JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 97, NO. El2, PAGES 20,883-20,897, DECEMBER 25, 1992 Effects of Dispersed Particulates on the Rheology of Water Ice at Planetary Conditions WILLIAM B. DURHAM Lawrence Livermore National Laboratory, Livermore, California STEPHEN H. KIRBY AND LAURA A. STERN U.S. Geological Survey, Menlo Park, California We have investigatedthe effects of initial grain size and hard particulate impurities on the transient and steady state flow of water ice I at laboratory conditions selected to provide more quantitative constraints on the thermomechanicalevolution of the giant icy moons of the outer solar system. Our sampleswere molded with particulate volume fractions, •b, of 0.001 to 0.56 and particle sizes of 1 to 150/am. Deformation experiments were conducted at constant shortening rates of 4.4 x 10 -7 to 4.9 X 10 -4 s-1 at pressures of 50and 100 MPaand temperatures 77to 223 K. Forthe pure icesamples, initial grain sizes were 0.2--0.6 mm, 0.75-1.75 mm, and 1.25-2.5 mm. Stress-strain curves of pure ice I under these conditionsdisplay a strength maximum tr u at plastic strains e -< 0.01 after initial yield, followed by strain softening and achievement of steady state levels of stress,trss, at e = 0.1 to 0.2. Finer starting grain size in pure ice generally raises the level of tr u. Petrography indicates that the initial transient flow behavior is associatedwith the nucleation and growth of recrystallized ice grains and the approachto trss evidently corresponds to the development of a steady state grain texture. Effects of particulateconcentrations •b< 0.1 are slight.At theseconcentrations, a small but significant reduction in tr u with respect to that for pure water ice occurs. Mixed-phase ice with •b -> 0.1 is significantly stronger than pure ice; the strength of sampleswith •b = 0.56 approachesthat of dry confined sand. The magnitude of the strengthening effect is far greater than expected from homoge- neous strain-rate enhancement in the ice fraction or from pinning of dislocations (Orowan hardening). This result suggests viscous drag occurs in the ice as it flows around the hard particulates. Mixed-phase ice is also tougher than pure ice, extending the range of bulk plastic deformation versus faulting to lower temperaturesand higher strain rates. The high-pressure phase ice II formed in •b = 0.56 mixed-phaseice during deformation at high stresses. Bulk planetary compositionsof ice + rock (•b = 0.4-0.5) are roughly 2 orders of magnitudemore viscousthan pure ice, promoting the likelihood of thermal instability inside giant icy moons and possiblyexplainingthe retention of crater topography on icy planetary surfaces. INTRODUCTION We have extended our laboratory study of the flow and fracture of water ice I [Durham et al., 1983; Kirby et al., 1987] to include ice + rock mixtures. Water is one of the more important volatile constituents of many of the low- density moons of the outer solar system, and since most of these moons have some rocky component, ice + rock mixtures must exist or have existed in at least some regions on these moons. Our experiments are intended to constrain evolutionary models of the icy moons. Of particular interest is the so-called Ganymede-Callisto dichotomy, which has puzzled planetologists since the Voyager flybys of Jupiter; althoughthe two giant moons have nearly identical compo- sitions, they have radically different external appearances [McKinnon and Parmentier, 1986; Mueller and McKinnon, 1988]. The difference between the moons may have resulted from a thermomechanical instability during the evolutionary process.Friedson and Stevenson [1983] have suggested that subtle differencesin ice/rock ratios can produce rheological differences that can drive an icy moon to differentiate (Ganymede?) or not differentiate (Callisto?). The determination of the rheology of ice + rock mixtures is also crucial to understanding the nature of planetary Copyright 1992 by the American Geophysical Union. Paper number 92JE02326. 0148-0227/92/92 JE-02326505.00 surfaces,particularly regarding viscous relaxation of craters and other topographic features. The surface layer is plausi- bly rich in rocky material because it has remained suffi- ciently cold to prevent differentiation by melting and be- cause it contains rocky dust that continuously rains down on surfaces.Varying ice/rock ratios are often cited as explana- tions for strong variations in albedo of icy satellites (see, e.g., McKinnon [1985] for a more detailed review of these topics). Knowledge of the rheology of ice + rock mixtures may eventually help constrain surface compositions, al- though interpreting viscously relaxed crater shapes is cur- rently controversial: some models of crater relaxation using our flow laws for pure ice suggest that pure ice I is too weak to support current crater topography [Croft, 1988; Thomas and Schubert, 1988], while others suggest that observed crater topographyis compatible with pure ice rheology [e.g., Schenk, 1991; Hillgren and Melosh, 1989]. Geissler and Croft [1988] and Thomas and Schubert [1988] argue that silicate particles can harden ice through the process of dispersionhardening. Dispersed particles harden the mate- rial by pinning glide dislocations and forcing them to climb over obstacles or to bow and pass around them. The purpose of the present study is to investigate the effects of hard particulates on the flow of ice I at conditions appropriate to the surfaces and interiors of large icy moons. Our results from several related studies, which have helped us interpret the results of this investigation of the effect of 20,883