Page 25 -

networks of the highest cited inventors in each technology-bounded dataset and
compared these networks to the same inventors’ "second-degree" networks extracted
from the entire USPTO dataset (Figure 7). In this way, we can gain a first approximation
of the full extent of social influence of top inventors (Figure 5). Comparing the networks
of the largest components and most highly cited inventors across technology areas
suggest differences in career connectivity of inventors in the different technology areas
(Figure 7), hinting at fruitful avenues for future research. Furthermore, we did a
preliminary comparison of the energy inventive space with that of software and
pharmaceuticals and observe massive differences in the collaboration networks. Here we
focus on energy patents so this thread of inquiry is not continued, but there are lessons to
be learned from these highly collaborative and innovative industries, which will be the
focus of future analysis.

Certainly, the historical sources of innovation, the industrial roots, and the resource
intensity of these technology areas affect these collaboration norms. Validating these
disparate network connectivity characteristics and comparing them to other industries
will require more extensive patent datasets, but we suspect that there are indeed
substantive differences between the technology fields and important lessons to be
learned.

3.2 Sources of Clean Energy Technical and Commercial Breakthroughs
The count-based methods used in this first analytical effort do not account for patent
quality, which misses most of the impact of patents, as the distribution of commercial
value is extremely skewed [68]. It is very difficult to predictively quantify the impact of
these patents, and this quandary is almost universal for bibliometric scholars of research
and innovation. While recent databases of patents and papers [42, 69] have greatly
facilitated large-scale analyses, they stop short of providing a comprehensive picture of
the later stages of application and commercialization.

Breakthrough innovations are fundamentally new technologies or services that change the
field of practice. This dramatic shift is relative in comparison to incremental innovations
that improve existing technologies and relatively useless innovations that do not
significantly advance the technical or commercial practice. Breakthrough energy
innovations have been deemed necessary to meaningfully compete with incumbent
technologies [70], overcome the significant market and political barriers to system
transformation [71], and play a meaningful role in mitigating climate disruption [72]. At
the same time, these innovations are likely to be uniquely difficult to achieve from a
technical perspective [73]. Given the disproportionate impact of breakthroughs on
technical advance and societal change [33, 34, 42], we now focus on these most
technically and commercially important patents as measured by citations and Web
presence, respectively. As previously mentioned, citations have a long history in the
bibliometric literature, and we will use them to identify and quantify technical
breakthroughs. However, patent citations are a limited measure of technical importance
and do not necessarily infer commercial relevance or likelihood of deployment.

Supplementing the use of patent citations, we propose to assess commercial value
through automated searches of the Web and content analysis of the resulting URLs. We

18