Report in reference to Trends in Sustainable Energy

lable at ScienceDirect

Renewable Energy 62 (2014) 657e671

Contents lists avai

Renewable Energy

journal homepage: www.elsevier.com/locate/renene

The production of scientific knowledge on renewable energies: Worldwide trends, dynamics and challenges and implications for management

Francesco Rizzi a, b, *, Nees Jan van Eck c, Marco Frey a

a Scuola Superiore Sant’Anna, Istituto di Management, 24 Piazza Martiri della Libertà, Pisa, Italy b Università Europea di Roma, 190 Via degli Aldobrandeschi, Roma, Italy c Leiden University, Centre for Science and Technology Studies, 62A Wassenaarseweg, Leiden, The Netherlands

a r t i c l e i n f o

Article history: Received 8 May 2013 Accepted 19 August 2013 Available online 19 September 2013

Keywords: Renewable energy Sustainable development Scientific knowledge Research strategy Resource management

* Corresponding author. Scuola Superiore Sant’Ann Piazza Martiri della Libertà, 56127 Pisa, Italy. Tel.: þ39 883936.

E-mail address: [email protected] (F. Rizzi).

0960-1481/$ e see front matter � 2013 Elsevier Ltd. http://dx.doi.org/10.1016/j.renene.2013.08.030

a b s t r a c t

The present study investigates renewable energy analyzing the last twenty years of worldwide scientific production and the dynamics of interest around relevant policies in this direction. Based on a review on the role of knowledge development in technology transitions, we coupled bibliometric and expert debate approaches to provide decision makers with a sound analysis of thematic and regional trends in the field. Results show that the level of activity of researchers in the field of solar energy is somewhat contrasted only by biomass and wind energy. Despite countries being embedded in a global virtual network, geographical differences still arise: while North America and Europe show isomorphism of national communities and a high diversification of vertical foci, emerging research communities (e.g. BRICS countries) reflect market strategies (e.g. China) and the natural environment (e.g. Brazil) with a higher directionality of researches.

Our findings provide an overall picture on world-wide development of competences as a relevant variable which policy makers should ideally consider in detail when setting integrated research, in- dustrial and energy policies and strategies.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Technological trends in the field of renewable energy sources (RESs) have been object of interest of analysts ever since the early discussions on sustainable development [1]. Among all drivers, scientific production (i.e. manuscripts published in peer reviewed scientific journals) is the most prominent on driving technology transitions. Indeed, scientific production informs those policies that are subject to an increasing integration between R&D and tech- nology transitions at national level and that are determined by the energy, environmental and socio-economic dimensions [2,3]. Knowledge development in the field of RESs is also a key aspect in sustainability transition studies [4], a growing field of research that includes transition management [5e7], strategic niche manage- ment [8e10], multi-level perspective on sociotechnical transitions [11e13], and technological innovation systems [14e16]. Researches

a, Istituto di Management, 24 (0)50 883972; fax: þ39 (0)50

All rights reserved.

from all these theoretical perspectives could benefit from comprehensive bibliometric analyses of sectorial scientific outputs.

Although bibliometric maps and expert knowledge present some limitations, bibliometric maps can support experts in improving their knowledge of a certain domain [17,18]. Biblio- metrics, the set of methods to quantitatively analyze scientific and technology literature, provide researchers (as well as public and private decision makers) measurements that can help (i) under- stand complex dynamics including, for example, the needs for balancing demand pull measures and direct public support towards challenging energy targets [19], (ii) forecast the productivity of national investments [20] and (iii) take strategic decisions for strengthening national innovation systems [21]. In that, a world- wide perspective can help capture changes occurring over time in the setting of knowledge generation and knowledge sharing and which influence the selection environments of technological tra- jectories [22].

Unfortunately, despite their relevance, scientific productions in the field of RESs are hardly ever analyzed longitudinally as a whole [23,24]. Manzano-Agugliaro et al. [25] made a first attempt to tackle the issue through a literature review, thus favouring logical con- nections over the comprehensiveness of the representation.

F. Rizzi et al. / Renewable Energy 62 (2014) 657e671658

Conversely, Romo-Fernandez et al. [23] made a first co-word the- matic analysis based exclusively on papers published in the journal Renewable Energy, whereas Celiktas et al. [24] adopted a country- specific approach. Our study aims to contribute to the literature by presenting a more comprehensive approach. Our paper is organized as follows: first, the “Background” section presents a literature review and the rationale and research questions for analyzing the field of RESs by means of bibliometric maps; the “Methodology” section provides details on how the data was ob- tained and processed; the final sections provide research results, discussion and conclusions, highlighting the contribution of our findings to current literature.

2. Background

Researches in the field of RESs are often perceived as strongly influenced by uncertainties in public R&D portfolios [26]. Research and industrial pressure on these portfolios may cause different expectations and constraints, especially in relation to incumbent technologies [27] that may have a recursive nature [28], i.e. the tendency to evolve through alternate phases of selection and of corroboration by use. As a result, the selection of technology pathways consists in a combination of individual and collective acts that are strongly interconnected with the creation and diffusion of ‘new’ knowledge and the formation of technology-specific advo- cacy coalitions. Even though, since the 1990s, such clusters and networks have started coupling research outcomes with the double-digit growth rate in the RESs market [15], it has been found that selection mechanisms which are associated to sustainability challenges are generally aggravated by strong path-dependencies and lock-ins [4,29]. From this perspective, along with the major role of co-evolution of long-lived infrastructures and technological clusters brought by “network-effects” [30], environmental and socio-economic issues have generated new and particularly rele- vant selection pressures. This fact has led, for example, to novel energy end-use applications that imply step-changes in improved energy efficiency, coupled with further electrification and the progressive use of hydrogen (e.g. in transportation, it has become one of the dominant directions for facing the climate challenge) [26]. In such context, a useful key for interpretation when dealing with relations between scientific progress and technological tran- sitions (whatever model is to be evaluated) can be found in the study of dynamics in the generation of scientific knowledge. Therefore, the first research question targeted in this study is:

RQ1: according to the evolution of world-wide environmental and socio-economic priorities, which domains of RESs show the highest generation of scientific knowledge?

In many cases, diffusion of clean technology is governed by endogenous mechanisms (epidemic learning and learning econo- mies) and by exogenous mechanisms [31] that co-evolve with local energy supply and, most of all, with energy demand systems [26]. Indeed, policies and policy jurisdictions can have a critical influence on the trajectory of technological regimes.

On the one hand, the basic thesis on the innovation gains from regional knowledge spill-overs is relatively easy to find in exemplar renewable energy regions where horizontal cross-fertilization op- portunities have shown to be relatively costless and easily turned into international knowledge portals [32]. At a local level, a related variety of industries can speed up lateral absorptive capacity be- tween neighbouring sectors and stimulate innovation cross-fertil- ization via knowledge spill-overs [33,34]. Similarly, shared infrastructures can deepen synergies of compatible technologies [30]. On the other hand, such local “forces” are not immune from generating trade-offs in knowledge development dynamics. In fact, while a knowledge-based economy develops as a dynamic system

at the global level e thus transcending national or geographical boundaries e national and regional systems of innovation may tend to retain wealth from knowledge at local level [22]. Furthermore, models of induced innovation often describe technological change as a cumulative process that leads to the phenomenon of path dependency [35] that could be emphasized at local level. As an example, lack of coordination in national trajectories in Europe had recently raised the need for a comprehensive and articulated approach to link inter-national research, innovation and deploy- ment which then led to the adoption of the Strategic Energy Technology Plan (SET-Plan) in 2007 [36]. Here, strategic objectives have been formulated based on technology roadmaps that identify priority actions for the forthcoming decade (2010e2020). These actions are linked to more specific three-year-period imple- mentation plans (i.e. 2010e2012 for the first edition) [19]. From a theoretical perspective, research development and demonstration (RD&D) policies in the energy sector should have a very long-term perspective and a balanced presence of competing R&D pathways [27], be continuous and persistent, stimulating creativity of research [15], and tolerant to failures by having “many eggs in the basket” [26]. However, the stability of policy measures varies in time and space with local attitudes. As an example, learning from the past, RD&D programmes emerged as significant contributors both to improved cost and performance [30] and to late adopters who can profit from important learning externalities [26]. In some cases, the complex set of impact of RD&D funding have character- ized national technology scenarios, like in Japan in the field of GHG- mitigating technologies [37]. In other cases, un-alignment has prevailed and fossil energy subsidies have practically only left spill- over opportunities for post-fossil energy supply technologies [38]. The fact that in many countries RD&D in renewables is much more dependent than RD&D in fossil fuels on the levels of private sector involvement [36] and of market penetration [39] further supports its geographical relatedness with natural and socio-economic resources.

Worldwide trends in scientific production can help understand these complex relations. Scientific publications are an important component of field-specific knowledge. Bibliometric studies can thus be very useful in understanding the relationship between the functional and socio-technical spaces of niche, regime and land- scape processes of sustainability transitions and other dimensions of space, such as territory, administration and communication (and their particular topologies) [13]. The second research question targeted in this study is therefore:

RQ2: are there national peculiarities in the field of RESs knowledge generation? Additionally, can a systematic analysis of these peculiar- ities draw directions of future theoretical and empirical studies in the field?

Innovations in the field of RESs not only compete towards incumbent technologies in search for an improved energy pro- ductivity [40], but also compete between each other. New tech- nological systems are faced with having to overcome a multitude of forces which favour an “incumbent energy system” and have to set in motion a process of cumulative causation which works in favour of the new technology [15]. A systematic analysis on RESs research can shed some light on where new knowledge is pushing tech- nology frontiers and, in a broader sense, how social constructivism applies to sectorial research communities [41]. This, in turn, can be useful in interpreting the impact of a variety of pressures on knowledge development. From this perspective, specific directions are often determined by the capability of demonstration projects and trials to capture and spread learning for the public good, to catalyze and strengthen national industry and develop national markets [42,39]. More generally, under the multi-level perspective of socio-technical transitions, knowledge development is to some

Table 1 List of journals used in the delineation of the field of RESs. Journals are organized per domain within the field of RESs.

Domain Journals

All RESs Renewable & sustainable energy reviews Iet renewable power generation Journal of renewable and sustainable energy Renewable energy

Biomass Global change biology bioenergy Biofuels bioproducts & biorefining-biofpr Bioenergy research Biomass & bioenergy

Geothermal Geothermics Solar Journal of solar energy engineering-transactions of the asme

Progress in photovoltaics Solar energy materials and solar cells Solar energy

Wind Wind energy

F. Rizzi et al. / Renewable Energy 62 (2014) 657e671 659

extent influenced by the integration between demand-side dy- namics and issue-centred policy areas [43]. As an example, the fact that policies inspired by prevailing innovation systems approaches mainly aim at optimizing the institutional environment of firm- based innovation processes and not at transforming whole sys- tems of production and consumption [44] supports the potential presence of directionality failures, i.e. of un-alignments between directional pressures on research and more targeted impulses to- wards defined corridors of acceptable development paths [45,46]. In that, knowledge development seems to be influenced also by energy service demand changes, which are frequently referred to as a pulling factor for drastic change in energy supply [38], especially in case of ready to “go big” solutions [26]. Consequently, energy end-use technologies are expected to attract more research efforts than energy supply technologies. The shape of scientific production clusters can help understanding these complex relations. Thus, the third research question targeted in this study is:

RQ3: Are there dominant exogenous pressures on knowledge development? In particular, are there dominant clusters in knowledge generation within each RES domain?

Besides informing the above mentioned theoretical framework, these RQs can help informing also more specific studies on tech- nological forecasting [47,48] and science and technology road- mapping [49,50].

Being aware that a systematic analysis of the scientific produc- tion in the last two decades in the field of RESs is just at its beginning stage, the sections that follow present a first contribution to answer the above RQs through a comprehensive bibliometric analysis approach.

3. Methodology

Given the overwhelming flood of information, understanding the overall structure of research is becoming more and more difficult [51]. Therefore, in order to answer the RQs above, we adopted a computer-aided bibliometric approach that consists of the following two steps:

I. Publication delineation of the field of RESs; II. Term map construction.

In step I, we identified publications in the field of RESs. Publi- cation data was taken from the Web of Science (WoS) bibliographic database. The period considered for analysis was 1992e2011. Publications were considered only if they were of the Web of Sci- ence document type article or review and if they were written in English. Step I consisted of three substeps. In the first substep, we selected all publications in the 14 journals listed in Table 1. The total number of publications in these journals was 19,782. In the second substep, publications were selected based on keywords occurring in their title and abstract. Table 2 lists the keywords we used. Each keyword was given a score as shown in parentheses in the table. Publications were selected only if the cumulative score of all key- words occurring in the title and abstract of the publication was at least 5. The first choice of the keywords and the corresponding scores was made and validated based on expert knowledge (2 re- searchers from the industry, 4 from university). After testing many different keywords and scores, the keywords and scores listed in Table 2 were found to provide a reasonably accurate delineation of the field of RESs (i.e. direct analyses on samples from the publica- tion delineation served the scope to verify its soundness). As shown in Table 2 a few keywords also had negative scores: the presence of such keywords in the title or abstract of a publication made it more difficult for the publication to be included in the selection. The keyword-based approach in the second substep yielded 18,419

publications in addition to the 19,782 publications identified in the first substep. In the third substep, we searched for additional publications in the Web of Science database that had not been selected in the first two substeps, but that do have strong citation links (i.e. number of direct citations) with the publications selected in these two substeps. The 9016 most strongly linked publications were added to the selection. In this way, we obtained an overall selection of 19,782 þ 18,419 þ 9016 ¼ 47,217 publications.

During the selection of publications in step I, we distinguished between five domains: biomass, geothermal, hydro, solar, and wind. As can be seen in Tables 1 and 2, most journals and keywords that we used relate specifically to one of these domains. Accord- ingly, most of the publications selected in step I could be assigned to one or more domains. The number of publications per domain is reported in Table 3. It is noteworthy to point out that some publi- cations were included in our selection because they were published in a general journal in the field of RESs or because they contained general RESs related terms. Such publications did not belong to any of the five domains.

Based on the publications selected in steps I and II, we con- structed a number of term maps using the VOSviewer software developed by Van Eck and Waltman [52,53] (freely available at www. vosviewer.com), and in particular the VOS mapping and clustering techniques [54,55] and the automatic term identification technique [53,56] implemented in this software. The idea underlying a term map is to provide a visual representation of a field of science by showing the most important terms occurring in the titles and ab- stracts of publications. Terms are identified by scanning the titles and abstracts of publications using natural language processing tech- niques. The most relevant terms are selected for inclusion in a term map by means of a term selection algorithm available in VOSviewer. General terms (e.g. ‘conclusion’, ‘method’, and ‘result’) and terms occurring in only a small number of publications are not included in a term map. Using the VOS mapping technique, terms that often co- occur within the same publications are located close to each other in a term map, while terms that have no or almost no co-occurrences are located further away from each other. In this way, the grouping of terms in a term map provides an indication of the main topics in a field. The larger the number of publications in which a term occurs, the more prominent the term is displayed in a term map. Frequently occurring terms are for instance indicated using a larger font size. For a detailed examination of a term map, it is best to use the VOSviewer software. This software offers extensive support to interactively explore a term map. All term maps presented in this paper can be examined online using the VOSviewer software.

Further details were gathered using so-called overlay visuali- zations. In an overlay visualization, the terms in a term map are

Table 3 Number of publications in each of the five domains within the field of RESs.

Domain No. pubs

Biomass 11,267 Geothermal 1439 Hydro 3192 Solar 26,045 Wind 5926 Total 47,217

Table 2 List of keywords used in the delineation of the field of RESs. The score of each keywords is reported within parentheses. Keywords are organized per domain within the field of RESs.

Domain Keywords (score)

All RESs “renewable Energ*” (2), “sustainable energ*” (2), “*power conver*” (1), “*power generat*” (1), “*power plant*” (1), “*power system*” (1), “alternative energ*” (1), “electricity generat*” (1), “electricity system*” (1), “energy conver*” (1), “energy generat*” (1), “energy system*” (1), “renewable*” (1), “sustainab*” (1)

Biomass “bioenergy” (3), “alternative fuel*” (2), “anaerobic digestion” and “energ*” (2), “biodiesel*” (2), “biofuel*” (2), “biogas*” (2), “biomass combustion” (2), “biomass gasification” (2), “biomass power” (2), “biomass production” (2), “biomass pyrolysis” (2), “biomass” and “energ*” (2), “biochar*” (1), “bioethanol” (1), “biomass energ*” (1), “biomass fuel*” (1), “biomass to energy” (1), “energy from biomass” (1), “palm oil” (1), “soybean oil” (1)

Geothermal “geotherm*” and “energ*” (4), “thermal water” (1) Hydro “hydro turbine*” (3), “hydro electric*” or “hydroelectric*” (2),

“hydro energ*” (2), “hydro generat*” or “hydrogenerat*” (2), “hydro power” or “hydropower” (2), “tidal energ*” (2), “tidal power” (2), “water power” (2), “hydro electric generat*” or “hydroelectric generat*” (1), “hydro” and “energ*” (1), “tidal stream” (1), “tidal” and “energ*” (1), “wave energ*” (1), “wave power” (1), “fish*” (�4), “hz” or “mhz” or “ghz” (�4), “pulse wave*” (�4), “salmon*” (�4), “sine wave*” (�4)

Solar “photovoltaic generator*” (2), “photovoltaic module*” (2), “photovoltaic system*” (2), “solar air” (2), “solar cell*” or “photovoltaic cell*” (2), “solar collector*” or “photovoltaic collector*” (2), “solar concentrator*” (2), “solar energ*” (2), “solar panel*” or “photovoltaic panel*” (2), “solar power” (2), “solar thermal” (2), “photovoltaic*” (1), “solar thermal collector*” (1), “solar” and “energ*” (1),

Wind “wind farm*” (3), “wind park*” (3), “wind turbine*” (3), “wind energ*” (2), “wind generat*” (2), “wind power” (2), “wind turbine generat*” (1), “wind” and “energ*” (1)

F. Rizzi et al. / Renewable Energy 62 (2014) 657e671660

coloured to indicate, for instance, the average year in which the publications containing a specific term were published or the proportion of the publications containing a specific term that are from a certain country [53]. Overlay visualizations make it possible to show time trends or to display how the scientific activity of a country is focused on specific topics.

4. Results and discussion

4.1. Production trends of scientific knowledge of RESs

Between 1992 and 2011, the scientific production in the field of RESs showed an exponential growth in both absolute and relative terms (Fig. 1), with the median of the publication distribution be- tween 2008 and 2009 for biomass, 2007 and 2008 for wind and hydro, 2006 and 2007 for solar and 2004 and 2005 for geothermal.

While the larger volume of manuscripts persistently referred to solar energy in a context where each RES was increasing yearly productions (Fig. 2), the relative speed of growth was steadily positive only for biomasses and wind (Fig. 3). This seems to overall suggest an increasing push from the research community towards diminishing society’s dependence on fossil fuel energy and, in detail, a sharp rise of scientific investments in parallel with the world economic crisis.

Looking at the sectoral productions, the term map based on all RESs manuscripts between 1992 and 2011 (Fig. 4) helps shedding some light on the different determinants of these trends. In the term map, four different colours can be distinguished. Each colour indicates a cluster of related terms. The clusters were produced using the clustering technique available in the VOSviewer software.

Fig. 4 clearly illustrates the presence of four main clusters with a strong polarization between the solar cluster (red (in web version))

and the ones referring to policy (green (in web version)), wind (blue (in web version)) and biomass (yellow (in web version)). For the solar cluster, links with terms that refer to infrastructure and planning, e.g. “grid”, “policy”, “demand”, “economy”, “climate change”, “market”, etc. were less relevant than the ones that sug- gest the prevalence of chemical and physical fundamental research, e.g. “acid”, “catalyst”, “atom”, etc. Wind and biomass clusters, on the contrary, show a more balanced presence of keywords that are typical in basic and applied research. In particular, biomass entails a sub-cluster on hydrogen energy and some bridging keywords from the chemical and physical. Because of the high number of con- nections in each direction of the map, biomass emerges as a largely interdisciplinary field where new streams of researches mainly come from a variety of scientific communities in biology, agronomy, process engineering, chemistry and environmental sciences.

In the overlay map (Fig. 5), each term is coloured based on the average publication year of the publications in which the term occurs. The closer the colour of a term is to blue, the older the publications in which the term occurs. Conversely, the closer the colour of a term is to red, the more recent the publications in which the term occurs. The overlay map highlights some areas of declining interest. Among these, silicon solar cells, geothermal and/ or solar heating systems for buildings seem to lose more attrac- tiveness than other fields of RESs. The search for higher perfor- mances for the earlier and the relative maturity of these technologies for the later seem to offer an explanation for these trends.

4.2. Geographical distribution of RES researches

First evidence from the analysis of geographical distribution of RES researches clearly concerned -in partial contrast with Manzano-Agugliaro et al. [25] the higher growth rates in Europe and East Asia than in North America and, in absolute terms, the higher volumes of production in Europe (Fig. 6). This could to some extent suggest the so-called “European Paradox” [57] to RESs in East Asian countries. However, in order to properly inform long term energy policy making, it is fundamental to thoroughly un- derstand the trends of RES research described below despite they are sometimes perceived as a low priority dimension for boosting markets in the short term [58].

Solar energy presents an interesting boom of researches in Asia after 2008 (Fig. 7). While the European countries still maintain the higher cumulative number of manuscripts, Asia currently achieved the leadership in terms of annual volumes of scientific production. Also, North America has the same trend of BRICS countries (Brazil, Russia, India, China and South Africa, i.e. a non-geographic aggre- gation of emerging economies) which perhaps is driven by the need to develop a number of technologies that are already available in the developed countries [21]. Such trends would suggest a tail of investments in new human and intellectual capital in emerging countries, although it is understood that further research is required in order to weight quality and potential impacts (for both

Fig. 1. Percentage of total population of Web of Science manuscripts between 1992 and 2011 that refers to RESs.

F. Rizzi et al. / Renewable Energy 62 (2014) 657e671 661

industry and academia) from the different scientific contributions. This is even more true if it is considered that our analysis does not include publications in languages other than English, which is in some countries (e.g. in China) not a marginal minority.

Beside China, which has clearly expressed a particularly clear strategic inflection point in 2008 in solar researches, the strategy of overall East Asia (especially, in order, China, South Korea, Taiwan and Japan) to focus on (retail-oriented) solar market is confirmed by the lower ranking in biomass research production (Fig. 8), where the EU-27 countries are largely the major contributors, followed by North America and the BRICS countries, where India and Brazil add a significant number of papers to China’s score. Apparently, EU-27 countries have started research investments on biomasses on average two years before the other continents.

The gap between EU-27 countries and other regions is even more increasing in the field of wind energy (Fig. 9), where Asia and North America have started investing more than the than EU-27 did three years later. Here it is possible to notice, given also the lower absolute number of manuscripts, how much national energy pol- icies matter. As an example, the Iranian research community (the major contributor in West Asia) is aligning the need of domestic

Fig. 2. Number of manuscripts bet

companies of entering into a market that is in line with US and European efforts to curb Iran’s nuclear program.

Similar trends are present also in the field of hydro energy, where the United Kingdom, Spain and Italy are the largest con- tributors to European leadership (Fig. 10) in a competition where USA and China are the other relevant players and Turkey and India are the new entries. Here, the inflection points are located around 2003 for EU-27, Asia and North America, just after the first eco- nomic slowdown in 2001, and show a 2e3 years delay in BRICS, as for the biomass and wind.

Among all the RESs, geothermal energy attracts the minor and less steady volumes of researches (Fig. 11). In this case, beside the leading role of Europe (mainly thanks to Germany, Italy and United Kingdom), USA, Japan and New Zealand have few well-established geothermal research centres, Turkey and Canada are the new entrants.

In sum, EU-27 Countries reflect the European Commission’s approach for research and innovation that, beside the non-uniform presence of specific excellences, tends to cover widely and hori- zontally all the issues that may produce a positive impact for Eu- ropean competitiveness. All the other World regions present

ween 1992 and 2011 per RES.

Fig. 3. Percentage of total population of RES manuscripts between 1992 and 2011 that refer to each RES.

F. Rizzi et al. / Renewable Energy 62 (2014) 657e671662

strategies oriented to more directional research priorities to address cutting-edge challenges. A straightforward example is given by China’s directionality towards solar technologies, that is expected to impact on sectorial structure since solar technologies are marked by complex international consolidation, often in a relatively short time, once breakthrough in markets occurs [39]. In contrast with the increasing gap in scientific production in other RESs domains, such directionality has produced a rapid achieve- ment of USA’s annual counts (Fig. 12) that, given the recent news of internationalization processes of domestic companies (e.g. Sunergy Co in Turkey, Hanergy Group in California and Germany, Suntech Power in India, etc.), makes China a potential (and imminent) worldwide leading supplier of both products and competences.

Considering that among the 20 more cited scientific publica- tions on solar topics (14 are from Europe, 5 from USA and 1 from Japan), further research will have to shed light on how research outcomes are capable to move out of the “ivory tower” in both developed and emerging countries [59,60]. What our data demonstrated is that high quality research (i.e. from the perspec- tive of the influence on the scientific community) in developed countries often comes from a limited number of highly specialized centres. Referring to the sample above, 4 manuscripts are authored/ co-authored by researchers who work in leading solar companies, 9 (i.e. 4 in Colorado and 5 in Denmark) in leading national

Fig. 4. Term map based on RESs manuscripts from the period 1992e2011. An interactive

laboratories, 4 in a leading university largely funded by the Swedish Energy Agency. Thus, both these and the Chinese experience seem to emphasize that sound and steady dyadic relations between sci- entific community and industry or national agencies/departments are fundamental for advancing science and transferring knowledge and innovations.

Such dyadic relations seem to co-occur with the persistent presence or the new introduction of national renewable energy research priorities (Table 4) and are often linked to the availability of specific ecosystem and socio-economic services reported in literature; the latter include the funded demonstration projects and field trials that, according to Brown and Hendry [61], helped USA, Japan and Germany in coordinating on-going R&D and in creating competitive solar socio-technical systems. Institutional environ- ment (sometimes combined with international cooperation, like in the case of Iran), intellectual capital and natural resources seem to be the major key drivers of the presence of these dyadic relations.

The experience of the European community provides inter- esting evidence that these dyadic relations are not necessarily effective only at local or national level, where the selection of leading research centres could take place. In fact, Table 5 clearly shows how the increasing orientation towards bringing together decision-makers from the Member States, industry and research and financial communities at community level �i.e. converging

version of the map is available online at www.neesjanvaneck.nl/renewable_energy/.

Fig. 5. Overlay map based on RESs manuscripts from the period 1992e2011. The colour of a term indicates the average publication year of the publications in which the term occurs (closer to red is more recent). An interactive version of the map is available online at www.neesjanvaneck.nl/renewable_energy/trend/. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

F. Rizzi et al. / Renewable Energy 62 (2014) 657e671 663

towards a coordinated and integrated Strategic Energy Technology Plan (SET-Plan) � is capable of reducing difference in national RESs research mix.

With exception of Europe, West Asia (including Turkey, that is the major contributor to the scientific production of West Asia, especially in the field of energy planning and energy potential assessment) and North America, the other world regions present an opposite trend (i.e. focalization on regional-specific energy mix). Only for hydro and geothermal energy the differences in terms of energy mix shares are decreasing in all the regions.

The optimal balance between a useful directionality and a counter-effective isomorphism is a fundamental point in devel- oping a knowledge-based economy and an interesting area for further research.

4.3. New solar trajectories as a playground for fundamental research

Since new scientific knowledge is by definition the result of combining internal (i.e. from a research team) and external (i.e. shared among the research community) contributions, the world- wide evolution of manuscript contents can provide useful infor- mation about alternative technology pathways.

Surfing in the term map for solar energy (Fig.13), it is possible to recognize four highly-polarized clusters that, differently to what

Fig. 6. Geographical publication

happens with the overall scientific production on RESs, do not significantly cover issues like costs, water use, energy efficiency, no environmental problems and social acceptance. In fact, on one side, keywords like “location”, “solar radiation”, “heat”, “battery” suggest the presence of research activities on operation and system inte- gration. These activities have weak links with the three clusters on the other side that are more close to fundamental research. Building on silicon sub-cluster, that is about the oldest and most diffused technology on the market, the green (in web version) area clearly refers to film engineering, with keywords like “band gap”, “heterojunction”, “interface”, “deposition”; the blue (in web version) area refers to photo-electrochemical efficiency of dye- sensitized solar cells, with research streams that vary from a number of sub-clusters on electrolytes and tio2 to more specific sub-clusters on ruthenium, iodide photo-catalytic activity and related different nano-structures for electron transport, in accor- dance with Kajikawa’s et al. [62] analysis stating that research on electrolytes and modelling is under rapid development. Finally, the yellow (in web version) area refers to polymeric researches on electricity transfer and power conversion efficiency that entails the emerging field of conjugated polymer [63]. Here, research streams cover a large number of organic photovoltaic polymers like pc60bm, pc71bm, polithiophene derivatives and related alternative reactive molecules, ultra-thin films of deposited oligomers, etc. More than in other RESs, this relevant related variety of

trends for the field of RESs.

Fig. 7. Geographical publication trends for the field of solar energy.

F. Rizzi et al. / Renewable Energy 62 (2014) 657e671664

technological trajectories stresses a certain lack of directionality within the scientific community and, in turn, poses questions about the geographical distribution of the different streams of research. This dimension of knowledge development is even more relevant if we consider, on the one hand, the huge amount of research efforts in the field and, on the other, the fact that an overlay-map similar to the one in Fig. 5 seems to suggest that the diversification of tech- nological frontiers (i.e. the development of new technological tra- jectories) is attracting more researches than system integration and design issues (i.e. the improvements of current market technolo- gies). Based on our data, what we can argue is that the way research communities interact with policy makers and industries will be soon determinant for boosting the competition for the creation of new market niches, e.g. organic solar films for functionality- improved products.

4.4. Resource efficiency in biomass research

A certain polarization of research clusters is present also in the field of biomass, where four clusters cover energy conversion issues and two other clusters cover policy and planning issues (Fig. 14). The dynamism of studies in energy conversion issues can be considered particularly high because keywords have a low density

Fig. 8. Geographical publication trend

but have a large number of connections (i.e. are used in a high variety of contexts). In detail, the cyan (in web version) cluster concerns the assessment of a number of biodiesel alternatives to fossil oil, with particular attention to the comparison of engine performances and environmental impacts. The pink (in web version) cluster, which refers to catalysts and reactions for biofuel production, is characterized by the presence of a certain number of alternatives in biodiesel trans-esterification processes and in the production different sets of fuels. The later field is highly linked with the yellow (in web version) cluster, which is about the development of different pyrolysis and gasification techniques, an alternative that seem more investigated than the blue (in web version) cluster on anaerobic digestion and fermentation processes (of sludge, waste, etc.). In sum, even though it is clear that the in- terest in biofuels is increasingly prevailing over other gaseous and solid derivatives (thus confirming the analysis from Kajikawa et al. [62]), there is a high dispersion of efforts among a large variety of technological alternatives. This is in line with some industrial diversification strategies (e.g. the biofuels from algae are being heavily invested in by US and Chinese producers) [32].

From the side of policy and planning issues, a high concentration of studies characterizes both the green (in web version) cluster concerning rotation and harvesting techniques for biomass

s for the field of biomass energy.

Fig. 9. Geographical publication trends for the field of wind energy.

Fig. 10. Geographical publication trends for the field of hydro energy.

1 An interactive version of the map is available online at www.neesjanvaneck.nl/ renewable_energy/biomass/brazil/.

F. Rizzi et al. / Renewable Energy 62 (2014) 657e671 665

production, and the red (in web version) cluster concerning envi- ronmental impacts of biomass energy policies, especially of those that refer to energy crops. Here, quite surprisingly, the need for a systemic perspective on biomass energy supply chain management results in a prevalent attention to climate change and life cycle assessment than on local socio-economic issues.

Further insights come from a dive in the overlay map based on biomass manuscripts between 1992 and 2011 (Fig.15). In sum, from a dive into the maps, while requirements and constraints placed on bio-energy systems are widely accepted (i.e. there is a low disper- sion in streams of investigation), the community of bioenergy re- searchers e inspired by resource efficiency ambitions e seems to be still underway in exploring new technological trajectories. From this perspective, production of bio-fuels and bio-diesel is attracting more research than conversion technologies and biomass supply. Furthermore, it is worth noting that studies in biomass processing are more focused on laboratory scale reactions than on industrial scale up.

Country-specific overlay maps provide this framework with interesting insights on some geographical research peculiarities. As an example, Brazilian research environment displays a strong directionality on supply- and demand-side sustainability of bio- fuels and bio-diesel for automotive industry, that is much more

influential here than in the world-wide framework.1 In fact, in Brazil it is recognized as an interesting exchange of knowledge and technique (bioethanol industries) with the advanced economies [32]. In this search for green technologies, particular attention is paid to the capitalization of local biomass potentials and, in detail, to sustainability of soybean, sunflower, canola, and Jatropha curcas supply and conversion.

4.5. Towards an improved reliability of wind installations

Studies in the field of wind energy are less dispersed than in other RESs (Fig. 16). Apart from a small yellow (in web version) cluster about multiple RESs integration, researches are grouped in only three main clusters. The red (in web version) one presents a high concentration of interest on policy issues. In particular, social acceptance and infrastructural scenarios (including transportation systems and hydrogen storage) seem to be central topics for the evaluation of technologies that are subject to improvements mainly in terms of design capabilities. Not surprisingly, there are strong

Fig. 11. Geographical publication trends for the field of geothermal energy.

Fig. 12. Comparison between the publication trends of the USA and China for the field of RESs.

F. Rizzi et al. / Renewable Energy 62 (2014) 657e671666

links with the green (in web version) cluster, covering design- related studies that range from wind data prediction to simula- tion accuracy, and with the blue (in web version) cluster, covering technological advancements in operation control. This cluster, in detail, collects studies on incremental knowledge in the field of electrical engineering, with particular reference to improved in- duction generation performances.

The overlay map built on wind manuscripts between 1992 and 2011 (Fig. 17) clearly shows that landscape and habitat protection are a key topics in policy issues, while generation control and grid connection (including storage) are emerging key topics in technical research.

Country-specific overlay maps highlight some geographical peculiarities in wind research trends. In particular, it is worth noting that in Denmark, where there is a large prevalence of wind research activity over solar and other RESs, wind technologies are covered through improvements for each component, from blades, to rotor, to turbine, converters and controllers.2 Beside the

2 An interactive version of the map is available online at www.neesjanvaneck.nl/ renewable_energy/wind/denmark/.

integration with a related variety of information and communica- tions technology and complementary green technologies [64], particular attention is paid to aerodynamic design and components fatigue in extreme conditions (e.g. sea applications). Habitat pro- tection and mortality of birds for collisions are well covered as well.

4.6. From local electrical case studies to diffused geothermal direct uses

Also in the field of geothermal energy the dispersion of RESs is thematically low. In fact, only three main clusters can be identified (Fig.18). The green (in web version) one entails keywords related to energy policy, with a specific focus on market barriers, availability of resources, and environmental footprint in energy mix design. The blue (in web version) one illustrates two main aspects: one refers to the design and the energy performance of geothermal district heating systems, an issue which is attracting an increasing interest from the geothermal community thanks to its potential impacts on new urban energy efficiency debates; the other refers to several studies on the new turn of green (in web version) energy strategies in Turkey for a sustainable integration of this technology into the national energy mix. The red (in web version) cluster,

Table 4 Evolution in country’s RES research focus (percentage of the specific RES on total RESs researches).

State years China (solar) Germany (solar) Japan (solar) Iran (wind) Malaysia (biomass)

1992e2001 69.0% 84.5% 81.2% 5.9% 39.1% 2002e2011 66.1% 67.8% 72.1% 31.7% 57.5% Key driver Institutional environment Intellectual capital Intellectual capital International cooperation Natural resources

State Years USA (solar) Finland (solar) Sweden (solar) Denmark (wind) Brazil (biomass)

1992e2001 58.8% 52.4% 43.6% 36.9% 27.6% 2002e2011 51.3% 55.5% 41.7% 48.0% 51.5% Key driver Intellectual capital Natural resources Intellectual capital Natural resources Natural resources

F. Rizzi et al. / Renewable Energy 62 (2014) 657e671 667

which concerns high enthalpy exploration, well drilling, and reservoir management, is still occupying the largest room for dis- cussion on geothermal science.

Differences in these clusters are not only in target markets. Surfing through the network of research items, it is worth noting that while the blue (in web version) cluster is about non site- specific design-oriented studies, the red (in web version) one is mainly about site-specific empirical studies. Because of the site- dependency of outcomes (i.e. uncertainties in their generaliza- tion), it is thus not surprising that, in contrast with the more dy- namic RESs, risks reduction in resource exploration and cultivation prevails over the development of either new or dominant tech- nology paradigms. Additionally, it is worth noting that technical contents are mainly from on-field experiences carried out (with similar frequencies of occurrence) in Larderello (Italy), Geysers (Nevada), Wairakei (New Zealand), Philippines, Soultz (Germany), Krafla (Iceland) and Kyushu and Kakkonda (Japan). Finally, just examining in depth this geographic perspective, the decline of the share of Italian contributions (from 10.3% between 1992 and 2001 to 3.9% between 2002 and 2011) suggests that availability of re- sources and the presence of market leaders is not a condition suf- ficient for preserving scientific leadership.

4.7. New applications of well-established technologies in hydro energy

Together with geothermal energy, hydro energy shows the lower dynamism among RESs in terms of openness to new disci- plines and technological frontiers.

Table 5 Coefficient of variation in RESs relative number of publications in leading European countries (in bold: Belgium, Denmark, France, Germany, Greece, Italy, The Nether- land, Spain, Sweden, Switzerland, United Kingdom) and emerging regions (in italic: World regions with the exclusion of Europe, North America, West Asia).

Wind Hydro Solar Biomass Geothermal

1992 1.06 0.98 1.69 1.70 0.38 0.80 0.83 0.58 2.56 1.58 1993 0.98 1.48 0.95 1.31 0.41 0.95 0.93 1.06 1.99 2.45 1994 1.32 0.97 1.54 0.84 0.39 1.13 0.76 1.00 1.32 2.02 1995 1.07 0.75 1.41 0.91 0.47 0.82 0.98 1.10 0.93 1.27 1996 1.27 0.76 1.54 0.87 0.44 0.73 0.89 0.54 1.85 0.60 1997 1.04 0.72 1.07 0.84 0.45 1.11 1.08 0.82 1.60 1.38 1998 0.99 0.96 0.64 0.94 0.28 0.92 0.66 0.75 1.36 1.40 1999 0.83 0.79 0.72 0.88 0.36 0.73 0.70 0.66 0.94 1.28 2000 0.98 0.83 1.03 0.93 0.29 0.88 0.54 0.84 1.79 1.26 2001 1.04 0.66 0.90 1.22 0.27 1.30 0.56 0.91 1.57 1.22 2002 0.94 0.72 0.93 1.09 0.34 1.09 0.57 0.74 2.19 1.34 2003 0.63 0.65 0.93 1.02 0.30 1.38 0.45 0.80 0.93 1.01 2004 0.83 0.72 0.57 1.46 0.26 1.06 0.57 0.87 0.97 0.85 2005 0.87 1.14 0.68 1.25 0.21 1.31 0.54 1.04 1.33 0.96 2006 0.84 1.29 0.44 1.32 0.29 1.36 0.59 1.13 0.97 1.01 2007 0.58 1.20 0.38 1.05 0.26 1.33 0.37 0.91 0.88 1.14 2008 0.55 1.02 0.35 1.04 0.34 1.44 0.25 0.78 0.54 0.87 2009 0.68 1.00 0.49 0.82 0.27 1.49 0.21 0.77 0.88 1.22 2010 0.74 1.04 0.44 1.06 0.25 1.57 0.30 0.71 0.94 0.90 2011 0.59 1.13 0.37 0.97 0.26 1.63 0.28 0.71 0.73 0.66

In the field of hydro energy (Fig. 19), besides a significant cluster of multi-RESs planning studies (red (in web version) cluster), in- cremental improvements of efficiency of traditional technologies are clearly the focus of the blue (in web version) cluster, that is in turn connected with a cluster concerning conventional hydropower sources (green (in web version)) and storage management (yellow (in web version)). Conventional sources consist of river basins and �with a higher independence from the other hydro studies � marine tides and waves. Storage capacity based on reverse pump- ing represents a relatively novel approach for improving economic efficiency of hydro-systems because of the prevalence of energy price volatility on physical energy losses.

It is worth mentioning that, looking into the data, after a certain equilibrium between scientific production in Nord America, Europe and Asia during 1992e2003, these last two regions have experi- enced a higher growth in scientific production during 2004e2011. From a geographical perspective, besides the United States and United Kingdom, which are the countries with the highest tradition in hydro researches, China, Turkey, Spain, Japan, India, Canada, Brazil and Italy represent recent but important newcomers.

5. Conclusions and policy implications

The present paper opens with a review on technology transition that demonstrates the importance of knowledge development dynamics in informing decision-making processes on behalf of the actors involved. Focussing on the scientific domain, our analysis uses a bibliometric approach to compare the advancement of knowledge among different RESs, both in time and in space. As a result, from a longitudinal perspective, solar, biomass, and wind energy clearly emerge as the research fields attracting most inter- est; in particular, such interest appears to be proportional to the distribution of resources (e.g. solar energy is widely used world- wide) as well as to the flexibility and scalability in terms of demand side actors. In this sense, other characteristics e e.g. being pro- grammable source or not e do not seem to systematically play a relevant role.

From a vertical perspective, while strategies of specialization (e.g. East Asia, South America, Central Asia, etc.) or diversification (e.g. Europe and North America) give a boost to regional peculiar- ities (which are often expression of socio-economic conditions and needs), globalization of markets and of explicit scientific knowl- edge does not seem to give the expected directionality to research streams within each RES. In fact, atomization and variety of research alternatives increase according to the same ranking above (i.e. from wind to biomass to solar energy). This means that the more R&D and energy policies can build on large volumes of sci- entific knowledge, the more social constructivism [41] can emerge as useful to orient researchers, industries and public authorities towards a limited set of technological goals that are embedded in priorities that are shared among stakeholders. In practical terms, considered the soar of research streams in some sectors and the associated threat of the “paradox of choice” (i.e. the loss of

Fig. 13. Term map based on solar manuscripts from the period 1992e2011. An interactive version of the map is available online at www.neesjanvaneck.nl/renewable_energy/solar/.

Fig. 14. Term map based on biomass manuscripts from the period 1992e2011. An interactive version of the map is available online at www.neesjanvaneck.nl/renewable_energy/ biomass/.

Fig. 15. Overlay map based on biomass manuscripts from the period 1992e2011. The colour of a term indicates the average publication year of the publications in which the term occurs (closer to red is more recent). An interactive version of the map is available online at www.neesjanvaneck.nl/renewable_energy/biomass/trend/. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

F. Rizzi et al. / Renewable Energy 62 (2014) 657e671668

Fig. 16. Term map based on wind manuscripts from the period 1992e2011. An interactive version of the map is available online at www.neesjanvaneck.nl/renewable_energy/wind/.

F. Rizzi et al. / Renewable Energy 62 (2014) 657e671 669

confidence and productivity in front of an unmanageable number of alternatives), our data show how the energy research planner is increasingly asked to maintain the control over the worldwide picture of scientific and technological research. Such condition is particularly important to grasp effective investments in promising and emerging technologies that, especially with limited resources, respond to the need to limit the local number of failures and lock- ins. In so doing, monitoring and benchmarking worldwide research production could help in the transition from uncertainty to risk management.

Regarding the other RESs (i.e. hydro and geothermal), well established real or virtual innovation systems seem to introduce barriers to radically new solutions. In these cases, energy and research policy makers may, therefore, find it useful to (i) strengthen some existing (or even latent) advocacy coalitions by creating favourable conditions for private capital, and (ii) support the work of various interest groups associated with alternative

Fig. 17. Overlay map based on wind manuscripts from the period 1992e2011. The colour o occurs (closer to red is more recent). An interactive version of the map is available onlin references to colour in this figure legend, the reader is referred to the web version of this

technologies, in order to prevent incumbents from delaying (in- cremental and, even more, fundamental) scientific development.

Relevant implications for R&D and energy policies and related technology transitions emerge also in terms of key competences required for creating value from progress in scientific knowledge. In fact, our representation of the contrast between international isomorphism and national specialization sheds light on the importance of determining the appropriate directionality in research, energy and industrial policies integration. Based on our findings, we recommend that policy makers and R&D managers recognize the importance of achieving a deep understanding of the forces that shape local energy research mix.

Of course, some limitations in our research should be consid- ered. Firstly, delineation of publications can present some limita- tions despite it being cross-checked by experts. Additionally, considering that term maps are a simplified representation of the contents of manuscripts, this could lead to a partial representation

f a term indicates the average publication year of the publications in which the term e at www.neesjanvaneck.nl/renewable_energy/wind/trend/. (For interpretation of the article.)

Fig. 18. Term map based on geothermal manuscripts from the period 1992e2011. An interactive version of the map is available online at www.neesjanvaneck.nl/renewable_energy/ geothermal/.

Fig. 19. Term map based on hydro manuscripts from the period 1992e2011. An interactive version of the map is available online at www.neesjanvaneck.nl/renewable_energy/ hydro/.

F. Rizzi et al. / Renewable Energy 62 (2014) 657e671670

of the scientific knowledge. Finally, expert based interpretations can be subject to the fashion of the time (e.g. bias in implicit knowledge). Nevertheless, the bibliometric approach we propose offers powerful ways to view the overall structure of research in a manner that an expert cannot perform. Thus we consider this ex- ercise a useful incipit e not a conclusion e to a wider discussion which, given the continuously evolving nature of the topic, simply aims at increasing the understanding of energy and research policy implications.

Thus a number of studies could be usefully developed building on our results. First, further insights on the correlation between scientific production and technology transitions could be achieved through a deeper examination of specific (sectorial or regional) R&D and energy policies. Patent analysis and investigations on best available technologies could be usefully performed to this end. Second, the value chain of scientific information could be explored

in order to inform better the integration of energy, research, and industrial policies.

References

[1] World Commission on Environment and Development (WCED). Our common future. New York: Oxford University Press; 1987.

[2] Aghion P, Howitt P. A model of growth through creative destruction. Econo- metrica 1992;60:323e51.

[3] Romer PM. Increasing returns and long-run growth. J Political Econ 1986;94: 1002e37.

[4] Markard J, Raven R, Truffer B. Sustainability transitions: an emerging field of research and its prospects. Res Policy 2012;41:955e67.

[5] Kern F, Smith A. Restructuring energy systems for sustainability? Energy transition policy in the Netherlands. Energy Policy 2008;36:4093e103.

[6] Loorbach D. Transition management for sustainable development: a pre- scriptive, complexity-based governance framework. Governance 2010;23: 161e83.

F. Rizzi et al. / Renewable Energy 62 (2014) 657e671 671

[7] Rotmans J, Kemp R, van Asselt M. More evolution than revolution. Transition management in public policy. Foresight 2001;3:15e31.

[8] Kemp R, Schot J, Hoogma R. Regime shifts to sustainability through processes of niche formation: the approach of strategic niche management. Technol Anal Strateg Manag 1998;10:175e95.

[9] Raven R, Geels FW. Socio-cognitive evolution in niche development: comparative analysis of biogas development in Denmark and the Netherlands (1973e2004). Technovation 2010;30:87e99.

[10] Smith A. Translating sustainabilities between green niches and socio-technical regimes. Technol Anal Strateg Manag 2007;19:427e50.

[11] Geels FW. Technological transitions as evolutionary reconfiguration processes: a multi-level perspective and a case-study. Res Policy 2002;31: 1257e74.

[12] Geels FW, Schot J. Typology of sociotechnical transition pathways. Res Policy 2007;36:399e417.

[13] Smith A, Voß JP, Grin J. Innovation studies and sustainability transitions: the allure of the multi-level perspective and its challenges. Res Policy 2010;39: 435e48.

[14] Bergek A, Jacobsson S, Carlsson B, Lindmark S, Rickne A. Analyzing the func- tional dynamics of technological innovation systems: a scheme of analysis. Res Policy 2008;37:407e29.

[15] Jacobsson S, Johnson A. The diffusion of renewable energy technology: an analytical framework and key issues for research. Energy Policy 2000;28: 625e40.

[16] Hekkert M, Suurs RAA, Negro S, Kuhlmann S, Smits R. Functions of innovation systems: a new approach for analysing technological change. Technol Forecast Soc Change 2007;74:413e32.

[17] Kostoff RN, Tshiteya R, Pfeil KM, Humenik JA, Karypis G. Power source roadmaps using bibliometrics and database tomography. Energy 2005;30(5): 709e30.

[18] Heersmink R, van den Hoven J, van Eck NJ, van den Berg J. Bibliometric mapping of computer and information ethics. Ethics Inf Technol 2011;13(3): 241e9.

[19] Newbery DM, Olmos L, Ruester S, Liong SJ, Glachant JM. Public support for the financing of RD&D activities. In: New clean energy technologies 2011. A report for the European Commission, Directorate General Energy within the EC FP 7 project THINK.

[20] Rodrik D, Subramanian A, Trebbi F. Institutions rule: the primacy of in- stitutions over geography and integration in economic development. J Econ Growth 2004;9:131e65.

[21] Chan L, Daim T. Exploring the impact of technology foresight studies on innovation: case of BRIC countries. Futures 2012;44:618e30.

[22] Leydesdorff L, Zawdie G. The triple helix perspective of innovation systems. Technol Anal Strateg Manag 2010;22(7):789e804.

[23] Romo-Fernández LM, Guerrero-Bote VP, Moya-Anegón F. Co-word based thematic analysis of renewable energy (1990e2010). Scientometrics 2013 [Available online].

[24] Celiktas MS, Sevgili T, Kocar G. A snapshot of renewable energy research in Turkey. Renew Energy 2009;34(6):1479e86.

[25] Manzano-Agugliaro F, Alcayde A, Montoya FG, Zapata-Sierra A, Gil C. Scientific production of renewable energies worldwide: an overview. Renew Sustain Energy Rev 2013;18:134e43.

[26] Grubler A. Energy transitions research: insights and cautionary tales. Energy Policy 2012;50:8e16.

[27] Jacobsson S, Bergek A. Transforming the energy sector: the evolution of technological systems in renewable energy technology. Ind Corp Change 2004;13(5):815e49.

[28] Perkmann M, Walsh K. The two faces of collaboration: impacts of university- industry relations on public research. Ind Corp Change 2009;18(6):1033e65.

[29] Safarzynska K, van den Bergh JCJM. Demand-supply coevolution with multi- ple increasing returns: policy analysis for unlocking and system transitions. Technol Forecast Soc Change 2010;77:297e317.

[30] Grubler A, Nakicenovic N, Victor DG. Dynamics of energy technologies and global change. Energy Policy 1999;27:247e80.

[31] Kemp R, Volpi M. The diffusion of clean technologies: a review with sugges- tions for future diffusion analysis. J Clean Prod 2008;16(1):S14e21.

[32] Cooke P. Regional innovation systems: development opportunities from the ‘green turn’. Technol Anal Strateg Manag 2010;22(7):831e44.

[33] Boschma R, Wenting R. The spatial evolution of the British automobile in- dustry: does location matter? Ind Corp Change 2007;16:213e38.

[34] Cantwell J, Iammarino S. Multinational corporations and European regional systems of innovation. London: Routledge; 2003.

[35] Binswanger HP. Induced technical change: evolution of thought. In: Binswanger HP, Ruttan VW, editors. Induced innovation: technology, in- stitutions and development. Baltimore: Johns Hopkins University Press; 1978. p. 13e43.

[36] Soriano FH, Mulatero F. EU research and innovation (R&I) in renewable en- ergies: the role of the strategic energy technology plan (SET-Plan). Energy Policy 2011;39:3582e90.

[37] Suwa A, Jupesta J. Policy innovation for technology diffusion: a case-study of Japanese renewable energypublic support programs. Sustain Sci 2012;7:185e97.

[38] Global energy assessment (GEA) e toward a sustainable future. Cambridge, UK: Cambridge University Press; 2012, ISBN 9780521182935; 2012.

[39] Hendry C, Harborne P, Brown J. So what do innovating companies really get from publicly funded demonstration projects and trials? innovation lessons from solar photovoltaics and wind. Energy Policy 2010;38:4507e19.

[40] Bashmakov I. Three laws of energy transitions. Energy Policy 2007;35: 3583e94.

[41] Pinch TJ, Bijker WE. The social construction of facts and artefacts: or how the sociology of science and the sociology of technology might benefit each other. Soc Stud Sci 1984;14:399e441.

[42] Harborne P, Hendry C. Pathways to commercial wind power in the US, Europe and Japan: the role of demonstration projects and field trials in the innovation process. Energy Policy 2009;37(9):3580e95.

[43] Weber KM, Rohracher H. Legitimizing research, technology and innovation policies for transformative change: combining insights from innovation sys- tems and multi-level perspective in a comprehensive ‘failures’ framework. Res Policy 2012;41:1037e47.

[44] Tukker A, Charter M, Vezzoli C, Stø E, Andersen MM, editors. System inno- vation for sustainability 1: perspectives on radical changes to sustainable consumption and production. Sheffield: Greenleaf; 2008.

[45] Azar C, Sandén BA. The elusive quest for technology-neutral policies. Environ Innov Soc Transit 2011;1:135e9.

[46] Jacobsson S, Bergek A. Innovation system analyses and sustainability transi- tions: contributions and suggestions for research. Environ Innov Soc Transit 2011;1:41e57.

[47] Bengisu M, Nekhili R. Forecasting emerging technologies with the aid of science and technology databases. Technol Forecast Soc Change 2006;73: 835e44.

[48] Daim TU, Rueda G, Martin H, Gerdsri P. Forecasting emerging technologies: use of bibliometrics and patent analysis. Technol Forecast Soc Change 2006;73:981e1012.

[49] Phaal R, Farrukh CJP, Probert DR. Technology roadmapping e a planning framework for evolution and revolution. Technol Forecast Soc Change 2004;71:5e26.

[50] Galvin R. Roadmapping e a practitioner’s update. Technol Forecast Soc Change 2004;71:101e3.

[51] Kajikawa Y, Takeda Y. Structure of research on biomass and bio-fuels: a citation-based approach. Technol Forecast Soc Change 2008;75:1349e59.

[52] Van Eck NJ, Waltman L. Software survey: VOSviewer, a computer program for bibliometric mapping. Scientometrics 2010;84(2):523e38.

[53] yan Van Eck NJ, Waltman L. Text mining and visualization using VOSviewer. ISSI Newsl 2011;7(3):50e4.

[54] yan Van Eck NJ, Waltman L, Dekker R, yan Van den Berg J. A comparison of two techniques for bibliometric mapping: multidimensional scaling and VOS. J Am Soc Inf Sci Technol 2010;61(12):2405e16.

[55] Waltman L, yan Van Eck NJ, Noyons EC:M. A unified approach to mapping and clustering of bibliometric networks. J Inf 2010;4(4):629e35.

[56] yan Van Eck NJ, Waltman L, Noyons ECM, Buter RK. Automatic term identi- fication for bibliometric mapping. Scientometrics 2010;82(3):581e96.

[57] Dosi G, Llerena P, Labini MS. The relationships between science, technologies and their industrial exploitation: an illustration through the myths and re- alities of the so-called ‘European Paradox’. Res Policy 2006;35:1450e64.

[58] Sovacool BK. The importance of comprehensiveness in renewable electricity and energy-efficiency policy. Energy Policy 2009;37(4):1529e41.

[59] Mowery D, Nelson R, Sampat B, Ziedonis A. Ivory tower and industrial inno- vation: university-industry technology transfer before and after the Bayh- Dole Act. Stanford Business Books; 2004.

[60] Tan X. Clean technology R&D and innovation in emerging coun- triesdexperience from China. Energy Policy 2010;38(6):2916e26.

[61] Brown J, Hendry C. Public demonstration projects and field trials: accelerating commercialisation of sustainable technology in solar photovoltaics. Energy Policy 2009;37(7):2560e73.

[62] Kajikawa Y, Yoshikawa J, Takeda Y, Matsushima K. Tracking emerging tech- nologies in energy research: toward a roadmap for sustainable energy. Technol Forecast Soc Change 2008;75:771e82.

[63] Kajikawa Y, Takeda Y, Matsushima K. Computer-assisted roadmapping: a case study in energy research. Foresight 2010;12(2):4e15.

[64] Stoerring D, Dalum B. Cluster emergence: a comparative study of two cases in North Jutland, Denmark. In: Cooke P, Schwartz D, editors. Creative regions: technology, culture and knowledge entrepreneurship. London: Routledge; 2007.