This paper was originally published and presented at The Society of Plastics Engineers’ ANTEC in 2013.
When a plastic part fails, a tough question is often asked, “Why are a limited number of parts failing?”. This is particularly true with seemingly random failures at significant, but low, failure rates. Two aspects are generally linked to such low failure rates, multiple factor concurrency and the statistical nature of plastic failures. Failure often only takes place when two or more factors take effect concurrently. Absent one of these factors, failure will not occur. Plastic resins and the associated forming processes produce parts with a statistical distribution of performance properties, such as strength and ductility. Likewise, environmental conditions, including stress and temperature, to which the resin is exposed through its life cycle is also a statistical distribution. Failure occurs when a portion of the distribution of stress on the parts exceeds a portion of the distribution of strength of the parts. This paper will review how the combination of multiple factor concurrency and the inherent statistical nature of plastic materials can result in seemingly random failures.
A useful definition for the purpose of understanding failure is an undesirable event or condition that results in the inability of a component to function properly or perform its intended function safely, reliably, and economically. In many cases, failure is catastrophic and results in a component that is completely inoperable. In other cases, however, the part may be partially operable, but not fully functional; or simply may be compromised to the point that it is deemed that further use is unreliable or unsafe
Failure within plastic components can take many forms, including:
- Deformation / Distortion
- Esthetic Alteration
This paper focuses on failure through fracture in illustrating the role of multiple factors in plastic part performance and the statistical nature of plastic part failure. However, it is important to note that the same principles apply to the other types of failure as well.
The mechanism of cracking within a plastic material principally occurs through disentanglement, whereby polymer chains slide past one another. The applied stresses, both internal formed-in and assembled-in together with those from external sources overcome inter-molecular forces, such as Van der Waals forces, London dispersion forces, hydrogen bonding, and dipole interactions. The mechanism is the same regardless whether the polymer is amorphous or semi-crystalline. Generally, the stresses responsible for cracking are insufficient to produce breakage within the covalent bonds of the polymer backbone.
While it might seem complex, it is important to remember that cracking is simply a response to stress. Fracture takes place as a stress relief mechanism. Ductile fracture is a bulk molecular response that occurs through yielding, a macro molecular rearrangement, followed by disentanglement. Conversely, brittle fracture is a micro molecular response where disentanglement is favored over yielding. The most common failure mechanisms of plastic components are:
- Short-term Overload
- Creep Rupture
- Environmental Stress Cracking
- Molecular Degradation
This paper addresses two fundamental questions of plastic component failure:
Why are a limited number of parts failing? – Statistical Distribution Considerations
Why are a limited number of parts failing? – Concurrent Factors Considerations
The cause of failure in plastic parts is not typically trivial. Parts can fail as the result of various factors concurrently reducing the expected strength and/or increasing the expected in-service stresses of the part. The number of factors affecting part performance produces a broad statistical distribution of the material properties and service conditions. The overlap of these two statistical distributions leads to sporadic and seemingly random failures. When working to identify the cause of component failure, it is important to consider the interaction between the various factors and the inherent statistical distribution of the performance and service conditions.