Asymmetric Damage Avalanche Shape in Quasibrittle Materials and Subavalanche (Aftershock) Clusters

International audienceCrackling dynamics is characterized by a release of incoming energy through intermittent avalanches. The shape, i.e. the internal temporal structure of these avalanches, gives insightful information about the physical processes involved. It was experimentally shown recently that progressive damage towards compressive failure of quasi-brittle materials can be mapped onto the universality class of interface depinning when considering scaling relationships between the global characteristics of the microcracking avalanches. Here we show, for three concrete materials and from a detailed analysis of the acoustic emission waveforms generated by microcracking events, that the shape of these damage avalanches is strongly asymmetric, characterized by a very slow decay. This remarkable asymmetry, at odds with mean-field depinning predictions, could be explained, in these quasi-brittle materials, by retardation effects induced by enhanced viscoelastic processes within a fracture process zone generated by the damage avalanche as it progresses. It is associated with clusters of sub-avalanches, or aftershocks, within the main avalanche. When slowly driven, many physical systems crackle [1], i.e. respond to slowly varying external conditions through intermittent discrete bursts, or avalanches, spanning a wide range of sizes and/or energies [2]. Classical examples are Barkhausen noise in magnets [3], sheared granular media [4-6], plasticity in crystalline [7,8] and amorphous [9] materials, domain walls [10], or martensitic phase transitions [11]. Damage, fracture and friction within heterogeneous materials also release the incoming mechanical energy through intermittent events, from the laboratory scale [12-14] to earthquakes [15]. In these systems, avalanches are characterized by power laws and scaling relationships indicating the emergence of scale invariance and criticality. A key question is whether these various systems belong to a unique universality class [1,16,17], with the potential consequence of making universal predictions from the analysis of a representative model. This has been largely addressed from the analysis of distributions of avalanche sizes, energies and durations, as well as scaling relations between these variables, hence defining a set of critical exponents. In particular, it has been argued that several of these systems could belong to the well-described universality class of interface depinning [18-20], for which mean field [21] and beyond mean-field [22] predictions are available, although the analogy between yielding and depinning remains controversial [2,23]. In these cases, avalanches are quantified from global variables (size, energy, duration,..). However, the shape, i.e. the internal temporal structure of the avalanche gives as well insightful information about the physical processes involved[24,25]. Universality is also expected in this case, meaning that the properly rescaled average avalanche velocity history () v t should be shared by systems belonging t