Jump directly to the page contents

Battery research

A Revolution with AI and Robotics

The use of artificial intelligence can accelerate the development of new batteries.Image: Messling, KIT

Helmholtz scientists want to make batteries more powerful, more environmentally friendly and cheaper. They are also using artificial intelligence, robots and supercomputers to search for new materials.

Electric cars have not yet been able to establish themselves in Germany.Image: Mikes photos. Pixabay

According to a study by the Federal Environment Agency (UBA), electric vehicles are the most cost-effective option for making the transport sector greenhouse gas neutral. But the German government has clearly missed its original target of one million electric vehicles on Germany's roads by 2020. According to the UBA, a total of 168,594 new electric cars were registered between 2008 and 2019, 63,281 of which were registered in 2019.

There are many reasons why e-vehicles have not been able to establish themselves so far: The high acquisition costs, the poorly developed charging infrastructure - and the incomplete market maturity. The batteries are a sticking point. Because meanwhile, the lithium-ion batteries commonly used in electric cars, smartphones and laptops can hardly be improved. Researchers are therefore looking for new technologies for powerful, environmentally friendly and inexpensive batteries.

AI and robots accelerate battery development

This research could soon receive a considerable boost: If scientists from the BIG-MAP (Battery Interface Genome - Materials Acceleration Platform) project have their way, new battery materials could be developed ten times faster in the future. To revolutionize battery research, they are using artificial intelligence (AI) and relying on consistent automation. "We feed AI with data on interfaces and materials of all kinds of batteries, whether based on lithium, sodium or other charge carriers," explains Maximilian Fichtner, Deputy Director of the Helmholtz Institute Ulm and spokesman of the CELEST research platform. The Karlsruhe Institute of Technology (KIT) is involved in both institutions and co-founded them together with partners. "The AI uses this input to understand the chemical behavior of the systems; it draws conclusions from it and recognizes complex relationships." Connected to this is autonomous robotics: The robots build novel battery components (prototypes) on the instructions of the AI, which are then tested by the researchers. Failures are also taken into account, from which the AI in turn learns.

Sketch of the "robotic road" from the BIG-MAP project. The robotics line is developed and built by Helge Stein's research group at the Helmholtz Institute Ulm (HIU). Visualization: KIT

"The robotics will be set up on a twelve-meter long and two-meter wide surface, with additional apparatus for synthesis and deeper characterization on the side," explains Helge Stein, who develops and implements the robotics line at the Helmholtz Institute Ulm together with his research group. The facility is scheduled to go into operation in 2021. "Because of the many variables, the battery models developed by the AI will be so complex that humans will hardly be able to grasp them," says Fichtner. "Our - so far still distant - goal is to tell the AI which properties we need and then to obtain a precisely fitting system.

Within the framework of BIG-MAP, a common European data infrastructure is to be created. This will enable researchers to work together across national borders and time zones. 34 institutions from 15 countries are involved. It is the largest single research project of the European research initiative for batteries BATTERY 2030+ - and important for international competition. Because only by joining forces can Europe remain competitive in battery research, Maximilian Fichtner knows: "A realignment of the existing discovery, development and manufacturing processes for battery technologies is necessary to enable Europe to compete with its main rivals in the USA and Asia.

By loading and unloading, the structure of crystalline silicon electrodes changes into a chessboard-like fracture pattern.Image: HZB

On the way to inexpensive sodium batteries

Researchers at the Helmholtz-Zentrum Dresden-Rossendorf (HZDR) have also taken an important step forward on the way to lithium alternatives. A German-Russian working group there has produced a study that shows, at least in theory, how a sodium battery could work. Replacing rare and therefore expensive lithium with the frequently occurring sodium would make the production of batteries much cheaper. The catch is that the graphite anode of the battery does not absorb enough sodium.

The study suggests that double layers of graphene - wafer-thin carbon - could store significantly more sodium atoms than in graphite. This would increase the storage capacity, as supercomputer simulations by the international research team show. "Thanks to the immense growth in computing power and the development of efficient algorithms, we now have very powerful methods at hand," explains Arkady Krasheninnikov, physicist at the HZDR. "They make it possible to predict detailed material structures and properties."

High-performance batteries with silicon

However, the aim is not only to replace lithium in batteries, but also to improve the widely used lithium-ion batteries with new materials. Silicon is the ideal material for this: As a material for electrodes in lithium-ion batteries, it promises a significant increase in capacity. In addition, the element is available in almost inexhaustible quantities in the earth's crust. But so far, the silicon has been slightly damaged by exposure to lithium during loading and unloading. Scientists at the Helmholtz-Zentrum Berlin für Materialien und Energie (HZB) have now been able to decipher how fractures occur in silicon - and how the material can still be used advantageously. "We made flashlight-like measurements, similar to a stroboscope," explains Sebastian Risse, who works on the analysis of storage materials at HZB. "The data records the complete period of time during which electricity flows through the battery, so that we now have a real film of what happened."

Risse and his team were able to show that during charging and discharging a chessboard-like fracture pattern develops and disappears again. "Although the fractures become slightly larger each time the battery is discharged, the pattern remains the same and no new fractures occur." This means: If the capacity of a silicon battery is reduced accordingly, its service life is extended many times over. Theoretically, silicon has eleven times the capacity of graphite, but in practice this advantage is reduced to three times. Silicon takes over the function of graphite as the anode, i.e. the negative pole of a battery, as in the case of a sodium battery. However, these significantly more powerful batteries would have to be replaced more often than conventional lithium-ion batteries. "Silicon, as the second most common element after oxygen, would be a cost-effective material," stresses Risse. "Moreover, the microparticles do not represent a burden on the environment when they are disposed of".

Cell phones would then have to be plugged in less often, and electric cars could cover longer distances with one battery charge. A drone manufacturer is also already showing interest in the high-performance batteries for future parcel delivery. The scientific foundation has now been laid.

New EU Project to Boost Battery Development

Hope for better batteries 

Readers comments

As curious as we are? Discover more.