A dislocation-based analysis of strain history effects in ice

Abstract

The cyclic loading response of ice specimens can be analysed with a dislocation-based model of anelasticity to produce an estimate of the effective mobile dislocation density. Moreover, the cyclic loading response is sufficiently sensitive to track the dislocation density changes that occur during creep straining. A combination of cyclic and creep loading experiments can thus be employed to gain crucial insight regarding the relationship between the dislocation density that evolves during creep straining and the anelastic and viscous components of strain. Creep and cyclic loading experiments have been conducted on laboratory-prepared saline and freshwater ice specimens to shed light on the effects of temperature, creep stress and accumulated strain on the mobile dislocation density and thereby to support the further development of a physically based constitutive model for ice. The findings indicate that, in addition to the expected stress and strain dependence, the dislocation density that develops during straining also depends on the temperature and microstructure. As a consequence of this temperature dependence, the apparent activation energy for creep when dislocation multiplication takes place is higher than that for creep in the absence of dislocation multiplication. Data from the literature are examined and found to support this finding. The results indicate that both the anelastic strain and the viscous strain rate vary with the mobile dislocation density. A preliminary set of results indicates that the power law stress exponent n is 3 when the applied stress is high enough to cause an increase in the dislocation density, but that n ≈ 1 when the dislocation density remains constant during straining. The results show that prior straining increases the stress level associated with the transition from n = 1 to n = 3. The findings provide support for the glide-controlled mechanism of ice creep, and a dislocation glide-based formulation for viscous straining is presented. The model agrees well with the experimental data in the pure flow regime.