The diffusion of technological knowledge through interlaced networks

The study of complex networks and the mechanisms underlying their formation gave an important direction to the research on the diffusion of information. It has been suggested that the network created by scientific citations follow a well-established picture of a scale-free network. Thus, citations should obey the basic mechanisms of preferential attachment, as in other examples of social networks. Technological development may be represented by the deposit of patents (USPTO), and the citation of patents by patents also creates a network. In this paper, we show that a network formed by the spreading of technological information, represented by the patent citing patents network, does not obey the expected power law behavior and, therefore, is not a scale-free network. Using a scientific paper as a starting point, we traced back the first patent on nuclear magnetic resonance. Following all the patents which cited this first one, and those which cite them and so on, we mapped a network formed by almost 600,000 patents, covering a 40 year period. Although the complete network is not scale-free, certain small portions within the whole network can be described by power laws. The complete network combines several fields of knowledge, but those power law portions belong to specific fields of technological knowledge. We introduce a mathematical model which explains the basic formation mechanism, showing that the maximum efficiency of diffusion occurs for branches which belong