Failure analysis, crazing

This topic has been mentioned in Section V, Failure, Defect and Contaminant Analysis, in Chapter 15, where a number of typical practical problem invetsigations were presented. Obviously the potential list of examples exhibiting different characteristics and requiring a different type of analysis is lengthy. When the sample is heterogeneous, e.g., a polymer blend or a composite, the study of the surface of a failed piece of material may reveal whether the problem is the interface of the components or that failure occurred within one of these. In particular in the case of crazing or necking orientation may have been induced, the way this can be analysed is discussed in Chapter 8. 

Other types of commercial coatings and membranes were also evaluated in the test facility at that time, and analysis indicated the performance of the sulfur composite was superior to all the others tested. At the conclusion of the test, all the other systems were in various stages of failure because of cracking, crazing, delamination, or tearing. 

The analysis of the fracture surface in Fig. 27b shows that most of the dPS is found on the PVP side of the interface even for high molecular weight MPS. This interesting result implies that the craze failure mechanism is controlled much more by the DP of the connecting chain than by the DP of the homopolymer. As one would expect, only when the molecular weight of the homopolymer drops below approximately 150 kg/mole (8-9 Ne for PS) does the molecular weight of the homopolymer affect Qc and the craze failure mechanism. These results have been qualitatively confirmed also on a shorter block copolymer (510-540). 

Fatigue failure in neat polymers or short fibre reinforced polymers involves a phase in which a defect zone initiates, with the development of crazes or microcracks, followed by crack propagation to final rupture [17], This statement means that an analysis of fatigue behaviour can use the tools already developed for metals. This will be demonstrated in the cases that follow. 

Porod analysis of the SAXS and SAEX measurements provided a quantitative estimate of the mean craze fibril spacing. Brown [40] subsequently made the key observation that the presence of the cross-tie fibrils has a profound effect on the failure mechanism of a craze because they enable stress transfer between broken and unbroken fibrils. Brown [40], and then Kramer [41], followed this idea through to produce a quantitative theory of craze failure of the molecular chains at the mid-rib of the craze. Brown s theory is a very ingenious mixture of the macroscopic and the microscopic. Starting at the macroscopic level the craze can be modelled as a continuous anisotropic elastic sheet. The stress on the craze plane in front of the crack is then... 

Values for the extension ratios, estimated by Kramer and colleagues [36, 42] and also by Ward and co-workers [18, 22] from analysis of optical interference patterns, compare reasonably well with estimates of the network extensibility from small-angle neutron scattering data [36] or stress-optical measurements [43]. It was therefore proposed that the criterion for craze failure, and hence crack propagation via a craze, is to assume that the entangled strands crossing the section of the craze at the crack tip break due to the development of the critical stress at the crack tip cTfaii,... 

The dependency of Np on a is extremely small. Crack initiation has a very long incubation period. Analysis of the fracture surfaces of fatigue failures in this region reveals furthermore that the slow crack growth mode has two phases. The first shows little evidence of crazing, the second resembles the region-II failure mode. 

---