Effect of disorder on temporal fluctuations in drying induced cracking

Abstract

We investigate by means of computer simulations the effect of structural disorder on the statistics of cracking for a thin layer of material under uniform and isotropic drying. The layer is discretized into a triangular lattice of springs. The drying process is captured by reducing the natural length of all springs by the same factor, and the amount of quenched disorder is controlled by varying the width ξ of the distribution of the random breaking thresholds for the springs. Once a spring breaks, redistribution of the load may trigger an avalanche of breaks. Our simulations revealed that the system exhibits a phase transition with the amount of disorder as control parameter: at low disorders, the breaking process is dominated by a macroscopic crack at the beginning, and the size distribution of the subsequent breaking avalanches shows an exponential form. At high disorders, the fracturing proceeds in small-sized avalanches with an exponential distribution, generating a large number of micro-cracks which eventually merge and break the layer. Between both phases a sharp transition occurs at a critical amount of disorder ξ_c = 0.40 \pm 0.01, where the avalanche size distribution becomes a power law with exponent τ = 2.6 \pm 0.08, in agreement with the mean-field value τ = 5/2 of the fiber bundle model. Good quality data collapses from the finite-size scaling analysis show that the average value of the largest burst <Δ_max> can be identified as the order parameter, with β/ν = 1.4 and 1/ν = 1.0, and that the average ratio of the second m2 and first moments m1 of the avalanche size distribution shows similar behaviour to the susceptibility of a continuous transition, with γ/ν = 1., 1/ν = 0.9. These suggest that the disorder induced transition of the breakup of thin layers is analogous to a continuous phase transition.