Nanotechnology research networks in Brazil Structure, evolution, and

Document Sample
Nanotechnology research networks in Brazil Structure, evolution, and Powered By Docstoc
					                     Nanotechnology research networks in Brazil
                           Structure, evolution, and policy concerns1

                                                 Luciano Kay



    1. Introduction

There is a growing body of literature related to scientific collaborations and, since more recently, focused
on the network analysis of scientific co-authorships. In particular, the analysis of social networks of
scientists has gained lots of attention and offered important insights to understand scientific collaborations
(Newman, 2001a, 2001b). Meanwhile, the study of inter-institutional and inter-country levels of
collaboration has remained relatively less explored, focusing on aspects like the nature of collaborations
(Katz & Martin, 1997), the rationale for research collaboration (Hagedoorn et al., 2000; Heinze &
Kuhlmann, 2008) , the intra-institutional organization of collaboration (Adams et al., 2005), the inter-
institutional organization of collaborations (Corley et al., 2006; Schummer, 2004), and, at the inter-
country level, the structure of international collaborations (Gomez et al., 1999). In general, the network
analysis approach has been less prevalent in the analysis of inter-institutional and inter-country scientific
collaborations than in the analysis of scientists’ networks.
        When exploring the literature of scientific collaborations through the lenses of policy implications
for developing countries, our knowledge lessens to some extent. In general, there is an agreement
regarding the importance of international scientific collaborations for those countries (Corder et al., 2002;
Gomez et al., 1999) which, typically, have less developed S&T systems and more priorities related to
social or economic development, making international collaboration an opportunity to, essentially, source
knowledge and resources. However, there are some new exceptions. Brazil, like a few other transition
economies, has followed an impressive development path in recent years. Between 2000 and 2007, this
South American giant has grown 43 percent in terms of GDP per capita and 51 percent in terms of total
GDP to reach $1,800 billion.2 Its total S&T expenditure is half-and-half split between industry and public
funding and still totaling only about 1 percent of the GDP, but making this country the leader in science
and technology in Latin America and the 15th largest producer of scientific publications in the world.
Remarkably, the rise in research output has outpaced the patenting activity, what suggests weaknesses in
the commercialization of knowledge (Bound, 2008; Hill, 2007).


1
  Work in progress, presented at the Workshop on Original Policy Research (WOPR), School of Public Policy,
Georgia Institute of Technology.
2
  PPP values according to World Bank country data.

                                                     -1-
        In terms of scientific research output, Brazil has increased its visibility in global networks and
productivity (Meneghini, 1996; Packer & Meneghini, 2006). Moreover, international collaboration
increased the impact of Brazilian publications (Leta & Chaimovich, 2002; Meneghini, 1996) and has been
considered, with some caveats, as instrument of economic policy and geopolitics (da Silva, 2007).
Specifically in terms of nanotechnology, increasing research has raised some questions about the
magnitude and distribution of the potential benefits of this new technology (Invernizzi, 2007) and about
the potential strategies that this country may pursue based on scientific collaborations (Kay & Shapira,
2008). In Latin America, Brazil has been a leader in nanotechnology research and the first country to
implement public programs to support its development. Interestingly, the federal government
nanotechnology policy has displayed a network approach since its origin, to promote research, integrate
efforts, and increase cooperation within the country and internationally. That inspires the analysis pursued
here and grows the expectations to contribute the literature concerned with inter-institutional research
collaborations and their policy implications, based on the analysis of the Brazilian network of
nanotechnology research, its recent evolution, and the potential implications for policy-making.



    2. Nanotechnology research policy in Brazil

According to research publications, nanotechnology research in Brazil began in the early 1990s and
started growing significantly towards the end of that decade, without the support of any specific program
or policy (Kay & Shapira, 2008). Nanotechnology policy effectively started in 2001, when the federal
government decided to articulate a coordinated development of nanotechnology through the
implementation of programs that can take advantage of the existing infrastructure within a strategy of
long-term development of nanosciences (MCT, 2002). The first step was supporting four proposals of
institutional, multidisciplinary networks recommended by CNPq3 aimed at promoting research in specific
areas of nanotechnology (Table 1). The main goal of this initiative was to promote the creation and
consolidation of integrated cooperative networks of basic and applied research, including the participation
of industry (MCT, 2002).
        That initial step was followed by the creation of four research institutes under a broader program
of S&T development (PADCT) and the direct support to government institutions undertaking
nanotechnology research through “sectoral funds,” a development tool that has been proven successful in
Brazil to promote R&D and university-industry collaboration (Bound, 2008). More recently, the Rede
BrasilNano program has supported the creation of ten new research networks, adding in this case a

3
 Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) is a government agency created in 1951
whose mission is to promote and stimulate the scientific and technological development of the country and
contribute to the formulation of national S&T policy.

                                                    -2-
linking component between the program to develop nanotechnology and broader industry, technology,
and trade policies (Invernizzi, 2007; MCT, 2006). These initiatives are combined under the umbrella of
the National Program of Nanotechnology which, interestingly, includes goals like reaching one percent of
the global markets for materials, products, and processes based on nanotechnology and exports of about
$10 billion within 10 years (Goncalves da Silva, 2003). Another interesting and recent program is the
joint initiative to create a nanotechnology research center with Argentina (CBANN) in 2005, explicitly
following the model of a biotechnology center created in 1987 (Centro Argentino Brasileño de
Biotecnologia (CABBIO), 2008) and other antecedents of bilateral cooperation with that country in areas
like nuclear energy (de Oliveira, 1998). Among other activities, this center has organized workshops and
schools of nanotechnology in both Argentina and Brazil with the participation of scholars from other
countries. The most recent nanotechnology research initiative has been the creation of a research network
by Petrobras, a very large state-owned oil company (Bound, 2008).
         An evaluation made by the federal government (Ministry of S&T) revealed that the research
networks supported originally by the “Redes Nacionais de Nanotecnologia” program, have involved
between 2001 and 2005 about 260 researchers, 90 research institutions (including 13 firms,) more than
990 publications, and 97 nanotechnology patents (MCT, 2006). That evaluation also pointed out the need
of intensifying international collaborations since few of them are active, including the bilateral initiative
with Argentina and collaborations with France, India, and South Africa, while other collaborations with,
for example, the US and Japan, were demanding larger projects. Latest reports estimate about 2,200
authors undertaking nanotechnology research in all the networks (Bound, 2008). Overall, the total
investment in different nanotechnology programs reached $100 million for the period 2001-2007 (MCT,
2006).


                                                    TABLE 1
                           Main programs related to nanotechnology in the last 10 years

         Program                        Description                          Targets              Funding ($
                                                                                                   million)a
   Edital CNPq 01/2001     Support to four nanotechnology          Mapping national                  5.0
   “Redes Nacionais de     projects in the areas of                competences for
   Nanotecnologia”         nanobiotechnology and its               nanotechnology.
   (2001-2003)             applications (NANOBIOTEC),
                           nanostructured materials
                           (NANOMAT), semiconductors
                           (NANOSEMIMAT), and molecular
                           nanotechnology and interfaces
                           (RENAMI).
   Institutos do Milênio   Creation of research institutes like:   Creation of nanotechnology        27.2
   do PADCT (2001-         Instituto do Milênio de Materiais       research institutions within
   2006)                   Complexos, Instituto de                 the “Programa de Apoio ao


                                                        -3-
                            Nanociências, Rede de Pesquisa em         Desenvolvimiento Cientifico
                            Sistema em Chip, Microssistemas e         e Tecnologico” (PADCT.)
                            Nanoeletrônica, and Instituto
                            Multidisciplinar de Materiais
                            Poliméricos.
    Sectoral funds (2003-   Support to strategic national                                                   9.7
    2006)                   laboratories like LNLS, INMETRO,
                            and EMBRAPA.4
    2004-2007 Program       Support to ten nanotechnology             Generation of patents,               58.9
    "Desenvolvimento da     projects in diverse nanotechnology        products, and processes,
    Nanociência e da        areas5 to promote basic research,         support to basic research,
    Nanotecnologia" and     generation of patents, and new            cooperation between
    creation of Rede        products and processes.                   industry and other research
    BrasilNano (2004-                                                 institutions, and
    2007)                                                             strengthening of existing
                                                                      infrastructure.
    Centro Brasileiro-      Creation of a center for cooperation      Develop joint projects, raise         n/a
    Argentino de            in nanotechnology research between        human resources capacity,
    Nanotecnologia          Brazil and Argentina.                     create interchange grants for
    (CBANN) (2005)                                                    researchers, organize
                                                                      activities like forums and
                                                                      conferences, and increase
                                                                      interactions with industry.
           Source: (Almeida, 2005; Bound, 2008; Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq),
           2006; Goncalves da Silva, 2003; MCT, 2006). Notes: a. estímate in $ million.


         At least formally, the common denominator in Brazil nanotechnology programs has been the
cooperative component, aimed at integrating efforts and leveraging existing resources through networks
of research institutions, approach that is in line with the trend already observed worldwide in scientific
research, which is increasing inter-institutional and international scientific collaborations.
Nanotechnology research is uniquely related to that trend, particularly in areas like Physics or Materials
Science where expensive equipment and data collection are needed (Meneghini, 1996; Schummer, 2004).
A network analysis approach seems adequate and even interesting having data for a 10-year timeframe.
This analysis allows testing a hypothesis of increasing inter-institutional cooperation and institutional
involvement from different types of institutions in nanotechnology networks during that timeframe,
seeking to answer, first of all, the question of how nanotechnology research networks look and evolve in
this country and, secondly, what the implications for policy are.

4
  LNLS stands for Associação Brasileira de Tecnologia Luz Síncroton, INMETRO for Instituto Nacional de
Metrologia (under the Ministry of Development, Industry, and Trade,) and EMBRAPA for Empresa Brasileira de
Pesquisa Agropecuária (under the Ministry of Agriculture.)
5
  Areas of nanostructured optical materials and nanobiofotonics application; nanobiotechnology; nanotubes
properties; development of nanotechnology for the production of cosmetics and dermatologic products; diffusion of
the scanning probe microscopy technique; nanostructured materials based on natural polysaccharides; development
of nanostructured coatings; development of numeric methods for nanostructures modelling; and, development of
biocomposites and biocompatible magnetic fluids.

                                                         -4-
    3. Data and methods

The empirical analyses presented in this paper draw on global databases of nanotechnology publications
developed at Georgia Institute of Technology, using the definition of nanotechnology and methods
described in (Porter et al., 2007). A two-stage modularized Boolean approach to defining
nanotechnology combined with expert panel review was used to operationalize a definition of
nanotechnology and develop publication and patent datasets for the 1998 to 2007 time period. This
approach identified more than 400,000 records in the Web of Science’s Science Citation Index (WOS-
SCI) in this same timeframe. A subdataset was created for Brazil, comprising records for 6,670
publications where at least one author affiliation is located in that country.
         Different dimensions of the data are analyzed. For the purpose of the analysis, the set of national
and international research institutions and their collaborations (i.e. co-authorships) are considered a
research network. There are a few antecedents of analysis of evolution of research networks at the
institutional level. Some inspiration for defining the methodology of analysis was found in prior research
works applying and suggesting methodologies for network analysis (Iori et al., 2005; Magnien et al.,
2005; Newman, 2001a). In this work, an approach based on comparing different measures for each year
during the period 1998-2007 is explored. The main advantages of this approach are its simplicity and the
possibility of doing a relatively detailed analysis using the available data. The main drawback is that the
analyzed network cannot be considered static when looking at specific years due to the very dynamics of
the collaborative research activity. For instance, it may be the case that, although some collaborative
research takes place during a specific year, its output is only reflected in publications after some time as
the process of cooperation and publication evolves. The approach used here cannot distinguish such
situations of active collaboration without publications in specific years, for which other complementary
methods can be used in future research. As far as the purpose of this work, all conclusions are based on
trends emerging along the entire 1998-2007 period considering that caveat.
         The network analysis approach required creating two main matrices from the nanotechnology
research publications dataset for each year: the adjacency matrix and the weighted connectivity matrix.
The elements aij of the adjacency matrix indicate whether there is a co-authorship between institutions i
and j (or between countries i and j.) The elements wij of the weighted connectivity matrix represent the
total number of articles co-authored between institutions i and j (or between countries i and j.) These
matrices define non-directed graphs when analyzing this research network, i.e. the links representing co-
authorships are bi-directional or aij = aji and wij = wji.




                                                        -5-
    4. Main findings

Clearly Brazil is increasing its research output in nanotechnology and involving a higher number of
national and international research institutions (Figure 1). During the period 1998-2007 Brazil has had
average annual increases of 17 percent in its total research output. The total number of research
institutions and authors involved in nanotechnology has also grown steadily since 1998 as well as the
total number of links (collaborations) since 2000.


                                                                 FIGURE 1
                                      Main statistics of the Brazil nanotechnology research network.
                                              1,200                                                               4,500
                                                                                                                  4,000
                                              1,000
                                                                                                                  3,500




                                                                                                                          Links and Authors
                         Nodes and Articles




                                               800                                                                3,000
                                                                                                                  2,500
                                               600
                                                                                                                  2,000
                                               400                                                                1,500
                                                                                                                  1,000
                                               200
                                                                                                                  500
                                                 0                                                                -
                                                      1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
                    Nodes (All inst.)                 160   204   212   277   299   337   376   419   545   586
                    Nodes (Brazil inst.)              47    68    75    92    100   115   115   145   152   185
                    Articles                          270   377   387   538   704   674   802   864   983 1,071
                    Links (collaborations) 1,873 1,686 1,601 1,748 1,724 1,740 1,909 1,936 2,213 2,287
                    Authors                           746 1,091 1,152 1,582 1,830 1,931 2,343 2,770 3,687 3,823




        Furthermore, the increasing number of authors and research institutions per article reveals
increasing collaboration within nanotechnology research networks (Figure 2). For instance, while in 1998
about 33 percent of research articles were authored by single institutions, in 2007 only slightly more than
20 percent were the output of single institutions. Meanwhile, in the same period, the percentage of articles
co-authored by three different research institutions increased from 16 percent to 27 percent. The
increasing number of authors per article also implies increasing collaboration, although this is an increase
that may take place within the same research institution.
        From a network analysis perspective it is interesting to know whether there is an increasingly
distributed research activity in nanotechnology. The research output of the top 20 research institutions
reveals how they continue concentrating the major part of the total nanotechnology research in terms of
publications, characteristic that is not unique to nanotechnology but to all scientific output in Brazil
(Bound, 2008; Packer & Meneghini, 2006). As of 2007, the top 5 research institutions were co-authors in

                                                                        -6-
60 percent of the published articles, while the next 15 research institutions were co-authors in 55 percent
of them. On the other hand, research institutions that are not among those top 20 are increasing their
participation in the total research output, from about 15 percent in 1998 to almost 50 percent in 2007
(Figure 3).6

                                                                FIGURE 2
                                               Different measures of collaboration per article

             Authors per article                           Institutions per article                                     Countries per article
    100%                                           100%                                                          100%
     90%                                            90%                                                           90%
     80%                                            80%                                                           80%
     70%                                            70%                                                           70%
     60%                                            60%                                                           60%
     50%                                            50%                                                           50%
     40%                                            40%                                                           40%
     30%                                            30%                                                           30%
     20%                                            20%                                                           20%
     10%                                            10%                                                           10%
      0%                                             0%                                                            0%



              1   2   3   4   5    6 or more                  1   2   3   4            5 or more                          1   2   3     4   5 or more




                           FIGURE 3                                                                   FIGURE 4
      Institutional participation in total nanotechnology                     Participation of different types of research institutions in
      research output for top Brazil research institutions                                nanotechnology research in Brazil
     80.0%                                                                                             100%                                        3.0%
                                                                              Academy and government




     70.0%                                                                                              90%
                                                                                                                                                   2.5%




                                                                                                                                                          Industry and others
                                                                                                        80%
     60.0%
                                                                                                        70%                                        2.0%
     50.0%                                                                                              60%
     40.0%                                                                                              50%                                        1.5%
     30.0%                                                                                              40%
                                                                                                        30%                                        1.0%
     20.0%
                                                                                                        20%                                        0.5%
     10.0%
                                                                                                        10%
      0.0%                                                                                               0%                                        0.0%



                                                                                                              Academy                 Government
                  top 5           next 15      remaining                                                      Other                   Industry




             Brazil nanotechnology research is dominated primarily by universities, followed by government
research centers at the national and state level. The participation of both types of institutions has been

6
    These percentages do not add up to 100 percent due to collaborations.

                                                                      -7-
relatively stable in the period 1998-2007, with universities co-authoring more than 90 percent of the
articles and government institutions slightly more than 10 percent. There is also a slow but steady
increase for industry and other type of institutions (independent foundations, hospitals, etc.), which
overall have still a relatively low participation in the total research output (co-authoring only about 4
percent of articles) (Figure 4).
        The number of different institutions that collaborate with each research institution (i.e. its number
of connections) is popular measure that characterizes the structure of networks. This number is defined as
the degree centrality of the node in network analysis (in this case, a non-directed network), given by the
formula:


                                                    =
                                                            ∈ ()


        where aij represents the existence of a co-authorship between institutions i and j and V(i)
represent the set of neighbors of (those who collaborated with) the institution i. The normalized degree
centrality is equal to di divided by the total number of nodes (i.e. institutions) in the network minus 1.
Both measures of degree are calculated based on the adjacency matrices.
        The mean degree for nanotechnology research institutions in Brazil increased from almost 5 in
1998 to 6 in 2007 but its standard deviation has doubled. This is a typical characteristic of most real-
world growing networks, where the distribution of the degree is heterogeneous and its mean is not
representative (Magnien et al., 2005). Typically, there are two different groups of nodes (institutions): a
large group of small-degree nodes and a small group of large-degree nodes. In terms of the
nanotechnology network analyzed here this implies that, although there is an increase in the number of
research institutions involved in nanotechnology (as observed before,) the most connected, as expected,
are still the few largest research institutions. Accordingly, the distribution function of the normalized
degree seems to follow a power law (inverted in Figure 5) where there are many research institutions with
low degree and a fat tail consisting of a small number of institutions with a relatively high degree. This
concentration of links around institutions with the highest research output is also predicted by an OLS
regression (Figure 6).


                      FIGURE 5                                                       FIGURE 6
  Normalized degree centrality distribution for nodes in           Relation between degree and research output for
       Brazil nanotechnology research network                        Brazil nanotechnology research institutions




                                                           -8-
                                       1.00

        Probability Density Function
                                                                                                                           45
                                       0.90




                                                                                            Normalized degree centrality
                                                                                                                           40         R² = 0.8516
                                       0.80                                                                                35
                                       0.70                                                                                30
                                                                  1998                                                     25
                                       0.60
                                                                  2001                                                     20
                                       0.50
                                                                  2004                                                     15
                                       0.40
                                                                  2007                                                     10
                                       0.30
                                                                                                                            5
                                       0.20                                                                                 0
                                              1   6   11 16 21 26 31 36                                                         0       100      200        300

                                              Normalized Degree Centrality                                                          Publications per year



              Using weighted connectivity matrices it is possible to assess the concentration of more dense
patterns of collaboration (i.e. those involving more co-authored articles) around some institutions. The
strength of nodes is defined by the formula:

                                                                                 =
                                                                                     ∈ ()


              where wij represents the number of articles co-authored between institutions i and j and V(i)
represent the set of neighbors of (those who collaborated with) the institution i. This analysis of the
strength of the nodes in the network shows a relative increase in the strength of top institutions when
comparing for example 1998 with 2007 (Figure 7) which reflects their increasing number of research
partners and a stable or increasing research output. Once again, the predominant role of the largest
institutions is evident.
              Another measure that helps in understanding the structure of this research network is the
clustering coefficient. It is a measure of the density of connections around a node or, in other words, the
proportion of nearest neighbors of a node that are linked to each other (Iori et al., 2005). The coefficient is
defined by the formula:
                                                                    2
                                                             =                                                             0≤       ≤1
                                                                  ( − 1)
                                                                             ,


              Normalized clustering coefficients are used here to account for a changing network and allow
comparisons within the 1998-2007 timeframe. In a completely connected network coefficients are equal
to 1. As shown in Figure 8, the nanotechnology network has median coefficients that are very low and
decreasing, which indicates decreasing interconnectedness in the network in that timeframe. This may be
explained by increasing number of institutions with low number of co-authored publications.




                                                                                     -9-
                      FIGURE 7                                                                          FIGURE 8
 Strength of nodes (research institutions) in the Brazil                            Median clustering coefficient for nodes in the Brazil
          nanotechnology research network                                                   nanotechnology research network
                                      1.00                                                0.040
       Probability Density Function

                                      0.90                                                0.035
                                      0.80                                                0.030
                                      0.70                                                0.025
                                                                  1998
                                      0.60                        2001                    0.020
                                      0.50                        2004                    0.015
                                      0.40
                                                                  2007
                                                                                          0.010
                                      0.30                                                0.005
                                      0.20                                                0.000
                                             1    10       100      1000
                                                 Strength (log)




                   The share of international collaborations has not increased substantially if looked in terms of the
number of countries co-authoring research articles, although it changed its composition. While the share
of articles co-authored internationally with institutions of other country (i.e. Brazil and a second country)
decreased slightly from 28 to 25 percent in the period 1998-2007, the share of articles co-authored by
three different countries (one of them Brazil) raised from 6 to 8 percent in the same period. This growing
trend in multilateral collaborations is not a unique characteristic of nanotechnology, since it has been
identified for other Latin American countries and scientific fields (Gomez et al., 1999). Overall, the effect
is marginally negative, with a decreasing share of internationally co-authored articles from 37 to 35
percent of the total nanotechnology research output. That share is still similar and somewhat higher than
the 30 percent found by prior research analyzing all scientific collaborations of Brazil with other countries
(Leta & Chaimovich, 2002).
                   The most important partners in research have been USA, France, and Germany, which have
increased their collaborations with Brazil in the period 1998-2007. In the region, there are increasing
regional collaborations with Argentina and relatively stable collaborations with Chile. Meanwhile,
collaborations with other neighbor countries like Cuba and Colombia decreased when comparing 1998
with 2007. Looking at the European countries within the list of top 15 international collaborators, it is
observed an increase in articles co-authored with European countries except with Italy and Belgium,
which have had a decreasing role in collaborations though they are still important research partners (Table
2). Interestingly, collaborations with Japan reached a 9 percent high in 2003 and decreased to less than 5
percent in 2007. At an aggregate level, international collaborations are concentrated primarily in Europe
(77 percent) and North America (27 percent) with a participation of Latin American countries (averaging



                                                                           - 10 -
16 percent for the 1998-2007 timeframe.) Collaborations with Asian countries decreased from 18 percent
in 1998 to 7 percent in 2007 (Figure 9).

                                                       TABLE 2
  International collaborations: shares (percentages) of internationally co-authored nanotechnology research articles
                                                  (top 15 countries)a

               Country                    1998     1999    2000       2001    2002     2003      2004     2005    2006    2007
USA                                        20.8     24.4    29.6       23.2    27.3     23.1      27.0     29.7   25.9     23.7
France                                     10.9     13.4    15.6       14.1    13.1    16.1      14.3     18.4    15.7     14.7
Germany                                    10.9     11.8    11.9       15.8      8.8    11.6      10.6     12.6    14.0    16.6
Spain                                       5.0    10.2      6.7        8.5     7.7      5.4       2.7      4.4     7.2     9.2
England                                     6.9      3.9     5.9        6.8      3.5      7.0      3.1      5.5     7.2      8.9
Japan                                       2.0      0.8     7.4        6.8     6.9      8.7       5.8      6.1     6.9     4.7
Italy                                      11.9      7.9     5.2        5.6     4.2      3.3       6.8      6.1     5.5     5.8
Argentina                                   3.0      3.1     2.2        4.0     4.2      5.0       4.8      5.8     4.7     6.3
Canada                                      5.0      2.4     3.0        3.4      6.9      3.7      5.8      5.1     3.3      3.7
Portugal                                    1.0      3.1     1.5        4.5      1.9      4.1      5.1      3.4     6.1      4.5
Cuba                                        4.0      5.5     8.1        2.8     4.6      2.5       3.8      3.1     4.7     2.1
Russia                                     10.9      3.9     5.9        3.4      3.8      2.9      4.1      2.0     1.9      2.6
Chile                                       0.0      3.1     2.2        2.3     5.4      4.1       2.0      3.4     4.7     3.7
Belgium                                     5.9      4.7     0.7        2.8     3.5      1.7       2.0      3.1     2.5     2.9
Colombia                                    5.0      2.4     1.5        1.1     1.9      1.2       1.4      1.0     3.6     2.6
Share of other countries                   31.7    16.5    16.3       18.6    16.5     20.7      26.6     23.5    27.0    27.9
Total intl. collaboration (articles)       101      127     135        177      260      242      293      293     363      380
Share of total research output (%)         37.4     33.7   34.9       32.9    36.9     35.9      36.5     33.9    36.9    35.5
Source: own data. a. total shares may exceed 100 percent because some articles are the result of collaboration with more than one
country.



                                                       FIGURE 9
                                 Share of co-authorships with different regions worldwide
                       90%
                       80%
                       70%
                       60%                                                                      North America
                       50%                                                                      Europe
                       40%                                                                      Asia
                       30%                                                                      Latin America
                       20%                                                                      Africa
                       10%
                                                                                                Japan
                         0%




         Other analyses of how different areas contribute nanotechnology research have found that the
majority of research output worldwide is concentrated in a handful of areas like Physics, Materials
Science, and Chemistry, pattern that Brazil also follows in general (Glanzel et al., 2003; Meyer &


                                                             - 11 -
Persson, 1998) (Table 3). It is not completely clear in the data how nanotechnology research relates to the
so called “bioenvironmental model” of research of Brazil, characterized by concentrations of research
publications in Agriculture, Biology, and Earth and Space Sciences (Bound, 2008). For instance, there
may be links between some of the top 15 nanotechnology areas and that research model, like between
Materials Science / Engineering and Space Sciences and between Biochemistry & Molecular Biology and
Biology. Yet the relations with Agriculture and Earth Sciences are not clear when considering the top 15
nanotechnology areas. However, there is also an increasing share of the more than 90 remaining
nanotechnology areas that do include, for example, Energy & Fuels, Environmental Sciences,
Biotechnology & Applied Microbiology, and Agriculture. Those 90 secondary areas increased their share
from about 12 to more than 22 percent of the total nanotechnology research output in the period 1998-
2007.

                                                       TABLE 3
              Share (percentages) and total research articles in the top 15 nanotechnology research areasa

        Subject area              1998     1999    2000     2001      2002    2003    2004    2005    2006     2007    Total
Physics                            65.2     49.9    45.0     45.2      44.7   35.9    41.4    37.5     45.6     37.3   2,842
Materials Science                  19.6     30.8    27.1     31.4     22.9    25.1    32.4    25.2     24.7     27.8   1,792
Chemistry                          12.2     11.7    17.1     18.4     20.5    22.0    24.2    23.0     26.4     26.1   1,466
Engineering                         7.4      5.6     8.5      3.2       5.5   10.4     6.0     9.5      6.1      6.4     459
Polymer Science                     4.4      9.5     5.4      8.4       6.8    8.2     5.2     6.9      5.8      5.2     432
Electrochemistry                    2.6      2.7     2.1      3.5       2.4    4.0     3.7     3.0      5.6      4.9     252
Nanoscience & Nanotech.             0.0      0.0     0.0      0.0       0.0    0.0     0.0      0.0     6.3      7.2     139
Dentistry                           0.7      0.8     1.6      1.7       1.4    3.1     2.5     3.1      2.5      2.6     151
Instruments & Instrum.              2.2      1.9     0.8      3.3       1.4    1.9     1.6     1.5      1.4      1.9     117
Metallurgy & Metal. Eng.            1.5      1.1     0.8      1.5       2.3    1.6     2.6     2.1      1.0      1.9     115
Pharmacology & Pharmacy             0.4      1.3     2.1      1.1       0.6    1.0     1.0      1.7     1.8      2.8     102
Crystallography                     0.7      1.6     1.6      1.3       3.7    2.7     1.0     0.7      1.0      0.9      99
Biochemistry & Mol. Biol.           0.7      1.1     1.0      0.7       0.7    1.6     1.0     1.3      1.5      2.1      87
Optics                              4.1      0.5     1.8      0.2       0.9     1.0     1.0     1.5     1.2      1.7      85
Nuclear Science & Tech.             4.1      1.6     0.3      3.5       1.3    1.3     0.9      0.7     0.9      1.4      92
Other areas                        12.2     15.1    14.5     13.0       9.7   10.1    16.0    15.2     20.8     22.3   1,054
Source: own data. a. total shares may exceed 100 percent because some articles are assigned to more than one research area.




         The geography of nanotechnology research in Brazil shows how South-Eastern states concentrate
most of the total research output (Figure 10). Sao Paulo is the leading state in both S&T (Packer &
Meneghini, 2006) and nanotechnology in particular. This is the most developed state, contributing about
30 percent of the country’s GDP and 45 percent of the total public funding for science and technology
(Bound, 2008; Instituto Brasileiro de Geografia e Estatstica (IBGE), 2007). Almost 30 percent of all
Brazil research institutions in the dataset are located in this state, authoring or co-authoring about 65
percent of the total research output in the period 1998-2007 (a proportion of research output higher than


                                                             - 12 -
the average for all sciences.) Using the same benchmarks, Rio de Janeiro is the second state,
concentrating almost 13 percent of the institutions and authoring or co-authoring about 14 percent of the
nanotechnology research articles in that period.


                                                   FIGURE 10
                   States share of nanotechnology research output in Brazil (period 1998-2007)


                                           Roraima
                                                                 Amapa




                                       Amazonas                 Para               Maranhao
                                                                                                  Ceara
                                                                                                       Rio Grande do Norte
                                                                                                        Paraiba
                                                                                         Piaui     Pernambuco
                         Acre                                                                           Alagoas
                                       Rondonia                           Tocantins                    Sergipe

                                                                                          Bahia
                                                      Mato Grosso


                                                                       GoiasDistrito Federal
                    brazil layer
                                                                                   Minas Gerais
                    PUBS                                                                   Espirito Santo
                          0-3                        Mato Grosso do Sul
                          4 - 12                                       Sao Paulo      Rio de Janeiro
                          13 - 39
                          40 - 127                               Parana
                          128 - 270
                          271 - 496                              Santa Catarina
                          497 - 941
                                                        Rio Grande do Sul
                          942 - 4360




        The growth of the nanotechnology research network in Brazil is perceived when it is visualized
based on co-authorship data for different periods. Figure 11 shows how the network has recently evolved
comparing four selected years. The structure of the network cannot be distinguished as clear as different
network measures have shown, but it is noticeable the increasing number of institutions (Brazilian
institutions in darker color) and co-authorships (links.) Increasing international collaborations
(represented by links to nodes in light grey) are also observed, most of them with nodes that appear barely
connected in the periphery of the graphs. This indicates that there are several international institutions that
share, individually, a few co-authorships with Brazilian partners, represented in the graph by low numbers
of thin links. Although there are research collaborations with international institutions that yielded several
co-authored articles, the majority of strong inter-institutional collaborations are held within the country
(not clearly appreciated in the core of the network graphs.)


                                                               - 13 -
                                                    FIGURE 11
                                                                                              ;             co
     Brazil nanotechnology research network evolution (Brazil institutions in darker color; links represent co-
                                                                                                parentheses
            authorships; total number of research institutions, national and international, in parentheses)

                    1998 (N=160)                                                 2001 (N=277)




                    2004 (N=376)                                                 2007 (N=586)




    5. Policy implications

Overall, the data show that the nanotechnology network in Brazil has evolved in the sense that is growing
                                                                                       collaborations.
in terms of nodes and links and changing slightly its structure reducing international collaborations In
other words, collaborations between national research institutions have increased their number, as
demonstrated by increasing number of authors and institutions per publication. The extensive
                                                                                                 research,
participation of universities in the research output is not surprising when comparing with prior research
but the share of government institutions and industry are about a half and a quarter, respectively, of the
                           de
average share found worldwide in top nanotechnology journals (Schummer, 2004). The role of


                                                       - 14 -
government laboratories have grown considerably between 1998 and 2007. Programs like “Institutos do
Milênio” or the sectoral funds may have encouraged that increasing share in recent years, not only for the
institutions directly targeted but also for other government laboratories at the state level. Meanwhile, the
industry role in research appears reasonable considering that Brazil is still a transition economy, but there
may be other more important factors responsible for this, including: low patenting level that suggests
weaknesses in system of commercialization of knowledge (Bound, 2008; Kay & Shapira, 2008), an
inward-looking industry innovation system (Bound, 2008), and the poor returns to private investment in
new technology, which have augmented the doubts about the future of Brazilian technological
development in the past (Pinheiro-Machado & de Oliveira, 2001). Therefore, industrial nanotechnology
research appears as a potential target of future programs, primarily through the reform or improvement of
the current intellectual property system and increasing support of university-industry collaborations in the
context of international projects. In the future, an increasing role of industry research will be an indicator
of successful network policies and an important base for potential commercialization of research findings.
        The data reveal that place still matters in nanotechnology research in Brazil, something that is
directly related to the discussion of the previous paragraph. It is remarkable the uneven participation of
different regions or states, being the vast majority of the research activity concentrated in the South-East,
particularly in Sao Paulo, Rio de Janeiro, and Minas Gerais, following the general pattern of S&T
concentration in the country. A process of decentralization of the S&T activity has begun in recent years
(Bound, 2008) and the nanotechnology network policy may contribute to some extent the aim of
developing lagging states. That is a reasonable target, even within the limits of the current budget for
nanotechnology. For this may be interesting to explore how many research institutions have capabilities
to integrate nanotechnology research networks in those regions and what regional spillovers that activity
may generate. Those institutions can be increasingly targeted in future network programs, coordinating
nanotechnology research with other industry and local development plans. Potential targets may be, for
example, the Northern states of Amazonas and Para, and the North-Eastern state of Maranaho, which
concentrate about 9 percent of the country’s population, 5 percent of its GDP, but only 2 percent of the
nanotechnology research output (Instituto Brasileiro de Geografia e Estatstica (IBGE), 2007).
        From an outward perspective, international collaborations are still critical because they increase
the impact of Brazilian publications in science and grow recognition and international visibility for the
country and the region (Leta & Chaimovich, 2002; Packer & Meneghini, 2006). Important variations in
co-authorships with different world regions at an aggregate level are evidence of selective collaborations
with European countries, in spite of still relatively strong collaborations with North America. This finding
confirms to some extent more broader collaborations in S&T based on bilateral collaborations with
countries like the US, France, Germany, and Japan (Bound, 2008). Overall, the data suggest that, although

                                                    - 15 -
the set of international partners is diverse, Brazil is part of the group of countries with collaborations
based on bilateral relations like those mentioned above, with an increasing share of trilateral relations
(Meyer & Persson, 1998). In the long term, the decrease (today insignificant) of international co-
authorships may be probably the consequence of the network effects through which Brazil successfully
has sourced knowledge, ideas, and other resources to develop its own research capabilities and
infrastructure. For that reason, a decrease in international collaborations or, more probably, a new change
in their configuration may be expected and considered positive. However, the meaning of the observed
patterns of international collaboration is not clear without further research, since different regions may be
better allies for developing capabilities or sourcing resources in different nanotechnology areas that are
not clearly related to specific regions or countries (Schummer, 2004).
        Collaborations with neighbor countries deserve some specific comments. Although the
nanotechnology policy of Brazil has been grounded mostly in national research and collaboration with
leading countries, it has been also to some extent built upon a research partnership with Argentina, the
largest neighbor, through the Centro Brasileiro-Argentino de Nanotecnologia (CABNN), a virtual
nanotechnology center bringing together research from groups working in both countries (Almeida,
2005). The trend of co-authorships with Argentina confirms that. This bilateral cooperation has
antecedents in biotechnology (Centro Argentino Brasileño de Biotecnologia (CABBIO), 2008) and
nuclear technology research (de Oliveira, 1998) and evolves towards a closer integration between the
S&T systems of both countries and, potentially, with all the countries of the MERCOSUR (Corder et al.,
2002).7 Furthermore, in the long-term, that regional partnership may be extended to other countries like
Chile, another research partner, and become a leading initiative for integration and technological
development in the region, something suggested by the author as a potential strategy for Brazil for
nanotechnology development (Kay & Shapira, 2008).
        Since “Redes Nacionais de Nanotecnologia,” the first nanotechnology program implemented in
2001, the federal government has followed a network approach to leverage existing resources and
increase the involvement of different actors. Interestingly but not surprisingly, the data reveal that the call
for research projects to be funded by that first program included a majority of research institutions that
have always dominated the scenario of nanotechnology research in Brazil (at least, during the 1998-2007
timeframe.) Most of the 36 institutions involved in that program have been universities and co-authored
87 percent of the total nanotechnology research output of Brazil in the period 1998-2007 (Table 4). This
is in line with other findings and raises the question of to what extent is possible to integrate more
government laboratories and firms, more states, and universities that are not among, for example, the top
7
  MERCOSUR stands for “Southern Common Market.” It was created by Argentina, Brazil, Paraguay, and Uruguay
in March 1991. The agreements of these countries include goals like gradual elimination of tariff barriers and
harmonization of the macroeconomic policies.

                                                     - 16 -
20. The list of nanotechnology projects approved under the more recent program "Desenvolvimento da
Nanociência e da Nanotecnologia" (2004-2007) gives additional insights. In addition to the most active
universities and federal government laboratories, a set of firms (still a few) are also included as
intervening institutions (MCT, 2006), which may increase the share of industry in nanotechnology
research and, even more interestingly, patenting activity.


                                                   TABLE 4
    Total research output and share of the research institutions involved in the first nanotechnology projects
(NANOBIOTEC, NANOMAT, NANOSEMIMAT, and RENAMI) under the “Redes Nacionais de Nanotecnologia”
                                                program in 2001.

                                    Institution                                  Publicationsa      Share (%)b
           Universidade de São Paulo, Sao Paulo                                         1,291              19.4
           Universidade Estadual de Campinas, SP                                        1,108              16.6
           Universidade Estadual Paulista Julio Mesquita Filho, SP                         565              8.5
           Universidade Federal do Rio de Janeiro, RJ                                      508              7.6
           Universidade de São Paulo em Sao Carlos, SP                                     506              7.6
           Universidade Federal do Rio Grande do Sul, Rio Grande                           413              6.2
           Universidade Federal de Minas Gerais, MG                                        395              5.9
           Universidade de Brasília, SP                                                    329              4.9
           Universidade Federal do Paraná, Parana                                          274              4.1
           Universidade Federal do Ceará, Ceara                                            266              4.0
           Universidade Federal de Pernambuco, PE                                          241              3.6
           Associação Brasileira de Tecnologia Luz Síncroton, SP                           202              3.0
           Pontificia Univ Catolica Rio De Janeiro, Rio De Janeiro                         183              2.7
           Universidade Federal de Santa Catarina, Sao Carlos                              174              2.6
           Universidade Federal de Uberlândia, MG                                          132              2.0
           Universidade Federal Rio Grande Norte, RN                                       117              1.8
           Universidade Federal Goias, Go                                                  109              1.6
           Universidade de São Paulo em Ribeirao Preto, SP                                  92              1.4
           Universidade do Estado do Rio de Janeiro, RJ                                     88              1.3
           Instituto de Pesquisas Energéticas e Nucleares, Sao Paulo                        87              1.3
           Centro Brasileiro de Pesquisas Físicas, Rio De Janeiro                           86              1.3
           Comissão Nacional de Energia Nuclear, MG                                         83              1.2
           Universidade Federal Santa Maria, SP                                             74              1.1
           Universidade Federal da Bahia, BA                                                57              0.9
           Universidade Federal de Juiz de Fora, Minas Gerais                               41              0.6
           Universidade Federal de Sergipe, SE                                              33              0.5
           Universidade Estadual de Ponta Grossa, PR                                        28              0.4
           Universidade Federal de Ouro Preto, MG                                           24              0.4
           Universidade de Mogi das Cruzes, SP                                              24              0.4
           Centro Federal de Educação Tecnológica do Maranhão, MA                           17              0.3
           Escola Politécnica da USP, SP                                                    16              0.2
           Universidade Federal Paraiba, SP                                                 16              0.2
           Instituto Pesquisas Tecnol Estado Sao Paulo Sa, Sao Paulo                        11              0.2
           Universidade Estado Rio Grande Do Norte, RN                                      11              0.2
           Universidade Federal Alagoas, AL                                                 11              0.2
           Univ Catolica Pernambuco, PE                                                      2              0.0
           Total for the group of 36 institutions                                       5,800              87.0
         Source: based on own data. Notes: a. period 1998-2007. b. share of each research institution in the total
         nanotechnology research output of the country for the period; totals exceed 100 percent due to collaborations.


                                                           - 17 -
    6. Conclusions

The analysis has portrayed the general structure of the nanotechnology research network in Brazil and
given some insights for understanding the policy implications of its recent evolution. Generally speaking,
the network has evolved in the sense that has grown in terms of nodes and links, while changing slightly
its structure to reduce and change international collaborations and increase the relative participation of
industry laboratories and industry. The increasing number of authors and institutions per article illustrates
also the growing density of the network. Interestingly, the share of different research areas within
nanotechnology has changed towards more diversity. From the policy perspective, nanotechnology
programs have shown to be aligned with those general trends, although some caveats apply regarding the
geographic distribution of research output and the share of government and industry research. The
evidence neither allows conclusive remarks about the effectiveness of the network policy in Brazil nor
understates its potential. Future analyses may be more instructive for assessing the effectiveness of the
network effects generated by policies when looking back and taking this research work as a baseline.



    7. References

Adams, J. D., Black, G. C., Clemmons, J. R., & Stephan, P. E. (2005). Scientific teams and institutional
        collaborations: Evidence from U.S. universities, 1981-1999. Research Policy, 34(3), 259-285.
Almeida, C. (2005). Brazil and Argentina launch joint nanotech centre. Retrieved November 29, 2007,
        from http://www.scidev.net/news/index.cfm?fuseaction=readnews&itemid=2537
Bound, K. (2008). Brazil. The natural knowledge economy. London: Demos.
Centro Argentino Brasileño de Biotecnologia (CABBIO). (2008). Centro Argentino Brasileño de
        Biotecnologia - CABBIO. Retrieved November 2, 2008, from
        http://www.mincyt.gov.ar/cabbio2.htm
Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq). (2006). Edital CNPq Nano nº
        01/2001. Retrieved September 21, 2008, from
        http://www.memoria.cnpq.br/servicos/editais/ct/edital_012001_nano_termoaditivo.htm
Corder, S., da Costa, M. C., Gomes, E., & Velho, P. E. (2002). MERCOSUR. Cooperacion en ciencia y
        tecnologia. Nueva Antropologia: Revista de Ciencias Sociales, 18(60), 9-28.
Corley, E. A., Boardman, P. C., & Bozeman, B. (2006). Design and the management of multi-institutional
        research collaborations: Theoretical implications from two case studies. Research Policy, 35(7),
        975-993.
da Silva, D. H. (2007). International cooperation in science and technology: opportunities and risks.
        Revista Brasileira De Politica Internacional, 50(1), 5-28.
de Oliveira, O. M. (1998). A integração bilateral Brasil-Argentina: tecnologia nuclear e Mercosul. Revista
        Brasileira De Politica Internacional, 41(1), 5-23.
Glanzel, W., Meyer, M., Du Plessis, M., Thijs, B., Magerman, T., Schlemmer, B., et al. (2003).
        Nanotechnology, Analysis of an Emerging Domain of Scientific and Technological Endeavor.
        Leuven: O&O Statistieken.



                                                    - 18 -
Gomez, I., Fernandez, M. T., & Sebastian, J. (1999). Analysis of the Structure of Intenational Scientific
         Cooperation Networks Through Bibliometric Indicators. Scientometrics, 44(3), 441-457.
Goncalves da Silva, C. (2003). O Programa Nacional de Nanotecnologia e o Centro Nacional de
         Referencia em Nanotecnologia. Retrieved December 12, 2007, from
         http://www.lnls.br/info/programaNano_a.pdf
Hagedoorn, J., Link, A. N., & Vonortas, N. S. (2000). Research partnerships. Research Policy, 29(4-5),
         567-586.
Heinze, T., & Kuhlmann, S. (2008). Across institutional boundaries?: Research collaboration in German
         public sector nanoscience. Research Policy, 37(5), 888-899.
Hill, D. (2007). Brazil, China, India, Russia, and Taiwan Lead S&E Article Output of Non-OECD
         Countries: NSF.
Instituto Brasileiro de Geografia e Estatstica (IBGE). (2007). IBGE divulga as Contas Regionais 2002-
         2005. Retrieved November 2, 2008, from
         http://www.ibge.gov.br/home/presidencia/noticias/noticia_visualiza.php?id_noticia=1039&id_pa
         gina=1
Invernizzi, N. (2007). Los científicos brasileños legitiman las nanotecnologías. Retrieved December 12,
         2007, from http://estudiosdeldesarrollo.net/relans/documentos/Noela-Visiones-esp.pdf.
Iori, G., de Masi, G., Precup, O. V., Gabbi, G., & Caldarelli, G. (2005). A Network Analysis of the Italian
         Overnight Money Market. City University London.
Katz, J. S., & Martin, B. R. (1997). What is research collaboration? Research Policy, 26(1), 1-18.
Kay, L., & Shapira, P. (2008). Developing Nanotechnology in Latin America. Journal of Nanoparticle
         Research, 10.1007/s11051-008-9503-z.
Leta, J., & Chaimovich, H. (2002). Recognition and international collaboration: the Brazilian case.
         Scientometrics, 53(3), 325-335.
Magnien, C., Mariadassou, M., & Roth, C. (2005). A basic toolbox for the analysis of dynamics of
         growing networks. Paper presented at the AlgoTEL 7th "francophone summit on algorithms for
         telecommunications".
MCT. (2002). Nanoredes. Retrieved October 20, 2008, from
         http://www.comciencia.br/reportagens/nanotecnologia/nano11.htm
MCT. (2006). Relatório Nanotecnologia Investimentos, Resultados e Demandas: Secretaria de
         Desenvolvimento Tecnológico e Inovação (SETEC) -- Coordenação-Geral de Micro e
         Nanotecnologias (CGNT).
Meneghini, R. (1996). The Key Role of Collaborative Work in the Growth of Brazilian Science in the
         Last Ten Years. Scientometrics, 35(3), 367-373.
Meyer, M., & Persson, O. (1998). Nanotechnology Interdisciplinarity, Patterns of Collaboration and
         Differences in Application. Scientometrics, 42(2), 195-205.
Newman, M. E. J. (2001a). Scientific collaboration networks. I. Network construction and fundamental
         results. Physical Review E, 64(1), 016131.
Newman, M. E. J. (2001b). Scientific collaboration networks. II. Shortest paths, weighted networks, and
         centrality. Physical Review E, 64(1), 016132.
Packer, A. L., & Meneghini, R. (2006). Articles with authors affiliated to Brazilian institutions published
         from 1994 to 2003 with 100 or more citations: I - The weight of international collaboration and
         the role of the networks. Anais Da Academia Brasileira De Ciencias, 78(4), 841-853.
Pinheiro-Machado, R., & de Oliveira, P. L. (2001). The Brazilian investment in science and technology.
         Brazilian Journal of Medical and Biological Research, 34, 1521-1530.
Porter, A., Youtie, J., Shapira, P., & Schoeneck, D. (2007). Refining search terms for nanotechnology.
         Journal of Nanoparticle Research.
Schummer, J. (2004). Multidisciplinarity, interdisciplinarity, and patterns of research collaboration in
         nanoscience and nanotechnology. Scientometrics, 59(3), 425-465.



                                                  - 19 -