The History of Education in Computational Thermodynamics – the KTH Experience

Computational thermodynamics is a rapidly developing field at the forefront of materials design. But did you know that the field is already over 40 years old? This year at the TMS Annual Meeting in San Antonio, Texas, John Ågren, one of the original developers of Thermo-Calc, gave a presentation on the history of computational thermodynamics at the Royal Institute of Technology (KTH) in Stockholm, Sweden, one of the earliest schools to teach computational thermodynamics. In his fun and fascinating presentation, he discusses how the education of computational thermodynamics started, which issues arose and how they were solved.

A Nord 10 16-bit mini computer
A Nord-10 mini computer similar to the once used by the Materials Science and Engineering (MSE) department at KTH in the 1970s.

It all started in the late 1970s when the Materials Science and Engineering (MSE) department at KTH got a minicomputer which was used in both research and education. But in the early days they came across some unexpected issues. It appeared that most students disliked computers and couldn’t do much coding themselves, which resulted in teachers spending a lot of time debugging codes. Besides that, there were also technical issues. About 30 students were working on the same computer at the same time, which made the response times very long. These difficulties made it hard for the students to understand the point of the computational exercises. As an attempt to solve these problems, teachers prepared codes for the students and handed out some written material about the underlying physics. Despite this, most time was spent on making correct inputs, which made the students not likely to understand the role of computers in materials science.

It started to change in the 1980s when computers became more powerful and the newly developed Thermo-Calc allowed more flexible calculation set-ups. It was still difficult to use, but after learning the basics of the interface, it was possible for the students to define their own problems. This made it easier for the students to understand what they were doing. It was also at this time that industrial firms wanted to be able to perform calculations in-house. This increased the demand for a new type of education that focused on calculating various types of phase diagrams and how to handle calculational difficulties.

Meanwhile at the university, minicomputers became more powerful and more common. The university started investing in new technology, such as a common terminal network. Students and teachers started to appreciate the new technology and some laboratory classes were replaced with computer exercises. This was also the time when the DICTRA code, which is now called the Diffusion module (DICTRA), was developed and introduced in education.

Traditional education of thermodynamics and kinetics was given to the students during the first and second year. However, the teachers had no experience of the new CALPHAD technology. This kind of education was at the time of no value for the students because it couldn’t serve as a basis for the emerging field of computational thermodynamics. In fact, it did more harm than good. The conclusion of this was that the education was in need of revision.

The materials education at KTH had great support for revision, but there was no common vision of how. Outside the field of materials, a revision usually meant getting rid of metallurgy and replacing it with more chemistry and physics. The problem was that in this case it was the education of chemistry and physics that was in need of revision.

The first course in materials design was launched by Mikael Lindholm at KTH in 1995. The format was simple: a couple of lectures, a project defined by the industry and examination in the form of a written report with an oral presentation. This course quickly became the most popular course in materials education at KTH.

John Ågren writing on a chalkboard
John Ågren teaching a course in Thermo-Calc and DICTRA in 1994 at Northwestern University in connection with the CALPHAD conference in Madison that year.

The technology continued to develop, PCs arrived to the market and the terminals disappeared. In 1997 Thermo-Calc was founded as a commercial company and education for professionals in the industry and PhD students was taken over by Thermo-Calc. Today, courses are offered all around the world. The first industrial course in Thermo-Calc was held for a period of one week. Nowadays, the courses are also held for one week but include Thermo-Calc, the Diffusion module (DICTRA) and the Precipitation module (TC-PRISMA). Thanks to the graphical interface of Thermo-Calc, students are able to perform advanced thermodynamic and diffusion simulations with no prior experience of the software and with less than one hour of introduction.

Around the year 2000, the MSE department at KTH started to build up competence in Ab-initio and density functional theory (DFT) calculations. In the beginning there was a competition between DFT and CALPHAD but it was soon realized that both sides needed each other. In 2003, KTH changed the name of its materials education to “Materials Design” due to another reorganization. The Hero-m center was launched in 2007 with the initial aim to develop tools for materials design in ICME. A course in DFT named Quantum Metallurgy, was developed and DFT is now an important part in the education of materials design.

Today, the software is easy to use, which means that the students can focus on the true engineering and scientific issues. However, Prof. Ågren states that it is important that students learn to master many computational tools such as CALPHAD, finite element method (FEM), DFT and Ab-initio, and that they also should learn how they are connected. He also points out that the students should understand what type of engineering problems each tool is useful for without knowing all scientific details.

Prof. Ågren also issues a warning for teachers and students. The increased usability of computational tools has led to a decrease in experimental activities for the students. Therefore, it is important that the students understand the importance of experimental work and that computations can never completely replace it.

See the presentation slides »