Building materials of the future in computer models

Electronic components radiate heat – a growing problems as they become ever smaller. But some materials can convert heat into electricity. Wallenberg Academy Fellow Paul Erhart at Chalmers University of Technology is using computer models to study how the heat can be reduced or converted into an asset.

Dr Paul Erhart

Department of Materials and Surface Theory, Chalmers University of Technology

Wallenberg Academy Fellow 2014

Chalmers University of Technology

Research field:
Materials technology, computer modeling, electrical and thermal conductivity

Large electronic appliances, such as personal computers, have fans or surfaces designed to conduct the heat away. Without them, systems would rapidly overheat. But nowadays many electronic circuits and components are so small that there is no room for cooling systems. One solution may be to design the material itself so less heat is generated. This is a practical issue, but the path to the answer is full of theory. Paul takes a deep breath before he starts to explain.

“Heat is really vibrating atoms. But thermal transport can be seen as a group of particles carrying heat. We call these particles phonons.”

Phonons are quasiparticles. A quasiparticle is not represented by a “real” particle, but by something that acts like one – at any rate sufficiently so that theories and models can be based on it.

“Phonons can be disturbed at each interface in the material. If we can gain a truly accurate picture of how thermal transport works, we will be able to design materials that block it, or that convert it into some other form of energy transport – namely electricity,” Paul says.

This would make the electronic equipment of the future much more energy efficient.

Transport impacted by both chemistry and physics

The way heat and electricity move in a material is influenced by many factors. One is chemical composition. Another is physical structure, all the way down to atomic level. In most of the materials Paul is studying, the atoms form crystals, i.e. regular patterns. But within the crystals, additional structures may be formed, in which parts of the material gather to form grains or lumps. This also impacts the transport of heat and electric current.

Paul smiles wryly, pointing out that he began his career as a material scientist, not a physicist, which is reflected in how he approaches problems. He explains that theoretical physicists are not particularly keen on working with defective materials – it disturbs their equations. But to a material scientist, it is the defects that are the most interesting thing. Deviations from the norm often produce the properties that are desired.

“It’s like when you dope silicon by introducing impurities into it, which changes its behavior. Without this ‘defect’, we wouldn’t have any transistors. We must understand how defects in a material work, since they give us the ability to control and disperse the heat-carrying phonons.”

Computer models of the future making their own suggestions

Analyzing and designing a new material at atomic level in the lab is very difficult and time-consuming. Paul is using computer models instead. These enable him to work out how a material should be designed to exhibit the desired properties. This is known as computational design. Nowadays it is possible to feed data about a potential material into the computer, which will then produce information about the material’s properties. But Paul hopes it will soon be possible to do the opposite – to input desired thermal and electric conductivity, for instance, and obtain a suggestion as to how the material should be designed.

“We can’t do it yet, but it’s just over the horizon.”

“I didn’t expect to be admitted as a Wallenberg Academy Fellow. It will be a huge help to me. It represents recognition of my research, and raises my profile in the research world. It may also improve my prospects of receiving other funding in the future.”

Revelations beat everything

Paul does not perform any practical experiments. He thinks that everyone should do what they are best at, and in his case that is computer modeling. He has always used computers. When he started studying material science he was not really satisfied; it did not seem to be precise enough, and he turned to physics instead. But he changed track again after spending a year working as an exchange program research assistant in Illinois alongside his studies. His boss sent him on to a summer graduate school in material science at Lawrence Livermore National Laboratory in California.

“That was what did it for me. It was then I realized what I wanted to do, and even before I had completed my master’s degree I had ideas about where I wanted to do my postdoc.”

He had heard about Chalmers during his early student days. He says the university has a fantastic history in the computation of electronic structures. When he attends conferences everyone knows about Chalmers. When asked what he likes most about his research, Paul replies “revelations” – the moment it suddenly becomes quite clear what the problem is – or what the solution is. He also enjoys lecturing.

“I like telling good stories. But writing a really good scientific article – now that’s a struggle. If it’s not hard work, it won’t be any good. You have to be something of a masochist…”

Text Lisa Kirsebom
Translation Maxwell Arding
Photo Magnus Bergström