Molecular Aspects of the Fatigue and Fracture of Rubber

Abstract

Abstract This short review highlights molecular mechanisms controlling the fracture of rubber and rubber reinforcement, with emphasis on the importance of mechano-chemistry. The manuscript is an extension and update of a previous short review published in Rubber Chemistry and Technology in 1991. All solids, including vulcanizates, contain inevitable ubiquitous flaws of various shapes and sizes. When a solid is subjected to a global stress, the local stresses at the tips of these flaws are magnified and can be many times larger than the average applied stress. Generally, there will be a single flaw at which the stress is magnified to the greatest degree and where fracture begins. Macroscopic fracture of vulcanized rubber is a process in which network chains are broken and new free surface area is created as a result of mechanical loading. Loading may be continuously increasing up to fracture (monotonic fracture) or it may be applied periodically, typically at much lower levels, until fracture ensues (fatigue fracture). For example, a specimen which breaks at, say, 20 MPa when loaded monotonically in tension may fracture, nonetheless, at only 5 MPa, if this load is repeatedly applied. In many rubber articles, two important types of fatigue fracture are fatigue crack growth and abrasion. With the former a (bulk) crack initiates (perhaps at an included impurity or microvoid) and grows as a result of “far-field” loading, whereas, with abrasion, fracture is caused by the direct action of frictional, sliding forces. The events occurring at the tip of a crack are quite important in controlling its growth. In particular, if a crack tip becomes blunted during deformation, or if there are other processes occurring which reduce the load borne by the molecular chains at the crack tip, then stress concentration will be reduced and fracture inhibited. If an elastomeric network is capable of dissipating input energy into heat through irreversible molecular motions, less elastic nergy will be available to break network bonds apart, and fracture energy is increased. More on the role of energy dissipation in fracture is given later. Both fatigue crack growth and abrasion are the culmination of accumulated damage due to mechano-chemical processes. Consider a rubbery article which contains a distribution of chain lengths between crosslink points, and which is subject to fatigue. When deformed, chains align, and the load is inequitably carried by the network strands. The network strives to distribute the stress among the chains, but it is limited from completely doing so because of the complex topology. At sufficiently low elongation, no chains are broken, but as deformation progresses, one network chain eventually ruptures. The force that the chain was carrying prior to breakage is quickly distributed among neighboring chains. This results in the overloading and rupture of some of these chains. (Electron spin resonance spectroscopy has been used to detect free radicals resulting from homolytic cleavage of network chains.) At this point, there has been molecular chain breakage (network damage), but no macroscopic fracture (creation of new free surface area). Chain breakage is not random, but rather is more prevalent in those “elements” where chains broke in the first place. If deformation were monotonously continued to a high level, there would be a particular element which experienced more chain rupture than any other, and a macroscopic crack would open there (far-field loading) or a portion of the material would be removed (frictional loading). However, in (mild) fatigue the article is unloaded well before sufficient damage has occurred in the first cycle to cause crack growth or abrasive loss. After the first cycle, the article contains elements with varying degrees of damage. In subsequent cycles there is more chain rupture (damage accumulation) and eventually macroscopic fracture ensues. It is interesting to calculate the number of chains which must break at one location in order to create 1 µm2 of fracture surface. Assuming a chain cross-sectional area of 0.5 nm2, this would require the rupture of 2(106) chains. Nonetheless, the number of localized molecular chain ruptures necessary to constitute the onset of macroscopic fracture is unclear.