ºÚÁϸ£ÀûÍø

SEPTEMBER 24, 2025

Formulating new theories and conducting experiments are not the only pathways to discovery. A third approach is computational research. Within EIRES, the Computational Energy Research (CER) focus area is centered around using advanced computational methods and artificial intelligence to find and optimize materials and processes for energy applications.

‘Through computational research, we are accelerating the discovery of materials and molecules for energy applications,’ say focus area leaders Süleyman Er, department head of Chemical Energy at DIFFER, and Peter Bobbert, full professor at ºÚÁϸ£ÀûÍø’s Department of Applied Physics and Science Education.

Er: ‘We work on systems ranging from the atomic level all the way up to the level of millions of interacting atoms, which is directly relevant to actual experiments. Our work is a complementary piece of the research puzzle, acting as a glue keeping the other pieces together. Computational research can help predict the behavior of promising molecules, materials, or processes.’ Bobbert adds: ‘As computational researchers, we need scientists from the other EIRES focus areas to alert us on interesting research directions’. That is why CER is positioned as a capability that cross-cuts the other four focus areas of the institute.

Fast route
Whether it is about the ideal composition of a solar cell, the chemical reactions one might expect to occur at the interface between an electrolyte and a catalyst, or the heat capacity of a certain phase-changing material for a heat battery: advanced simulations are an indispensable tool to reach answers to pressing questions fast, both computational scientists explain. Bobbert: ‘With our models and methods we can work on problems that cannot be solved by pen and paper alone, simulating a material’s behavior in a way that closely resembles what happens in reality.’ Er:’ And since we can solve multiple problems in parallel using highly automated calculation workflows, we can reach an answer way quicker than would be possible through any lab experiments.’

Speeding up
These are interesting times for the computational sciences. Er: ‘The rise of artificial intelligence (AI) is helping us to expand our toolkit immensely. Ultimately, we want to understand why a material behaves in a certain way to find or develop new materials and properties. AI enables us to dramatically accelerate our computational studies so that we can swiftly identify promising material candidates across multiple research directions simultaneously. Take the example of looking for a material that can act as a catalyst for a certain chemical process. Previously, we would start to review the existing literature, insert some intuition and previous findings, and then use computations to put our initial ideas to the test. With AI, we can investigate a multitude of interesting candidates in a short time.’

And we are at the brink of an even greater breakthrough, Bobbert adds. ‘At the moment, AI systems are particularly good at interpolating between existing data. What they cannot do well enough yet is extrapolate. However, we see the first possibilities of such generative AI emerging, with systems that can generate art for example. In our own field of research, an important step has been made by introducing on-the-fly machine learning of force fields, where we feed a neural network with examples of quantum-mechanically calculated energies, and then it gradually improves the force field until no more quantum calculations are required. By combining data, models, and AI in this field, we can speed up our calculations tens of thousands of times, opening up an entirely new realm of possibilities.’

Broadening the base
Now that the community of computational energy researchers has officially joined EIRES, Bobbert and Er hope to unlock a lot of new potential. Bobbert: ‘Our community here in Eindhoven is rather young. DIFFER and ºÚÁϸ£ÀûÍø brought together their computational energy research in CCER in 2017 (see box) and COVID-19 struck just a few years later, making it hard for people to meet each other and establish new connections. Fortunately, especially through our open, biweekly seminars, we have managed to build a community of people taking a computational approach to energy-related topics. Now, we hope to take the next step and connect our community to the wider group of energy researchers here in Eindhoven.’

Er: ‘Through our involvement in EIRES, we obtain a clearer overview of complementary research activities, which enables us to realize targeted and fruitful collaborations with theorists and experimentalists.’ Bobbert: ‘There is a lot of underused potential when it comes to integrating computational approaches in energy research projects. That is why, for example, we are talking to people from the Greening the Process Industry focus area to explore the possibilities of applying on-the-fly machine learning to gain a better grip on the burning of iron. The Wax+ project is a good example of the collaborative approach we are striving for. In that project, we are developing materials that combine a high heat capacity with a high heat conductivity for heat storage applications.’

The focus on applications is what makes the Eindhoven approach to computational energy research so unique, states Er. ‘Our diversified computational approaches directly support the strategic vision of EIRES by rapidly identifying and developing molecules and materials that are critical for accelerating the energy transition. With our proven tools to identify promising molecules, materials, and systems, we can help both researchers and industry to obtain results far faster.’ Bobbert ends with a heartfelt appeal: ‘If you are working in the field of energy technology in any way, think about how we can help you. We’re actively looking for opportunities to integrate our computational expertise into practical applications and strengthen partnerships across academia and industry.’

The researchers develop and use a variety of models and techniques to accelerate the discovery of molecules and materials for energy applications. These include density-functional theory, molecular dynamics, dissipative particle dynamics, phase-field modeling, lattice-Boltzmann, magnetohydrodynamics, computational fluid dynamics, continuum mechanics, moment methods, coarse-graining, machine learning, and reduced-order modeling.

In 2017, ºÚÁϸ£ÀûÍø and DIFFER decided to join forces in the field of computational energy research. They launched their joint Center for Computational Energy Research (CCER) with the aim of exploring pathways to future energy systems, such as solar cells, flow, ion and heat batteries for energy storage, catalysts for energy conversion, electrochemical interfaces, energy conversion materials, porous materials for gas storage, and plasma-assisted chemical conversion.

In 2024, CCER was officially integrated into EIRES as the institute’s fifth cross-cutting focus area, Computational Energy Research.