Microscopy of Polyoxymethylene
by Steven MacLean
Department of Material Science, University of Rochester, Rochester, NY
Polyoxymethylene (a.k.a. polyacetal or POM) is versatile engineering thermoplastic resin manufactured by several raw material suppliers, including DuPont, BASF and Ticona. Homopolymer POM is traditionally made from ring opening polymerization of formaldehyde and is typically sold in pellet form. Pelletized resin can be formed into complex three dimensional shapes via standard conversion processes such as injection molding or sheet extrusion. Due to its excellent physical properties, similar to those of Nylon 66, POM is often used in load bearing applications such as gears, pump housing, impellers and automotive components. An important distinction between POM and Nylon is that POM is hydrophobic and does not experience moisture uptake issue.
As engineered materials, like POM, have become poplular choices for structrual applications, there has also become a need to better understand thier long-term performance characteristics. In particular, it is crucial to understand how these materials fail under certain loading conditions such as stress overload or fatigue loading. Over the past 20 years, microscopic analysis of polymeric fracture surfaces has been a vital tool aiding the failure analyst to ascertain the root cause of product failures. This web site explores the resultant fracture surface topography of POM under known failure conditions. In this investigation, ASTM Type 1 tensile specimens were tested under tensile loading (ASTM D638) and fatigue loading (modified ASTM D790). Also, two additional sets of samples were exposed to a citric acid cleaner for 30 days prior to performing the tensile and fatigue tests to better understand the cleaners influence on crack initiation and crack propagation. This is a common cleaner used in salt water pumps and often comes in direct contact with POM during the pump's normal operating conditions.
Tensile Specimen Fracture Surfaces
Figure 1: Control Sample
Figure 2: Acid Cleaner Sample
Standard tensile testing is comprised up pulling a tensile specimen (dog bone shaped) at a constant strain rate (velocity) until the specimen fractures into two distinct pieces. Once the specimen has catastrophically failed, the newly created fracture surfaces can be examined in an effort to identify key features such as crack origin, crack growth direction and qualitatively assess the ductility or brittleness of the material. Figure 1 is an SEM image of a crack origin for a tensile specimen tested under ambient conditions. The image was captured at a working distance of ~16mm with 4.0 kV accelerating voltage at 200X magnification. Noteworty features include a true crack nucleation site (bullseye) in the center of the image along with symmetric, rapid crack growth resulting in an "inverted volcano" three dimensional structure. Outside of the circular region, the reader should make note of the change in surface topography; this is where the crack growth rate has decelerated and the fracture surface began to progress on multiple planes (especially on the left hand side of the image). In general, this is a classic crack initation site for ductile materials like POM.
In contrast, Figure 2 represents a typical brittle crack initiation site. In this case, the embrittlement of the POM was facilited by exposing the tensile specimen to a citric acid cleaner - a common fluid used to defoul salt water pumps. In this image, the crack origin is located at the intersection of two outside surfaces of the tensile specimen and radiates toward the geometric center of the fracture surface. If the reader examines the image closely, one will notice faint lines in the clam shell area running radial outward from the crack initiation site. These features are commonly referred to as "hackle marks" and can also be easily seeen in Figure 1 and are indicative of fast, crack growth. The other most noteable difference between these two tensile specimen images is the areas outside of the crack origins. As previously mentioned, the surrounding area in the control sample is feature-rich with multiple fractures planes - a good indication of ductility on the microscopic level. Whereas the surrounding area outside of the crack initiation site in Figure 2 is very flat, featureless and somewhat glossy. These characteristics are indicitive of brittle crack growth which is common of cracks that have initiated in the presence of a foreign chemical or reagent.
Fatigue Specimen Fracture Surfaces
Fatigue Specimen Fracture Surfaces
Figure 3: Control Sample
Figure 4: Acid Cleaner Sample
Figure 5: Fatigue Striations
Figures 3, 4 and 5 are SEM images of fracture surfaces created from specimens subjected to fatigue loading. Fatigue loading is the application of a low or moderate cyclic force (or stress) on a test specimen. Depending on the size of the applied cyclic load, it can often take thousands, if not millions, of loading cycles before the specimen fails catastrophically. Figure 3 is an image of the crack initiation site for a control fatigue sample. Overall, it is similar to the tensile control crack initiation site, but there are subtle, yet important differences that will be discussed later.
Figure 4, the acid cleaner sample, also resembles its tensile counterpart, with the exception of new circumfrential lines found in the clam shell region of the crack origin. These circumfrential lines are commonly referred to as fatigue striations and often are used to help distinguish fatigue failures from other possible failure mode such as stress overload or creep. Figure 5 is a 100X magnification of the fatigue striations found in the clam shell area. Using the marker length in the image, the distance between fatigue striations is approximately 50 micrometers. This distance represents the amount of crack growth due to a single loading cycle. With this quantitative information, crack growth rates can be established - a common and very important fracture mechanics parameter.
Topography Differences Between Tensile and Fatigue Fracture Surfaces
One of the main goals of this investigation was to be able to differentiate between stress overlaod (tensile) and fatigue failures by examining the corresponding fracture surfaces of each loading type. At first glance, the individual crack initiation sites look quite similar, as can be seen in Figures 1 and 3. Both failure modes induced a circular origin site with a dark central nucleus. However, upon closer inspection, the reader should note the topography differences between these circular regions. The tensile specimen is feature-rich with many hackle marks emanating radially from the dark core. Whereas the fatigue specimen is smooth and featureless and does not posses any hackle marks. Moreover, there are faint fatigue striation circumfrentially located throughout the circular region.
Outside of the circular region there are significant differences as well. The tensile specimen's topography is very jagged with many peak and valleys - a clear indication of rapid and random crack growth. In contrast, the material away from the crack origin in the fatigue sample is rather smooth and flat, almost planar in nature. This is indicative of slow moving crack groth and is typical of fatigue failure.
Acid Cleaner Effects
Figure 6: Control Sample
Figure 7: Acid Cleaner Sample
Another goal of this investigation was to assess the changes in fracture surface morphology due to environmentallly assisted crack growth. This was accomplished by subjecting a second set of tensile and fatigue samples to an aggressive citric acid cleaner for 30 days prior to loading each set to failure. Figures 6 and 7 are fracture surface images (far away from the crack origin) of exposed samples. It is easily seen that there are appreciable differences between the two resultant surfaces. The unexposed sample is, once again, feature-rich with indications of stress whitening (ductility). The fracture surface from the exposed sample is glossy and flat with no signs of ductility.
TEM ImagingTransmission Electron Microscopy (TEM) is often used to evaluate crystal structures in polycrystalline materials like metals and alloys. Similar concepts, to some degree, can be translated to evaluate the resultant crystal structure of a semi-crystalline polymer. Unlike FCC and BCC metals, there is not a dominant, uniform crystal structure, in polymeric materials. In some cases polymeric materials are completely amorphous and possess no degree of crystallinity. On the other hand, semi-crystalline materials, such as POM, do possess crystal domains randomly distributed in and around the amorphous regions of the microstucture. Figure 8 is a bright field TEM image of a microtomed POM sample (thickness of ~100 nm). Although the sample is being slightly degraded from the electron beam, the image clearly shows two distinct regions in the sample. The first is dark circular regions sporadically distibuted throughout the sample - in three dimensional space these are spherical crystal domains of the material. Between the dark circular regions are lighter areas which represent the amorphous phase of the polymer. Qualitatively, the material appears to have a ~50% degree of crystallinity, which is typical for POM materials. This is a strong indication that the material was properly processed during extrusion or injection molding.
EM Degradation of POM
Figure 9: SEM Degradation at 500X
Figure 10: SEM Degradation at 3000X
Figure 11: TEM Degradation at 2000X
As with any experiment, there author did experience some minor set backs along the way. The most noteable issue was rapid and severe degradation of the POM at high magnifications or high accelerating volatages. Typically, when the magnification level was set higher than 200X the material began to burn and small burn holes formed throughout the imaged area, as can be seen in Figure 8. When magnification levels in excess of 1000X were employed, such as the image in Figure 9, rapid degradation ensued creating large voids throughout the sample. The author would have enjoyed viewing some of the previously mentioned features, such as hackle marks and fatigue striations at higher magnifications, but was simply unable to due to this phenomenon. Likewise, in the TEM aggressive degradation took place at magnification levels at 2000X and higher. When degraded the sample would resemble a piece of paper being burnt by fire around the edges as is shown in Figure 10.
Electron microscopy is a powerful tool for investigating fracture surfaces of failed polymeric components. Often, the failure analyst relies heavily on the techniques presented in this study to identify key surface features such as crack initiation sites, direction of crack growth and loading type that are critical in any root cause investigation. Proper identification of the dominant failure mode is essential when determining an appropriate corrective action plan such as redesigning the failed component or selecting a more suitable material.
The primary takeaway from this study was that extreme caution must be exercised when examining polymeric materials in an SEM or TEM. Highly energized electrons, which are part of the imaging beam, can quickly degrade the material to the point where surface structure and features are substantially compromised, which could lead to misinterpretation of the imaged area. The key to successful imaging of these materials, is to use low accelerating voltages, view at low magnifications and ensure that proper coating techniques are employed to minimize degradation and charging.
In closing, the author would like to thank Brian McIntyre, the instructor, for his guidance and support throughout this project. Many of these images would not have been captured without his assistance.