Analytical methods used to determine failure modes in rubber-to-metal parts.
Failure determination in rubber-to-metal bonded parts can be a daunting task. Multiple layers of polymeric and inorganic coatings tie the rubber to the metal. An understanding of the failure must start with the identification of which layer failed and where in the layer the failure occurred. We have found that a systematic approach, starting with characterization of both sides of the failure to identify where the failure occurred, followed by characterization of the chemistry of the failed layer, leads to a root cause determination in many situations.
Often, the first approach in analyzing a failure is optical microscopy. Although the need to look at the surfaces is obvious, this step can be decisive in determining the next most logical step in diagnosis. One must proceed with caution, how ever, since the visual appearance may not correctly indicate the interface of failure if layers are very thin or the colors are not distinct. Many of the adhesives, primers and inorganic coatings (phosphates) contain heteroatoms not found in the surrounding layers. Energy dispersive spectroscopy (EDS) is an elemental surface analytical tool coupled to some scanning electron microscopes (SEM).
Comparison of the chemistry of the rubber side (meaning the side of the failure originating on the rubber, not necessarily that rubber is at the surface) and the metal side helps to quickly determine if the failure is cohesive (same chemistry on both sides) or adhesive (each side has chemistry related to different layers of the structure). This also provides the first opportunity to compare the chemistry of the failed layer to that of a good performing standard. From here we can decide if the use of surface sensitive techniques like infrared spectroscopy (IR), x-ray photoelectron spectroscopy (XPS) or secondary ion mass spectrometry (SIMS) is appropriate.
Blisters are often a sign of delamination of layers or particulate contamination. Carefully preparing the blister and examination of the blister both visually and chemically can lead to the root cause of blister formation. In the example that follows, blisters were seen on a part consisting of metal, phosphate, primer, adhesive and rubber layers. In figure 1, a particle can clearly be seen in the SEM image of the rubber side (figure 1b) of the blister and the adhesive coating is intact on the metal side (figure 1a). Using EDS, as shown in figure 2, the particle was found to contain Ti, an element from the primer. The metal side of the blister shows the same chemistry as the adhesive, leading to the conclusion that the particle of primer and adhesive was deposited on the surface of the coated adhesive part prior to bonding with rubber. Further investigation found a source of these particles at the manufacturing facility.
[FIGURES 1-2 OMITTED]
Cross sectional analysis
The deposition of the proper thickness of each layer in rubbe-t-o-metal bonded parts is crucial to good performance. Layers that are too thick are often associated with cohesive failures of that layer; whereas layers that are too thin have more variable failures. A failure was presented where the rubber side of the failure had a visually brown appearance in some areas and black in others. Again, this part had a metal, phosphate, primer, adhesive and rubber construction. In this system, the rubber and adhesive were black, and the primer and phosphate layer were dark gray. Although the visual appearance of a failure is not always indicative of the cause, in particular where there are very thin surface coatings, the observation of the brown color was of concern since it often is associated with rusting. The EDS spectra from the black and brown regions from the surface of the rubber side of the failure are shown in figure 3. The black region appears to be similar to the chemistry of the adhesive, whereas the brown region has elements from both the primer (Ti, CI) and the phosphate layer (Zn, P, Fe). A cross section of the bonding layers was prepared and characterized using EDS imaging. In figure 4, the metal, phosphate, primer, adhesive and rubber layers can be identified. In some regions, the phosphate layer is either very thin or discontinuous. X-ray photoelectron spectroscopy (XPS) was also used to characterize the top 10 nm of the surface of the rubber. Using this technique, inorganic CI can be differentiated from organic CI, and it is important to determine if there is any inorganic CI on surfaces that visually appear to contain rust. In the brown region, most of the CI was organic; however, no Ti was detected on the rubber surface. This implies that the Ti seen in the EDS data is subsurface and greater than 10 nm deep, the sampling depth of XPS. Taken together, these data suggest failure within the phosphate layer, but probably within 1-2 microns of the primer. By XPS, potassium (K), sodium (Na) and sulfur oxides (S[O.sub.x]) were also detected. These species are most likely from the Cleaning system and may be from insufficient rinsing of soaps from the surface. It is suspected that these species would form a weak boundary layer that could lead to failure.
[FIGURES 3-4 OMITTED]
Based on the analytical characterization of the failure, there are two potential mechanisms for bond failure. The first is the very thin phosphate layer, and the second is the surface contamination of the phosphate. Either mechanism could cause failure, but the presence of both issues could make bonding difficult. There were no chemical indications of rust.
Tracing in-plant contamination
Identification of a contaminant on a surface is almost always followed by speculation as to the source of the contaminant and the mechanism of contamination. In one case, a phosphate bath was depositing fatty acids as well as phosphate on surfaces in a manufacturing facility. Secondary ion mass spectrometry (SIMS) was used to characterize part failures, dried films of the phosphate bath and lubricants used in the manufacturing plant. SIMS detects the elemental and molecular chemistry of the top 2.5 nm of a surface. Figure 5 shows spectra from the phosphate bath, as well as from three lubricants. The spectrum from the phosphate bath identifies fatty acids, and lubricant 1 is a very good match for the fatty acid contaminant.
[FIGURE 5 OMITTED]
In some instances, we are asked to determine which contaminants could impact the cure or performance of an adhesive. Thermal transitions associated with cure are investigated by differential scanning calorimetry (DSC). Using this technique, the amount of heat required to raise the temperature of a sample is monitored. If the temperature of the sample rises when no additional heat is applied, that is termed an exotherm and can indicate cure. The area under the exotherm curve indicates the amount of reaction. In figure 6, there are three DSC curves showing a Rohm and Haas control, a customer retain and the Rohm and Haas control purposefully contaminated with Mn phosphate. The cure exotherm extends from approximately 170[degrees]C to 230[degrees]C in the control sample. It is apparent that addition of Mn phosphate interferes with cure, and that the retain sample of the same adhesive provided by the customer has been compromised. Other potential contaminants found in the plant did not adversely affect cure: however, other deleterious effects could occur from any contamination.
[FIGURE 6 OMITTED]
Characterization of rubber-to-metal failures can be a challenging task. The use of a well equipped analytical characterization facility can conclusively determine the cause of the failures and lead to the prevention of future failures.
|Printer friendly Cite/link Email Feedback|
|Author:||Pacholski, Michaeleen L.|
|Date:||Jan 1, 2007|
|Previous Article:||Overview of capillary rheometry--part 1.|
|Next Article:||Independent testing laboratories directory.|