Technology life cycle analysis based on patent co-classification. A dynamic network analysis of hydrogen-related technologies

We propose a method to analyze the technology life cycle from a network perspective, based on the co-classification of patent subclasses. By combining dynamic network analysis with network stability measures and community detection algorithms, our method enables users to analyze technological trajectories, identify technological discontinuities, and assess patterns of knowledge recombination and knowledge specialization. We illustrate the method with the case of hydrogen-related technologies. Our dataset comprises 668,109 patents, distributed across 475 subclasses, published between 1980 and 2021. Our analysis reveals that the first two decades were characterized by high levels of network instability and knowledge recombination, whereas the last two decades witnessed a shift toward knowledge specialization. We discuss our findings from the perspective of technology life cycle theory.