Powder metallurgy processing of high entropy alloys: Bibliometric analysis and systematic review

2.1 Data mining strategy

The schematic flow chart in Figure 1 shows the steps undertaken to complete the bibliometric analysis and systematic review in this work. All the articles published between 2007 and August 2022 were searched and retrieved from the Scopus database. Although Web of Science offers similar coverage, Scopus offers a wider and more extensive coverage of the abstracts and citations of the peer-reviewed literature [51]. Some keywords related to PM and HEA/MCA were utilized in searching the title, abstract, and keyword fields in Scopus. The query string utilized was (TITLE-ABS-KEY (“high entropy alloy*” OR “multicomponent alloy*” OR “multiprincipal element alloy*” OR “composit* complex alloy*”) AND TITLE-ABS-KEY (“powder metallurgy” OR “powder technology” OR “hot press*” OR “hot isostatic press*” OR “HIP” OR “hot consolidat*” OR “sinter*” OR “spark plasma sinter*” OR “SPS” OR “ball mill*” OR “mechanical mill*” OR “mechanical alloy*” OR “mechanical activation” OR “solid state processing” OR “powder metallurg*” OR “PM” OR “powder process*”) AND NOT TITLE-ABS-KEY (“binary” OR “glass” OR “glass form*” OR “glass-form*” OR “steel” OR “equilibrated” OR “binary alloy” OR “ternary” OR “equivalent” OR “super alloy*: OR “superalloy” OR “Inconel”) AND NOT SRCTITLE (“glass*” OR “glass-form*” OR “glass form*”) AND LANGUAGE (English)) AND PUBYEAR > 2003 AND (LIMIT-TO (SRCTYPE , “j”)) AND (LIMIT-TO (DOCTYPE, “ar”)) AND (LIMIT-TO (LANGUAGE, “English”)). With this query string, 700 articles were extracted from the database. The metadata of the selected literature were extracted from the database on 22nd August 2022.

Figure 1

Schematic flow chart for bibliometric analysis.

2.2 Text cleaning/data reconciliation

The title abstract, and keywords of the retrieved 700 articles were screened to detect and eliminate unrelated ones. No irrelevant article was found, so all the700 articles were subjected to further text cleaning and bibliometric analyses. In the text cleaning step, various forms of author keywords were unified. For instance, keywords like “high-entropy,” “HEA,” “high entropy,” and “high entropy alloy” were combined and treated as high entropy alloys. Likewise, “powder metallurgy,” “powder processing,” and “powder technology” were combined as powder metallurgy. Furthermore, author keywords such as (“SPS,” “spark plasma sintering); (“oxidation,” “oxidation kinetics,” “oxidation resistance”); and (“wear,” “wear behaviour,” “wear resistance,” “wear mechanisms”) were grouped as spark plasma sintering, oxidation behavior, and wear, respectively. Text cleaning was accomplished with a freely available software, OpenRefine, v. 3.4.1 [54].

2.3 Bibliometric analyses

After cleaning the data of the articles, they were analyzed to investigate the trend of publication growth over the years. Further, bibliometric analyses comprising network maps of co-occurrence of article keywords (i.e., author keywords), co-citation between articles, co-citation between scientific journals, and co-authorship between countries were undertaken. These were deployed to identify the main research interest and direction, citation pattern, influential outlets for scientific works, and scientific collaboration between institutions and countries. For these analyses, a free, open-source science mapping software VOSviewer 1.6.17 was utilized [53]. VOSviewer is a distance-based map wherein the relationship between items is reflected by the length of the distance between them. Invariably, a shorter distance between two items indicates a strong relationship between them. Thus, it is easier to identify clusters of related items [53]. Furthermore, each item of analysis (journal, author, country, etc.) is represented by a node (circle or rectangle). The larger it is, the more influential/important the item is. The importance of an item may be assessed based on for instance, the number of documents, citations, average citations, etc. Items with the same color belong to the same cluster and thus, indicates their co-citation of HEA/MCA research. Maps displayed based on density visualization, e.g., for journals, indicate that they received a lot of citations or published more articles [53].

3.1 Trend of scientific publication on PM processing of HEA/MCA from 2004 to August 2022

Figure 2 shows the annual scientific outputs on PM-based processing of HEA/MCA. According to the Scopus database, one article is published per year from 2007 to 2011. However, in 2013, the number of research articles have increased significantly to 12. This increasing trend is equally observed in subsequent years until 2020 when the number of annual publications is lower than 2019. Plausibly, this may be attributed to the corona virus pandemic and its attendant lockdown and restrictions that may have prevented researchers from accessing their laboratories and pandemic-induced delays in the editorial review process. Cumulatively, 700 research articles are indexed in the Scopus database as of 14th August 2022. It is expected that this increase will continue in the following years as researchers design more developmental alloys and more application areas are explored.

Figure 2

Annual scientific outputs on PM processing of HEA/MCA.

3.2 Preferred journals for publishing PM-processed HEA/MCA

It is of utmost importance to identify the most preferred journals for publishing works on PM-processed HEA/MCA research. Researchers and funders can easily recognize important up-to-date sources of information and key outlets for disseminating new ideas and gaining recognition for their efforts. High journal reputation has been shown to enhance article citation [55,56].

Figure 3 shows the article network map of the top 20 research journals for publishing PM-processed HEA/MCA works. The map is generated by setting the minimum number of articles in a journal and minimum number of citations of a journal to 5 and 20, respectively. It should be pointed out that the bibliometric literature does not contain standardized criteria for generating this type of network map [46]. However, we have chosen these criteria for easier analysis of the 142 unique scientific journals.

Figure 3

Network map of the top research journals for publishing PM-processed HEA/MCA.

Consequently, 37 of these journals meet these criteria and are shown in Figure 3. A ranking of the top 20 journals based on the number of published articles, citations, and total link strength is shown in Table 1 below. The total link strength is an attribute that measures the total strength of the links of an item (e.g., journal, author, etc.) with other items under analysis [57].

In Figure 3, there are five clusters of journals indicated in different colors. Cluster 1 (red; ten journals), cluster 2 (green; nine journals), cluster 3 (blue; five journals), cluster 4 (yellow; five journals), and cluster 5 (purple; four journals). Journals clustered in the same group have higher co-citation than journals in other clusters. Besides, the size of a node assigned to a journal indicates its influence based on the number of publications (or citations). In addition, the shorter the length of line between two journals, the stronger the relationship between them [53]. Therefore, the most preferred journal is the Journal of Alloys and Compounds with the highest total link strength (41), number of published articles (116), and citations (2,772). This journal publishes more than twice the number of articles in the next ranked journal, i.e., Materials Science and Engineering A (50 articles, 1,641 citations). The next three are Intermetallics, Materials Letter, and International Journal of Refractory Metals and Hard Materials. Put together, these five journals have published a combined 34% of all PM-processed HEA/MCA works. Although these journals appear to be the most preferred outlets due to the number of publications, Acta Materialia has 1,276 citations from only 8 publications. This strongly indicates that the quality of research works in this journal is quite high. Other factors that may contribute to high citation counts include novelty, reputation of publishing journal, and journal publication model (open access or subscription-based) [55,56,58].

3.3 Scientific collaboration networks

3.3.1 Analysis of co-authorship of researchers

Scientific cooperation among researchers is important for enabling knowledge exchange and technology transfer, gaining access to expensive and specialized research facilities, and it is also an important indices for grading universities [45,51]. In VOSviewer, co-authorship cluster density map of authors is generated by using the criteria “minimum number of documents of an author = 5” and the “minimum number of citations of an author = 10.” Figure 4 shows the cluster density map. One hundred and forty-one authors of the 1,961 researchers in the database meet these criteria. However, only the top 113 connected researchers are shown in the co-authorship cluster density map in Figure 4, while the top 10 authors based on total link strength and other ranking attributes are shown in Table 2. From Figure 4, researchers are grouped into seven clusters represented by red (Cluster 1; 25 researchers), green (Cluster 2; 19 researchers), blue (Cluster 3; 19 researchers), yellow (Cluster 4; 18 researchers), purple (Cluster 5; 16 researchers), cyan (Cluster 6; 9 researchers), and orange (Cluster 7; 7 researchers). Scientists in each cluster represent a collaborating group. The brightness of a node associated with a researcher implies a stronger, collaborative network. Apparently, the most influential researchers in each of these clusters are Liu Y. (cyan; 48 articles; 45 collaborators), Murty B. S. (green; 22 articles; 7 collaborators), Zhang M. (blue; 18 articles; 24 collaborators), Fu Z. (yellow; 21 articles; 18 collaborators), Li X. (purple; 11 articles; 22 collaborators), Liu X. (red; 22 articles; 23 collaborators), and Kumar A. (brown; 12 articles; 3 collaborators. Of all these clusters, the overall most prominent author is Liu Y. He has research collaborators in all other clusters except the green and brown clusters. In most cases, his collaborators are the most prominent in their respective clusters. There are several factors that may affect the diversity of research collaboration. Some of these are the availability of a large number of foreign postgraduate students/visiting scholars and big research funds [51].

Figure 4

Co-authorship clustering map of authors based on number of publications.

Table 2

Analysis of co-authorship of authors

S. No.	Author	Cluster	Number of articles	Citations	Institution/country
1	Liu Y.	2	48	1,531	Sichuan University, Chengdu, China
2	Liu B.	2	24	1,074	Central South University, Changsha, China
3	Liu X.	3	22	314	Southwest Jiaotong University, Chengdu, China
4	Murty B.S.	1	22	1,296	Indian Institute of Technology Madras, Chennai, India
5	Fu Z.	4	21	934	Wuhan University of Technology, Wuhan, China
6	Zhang Y.	3	21	319	Nanjing Institute of Technology, Nanjing, China
7	Li Z.	3	20	286	Chinese Academy of Sciences, Hefei, China
8	Zhang M.	5	18	181	Central South University, Changsha, China
9	Chen J.	3	17	255	Xi'an Technological University, Xi'an, China
10	Chen W.	3	17	500	South China University of Technology, Guangdong, China

Table 2 presents more details about the collaborative network of the top 10 most prolific authors. The five most prolific authors are Liu Y. (48 articles), Liu B. (24 articles), Liu X. (22 articles), Murty B. S. (22 articles), and Fu Z. (21 articles). However, the positions of these authors change when they are ranked based on citations. Thus, the top five are Liu Y. (1,531 citations), Murty B. S. (1,296 citations), Kottada R. S. (1,124 citations), Liu B. (1,074 citations), and Fu Z. (934 citations). It is clear from this new list that many published articles by an author do not necessarily imply more citation counts. High citation counts are more likely to be associated with the quality of a research work, novelty, reputation of publishing journal, and access type (open access or subscription-based) [55,56].

3.3.2 Analyses of institutional and international co-authorship

Results in Table 3 show the analysis of institutional co-authorship. The criteria for generating the results include setting “minimum number of documents of an institution” and “minimum number of citations of an institution” to 5 and 10, respectively. Sixty-seven out of 964 institutions satisfy these conditions, but only the top ten institutions are indicated in Table 3. Furthermore, only 19 of these 67 institutions have collaborative links with the other institutions, suggesting that there is a low inter-institutional collaboration. The highest number of articles (31) and citations (1,509) are linked to the Indian Institute of Technology (IIT) Madras, India. This is followed closely, in terms of number of publications, by the Central South University (CSU), China. Although IIT has published ten more research articles than CSU, the citations count of IIT is more than 400% higher than that of CSU. It is highly likely that the articles from IIT are of higher quality and in highly reputed journals than those of CSU.

Table 3

Analyses of institutional co-authorship

S. No.	Institution	Country	Number of published articles	Links	Total link strength	Citations
1	Indian Institute of Technology Madras (Department of Metallurgical and Materials Engineering), Chennai	India	31	5	15	1,509
2	Central South University (State Key Laboratory of Powder Metallurgy), Changsha	China	21	7	10	362
3	Chinese Academy of Sciences (State Key Laboratory of Solid Lubrication, Lanzhou Institute of Chemical Physics), Lanzhou	China	18	2	7	434
4	Zhengzhou University (School of Physics and Engineering), Zhengzhou	China	13	2	8	161
5	Ajou University (Department of Materials Science and Engineering), Suwon	South Korea	12	2	11	52
6	Russian Academy of Sciences (Merzhanov Institute of Structural Macrokinetics and Materials Science), Chernogolovka	Russian Federation	12	2	7	60
7	Yuanmeng Precision Technology (Shenzhen) Institute, Shenzhen	China	10	6	17	151
8	Nanjing Agricultural University (College of Engineering), Nanjing	China	9	3	10	52
9	Ajou University (Department of Energy Systems Research), Suwon	South Korea	9	1	9	50
10	Pohang University of Science and Technology (Department of Materials Science and Engineering), Pohang	South Korea	9	2	6	88

However, the total link strengths of all the institutes are relatively low. The total link strength is an attribute that indicates the strength of co-authorship of an item (e.g., institutions) with others [57]. This implies that most of these institutions have reatively low collaboration with the other institutions. Likewise, of the ten institutions listed in Table 3, half are located in China based on the total number of research articles. To advance knowledge in this field, it is imperative to encourage and establish inter-institutional and international collaborations.

Other than inter-institutional collaboration, international research collaborations are equally important means of enhancing knowledge sharing and technology transfer. Figure 5 shows the density map of the countries with the highest number of publications. To generate this figure in VOSviewer, the “minimum number of documents of a country” is set to 5 while the “minimum number of citations of a country” is set to 10. All PM-based HEA/MCA articles are affiliated with 58 countries according to the data retrieved from the Scopus database but only 31 meet this threshold, as shown in Figure 5. However, two countries – Egypt (Links 1, Research articles 9; Citations 36) and South Africa (Links 0, Research articles 7; Citations 37) – are excluded from the map due to not having significant number of co-authorships with other nations. The brightest and largest spot is linked to China (299 articles), followed by India (115 articles), and then South Korea (56 articles). The brightness of a spot indicates the density (or number) of articles published by the country associated with the node.

Figure 5

Density map of co-authorship of countries measured by number of scientific publications.

All the countries are connected showing that international collaboration is well established. A summary of the statistics of the top ten most prolific countries is presented in Table 4.

Table 4

Top ten most prolific countries in the field of PM-processed HEA/MCA

S. No.	Country	Number of articles	Citations
1	China	299	5,634
2	India	115	2,641
3	South Korea	56	765
4	United States	56	2,777
5	Germany	39	862
6	Russian Federation	24	201
7	Czech Republic	23	402
8	France	22	324
9	Iran	22	108
10	Turkey	17	171

The highest number of publications (299) and citation counts (5,634) emanates from China. In fact, Chinese publications in the period under review exceed the sum of publications from the next top four countries, i.e., India, South Korea, United States, and Germany.

3.4 Citations analysis of most influential articles on PM-processed HEA/MCA

Figure 6 below shows the citations analysis of the most influential articles on PM-processed HEA/MCA. This section presents the results of the analysis of the most influential publications. The minimum number of citations of a document is selected to be 20 with 195 of the 700 articles meeting this criterion. However, only the top 50 are shown in Figure 6. Despite that the first articles in this field appear pre-2010 [37,59,60], the top five most cited are published post-2011 [61,62,63,64,65]. This is a bit strange; pioneering research articles usually have the highest citation counts as newer articles continue to refer to them. In the broad field of HEA/MCA, the most cited articles are those of Yeh et al. [2] and Cantor et al. [1], the pioneering publications on HEA/MCA. These two articles have more than 6,800 citations between them.

Figure 6

Network map of most-cited PM-processed HEA/MCA research articles.

In the sub-field of PM-processed HEA/MCA, the most influential article in terms of citation count is Liu et al. [61] with 456 citations, Followed by Senkov et al. [62] with 420 citations and Youssef et al. [63] with 368 citations. Completing the top five most cited articles are Praveen et al. [64] (274 citations) and Senkov et al. (272 citations) [65]. More details about these most-cited articles and publication outlets are contained in Table 5.

3.5 Co-occurrence network of author keywords

A co-occurrence network is useful for recognizing important research areas and potential future research directions [46]. In VOSviewer, the co-occurrence network map of author keywords is generated by setting the minimum number of occurrences of a keyword at 5. In all, there are 684 keywords of which 86 meet this threshold. Each circular node in Figure 7 represents a keyword, while the size of a node is proportional to the number of articles in which the keyword occurs. The distance between any two keywords is a measure of their relatedness. This implies that related keywords have shorter distance between them. Finally, a group of keywords may have the same color and therefore, belong to the same cluster. This implies that such keywords in the same cluster tend to co-occur together more than keywords in a different cluster [48].

Figure 7

Co-occurrence network map of author keywords.

All the 86 keywords are grouped into seven clusters. The cluster with largest number of keywords (i.e., 18) is cluster 1 in red and has “high entropy alloy” as the most prominent (492 occurrences and 85 links with other keywords). Some of the other keywords in the same cluster are HEA composites, laser sintering, compressive properties, microwave sintering, intermetallic compounds, deformation behavior, plastic deformation, etc. Cluster 2 is in green with 18 keywords. The most frequently occurring here is mechanical alloying (210 occurrences, with 70 links). This indicates that the preferred production method is by mechanical alloying, rather than powder mixing or additive manufacturing, for instance. Other co-occurring keywords in the same cluster with “mechanical alloying” are phase evolution, annealing, nanocrystalline HEAs, process control agents, oxide dispersion strengthening, molecular dynamics, etc. All these keywords have strong correlation with “mechanical alloying.” For instance, mechanical alloying is a prominent method to obtain nanocrystalline structure, which often requires the use of “process control agents” to minimize cold welding.

Cluster 3 appears in blue and contains a group of 13 keywords. Some of these include vacuum hot pressing (5 occurrences, 11 links), powder metallurgy (128 occurrences, 60 links), HEA carbides (6 occurrences, 12 links), HEA ceramics (6 occurrences, 10 links), friction and wear behavior (129 occurrences, 84 links), etc. The yellow cluster is Cluster 4 with 13 keywords including corrosion behavior, HEA coatings, heat treatment, magnetron sputtering, etc. Of all these Cluster 4 keywords, the most frequently used is corrosion behavior with 40 mentions and 37 links. Clusters 5 (9 keywords), 6 (8 keywords), and 7 (6 keywords) are coded purple, cyan, and brown, respectively.

Going by the number of articles retrieved from the Scopus database, PM-processing of HEA/MCA is still evolving. Therefore, other than the conventional “sintering” techniques, others like “spark plasma sintering,” “microwave sintering,” and “vacuum hot pressing” have been investigated in the literature. Among the different consolidation techniques, the most preferred is the SPS (142 mentions, 63 links). This may not be unconnected to its unique advantages including rapid sintering in a matter of minutes, lower sintering temperature, which is beneficial for minimizing phase transformation, and its ability to maintain the nanocrystalline structure obtained from the high-energy/mechanical milling process [25,26]. As indicated by its high occurrence, this method has been well researched in the literature.

Other than this consolidation technique, there are also a few reports on using the microwave sintering technique (14 occurrences). One of the first works using this approach is reported by Veronesi and co-workers [36]. Given that it is a very rapid sintering process (e.g., 180 s of sintering time [36]), this technique offers similar advantages associated with SPS. Therefore, it has been widely used to consolidate various traditional metal matrix composites (MMCs) [66,67,68]. Nonetheless, it may be of scientific interest to compare the properties of HEA/MCA materials consolidated via SPS technique with the microwave sintering technique.

Majority of the studies on PM-processed HEA/MCA have focused on investigating mechanical properties. This explains the relatively high occurrences (keyword variants include compressive behavior, microhardness, tensile, mechanical testing, etc.) in the literature. Nonetheless, other functional properties such as tribological properties (56 occurrences), electrochemical impedance (five occurrences), electromagnetic (two occurrences), magnetic (17 occurrences), thermo-electric (14 occurrences), and thermal analysis (14 occurrences) have also been investigated. However, the low number of occurrences for electrochemical, thermo-electric, and electromagnetic indicates that future research directions may consider these properties for investigation and application areas.

Departing from the traditional applications (e.g., refractory materials, hard materials, etc.), some functional applications of PM-processed HEA/MCA have also appeared in the literature. This explains why keywords like hydrogen storage and electromagnetic properties, have been investigated. However, all these keywords have been used less than ten times in the examined literature. More future research works may be needed in these directions to fully harness the potentials offered by HEA/MCA for these applications.

The bulk of the works retrieved are biased towards laboratory experiments. To predict stable phases and their relative amounts, reduce cost, and minimize developmental time of important alloys, machine learning/computational approaches may be adopted [69,70]. In traditional PM processing, machine learning approach has been successfully deployed to predict the particle sizes of mechanically-milled magnesium-based metal powder [71] and discover novel nickel superalloys [72]. To this end, “calphad” (CALculation of PHAse Diagrams) has been used in the PM-based HEA/MCA literature about seven times. Some of these works are in the cited references [73,74]. In the broad field of HEA/MCA, it is encouraging that some recent works have adopted this approach [70,75,76]. To fully harness the potentials of this computational approach, more future research efforts may be devoted to this niche.

One of the core strengths of PM-processing is its versatility in producing difficult-to-melt and refractory alloys. For this, a keyword like refractory high entropy alloy has been used about 41 times in the surveyed literature on PM-HEA/MCA. Other associated keywords with similar use are cermets, cemented carbides, wear, and oxidation behavior. It is equally important to note that some keywords have been excluded from analysis due to not meeting the minimum of 5 occurrences threshold. Some of these include keywords associated with physical properties (e.g., dielectric properties, machinability – one occurrence each), processing techniques (e.g., severe plastic deformation, hot forging, metallothermic reduction, self-propagating high-temperature synthesis – one occurrence each), theoretical modeling (e.g., density functional theory, thermo-calc etc.). The occurrence of these keywords less than five times is an indication that they are potential future research directions.

3.6 Limitations of bibliometric study

One of the limitations of the bibliometric analyses in this study is that we have limited literature retrieval to a single database (i.e., Scopus). It is likely that some important articles may have not been indexed in this database. Future works may consider combining articles from several databases. Furthermore, our data exclude articles not published in English language. Finally, other academic sources, such as conferences, books, encyclopedia, etc., have not been considered in our analysis.

4.1 Tribological application

PM techniques have provided opportunities to develop a reliable and efficient production of advanced materials, such as HEAs [80]. The process is well suited for producing fine-grained homogenous microstructures and wear resistant materials from HEA-based composites [81]. Hard ceramic particles provide a wide range of improved mechanical properties, such as hardness, wear resistance, and compressive strength when combined with MMCs [82,83]. An FCC phase and a Cr₇C₃ phase were formed during the sintering of CoCrFeNiMn HEA after the addition of Cr₃C₂ to improve the tribological properties of the material [84]. When CoCrFeNiMn HEA was prepared without the addition of Cr₃C₂ particles, the Vickers hardness was 181.2 HV, whereas increasing the amount of Cr₃C₂ by 10, 20, and 40% increased the hardness to 323.2, 417.0, and 681.6 HV, respectively. Figure 8 depicts the friction coefficients of CoCrFeNiMn and CoCrFeNiMn-Cr₃C₂ at various temperatures under different conditions. According to Figure 8a, the friction coefficient of the CoCrFeNiMn HEA was 0.62 at room temperature, and it further decreased as the temperature rose to 200, 400, 600, and 800°C.

Figure 8

(a) Friction coefficients and (b) wear rates of CoCrFeNiMn HEA and its composites at RT, 200, 400, 600, and 800°C. Reproduced with permission [84]. Copyright 2021, Elsevier.

This decrease in friction coefficient was primarily due to the oxide lubrication and matrix softening effects at higher temperatures. With the increase in the Cr₃C₂ content (Figure 8a), the friction coefficient decreased until it reached 20%, after which it increased until it reached 40%. As a result, CoCrFeNiMn HEA matrix composites exhibited excellent ability to reduce friction. CoCrFeNiMn HEA wear rate (Figure 8b) decreased with the increase in the temperature, from 5.60 × 10⁻⁴ mm³·Nm⁻¹ at room temperature to 7.17 × 10⁻⁶ mm³·Nm⁻¹ at 800°C. When the content of Cr₃C₂ in CoCrNiMn HEA was increased by 10, 20, and 40%, the wear rate at room temperature decreased from 5.60 × 10⁻⁴ mm³·Nm⁻¹ to 0.98 × 10⁻⁵ mm³·Nm⁻¹, and 0.56 × 10⁻⁵ mm³·Nm⁻¹. This was primarily due to the improvement in hardness [84]. The wear rates of CoCrFeNiMn HEA, CoCrFeNiMn-10% Cr₃C₂ composite, and CoCrFeNiMn-40% Cr₃C₂ composite all remained constant at 800°C, but the wear rate of CoCrFeNiMn-20% Cr₃C₂ composite was significantly higher. In general, the CoCrFeNiMn-10% Cr₃C₂ composite exhibited excellent wear resistance from room temperature to 800°C.

Using AlCoCrFeNi–ZrO₂ HEA composites, Ghanbariha et al. [85] investigated the effect of ZrO₂ particles on the nanomechanical properties and wear behavior. The AlCoCrFeNi–ZrO₂ composite was synthesized using a mechanical alloying process combined with SPS at 1,000°C. The sintered samples comprised two main phases: FCC and BCC, and minor phases: Al-rich and Cr-rich. This was due to the short duration of the SPS process, which did not allow for the conversion of the non-equilibrium BCC phase into the equilibrium FCC phase [85]. As a result, a combination of FCC and BCC phases were found. For the effect of zirconia content on the wear resistance of AlCoCrFeNi–ZrO₂, it was discovered that zirconia reinforcement content from 0 to 5 wt% only slightly improved the wear resistance due to a balance of positive effects of ZrO₂ particle addition and negative effects of BCC phase reduction. When there was an increase in zirconia particles to 10 wt%, the elastic to strain ratio (H/E) of the FCC phase remained constant, while the H/E ratio of the BCC phase moderately increased. Despite the H/E effect, it appeared that the positive effect of the addition of 10 wt% zirconia played a more significant role in the improvement of wear resistance. According to the wear width analysis, the pure HEA sample had the largest wear width (1932.58 μm), followed by the 5% ZrO₂ sample (1825.84 μm), and the smallest wear width was found in the 10 wt% zirconia particles sample (1433.20 μm). Consequently, the AlCoCrFeNi-10 wt% ZrO₂ sample exhibited the best wear performance of the three samples tested.

Kang et al. [81] investigated the effect of TiB₂ on the tribological properties of AlCoCrFeNi using ball milling and SPS at 1,200°C, and detected the following phases: B2, BCC, FCC, and TiB₂. The microstructure consisted of spherical or near-spherical particles, which become more pronounced with increased TiB₂ content, indicating that TiB₂ could inhibit the bonding of the AlCoCrFeNi powders during the sintering process. Furthermore, when the TiB₂ content of the AlCoCrFeNi particles was increased, a significant agglomeration phenomenon was observed between the AlCoCrFeNi particles. It was discovered that increasing the TiB₂ content in AlCoCrFeNi/TiB₂ composites caused an increase in the amount of boron oxide present in the tribological film, which resulted in a decrease in the values of coefficient of friction (COF). When the TiB₂ content increased from 0 to 30%, the wear rate of the composites decreased from 5.84 × 10^–4 to 0.93 × 10^–4 mm³·m⁻¹, indicating a decrease in the wear rate of the composite. It was also reported in additive manufactured CoCrFeMnNi/TiB₂ composites that wear rate decreased from 3.42 to 1.48 mm³·Nm⁻¹ due to TiB₂ addition [86]. This clearly demonstrates that increasing the TiB₂ content in AlCoCrFeNi/TiB₂ composites increases the wear resistance. Additionally, the boron oxide formed during the tribological oxidation of TiB₂ played a significant role in the formation of solid lubricants.

Wang et al. [86] studied the in situ formed graphene which provided lubricity for the FeCoCrNiAl-graphene (FeCoCrNiAl-G) based composite. The COF of FeCoCrNiAl increased with the increase in the force and frequency at 5 N and 1 Hz, respectively, and remained stable at a high value. However, when the load was greater than 20 N and the velocity was greater than 3 Hz, the COF showed a decreasing trend in the running period. This was especially noticeable when the load was greater than 40 N and the frequency was greater than 6 Hz [86]. When the HEA-G was operated at a load less than 30 N, the COF decreased initially from 0.2 to approximately 0.16 in the steady state; when the load was greater than 40 N, the COF was also reduced to a steady state after the initial 300 cycles. The study revealed that the original FeCoCrNiAl HEA had poor tribological properties when subjected to low loads and velocities. The addition of graphite nanoplate improved the tribological properties of the original HEA significantly under low loads and velocities, with a reduction in COF of 80% at a load of 5 N (0.15 vs 0.84) and a reduction in wear of nearly two orders of magnitude. The sliding and shearing produced by graphene were primarily responsible for the dramatic reduction in friction and wear (below ten layers). Another way that the tribo-chemical reaction improved tribological properties was by forming a tribo-layer that had metallic and partially oxidized metal oxide nanoparticles in it, which could also help improve the tribological properties. Similarly, tribological properties improvement through tribo-layer containing oxidized metal oxide was reported in graphite-reinforced FeNiCrCuMo high-entropy alloy [87].

4.2 Oxidation resistance

HEAs are potentially applicable in high-temperature applications because of their exceptional oxidation resistance [88,89]. According to Butler and Weaver [89], the good oxidation resistance of HEA is because they are less compositionally constrained than conventional structural alloys and accommodate higher concentrations of the elements that are necessary to form protective external oxide scales such as Al or Cr. In addition, HEAs exhibit sluggish diffusion kinetics that could enhance their oxidation behaviors by inhibiting the formation of non-protective transient oxides [90]. Vilémová et al. [91] investigated the oxidation properties of CoCrFeMnNi and FeCoCrNi HEA alloys containing a Y-Ti-O particle produced by MA+SPS for 1–150 h at temperatures ranging from 750 to 950°C. The oxidation performance of the CoCrFeMnNi exhibited a two-stage oxidation behavior at 750°C. The oxidation rate in the second stage did not decrease with the thickness of the oxide layer; therefore, the alloy did not function as an effective oxidation barrier. However, after 150 h at 750°C, no significant oxide spallation was observed. It was discovered that removing Mn from the alloy composition significantly reduced the weight gain of FeCoCrNi and that the oxide served as a protective barrier against further oxidation after 150 h at 750°C. When heated to 950°C for 5 h, CoCrFeMnNi exhibited significant spallation. The spallation phenomenon was also observed in as-cast HEAs such as: Fe_22.63Ni_26.06Co_26.82Mn_24.47 at 750°C for 100 h [92]; AlCoCrFeNi at 10–30 h [89]; Al_0.9FeCrCoNi [93]; and Al_0.25CoCrCuFeNiMn at 700 and 900°C [94]. Spallation was attributed to high thermal stresses that developed within the substrate and the oxide layer during the cooling part of the exposure [93,94]. Similarly, Vilémová et al. [91] discovered that removing Mn from the FeCoCrNi alloy composition resulted in an improved protection of the alloy even at 950°C after 5 h.

The role of Al content on the cyclic oxidation of resistance of sintered CoCrFeNiAl _x (x = 0.7, 0.85, and 1 mol) was investigated in another study at 800, 875, and 950°C for 120 h [95]. Because of the higher Al content in CoCrFeNiAl₁, the parabolic rate constant (Kp) of CoCrFeNiAl₁ was lower than that of CoCrFeNiAl_0.7 and CoCrFeNiAl_0.85. The oxidation kinetics were similar to that reported in as-cast Ni₂FeCoCrAl _x alloy during dry air-oxidation at T ≥ 700°C [96]. This suggests that the addition of Al enhances the oxidation resistance of CoCrFeNiAl _x HEAs. An increase in Al content also resulted in better oxidation resistances for arc melted AlCoCrFeNi HEA [89]. Furthermore, the Kp values increase as the oxidation temperature is raised, but they decrease as the Al content is raised. The CoCrFeNiAl₁ Kp value (0.01985 mg²·cm⁻⁴·h⁻¹) increased slightly when heated to 800°C to 0.02954 mg²·cm⁻⁴·h⁻¹ for the CoCrFeNiAl_0.7 alloy and further to 0.1399 mg²·cm⁻⁴·h⁻¹ for the CoCrFeNiAl_0.85 alloy. The oxidation results obtained at 950°C for 120 h revealed an outer layer composed of Cr₂O₃ and an inner discontinuous layer (Al₂O₃) formed on the HEAs CoCrFeNiAl_0.7 and CoCrFeNiAl_0.85. This outer layer (Cr₂O₃) and inner discontinuous layer (Al₂O₃) were found in as-cast AlCoCrFeNi HEA oxidized at 1,050°C for 100 h [89] and as-cast AlCoCrFeNi oxidized at 1,000°C for 48 h [97]. The oxide scale of CoCrFeNiA_l contained significantly more Al₂O₃ than CoCrFeNiAl_0.7 and CoCrFeNiAl_0.85. The increase in Al concentration led to an increase in the continuity of the Al₂O₃ subscale [89].

Zhang et al. [98] also varied Al content in spark plasma sintered Al _x CrTiMo (x = 0.25, 0.5, 0.75, 1) refractory HEAs. During the process of oxygen corrosion, the oxidation kinetics resulted in a moderate mass gain in the Al_0.5CrTiMo, Al_0.75CrTiMo, and AlCrTiMo refractory HEAs. However, there was a reduction in the rate of mass increase that occurred twice during the oxidation process of Al_0.25CrTiMo. This occurred because of the absence of protective alumina oxide layers forming on the surface of low Al-containing samples, or it could be attributed to a few spallation and delamination that occurred on the surface of low Al-containing samples. The oxidation layers, which extended from the exterior to the interior of the alloys, were composed of TiO₂, Al₂O₃, and Cr₂O₃. The findings of the study confirmed the importance of Al content in the oxidation resistance of HEAs.

4.3 Corrosion resistance

HEAs are being considered for potential corrosion-resistant applications because most elements in HEAs are corrosion-resistant and passivating. They are usually free of common corrosion initiating sites, such as impurities and inclusions [99]. The PM-HEAs have exhibited considerable compositional precision, especially the minimization of elemental segregation, which has been reported to improve the corrosion resistance of HEAs. For example, reported as-cast Cu₄₅Mn₂₅Al₁₅Fe₅Cr₅Ni₅ [100] and CuCrFeNiMn [101] HEAs showed decreased corrosion resistance due to elemental segregation. Zhou et al. [102] investigated the effect of annealing temperature on the corrosion properties of SPS-ed AlCoCrFeNi HEA in 0.5 mol·L⁻¹ H₂SO₄. In SPS-ed AlCoCrFeNi, B2, BCC, and FCC phases were discovered. During the annealing process at temperatures of 700, 800, and 900°C, the SPS-ed AlCoCrFeNi transformed to B2 + BCC + FCC + sigma. It was found that the corrosion potential and current density increased, and the polarization resistance decreased with the increase in the annealing temperature. The AlCoCrFeNi alloy annealed at 900°C showed the highest corrosion potential (0.315 V), current density (108 μA·cm⁻²), and polarization resistance (401.2 Ω). The formed sigma phase during the annealing process decreased the corrosion resistance of the AlCoCrFeNi alloy. The sigma phase also dominated the corrosion process of as-cast CoCrFeNiMo_0.3 [103], with the corrosion resistance decreasing rapidly due to the sigma phase precipitation. In an as-cast CoCrFeMnNi HEAs [104], the sigma phase formed in the ultra-fine grain alloy after the annealing process resulted in galvanic corrosion between the sigma phase and its neighboring matrix phase.

Parakh et al. [105] investigated the crystal structure and grain size effects on the corrosion properties of SPS-ed AlNiCoCrFe, FeCrNiAlCo, CoNiFeCrAl, and as-cast AlCoCrFeNi HEAs in 3.5 wt% NaCl solution. The SPS-ed AlNiCoCrFe, FeCrNiAlCo, and CoNiFeCrAl alloys contained BCC and FCC phases. It was found that there was an improvement in corrosion resistance of FeCrNiAlCo and CoNiFeCrAl HEAs when the FCC content was 38 and 16%. For the AlNiCoCrFe alloy, an increase in FCC content to 62% led to a drop in corrosion resistance. Amongst the SPS-ed alloys, FeCrNiAlCo with an optimum FCC content of 38% had the best corrosion resistance. This is equivalent to the as-cast AlCoCrFeNi alloy. It was found that the as-cast AlCoCrFeNi alloy had better corrosion resistance than the SPS-ed alloys because the grains were bigger. It was also confirmed in as-cast CoCrFeMnNi HEAs where the coarse grained (48 μm) sample exhibited a better corrosion resistance than the ultra-fine-grained (0.689 μm) in 3.5 wt% NaCl solution [104]. The authors concluded that the corrosion resistance of an alloy was affected by three things: phase fraction, grain size, and dislocation density. Of these three, phase fraction is the most significant, while dislocation density is the least.

Thaha et al. [106] investigated the corrosion behavior of sintered MgZnFeCuCo alloys in Hanks’ solution with and without 0.01 M coumarin. It was discovered that Mg₁₉Zn₁₇Fe₂₈Cu₁₈Co exhibited main phases of Cu, Zn-rich phase regions of MgZn₂, Mg₂Cu, and CuZn₅. The microstructure of the Mg₁₁Zn₂₆Fe₇Cu₄Co and Mg₁₁Zn₆Fe_0.2Cu_0.3Co alloys contained MgZn₂, Mg₂Cu, CuZn₅, and FeZn, which were distributed along the grain boundaries of the respective alloys. The corrosion current density of Mg₁₉Zn₁₇Fe₂₈Cu₁₈Co was the lowest when compared to Mg₁₁Zn₂₆Fe₇Cu₄Co (313.5 µA·cm⁻²) and Mg₁₁Zn₆Fe_0.2Cu_0.3Co (345.6 µA·cm⁻²). As a result, it demonstrated a superior corrosion resistance to Mg₁₁Zn₂₆Fe₇Cu₄Co and Mg₁₁Zn₆Fe_0.2Cu_0.3Co in Hanks’ solution and Hanks’ solution + 0.01 M coumarin. Furthermore, the corrosion resistance of MgZnFeCuCo alloys increased with an increase in the amounts of MgZn₂, Mg₂Cu, and CuZn₅ present in the α-Mg matrix of the alloy. The addition of 0.01 M coumarin inhibitor to the Hanks solution increased the charge resistance of all MgZnFeCuCo alloys regardless of their composition. This suggested that coumarin acted as a good corrosion inhibitor to protect the magnesium based HEAs Mg₁₉Zn₁₇Fe₂₈Cu₁₈Co, Mg₁₁Zn₂₆Fe₇Cu₄Co, and Mg₁₁Zn₆Fe_0.2Cu_0.3Co from corrosion.

Singh et al. [107] conducted research to improve the corrosion resistance of mechanically alloyed FeCoCrNiCu HEA in 3.5% NaCl solution by the incorporation of carbon nanotubes (CNT). In the FeCoCrNiCu-CNT composites, two FCC phases (Cu-rich and Cr-rich) and a minor sigma phase were observed. The authors discovered that increasing the number of CNTs caused a progressive decrease in phase separation. According to the corrosion resistance behavior of FeCoCrNiCu-CNT, it was observed that the corrosion rate increased with the addition of more CNTs to the composition. Basically, this implied that an optimal amount of CNT (2 wt%) could be used to incorporate into the HEA matrix to achieve a significant 88.6% reduction in the rate of corrosion. The improvement in corrosion resistance was attributed to an increase in chemical homogeneity, which reduced the likelihood of galvanic coupling occurring. Furthermore, the re-emergence of chemical heterogeneity and the evolution of the Cr₂₃C₆ phase were attributed to the increase in corrosion rate beyond the optimum CNT fraction.

A recent study conducted by Wang et al. [108] on the corrosion resistance of ball milled CuZrAlTiNiW/Al HEAs (designated as Al₁₀, Al₂₀, and Al₃₀) in artificial sea salt and distilled water (mass ratio = 1:30) led to an intriguing finding. A single BCC solid solution phase was found in the as-milled CuZrAlTiNiW, while the SPS-ed contained an ordered BCC phase (B2) along with WAl₁₂ intermetallic compound and a few unknown phases. Increases in the contents of B₂ and WAl₁₂ phases have been observed in HEAs as the volume fractions of the HEA alloys are increased from 10–30%. Potentiodynamic polarization curves revealed that Al₁₀ had the lowest corrosion current density (i _corr) of 0.86 × 10⁻⁵ A·mm⁻² and the most positive corrosion potential (E _corr) of −1.04 V and lowest penetration rate of 0.827 mm per year, demonstrating that the use of high-efficiency additives could be beneficial in improving the general corrosion resistance of HEA composite materials. When compared to pure Al, the pitting resistance of the SPS-ed Al-HEA composites was significantly improved in seawater solution. The Al₃₀ composite, on the other hand, had the highest ΔE value (0.89 ± 0.05 V), indicating that it had a good pitting resistance.

4.4 Hydrogen storage

Hydrogen is being considered as a viable energy source in the development of cleaner energy to replace fossil fuels. Further improvements in this field are hinged on the development of safe and efficient hydrogen storage systems. It is possible to use metal hydrides in renewable energy systems because their compactness allows them to store large amounts of energy in relatively small spaces. They are also used in thermal energy storage systems and hydrogen compression [109]. In terms of properties, such as gravimetric/volumetric hydrogen storage capacity, heat of reaction, operating pressure, and temperature, selecting a suitable metal hydride is essential. The problem with metal hydrides composed of only transition metals is their limited capacity with a hydrogen/metal (H/M) gravimetric ratio ≤2 [110]. Because of the vast compositional space that can be evaluated in multicomponent systems, HEAs have attracted attention because of their ability to display a high hydrogen storage capacity with a H/M ratio of 2.5 [110,111]. They also have high lattice distortion, leading to additional lattice strain favorable for hydride formation. HEAs have shown a considerable amount of high hydrogen storage capacity of about 2.7 wt% using combinations of elements such as Ti, V, Zr, Nb, Ta, Hf, Nb, Al, and Ni as shown in Table 6.

Currently, the use of metal hydrides, such as magnesium (Mg) has attracted attention because of properties that favor hydrogen applications including high gravimetric capacity, lightweight, low cost, and abundance [112,113,114]. A few HEAs containing Mg synthesized by high-energy ball milling have already been reported and have displayed promising high hydrogen storage, as shown in Table 6. For example, Mg_0.68TiNbNi_0.55 synthesized by high-energy ball milling had a high hydrogen storage of 3.3 wt% [112].

Strozi et al. [113] reported a method for designing a single-phase Mg-containing HEAs for hydrogen storage applications, which was performed via a high energy ball milling method for 24 h. A single-phase BCC crystal structure was discovered in the synthesized alloy (Mg₁₂Al₁₁Ti₃₃Mn₁₁Nb₃₃) using XRD, and further characterization with EDS revealed that the sample exhibited high chemical homogeneity. It was discovered that the lower temperature at which absorption occurs was 275°C for the initial pressure-composition-isotherm (PCI) test on Mg₁₂Al₁₁Ti₃₃Mn₁₁Nb₃₃ under hydrogen pressure of 30 bar. Based on their findings, it was estimated that the maximum hydrogen absorption capacity was approximately 1.75 wt% (H/M = 1.05). This value is higher than the hydrogen absorption capacity of reported Ti_0.20Zr_0.20Hf_0.20Nb_0.40 (1.5 wt% at 300°C) [114], MgZrTiFe_0.5Co_0.5Ni_0.5 (1.2 wt% at 350°C) [114], TiZrNbTa (1.25 wt% at 220°C) [115], and MgTiNbCr_0.5Mn_0.5Ni_0.5 (0.8 wt% at 300°C) [112]. The BCC phase found in Mg₁₂Al₁₁Ti₃₃Mn₁₁Nb₃₃ expanded following the PCI test. A mass loss of approximately 1.75 wt% was detected using TGA during the hydrogen desorption test, indicating that the hydrogen was completely dehydrated. After undergoing a hydrogen desorption test, a single BCC phase was revealed with two endothermic peaks that overlapped each other. There are two endothermic peaks that are associated with hydrogen release, the first of which is associated with the phase transformation of the undistorted BCC hydride (α-phase) and the second of which is associated with the hydrogen release from the α-Phase.

Cardoso et al. [116] produced another magnesium HEA (MgAlTiFeNi) by mechanical alloying (30 h at 200 rpm) and reactive milling (RM; 3 MPa for 24 h at 600 rpm). A primary BCC phase, an FCC phase (TiH₂), and a hydride phase (Mg₂FeH₆) were formed during the RM process. A BCC phase was formed during the mechanical alloying procedure. Using a DSC scan, it was discovered that the reactive milled-MgAlTiFeNi exhibited a hydrogen desorption peak at around 326°C. However, this was still lower and closer to the 378°C temperature determined for the pure complex hydride Mg₂FeH₆ [117]. The onset temperature for hydrogen desorption was 286°C, which was lower than the onset temperature for commercial magnesium hydride (MgH₂), which is between 426 and 442°C [118]. It was further discovered that the amount of hydrogen absorbed in the MgAlTiFeNi alloy by RM was 0.87 wt%, which increased to 0.94 wt% at 325°C after the second absorption, using the TG curve. These values were higher than reported hydrogen absorption capacities of high energy ball milled MgVAlCrNi at 0.3 wt% [119], MgTiNbCr_0.5Mn_0.5Ni_0.5 at 0.8 wt% [112], and MgTiVCrFe at 0.2 wt% [120]. However, the hydrogen absorption capacity was lower than as-cast Ti_0.30V_0.25Zr_0.10Nb_0.25Ta_0.10 at 2.5 wt% [121], Al_0.10Ti_0.30V_0.25Zr_0.10Nb_0.25 at 2.6 wt% [111], and Ti_0.20Zr_0.20Hf_0.20Nb_0.40 [122]. After desorption, it was discovered that the hydride-related peaks of Mg₂FeH₆ had disappeared, which indicated that the hydride had decomposed into Mg and Fe, as well as the release of hydrogen. All of the hydrogen that was removed from the reactive milled MgAlTiFeNi sample and detected by TG came from the decomposition of the Mg₂FeH₆.

4.5 Radiation application

Several promising properties of HEAs have been reported, including excellent combination of strength and ductility, as well as promising radiation-tolerant behavior [123,124]. Radiation resistance has been attributed to the complex intrinsic transport properties of HEAs where the increased compositional complexity can reduce effective interstitial mobility while increasing vacancy-interstitial reaction [125,126]. The mixing of various elements leads to the possibility of high irradiation resistance via unique “self-healing” mechanisms [127]. Similarly, Xia et al. [123] found that the "self-healing" process makes HEAs more stable against irradiation than other materials, like amorphous alloys and nano-structured alloys.

In the study of SPS-ed CuCrFeTiV irradiated at 300 keV at room temperature, Dias et al. [128] found that irradiation of CuCrFeTiV did not affect the microstructure. This behavior was reported by Kumar et al. [129] for an as-cast FeNiMnCrHEA, which retained its original crystalline structure after irradiation and dose up to 3 dpa at room temperature. In the as-cast HfNbTaTiZr alloy studied by Chang et al. [130], no phase transformation occurred after being irradiated with 300 keV Ni⁺. These showed the high stability and resistance of CuCrFeTiV HEA to irradiation damage by energetic Ar⁺ ions and to neutron irradiation in fusion reactors. Furthermore, Cui et al. [131] investigated the irradiation resistance of MA+SPS-ed W₉₀(TaVCrTi)₁₀, W₈₀(TaVCrTi)₂₀, W₇₀(TaVCrTi)₃₀, and W₆₀(TaVCrTi)₄₀ HEAs when exposed to 60 eV He⁺ ion. The HEAs showed excellent surface stability when exposed to 60 eV He⁺ ion irradiation. However, the W₇₀(TaVCrTi)₃₀ showed the best surface stability among the four HEAs. It was discovered that when the irradiation fluence was increased, the HEAs displayed a nearly flat surface irrespective of the irradiation fluence. The authors attributed the higher surface stability of the HEAs to their self-healing mechanism and ultrafine grain structure. The irradiation temperature did not significantly affect the HEAs, such that the irradiation damage of the W₆₀(TaVCrTi)₄₀ was slightly lower than that of the W₈₀(TaVCrTi)₂₀ and W₉₀(TaVCrTi)₁₀ alloys.

Zhou et al. [132] found that nanocrystalline Al _x CoCrFeNi high-entropy alloys with varying Al concentrations (x = 0, 1, 2) exhibited grain growth when subjected to a high-entropy radiation treatment. The HEAs were synthesized using a high-energy ball milling process that lasted 15 h, with ethanol serving as the wetting agent. High-entropy alloys made of nanocrystalline Al _x CoCrFeNi were irradiated with 1 MeV Kr²⁺ ions at room temperature. In the case of the nanocrystalline Al _x CoCrFeNi HEAs, the crystal structure for x = 0 was a simple FCC crystal structure, whereas the crystal structures for the other two Al concentrations (x = 1 and x = 2) were BCC + FCC crystal structures. The Al _x CoCrFeNi HEAs studied by Zhou and co-workers experienced significant grain growth under 1 MeV Kr²⁺ ions, with grain sizes increasing with irradiation dose from 13.8 ± 3, 7.4 ± 1, and 11± 1 nm before irradiation to 36 ± 8, 25 ± 5, and 26.6 ± 3 nm at an irradiation dose of 5.625 dpa. However, no phase transformation was found for the Al-0 alloy. While under ion irradiation, it was discovered that Al-2 alloy had the highest grain growth rate due to its lowest cohesive energy, which resulted in the lowest activation energy for atomic jump. According to the results of the irradiation test, a general trend of radiation-induced growth for nanocrystalline Al _x CoCrFeNi high-entropy alloys was observed. The initial rapid grain growth of Al _x CoCrFeNi high-entropy alloys was attributed to a disorder-driven mechanism caused by the loss of crystalline order because of ion-irradiation-induced large lateral damage volume. Relatively slow grain growth was attributed to the defect-driven mechanism near grain boundaries created by the radiation-induced point defects.

4.6 Thermoelectricity

Thermoelectric technology has attracted the interest of the scientific community due to its ability to convert heat into electricity in a solid state. It possesses a wide range of advantageous characteristics, including environmental friendliness, virtually no maintenance, zero noise, scalability, portability, and a long service life span [133]. High-entropy and medium-entropy alloys have intrinsically low lattice thermal conductivity due to effective scattering of phonons caused by lattice disorder, which results from the formation of lattice disorder in these alloys. The lead and tin chalcogenides PbTe, PbS, PbSe, SnTe, and SnS have all been proposed as promising and potentially thermoelectric materials, and their properties have been investigated.

According to Raphel et al. [134], the high entropy phenomena that resulted in low thermal conductivity in the BiSbTe₁.₅Se_1.5 thermoelectric alloy are discussed. By combining mechanical alloying (5 h at 300 rpm) with SPS (325°C at 50 MPa), the BiSbTe_1.5Se_1.5 thermoelectric HEA was produced. In the case of mechanically alloyed HEA powder, the XRD peaks confirmed the presence of a monophase BiSbTe_1.5Se_1.5 alloy with a rhombohedral crystal structure. Since the alloy's absolute Seebeck coefficient (S) value was negative, it could be assumed to exhibit n-type semiconducting properties due to the contribution of electrons. In addition, the Seebeck coefficient increased with the increase in the temperature, rising from 117 μV·K⁻¹ at 373 K to 124.97 μV·K⁻¹ at 438 K. As the temperature was further increased, S decreased slightly, reaching 119.15 μV·K⁻¹ at 523 K. Electrical conductivity (σ) increased monotonically from 2.85 × 10⁴ S·m⁻¹ at 373 K to 5.51 × 10⁴ S·m⁻¹ at 523 K, with the highest value occurring at 373 K. The enhanced configuration in BiSbTe_1.5Se_1.5 was due to the random and homogeneous atomic distribution, which could improve the crystal structure symmetry and electronic transport properties of the material [134]. In general, the high entropy and nanocrystalline features have a significant impact on the high configurational entropy because of the higher chemical disorder and lattice distortion in the material.

The thermoelectric properties of the block textured BiSbTe_1.5Se_1.5 HEA were investigated by Ivanov et al. [135]. The HEA was synthesized using SPS at 40 MPa and a temperature of 723 K for 15 min. The texturing axis was aligned with the direction of the SPS pressurization. The lamellar grain structure was formed because of the texturing process. The lamellar sheets did not cover the entire volume of the textured sample and are not continuous. Although there were blocks that contained continuous lamellar sheets that were oriented in a specific direction, the orientations of the sheets in neighboring blocks differed from one another. Specific electrical conductivity and thermal conductivity were measured perpendicularly and parallel to the texturing axis to determine the development of anisotropy because of the texturing process. Due to its low lattice thermal conductivity, the thermoelectric performance of the material was found to be suitably promising.

Kush et al. [136] studied the thermoelectric behavior of mechanical alloyed nickel base Ni₂CuCrFeAl _x HEA and discovered that when the Al = 0.5, the Seebeck coefficient was 256.66 μV·K⁻¹ at 660 K and then decreased to 35.06 μV·K⁻¹ at 850 K. It was also found that when the Al content was greater than 0.5, there was an exponential rise in figure of merit (ZT) at ∼0.2731 at 850 K. This indicated that the ZT of Ni₂CuCrFeAl_0.5 alloy was strongly affected by Al-concentration. Tin telluride (SnTe) is a representative compound with a narrow bandgap and promising ZT values. Yang et al. [137] enhanced the thermoelectric performance in Sn_0.25Pb_0.25Mn_0.25Ge_0.25Te HEAs through SnTe alloying and composition tuning. A ZT value of ∼1.4 at 823 K was achieved, which was higher than the ZT value of 1.07 in (Sn_0.7Ge_0.2Pb_0.1)_0.9Mn_0.11Te HEAs [138]. At room temperature, the Seebeck coefficient increased to ∼103 μV·K⁻¹ higher than the 82 µV·K⁻¹ recorded for (Sn_0.7Ge_0.2Pb_0.1)_0.9Mn_0.11Te HEAs [138].

In a separate study, Raphel et al. [139] investigated the synthesis of nanocrystalline PbSnTe HEA and PbSn₀.₈₇₅TeSeBi_0.125 HEA through mechanical alloying (5 h) and SPS (325°C). It was observed that the XRD peaks of sintered HEA samples corresponded to single-phase PbSnTeSe HEA with a NaCl type FCC crystal structure. The absolute Seebeck coefficient (S) result revealed that both the pristine and Bi-doped PbSnTeSe HEAs had S values in the positive range. This confirmed the p-type semiconducting behavior of the HEAs, with holes serving as the major charge carriers in both samples. The p-type semiconducting behavior was also observed by Fan et al. [140] in their study on the thermoelectric performance of PbSnTeSe high-entropy alloys. The peak value of S was also recorded at 623 K, with a value of 159.66 μV·K⁻¹. In the nanocrystalline PbSnTeSe, substitutional doping of Bi caused the S to increase dramatically, reaching a maximum value of 194.85 μV·K⁻¹ at 623 K. The increase in Seebeck coefficient that the thermoelectric alloy experienced was attributed to an increase in configurational entropy, which is believed to aid the modification of the alloy's band structure by increasing the phonon disorder. The electrical conductivity (σ) of the PbSnTeSe HEA gradually decreased from 6.13 × 10⁴ S·m⁻¹ at 373 K to 2.63 × 10⁴ S·m⁻¹ at 623 K, whereas the PbSn₀.₈₇₅TeSeBi_0.125 HEA gradually decreases from 5.95 × 10⁴ S·m⁻¹ at 373 K to the lowest value of 2.63 × 10⁴ S·m⁻¹ at 623 K. The high configurational entropy, higher chemical disorder, lattice distortion, and nanocrystalline features of both synthesized HEAs all contributed to achieving an ultralow thermal conductivity of less than 0.9 W·m⁻¹·K⁻¹ in both HEAs. It had a higher ZT of 0.71 than the PbSnTeSe HEA and very low thermal conductivity. This is because the band engineering makes it possible to get a high Seebeck coefficient and low thermal conductivity.

4.7 Deformation studies on PM processed-HEAs

To manufacture sheets, pipes, bars, and intricate parts from metallic alloys, hot forming technology, such as hot rolling or hot forging, is typically used. When metallic materials are subjected to hot deformation processes, their microstructures and properties can be enhanced, and the optimum parameters for processing such materials can be established. For example, good strength and ductility are often achieved during hot deformation processes due to grain refinement. It is for this reason that several researchers have used both experimental and computational techniques to provide further understanding of forming processing in many metallic systems. With respect to the processing of HEAs, most hot deformation studies were conducted on arc melted HEAs, with only a few studies on the hot deformation behavior of PM-HEAs. This may be ascribed to the near-net shaped processing nature of most PM components. The phenomena observed during hot forming behavior of PM-HEAs is like those of arc melted HEAs and conventional alloys. Some of the recent studies that focused on this topic are discussed in the subsequent paragraphs.

While assessing the hot deformation behavior of a MoNbTaTiV refractory HEA fabricated by PM, Liu et al. [141] identified that grain boundary gliding and rotation were the dominant deformation mechanisms when the alloy was deformed in the temperature range of 1,100–1,300°C and strain rate range of 0.0005–0.5 s⁻¹. After hot uniaxial compression testing, they found that the flow stresses were sensitive to the change in strain rate and deformation temperatures. When the deformation temperature was raised from 1,200 to 1,300°C, the maximum compressive stress was rapidly reduced at strain rates of 0.05 and 0.005 s⁻¹. Similar trends of compressive stress reduction at high temperatures were also reported in as-cast refractory HEAs (RHEAs) [142,143,144]. Furthermore, the microstructural characteristics of the deformed MoNbTaTiV alloy shown in Figure 9a reveal nearly-equiaxed grains, except for some deformation microstructures observed at 1,100–1,200°C and 0.05 s⁻¹ strain rate. Equiaxed grains microstructure have been reported for as-cast HEAs hot deformation at temperature of 1,000–1,100°C [145,146,147].

Figure 9

(a) Inverse pole figure maps of the deformed MoNbTaTiV RHEA at different deformation temperatures and strain rates and TEM images of the deformed MoNbTaTiV RHEA showing (b) and (c) bulging grain boundaries and (d) discontinuous dynamic crystallized grains and continuous dynamic crystallized grains. Reproduced with permission [141], Elsevier.

With the increase in deformation temperature and decrease in the strain rate, grain sizes increased though they were still ultrafine, suggesting the thermal stability of the RHEAs. This phenomenon was attributed to the characteristic sluggish diffusion effect caused by severe lattice distortion. As shown in the TEM images in Figure 9b, the grain boundaries of the deformed MoNbTaTiV RHEA are the nucleation sites for the ultrafine grains. Furthermore, smaller bulging shapes are observed to be present in Figure 9c. Consequently, several ultrafine discontinuous dynamic recrystallized grains are produced (Figure 9d). The bulging of grain boundaries, which was caused by the nucleation of new grains and subsequent grain growth, distinguished discontinuous dynamic crystallized grains from other types of crystallized grains [148]. The bulging phenomenon was also observed in some studies on the hot deformation of arc-melted RHEA alloys [143,144,149]. Severe extreme lattice distortion at the local grain boundaries, which has a strong pinning effect on the movement of the dislocations, can cause this phenomenon to occur. There is a restriction on the movement of grain boundaries during the bulging process, which leads to a very small bulging size [141].

Abhijit et al. [150] studied SRS in sintered CoCrFeMnNi HEA by nanoindentation. It was found that the SRS was negative (−0.0206), which was attributed to a cumulative effect of dislocation–grain boundary interaction, dislocation–interphase boundary interaction, dislocation–twin boundary interaction, and a highly frictional HEA lattice. Rymer et al. [151] investigated the SRS deformation behavior of sintered Al₀.₃CrFeCoNiMo_0.2 HEA under tension and compression at strain rates of 10⁻³, 10, and 10² s⁻¹. It was discovered that the SRS of the deformed HEA was quite low, particularly when tensile loading was applied to it; 0.0072 at a plastic strain of 5% and 0.0066 at a plastic strain of 10%. For each investigated plastic strain, the SRS (m) value was 0.0149 at 5%, 0.0121 at 10%, and 0.0101 at 20% under compression. This was almost twice as high as deformation under tensile loading. A further observation was that, for both tension and compression, the SRS value showed a slight decrease with the increase in the plastic strain.

This is demonstrated by the fact that the m value of the deformed HEA at a plastic strain of 20% under compression is approximately 33% lower than the m value at 5% plastic strain. In as-cast Al_0.1CoCrFeNi HEA, the SRS of the flow stress changed from positive to zero to negative when tensile loading was applied. Therefore, the negative SRS was attributed to dynamic strain ageing of dislocations by the solutes created by aluminum atoms (Komarasamy et al., [152]). Likewise, in as-cast Al_0.3CoCrFeNi HEA, both negative SRS and dynamic strain ageing were observed [153]. This discovery demonstrated that the strain rate sensitive behavior was asymmetric in both tension and compression.

4.8 Summary of functional applications of PM-processed HEAs

The summary of important findings in some published articles (2019–2022) relating to corrosion, thermoelectric, oxidation, tribological, deformation, hydrogen storage, and radiation properties of PM-HEAs are presented below.

• It was revealed that the deformation temperature and the strain rate had a significant effect on the flow stresses. The dominant deformation mechanism was the discontinuous dynamic recrystallization mechanism. In addition, equiaxed grain microstructure and grain boundary bulging were found, which were comparable to as-cast HEAs during high temperature deformation. Negative SRS occurred in sintered HEAs like what was reported for as-cast HEAs.

• Existing PM-HEA literature showed that composition tuning and the addition of dopants, such as Bi and Ga, were instrumental in attaining a higher ZT value in HEA thermoelectric alloys. The dopants increased the bandgap and created a directional anisotropy effect, which increased the absolute Seebeck coefficient and ultralow thermal conductivity that resulted in a higher ZT.

• The addition of 10 wt% of TiB₂, Cr₃C₂, and ZrO₂ reinforcement particles in the CoCrFeNi HEA matrix increased its wear resistance. In addition, the hardness of HEAs was enhanced with the increase in the weight fractions of hard ceramic particles. The decrease in friction coefficient of the composite containing HEAs could be attributed to oxide layer formation and matrix softening effects at higher temperatures.

• As appears in as-cast HEAs, spallation due to high thermal stresses developed within the substrate and the oxide layer in PM-HEAs CoCrFeMnNi. The addition of Mn to FeCoCrNi resulted in the formation of Mn-based oxide scales, which did not provide protection against oxidation.

• It is worth noting that, due to its lightweight and its monohydride formability, the gravimetric capacity of the Mg₁₂Al₁₁Ti₃₃Mn₁₁Nb₃₃ alloy is still competitive with other refractory HEAs that form dihydrides. After hydrogen desorption in MgAlTiFeNi, it was found that the hydride-related peaks of Mg₂FeH₆ had disappeared, which meant that the hydride had decomposed into Mg and Fe with the simultaneous release of hydrogen gas.

• As the corrosion potential and current density increased, polarization resistance decreased with the increase in the annealing temperature. The formed sigma phase during the annealing process decreased the corrosion resistance of PM-HEAs. The addition of 0.01 M coumarin to Hanks’ solution acted as a corrosion inhibitor and increased the charge resistance of all MgZnFeCuCo alloys regardless of their composition. A carbon nano tube of 2 wt% reduced the corrosion rate of FeCoCrNiCu HEA in a 3.5% NaCl solution.

• During the irradiation of SPS-ed CuCrFeTiV, higher surface stability was attributed to self-healing mechanism and ultrafine grain structure of HEA.