Craze nucleation

Also shown on Fig. 1 is the important effect of the state of stress which may be characterized by the negative pressure CTq and the deviatoric stress Sq. These are given by 

Despite this lack of reproducible experimental evidence on kinetics, one may postulate several microscopic steps involved in craze nucleation. Imagine a polymer surface under simple tension as shown in Fig. 2 a. The first is logically plastic 

Kawagoe and Kitagawa proposed a somewhat different microscopic model in which two shear bands or patches are assumed to interact at their intersection to form a microcrack which expands elastically to a cylinderical pore. In this treatment although the surface energy of the pore is considered, it is supposed that no thermal activation of pore formation is possible. 

Gent proposed a quite different model for craze nucleation. The local hydrostatic stress Go (concentrated by the presence of flaws) was supposed to decrease Tg of the glass to the ambient temperature. Upon reaching the rubbery state, the polymer will cavitate easily to form voids. The main problem with this mechanism is the large stress concentration factors that are necessary for it to operate at room temperature it also cannot easily account for the time dependence of craze nucleation under constant stress. It provides a possible explanation, however, for the nucleation of crazes II In view of the fact that the work of Kausch and Dettenmaier on craze nucleation and intrinsic crazing will be treated in detail in Chapter 2, it will not be discussed further here 

---

Craze Nucleation Theory. In various ways it has been suggested that the role of the rubber particles is that of stress concentrators. Thus, Schmitt and Keskkula (33) believe that the multiplicity of stress concentrators (i.e., a multitude of weak points) produce a large number of small cracks rather than a few large ones more energy is needed to propagate a large number of small cracks, and stress fields of the various... 

Further, the Goodier equations predict that hard particles and voids produce higher stress concentrations (i.e., stronger craze nucleation) than rubbers, and thus hard particles and voids should toughen even better than rubbers if nucleation were the operative mechanism. This is not observed experimentally. The nucleation theory is thus seen to have substantial drawbacks. 

Experimental Evidence. Morphology. Figure 3 (33) shows in phase contrast microscopy the development of crack or craze patterns around rubber particles in a toughened polystyrene. The lack of dependence of crack inclination on direction of stress is especially marked in this micrograph, and can be explained only by reference to dynamic branching rather than to crack or craze nucleation by stress raisers. Schmitt and Keskkula refer to the lines as craze cracks and cracks. ... 

The craze nucleation in bulk polymers results from nucleation of voids in the plane strain region of the sample to relieve the triaxial constraints. 

Figure 5. A uniform distribution of thick crazes nucleated from a good dispersion of rubber particles. Large number of crazes.

This study (34) implies that a right dispersion of rubber particles may permit optimum stress field overlap that affords lower craze-initiation stresses and therefore can rapidly dissipate the strain energy in the HIPS. A more homogeneous spatial distribution of rubber particles allow for a uniform development of crazes. Prevention of the strain localization phenomenon to avoid the detrimental situation, where crazes prefer to develop in certain areas and quickly lead to a catastrophic crack, could result in a larger total volume of crazed material. Further, Donald and Kramer (22) discovered no crazes nucleating from an isolated rubber particle with diameter smaller than 1 urn because of an insufficient size of stress-enhanced zone. Since Sample-A has a small average particle size it should contain a large number of small rubber particles. Two small rubber... 

The cohesive surface considered in the foregoing is based on observations made under quasistatic conditions. In particular, the incubation time for craze initiation is neglected and a critical stress state for craze nucle-ation is used (Eq. 11). For dynamic loading, a time-dependent craze initiation criterion is to be included in the kinetics, since the characteristic timescale associated with the loading can be comparable to that involved in the craze nucleation process. If the time for craze initiation is accounted for, another timescale is involved in the competition between crazing and shear yielding that determines whether or not crazing takes place. Therefore, a switch... 

Early in the study of crazing it was recognized that there was usually a time delay between the application of stress and the visual appearance of crazes. This delay time is evidence of a barrier to craze nucleation. The experimental situation may be suimnarized by Figure 1, which shows the density of crazes as a function of time under various states of stress in polystyrene (PS) from the work of Argon and... 

Fig. 2a—c. Schematic drawing of several postulated microscopic steps in craze nucleation a Formation of a localized surface plastic zone and buildup of significant lateral stresses, b Nucleation of voids in the zone to relieve the triazial constraints, c Further deformation of polymer ligaments between voids and coalescence of individual voids to form a void network... 

There are two important questions about craze growth, namely what are the mechanisms of craze tip advance (expansion of the craze periphery generating more fibrils) and craze thickening (normal separation of craze surfaces lengthening the craze fibrils). Unlike the cloudy experimental situation regarding craze nucleation, that regarding craze growth now seems quite clear. 

It is observed that the normal craze fibril structure can be observed just behind the craze tip where the craze is as thin as 5—lOnm . This observation was difficult to reconcile with early models of craze tip advance which postulated that this occurred by repeated nucleation and expansion of isolated voids in advance of the tip. One problem was to explain how the void phase became interconnected while the craze was still so thin. Another was that the predicted kinetics of craze growth appeared to be incorrectly predicted indeed since this mechanism almost involves the same steps as the original craze nucleation, it is hard to understand how craze growth could be so much faster than craze nucleation as observed experimentally. 

Craze nucleation appears controlled by the nucleation of voids in localized regions undergoing large unstable plastic deformation. The sensitivity of this process to the nature of the flaw structure of the surface makes a detailed comparison of data between different experimental groups or between experiment and theory very difficult. 

In simple tension and tension-compression fatigue, HIPS deforms by craze nucleation and growth while ABS deforms primarily by shear. Crazes develop in ABS prior to fracture but at a later stage than does shear deformation. 

---