Saudi Arabia Energy Transition in the Context of Scholarly Publications

The article is devoted to a bibliometric analysis of scientific publications reflecting trends in scientific publications related to the energy transition of leading universities in Saudi Arabia. The universities were selected based on the presence of joint publications with the Saudi Arabian Oil Company, they are: King Fahd University of Petroleum and Minerals, King Abdullah University of Science and Technology, and King Saud University. The importance of studying the experience of Saudi Arabia for Russia in the context of the energy transition was highlighted. The analysis was based on Scopus platform data related to Subject Areas: Physical Sciences. Scientists from Saudi Arabian universities are increasingly publishing with co-authors from China, India, Egypt, Pakistan, and Malaysia, while maintaining high interaction with co-authors from the United States, South Korea, and the United Kingdom. Authors from the Russian Federation publish poorly with coauthors from Saudi Arabia. The trends in the topics of scientific publications were investigated, and their connection with the renewable energy topic was shown. The VOSviewer software was used to determine the main research areas using the cluster analysis method based on the co-occurrence of key terms. The analysis was done for both Author's keywords and Index keywords of the Scopus system. Graphical representation of the final choice of terms to describe research trends was proposed.


Introduction
In the context of growing contacts between Russia and Saudi Arabia in the framework of OPEC+ decisions, it is reasonable to consider the priorities and trends in scientific research of this partner. 1 Since the 2010s, Russia has made significant progress in developing closer economic, political and security relations with the Gulf countries [1].
The initiative of national leaders to create a Charter for Cooperation among oil-producing countries will promote dialogue and cooperation among oil-producing countries at the ministerial and technical levels for the benefit of both oil-producing and consuming countries and the world economy. 2 The Gulf countries need to diversify their economies away from dependence on oil and gas, which involves creating a next-generation energy market and improving access to markets in China, India and other developing countries. [2]. This trend can be traced not only at the political level, but also at the level of research cooperation, which can be assessed by analyzing bibliometric data on scientific publications.
Fossil fuels are the main source of economic income in the Gulf countries. On the other hand, climate change is closely linked to the use of fossil energy, and this has motivated the Gulf countries to look for alternative solutions, such as renewable energy technologies to comply with the Paris Agreement [3].
Countries have unequal natural resources, economic power and capacity. Achieving the Sustainable Development Goals is about making the best use of a country's resources and capabilities [4].
On the WoS platform I get: 1,973 results for OG=(Saudi Aramco) Indexes=SCI-EXPANDED, ESCI Timespan=1975-2021 (In the classic WoS interface) Thus, the number of articles indexed in Scopus and WoS for the period 1975-2021 is almost the same. The advantage of Scopus is the large number of indexed conference proceedings and the greater number of Index Keywords compared to the Keywords Plus of the WoS platform.
To understand which topics are of most interest to Saudi-Aramco, let's analyze Index Keywords of the 2000 most relevant conference papers obtained by the query: AF-ID ("Saudi Arabian Oil Company" 60025307) AND PUBYEAR > 2011 AND (LIMIT-TO (DOCTYPE, "cp"). The predominant Index Keywords is shown in Tab. 1. Oil and gas companies stick to classic topics, even in conference proceedings, which reflect new topics more quickly than peer-reviewed articles.
Usually, universities develop new research areas. Therefore, it is useful to identify those that most often publish joint articles with Saudi Aramco, which can be done by analyzing the affiliation of the co-authors.
To identify such universities, I query the Scopus database with the following filters: years -2012-2021; language -English; article typejournal article; Subject Arears -Physical Sciences; affiliationthe Saudi Arabian Oil Company organization, the following affiliations for co-authors in the format Affiliation number of publications are obtained: For further analysis, we will limit it to the Arab universities. Below is a summary of four institutions (Saudi Arabian Oil Company, King Fahd University of Petroleum and Minerals, King Abdullah University of Science and Technology, and King Saud University), whose publication activities will be discussed later. A summary of the Subject Areas for each of the organizations is shown in Fig. 1.   1 shows that the first three organizations are mostly published in natural sciences disciplines, so the number of joint publications of King Fahd University of Petroleum and Minerals, King Abdullah University of Science with Technology and Saudi Arabian Oil Company is greater than that of King Saud University, which pays more attention to research in the field of medicine.
But King Saud University is great, and the total number of publications on the topic of interest to us is also large: Creation the main query to the abstract database Scopus, the data on which will be further exploited The assignment of a publication to a particular institution was determined by the affiliation ID in Scopus.
The main parameters of the publication activity of the given organizations, which correspond to the request, are presented in Fig. 2 and 3.  Scopus data -Analyze search results.
The main Subject Areas of research are: Chemistry, Engineering and Materials Science. Energy takes up a modest 4.5%. Computer Science (7.1%) and Mathematics (5.7%) suggest a significant role for digitalization in the scholarly research of these institutions.  The relative number of Chemistry publications declined between 2019 and 2021, with Engineering and Materials Science coming in first and second during this period.
The number of publications in Computer Science has been increasing since the second interval, while the number of publications in Mathematics has been decreasing.
Recently, energy topics have been attracting more and more attention. It can be assumed that such changes reflect the process of the energy transition, which requires new engineering solutions and materials (especially for renewable energy storage systems), as well as methods for optimizing operations in energy systems, which requires new solutions in the field of computer science.
While Tab. 2 shows the ranking of topics by publication activity in a particular time interval, Tab. 3 shows the distribution of publications by year for each of the Subject Arears. The following list reflects the increase in the relative number of publications in 2020 compared to 2012 for Subject Arears that I believe are particularly important for the energy transition: While Engineering and Materials Science have grown in absolute terms, Energy has seen the highest relative growth, followed by Computer Science, Chemical Engineering, and Environmental Sciences-the picture is not contradicted by the growing interest in the energy transition to clean energy.
For clarity, the data in Table 3 are shown graphically in Fig. 4. In order to catch trends in the development of research, it is crucial to know which countries are cooperating with, thus determining the potential for expanding the geographical scope of the topic. Table 4 shows the distribution by year of co-authorship for 10 countries. Data as of August 6, 2021.  Table 4 shows that prior to 2018, the majority of co-authors were from the U.S. Beginning in 2019, coauthors from China dominate, with an increase in the number of publications with U.S. co-authors continuing.
Authors from India, Egypt, Pakistan, and Malaysia are consistently in print along with authors from Saudi Arabia, a rate that increases significantly in 2020.
Publications with co-authors from South Korea and the United Kingdom steadily increasing over a whole ten-year period.
Authors from the Russian Federation publish poorly with co-authors from Saudi Arabia, even though there has been a relative increase over the years. The potential for cooperation between the Russian Federation and Saudi Arabia in the field of joint research remains unrealized, despite the extensive cooperation between the countries within the framework of OPEC+.

Using VOSviewer for Analysis of Scientific Publications Landscape
Bibliometric data of the 2,000 most-cited publications for each year were used to analyze trends in scientific research based on co-occurrence of key terms. Data as of September 09, 2021.

Why it makes sense to consider Author and Index keywords separately
Here is the Scopus definition for Author and Index keywords: • Author keywords: These are keywords chosen by the author(s) which, in their opinion, best reflect the contents of their document. To elaborate, here are two records:
The validity of this expansion of the list of Indexed Keywords requires a separate study, which is beyond the scope of this publication. In this case, we can only assume that the Indexed Keywords broadly reflect the potential fields of research to which a particular article may be assigned.

Analysis of Scientific Publications Landscape by Index Keywords for 20,000 Records
As noted earlier, the main query yielded metadata of 45,988 documents.
For constructing and visualizing Index keywords networks, 2,000 metadata records of the most cited publications were taken for each year (20,000 records in total). Then, using the VOSviewer software, a network was constructed using the specified metadata, based on the co-occurrence of Index keywords.
The total number of Index keywords in this sample was 79295, of which 12598 occur more than 5 times and 6318 occur more than 10 times; 1000 Index keywords out of 6318 with maximum link strength are further used to construct a term co-occurrence network.
After an initial review of the results, a partial replacement of the most frequent terms was performed, e.g., terms can occur in plural or singular form in different records, or have a hyphen between words or not, etc. After this substitution, the total number of terms was 79250 of which 6273 occurred more than 10 times. The condition more than 10 in cluster was used to enlarge the clusters; this yielded 4 clusters, which are shown in Figure 5. The following list shows the frequency data for each of the clusters in the format: number of unique terms/total number of terms:  The first cluster theme: • adsorption, catalysis, morphology, solar cells, thin films, heterojunctions • nanoparticle, carbon, nanocomposites, graphene, metal, chlorine compounds, zinc oxide, titanium dioxide, perovskite • scanning electron microscopy, x ray diffraction, transmission electron microscopy, fourier transform infrared spectroscopy The second cluster theme: • chemistry, metabolism, synthesis, drug synthesis, chemical structure, enzyme activity, oxidative stress, antiinfective agent, male, female • unclassified drug, animal, human cell, metal nanoparticles, antioxidant, escherichia coli • infrared spectroscopy, controlled study, in vitro study, animal experiment The third cluster theme: • optimization, energy efficiency, energy utilization, internet of things, solar energy, • carbon dioxide, polymer, water, membranes, silica, ethylene, fuel • kinetics, algorithm, thermodynamics, chemical analysis, performance assessment, forecasting The fourth cluster theme: • Saudi Arabia, wastewater treatment, sea water, waste water, water management • bacteria, copper, nitrogen, zinc, aluminum, lead, biomass, organic compound • mass spectrometry, Ph, concentration, isolation and purification, limit of detection Terms do not occur evenly over time; the simplest estimate of this is to use the average year of occurrence of the term in the metadata of publications, <Avg. pub. Year>, the results for some of the terms in the four clusters are shown in Table 6. This parameter will be used later in the construction of the final graphs: Fig. 15. and Fig. 16. To identify trends in the topics of publications described by keywords, in addition to the average year of appearance of the term, it is important to assess the citation rate of publications in which the term is found. Given that the citation rate of publications increases over time, it is advisable to use citation normalization, e.g., over a certain period of time. I used the parameter <Avg. norm. Citations>, which is available in the map files exported by VOSviewer. The results for the terms with the highest Avg. norm. Citations for the four clusters are shown in Table 7. The data given in Table 7 will be used to build the final graphs shown in Fig. 15 and Fig. 16.
The gradation of key terms can be very different, depending on the task at hand. For example, for a more detailed description of cluster topics, keywords can be assigned to different categories, conventionally labeled: research subject, materials, methods. In the lists below for each of the clusters, I have tried to intuitively divide the key terms into these categories. This is solely the author's categorization, given only to show that this division of terms can improve understanding of the topics of each of the clusters. To reduce biased judgments, a reasonable formalization of such categorization is necessary, which is beyond the scope of this paper, but can be accomplished by compiling for each of the categories its own vocabulary based on the occurrence of the terms in the corresponding sections of the full texts of the publications: introduction, materials, methods.
When selecting keywords to describe the cluster, it is important to consider what criteria were used to rank keywords. For example, in Table 5 they were ranked by occurrence, in Table 6 there was an additional ranking of frequently occurring keywords by <Avg.pub. Year>, and in Table 7 they were ranked by <Avg. norm. citation>. The leading keyword samples will be somewhat different, as will the cluster descriptions. At the same time, some commonality in the subject matter of the clusters is retained. This can be seen from the cluster descriptions in Table 5 below and the following cluster descriptions in Tables 7 and 6: 1 cluster, conditional division of key terms into categories: subject of research, materials, methods.

cluster
• water treatment, water purification, biotechnology, ecology, water supply, environmental impact, waste water management • sulfate, biochar, charcoal, nanotubes, inorganic compound, copper, potassium, lead • sorption, adsorption kinetics, pollutant removal, oxidation reduction reaction, waste component removal The most affordable method of viewing the change in occurrence of terms over time is to use the VOSviewer Overlay graph for the time parameter. An example is shown in Fig. 6.

Analysis of Scientific Publications Landscape by Author Keywords for 20,000 Records
The data processing procedure for Author Keywords was the same as for Index keywords, described in the previous section.
The total number of Author Keywords in this sample was 40369, which is significantly less than the Index Keywords -79295, of which 2455 occurred more than 5 times.
In the absence of cluster size restrictions, I got 16 clusters. Therefore, a limit of at least 120 terms per cluster was implemented, resulting in 6 clusters.
The connectivity of clusters for Author Keywords is significantly lower than for Index Keywords (Fig. 7.). There may be several reasons: a much smaller total number of Author Keywords compared to Index keywords, authors having their own opinion on the choice of keywords, for example, the terms: deep leaning, internet of things, and multi-step methods (green cluster) are more often used by authors than in Index keywords.   Fig. 7. Fig. 9 shows the evolution of the occurrence of terms over time using the VOSviewer Overlay graph for the time variable. Fig. 9. Variation of occurrence of Author's keywords over time Figure 9 shows that Author Keywords are likely to be indicative for terms that appear more frequently in new publications. I believe that it is the Author Keywords that are better suited for identifying emerging trends in scientific publications, and the Index Keywords are more interesting for identifying potential directions for expanding the application of results obtained in emerging trends.
VOSviewer allows us to detect not only the change of key terms over time, but also to show the distribution of citations of publications in which a particular term occurs. In my opinion, in the available publications on bibliometrics this feature is rarely used, but it serves as an indirect expert assessment of the topics described by the terms given in this graph (Fig. 10).

Fig. 10. Average Citation Rates of publications containing various Author keywords
The conjoint use of Figures 9 and 10 makes it possible to find keywords whose use is more common in new publications, and whose publications are highly cited. Figure 11 shows fragments of Figures 9 and 10 related to the term Visible light, which satisfies the above requirements. Fig. 11. The term Visible light. On the left is the occurrence over time, on the right the average citation rate of publications with the given keyword.
The figure shows that the term Visible light is used in the context of the keywords Photocatalysis and Nanocomposites. Figure 12 shows fragments of Figures 9 and 10 related to the term Chitosan, which satisfies only one requirement. Fig. 12. The term Chitosan is often found in new publications, but publications containing this keyword have an average citation rate.
As can be seen in Figures 11 and 12, using the average publication time of articles containing a given term (score_Avg_pub_year) and their average citation rate (score_Avg._citations) allows us to clearly assess possible research trends in scientific publications. The simplest formalization of the term selection process for such consideration may look as follows: we set a query to a file containing, for example, the Author's keywords and their score parameters. At the same time, we set restrictions on the threshold of occurrence of a key term, the average year of occurrence, and the average citation rate.
Let's look at an example result of such a query: select * from "Author KWs map csv" where "weight_Occurrences" > 10 AND "score_Avg_pub_year" > 2018 AND "score_Avg._citations" > 50. The term composite pcm (phase change material (PCM)) is found in the most recent publications (Avg_pub_year 2019.36). Phase change materials are used to store thermal energy, which is the second most important task after electrical energy storage. Examples of papers exposing this topic are [8,9] The disadvantage of the considered approach is the use of Avg._citations indicator, which increases over time, to assess the importance of key terms it is more appropriate to use a normalized estimate of the average citation (Avg. norm. citations). It is also necessary to propose a graphical representation that is convenient for the final assessment of the importance of terms for the description of research trends.

Graphical representation of the final choice of terms to describe research trends
The structure of scientific research landscapes well reflects the overall picture of publication activity, clustering based on the co-occurrence of key terms makes it possible to identify the main areas in scientific research. The use of an overlay makes it possible to trace changes in the use of terms over time and determine the citation rate of publications containing certain terms. When analyzing trends in scientific publications, VOSviewer allows you to browse flexibly through particular areas of the scientific research landscape, as shown in Figure 5. and 7. But when compiling an analytical report based on bibliometric data, the interactivity provided by analytical platforms and programs is lost; moreover, the report should provide summaries highlighting the main conclusions of the bibliometric study. Therefore, it is advisable to supplement the overall picture of publication activity, relevant to the main query to the abstract database, with illustrations reflecting the main results.
In my opinion, in order to identify trends in the topics of scientific publications, it is advisable to reflect the following data on the final graph: key terms and their occurrence over time; the average normalized citation rate of publications containing the selected key terms. The above parameters can be retrieved from the data export files provided by VOSviewer. In such files, they are stored in the record fields: label, weight<Occurrences>, score<Avg. pub. year> and score<Avg. norm. citations>.
Thus, the final graph should reflect the four parameters and be two-dimensional for readability. Bubble charts meet these requirements and are further used to present the final results in Figures 13. and 14.
Given that different time intervals are characterized by the use of different key terms, and that too many bubbles in the graph make it unreadable, the bibliometric data were divided into two periods -before 2017 and after 2017 (horizontal axis).
The second parameter I used to limit the choice of terms for the final graph was the average normalized citation (vertical axis), which was taken above 1.0. Next, the 15 most frequent key terms were selected. Each term was represented by its color, and the size of each term corresponded to its frequency of occurrence in the sample.
The final graphs for Author keywords and Index keywords are shown in Figures 13 -14   Water splitting, activated carbon and hydrogen production were the main terms in the publications and had an average normalized citation rate. Photocatalytic activity, taxonomy and phylogeny had the highest citation in the given time interval.
Examples of publications containing the Author keywords-Water Splitting: • Photocatalytic Water Splitting: Quantitative Approaches toward Photocatalyst by Design [10] • An Oxygen-Insensitive Hydrogen Evolution Catalyst Coated by a Molybdenum-Based Layer for Overall Water Splitting [11] • Enhanced Photoelectrochemical Solar Water Splitting Using a Platinum-Decorated CIGS/CdS/ZnO Photocathode [12] • Vertically aligned Ta3N5 nanorod arrays for solar-driven photoelectrochemical water splitting [13] • Tungsten carbide nanoparticles as efficient cocatalysts for photocatalytic overall water splitting [14] Definition: «In photoelectrochemical (PEC) water splitting, hydrogen is produced from water using sunlight and specialized semiconductors called photoelectrochemical materials, which use light energy to directly dissociate water molecules into hydrogen and oxygen» 5 Examples of publications containing Author keywordtaxonomy. The context for the term taxonomy is not obvious, but the titles of the articles disclose it:  [19] In Deep learning, Features extraction, anomaly detectiontaxonomy is a structuring and conceptualization of data in a particular subject area, which allows to reduce the dimensionality of the task and increase the stability of the analysis results. This time interval is most characterized by topics described by the keywords visible light, solar cells, perovskite, and biochar.
For identifying emerging trends in scientific publications, the terms Mxene and composite PCM are of most interest. The first is characterized by a high citation rate and the middle of the time interval, the second by an average citation rate, but the prevalence in newer publications.
Examples of publications containing the key term Mxene: • Versatile N-Doped MXene Ink for Printed Electrochemical Energy Storage Application [20] • Asymmetric Flexible MXene-Reduced Graphene Oxide Micro-Supercapacitor [21] • Large Dielectric Constant Enhancement in MXene Percolative Polymer Composites [22] • MXene Ti3C2: An Effective 2D Light-to-Heat Conversion Material [23] • MXene-on-Paper Coplanar Microsupercapacitors [24] Examples of publications containing the key term Composite PCM: • Silica fume/capric acid-palmitic acid composite phase change material doped with CNTs for thermal energy storage [25] • Preparation, characterization and thermal regulation performance of cement based-composite phase change material [26] • Thermal characteristics of expanded perlite/paraffin composite phase change material with enhanced thermal conductivity using carbon nanotubes [27] • Thermal energy storage characteristics of bentonite-based composite PCMs with enhanced thermal conductivity as novel thermal storage building materials [28]  In this case, the main topics are described with the words metal organic frameworks, electrochemistry, monolayers, decomposition and crystallization.
Sample publications: • A microporous metal-organic framework with naphthalene diimide groups for high methane storage [29] • A microporous metal-organic framework with basic sites for efficient C2H2/CO2 separation [30] • Effects of structural crystallinity and defects in microporous Al-MOF filled chitosan mixed matrix membranes for pervaporation of water/ethanol mixtures [31] • Light Hydrocarbon Adsorption Mechanisms in Two Calcium-Based Microporous Metal Organic Frameworks [32] Electron and electron transportterms that appear in the most cited articles.  Perovskite, perovskite solar cells and power conversion efficiencieshighly cited topics well described by these Index keywords.
Sample titles of publications that reflect well on perovskite solar cells: • Managing grains and interfaces via ligand anchoring enables 22.3%-efficiency inverted perovskite solar cells [36] • Single-Crystal MAPbI3 Perovskite Solar Cells Exceeding 21% Power Conversion Efficiency [37] • Inorganic CsPbI2Br Perovskite Solar Cells: The Progress and Perspective [38] • Phase Transition Control for High Performance Ruddlesden-Popper Perovskite Solar Cells [39] • Highly efficient perovskite solar cells based on a nanostructured WO3-TiO2 core-shell electron transporting material [40] Transition metals -it is advisable to disclose the use of this term in more detail, since its context is not obvious from the above chart. For this purpose, the publications in which the term Transition metals appears both in the Index keywords and in the titles of the articles have been picked.
Samples of article titles that give a good context for the term Transition metals: • Synthesis and characterization of binary selenides of transition metals to investigate its photocatalytic, antimicrobial and anticancer efficacy [41] • Heteroatom-doped magnetic hydrochar to remove post-transition and transition metals from water: Synthesis, characterization, and adsorption studies [42] • Unraveling the role of entropy in tuning unimolecular vs. bimolecular reaction rates: The case of olefin polymerization catalyzed by transition metals [43] • Complexation of trichlorosalicylic acid with alkaline and first row transition metals as a switch for their antibacterial activity [44] • Molecule-Level g-C3N4 Coordinated Transition Metals as a New Class of Electrocatalysts for Oxygen Electrode Reactions [45] • Liquid-phase exfoliated MoS2nanosheets doped with: P -type transition metals: A comparative analysis of photocatalytic and antimicrobial potential combined with density functional theory [46] From the list above we can see that the context for transition metals is the terms photocatalytic, antimicrobial and anticancer efficacy, olefin polymerization, and remove post-transition and transition metals from water which agrees well with the term photo catalytic degradation in Fig. 16.

Conclusions
A bibliometric analysis of scientific publications showing trends in scientific publications from leading universities in Saudi Arabia related to the energy transition was conducted.
The following universities were selected based on the presence of joint publications with the Saudi Arabian Oil Company they are: King Fahd University of Petroleum and Minerals, King Abdullah University of Science and Technology, and King Saud University.
According to data from the Scopus platform related to the subject areas: Physical Sciences, scientists from Saudi Arabian universities are increasingly publishing with co-authors from China, India, Egypt, Pakistan, and Malaysia, while maintaining a high level of interaction with co-authors from the US, South Korea, and the UK. Authors from the Russian Federation publish poorly with co-authors from Saudi Arabia.
A significant difference between the Scopus platform Index keywords and Author keywords has been shown, which should be taken into account when identifying trends in scientific publications.
The main subject areas of research are: Chemistry, Engineering, and Materials Science. Energy occupies a modest 4.5%. Computer Science (7.1%) and Mathematics (5.7%) indicate a significant role for digitalization in research at these institutions.
In recent years, the relative number of chemistry publications has declined, while engineering and materials science has taken the lead.
There has been a relative increase in attention to energy topics, reflecting the energy transition process, which requires new engineering solutions and materials (especially for renewable energy storage systems), as well as methods for optimizing performance in energy systems, which requires new computer science solutions.
Based on the co-occurrence of Index keywords, four sub-themes are identified, which can be briefly described by the terms: • adsorption, catalysis, morphology, solar cells, thin films, heterojunctions, nanoparticle, carbon, nanocomposites, graphene, metal, chlorine compounds, zinc oxide, titanium dioxide, perovskite • chemistry, metabolism, synthesis, drug synthesis, chemical structure, enzyme activity, oxidative stress, anti-infective agent, unclassified drug, animal, human cell, metal nanoparticles, antioxidant • optimization, energy efficiency, energy use, internet of things, solar energy, carbon dioxide, polymer, water, membranes, silica, ethylene, fuel • Saudi Arabia, wastewater treatment, seawater, wastewater, water management, bacteria, copper, nitrogen, zinc, aluminum, lead, biomass, organic compound Shows the feasibility of using additional keyword ranking for <Avg.pub. Year> and <Avg. norm. citation> to identify promising areas of research.
The use of bubble charts to present the final results of bibliometric studies to identify emerging research topics has been proposed.
Based on the analysis, two areas of research that are advisable to study in more detail are: composite phase change materials (PCM) and MXenes -carbides and nitrides of transition metals, a fast-growing and already very large family of 2D materials.
In methodological terms, it is advisable to formalize in more detail the approaches proposed in the article to the graphical representation of the final data on the identification of promising areas of research and additional separation of keywords in the clusters by the categories Subject, Materials, and Methods.