Periodic Reporting for period 3 - TriboKey (Deformation Mechanisms are the Key to Understanding and Tayloring Tribological Behaviour)
Période du rapport: 2021-09-01 au 2023-02-28
The science of friction, wear and lubrication is called tribology. Friction and wear are responsible for more than 20 % of mankind’s primary energy usage. Therefore CO2 emissions could be significantly reduced if better tribological practices were being available. This could in large part be achieved by designing new materials to be employed in tribological systems. Such a strategic materials development is currently however not possible. This is in part true as the so-called elementary deformation processes acting in materials – and we are focusing on metals here – are not sufficiently known and understood.
The overall objective of this research therefore was and is to identify elementary deformation mechanisms acting in metals that are subjected to a tribological load, with the long term goal to be able to provide guide lines for alloy and microstructure development.
In order to achieve this goal, we mainly focus our attention on model materials like high-purity copper, as well as on copper alloys, as here these mechanisms can be studied. This focus on elementary mechanisms during the implementation of the action was extended to the field of tribologically induced chemical changes, mainly tribo-oxidation, which is also an important failure mechanism.
One of the challenges that such materials tribology research faces is that these deformation mechanisms are difficult to detect. This is mainly true as every tribological contact constitutes a so-called “buried interface”, meaning that it is normally not possible to directly observe the processes acting in the contact itself. Traditionally elementary mechanisms are therefore studied by interrupting a tribological experiment after a pre-defined sliding distance and the material is then investigated by electron microscopy.
The area of the materials that one is focusing its attention on hereby is not the surface itself, but the material under the surface, down to a subsurface depth of several micrometers. We made extensive use of this approach during this action. However, we also aim at developing a completely new way of in situ probing the subsurface structure of metals subjected to a tribological load. To this end we are designing and building a new tribometer that probes the subsurface structure of the metal in the contact by means of non-destructive testing methods, such as ultrasound and acoustic emission. With this new tribometer at hand, we aim at studying elementary deformation mechanisms for a range of copper-based model materials with the ultimate aim of being able to come up with design guidelines for alloy development. It needs to be mentioned that the latter is a very ambitions goal, which might not be achieved through the course of this action, while at the same time the results obtained will be crucial steps towards this aim.
The overall objective of this research therefore was and is to identify elementary deformation mechanisms acting in metals that are subjected to a tribological load, with the long term goal to be able to provide guide lines for alloy and microstructure development.
In order to achieve this goal, we mainly focus our attention on model materials like high-purity copper, as well as on copper alloys, as here these mechanisms can be studied. This focus on elementary mechanisms during the implementation of the action was extended to the field of tribologically induced chemical changes, mainly tribo-oxidation, which is also an important failure mechanism.
One of the challenges that such materials tribology research faces is that these deformation mechanisms are difficult to detect. This is mainly true as every tribological contact constitutes a so-called “buried interface”, meaning that it is normally not possible to directly observe the processes acting in the contact itself. Traditionally elementary mechanisms are therefore studied by interrupting a tribological experiment after a pre-defined sliding distance and the material is then investigated by electron microscopy.
The area of the materials that one is focusing its attention on hereby is not the surface itself, but the material under the surface, down to a subsurface depth of several micrometers. We made extensive use of this approach during this action. However, we also aim at developing a completely new way of in situ probing the subsurface structure of metals subjected to a tribological load. To this end we are designing and building a new tribometer that probes the subsurface structure of the metal in the contact by means of non-destructive testing methods, such as ultrasound and acoustic emission. With this new tribometer at hand, we aim at studying elementary deformation mechanisms for a range of copper-based model materials with the ultimate aim of being able to come up with design guidelines for alloy development. It needs to be mentioned that the latter is a very ambitions goal, which might not be achieved through the course of this action, while at the same time the results obtained will be crucial steps towards this aim.
As described above, we focus our attention on the elementary mechanisms of microstructural changes in materials subjected to a tribological load. To this end we subjected ultra-high purity copper and its alloys to mainly reciprocating tribological experiments. Our main counter body material was sapphire, as here we can be sure that sapphire does not take part in any microstructural changes nor in chemical processes taking place at the interface.
For pure copper we investigated the influence of different counter body sizes as well as normal loads. We saw that for all experimental conditions chosen, at the very beginning of sliding a dislocation self-organization phenomenon takes place, which leads to a horizontal dislocation feature about 100 to 400 nm under the sliding contact. The named this feature the “dislocation trace line”, or DTL. The DTL is the first discontinuity in the microstructure and therefore of key importance for any further changes to the materials. This is also true for the position where as sliding progresses material delaminates from the sample and therefore wear occurs.
For low normal loads and independent of counter body size we found that additional dislocation activity is happing below the DTL, while for higher normal loads there is significant change to the microstructure above the DTL. These experimental results were complemented by dislocation dynamics simulations performed by colleagues from Imperial College London.
In another stet of experiments, we made use of a preexisting lattice defect, namely a so-called twin boundary. By a detailed electron microscopy characterization of the sample after the tribological experiment, we were able to identify three distinct elementary mechanisms acting during the early stages of sliding. These processes are i) a rotation of the material above the DTL in sliding direction around an axis parallel to the transvers direction, ii) a simple shear of the material above the DTL in sliding direction and iii) a highly localized shear at the DTL, again in sliding direction. These results were highly unexpected and in our view are a major achievement.
For studying the effect that different alloying strategies might have on the deformation processes, we made use of Cu-Zn and Cu-Mn alloys. The brasses were chosen as here the effect of the so-called stacking fault energy could be studied. In these experiments we found a significant influence of the stacking fault energy on both friction and wear as well as on the microstructure that developed due to the sliding contact. Exemplarily it could be mentioned that for higher zinc concentrations, thus lower stacking fault energies, no DTL was found, but slip-band like features. In the case of the Cu-Mn alloys, a DTL was found for all manganese concentrations, which fits well our thinking as here the stacking fault energy does not change with alloying. However, there was a significant influence of solid solution hardening due to an increase in Mn concentration. This was also reflected in the friction and wear properties.
As mentioned above, tribological oxidation also is an important failure mechanism, as well as an interesting model experiment also for other tribologically induced chemical changes to a metal. We found that even under very mild sliding conditions, oxidation is about a factor of 500 faster for copper subjected to a tribological load, compared to native oxidation. Through a very detailed electron microscopy and a fruitful collaboration with Dr. Baptiste Gault (ERC-CoG-Shine) from the Max Planck Institute for Iron Research, where Atom Probe Tomography was performed on samples from us, we identified on an atomic level, that the rate dependent step is the diffusion of oxygen and metal ions along dislocations (dislocation pipe diffusion). This explains why for lower sliding speeds we saw an increase in oxide thickness; a result that is contrary to existing explanations for tribo-oxidation, but that we can convincingly explain.
In this research we also studied the tribological behavior of CuO and Cu2O as well as their evolution and chemical changes with sliding distance.
Finally, we worked extensively on setting up our non-destructive subsurface sensitive in situ tribometer. We here first screened potential experts in the field of non-destructive testing and after talking to a number of them, are now working together with colleagues from the Fraunhofer Institute for Non-Destructive Testing in Saarbrücken. Together with them we performed a number of preliminary experiments and identified that acoustic emission, ultrasound and optical light scattering are among the most promising candidates to probe our contacts non-destructively during sliding. It needs to be mentioned that these preliminary experiments have also shown that what we are intending to do is right at the edge of what is technologically possible. This is part of the reason why this part of the research has progressed slower than initially anticipated. We also realized that in order to be able to apply machine learning methods to this data – which is the long term goal – we need to standardize the meta information ascribed to every experiments. To this end we generated a “tribo wiki” where we defined standards how to describe a tribological experiment. In order to translate this knowledge, we in addition came up with an ontology of how to describe our tribological experiments.
In addition we built a new tribometer which is so precise that we were able to measure the very minute differences of different grain orientations. This in itself is a major achievement.
For pure copper we investigated the influence of different counter body sizes as well as normal loads. We saw that for all experimental conditions chosen, at the very beginning of sliding a dislocation self-organization phenomenon takes place, which leads to a horizontal dislocation feature about 100 to 400 nm under the sliding contact. The named this feature the “dislocation trace line”, or DTL. The DTL is the first discontinuity in the microstructure and therefore of key importance for any further changes to the materials. This is also true for the position where as sliding progresses material delaminates from the sample and therefore wear occurs.
For low normal loads and independent of counter body size we found that additional dislocation activity is happing below the DTL, while for higher normal loads there is significant change to the microstructure above the DTL. These experimental results were complemented by dislocation dynamics simulations performed by colleagues from Imperial College London.
In another stet of experiments, we made use of a preexisting lattice defect, namely a so-called twin boundary. By a detailed electron microscopy characterization of the sample after the tribological experiment, we were able to identify three distinct elementary mechanisms acting during the early stages of sliding. These processes are i) a rotation of the material above the DTL in sliding direction around an axis parallel to the transvers direction, ii) a simple shear of the material above the DTL in sliding direction and iii) a highly localized shear at the DTL, again in sliding direction. These results were highly unexpected and in our view are a major achievement.
For studying the effect that different alloying strategies might have on the deformation processes, we made use of Cu-Zn and Cu-Mn alloys. The brasses were chosen as here the effect of the so-called stacking fault energy could be studied. In these experiments we found a significant influence of the stacking fault energy on both friction and wear as well as on the microstructure that developed due to the sliding contact. Exemplarily it could be mentioned that for higher zinc concentrations, thus lower stacking fault energies, no DTL was found, but slip-band like features. In the case of the Cu-Mn alloys, a DTL was found for all manganese concentrations, which fits well our thinking as here the stacking fault energy does not change with alloying. However, there was a significant influence of solid solution hardening due to an increase in Mn concentration. This was also reflected in the friction and wear properties.
As mentioned above, tribological oxidation also is an important failure mechanism, as well as an interesting model experiment also for other tribologically induced chemical changes to a metal. We found that even under very mild sliding conditions, oxidation is about a factor of 500 faster for copper subjected to a tribological load, compared to native oxidation. Through a very detailed electron microscopy and a fruitful collaboration with Dr. Baptiste Gault (ERC-CoG-Shine) from the Max Planck Institute for Iron Research, where Atom Probe Tomography was performed on samples from us, we identified on an atomic level, that the rate dependent step is the diffusion of oxygen and metal ions along dislocations (dislocation pipe diffusion). This explains why for lower sliding speeds we saw an increase in oxide thickness; a result that is contrary to existing explanations for tribo-oxidation, but that we can convincingly explain.
In this research we also studied the tribological behavior of CuO and Cu2O as well as their evolution and chemical changes with sliding distance.
Finally, we worked extensively on setting up our non-destructive subsurface sensitive in situ tribometer. We here first screened potential experts in the field of non-destructive testing and after talking to a number of them, are now working together with colleagues from the Fraunhofer Institute for Non-Destructive Testing in Saarbrücken. Together with them we performed a number of preliminary experiments and identified that acoustic emission, ultrasound and optical light scattering are among the most promising candidates to probe our contacts non-destructively during sliding. It needs to be mentioned that these preliminary experiments have also shown that what we are intending to do is right at the edge of what is technologically possible. This is part of the reason why this part of the research has progressed slower than initially anticipated. We also realized that in order to be able to apply machine learning methods to this data – which is the long term goal – we need to standardize the meta information ascribed to every experiments. To this end we generated a “tribo wiki” where we defined standards how to describe a tribological experiment. In order to translate this knowledge, we in addition came up with an ontology of how to describe our tribological experiments.
In addition we built a new tribometer which is so precise that we were able to measure the very minute differences of different grain orientations. This in itself is a major achievement.
At the very end all the results described above go beyond the state of the art. At the same time we will choose three aspects that not only go beyond the state of the art, but are also unexpected.
The first is the use of a twin boundary as a “marker” for subsurface processes. We here were able to identify three separate, but congruent elementary mechanisms acting during the very early stages of deformation: A simple shear, a highly localized shear and a rotation of the lattice above a dislocation self-organization phenomenon. This is completely new knowledge and most definitely goes beyond the state of the art.
Second, we were able to build a new tribometer, which allows to test single and bicrystalline samples in different, highly defined sliding and therefore crystallographic directions, which allows us to measure the difference in friction forces as a function of crystallography. This was something which in the past was only possible in ultra-high vacuum conditions, and we are now able to do so in ambient and controlled atmosphere.
Third, classical common knowledge about tribologically enhanced oxidation is that due to sliding there is a temperature increase in the contact – concept of flash temperature – and this temperature increase is responsible for the accelerated oxidation. With very systematic experiments we could show that this classical picture is wrong. It is the diffusion of oxygen and metal ions along lattice defects that is rate controlling. We could show the decoration of for example dislocations with oxygen with near atomic scale resolution.
Until the end of the project, we anticipate that the non-destructive subsurface in situ tribometer will be built and first experiments will be conducted.
For the tribological experiments focusing on tribologically enhanced oxidation, we anticipate that we will gain a much deeper understanding of the diffusion pathways as well as their temporal sequence during different stages of a tribological experiment. This will be achieved through the ongoing collaboration with Dr. Gault from the Max Planck Institute for Iron Research.
We also anticipate to gain additional knowledge about the role of dislocation grain boundary interactions under a tribological load. Here we will make use of the concept of the so-called transmission coefficient and systematically test different large angle grain boundaries. We will also shed light on how the rotation of the material right below the contacting surface takes place and start to understand how the material accommodates this rotation.
We are in addition planning at the end of the action being able to connect our tribometers directly to an electronic lab notebook. In addition we are planning to expand the ontology to all our experiments in order to be able to fully describe our tribological research in a FAIR manner.
The first is the use of a twin boundary as a “marker” for subsurface processes. We here were able to identify three separate, but congruent elementary mechanisms acting during the very early stages of deformation: A simple shear, a highly localized shear and a rotation of the lattice above a dislocation self-organization phenomenon. This is completely new knowledge and most definitely goes beyond the state of the art.
Second, we were able to build a new tribometer, which allows to test single and bicrystalline samples in different, highly defined sliding and therefore crystallographic directions, which allows us to measure the difference in friction forces as a function of crystallography. This was something which in the past was only possible in ultra-high vacuum conditions, and we are now able to do so in ambient and controlled atmosphere.
Third, classical common knowledge about tribologically enhanced oxidation is that due to sliding there is a temperature increase in the contact – concept of flash temperature – and this temperature increase is responsible for the accelerated oxidation. With very systematic experiments we could show that this classical picture is wrong. It is the diffusion of oxygen and metal ions along lattice defects that is rate controlling. We could show the decoration of for example dislocations with oxygen with near atomic scale resolution.
Until the end of the project, we anticipate that the non-destructive subsurface in situ tribometer will be built and first experiments will be conducted.
For the tribological experiments focusing on tribologically enhanced oxidation, we anticipate that we will gain a much deeper understanding of the diffusion pathways as well as their temporal sequence during different stages of a tribological experiment. This will be achieved through the ongoing collaboration with Dr. Gault from the Max Planck Institute for Iron Research.
We also anticipate to gain additional knowledge about the role of dislocation grain boundary interactions under a tribological load. Here we will make use of the concept of the so-called transmission coefficient and systematically test different large angle grain boundaries. We will also shed light on how the rotation of the material right below the contacting surface takes place and start to understand how the material accommodates this rotation.
We are in addition planning at the end of the action being able to connect our tribometers directly to an electronic lab notebook. In addition we are planning to expand the ontology to all our experiments in order to be able to fully describe our tribological research in a FAIR manner.