Effects of Size Scale on Fracture Processes in Engineering Materials

Abstract

This paper treats two aspects of size-scale. The first is concerned with the size-scale at which fracture events are analysed. Although there are substantial overlaps, four main size-scales may usefully be recognised, as follows:

• i) the MACRO -scale. This is concerned with events at the “engineering” level, and with material properties treated as those of a continuum. Generally, the size-scale is upwards of a few mm but, in some situations, such as that of a single dominant crack in a high-strength steel of homogeneous microstructure, continuum concepts can be carried down to a defect size of 0.2mm.

• ii) the MESO-scale. This comprises inherent “defects” or inhomogeneties, produced by processing or fabrication, which are smaller than the non destructive-testing (NDT) limit. Such defects could be grain-boundary voids in a ceramic; non-metallic inclusions in wrought metallic alloys; “brittle patches” in multi-pass welds or in dual-phase steel microstructures. A very rough estimate of the size-range is 20mm-0.2mm. In ceramics, defects of length 501.tm can produce catastrophic failure at a stress of only 160 MPa. For ultra-high strength maraging steel, a defect of length 75μm could produce catastrophic failure at a stress of 2GPa, but, generally, in engineering alloys, defects of length less than 100μm are of significance only under fatigue loading.

• iii) the MICRO-scale. This is associated with microstructures designed to produce a given combination of flow stress, work-hardening characteristics and fracture resistance. The average properties (such as 0.2% proof stress) may be determined by the dimensions of the metal’s grains, from a few.tm to more than 100µm, but the brittle particles which initially trigger off cleavage fracture are usually less than 10µm in size; the smallest size of significance is of order 10nm.

• iv) the NANO-scale. This involves events at a scale less than some 2nm: a “few” atomic spacings. Typical examples concern the structure of the “core” of a dislocation; the co-ordination of nearest or next-nearest neighbours in a grain-boundary or interfacial “site” for an impurity atom; or events in the region of the tip of an atomically sharp crack as the load applied to it increases. The question here is whether a crack propagates in an “atomically sharp” manner, or whether it blunts, by the emission and gliding-away of crack-tip dislocations.