Grain Boundary Sliding in Ice

It is well known that grain boundary processes intercrystdline sliding and grain boundary migration make an important contribution to high temperature deformation of polycrystalline materials, particularly for metals. More recently, these processes have been correlated with the atomic structure of the boundaries. Our goal is to study sliding and migration of ice boundaries, with respect to temperature, stress, and the crystallographic parameters which characterize the boundaries. We here present our fmt results, obtained on bicrystals with high angle boundaries. The importance of grain boundaries in the high temperature deformation of polycrystalline materials is now well established. In metals, grain boundary sliding has been measured to make an important contribution to high temperature creep. Recent studies have been able to correlate the structure of grain boundaries to their migration and sliding during deformation. Our purpose is to study the kinetics of grain boundary sliding and migration in ice as a function of stress, temperature, and the internal parameters which characterize the atomic structure at the boundary. This is done by submitting bicrystals of different orientations to creep shear stresses. In this paper we present our first results on randomly oriented bicrystals submitted to a constant shear stress parallel to the grain boundary macroscopic plane. 2. Exverimental Procedure The bicrystals used for these experiments were grown by a modified Czochralski solidification technique, used by Landauer [I] and by Homer and Glen [2]. A Peltier cell, which acts as a heat sink, is placed in contact with a plate, onto which are frozen two preoriented single crystal seeds, or a bicrystal seed. The Peltier cell and plate are placed on top of a beaker of water so that the seeds just touches the water surface. The water in the beaker is kept at constant temperature. A temperature gradient between the water and the ice seed is then established, driving the seed's growth into the liquid The growth rate is controlled by slowly removing water from the beaker through an outlet. The crystallographic orientations of the bicrystals were determined by indexing the optical back reflections obtained from hoar frost crystals grown on the surface of a transverse cut at the en& of the bicrystal. Etch pits were also grown in order to confirm the measurements. The actual specimens for the shearing experiments were obtained by hot die extrusion, which produces a long ingot with a rectangular cross section, divided in half by the grain boundary. This method has been used by Itagaki [3], and produces a low dislocation density near the surface. Samples were then cut from the ingot by Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1987127 C1-190 JOURNAL DE PHYSIQUE perpendicular cuts, followed by a slight melting of the cut surfaces. (Figure 1) Their dimensions are 15 mm height, and a 24 mm x 12 mm section. The grain boundary plane has a mean surface of about 15 x 12 mm. Scratches were made on the sample surface so as to reveal the relative displacement between the crystals. The creep device used in our experiments has grips which allow the application of a stress state in which the shear stress is concentrated on the grain boundary plane. (Figure 2) The average shear stress on the boundary is specified by the load which must be transmitted across the boundary from one grip to the other. Tension and compression stresses may be present parallel or perpendicular to the boundary, particularly if there is deformation within the grains. Parasitic torsion and bending stresses are minimized by the symmetrical arrangement of the sample in the grips. The displacement was measured bv an LVDT and recorded continuously. The bicrystals were placed between crossed polarizing sheets, &d backlighting allowed observationskith crystdine contrast during the test. Photographs were taken at regular intervals during the tests. The entire experimental setup was placed inside an insulated box in a colvd room. Near the sakple the temperature osdlations were liss than lo C. Figure 1. (a) Bicrystalline grown ingot. Figure 2. Two view of the shearing device (b). Parallelipiped hot extruded ingot. A and B 1-Fixed Grip, 2-Mobil Grip, 3-Bicrystal sample. are the crystals; I corresponds to a support of ice White arrow shows GB migration during a prelimfor the extrusion. inary test. Lower bicrystal shows cracking that occurred when the load was removed from Bicrystal2. 3. Experimental Results. The experimental set up described above was used for a series of shearing creep experiments. We will describe the results for three separate bicrystals (identified as 1,2, and 3). All three experiments were done at the same temperature: -3.5 + 0.5 "C (0.986 T ) The nominal applied shear stresses were 0.1 MPa or 0.4 MPa, applied alternatively during the tests?reep curves for each sample are shown in Figure 3,4, and 5. The evolution of the bicrystals during the creep tests was followed by sequential macrographs, as shown on the creep curves. Figures 3 5 also show the orientations of the crystals and the position of the boundary in stereographic plots. The direction of the shear must be properly specified with respect to the orientations of the crystals. The point marked CA is the c-axis of the crystal on the right side, which has the stress forcing it down. Samples 1 and 2 were cut from the hot die extrusion so that the shear stress acted perpendicular to the growing direction. For sample 3 the shear stress acted parallel to the growing direction. The three bicrystals showed remarkably different behavior. Bicrystal 1 showed clear evidence of grain boundary sliding at 0.1 MPa. The boundary of Bicrystal 2 showed no sliding at 0.4 MPa, but became wavy and was eventually consumed by the growth of newly nucleated grains. Bicrystal3 Figure 3. Bicrystal 1. Left: stereographic projection of the orientations of the crystals and the plane of the boundary. Right: creep curve with macrographs at different times. Black arrows show GB sliding; white arrows show GB migration. TlME [HOURS] TIME [HOURS) TlME [ I 0 4 SECONDS]