±¬ÁÏTV

Using ThermoCalc’s Diffusion Module (DICTRA), work was undertaken to see if the simulations could accurately predict the interface between automotive steel and a protective coating during various heat treatments. The work aimed to improve upon the basic linear and step functions offered by the software to improve the starting compositional profiles of elements to better follow the in-situ diffusion of individual elements. Two coatings were investigated, one being a mainstay of the coating industry, Al-10Si. Results showed an improvement in the simulation to follow the diffusion pathway more closely when using a Heaviside function-based profile and laid the groundwork to predict the diffusion zones of experimental coatings more accurately.

The ThermoCalc program is a useful tool for simulating the thermodynamic properties of metal systems. Its built-in diffusion program (DICTRA) offers the ability to create a compositional profile for elements in three main ways: linear, step and function-based. The first two methods allow quick profiles to be set up for simplistic compositional profiles; however, they can fall short when looking at more complex starting points. For example, with a step profile, elements can make only one change in composition at a single point. This means gradients and fluctuations cannot be modelled accurately. This causes an issue when considering the fact that the interface of a coated material is not a perfect boundary. During the coating process, minor levels of diffusion occur as the coating cools, with the cooling profile being dependent on the coating technique used. With the limitations of using a step profile, this initial diffusion cooling, as well as compositional fluctuations throughout the coating, have to be approximated. 

ThermoCalc allows the user to formulate any function with respect to the distance, such as the Heaviside step-function. The use of Heaviside functions allows the combination of multiple step functions into a single expression; these functions can also incorporate linear and logarithmic gradients into the expression, which allows compositional fluctuations to be mimicked more accurately. By incorporating compositional profiles based on these functions, starting points of simulations will be able to better portray the initial interface between the coating and substrate. 
 

I would like to thank Ciaran for creating phase diagrams and Dictra simulations. This really helped me in understanding the diffusion process and gave me an idea of the phases that were formed. Thanks to his work, I’m able to link my experimental results from EDX-Linescans and -Mappings to theoretical simulations. I will use the results further for EBSD and XRD investigations and my PHD thesis.

Patrick Veyel MSc

Group Innovation, Sustainability Solutions, High Performance Metals (K-GERS/M), Volkswagen Aktiengesellschaft
 

As mentioned, the use of Heaviside functions allows a much more accurate depiction of compositional profiles compared to a combination of the in-built linear function. An example of this is seen in the provided linescan and the corresponding compositional profiles. The linescan depicts the changes in the composition of a tantalum, vanadium, titanium, iron and tungsten (TaVTiFeW) alloy and its inclusions, carbon, oxygen and nitrogen (C, O, N). As seen in the contrasting compositional profiles, the use of the functions has allowed the profile to follow the linescan more accurately, notably the spike caused by the different phase and fluctuations in the latter half of the scan as the material returns to the bulk.
 

By having these points of interest incorporated before the running of simulations, the results will be able to better represent the changes to the material and not be put down to potential anomalies. The plan is that by utilising the functions to better capture the starting composition, simulations can be run on potential coatings to capture the interface during the coating application, as well as future heat-treatments. Being able to model the composition prior to a certain process will allow the results to show the extent of the diffusion. This will help to pick up better the back diffusion of elements as well as the potential formation of intermetallic phases via spikes in elements of interest. In doing so, it may provide an alternative a quicker and more accurate computational alternative to testing multiple coatings/conditions without the need to manufacture them.

Equipment used

• ThermoCalc-2024a
• ThermoCalc-2025a 
• Dictra add-on module 

Biography

Ciaran Miles is a final year EngD student in the Department of Materials Science and Engineering, with his work focusing on coatings, in conjunction with Volkswagen Innovation Group. Ciaran has just come back from a two-month placement with Volkswagen, where he used the Heaviside functions to aid in simulations of novel coatings. Prior to his doctorate, he obtained a Master’s degree in Chemistry, including a year studying at the Australian National University.

Outside of his research, Ciaran is heavily active in sports, coaching a local kids’ American football team, having played a combination of it and rugby throughout most of his life. When not coaching, he enjoys reading, baking and most notably working on his shared project cars.