GP Plasma introduces new industrial friendly technique to measure stress profiles within coatings
Leveraging a simple calotte grind and high-quality micro indenter, we reveal a novel approach to measuring mechanical property profiles such as bi-axial and shear stress, particularly around the critical interface regions. Working in close collaboration with your application engineers, we use state of the art materials analysis and measurement techniques to deepen your understanding and optimize your product.
A series of indentations are made within a calotte crater
Our collaboration between Micro Materials and SiOMEC provided a novel approach to measuring bi-axial and shear stress profiles, particularly around critical interface regions.
An experimental coating deposited by PECVD suffered from complete delamination in an aggressive wet abrasion performance test, rather than the gradual wear typically observed after many hours of dry abrasion testing. The on/off type performance behavior from a low-stress coating that can support a thickness up to 250 microns without delamination was puzzling. While the cause is still under investigation, one avenue to explore was the stress state of the coating. We hoped that by measuring the coating properties as a function of thickness, we might understand if we were missing something such as any critical stresses that might suggest a performance risk.
Figure 1. An example calotte-wear crater made by a ball crater wear tester can be used to create a depth profile through a coating (left). Coating thickness ~4.5 microns and choice of polishing grit was critical to achieving a low roughness. A series of indents is then mapped from the outer surface to the center (right). Calo image courtesy of Jose A. Santiago at Nano4Energy.
The realization we could measure localized property information via micro indentation as a function of depth was inspirational in exploring what information we might be able to access with respect to understanding coating failure. Taking special care to achieve a very low surface roughness within the Calo (see Figure 1), we were able to measure the mechanical property changes (importantly the stress) with the resolution we never thought possible in such a short time and accessible cost compared to alternatives. A key advantage of localized property information is that the Young’s Modulus, Yield Strength and Stress are all determined rather than assuming some as constant values. However, in the case of a normally loaded indenter rather than a mixed load one, the Poisson’s ratio is assumed constant in each layer.
The sample is a PECVD coating of approximately 4.5 microns thick deposited on A2 tool steel. Empirically, it was known these samples typically have slight compressive stress (on average) by the curvature induced when a metal sheet is coated, either from the coating growth stress and/or the thermal coefficient of expansion differences.
Figure 2. Hardness profile (GPa) showing the regions of coating and substrate measured across the Calo profile depth (μm). The surface region outside the polished wear scar was discarded due to excessive roughness preventing reliable results with the low indentation load used. For the initial coating profile analysis, we also left out the measurements inside the substrate. However, it is possible to analyze stress profiles across such transitions. Chart courtesy of Adrian Harris at Micro Materials.
The hardness profile was reviewed within the indenter software (Figure 2) to highlight regions of interest. Only data within the red lines were used for further analysis. While it was originally intended to use the surface measurements as a reference set, the unpolished surface roughness was too large. The substrate data will be analyzed at a later date and is not included in the analysis below. The plateau region can be successfully analyzed with coating and substrate contributions separated with the FilmDoctor analytical software.
Curve selection & Analysis
The load-depth curves were checked and selected within the FilmDoctor software rather than traditionally within the indenter software. This provides an option of making your own assumptions on the failure behavior and causes of suspect curves rather than solely relying on the interpretation of the indenter user.
Following curve review and selection, they were then analyzed to extract the mechanical property profiles. The biaxial stress results are presented below in Figure 3.
Figure 3. Relative stress values extracted from a series of micro indents along a Calo wear scar, compared to an arbitrarily chosen indent as a reference that is closest to the outermost surface (left side of the graph) with the profile terminating at the steel substrate interface (right side of the graph). These results present a purely bi-axial stress assumption for each indent location and a constant Poisson’s ratio within the coating. The red box at right highlights the increased error as the interface is approached due to surface roughness having a greater influence on the indentation measurement as the coating becomes thinner. Chart courtesy of Nick Bierwisch at SiOMEC.
For this analysis, a biaxial stress assumption was selected and the resulting profile extracted from the load-depth curve data as shown in Figure 3 above, the left side represents the surface of the coating, and the right side the interface. It is important to understand that being a relative measurement (not absolute) the numerical value of the stress is determined by reference to an indentation measurement point within the data set. This might be chosen within the substrate itself or on the top surface of the coating. In this example, we chose a point nearest to the coating surface and assumed a ‘zero’ stress state for this measured location point. The magnitude of the measured differences between points at each location is accurate, as is the trend of the stress profile, however, the vertical (absolute) positioning of the entire dataset is subject to the chosen reference point. While the color band shows the error of the average stress profile for ‘all’ uncertainties and treats the data points as an ensemble rather than individually (note: typically error bars are only reported for the unloading portion within the load-depth curve for each point).
Notwithstanding the above assumptions, we have observed increasing biaxial tensile stress towards the interface. The relative average ranges from ~180 MPa near the top surface of the coating to ~1.3 GPa (tensile direction) close to the substrate interface. We can also explore a 100% shear stress assumption or any combination of shear and biaxial stress for a more detailed analysis at each measured location. It is worth pointing out that if an absolute stress value is known for the coating (average stress across the coating for example) then the vertical positioning of the relative stress values can be adjusted to match it.
The significant stress increase in the tensile direction measured at the interface would imply a weakened interface likely to suffer adhesion problems or early failure in service, particularly if the maximum Von Mises stress of the applied load reaches this weakened region. Another potential consideration is an aqueous corrosion reaction within this interface zone, where the reaction by-product might lead to a volume expansion creating a further increase in tensile stress. The potential to expose the sample to aqueous corrosion and repeat the test to measure its change in stress state will be an interesting future consideration.
The method is based on a fundamentally derived theory of elasticity using analytical physics to extract additional information from the stress-strain curves from a normally loaded indenter. More discussion regarding the mathematical approach, assumptions, and comparison to other methods will be presented at the upcoming SVC Techcon 2021 Nashville and the ICMCTF 2021 San Diego conferences.
Special Thanks To:
Jose A. Santiago at Nano4Energy for the Calowear, Adrian Harris at Micro Materials UK for micro-indentation with the Nanotest Vantage, and Nick Bierwisch at SiOMEC for the stress analysis with the FilmDoctor software.