Artificial Intelligence Reinventing Materials Engineering: A Bibliometric Review
1
Technology, Instruction and Design in Engineering and Education Research Group (TiDEE.rg), Catholic University of Ávila, C/Canteros s/n, 05005 Ávila, Spain
2
Department of Applied Informatics, University of Macedonia, 54636 Thessaloniki, Greece
3
Department of Education, University of Nicosia, 1700 Nicosia, Cyprus
*
Author to whom correspondence should be addressed.
Appl. Sci. 2024, 14(18), 8143; https://doi.org/10.3390/app14188143
Submission received: 19 July 2024
/
Revised: 15 August 2024
/
Accepted: 20 August 2024
/
Published: 10 September 2024
Abstract
:The use of artificial intelligence (AI) is revolutionizing many professions and research fields. Thus, the present study focuses on the implications that AI is having on research in materials science and engineering (MSE). To this end, a bibliometric review has been conducted to analyze the advances that AI is generating in MSE. Although expectations for AI advances in the field of MSE are high, the results of this study indicate that we are still at a preliminary stage of development. It is worth highlighting that despite the progress made, the potential of AI in MSE has not been fully exploited and numerous challenges remain to be overcome to achieve effective and widespread implementation. It should be noted that the subarea “Materials structure, processing, and properties” is the one that currently presents the largest number of research works linked to AI. It appears that the United States and China are currently the countries with the greatest involvement in the use of AI in the field of MSE. The emerging themes and thematic map of the topic are revealed, and future research directions are provided.
1. Introduction
Materials science and engineering (MSE) is a domain of knowledge characterized by high scientific production. This field encompasses a multiplicity of aspects, which has led the Materials Research Society to developing a classification that segregates this field of research into several specific subareas [1]. In the current landscape of scientific research, there is a marked trend towards the application of artificial intelligence (AI) in various disciplines. In this context, there is a significant opportunity to integrate AI into the field of MSE, thus enabling improved research and development capabilities within this field. To date, many MSE studies are based on experimentally obtained data and calculations are performed using specific computer programs. However, these calculations and the management of these data can now be optimized through the application of AI. This transformation has the potential to revolutionize research in the field of MSE. This study seeks to analyze recent advances in the integration of AI in MSE, highlighting the implications and benefits that this synergy could bring to the development and innovation of this discipline.
Therefore, the main aim of this study is to provide an overview regarding the use of AI in materials science through a bibliometric review. The main contributions of this study are that it offers a representation and analysis of the existing literature, reveals the most developed research lines, identifies strengths and weaknesses of the literature, and provides future research directions. The structure of this study is organized as follows: The materials and methods used are described in Section 2 and the results are analyzed and presented in Section 3. In Section 4, the outcomes are further discussed and in Section 5 implications and future research directions are presented.
2. Materials and Methods
To provide an overview of the state of the art of a specific topic, an approach such as bibliometric analysis is deemed appropriate [2]. Additionally, to ensure the integrity and validity of a bibliometric analysis, specific guidelines and approaches should be followed [3,4]. In this study, we adopted the guidelines presented in [4,5] and used the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) protocol to increase reproducibility and accuracy of the documents identified and selected [6]. To carry out this study, the open-source R (v. 4.3.3) package “Bibliometrix” (v. 4.2.1) was utilized [5]. To identify documents relevant to the topic, the Scopus and Web of Science (WoS) databases were used due to their high impact, the large number of scientific documents from established sources that they contain, and the ability to extract information that can be used by the Bibliometrix tool [7,8]. To identify relevant documents, a detailed query was used: (“artificial intelligence” OR “ai”) AND (“materials science” OR “materials engineer” OR “materials science and engineering” OR “materials characterization” OR “materials test*”). The query used keywords related to AI and material sciences. As the field of AI has undergone drastic advancements in the last few years, only English documents from 2019 to 2023 were searched for at a topic level (e.g., title, abstract, and keywords). As 2024 was still ongoing during the performance of this study, the documents published in 2024 were not included as, by missing data from a whole semester, the outcomes for 2024 would not have been valid. The detailed document processing flowchart is presented in Figure 1.
Specifically, 889 documents (Scopus: 603 and WoS: 286) that are relevant to the topic were identified in May 2024. The removal of duplicates was performed both automatically through Bibliometrix as well as manually, and 237 duplicates were found and removed. Hence, the initial screening for eligibility involved 652 documents. Some documents (Reasons 2–7) were removed due to their being proceedings books (10), editorials (10), retracted documents (3), errata (1), and letters (1). Additionally, using the keywords and automatically searching within the titles and abstracts of the documents, 209 documents were removed as they were out of the scope of this study and did not focus on the use of AI in the greater context of materials science. Thereafter, the remaining 418 documents were manually processed and assessed for eligibility through the examination of their full text. The inclusion criteria were for a document to focus on, examine, or use AI or its subfields in materials science, materials engineering, material characterization, or materials testing. Through this process, 148 documents were removed as they did not meet the inclusion criteria. As a result, the document collection examined and analyzed in this study consisted of 270 documents.
It is worth noting that Scopus and WoS are considered two of the most impactful databases and most appropriate to conduct bibliometric analyses [7,8], and their indexed documents are highly relevant to the topic. Nonetheless, the use of only two databases can be regarded as a limitation of this study. Additionally, although bibliometric analysis enables us to present a descriptive representation of the current state of the art, it does not allow for an in-depth analysis of each document. To address this, we analyzed the most impactful documents arising from the analysis. Biases can also be introduced in such studies. To counter that, we opted to use the most widely used databases which contain the documents most relevant to the topic, we selected to use a tool widely utilized and specialized for this task, and, in addition, we included all types of documents. Therefore, the metrics presented in this study derive directly from the datasets generated from the two databases.
3. Results
The descriptive statistics of the document collection examined in this study are presented in Table 1. Specifically, the document collection consisted of 270 documents which were published from 2019 to 2023 in 208 sources. Despite the documents examined being published in the last 5 years, the average age of the documents was 2.27, which in combination with the high annual growth rate of 48.02% highlights the importance of, and interest in, the use of AI in materials science. This increase is in line with the advancement of the field of AI. A total of 1350 authors from 50 countries contributed to the documents of the collection examined. Each document received 28.16 citations on average and was created by 5.86 co-authors on average. Only 20 documents (7.41%) were created by a single author. Additionally, collaborations among stakeholders from different countries and continents were observed to have an international co-authorship rate of 14.07%. The majority of documents (49.3%) were published as research articles in journals, while a large number of documents (34.4%) were review studies. In total, 35 documents (13.0%) were published in conferences/proceedings and 9 documents (3.3%) were published as book chapters.
As the field of AI is advancing, its adoption and use in different domains is increasing. The high annual growth rate of 48.02% highlights that the use of AI to enrich and transform materials science is no exception to this trend. This fact can be further validated by the annual scientific production presented in Figure 2 and Table 2. Specifically, the number of documents being published that are relevant to the topic is increasing annually, something that is expected to continue in the future. Most of the documents were published in 2023 (freq.: 96, pct.: 35.6%), followed by 2022 (freq.: 74, pct.: 27.4%), which is a drastic increase when considering that only 20 (7.4%) relevant documents were published in 2019. These findings are in line with the increasing interest in AI and its being widely adopted and examined. Moreover, the documents from 2019 followed by those published in 2020 have the largest mean total number of citations per document (MeanTCperArt), with 70.6 and 61.97, respectively. When examining the average total number of citations received per year (MeanTCperYear), which reveals the impact and relevance of a document over time, the documents published in 2020 (12.39 MeanTCperYear) present the highest MeanTCperYear, followed by those published in 2019 (11.77 MeanTCperYear).
Furthermore, the sources in which the documents were published were examined using Bradford’s law, which “estimates the exponentially diminishing returns of searching for references in science journals” [5,9]. In particular, the documents were classified into three clusters, with Cluster 1 consisting of the most impactful sources. Cluster 1 was composed of 31 sources (14.9%) in which 90 documents (33.3%) were published. Cluster 2 consisted of 88 sources (42.3%) in which 91 documents (33.7%) were published. Cluster 3 was composed of 89 sources (42.8%) in which 89 documents (33.0%) were published. The most impactful sources based on Bradford’s law are presented in Table 3. Specifically, ranks 1–6 of Cluster 1 belonged to the following sources: Advanced Materials (number of published documents: 7 and impact factor: 27.4), MRS Bulletin (number of published documents: 5 and impact factor: 4.1), Nanomaterials (number of published documents: 5 and impact factor: 4.4), Materials (number of published documents: 4 and impact factor: 3.1), MRS Communications (number of published documents: 4 and impact factor: 16.8), and WIREs Computational Molecular Science (number of published documents: 4 and impact factor: 1.8). The remaining sources had fewer than three relevant documents published. Table 4 depicts the most impactful sources based on their published documents’ h-index and total number of citations on the topic. It is worth noting that only three sources (Advanced Materials, MRS Communications, and MRS Bulletin) started publishing documents on this topic in 2019.
The most impactful documents that greatly contribute to the development of this topic can be identified by examining the total number of citations, the total number of citations per year, and the normalized total number of citations received globally. The related outcomes are presented in Table 5. Focusing on the total number of citations, it can be observed that the top 10 documents have all received over 100 citations, with most of them having over 200 citations. The study of [10] (382 citations) has received most citations, followed by the study of [11] (322 citations). When focusing on the total number of citations received per year to identify the document’s annual influence, the study of [10] (76.4 citations received per year) and the study of [12] (71.75 citations received per year) were the only ones that received over 70 citations per year. When considering the normalized total number of citations which eliminates temporal biases, thus allowing for a fairer comparison between older and recent studies, the studies of [11] (6.59) and [10] (6.16 normalized total citations) are the most influential ones, followed by the studies of [13] (5.23 normalized total citations), [11] (5.2 normalized total citations), and [14] (4.89 normalized total citations).
Moreover, when taking the corresponding author’s country into account, the countries whose authors focus on this topic can be identified. Specifically, when considering the number of documents published, the United States (74 documents), China (57 documents), Germany (14 documents), India (11 documents), and South Korea (11 documents) emerge as the countries whose authors have contributed the most documents related to the use of AI in materials science. The related information is presented in Table 6. This may be due to the fact that the United States and China are among those that invest the most in research and innovation worldwide (according to data from the Global Innovation Index 2023 Ranking [20]) and that, furthermore, both countries lead the world in AI-research investment strategy [21].
Focusing on the total number of citations received, the countries whose authors have contributed the most impactful documents can be identified as shown in Table 7. In particular, the United States (2582 citations), China (1582 citations), Germany (685 citations), South Korea (381 citations), and Canada (319 citations) are the countries whose authors have received the most citations. However, it should be noted that authors from countries such as Austria and Portugal have contributed a single but impactful document. Hence, the average number of citations received per document should also be considered. Furthermore, considering the international co-authorship rate (14.07%) and the fact that authors from different countries and continents are actively collaborating, the country collaboration network is displayed in Figure 3.
To obtain a better understanding of the thematic map of the topic, the keywords, both keyword-plus and author keywords, were also examined and are presented in Table 8. Based on the outcomes, the role of AI and its subfields (e.g., machine learning, deep learning, neural networks) in materials science is highlighted. This role becomes more evident when considering the keyword co-occurrence network presented in Figure 4. Specifically, three main clusters arose. Cluster 1 (blue color) refers to machine learning and deep learning and includes terms such as learning algorithms, data mining, forecasting, etc., which are being more widely used in materials science. Cluster 2 (red color) highlights the role of AI and its versatile nature, which enables it to be used for different purposes and in different domains (e.g., drug discovery, simulations, automation, industrial research, etc.). Cluster 3 (green color) highlights the key aspects which the integration of AI in materials science can affect, such as prediction, optimization, design, discovery, generation, etc. Some smaller clusters related to different types of neural networks and material properties are also noticed but are less impactful overall. The keywords were also used to examine the trend topics presented in Figure 5. Even during the five-year period (2019–2023) explored in this study, some trends emerged. Specifically, the focus shifted to automation, material discovery, and machines during 2020–2022, and thereafter (2022–2023) more emphasis was placed on machine learning, natural language processing, graph neural networks, materials science, and properties. The focus on artificial intelligence and its role in materials science has become more obvious since 2021. Additionally, since 2020 emphasis has been placed on drug discovery.
When clustering the documents using document coupling and focusing on the keywords, six main clusters emerged which can be seen in Figure 6. Specifically, the first cluster was related toin-vitro, qsar, and bacteria-driven microswimmers, the second cluster was associated with neutral networks, deep learning, and deep neural networks, and the third cluster was related to artificial intelligence, materials science, and artificial neural networks. The remaining three clusters revealed a higher impact and were related to density-functional theory, materials informatics, and scaling relations (cluster 4), design, discovery, and optimization (cluster 5), and machine learning, deep learning, and artificial intelligence (cluster 6). Furthermore, focusing on the use of keywords, the thematic map revealed eight themes as can be seen in Figure 7. Four themes were characterized as emerging or declining themes and refer to (i) memory, dynamics, and networks; (ii) in-vitro, density, and hydroxyapatite; (iii) differential phase contrast; and iv) irradiation, whereas the remaining four themes were characterized as motor themes and refer to (i) AI, natural language processing, and chemistry; (ii) machine learning, materials science, and deep learning; (iii) industrial research, additives, and biomedical applications; and (iv) design, neural networks, and optimization. The thematic evolution was also examined focusing on the periods of 2019–2021 and 2022–2023. The specific outcomes can be seen in Figure 8. The greater emphasis on artificial intelligence and related technologies as well as the focus on specific aspects of materials science during 2022–2023 were observed.
4. Discussion
Based on the findings, it can be concluded that although AI can significantly enrich and transform materials science, its integration in materials science is still in its early stages. Nonetheless, given the significant annual growth rate (48.02%) of related documents being published in recent years (2019–2023), and with the majority having been published in the last two years (63.0%), the importance of the topic is highlighted and is expected to be further advanced in the near future. Sources of various types, such as journals, conferences, and books, have been utilized to publish relevant documents, with 31 sources (14.9%) being regarded as highly impactful. Authors from 40 countries have contributed to the creation of the documents examined. Several international collaborations, even across continents, emerged. This fact highlights the recency and significance of the topic. Authors from the United States, China, Germany, and South Korea have both the largest number of published documents on the topic and the highest number of citations received.
Furthermore, to obtain a better understanding about which categories of materials science AI is most applied in, the existence of specific keywords within the title, abstract, author keywords, and keywords-plus of each of the 270 documents was examined. It should be mentioned that due to the nature of the topic, a document can belong to more than one category. The eight areas of materials science used were presented in Extremera et al. [22] and are based on the classification provided by the Materials Research Society [1]. The outcomes are summarized in Table 9. Specifically, the materials science area in which AI is most being applied in is “Materials structure, processing, and properties” (70.0%). Particular emphasis is also placed on its use in “Electronics, optics, and quantum” (29.3%) as well as in “Materials computing and data science” (29.3%). The use of AI in “Structural and functional materials” (25.9%), “Material characterization” (24.1%), “Carbon-based nanocomposite materials and applications” (22.2%), and “Energy and sustainability” (21.5%) is also being examined. Finally, the use of AI in the area of “Biomaterials and soft materials” (12.2%) is examined in the context of materials science to a lesser extent. These results provide a clearer, albeit estimated, representation of the areas of materials science where AI is most being examined and applied, which is consistent with previously presented results on virtual reality (VR) applications in Extremera et al. [22].
As an estimate, it can be seen in Figure 9 that there are fields in which many applications of AI are being found, especially in engineering. Similarly, it can also be seen that in the field of MSE there is hardly any significant progress, so this shows the potential that AI still has in this field and the margin for research that exists. In spite of this, there are recent studies that predict a strong development of this sector in the coming years. Thus, one can highlight the work of Choudhary et al. [23], which shows the potential of deep learning in working with images and spectral data of large da-tabs materials; Liang et al. [24] assemble machine learning models to predict the creep behavior of concrete; in this field of mechanical properties, Ni and Gao [25] apply deep learning to identify elastic modulus, thus providing a less costly method than traditional non-destructive evaluation (NDE) techniques; Kennedy et al. [26] evidence the application of AI in the research and development of applications related to both acoustics and mechanics; Morgan and Jacobs [27] highlight the importance of creating open-source software packages to advance the implementation of AI in MSE; and even in the paper by Hanxun et al. [28], new concepts for materials which can respond autonomously in real time to certain external inputs (decision-making materials) are planned; on the other hand, Orosa et al. [29] trained neural networks to predict interior environments from the design of interior covering materials, thus designing an original methodology to optimize these environments.
5. Conclusions
According to the results found in this study, it can be stated that the use of AI in MSE is still in an incipient phase, so that, although there seems to be a growing tendency to advance in this line of work, this area of research still has great potential for exploitation. Among all the subareas of MSE, it is precisely the “Structure, processing and properties of materials” in which, so far, the greatest applications of AI seem to have been found, which coincides with the results found in previous studies linked to the use of VR. In this sense, there seems to be a greater tendency towards innovation in this subarea than in others. Moreover, the results suggest that it is in the United States and China where the greatest efforts are being made to find AI applications in MSE.
Author Contributions
Conceptualization, D.V.; methodology, D.V. and G.L.; software, G.L.; validation, D.V., G.L. and P.F.-A.; formal analysis, D.V., G.L., P.F.-A. and Á.A.-S.; investigation, D.V., G.L. and P.F.-A.; resources, D.V. and G.L.; data curation, G.L.; writing—original draft preparation, D.V., G.L., P.F.-A. and Á.A.-S.; writing—review and editing, D.V., G.L., P.F.-A. and Á.A.-S.; supervision, D.V., G.L. and P.F.-A. All authors have read and agreed to the published version of the manuscript.
Funding
This research received no external funding.
Institutional Review Board Statement
Not applicable.
Informed Consent Statement
Not applicable.
Data Availability Statement
The dataset created and examined is available from the authors upon reasonable request.
Conflicts of Interest
The authors declare no conflicts of interest.
References
- Materials Research Society—2021 MRS Fall Meeting & Exhibit. Available online: https://www.mrs.org/past-fall-meetings/2021-mrs-fall-meeting (accessed on 2 June 2024).
- Ellegaard, O.; Wallin, J.A. The bibliometric analysis of scholarly production: How great is the impact? Scientometrics 2015, 105, 1809–1831. [Google Scholar] [CrossRef] [PubMed]
- Gusenbauer, M.; Haddaway, N.R. Which academic search systems are suitable for systematic reviews or meta-analyses? Evaluating retrieval qualities of google scholar, PubMed, and 26 other resources. Res. Synth. Methods 2020, 11, 181–217. [Google Scholar] [CrossRef]
- Donthu, N.; Kumar, S.; Mukherjee, D.; Pandey, N.; Lim, W.M. How to conduct a bibliometric analysis: An overview and guidelines. J. Bus. Res. 2021, 133, 285–296. [Google Scholar] [CrossRef]
- Aria, M.; Cuccurullo, C. Bibliometrix: An r-tool for comprehensive science mapping analysis. J. Informetr. 2017, 11, 959–975. [Google Scholar] [CrossRef]
- Page, M.J.; McKenzie, J.E.; Bossuyt, P.M.; Boutron, I.; Hoffmann, T.C.; Mulrow, C.D.; Shamseer, L.; Tetzlaff, J.M.; Akl, E.A.; Brennan, S.E.; et al. The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. Int. J. Surg. 2021, 88, 105906. [Google Scholar] [CrossRef] [PubMed]
- Mongeon, P.; Paul-Hus, A. The journal coverage of web of science and scopus: A comparative analysis. Scientometrics 2015, 106, 213–228. [Google Scholar] [CrossRef]
- Zhu, J.; Liu, W. A tale of two databases: The use of web of science and scopus in academic papers. Scientometrics 2020, 123, 321–335. [Google Scholar] [CrossRef]
- Bradford, S.C. Sources of information on specific subjects. Engineering 1934, 137, 85–86. [Google Scholar]
- Muratov, E.N.; Bajorath, J.; Sheridan, R.P.; Tetko, I.V.; Filimonov, D.; Poroikov, V.; Oprea, T.I.; Baskin, I.I.; Varnek, A.; Roitberg, A.; et al. QSAR without borders. Chem. Soc. Rev. 2020, 49, 3525–3564. [Google Scholar] [CrossRef] [PubMed]
- So, S.; Badloe, T.; Noh, J.; Bravo-Abad, J.; Rho, J. Deep learning enabled inverse design in nanophotonics. Nanophotonics 2020, 9, 1041–1057. [Google Scholar] [CrossRef]
- Xu, Y.; Liu, X.; Cao, X.; Huang, C.; Liu, E.; Qian, S.; Liu, X.; Wu, Y.; Dong, F.; Qiu, C.-W.; et al. Artificial intelligence: A powerful paradigm for scientific research. Innovation 2021, 2, 100179. [Google Scholar] [CrossRef]
- Guo, K.; Yang, Z.; Yu, C.-H.; Buehler, M.J. Artificial intelligence and machine learning in design of mechanical materials. Mater. Horiz. 2021, 8, 1153–1172. [Google Scholar] [CrossRef] [PubMed]
- Westermayr, J.; Marquetand, P. Machine learning for electronically excited states of molecules. Chem. Rev. 2021, 121, 9873–9926. [Google Scholar] [CrossRef] [PubMed]
- Schlexer Lamoureux, P.; Winther, K.T.; Garrido Torres, J.A.; Streibel, V.; Zhao, M.; Bajdich, M.; Abild-Pedersen, F.; Bligaard, T. Machine learning for computational heterogeneous catalysis. ChemCatChem 2019, 11, 3581–3601. [Google Scholar] [CrossRef]
- Draxl, C.; Scheffler, M. The NOMAD laboratory: From data sharing to artificial intelligence. J. Phys. Mater. 2019, 2, 036001. [Google Scholar] [CrossRef]
- Yang, J.-Q.; Wang, R.; Ren, Y.; Mao, J.-Y.; Wang, Z.-P.; Zhou, Y.; Han, S. Neuromorphic engineering: From biological to Spike-Based hardware nervous systems. Adv. Mater. 2020, 32, 2003610. [Google Scholar] [CrossRef] [PubMed]
- Pollice, R.; dos Passos Gomes, G.; Aldeghi, M.; Hickman, R.J.; Krenn, M.; Lavigne, C.; Lindner-D’Addario, M.; Nigam, A.; Ser, C.T.; Yao, Z.; et al. Data-Driven strategies for accelerated materials design. Acc. Chem. Res. 2021, 54, 849–860. [Google Scholar] [CrossRef] [PubMed]
- De Almeida, A.F.; Moreira, R.; Rodrigues, T. Synthetic organic chemistry driven by artificial intelligence. Nat. Rev. Chem. 2019, 3, 589–604. [Google Scholar] [CrossRef]
- Global Innovation Index 2023 Ranking. Available online: https://www.statista.com/statistics/1102558/most-innovative-countries-gii-score/ (accessed on 14 August 2024).
- World Economic Forum. Available online: https://www.weforum.org/agenda/2024/05/these-5-countries-are-leading-the-global-ai-race-heres-how-theyre-doing-it/ (accessed on 14 August 2024).
- Extremera, J.; Vergara, D.; Rodríguez, S.; Dávila, L.P. Reality-Virtuality Technologies in the Field of Materials Science and Engineering. Appl. Sci. 2022, 12, 4968. [Google Scholar] [CrossRef]
- Choudhary, K.; DeCost, B.; Chen, C.; Jain, A.; Tavazza, F.; Cohn, R.; Park, C.W.; Choudhary, A.; Agrawal, A.; Billinge, S.J.; et al. Recent advances and applications of deep learning methods in materials science. npj Comput. Mater. 2022, 8, 59. [Google Scholar] [CrossRef]
- Liang, M.; Chang, Z.; Wan, Z.; Gan, Y.; Schlangen, E.; Šavija, B. Interpretable Ensemble-Machine-Learning Models for Predicting Creep Behavior of Concrete. Cem. Concr. Compos. 2022, 125, 104295. [Google Scholar] [CrossRef]
- Ni, B.; Gao, H.J. A Deep Learning Approach to the Inverse Problem of Modulus Identification in Elasticity. MRS Bull. 2021, 46, 19–25. [Google Scholar] [CrossRef]
- Kennedy, J.; Lim, C.W.; Gulzari, M. Machine Learning and Deep Learning in Phononic Crystals and Metamaterials—A Review. Mater. Today Commun. 2022, 33, 104606. [Google Scholar] [CrossRef]
- Morgan, D.; Ryan, J. Opportunities and challenges for machine learning in materials science. Annu. Rev. Mater. Res. 2020, 50, 71–103. [Google Scholar] [CrossRef]
- Jin, H.; Zhang, E.; Espinosa, H.D. Recent advances and applications of machine learning in experimental solid mechanics: A review. Appl. Mech. Rev. 2023, 75, 061001. [Google Scholar] [CrossRef]
- Orosa, J.A.; Vergara, D.; Costa, Á.M.; Bouzón, R. A novel method based on neural networks for designing internal coverings in buildings: Energy saving and thermal comfort. Appl. Sci. 2019, 9, 2140. [Google Scholar] [CrossRef]
Figure 1.
Document processing flowchart.
Figure 2.
Annual scientific production.
Figure 3.
Country collaboration network.
Figure 4.
Keyword-plus co-occurrence network.
Figure 5.
Trend topics.
Figure 6.
Documents coupling.
Figure 7.
Thematic map.
Figure 8.
Thematic evolution.
Figure 9.
Number of manuscripts published in Scopus by using the search string indicated in the figure legend.
Figure 9.
Number of manuscripts published in Scopus by using the search string indicated in the figure legend.
Table 1.
Document collection information.
| Description | Results | Description | Results |
|---|---|---|---|
| Main information about data | Document types | ||
| Timespan | 2019:2023 | Article | 133 |
| Sources (journals, books, etc.) | 208 | Βook chapter | 9 |
| Documents | 270 | Conference/Proceedings paper | 35 |
| Annual growth rate % | 48.02 | Review | 93 |
| Document average age | 2.27 | Authors | |
| Average citations per doc | 28.16 | Authors | 1350 |
| References | 14,202 | Authors of single-authored docs | 19 |
| Document contents | Author collaboration | ||
| Keywords plus (ID) | 1735 | Single-authored docs | 20 |
| Author’s keywords (DE) | 825 | Co-authors per doc | 5.86 |
| International co-authorships % | 14.07 |
Table 2.
Annual scientific production and citations.
| Year | MeanTCperArt | No. of Documents | MeanTCperYear | CitableYears |
|---|---|---|---|---|
| 2019 | 70.6 | 20 | 11.77 | 6 |
| 2020 | 61.97 | 30 | 12.39 | 5 |
| 2021 | 43.56 | 50 | 10.89 | 4 |
| 2022 | 20.89 | 74 | 6.96 | 3 |
| 2023 | 6.33 | 96 | 3.16 | 2 |
Table 3.
Most impactful sources based on Bradford’s law.
| Source | Rank | Freq | cumFreq | Cluster |
|---|---|---|---|---|
| Advanced Materials | 1 | 7 | 7 | Cluster 1 |
| MRS Bulletin | 2 | 5 | 12 | Cluster 1 |
| Nanomaterials | 3 | 5 | 17 | Cluster 1 |
| Materials | 4 | 4 | 21 | Cluster 1 |
| MRS Communications | 5 | 4 | 25 | Cluster 1 |
| WIREs Computational Molecular Science | 6 | 4 | 29 | Cluster 1 |
| Acta Materialia | 7 | 3 | 32 | Cluster 1 |
| Advanced Energy Materials | 8 | 3 | 35 | Cluster 1 |
| Advanced Intelligent Systems | 9 | 3 | 38 | Cluster 1 |
| Applied Sciences (Switzerland) | 10 | 3 | 41 | Cluster 1 |
Table 4.
Most impactful sources based on their published documents’ h-index and total number of citations on this topic.
Table 4.
Most impactful sources based on their published documents’ h-index and total number of citations on this topic.
| Sources | h-Index | g-Index | m-Index | TC | NP | PY-Start |
|---|---|---|---|---|---|---|
| Advanced Materials | 5 | 7 | 0.833 | 393 | 7 | 2019 |
| Nanomaterials | 4 | 5 | 1 | 59 | 5 | 2021 |
| Materials | 4 | 4 | 1.333 | 41 | 4 | 2022 |
| npj Computational Materials | 3 | 3 | 1 | 161 | 3 | 2022 |
| Advanced Intelligent Systems | 3 | 3 | 0.6 | 148 | 3 | 2020 |
| MRS Communications | 3 | 4 | 0.5 | 144 | 4 | 2019 |
| Wiley Interdisciplinary Reviews: Computational Molecular Science | 3 | 4 | 0.6 | 114 | 4 | 2020 |
| MRS Bulletin | 3 | 5 | 0.5 | 113 | 5 | 2019 |
| Journal of Applied Physics | 3 | 3 | 0.6 | 110 | 3 | 2020 |
| Advanced Energy Materials | 3 | 3 | 1 | 99 | 3 | 2022 |
Table 5.
Most impactful documents based on the total citations received.
| Document | DOI | Total Citations | Total Citations per Year | Normalized Total Citations |
|---|---|---|---|---|
| [10] | https://doi.org/10.1039/d0cs00098a | 382 | 76.4 | 6.16 |
| [11] | https://doi.org/10.1515/nanoph-2019-0474 | 322 | 64.4 | 5.2 |
| [12] | https://doi.org/10.1016/j.xinn.2021.100179 | 287 | 71.75 | 6.59 |
| [13] | https://doi.org/10.1039/d0mh01451f | 228 | 57 | 5.23 |
| [15] | https://doi.org/10.1002/cctc.201900595 | 214 | 35.67 | 3.03 |
| [14] | https://doi.org/10.1021/acs.chemrev.0c00749 | 213 | 53.25 | 4.89 |
| [16] | https://doi.org/10.1088/2515-7639/ab13bb | 200 | 33.33 | 2.83 |
| [17] | https://doi.org/10.1002/adma.202003610 | 185 | 37 | 2.99 |
| [18] | https://doi.org/10.1021/acs.accounts.0c00785 | 180 | 45 | 4.13 |
| [19] | https://doi.org/10.1038/s41570-019-0124-0 | 156 | 26 | 2.21 |
Table 6.
Top 10 countries based on the number of documents published on the topic.
| Country | Articles | SCP | MCP | Freq | MCP Ratio |
|---|---|---|---|---|---|
| United States | 74 | 63 | 11 | 0.274 | 0.149 |
| China | 57 | 51 | 6 | 0.211 | 0.105 |
| Germany | 14 | 12 | 2 | 0.052 | 0.143 |
| India | 11 | 10 | 1 | 0.041 | 0.091 |
| South Korea | 11 | 11 | 0 | 0.041 | 0 |
| Japan | 7 | 7 | 0 | 0.026 | 0 |
| Australia | 6 | 6 | 0 | 0.022 | 0 |
| United Kingdom | 6 | 3 | 3 | 0.022 | 0.5 |
| Brazil | 5 | 4 | 1 | 0.019 | 0.2 |
| Canada | 5 | 4 | 1 | 0.019 | 0.2 |
Table 7.
Top 10 countries based on the total number of citations received.
| Country | TC | Articles | Average Article Citations |
|---|---|---|---|
| United States | 2582 | 74 | 34.9 |
| China | 1582 | 57 | 27.8 |
| Germany | 685 | 14 | 48.9 |
| South Korea | 381 | 11 | 34.6 |
| Canada | 319 | 5 | 63.8 |
| United Kingdom | 282 | 6 | 47 |
| Austria | 213 | 1 | 213 |
| Portugal | 156 | 1 | 156 |
| Spain | 151 | 3 | 50.3 |
| Iran | 144 | 3 | 48 |
Table 8.
Most commonly used keywords.
| Keyword-Plus | Author Keywords | ||
|---|---|---|---|
| Keywords | Occurrences | Occurrences | |
| artificial intelligence | 66 | artificial intelligence | 81 |
| machine learning | 52 | machine learning | 77 |
| materials science | 33 | deep learning | 26 |
| deep learning | 32 | materials science | 19 |
| design | 31 | materials | 11 |
| neural networks | 30 | artificial neural networks | 9 |
| optimization | 24 | additive manufacturing | 7 |
| prediction | 21 | materials discovery | 7 |
| discovery | 20 | materials informatics | 7 |
| deep neural networks | 16 | neural networks | 7 |
Table 9.
Materials science areas.
| Category | Frequency | Percentage |
|---|---|---|
| Biomaterials and soft materials | 33 | 12.2% |
| Carbon-based nanocomposite materials and applications | 60 | 22.2% |
| Electronics, optics, and quantum | 79 | 29.3% |
| Energy and sustainability | 58 | 21.5% |
| Material characterization | 65 | 24.1% |
| Materials computing and data science | 79 | 29.3% |
| Materials structure, processing, and properties | 189 | 70.0% |
| Structural and functional materials | 70 | 25.9% |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
MDPI and ACS Style
Vergara, D.; Lampropoulos, G.; Fernández-Arias, P.; Antón-Sancho, Á. Artificial Intelligence Reinventing Materials Engineering: A Bibliometric Review. Appl. Sci. 2024, 14, 8143. https://doi.org/10.3390/app14188143
AMA Style
Vergara D, Lampropoulos G, Fernández-Arias P, Antón-Sancho Á. Artificial Intelligence Reinventing Materials Engineering: A Bibliometric Review. Applied Sciences. 2024; 14(18):8143. https://doi.org/10.3390/app14188143
Chicago/Turabian StyleVergara, Diego, Georgios Lampropoulos, Pablo Fernández-Arias, and Álvaro Antón-Sancho. 2024. "Artificial Intelligence Reinventing Materials Engineering: A Bibliometric Review" Applied Sciences 14, no. 18: 8143. https://doi.org/10.3390/app14188143
APA StyleVergara, D., Lampropoulos, G., Fernández-Arias, P., & Antón-Sancho, Á. (2024). Artificial Intelligence Reinventing Materials Engineering: A Bibliometric Review. Applied Sciences, 14(18), 8143. https://doi.org/10.3390/app14188143
Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.
