Robot accelerates the search for energy-storage materials
A new robotic platform automates electrochemical experiments and helps researchers gather information more efficiently about materials suitable for energy storage.
“We have developed new robotic tools that are not particularly complicated, but which allow us to speed up or even completely change the way research is conducted,” says Associate Professor Pekka Peljo.
The robotic platform developed by the researchers combines a liquid-handling robot, a consumer-grade 3D printer modified by the researchers, and an electrochemical measurement system. The liquid-handling robot prepares samples in a 96-well plate, while the electrochemistry robot, modified from a 3D printer, positions electrodes in the wells for measurements.
The number of possible combinations of materials is so vast that testing all combinations manually, one experiment at a time, is extremely slow.
The robot itself does not perform an individual measurement much faster than a human, but researchers can program the experiments in advance and leave the system to work independently around the clock. In addition, preparing samples with the liquid-handling robot is considerably faster.
Li Chuyue, doctoral researcherMeasurements that might take a researcher two weeks to carry out manually can be completed in about one day
Automation also makes it easier to study how different conditions affect a material. For example, researchers can systematically vary the pH level and monitor how it affects the material’s performance in electrochemical measurements.
“We can perform this kind of optimization more easily with the robot. We can also more easily examine how different parameters affect the response.”
The robotic platform also saves resources because the measurements can be carried out using considerably smaller sample volumes, which in turn produces less waste.
Why data matters
Researchers are interested in using machine learning to identify promising materials. However, this requires sufficient amounts of reliable experimental data with which to train the models.
“Machine learning is only as good as the data it has been trained on,” Peljo says.
If a model has been trained using only a limited range of materials, it will produce unreliable predictions when it encounters a material or condition that differs from those represented in the training data. This creates a practical problem: researchers need experimental data to build better predictive models, but producing large datasets experimentally can itself be a slow process.
“The question is how we can actually make these experiments faster so that we can obtain enough data to train AI and accelerate the process,” Peljo says.
Better materials for long-duration energy storage
One potential application is flow batteries. Unlike conventional batteries, flow batteries store energy in liquid electrolytes. They are a particularly promising solution for long-duration energy storage, in which electricity can be stored for several hours and used later.
“Vanadium is currently used as a material in many flow battery systems. It works well and enables a long service life, but its cost is a challenge. That is why we have been investigating materials that should be more affordable,” Peljo explains.
Vanadium is a silvery-white metal and an important element used, among other things, in the chemical and steel industries.
The purpose of the platform is to support this search by making it easier to test multiple materials and different conditions. However, the researchers are not yet claiming to have found a replacement material:
“As far as flow batteries are concerned, we do not really have any promising alternatives yet,” Peljo says.
The immediate result is therefore, above all, a new way of conducting experimental research.
The robotic platform is suitable for a wide range of electrochemical research
Although flow batteries are one potential application, the platform has been designed more broadly for electrochemical measurements involving liquids.
The researchers have also demonstrated the system using various chemical compounds and carried out systematic pH measurements. The platform automates sample handling, electrode positioning, measurements, and part of the data processing.
The hardware design and software have been published as open source. The aim is to make automated electrochemical research readily accessible to other laboratories as well.
“We are reaching the point where building a system like this is almost within everyone’s reach,” Peljo says.
The longer-term goal is to reduce the dependence of materials research on slow, manual experimental work: to automate more measurements, generate better datasets, and use them to improve our understanding and prediction of material behaviour.
The work is part of the Marie Skłodowska-Curie (MSCA) network, which includes 17 doctoral students from across Europe. The shared goal of the network is to accelerate the development of materials suitable for energy storage. At Aalto University, one of the research priorities is to speed up experimental work.
Further information:
Read more news
Everyday choices: Ghassem Gozaliasl, how do galaxy groups reflect the need for human connection?
An astrophysicist with the Aalto High-Performance Computing Lab (HPCLab), Gozaliasl uses the James Webb Telescope to study how galaxies have formed over the last 12 billion years.
"Translate Your Research": CVs & Cover Letters for the Job Market” Workshops in Autumn
Hands-on workshops for doctoral students and researchers on CV & Cover Letter for the Job Market
Master the Room: Networking for Researchers Workshops in Autumn
Hands-on workshops for doctoral students and researchers on building professional networking skills on 11.11. and 28.1 and 15.4.