Fem: New PVD process shortens thin-film development time

The development of new coatings is a critical and risky process for many companies. This is because they are still created in complex individual steps that require a lot of time, materials and costs, with no guarantee of an optimal result. At the same time, ever shorter innovation cycles mean that more efficient and sustainable thin films must be available ever more quickly. The combinatorial PVD process for developing new thin films has therefore been methodically expanded at the Fem Forschungsinstitut, enabling systematic and significantly more efficient material development. "We are transferring layer development from a complex trial-and-error principle to systematic screening. This allows complex material systems to be investigated in a fraction of the time previously required," explains Dr. Martin Fenker, Head of the Plasma Surface Technology Department at the Fem Research Institute.

Data-driven layer development with AI support

Combinatorial magnetron sputtering (CMS), in which several magnetron sputtering sources are operated simultaneously, is at the heart of the project. The targeted arrangement of the sources creates defined lateral composition gradients on a substrate. This principle transforms a single sample into a material library: instead of isolated individual tests, a large number of different material variants are created in a single process run, which can be systematically analyzed. Properties such as hardness, corrosion behavior or optical parameters can be directly assigned to the respective composition along these gradients. In contrast to conventional development, which is based on an iterative approach, this approach enables a targeted and structured screening of material systems. The experimental effort with regard to layer depositions is reduced by a factor of at least 10 to 100, while at the same time reliable correlations between composition, phase formation and functional properties are obtained. The spatially resolved characterization - for example via CIE-Lab colour measurements, instrumented hardness tests or corrosion tests - generates structured data sets. These form the basis for data-driven material development and the use of material informatics and artificial intelligence methods. The efficiency of the process has been demonstrated in several projects. In the EU project "Colored Gold", combinatorial co-sputtering was used to specifically generate different color states in metallic layers. Development cycles for the layer depositions were reduced from many years to just a few months.

Faster industrial application of new coatings

In the IGF project "RefMagS", the method was transferred to complex metal nitride systems. Here it was shown that mechanical and chemical properties can be specifically adjusted via the magnesium content. At the same time, systematic correlations between composition, microstructure and performance were identified. For industrial practice, this means a clear competitive advantage: companies can evaluate material systems more quickly, reduce development risks and significantly shorten their time-to-market. "A single coating run replaces numerous individual tests. This significantly reduces costs and at the same time increases the hit rate when developing new coating systems," says Fenker. This approach is particularly relevant for small and medium-sized companies. Without an extensive R&D infrastructure of their own, they can use combinatorial layer development as an upstream screening platform to identify suitable material solutions before they are transferred to series-relevant processes. Major bad investments can thus be avoided. The combinatorial approach is flexibly scalable and can be transferred to numerous applications - from wear and corrosion-resistant protective coatings and decorative surfaces to optical and functional coatings as well as applications in energy and medical technology. For medium-sized companies, the approach thus opens up significantly more efficient access to new materials and considerably shortens the path from idea to industrial application.