By Eric Guyer, Ph.D. and Joseph Lemberg, Ph.D., P.E.
It’s true: nothing lasts forever, certainly no useful engineering materials or products made of them do. Moreover, it is not reasonable to expect an engineer or scientist to design a product that does last forever. Indeed, all products have a finite service life. When things do eventually break, whether early or late in their life, it is sometimes important for a manufacturer of the product or an attorney who represents a manufacture to determine why it broke. That’s where Failure Analysts come in. Failure Analysts can be of many engineering and scientific disciplines; the authors of this article are both Metallurgists with specific training in the fracture, fatigue and corrosion of materials. We are often asked to describe the cause as to why something broke. Accordingly, we use tools such as optical microscopes and scanning electron microscopes to examine and study the features on fracture surfaces – this field of study is called “fractography.” We look at features on a fracture surface that are centimeters in size to nanometers (or one billionth of a meter). A common perception is that this field seems more like reading tea leaves as opposed to objective science.
Here we peel back the curtain and discuss briefly some of the features we as metallurgists, fractographers, and failure analysts examine to help us diagnose failures. In doing so, we hopefully shed some light on a process that at times can seem to be a black art.
Common Fracture Modes
When we first learn of a failure, specifically a fracture, we ask several questions in order to understand how and why it occurred:
- What type of material was it and how was it manufactured?
- What type of environment was it subjected to?
- How long was it in service before the fracture occurred?
- Was it subjected to loads or forces and were they sustained or cyclic?
- How was the product used and maintained?
- Is this a new design or old design?
- How many failures exist and how large is the total population of similar parts?
Such questions help us sort through potential failure modes and develop potential hypotheses so that we can apply the scientific method to our investigation. For instance, if a product is never subjected to cyclic loads, then fatigue is not a likely failure mode. Questions about the material help us understand, for instance, if it is generally going to be a brittle or ductile material and whether or not it may be susceptible to attack by various environments. As a general proposition, there are numerous failure modes that exist in reality, only the most predominant failure modes are discussed in this paper which include:
- Environmentally-assisted cracking
Each of these modes is now described as well as compared and contrasted. All leave behind tell-tale signs that a trained metallurgist can use to determine which type of failure occurred.
Briefly, the word “overload” represents a one-time excursion where the load-carrying capacity of the part is exceeded and the part breaks. Overload failures typically originate from a single location that may be a small surface nick, an inclusion in a part, or an area where stress concentrates as a result of the design of a part (a hole, for example). Depending on the properties of the material, a crack can propagate in what is known as a “brittle” manner or a “ductile” manner.
Brittle Overload Fracture
On the microscopic scale, metals are comprised of crystals which we call “grains” (see Figure 1). The morphology of these grains forms what we call the “microstructure” of the metal. Brittle fracture commonly occurs by two means: intergranular fracture (at the interface between adjacent grains) or transgranular fracture (meaning the fracture grows through a grain). The boundaries between adjacent grains may represent a natural weak point in the structure or potentially as a result of contamination of the boundaries as well as other factors. An example of intergranular fracture is shown in Figure 2. Transgranular fracture, also known as cleavage, occurs when a crack plows through a grain, and doesn’t follow the boundaries. Transgranular fracture typically occurs in very hard materials, like ceramics. An example of a transgranular fracture is shown in Figure 3.
Ductile Overload Fracture
Ductile fracture occurs by a different mechanism entirely. A ductile fracture presents a dimpled fracture surface, as shown in Figure 4. These dimples form as the result of tiny voids that form and grow together (coalesce) as the material is deformed. Typically, these voids form around local hard particles, where the nearby material can deform to a different extent than the hard particle; eventually, the hard particles separate from the softer material deforming around them, leading to the characteristic dimples. The crack propagates as these dimples link up in a process known as microvoid coalescence.
In contrast to an overload fracture, fatigue is the cyclic application of loads to a part; here, damage accumulates on a part, a crack is initiated and then it grows to the point of final fracture. It is a time dependent fracture mechanism. In fatigue, a crack can propagate a minute amount with every load-unload cycle. Though the vast majority of fatigue occurs as a result of ductile processes, it is possible to have fatigue of brittle materials. One of the hallmarks of fatigue is the presence of multiple origins. Local inhomogeneities, surface perturbations or surface damage can all lead to the propagation of fatigue cracks.
Most of the lifetime of a part is spent generating the small perturbations, while crack extension typically consumes only a small portion of the life of a part. The fact that a crack extends a small amount with every load cycle leaves behind features known as “striations.” These marks, which are another hallmark of fatigue, point back towards the origin of the crack (picture ripples in a pond from a rock that is dropped). As the crack grows, the spacing between striations increases. Striations start out very small, and require very high magnifications to observe (1000s of times magnification). Such fracture surfaces also generally contain larger features, known as “beach marks” for their similarity to ripples sometimes observed in the sand at the water line at a beach, do not necessary represent a single load-unload cycle (that is, many load-unload cycles may occur between beach marks). An example of striations on a fracture surface is shown in Figure 5.
Environmentally-Assisted Cracking: Stress-Corrosion Cracking
Both of the above fracture modes are primarily and typically related to loads and material properties (temperature can be an exception). However, the service or manufacturing environment can play a role in a fracture as well. One such mechanism is known as stress-corrosion cracking, or SCC. SCC is also a time dependent fracture mechanism and is the result of a sustained stress, rather than a sudden overload or the cyclic application of stresses. A combination of three factors are required for SCC to occur, as described below and shown schematically in Figure 6:
1. Stress: This can be an applied stress from installation or service, or residual stresses left over from manufacturing.
2. Susceptible material
3. Environment: A service or manufacturing environment containing a component that can attack a particular material is required. One common SCC agent for brass is ammonia.
SCC cracks tend to have multiple origins, and usually present as highly branched, intergranular cracks (though transgranular SCC cracks are also possible). An example of SCC cracks in brass is shown in Figure 7.
Fractography is a challenging field of study with many intricacies and subtleties that can impact the outcome of an analysis and accordingly requires a trained eye to accurately diagnose. Hopefully, although only a few of the high level features are examined here, this short article sheds some light on the types of features that we metallurgists look for when examining a failure. The fracture surfaces are therefore extremely important, and are sometimes the only information available to aid in determining what led to a failure.
Eric Guyer, Ph.D., Principal, Exponent, Inc., 3350 Peachtree Rd. NE Atlanta, GA 30326 firstname.lastname@example.org
Joseph Lemberg, Ph.D., P.E., Managing Engineer, Exponent, Inc., 3350 Peachtree Rd. NE Atlanta, GA 30326 email@example.comPrint this article.