Storing thermal energy in the home over long periods of time is important to optimize energy consumption and reduce heating costs. Porous materials such as zeolites and metal-organic frameworks (MOFs) can serve as efficient storage media.

Zeolites and MOFs are special materials with many small holes and a large surface area - the technical term is: porous crystalline materials. They can absorb and store a large amount of water or other substances. This produces heat that can be used for heating in winter. In summer, the sun's energy can be used to heat the materials and release the stored water.

Simulations make things faster and better

To achieve optimal heat absorption and dissipation, suitable zeolites or MOFs must be selected. Computer simulations can help predict the behavior of these materials and select the most promising candidates for fabrication in the laboratory.

However, the selection of materials must also take into account the conditions of the storage and retrieval process, such as the desired temperatures. Other aspects such as the stability and lifetime of the materials must also be included in the planning at an early stage.

Quite a lot of factors and quite tricky!

Give it a try on our slot machine. Can you beat the computer simulation?

Probably not, because without computer simulations it would not be possible to solve this complex optimization problem. You can divide the simulation into different subtasks: E.g., the behavior of the materials can be calculated in advance. The process simulation then uses this data. Simultaneous improvement of material and process is also possible, but requires very efficient and accurate methods. We are driving the development of such methods in the Cluster of Excellence "Data-integrated Simulation Science".

The use of computer-aided methods and simulations allows virtual design of materials with specific properties, e.g. for use in batteries or for energy storage. In contrast to the traditional empirical approach, in which materials are produced and tested experimentally, virtual material design can accelerate the development process.

This requires a cross-scale understanding of the physical processes. Cross-scale means analysis at multiple length and time scales: from the atomic level to the microstructure level to macroscopic dimensions. The different levels provide different information that must be combined in an appropriate way. In the following, we present two examples of data-integrated and cross-scale design of materials.

Example 1: A molecular approach to adsorbent optimization.

When selecting materials for adsorbents in adsorption heat pumps, the adsorption behavior plays an important role. The adsorbent can accumulate gaseous or liquid substances on its surface and thereby release heat.

The heat and the amount of substance that can be adsorbed depend on the molecular interactions between the solid and the fluid. Molecular simulations can be used to calculate adsorption isotherms that describe the relationship between the amount of substance adsorbed and the external pressure.

Due to the computational intensity, so-called surrogate models are developed, which use a simplified description of the fluid by a thermodynamic equation of state. The classical density functional theory allows a more efficient calculation of the adsorption isotherms. This allows the solid material to be optimized by taking into account various degrees of freedom, such as geometric or chemical properties, with respect to the with respect to the desired heat of adsorption. Current research questions deal with the development of such surrogate models and the efficient numerical solution of the mathematical equations.

Example 2: Data-integrated multi-scale simulation of lithium batteries

To improve the operational safety and performance of lithium-based batteries, intensive research is being conducted into so-called all-solid-state batteries. Unlike conventional batteries, these cells do not contain liquid electrolyte, which extremely reduces fire hazards. Instead, lithium ions move through a crystalline structure during charging and discharging. Differences in diffusion within the crystals and at the boundaries between the crystals, which are only µm (1 µm = 0.001 mm) in size, affect the charging and discharging behavior. Research is focused on lithium transport at the grain boundaries. This involves complex and time-consuming simulations at the atomistic level, the results of which are used in machine learning to investigate larger atomic configurations and time steps. With the limited available data from the atomistic/molecular level, data-integrated material models are then developed. At the same time, special modeling will be performed to include lithium transport along grain boundaries. These models will eventually allow to feed cell-level simulations with sound material models.

The information and data are continuously exchanged between cell and crystal level in the project to improve the quality of the predictions. This is done on all length and time scales.