return

Nanotechnology as a Field of Science: Its Delineation in terms of Journals and Patents

(revision, 24 September 2006)

<pdf-version>

Loet Leydesdorff * & Ping Zhou **

 

* Amsterdam School of Communications Research (ASCoR),

University of Amsterdam, Kloveniersburgwal 48, 1012 CX Amsterdam, The Netherlands

loet@leydesdorff.net; http://www.leydesdorff.net

 

** Institute of Scientific and Technical Information of China,

15 Fuxing Road, Beijing, 100038, P. R. China;

zhoup@istic.ac.cn; pzhou@fmg.uva.nl.

 

Abstract

The Journal Citation Reports of the Science Citation Index 2004 were used to delineate a core set of nanotechnology journals and a nanotechnology-relevant set. In comparison with 2003, the core set has grown and the relevant set has decreased. This suggests a higher degree of codification in the field of nanotechnology: the field has become more focused in terms of citation practices. Using the citing patterns among journals at the aggregate level, a core group of ten nanotechnology journals in the vector space can be delineated on the criterion of betweenness centrality. National contributions to this core group of journals are evaluated for the years 2003, 2004, and 2005. Additionally, the specific class of nanotechnology patents in the database of the U.S. Patent and Trade Office (USPTO) is analyzed to determine if non-patent literature references can be used as a source for the delineation of the knowledge base in terms of scientific journals. The references are primarily to general science journals and letters, and therefore not specific enough for the purpose of delineating a journal set.

 

Keywords:

nanoscience, nanotechnology, classification, interdisciplinarity, journal, patent

 

 

1. Introduction

 

The emergence of new fields of science and technology potentially upsets previously existing classification systems. Chan (1999, pp. 12-16) explains that the Library of Congress of the United States (LC), for example, is based on “literary warrant.” A classification scheme based on literary warrant is not logically deduced from some abstract philosophical system for classifying knowledge but inductively developed in reference to the holdings of a particular library, or to what is or has been published.  In other words, it is based on what the actual literature of the time warrants. The LC has a policy of continuous revision to take current literary warrant into account, so that new areas are developed and obsolete elements are removed or revised (Leydesdorff & Bensman, 2006).

 

Similarly, the U.S. Patent and Trade Office (USPTO) decided in 2004 to introduce a new category into its classification scheme devoted to “nanotechnology.” As defined by the USPTO (at http://www.uspto.gov/web/patents/biochempharm/crossref.htm), nanotechnology patents in this “Class 977” must meet the following criteria:

Patents issued before the new class was created have actively been reclassified by the office. In the meantime, the sub-classifications of Class 977 contain more than 250 categories.[1]  

 

In summary, these two catalogues are very detailed, but they potentially suffer from so-called indexer effects (Courtial et al., 1984, 1993; Healey et al., 1986; King, 1987; Leydesdorff, 1989). Indexes can be considered as second-order mechanisms of codification, while publication and citation practices by active scientists provide first-order updates of scientific literature (Leydesdorff, 2002). In the case of patent references, examiners add citations to the references provided in the applications, but one may expect this to be the case in terms of previous patents more than in the case of previous non-patent literature references (NPLR). NPLRs are less central to the legal upholding of a patent when litigated in court (Granstrand, 1999; Jaffe & Traitenberg, 2002; Meyer, 2000, forthcoming). Can one use aggregated citations among journals and/or in classes of patents for the delineation of a nano-relevant and core nano set of journals?

 

In this study, we update on a previous attempt (Zhou & Leydesdorff, 2006) to use the Journal Citation Reports of the Science Citation Index 2003 for the construction of a nano-relevant set of journals. We use the JCR-data of 2004,[2] and extend the previous analysis by using betweenness centrality as a measure of interdisciplinarity (Leydesdorff, 2006a). “Betweenness centrality” will be analyzed both in the set of journals cited by seed journals in nanoscience and nanotechnology and in the set which is citing this set. Finally, we use non-patent literature references (NPLR) in patent class 977 to examine whether and how a bridge with the relevant journal literature might be provided (Meyer & Persson, 1997).

 

2. Methods and materials

 

The aggregated journal-journal citation data was harvested from the Journal Citation Reports of the Science Citation Index 2004. This data was brought under the control of relational database management. This enables us to generate files that can be imported into programs for statistical analysis and visualization. We use SPSS, UCINet, and Pajek for the statistical analysis, and the latter program also for the visualization.

 

The data allows us to generate citation environments for individual source journals or for a list of such journals at a variable threshold level. (In most analyses below the threshold was one percent of the total citations in the respective dimension (He & Pao, 1986; Leydesdorff & Cozzens, 1993).) The data matrix of aggregated citations among journals is asymmetrical and therefore contains structures in both the “citing” and the “cited” directions. These structures are analyzed using factor analysis with Varimax rotation.

 

Visualizations are based on the vector-space model, using the cosine between vectors as the similarity measure (Salton & McGill, 1983) and the spring-embedded algorithm of Kamada & Kawai (1989) for the representation. The visualizations correspond by and large with the results of factor analysis, since the Pearson correlation coefficient—which is basic to factor analysis—and the cosine are similar, except that the latter normalizes on the basis of the geometrical mean while the former uses the arithmetic mean (Jones & Furnas, 1987; Ahlgren et al., 2003; Chen, 2006).

 

The patent data was downloaded from the USPTO database on June 20, 2006, by using the Internet module available in Visual Basic (Leydesdorff, 2004, at p. 1001) and the search string “CCL/977/$ and ISD/$/$/2005”.[3] The data was then brought under the control of a database manager for further processing. The descriptive statistics of this data are provided in Table 1.

 

“CCL/977/$ and ISD/$/$/2005”

 

Number of patents retrieved

336

Nr of assignees

352

Nr of inventors

1027

Nr of patent references

4830

NPLR

1948

 

Table 1: Patents assigned under the category “nanotechnology” in the USPTO database during 2005.

 

 

3. Nanotechnology journals

 

Zhou & Leydesdorff (2006) used three journals included in Science Citation Index 2003 with the stem “nano” in their title. In 2004, six such journals could be retrieved in the Science Citation Index.[4] Table 2 shows the aggregated citation matrix among these six journals.

 

 

citing →

Fullerenes Nanotubes and Carbon Nanostructures

21

0

0

0

0

0

IEEE Transaction on Nanotechnology

0

33

0

3

15

18

Journal of Nanoparticle Research

0

2

22

7

13

7

Journal of Nanoscience and Nanotechnology

2

2

0

30

10

9

Nano Letters

0

23

9

96

727

160

Nanotechnology

2

24

7

23

107

247

 

Table 2: Aggregated citation matrix among six journals with the stem “nano” in their main title and included in the SCI 2004.

 

This citation matrix reveals upon inspection that Fullerenes Nanotubes and Carbon Nanostructures is not cited by the other five journals, and articles in this journal rarely cite papers in the other ones. Fullerenes Nanotubes and Carbon Nanostructures is an older journal; the journal is a prolongation of Fullerene Science and Technology, which had published its first volume in 1992. (Nanotubes were discovered as a specific form of fullerenes in 1991.)

 

Figure 1: Citation Impact Environment of 14 journals citing Fullerenes Nanotubes and Carbon Nanostructures more than once (cosine ≥  0.2).

 

Figure 1 shows that Fullerenes Nanotubes and Carbon Nanostructures is firmly integrated in a set of chemical-physics journals. However, we shall see below that the journals citing and cited by this journal are heavily interwoven with the journals in the environment of the other five journals. Let us first combine the sets of journals citing or cited by these six journals.

 

3.1 The citation environment of the combined set

 

The local citation environments of the core journals in nanotechnology are sometimes very large. For example, Nano Letters is cited by articles in 305 journals more than once.[5] However, only 17 of these journals cite Nano Letters to the extent of more than one percent of its total citation rate of 7,349. Authors in Nano Letters themselves cite 372 journals, of which only 16 to the extent of more than one percent of the journal’s total references (12,131). In order to discard these large tails of the distributions, we shall use this one-percent threshold for the delineation (He & Pao, 1986; Leydesdorff & Cozzens, 1993).

 

Thirty-seven journals cite one of the six seed journals with the stem “nano” in their title above the threshold, and 53 journals are cited by them. Since there is an overlap of 23 journals among these two subsets, 67 journals can be considered as “nano-relevant” journals. Using the same threshold of one percent, Zhou & Leydesdorff (2006) found 85 journals to be “nano-relevant” in 2003 using only three instead of six seed journals. Therefore, the conclusion seems justified that the nano-relevant environment among scientific journals is an increasingly focused set.

Figure 2: Screeplot of 67 journals in the relevant citation environment of six core nano-journals.

 

When the citation matrix among these 67 journals is analyzed in terms of being-cited patterns, the screeplot (Figure 2) suggests the extraction of eight factors explaining 56.5% of the variance in this matrix. Table 3 provides the eight-factor solution using Varimax rotation and Kaiser normalization.


 

                                                                             Rotated Component Matrix(a)

 

 

Component

 

1

2

3

4

5

6

7

8

SOLID STATE ELECTRON

.868

-.115

-.122

 

 

 

 

 

IEEE ELECTR DEVICE L

.811

-.121

-.169

-.182

 

 

-.185

 

P IEEE

.788

-.168

-.109

 

 

 

-.164

 

APPL PHYS LETT

.777

 

.113

.392

 

 

.169

 

IEEE T ELECTRON DEV

.751

-.119

-.186

-.186

 

 

-.217

.107

J APPL PHYS

.748

 

 

.452

 

 

.233

 

JPN J APPL PHYS

.658

 

 

 

 

 

.148

 

J VAC SCI TECHNOL B

.632

 

 

.120

 

 

 

 

SEMICOND SCI TECH

.629

 

 

.562

 

 

 

 

MRS BULL

.612

 

.459

.387

 

.112

.380

 

IEEE T NANOTECHNOL

.557

-.128

.149

.193

 

 

-.183

-.111

APPL SURF SCI

.417

-.102

.181

.294

 

-.146

.176

-.158

J CRYST GROWTH

.334

 

 

.103

 

 

.248

 

FULLER NANOTUB CAR N

-.180

 

 

 

 

-.113

 

-.166

AEROSOL SCI TECH

-.124

 

 

 

 

 

 

 

POWDER TECHNOL

-.113

 

 

 

 

 

 

 

RUSS J INORG CHEM+

-.108

 

 

 

 

 

 

 

J ORG CHEM

 

.925

-.124