Dynamics and agglomeration phenomena in nano sciences and technologies, results from the NANODISTRICT project
The project focused on how a new scientific and technology wave develops and penetrates the economy. In a special session organised at the Turino Triple Helix conference (2005), we opened the discussion on whether we can learn from bio to characterise nano dynamics. This drove the project to focus on 3 major questions.
1) Would we witness a central role of clusters at least at the initial phase ? In turn this entailed two complementary questions : on what factors would such agglomeration phenomena be warranted ? Would the world geography remain similar ?
2) Would start-up firms be the road for industrial development and initial commercialisation, and would we witness a nano industry as we thought we had a bio industry ? What role then for incumbent firms, whether old start-ups or large existing mostly international firms ?
3) What policies and governance mechanisms for framing and shaping nanotechnology developments – knowing that there is a long tradition of reusing past policies that worked in the preceding wave (often in other countries than those concerned) ?
The project mixed both qualitative approaches (with case studies of clusters and firms) and quantitative approaches using scientometrics and combining publications and patents. Before looking at major results, it is important to reflect on its products and dynamics.
The quantitative element required methodological developments (in particular in delineating a new emerging field at the encounter of different disciplines) and costly database development. As soon as the demonstration was made of the interest of such a DB, the new investments required led to a concentration of efforts in an alliance between 2 groups and to look for new funds : these were provided first by one large research organisation (2006-07) and then by the French funding agency (ANR, 2008-2011).
Similarly initial explorations enabled to identify a series of interesting paths to follow, and this drove towards partner specialisation along different lines : on the involvement of large firms, on nano-based innovation in ‘traditional’ or ‘well established’ sectors, in market shaping processes (and in particular the new dynamics in standardisation activities), on agglomeration phenomena (and in particular on differences in cluster performance), and on how regulatory developments deploy (building a strong link with the very visible and political ethical and societal dimensions). In one word, the project actively participated in shaping an agenda, with three important ‘turning point’s through first the academics-stakeholder conference held in Grenoble (March 2006), the Special Issue in Research Policy (July 2007, see below) and third the doctoral school held at Pinson (February 2008, see key events). Let us now turn to the main issues addressed and results arrived at.
The initial work on the database was done by Zitt et al. (2006) who defined a method to delineate the nanotechnology field. The approach proposed a shift away from expert-based delineations. It has however one important drawback though being probably still the most exhaustive today, it requires a full and local access to the WoS, which makes it accessible to only very few scholars. A new development was thus made which was based on a modular approach coupled with specificity measures (Mogoutov and Kahane, Research Policy 2007). This development has been tested for reproducibility (with US colleagues working for the DOE) and has been considered as robust by scholars comparing existing delineations (Huang, 2008). This was applied both on the Web of Science (publications) and USPTO and Patstat (patents) giving a first full nano database beginning of 2007 (it has been updated since once and will be again end of 2010).
Nano science and technology dynamics were then looked at considering the conceptual framework proposed by Bonaccorsi (see 2005 Prime annual conference, and Minerva 2008). Following the ‘search regime’ approach, three properties characterise knowledge production : the rate of growth, the number of options and directions which the researchers explore (and thus the degree of cumulativeness of knowledge), and the nature of complementarities required. These requisite complementarities may be cognitive (e.g. interdisciplinary), technical (e.g. big science and large facilities such as ITER) or institutional (mixing producers from different institutional background, e.g. university–industry).
One first result lied in the identification of the very rapid growth of nano-science : 14% against an average of 2% (see Bonaccorsi & Thoma 2007, Laredo et al. 2009a and b). However there has been debates about the interpretation of the very different pattern observed for patents, where after an initial rapid growth, the global output has remained on a plateau since 2001-02. One interpretation (see Laredo, 2009b) is that we are still at an explorative stage so that ‘families’ (that is technical and/or geographical extensions) remain small, marking the fact that there is limited technological and market knowledge accumulation.
Convergence has been more difficult to address. This has not been a priority since early markers - the yearly rate of appearance of new keywords, consistently over 40% for the whole period (Bonaccorsi, 2005) - were enough to demonstrate high divergence. New theory-based indicators have been since developed to deepen the measure of initial divergence (Bonaccorsi and Vargas, 2008).
Complementarities have been a central focus of investigation.
A first focus has been on institutional complementarities with an exploration based on more than 8000 inventors identified through a first download of the USPTO database. Bonaccorsi and Thoma (2007) proposed a simple taxonomy, based upon separating inventors on the basis of publications in academic journals. They thus classify patents in three classes : patents which gather inventors that are “inventors only”, patents in which all inventors are also “authors” of academic articles, and patents that mix the two. Nanotechnology is at a very early stage, inasmuch as two-thirds of patents have at least one inventor who has also published academic articles. Looking then at performance (the highest patenting authors) and at quality of patents (mixing diversity, breadth and extension), they show that hybrid “author–inventor” patents are the most promising, underscoring the critical importance of institutional complementarities.
A second focus has been on technical complementarities, building upon the work of Mangematin and colleagues on biotechnology technological platforms. A case study approach was selected comparing the emergence of Mesa + (Twente, the Netherlands) and Minatec (Grenoble, France). In both cases we witnessed a strong organisational transformation, internal to one university in one case, reshaping the relations between organisations in the other. Delemarle (2007) analyses the phenomenon and shows the critical role of ‘institutional entrepreneurs’ in such dynamics, explaining partly the large variety we observe between ‘nano science districts’. Robinson et al. (2007) focus on factors that underpin such new organisational developments. They suggest that a major difference (especially compared to biotechnology) deals with how to manipulate and produce at the nanoscale. They thus highlight the key role of research facilities. Their assertion is that “technological agglomeration is the effect of technological platforms being set-up, used and expanded.”
Trying to delineate the issue of ‘cognitive complementarities’, that is the mobilisation of different disciplines or technological areas, two parallel efforts were undertaken, one considering publications and the other patents. Using patents, Avenel et al. (2007) delineate knowledge building strategies of firms. They confirm a convergence of the nano knowledge base of firms through the examination of the technological diversity of both patents and the patent portfolio of firms. They show, however, that small and large firms deploy different strategies : convergence is at the level of individual patents for small firms, while it is mostly at the level of the patent portfolio for large firms. This suggests two trajectories of innovation, through hybridizing the existing knowledge base for large firms and via the search for breakthrough innovations in small firms. For publications a two stage strategy was taken : a first inquiry was to look into the coverage of scientific domains at the district level, gathering production under three large domains, nanoelectronics and physics, nanomaterials and chemistry, nanobiotechnology and life sciences. We considered that a simultaneous presence of the three domains would warrant the possibility of cross fertilising and thus of “long distance” interdisciplinarity. This proved a simple but powerful proxy in examining clusters (see below). Meanwhile work done by Rafols (in particular on kinesin, see 2008 ENID conference) demonstrated that it was important to go beyond such a rough marker. This faced us with a very difficult methodological problem and the need for very costly new developments. Thanks to the creation of the IFRIS platform, we have been able to develop a new software (keywords Lab, see 2010 ENID conference for a first presentation) which is still under testing and should generate results in 2011.
This first knowledge base on the dynamics of nano sciences and technologies is a major result per se that has given rise to numerous presentations (in seminars in Manchester, Seoul and Atlanta, in stakeholder conferences in Grenoble, Paris, ...), to one book chapter (Laredo 2009), and one article in a journal widely disseminated in the nano community (Nano Law and Business, 2009). It has served for training days (at IHEST, 2010). And there is now a 5 minutes video available on request (as are the different presentations made).
Equipped with this first knowledge base on nanodynamics, we could start addressing the three central issues identified : agglomeration phenomena, role of start-up firms, policy implications.
A clear first development lies with agglomeration phenomena. Geolocalising all addresses of authors and inventors has required completely new methodological developments. They enabled us to demonstrate that the core of scientific production is concentrated in 200 clusters worldwide. We also demonstrated that if the US geography bears much upon this inherited from previous scientific waves (as shown by Zucker et al. 2007, and nuanced by Shapira et al. 2008 highlighting an impact of the large facilities created in particular by DOE), such was not the case worldwide, with a strong emergence of Asia, around very large publication clusters. Europe also witnesses a shifting map with classical strongholds of new scientific developments (e.g. Cambridge and Oxford) falling behind new places such as Leuven or Grenoble. It also proposes a very different articulation of districts. While the US witness a very hierarchical structure where 5 clusters drive all co-authored work, with relatively very limited international linkages, Europe exhibits more inter-country linkages (within Europe) than intra-country ones, which is a striking feature of the development of European level connectivity. Work is on-going to evaluate the potential for growth and breakthrough innovation of clusters (see presentations in Grenoble 2009 and at the 2010 ENID conference).
A second line of interrogation dealt with the involvement of economic actors. Contrary to what was expected we showed the very high involvement in patenting of large incumbent firms : around 60% of the largest R&D industry performers (using either the DTI or the IPTS scoreboard) have nanopatents. Though we cannot clearly characterise the role of start-up firms, this has driven us to hypothesise that the growth of nanotechnology would not warrant the emergence of a large new specialised industry. This also made us consider (see Laredo et al, 2009 c and d, and Laredo et al., 2010) that we were facing a new type of general purpose technology (Bresnehan and Trajtenberg 1995). This was warranted on the wide industrial coverage on one side, and on the very early presence of incumbent firms on the other. In previous cases, pervading the economy went through equipment goods (e.g. the steam engine), the emergence of a new utility industry (electricity) or via mass produced intermediary goods (computers and semiconductors). Here we seem to face “new methods of inventing” (to use Griliches 1957 definition). Pervasiveness would thus be at the R&D stage. In such a case, the new industry would focus on new tools and methods, it could also be based on R&D outsourcing and contract research. Both have limited chance to build a large industry. The only condition under which new firms might play a role is linked to societal and regulatory issues, large incumbents not wiling to be involved into controversies, and leaving to smaller firms the role of ‘demonstrating’ the value of the new products (before buying them, we would then be faced to a near to similar trajectory of biotechnology, see Rothaermel 2007). For the time being this is not what we are witnessing and the central phenomenon at the exploration level, is the high involvement of existing global firms in nearly all economic activities.
Putting together both phenomena drives to question the first generation on ‘nanotechnology initiatives’. Laredo et al. (2010) review them and underline four central challenges. Nanotechnology policies have been quite good at nurturing the development of instruments, concepts, and models needed to work at the nanoscale. However this is far less the case of the three other issues : (1) how to combine all over the board or even thematic calls with the agglomeration process and the need for complementary competences within districts ; (2) how to combine nano knowledge concentration and widely distributed sectoral industrial competences (just as an example there are over 60 industry related “poles” in France !) ? and (3) what policies to prepare widely disseminated abilities for undertaking R&D at the nanolevel ? None of these have been yet seriously addressed (with may be the exception of the second issue and German sector based leading edge innovation projects).
Some outputs related to the nanodistrict project
The special issue in Research Policy (vol 36, issue 6)