Global R&D Collaboration in the Development of Nanotechnology: The Impact of R&D Collaboration Patterns on Patent Quality

: Nanotechnologies have been regarded as general-purpose, and interdisciplinary technology converged with di ﬀ erent implications of technologies. This paper aims to explore the global Nano collaboration of research and development (R&D) by patent analysis to realize the sustainable development of nanotechnology and investigate the inﬂuence of global R&D collaboration on patent quality. The primary ﬁndings indicated that the number of issued patents grew sharply since 2002, and the major collaborative ﬁelds in terms of nanotechnology are nanoelectronic and nanomaterials. Finally, the collaboration among star assigners, global partnership, and the indicator of patent centrality had a positive impact on patent quality, except that university-industry collaboration did not show a signiﬁcant e ﬀ ect on patent quality. One part of the reason is that industries have not yet fully recognised nanotechnology opportunities with a risk-averse attitude toward market uncertainties. Overall, the relationship of global R&D collaboration is an essential factor in promoting the sustainable development of the patent quality of nanotechnology. An e ﬀ ective patenting strategy and high-quality application of nanotechnology R&D could secure protection for innovations to reinforce core competitiveness in the business. This explores the inﬂuence of global R&D collaboration on patent quality. This paper investigates the impact of four types of variables on patent quality. While the number of claims was used as a proxy for patent quality, the four types of variables were as follows: (a) the presence of star assigners; (b) repeated collaboration; (c) global collaboration; (d) university-industry collaboration.


Introduction
Nanotechnology is one of the emerging and breakthrough general-purpose technologies that can be applied to several technological convergences as the fourth wave of the industrial revolution is started [1]. Many of nanotechnological innovations, therefore, are through a variety of global collaborative activities. Global research and development (R&D) collaborations among academic, industrial, and research institutes have become more and more essential measures of innovation policies to develop nanotechnology. Notably, the dominant business operating paradigms have changed from a self-reliance strategy to partner-dependence strategy [2]. Since an innovation paradigm has encouraged a new method of R&D collaborating with external sources to create new products or services at the beginning of the 2000s, the influence of collaboration on innovation activities enables us to understand more specifically the transformation of R&D collaboration. However, much of the debates on R&D collaboration have investigated the questions about the extent that an incentive for universities and industry collaboration has shifted the resources toward more applied science and technology when patents are easily obtained [3]. These studies, therefore, lead to a doubt as to whether the impact of R&D collaboration patterns has a positive influence on the patent quality of nanotechnology.
While technological innovation cannot be accomplished in solitude, collaboration among inventors without organizational boundaries is required to take into consideration. Specifically, opportunities Since the upcoming technological globalization has been inevitable, the international collaboration amongst multinational firms has become an important trend in the commercialization of nanotechnology. Global collaboration is a phenomenon of collective inventions [12]. Open sharing of information has brought about an instant knowledge accumulation with higher innovation rates and has added a higher value of technology [13]. Several studies have shown the advantages of R&D collaborations as follows: scientists' productivity [14], firms' innovation capability [15], regions' technology competitiveness [16], and industries' knowledge creation [17]. On the other hand, other studies have instead shed light on the flow of knowledge and the actor's performance in the network [18]. Therefore, for a country's economic development and niches in the era of the knowledge economy, global collaboration has become one of the major concerns.
In general, R&D collaboration describes various research projects in which two or more researchers, including people or organizations, interact with each other to co-produce some types of outcomes. The word "global" means that more than one country is somehow involved. More and more global collaborations nowadays exist all over the world. Global collaborative efforts have been among a large number of multinational enterprises or subsidiaries that disseminate all over the world [19,20] or partly due to the intra-and inter-industry that crosses international borders in quest of worldwide competitiveness [21]. According to the above studies, technological innovation is not only accomplished independently but also needs R&D collaboration among inventors without boundaries to obtain economic and competitive advantages.
Since Zhang et al. [22] confirmed that at least 90% of the total number of patents by year are issued by organizations from just one country, the related empirical works in terms of collaboration reported in the literature are classified in three major categories, including cooperation among firms, nations, and industries. First, the collaboration between firms means that firms should have a certain degree of commonality in their knowledge to have a successful collaboration to produce innovation [23,24]. Cohen and Levinthal [25] noted that firms could enhance R&D capability not only by investing in R&D inside the firm but also by hiring the best potential researchers as so-called "star scientists." Similarly, the paper in Zucker et al. [24] proposed that the firm performance in biotechnology is positively influenced by the movement of star scientists. Moreover, this paper intends to regard star scientists as star assigners. Therefore, the star assigner is one of the indicators of collaboration.
In general, global collaboration enables us to obtain more scientific value, and more dispersed R&D teams are more likely to commercialize expensive technologies [2]. As Wang et al. [26] noted, foreign ownership of a patent has a positive impact on patent quality. Therefore, the study used a global collaborative patent as an indicator of patent quality. In addition, repeated collaboration (long-term collaboration), which can be identified from the patent assigner's column, is another type of collaboration. Previous collaborations are found to be determinant of innovative products and could promote mutual trust and confidence [13,27]. Therefore, repeated collaboration was used as one of the indicators in this study.
The last indicator of collaboration is a university-industry collaboration. In general, universities may be a significant source of new knowledge and innovation. The university-industry collaboration may have different approaches such as joint research project, licensing, and technology transferring of university patents, etc. [4,6,28]. The study of Motohashi and Muramatsu [6] shows that university-industry collaboration will have a positive influence on patent quality. Therefore, university-industry collaboration was selected as one of our indicators in terms of patent quality.
In addition to industrial collaboration, university-industry collaboration is a way of cooperation. Nowadays, universities are important channels of new knowledge, particularly in advancing science and technology [13,28]. Universities often tend to act as advancing institutes, offering a platform to collaborate R&D from multiple firms and technological fields [29,30]. The various platforms of university-industry collaboration are regarded as an important channel for economic development by providing businesses and research opportunities for their innovation activity [31]. At the same time, more and more literature has noted that universities are crucial institutions for motivating or enhancing the progress of the national economy [32,33]. Therefore, universities not only create new knowledge but also contribute more directly to the platform of R&D collaboration.
The quality of R&D collaboration strongly depends on technology transfer, characteristics of technology, and various environmental factors during R&D collaborations [34,35]; for example, incubators and geographic location can significantly influence the organizations' propensity to collaborate [36,37]. Other factors, such as incentive mechanisms intermediary agents and faculty characteristics also play a crucial role in making technology or patents more commercial than others [28,38,39]. Therefore, R&D collaboration is an important factor in enhancing the quality of R&D collaboration.
Numerous studies have employed patent quality measures as R&D performance for patent value. This measure explores what influences the technological or economic importance of a patent. Patent quality has been particularly influential in forecasting technology trends, using various indicators such as citations [40], patent family size [41], patent renewal decisions, the number of claims [42], or complex combinations of the above indicators to investigate the value of the patent. Moreover, previous studies have shown that patent documents offer important information related to global collaboration [40,43]. This paper constructed the dataset from the United States Patent and Trademark Office (hereafter referred to as USPTO). The patent is registered in the USPTO which is one of the important databases all over the world in that it enables us to explore the impact of global collaborations and the technological competitiveness of a country or organization. For example, the analysis of patent application enabled us to identify the name of the inventor, the inventor's country of residence, and the location of inventive activity [41]. Therefore, several inventorships indicate that global collaboration and indicators are used in the study.
Inventorship is a fundamental right in patent law. Inventors are those who make an intellectual contribution to the claims of patentable inventions. Inventorship can be sole or joint. Inventorship is the indicator of several inventors collaborating to complete a patent. At least two possible variants exist. First, all inventors share the same country of residence. Second, members of the R&D team declare different countries of residence. In this case, the occurrence of joint inventors is a clear indicator of inventive collaborative activities. As inventors share or state different countries of residence, the literature suggests the location of inventive activity with global collaboration [41].
Similarly, as for patent assignees, multiple assignees are regarded as a given patent belongs to multiple assignees. Empirical evidence suggests that some inventions carried out at multinational enterprises or subsidiary locations are either jointly or exclusively assigned to the headquarters for specific strategic reasons [41,43]. This paper assumes that a single patent may be assigned to either single or multiple entities, including the inventors and assignees. Global R&D collaboration is regarded as no significant difference between the declared different country of residence between joint inventor and assigners. In a word, either inventors or assignees with their associated different countries of residence are stated on the patent documents [41]. Therefore, with the record of patent, it is possible to operate the indexes of the joint invention and joint assignment for the issued patent in the USPTO.
According to the literature discussed above [2,10,27], global collaboration is inevitable in an increasingly globalized world. Therefore, this paper intends to investigate how patent inventors and assigners collaborate with each other between the linkage of university and industry. In short, the case of R&D collaboration has been occurring more and more, not only between enterprises but also between universities, industries, governments, etc. Furthermore, technological globalization is inevitable, especially in nanotechnology that is in a high degree of collaboration. In addition to the collaboration of the inventors, the collaboration of assigners also needs to be explored. Therefore, the purpose of this paper is to examine the relationship between various forms of global R&D collaboration and patent quality. To sum up, with this in mind, the effort of global R&D collaboration between one country and others at the beginning development of nanotechnology can be explored.

Materials and Methods
The nature of this paper is to explore the collaborative nature of nanotechnology by analyzing secondary data from the USPTO patents. To do so, the study employed patent citation and negative binomial regression analysis to understand the quality of the nanotechnology patent. Patents were searched directly on the USPTO official website and then negative binomial regression analysis was utilized as a statistics tool to explore the effect and quality of nanotechnology.
The network analysis includes a model composed of individual nodes and their connections. This approach provides the best way of understanding the relationship in terms of the connections between patents. Two major reasons to use formal network analysis are that network analysis enables us to describe the characterizing technology group and quantitative measures of relationships. Furthermore, the measurement allows us to test the statistical regression of relationships and structure [11]. In this study, we used the technique of network analysis to explore the patterns of technology evolution. Therefore, this study searched relevant patents of nanotechnology on the USPTO website and employed a software package (UCINET) as a research tool to draw its citation network further. Finally, we used an eigenvector as a standard of centrality.

Research Model and Hypothesis Development
This paper explores the influence of global R&D collaboration on patent quality. This paper investigates the impact of four types of variables on patent quality. While the number of claims was used as a proxy for patent quality, the four types of variables were as follows: (a) the presence of star assigners; (b) repeated collaboration; (c) global collaboration; (d) university-industry collaboration. In light of the reasoning above, the following hypotheses were proposed as follows :   Hypothesis 1 (H1). Different types of collaboration have a positive influence on patent quality.
Zucker et al. [23] showed that collaboration with star scientists has a positive influence on the number of products in development and on the market. This paper defined the top 10 per cent assigners as star assigners and hypothesized as follows :   Hypothesis 1a (H1a). The presence of star assigners in the patent team enhances patent quality.
Gittelman [2] argued that the geography of the research collaborations has a significant impact on the firms' scientific influence and their inventive output. In addition, Gergő and Lengyel [2] highlighted the positive impact of inter-firm co-inventor networks on the quality of innovation. The corporations create more knowledge and commercial value by knowledge flow. Therefore, the study predicted: Hypothesis 1b (H1b). The global collaboration of patent assigners could enhance patent quality.
Previous studies [23,26] argued that previous collaborations also lead to a higher percentage of successful collaboration, while Lee et al. [16] argued that the experience of an inventor's past collaboration would strongly inspire the following productivity. Not only the re-collaborations have an encouraging influence on the firm's innovation but also influence the range of patents. Therefore, the study hypothesis was as follows: Hypothesis 1c (H1c). The presence of assigners that repeatedly have collaboration in the patent team is likely to have a positive influence on patent quality.
Several studies [3,6] highlighted that university-industry collaborations have resulted in a favorable result. Other studies, such as Petruzzelli [33], highlighted that geographical distance between universities and industries both have positive relations to the achievement of higher innovative outcomes. Therefore, the study hypothesized as follows:

Hypothesis 1d (H1d). University-industry collaboration is likely to have a positive influence on patent quality.
Wang et al. [42] showed that a high brokerage (intermediary position measured by betweenness centrality) harms the patent renewal decision by using an analysis of the patent citation network. Based on the above, the study predicts that a better network position of patents has a positive impact on patent quality: Hypothesis 2 (H2). Patents in a more central position have a positive influence on patent quality.

Definition of Variable
The definition of variables is summarised in Table 1. We used the duration of the patent application if the patent was applied from 2002 to 2006 as a control variable. The panel data of a five-year duration enabled us to reduce the time effect. Furthermore, a patent's technology filed based on the international patent code was used to control the effect of convergence. The dependent variable of patent quality was measured by the number of claims on a firm's granted patents in a given year [43]. Finally, dependant and independent variables are demonstrated in Table 1.

Patent Data Collection and Selection Procedure
The data of patent documents in the field of nanotechnology were collected from the United States Patent and Trademark Office (USPTO) database, and all these samples of patents were patents issued as utility patents. The USPTO database was extracted by keyword searches on the title, abstracts, and claims ranging from general strategies, such as "Nano", "Bionanotechnology", "Nanomaterials", "Nanomanufacturing", and "Nanoelectronics".
We preferred the USPTO because the USA has the most significant technology market in the world and has accounted for a significant tech sector of economic activity, thus providing us with a crucial dataset to explore the impact of global R&D collaborative efforts and the technological capability of a country or organization [41]. Since the USPTO and the European Cooperative Patent Classification scheme are both global classification systems for patent documents [44], we composed the dataset of collaborative patents in the USPTO to investigate who collaborate with each other between patent inventors and assigners, even the relationship of university and industry. We then reviewed the search results and eliminated the irrelevant ones, such as design patents. Moreover, the study also used the keyword, including nanoelectronic technology, nano-physics, nanobioloy, nanochemistry, and so on, in the map. A total of 2477 patents from 1973 to 2010 were collected. The subject set contained patents from U.S. Patent No. 3,985,578 to U.S. Patent No. 7,868,426.

Descriptions of Nanotechnology Samples
This paper extracted patent counts based on the panel data of the USPTO from 1973 to 2010. This is because there was a period of rising globalization, during which global R&D collaboration may play an essential part in the development of nanotechnology [1]. A total of 2477 patents were identified to realize the relationship of collaboration among nanotechnology networks. Figure 1 shows the general trend of samples in terms of nanotechnology patents, including applied date and issued date. The first patent was applied in 1973, and it was issued in 1976. From 1973 to 2010, over 3000 patents were applied in the U.S. patent office. The number of applied patents reached a peak in 2003. In general, the result indicated that the patent distribution is at the stage of growth. As for cross-border collaboration, a total of 205 patent collaborations were extracted from the above dataset and they were largely collaborative activities of patent assigners in the Asian region. There were 87 collaborative activities in Japan; the second-highest number belonged to the United States with a total of 34 patent collaborations. Noticeably, there were 53 patents of collaborative activities that were cross-national. In a word, Figure 2 shows the collaborative activities of patent assigners. This paper employed social network analysis to visualize the nanotechnology development.
Only necessary technology or patent may be selected or retained that further technological development went on with above-selected technology and patents, which was so-called path dependence [26].
This paper used 2477 issued patents as referred in Figure 1 above to draw technological networks of nanotechnologies. Firstly, 2477 patents were recoded into several variables, such as patent number, patent name, filed, and date of the patent in order to be structured information. Secondly, we rearranged each patent citation information of 2477 patents with other 2477 patents and transformed into adjacency matrix |αij|2477 × 2477. Figure 3 shows the networks of collaborative patents by UCINET 6 with the function of NetDraw. Moreover, Figure 4 further describes the overall cohesion of a network and in turn, allows us to understand the spread of distributing information and within the network.  In Figure 4, the red circle is about nanoelectromechanical switch systems, and the critical patent number is No. 7256063. Patent 7256063 means that structures around the mechanical manipulation of nanotubes are delivered, and its functionality includes automatic switches, adjustable diodes, amplifiers, inverters, variable resistors, pulse position modulators (PPMs), and transistors. Furthermore, Patent No. 7256063 has two independent claims about a method for operating a nanoelectromechanical transistor comprising.
In what follows, the green circle is another networking group, and it is about nanostructures with the number of the patent No. 7354850. The abstract of patent is that nano-whiskers are grown in a non-preferential growth direction by regulation of nucleation conditions to constrain growth in a preferential direction. In a preferred implementation, semiconductor nano-whiskers are grown on a semiconductor substrate surface by effectively inhibiting growth in the preferential direction. Furthermore, this patent has five independent claims about a method of growing a nano-whisker.
Finally, the third circle is about carbon nanotube devices at the top of the network, and the critical patent number is No. 7166325. Both nanotubes and nanotube-based devices are executed in a variety of applications, including an adapter for chemical and biological sensing, conveyed to the nanotube by means of one or more of a variety of materials coupled to the nanotube, such as metal particles, biological particles, and/or layers of the same. Moreover, this patent has three independent claims about a method of making a device comprising at least one carbon nanotube.

Hypothesis Test
To be consistent with our theoretical analysis, we further performed the Poisson regression model to measure their relationships between the variables and test hypotheses. Poisson regression is regarded as an appropriate tool for calculating rate data. The rate is a count of events taking place to a particular unit of observation, which are divided by some measure of that unit's exposure. More generally, event rates can be calculated as events per unit time that allows us to vary for each unit by the observation window to vary for each unit. We employed patent quality as a dependent variable while using collaborative characteristics and network characteristics as independent variables with three models. Firstly, in model 1, the study puts in control variables, such as the nanotechnology field and patent applied year, into consideration. Secondly, in model 2 under the control of model 1, the study discusses the impact of collaborative characteristics on patent quality. Finally, under the control of model 3, the study tests the hypothesis that network characteristics and the result are significant. Table 2 presents the results of our Poisson regression, estimated by STATA software. The independent variables are entered sequentially as follows. In model 1, two control variables have a significant effect, which indicates application year' and technology fields have an impact on patent quality. The number of applied patents grew sharply from 2001 to 2006, which means that this is a particular period of great development of nanotechnology. Furthermore, cross-international code means that this technology has a variety of filed nanotechnology. Therefore, these are all why two control variables have a significant effect. Model 2 tests H1. The results were significantly supported except for university-industry collaboration. The studies of Islam and Miyazakib [5] showed that the universities are the primary source for generating nanotechnology; this means that the development of nanotechnology in the industry is not very significant. In addition, some nanotechnology patent rights belong to industry or research institutes as they keep them secret and do not share with universities. In general, other various collaborative indicators, including star assigners, global collaboration, and repeated collaboration, were significantly supported.
The result of H2 is presented in model 3. Model 3 showed that centrality has a positive effect on patent quality. That probably means that patent locating more central positions would increase the development and specialization of nanotechnology. Therefore, more central patents advance patent quality.

Conclusions
The present study enabled us to have a better understanding of the relationship between different types of collaborations and patent quality. Firstly, it must be noted that the applied patents in terms of nanotechnology reached a peak in 2005 and the issued patents in 2010. One explanation is that the process of the patent issued has its time lag, but it also means that the nanotechnology was at the stage of growth when the number of issued patents grew steadily in recent years. In other words, nanotechnology has gathered great importance in every technology domain.
Secondly, as for global collaboration, a total of 205 patent collaborations were extracted from the USPTO in the same period. Japan and America had the most significant number of assigners. Furthermore, there were 53 patents of collaborative assigners that were multi-national. Global collaboration cannot be negligible when global collaboration may foster more commercially valuable innovation. On the other hand, when the major nanotechnologies were nano-electronic and nanomaterial, the other technologies, such as bio-nanotechnology and nano-manufacturing, were developed significantly.
Finally, the study investigated the impact of patent centrality on patent quality; the four types of variables, including star assigners, repeated collaboration, and global collaboration, were employed. In general, the results confirmed most of our hypotheses with one notable exception. Most indicators showed a positive impact on patent quality, except for university-industry collaboration. One of the reasons is that nano-relevant universities and industries not only do not fully realize the opportunities of nanotechnology yet but also take a risk-averse approach toward market uncertainties. While the newly constructed dataset of patents enables us to test more complicated models, the construct of a new dataset may have a new issue itself in terms of the dramatic change of nanotechnology.
With regard to the limitations of this study, the development of other patent centrality indicators or public-private partnerships [45] is required to expand the scope of analysis. In addition, we only used claims and citation as a proxy for patent value; further studies could be developed based on other intellectual property or financial funding. Furthermore, there were some shortcomings on claims; the number of claims also depends on the technology field and types of inventions, such as a significant difference of patents between the semiconductor and the machine. We expect that the finding from our paper is intriguing to invite further research on the topic of global nanotech collaboration and patent quality. Funding: This research received no external funding.

Conflicts of Interest:
The authors declare no conflict of interest.