations, in relatively pure Martian ice masses (i.e., those with dust contents up to a few percent by mass [5-7]).   Salt-dopants in such ice masses could have significant implications for ice-rheology near and above eutectic temperatures [8], and thus for the timescales of possible ice flow in a wide variety of surface features claimed to show evidence of such flow [9-14].   In particular, the perchlorate salts discovered at the Phoenix landing site [3] may be widespread, in view of their likely origins and evolution [15,16].  Eutectic temperatures reported for perchlorate salts range from 206-239K [15].  The lower part of this range coincides with that of modeled basal temperatures for midlatitude Martian ice masses [17].  Moreover, Lenferink et al. [18] report observations indicating that Mg(ClO4)2 in ice could weaken it significantly (in comparison to pure ice), even at 190K. These considerations motivate investigation of ways to detect salts generally, and perchlorate salts in particular, in Martian ice masses, independent of inferences based on ice rheology.   Salt-doping generally increases the attenuation of ice of radar waves at common sounding frequencies (1-60 MHz), relative to attenuation in pure ice, especially at temperatures near and above eutectics [19, 20].  Thus comparison of changes in observed attenuation with modeled ice temperature fields may reveal salt-doped ice.  Spatial variations in radar attenuation have recently been estimated using the Mars Reconnaissance Orbiter Shallow Radar (SHARAD) in a lobe of the North Polar Layered Deposits (NPLD) known as Gemina Lingula [7]. Most authors estimate basal ice temperatures in polar regions to be below even perchlorate eutectics [21], and the observed attenuation in Gemina Lingula is indeed small [7].  The higher modeled ice temperatures in mid-latitude ice masses, however, motivate investigation of prospective attenuation vs. temperature such as might be observed at those locations, whether using existing or future sounding radars. Here we report initial results from such an investigation based on a combination of laboratory data and dielectric modeling. Components of Radar Attenuation: Radar attenuation at 1-60 MHz in impure ice results from at least 4 distinct processes: (1)  Ohmic (direct current, or zero-frequency) conduction in the ice - Pure ice is a semiconductor in which a small direct-current (DC) conductivity results from effective motion of protons from one crystal defect to another.  In ice with sufficiently high salt concentrations (roughly 3 mM for chloride salts), and at super-eutectic temperatures, brine films at cystal interfaces cross a percolation threshold in which the films interconnect, which leads to DC conductivities increased by a factor of a 100 or more over the subeutectic case [20].  In this initial study, we assume that salt concentrations in Martian ice are below the percolation threshold. (2)  Relaxation (i.e., time-lagged polarization) associated with crystal-defects - This process results a broad relaxation centered at a few tens of kHz, with a tail at 1-60 MHz that causes attenuation even well below any eutectic.  Because dissolved Cl, F, and N ions also increase defect densities, they also increase attenuation from this mechanism and shift the relaxation frequency to somewhat higher values.  This process is evidently important for ice doped with Mg(ClO4)2 at 1 mM concentration, as reported below. (3)  At temperatures near and above a relevant eutectic, microscopic pockets of brine appear, primarily at ice-crystal triple junctions.  Relaxation due to the polarization of mobile charge carriers in these brine pockets (i.e., Maxwell-Wagner effects) lead a second  relaxation frequency at 10s to 100s of kHz at Martian temperatures, thus increasing attenuation at sounding frequencies.  Modeling of this attenuation mechanism depends on data for brine-conductivities at the relevant temperatures and pressures.  Those data are presently limited [22] and are lacking, in particular, for perchlorate brines.  Work to model this mechanism is therefore ongoing [19] and seeks to bound possible effects rather than to specifically predict attenuation in Martian cases. (4)  Also at super-eutectic temperatures, an additional relaxation occurs due to rotation of water molecules in the brine pockets.  In contrast to the relaxations discussed above, the relaxation frequency in this case occurs at frequencies higher than sounding frequencies (at 100s to 1000s of MHz).  Its effect on sub- to super-eutectic temperature variation in attenuation at sounding frequencies can therefore be large.  Modeling this effect depends on data for the resonant (Debye) parameters of the rotation resonance as a function of salt type and content.  Such data are apparently lacking for perchlorate at present [22], so modeling in this case is also limited to bounding possible effects in the Martian situation. Initial Observations:  Here we present initial observations of dielectric properties (and resulting attenuation estimates) for ice doped with Mg(ClO4)2 at a concentration of 1 mM, using methods similar to those of Grimm et al. [21]. Figure 1 shows measurements of the real and imaginary parts of the relative permittivity of ice doped with Mg(ClO4)2 as functions of frequency, for temperatures ranging from well below to well above the eutectic. The shift of the low-frequency relaxation to to higher frequencies with increasing temperature is associated with increased mobility of charge carriers associated with cystal defects, and the presence of perchlorate increases the density of such defects relative their density in pure ice.  Figure 1. Real and imaginary parts (upper and lower panel, resp.) of the complex relative permittivity of 1 mM-magnesim perchlorate-doped ice vs. frequency, for temperatures ranging from 184.4 K to 271.5 K.  Figure 2 shows attenuation for the SHARAD centerfrequency (20 MHz) based solely on a dielectric model resulting from the measurements in Figure 1, i.e., without a rotation-relaxation mechanism.  This is a point of comparison for additional model results to follow.  The calculation for Figure 2 assumes a meanannual surface temperature of 162 K and basal temperature of 212 K.  Although computed attenuation for perchlorate-doped ice considerably exceeds that for pure water ice, these computed attenuations are lower even than those reported by Grima et al. [7] for the polar case. Discussion: While the need for inclusion in our model of  a high-frequency contribution from rotationrelaxation is clear, the current predictions of such low attenuation relative even to polar measurements (at presumably subeutectic temperatures throughout the ice column there) requires further attention.  Layers of water molecules on hydrated mineral surfaces may play a role and may be significantly affected by salts as well [23].  Thus the next phase of this work will address this mechanism, in addition to those discussed above.   Figure 2.  Radar attenuation at the SHARAD center frequency accounting only for measured lowfrequency losses (cf. Figure 1), as a function of depth to the base of an ice deposit (i.e., ice thickness), assuming a linear temperature profile with surface and basal temperatures of 162K and 212K, resp.  References: [1] Calvin, W.M., and T.N. Titus (2008) Planetary and Space Science  56, 212-226. [2] Brown, A.J., et al. (2008), Icarus 196, 433-445.	  [3] Hecht, M.H. et al (2009) Science 325, 64-67. [4] Osterloo, M.M., et al., (2008) Science 319, 1651-1654. [5] Plaut, J.J. et al. (2009) Geophys. Res. Lett. 36, doi:10.1029/	  2008GL036379. [6] Holt, J.R. et al. (2008) Science 322, 1235-1238. [7] Grima, C. et al. (2009) Geophys. Res. Lett. 36, doi:10.1029/	  2008GL036326. [8] Paterson, W.S.B. (1994) The Physics of Glaciers. 3rd ed., Pergamon, Oxford. [9] Milliken, R.E., et al. (2003), J. Geophys. Res. 108(E6),  doi:10.1029/2002JE002005. [10] Arfstrom, J. and W.K. Hartmann (2005), Icarus 174, 321-335. [11] Li, H. et al (2006) LPS XXXVII, Abstract #2390. [12] Winebrenner, D.P. et al. (2008) Icarus, 195, 90105. [13] Kress, A.M. and J.W. Head, Geophys. Res. Lett. 35, doi:10.1029/	  2008GL035501. [14] Hubbard, B., et al. (2011) Icarus, in press. [15} Marion, G.M. et al. (2010) Icarus 207, 675-685.	  [16] Catling, D.C. et al. (2010) J. Geophys. Res. 115  doi:10.1029/	  2009JE003425.  [17] Arfstrom, J. (2004) LPS XXXV, Abstract #1105. [18] Lenferink, H. et al., 42nd Lunar and Planetary Science Conference. Abstract #1909. [19] Stillman, D.E. et al., 42nd Lunar and Planetary Science Conference, Abstract #2193. [20] Grimm, R.E. et al. (2008) J. Phys. Chem. B 112, 15382-15390, doi: 10.1021/jp8055366. [21] Pathare, A.V. and Paige, D.A. (2005), Icarus, 174(2), 419-443. [22] Hasted, J.B. (1973) Aqueous Dielectrics, Chapman and Hall, London. [23] Stillman, D.E. and R.E. Grimm (2010) J. Geophys. Res. in press.  Acknowledgements:  This work was supported by NASA Awards NNX09AI48G and NNX10AJ65G. 
