Diffusion creep in ice

To delineate the diffusion creep field for ice and develop constitutive laws for ice that describe both diffusion creep and dislocation creep, we have conducted a series of constant load experiments on fine-grained ice (25-80 μm) at 1-atm confining pressure. Ice powders with a particle size <25 μm, produced by quenching a fine mist of water from a pneumatic liquid atomizer in liquid nitrogen, were uniaxially hot pressed for 2-4 h at 100 MPa and T=0.7Tm, yielding samples with an average grain size of ~30 μm and a porosity of <<1%. This starting grain size was increased by annealing at 0.9Tm for 10 h. The temperature was then lowered for the deformation experiment; all temperature changes were made at a rate of 2-3°C/h to prevent microcracking due to thermal expansion mismatch. Hot-pressed samples were deformed in a high-resolution dead-weight creep rig at −37°C at differential stress of 0.5< σ <10 MPa and strain rates of https://www.w3.org/1998/Math/MathML" display="inline"> 10 − 8 < ϵ ˙ < 10 − 4 s − 1 . At sufficiently small grain sizes (25-40μm) and low stresses (<6 MPa), a transition occurs from a rheology characterized by n=2 to one with n=4. This change in stress exponent indicates 1) a transition from diffusion creep to dislocation creep, 2) a transition from dislocation-accommodated grain boundary sliding (GBS) to dislocation creep or 3) a transition from diffusion creep or GBS to a microcracking-dominated regime. The third interpretation can be ruled out because, at the highest stresses used in this study, samples exhibit strengths nearly identical with those of samples deformed at high confining pressures. Creep results for samples with grain sizes of 26 and 40 μm yield a grain size exponent of ~2. The creep data are consistent with models of dislocation-accommodated GBS used to explain superplastic flow.