International scientific cooperation is one of the key drivers in the contemporary science and technology development. Since the late twentieth century, a growing number of joint publications and international projects have contributed to faster accumulation of knowledge, formation of global research networks and increased citation impact of scientific results (Wagner & Leydesdorff, 2005). This trend is particularly significant in high-tech industries, where achieving results requires concentration of financial resources, equipment and expertise from several countries simultaneously. Research in the field of semiconductors gives a vivid example of this interdependence: key technological chains are spread across various world regions, and international collaborations provide access to critical knowledge and infrastructure (OECD, 2025). The United States and Europe continue to lead in architectural design and software development, Taiwan and Korea specialize in contract manufacturing, and China dominates the rapidly expanding sector of mass production and testing. Under these circumstances, the disruption of international relations threatens both scientific advancement and economic stability.

Since 2022, amidst escalating geopolitical tensions and imposition of sanctions on Russia and some of its partners, international scientific landscape underwent significant changes. Major universities in the United States and Europe have announced that they are suspending cooperation with Russian organizations. Many grant programs, including Horizon Europe, have been stopped for Russian researchers and access to international research infrastructure has been restricted (Plackett, 2022). All this affected not only Russia but also the BRICS countries and, more broadly, the “Global South” (Gueye et al., 2022; McManus et al., 2024; Saba & Pretorius, 2024).

Bibliometric data confirm the decline in publication activity and the sharp drop in joint works between Russia and its Western partners in the period 2022-2023 (Zhang et al., 2024). At the same time, there have been increases in joint publications within BRICS and intensified collaborations with universities and research centers in the Middle East, Africa, and Southeast Asia (Sokolov et al., 2025). This indicates the emergence of new patterns of interaction in global science, which can be described as polarization of the global scientific space. While in the 2000s and 2010s, science developed as an integrated and interconnected network, the 2020s have seen the growing importance of regional clusters and the weakening of global links. Sanctions and political barriers have not only limited BRICS cooperation with Western research centers but also stimulated the rise of peripheral nodes of scientific collaboration, such as Saudi Arabia and Egypt.

The aim of this paper is to reveal, based on bibliometric analysis of publications related to semiconductors, how sanctions have transformed the structure of international scientific cooperation. The research questions are as follows: is there a reduction in the connectivity of the global network and a constraining of traditional interaction channels? And, at the same time is there a stimulation of the development of new regional centers, leading to a gradual transformation of the global architecture of science?

The paper is structured as follows: Section 2 provides a review of relevant literature. Section 3 describes the hypotheses and chosen methodology. Section 4 presents the main findings of the study, including the visualization of networked collaboration in semiconductor research. Section 5 discusses the results compared to the hypotheses formulated in Section 3. The paper concludes with a discussion of the limitations of the study and suggestions for future research paths.

Drawing on theoretical perspectives of the impact of geopolitical factors on transnational scientific networks, we have formulated several hypotheses.

The first hypothesis posits that sanctions against Russia and related restrictions on academic cooperation have led to a decline in joint publications between the BRICS countries and leading Western scientific centers.

The second hypothesis suggests that under the sanctions pressure, polarization within scientific networks has intensified: the reduction in the BRICS linkages with the EU and the United States has been accompanied by growing collaboration among the BRICS countries and also with universities of the Global South — including the Middle East, Africa, and Southeast Asia.

The third hypothesis proposes that sanctions have facilitated the emergence of new peripheral nodes of international cooperation — such as universities in Saudi Arabia and Egypt — which previously played secondary roles but have since strengthened their presence in global publication networks.

To test these hypotheses, we conducted a bibliometric analysis of semiconductor-related publications, representing a strategically vital segment of high technology. The empirical base comprised data from the Web of Science Core Collection covering the period 1965–2025. Approximately 688 thousand publications containing the keyword semiconductor were identified. Two time periods were chosen for analysis: 2000–2021, characterized by relatively stable international cooperation, and 2022–2025, corresponding to the period of extensive sanctions.

The research methodology included constructing international co-authorship networks between countries and organizations using the VOSviewer software and the Python/NetworkX library for graph analysis.

VOSviewer constructs bibliometric networks based on a similarity matrix that reflect the strength of connections between elements such as authors, institutions, countries, or keywords. It processes data imported from Web of Science and computes link strength for each pair of elements according to the chosen analysis type — co-authorship, co-occurrence, co-citation, or bibliographic coupling. To ensure comparability, the software applies the association strength normalization, which compensates for differences in publication volume across nodes and produces a weighted network where vertices represent analytical objects and edges capture the intensity of their interactions. Node placement on the map is determined by the VOS algorithm, a modified multidimensional scaling technique that minimizes distances between strongly connected elements and increases distances between weakly connected ones, producing a stable and interpretable spatial layout. The software then applies modularity-based clustering (a variant of the Louvain algorithm) to identify groups of closely related elements. The visualization simultaneously conveys link strength via distances and edge thickness, and node relevance via size and colour, allowing users to analyze the structure of a research domain and identify its main participants (van Eck & Waltman, 2018).

The essential network-analysis metrics were calculated to assess the position of countries and institutions:

Network clustering was also performed to identify the main research centers and evaluate their evolution over time.

The comparative approach enabled us to juxtapose network structures across the two time intervals and assess changes in link distribution. Special attention was given to the share of the BRICS joint publications with the United States and the EU, the dynamics of Chinese, Russian, and Indian institutions, and the rise of peripheral actors that had previously occupied marginal positions in global science.

We acknowledge the limitations of this approach: bibliometric data capture only a segment of research represented in international peer-reviewed journals and omit local studies, technical reports, and projects under embargo. Nevertheless, the use of the Web of Science database provides the most reliable and comparable dataset for analyzing the transformation of global scientific networks under sanctions pressure.

This paper’s findings show that significant changes have also occurred in the overall network landscape.

An analysis of the interactions between the BRICS countries and the nations of East, Asia, and Africa reveals several notable trends:

Overall, the bibliometric analysis confirmed the initial hypotheses and revealed substantial structural changes in international scientific collaboration following the imposition of sanctions.

First, the role of China as the central node of the global publication network has strengthened markedly. The Chinese Academy of Sciences (CAS) has become the largest institution in terms of co-authored works, maintaining active partnerships with the BRICS members and with several countries of the Global South.

Whereas CAS was already a leading center before 2022, the centrality metrics of CAS have further increased since sanctions were introduced. The number of joint publications with India, Russia, Saudi Arabia, and Egypt has risen significantly. This demonstrates that China is gradually taking on the role of an integrator in international research networks, especially in the face of the weakening of traditional Western actors. Second, the analysis indicates a sharp decline in collaboration between the BRICS countries and Western scientific centers. The share of joint publications with researchers from the United States and the European Union has decreased by nearly half, as evidenced by the drop in betweenness centrality for countries such as Germany, Japan and Britain. Where these nations once served as bridges connecting different clusters, the linking function has shifted increasingly towards China and India since 2022.

This trend reflects a structural fragmentation of the global scientific network. Instead of a single, integrated core, there are now multiple regional clusters forming.

Third, the period under review shows the rising importance of Global South countries as emerging peripheral nodes of scientific collaboration. Universities in Saudi Arabia (e.g., King Saud University, King Khalid University) and Egypt (Egyptian Knowledge Bank and other centers) are now among the top 20 global partners in joint publications with institutions from the BRICS countries. Only a few years ago, these organizations were peripheral to the network and played a limited role in semiconductor research. Their rapid ascent can be explained by the reallocation of international research efforts towards countries not subject to sanctions.

Russia occupies a special place in this analysis. The Russian Academy of Sciences, which ranked among the most prominent network participants in the 2000-2021 period, showed a substantial decline in co-authored publications after 2022 — from more than 1,500 to roughly 700. Particularly sharp was the drop in collaboration with European institutions, which previously accounted for a significant share of Russia's international linkages.

Although Russia has maintained active cooperation with China and India and expanded contacts with Middle Eastern universities, the overall dynamics suggest a reduction in Russia’s role in the global scientific system and its gradual shift toward the periphery.

Taken together, the results confirm the polarization of global science. Sanctions have led to decreased network connectivity, weakened traditional nodes, increased participation of the Global South countries and the rise of new centers of influence based in China and India.

The global scientific system in semiconductor research appears to be entering a phase of structural transformation — from an integrated global network towards a constellation of regional clusters, each with its own core and sphere of influence.

The analysis shows that the sanctions against Russia, and their indirect effects on the BRICS countries, have become a significant factor in driving the transformation of the global science system in the area of semiconductor research.

Before 2022, global science was characterized by increasing integration into a unified network with relatively dense linkages between Western countries and emerging economies.

However, the trend has now reversed towards fragmentation and polarization.

1. First, the results support the hypothesis that there has been a decline in the number of co-authored publications between the BRICS countries and Western countries. Restrictions on access to international grants, the termination of partnerships and barriers to the exchange of equipment and data have all contributed to a reduction in scientific cooperation and a decrease in publication activity.

2. Second, under the pressure of sanctions, polarization of international scientific network intensified. Western countries have partly lost their intermediary role, while China and India increasingly assumed integral functions. This led to the emergence of new regional centers and a shift in emphasis towards the Global South.

3. Third, sanctions have stimulated the rise of new peripheral nodes.

4. Countries such as Saudi Arabia and Egypt, previously peripheral, have significantly strengthened their positions owing to the reallocation of research flows and targeted policies to develop science and innovation.

For Russia, the effects have been predominantly negative. The sharp decline in joint publications with the EU and the United States weakened its position in the global scientific network. Cooperation with China and India, which has been maintained, created a dependence on a limited number of partners and increased the risk of long-term scientific isolation.

On the whole, sanctions have not only reduced the connectivity of international science but also accelerated its structural transformation. The global scientific system is shifting from a formerly integrated network toward a regionalized architecture with distinct centers and zones of influence.

This supports the broader thesis that science is increasingly influenced by geopolitical and geoeconomic factors, with international cooperation closely linked to issues of global security and technological independence.

This study has certain limitations that suggest directions for future research. It inevitably operates within several methodological constraints, some of which arise from conscious decisions we made when developing a general approach to mapping the global scientific network.

First, our reliance on a subset of 100,000 highly cited publications puts the spotlight on established scientific powers and old papers — a familiar side-effect of citation-based selection. Expanding the dataset downward would, of course, include more recent and more diverse works, but it would also blur the comparability of periods and inject a large amount of noise. In this sense, the approach is less about capturing every single collaborative tie and more about ensuring that the structural contours of the network remain interpretable (and replicable).

Second, we did not implement extended keyword harmonization or semantic clustering, even though semiconductor research spans everything from perovskite solar cells to nanoelectronics. Such refinements could certainly reduce thematic “spillover,” but they would also shift the study from macro-mapping to field-level micro-taxonomy — a task for a different paper and, ideally, a different coffee supply.

A related limitation concerns the use of full counts for international co-authorships. Fractional counting often produces a more conservative and “fair” representation of contributions, but switching to this method mid-analysis disrupts comparability across different time periods and institutions. Since our goal was to outline a methodological pathway rather than adjudicate credit allocation in large multi-author teams, we preserved the full-counting logic as the simplest and most transparent option. Equally important is the fact that the Web of Science only captures the visible tip of global research. It leaves out local journals, industrial research and projects covered by confidentiality and export control regimes — an omission that is especially relevant for semiconductor research where classified and proprietary work takes up a large share of real effort. WoS indexing also expanded in the early 2020s, particularly for journals from the Global South, which means that part of the “rise” observed among new regional hubs may reflect adjustments in the database rather than dramatic shifts in scientific behavior.

Another limitation lies in the inherent difficulty of separating short-term geopolitical shocks from long-term structural trends. China’s ascent in the global collaboration networks began well before 2022. Although sanctions amplified the rate of reconfiguration, they were not its sole cause. Our methodological design allows us to trace transformations in co-authorship networks, but not to fully disentangle the share attributable to sanctions from other significant shifts in the system of global science. The clustering logic of VOSviewer adds its own subtleties: cluster boundaries depend on resolution parameters and threshold settings, and small adjustments may shift cluster composition at the margins. Finally, the presence of organizations with multiple or fragmented affiliations (multi-campus systems, hybrid institutes, or national labs) leads to a certain granularity that no bibliometric tool has yet fully resolved.

Taken together, these limitations do not reflect shortcomings in the approach, but rather the inherent constraints of any attempt to map a rapidly evolving, unevenly indexed, geopolitically influenced scientific landscape. Put differently, our goal was not to provide the final word on semiconductor collaboration, but to demonstrate a robust way of thinking about it — even if the global science system is moving faster than any of us (or any reviewer) might prefer.