Cyracle - Planning software for experiments

Your Virtual Guide for Materials Development
cyracle is a software for more efficient planning of experiments in material development. The software uses the latest machine learning algorithms
Many of today's industrial applications require chemical additives that fulfill specific tasks. In the case of anti-corrosion coatings, for example, small organic molecules block the surface from corrosive species or prevent the surface from dissolving. Since the REACH agreement, sustainable and environmentally friendly alternatives to chromate-based anti-corrosion coatings have been in demand. However, the chemical space is in principle infinite (around 1063 theoretically synthesizable compounds). The huge number of components to be tested also has an impact on other areas, such as the development of advanced battery systems, where chemicals are often used as additives to the battery electrolyte to increase performance.
This raises a fundamental challenge: how to find a suitable compound for the task at hand without testing all of them over literally thousands of years?
Experiments are time-consuming and resource-intensive for R&D departments in industry and research institutions, so the number of compounds to be tested and other experimental parameters to find a suitable material solution should be as low as possible. For industrial users in particular, this is crucial in order to work cost-efficiently and sustainably and to remain competitive. The planning and documentation of experiments in development is still mostly based on lengthy and cost-inefficient trial-and-error approaches, experience and intuition, which wastes the potential of digital tools in this regard. The results are still mostly organized and analyzed in Microsoft Excel and Origin, which makes it difficult to make full use of the available data in terms of linking different data sources.
Spin-off project at Hereon

Some platform solutions for experimental design already enable the use of existing data to plan experiments in material development and to identify new materials more efficiently and with high flexibility, but such solutions leave you alone with the problem and come with steep learning curves and quite high investments in terms of time and money, which prevents wide adoption. In addition, most solutions originate from companies outside Europe, leading to potential conflicts in terms of data regulations. This is particularly unfortunate as the current challenges of our time, such as the climate crisis, require novel material solutions, fast development processes and short time-to-market. We simply do not have the time and resources to search for the best solution in a huge haystack of possibilities. Therefore, digital tools that facilitate efficient trial planning are crucial to reduce the total number of trials and find the needle in the haystack faster - now more than ever.
The machine learning models developed at the Institute of Surface Science (MO) have already proven valuable in the search for new corrosion inhibitors or battery electrolyte additives. However, fine-tuning and using these models in everyday laboratory work is still tedious, as knowledge of Python is required and customization is not always easy. This gave rise to the idea of developing software that makes these models easy to use. In the long term, such software should be applicable for all possible scenarios in the selection of chemical compounds and lay the foundation for a spin-off.
The need for such software was already validated in early 2024 in the Field Study Fellowship as part of Helmholtz Enterprise (€30,000 funding). Based on the insights gained from the scientific and industrial sector, a working software prototype is to be developed by June 2025, which will be tested in collaboration with Sviatlana Lamaka and Darya Snihirova from the MOD department.
Internal funding of €37,000 from the Innovation Fund was applied for by the Department of Innovation and Transfer for the development of the prototype. Angelika Eichenlaub offers further support in all topics related to the journey as a sciencepreneur. Scientific support is provided by Daniel Höche, Christian Feiler and Mikhail Zheludkevich from MO.
By providing easy-to-use software that is customized to the user's needs, the prototype is intended to simplify test planning in the search for new material solutions. Instead of offering a multitude of functions and options that may not even be needed, the software is customized in terms of user needs and domain-specific features. The combination of domain knowledge and ease of use is the basis for a strong USP. The next step is to apply for the Helmholtz Enterprise spin-off program in order to continue working towards the planned spin-off.
Kontakt
