3.2. Subject Categories Co-Occurrence Analysis
An article that is indexed by the Web of Science(WoS) usually belongs to one or more subject categories. The co-occurrence analysis of subject categories contributes to detecting the disciplines involved in intellectual development and the interdisciplinary characteristics of a specific knowledge domain [
30]. Phytoremediation technology has obvious interdisciplinary characteristics; for example, it is well known to the public and involves plant science, environmental science, and agricultural science. In this section, the co-occurrence analysis of the subject category was used to obtain detailed categories involved in the research of phytoremediation of HMs and determine its interdisciplinary characteristics. In this research, the node/link colors were used to signify the corresponding year from 1989–2018, and the gray color from light to dark corresponds with 1989–2008, whereas purple to red corresponds with 2009–2018 (
Figure 4). It is worth noting that the links’/nodes’ colors in Figures 5 and 7–9 have the same means.
A total of 6842 records in the 6881 search results had valid subject categories. These subject categories belonged to 117 unique subject categories. A total of 68 nodes and 93 links were identified in the subject categories co-occurrence network. As shown in
Figure 5, we found that research on phytoremediation of HMs is multifaceted and covers a wide range of interests. Environmental Science and Plant Sciences are two primary subject category groups and have the characteristics of being multidisciplinary. The environmental science category group mainly includes “Environmental Sciences & Ecology”, “Environmental Sciences”, “Engineering, Environmental”, and “Toxicology”, whereas the plant sciences category group mainly includes “Soil Science”, “Plant Sciences”, “Agriculture”, and “Agronomy”. An increasing number of subject categories have been involved in this research field in recent years [
31,
32,
33]. According to the co-occurrence analysis of the subject category, “Environmental Sciences & Ecology”, “Environmental Sciences”, and “Plant Sciences” are obviously larger than other nodes, have more frequent co-occurrences, and can be identified as research hotspots among the subject categories.
On the basis of the colorful circle analysis, the earliest research was mainly conducted in the “Plant Sciences”, “Environmental Sciences & Ecology”, and “Environmental Science” fields, which had gray inner circles. On the other hand, nodes such as “Water Resources”, “Mining & Mineral Processing”, and “Mineralogy” had colorful inner circles, indicating that they were recently incorporated into the research of phytoremediation of HMs.
Furthermore, we also found that nodes such as “Chemistry, Inorganic & Nuclear”, “Chemistry”, “Biochemistry & Molecular Biology”, “Ecology”, and “Chemistry, Multidisciplinary” were marked with thick purple outer circles; they are connected to multiple nodes or serve as channels for other nodes, indicating that they play an important role in the subject category co-occurrence network. Without these nodes, the whole network would become very loose, with isolated nodes.
The top 10 most frequent categories and the turning points in the network according to the subject category co-occurrence analysis result are listed in
Table 1 and
Table 2. Furthermore, among the 117 unique subject categories, occurrence bursts were detected. Bursts were found in 25 subject categories, and the nodes with strong bursts are listed in
Table 3.
While the co-occurrence frequency of nodes represents the output efficiency of papers in a given category, the BC score means that the papers published in that category are given attention by the scientific peers involved in the topic of the category. Therefore, categories that have both high-frequency and high BC scores are category hotspots in the phytoremediation of HMs. Thus, according to
Table 1 and
Table 2, categories such as “Environmental Sciences & Ecology”, “Plant Sciences”, and “Agriculture” are basic supporting categories of research on phytoremediation of HMs. Categories such as “Chemistry, Inorganic & Nuclear”, “Biochemistry & Molecular Biology”, “Chemistry”, “Ecology”, “Chemistry, Multidisciplinary”, “Mining & Mineral Processing”, “Engineering, Chemical”, “Metallurgy & Metallurgical”, and “Engineering” also play a vital role in the interdisciplinary research of phytoremediation of HMs.
According to the burst detection results in
Table 3, “Plant Sciences” had the largest burst strength at 57.7754, the earliest burst beginning year (1991), and the longest burst duration of 16 years. This was followed by “Agronomy”, which had a burst beginning at the year of 1997, a burst strength of 15.9624, and a burst duration of 11 years. These two categories are the origin categories in the research of phytoremediation of HMs, and they have long been dominant in the research process of phytoremediation of HMs [
34,
35]. Similarly, on the basis of the subject category co-occurrence network burst detection, “Science & Technology-Other Topics”, which had a burst beginning at the year of 2016, and “Engineering, Multidisciplinary”, “Engineering, Chemical”, and “Green & Sustainable Science & Technology”, with burst beginning at the years of 2014–2016, represent a new trend in the research and development of phytoremediation of HMs, and these categories can be identified as trends in the subject categories. These subject categories are generally applicable, focused on engineering, and comprehensive, and may provide evidence for the application of engineering practices to phytoremediation of HMs [
36,
37,
38,
39].
3.3. Keywords Co-Occurring Analysis
Keywords can be seen as the soul of an article [
25], reflecting the core content of the article in a concise form. Keyword co-occurrence analysis is useful for tracking the development and evolution of research hotspots, as well as obtaining the research trends of a certain knowledge domain. A total of 9823 valid keywords were found in the 6881 search results. According to the keyword co-occurrence analysis results, the top 10 most frequently occurring keywords and the turning points in the network are listed in
Table 4 and
Table 5. Furthermore, the nodes with strong bursts are listed in
Table 6.
Among the top 10 keywords (
Table 4), “phytoremediation” was ranked first, with a frequency of 2833, followed by “heavy metals” (2648) and “phytoextraction” (2197). It can be found that the keyword co-occurrence results are greatly influenced by the human factor of the retrieval strategy. For example, “phytoremediation”, “heavy metals”, “phytoextraction”, and “cadmium” were the top four most frequently occurring keywords in the network, and they were also the words in the retrieval strategy. Thus, in this section, the word cloud maps (
Figure 6) of the authors’ keywords, keywords plus, titles, and abstracts were drawn on the basis of the text mining function of the Bibliometrix software. In this step, the search terms that appeared in the retrieval strategy were removed from the text mining results, and close synonyms such as “Cu” and “coper”, and singular and plural forms such as “plant” and “plants”, were combined. The high-frequency keywords obtained by the co-occurrence analysis function of CiteSpace greatly interfered with the retrieval strategy. Thus, in this research, the frequency analysis was based on the word cloud result. The most frequent keywords in
Figure 6a–d include “accumulation”, “plants”, “contaminated soils”, “tolerance”, “contaminated”, “soils”, “uptake”, “hyperaccumulator”, “
Thlaspi caerulescens”, “Indian mustard”, “potential”, “hyperaccumulator
Thlaspi caerulescens”, “toxicity”, “growth”, “effects”, “hyperaccumulation”, “EDTA” (ethylenediaminetetraacetic acid), “rhizosphere”, “translocation factor”, and “bioavailability”.
According to
Table 5, several keywords, such as “
Thlaspi caerulescens”, “hyperaccumulator”, “phytochelatin”, “organic acid”, and “nickel”, had high BC scores of 0.58, 0.52, 0.52, 0.35, and 0.31, respectively. According to
Table 4 and
Table 5, keywords such as “soil contamination”, “hyperaccumulator”, “EDTA”, “rhizosphere”, “bioavailability”, “oxidative stress”, “chelating agent”, and “organic acid”, which not only had a high BC score but also had a high frequency of co-occurrence, play key roles in connecting various research topics, have a significant influence on the development of phytoremediation of HMs, and can be recognized as hotspots in the research domain.
Among the 9823 keywords, 72 keywords had occurrence bursts from 1989–2018. Among them, the nodes with high burst strength, long burst duration, and the most recent burst are more attractive to researchers and were listed in
Table 6.
According to
Table 6, keywords such as “population” had the highest burst strength (27.1522) from 2001 to 2008. This was followed by “transport” and “plant growth”, with burst strengths of 25.9043 and 25.7854, respectively. Moreover, keywords such as “transport” had the longest burst duration of 16 years, followed by “Brassicaceae” and “cadmium uptake”, with a burst duration of 11 and 10 years, respectively. These findings indicate the prominent role of these research subjects in the phytoremediation of HMs. It was worth mention that several keywords occurred burstness in recent years: “plant growth” and “water” burst in 2015, with burst strengths of 25.7854 and 19.2366, respectively, and a burst duration of 4 years. Keywords such as “bioma”, “bacteria”, and “mine tailing” had a burst period from 2016 to 2018 and burst strengths of 18.7335, 12.9124, and 8.3637, respectively. This represents the attention paid by peer researchers to these fields and the research trends in recent years.
Furthermore, the characteristics of the knowledge structure evolution over time in the field of phytoremediation of HMs were explored on the basis of the keywords’ burst times.
In the period from 1989 to 1999, there was only one keyword (“plant”) that had an occurrence burst. However, there were 38 keywords with occurrence bursts from 2000 to 2010. In this period, keywords with occurrence bursts can be divided into three groups:
The plants or hyperaccumulator group included the keywords Thlaspi caerulescens, Arabidopsis, Silene vulgaris, Pteris vittata L, Salix, Zea may L, nickel hyperaccumulator, Brassicaceae, Holcus lanatus L, and Indian mustard.
The remediation mechanisms group (including physiology and biochemistry of plants) included the keywords metal tolerance, localization, compartmentation, leaves, HM detoxification, glutathione, response, cellular compartmentation, phytochelatin, iron, phytotoxicity, and expression.
The enhancements technology group included the keywords Saccharomyces cerevisiae, phosphorus, revegetation, population, phosphate, assisted phytoextraction, and chelating agent.
In the period of 2011 to 2018, 32 keywords had occurrence bursts, and these keywords could be divided into four groups:
The plants or hyperaccumulator group included the keywords Helianthus annuus, hyperaccumulator Thlaspi caerulescens, Arabidopsis halleri, fern, Zea mays L, and Sedum alfredii.
The enhancement technology group (exogenic substances) included the keywords citric acid, trace element, organic acid, EDD, availability, and enhanced phytoextraction.
The enhancement technology group (bacteria/microorganisms) included the keywords bioma, bacteria, Arbuscular mycorrhiza, plant growth, and amendment.
The engineering/field application group included the keywords phytomining and mine tailing.
Overall, considering the burstness detection results of the keyword co-occurrences, it can be concluded that the research hotspots of phytoremediation of HMs mainly have the following four points: HM hyperaccumulators, uptake mechanisms of HMs, the enhancements of technology for HM uptake, and the engineering/field application of phytoremediation of HMs.
The cluster analysis of the keyword co-occurring network can be used to obtain the distribution of the research on phytoremediation of HMs, and the timeline view can be used to more intuitively analyze its evolution (
Figure 7). In this research, the keyword co-occurrence network is divided into 12 co-citation clusters. On the basis of the log-likelihood ratio (LLR) algorithm and according to the titles of their own citing articles, the co-citation clusters were obtained and automatically labeled with the format “# + number + Label” with CiteSpace. As shown in
Figure 7, we were more interested in the clusters with colored labels. The mean years of these clusters were distributed over the last decade and they were defined as the active clusters in this study. We were particularly interested in these clusters because they are more likely to represent the research trends of phytoremediation of HMs.
In this research, the co-occurrence network clustering results had a mean
Q value of 0.7747 and a mean
S value of 0.9396, which indicated high reliability of the results. The largest cluster (#0) had 19 members, and the
S value was 0.947. It was labeled as “arsenic hyperaccumulator” by the LLR algorithm, “effect” by the TFIDF (term frequency-inverse document frequency) analysis, and “plant” by the MI (mutual information) analysis. The most active citation to cluster #0 was a research paper on the effects of different single-metal concentrations of Cd, Ni, and Cu and their subcellular distribution in
Brassica juncea L. var. megarrhiza [
40]. The second-largest cluster (#1) had 18 members and an S value of 0.869. It was labeled as “novel mechanism” by LLR, “plant” by TFIDF, and “different metal-bearing solid” by MI. It had two active citations: a research paper on zinc hyperaccumulation and cellular distribution in
Arabidopsis helleri [
41], and a research article that reports a newly observed novel mechanism of silicone uptake in plants (Zhao et al. 2000). This reflected that the mechanism of phytoremediation of HMs is the biggest research hotspot in this topic, and it is consistent with the conclusion of that based on the burst detection of the keywords. The third-largest cluster (#2) had 15 members and an S value of 0.762. It was labeled as “Pb-EDTA accumulation” by LLR, “zinc” by TFIDF, and “
Nicotiana tabacum” by MI. The most active citation to cluster #2 was research on the effect of EDTA on maize seedlings’ response to cadmium-induced stress [
42]. This reflects that the technological enhancements for phytoremediation of HMs is another research hotspot in this topic, and it is also consistent with the conclusion of that based on the burst detection of the keywords.
According to the timeline view of the co-occurrence network clustering results (
Figure 7), from 2017 to 2018, a total of four keywords (such as “wastewater” in cluster #0, “induced oxidative stress”, and “serpentine soils” in cluster #1, and “contamination” in cluster #7) had occurrence bursts from 2017 to 2018, which may represent the research trends in this field to some extent.
3.4. Reference Co-Citation Analysis
Scientific research needs to be based on the knowledge accumulated from relevant previous research. In other words, subsequently published papers usually rely on previously published literature and research results within a given subject or other related subjects as its references. It is generally believed that two papers that appear simultaneously in the same reference list have a co-citation relationship. Papers with co-citation relationships usually have intrinsic relations, and thus the relationship and structure in the academic field can be revealed with co-citation analysis.
In this research, a total of 6856 of 6881 publications in the local database had global citation records, whereas 1730 publications had local citation records. There was a total of 129,313 citation records, including 4339 references in the local database. On the basis of the analysis of the co-cited references from the documents,
Figure 8a displays a 404-node, 494-link reference co-citation network of the research on the topic of phytoremediation of HMs from 1989 to 2018. The reference co-cited network after the burst detection is displayed in
Figure 8b. In
Figure 8a, the colored lines represent the first co-citation that occurred in the last decade. The gray line represents that of 10 years ago.
There were several documents marked with purple circles or red centers, usually with a large BC score or high burstness, suggesting that they are probably landmark papers in the field and have attracted the most extraordinary degree of attention from its scientific community.
Thus, the nodes in the middle of
Figure 8a represent the early studies on the phytoremediation of HMs [
43,
44,
45], as the dominant color is gray (gray nodes and gray links), whereas both sides of the network, which are dominated by other colors (blue, green, yellow, or red links and colored nodes) represent the recent research [
46,
47,
48]. There are a large number of nodes with a large size in the early research period and some marked with purple rings, which represent the rich and extensive knowledge structure in the research of phytoremediation of HMs. The large number of large nodes appearing on both sides of the network represents the rapid development of this field in recent years. Furthermore, on the basis of the co-cited reference analysis, the burst detection, cluster analysis, and timeline view were carried out and displayed the evolution of the overall trends in the research of phytoremediation of HMs (
Figure 9).
According to the reference co-citation analysis result, the top 10 most frequently co-cited references (nodes with a large size) and the turning points (nodes with a high BC score) of the reference co-citation network from 1989 to 2018 are displayed in
Table 7 and
Table 8. Among 129,313 references, 235 references had occurrence bursts from 1989 to 2018. The top 10 strongest citation bursts are shown in
Table 9.
The most frequently co-cited document, with a co-citation frequency of 277, was a review article focused on the background, concepts, and future trends in the research of phytoremediation of HMs [
16] (listed at the top of
Table 7). This was followed by two other review articles about hyperaccumulators. van der Ent et al. clarifies the conditions for the use of the term hyperaccumulator and (re)defines some of the hyperaccumulator terminology [
49]; this publication had a co-citation frequency of 260. Krämer outlines the hyperaccumulator germplasm and reviewed the physiological, molecular, and genetic basis underlying metal hyperaccumulation and its evolution [
50]. These three review articles summarize the shortcomings of the current research, analyze the future trends of phytoremediation of HMs, and standardize the concepts and terminology related to phytoremediation and hyperaccumulators and the identification of a hyperaccumulator. These two pieces of research regulate the use of terms related to phytoremediation and hyperaccumulators and greatly promote research related to the phytoremediation of HMs.
The top-ranked document by BC score, with a BC score of 0.65, was a research article that was published in
Nature in 2008 (
Table 8) and clarified the contribution of HMA4 (heavy metal ATPase 4) to metal hyperaccumulation or hypertolerance [
57]. Van Zaal contributed the publication with the second-highest BC score (0.58) in 1999; he reported a novel
ZAT (Zn transporter of
Arabidopsis thaliana) gene that is involved in the transport of Zn in
Arabidopsis [
58]. The publication with the third-highest BC score (0.55) was contributed by Wu in 1999 and focused on the effect of exogenous chelate on the availability, uptake, and translocation of lead in the phytoextraction of Pb [
59].
An article by Ali et al. in 2013 had the strongest citation burst [
16]. Its burst lasted for 5 years, from 2014 to 2018, with a burst strength of 110.8981 (
Table 9). This article was also the most frequently co-cited document and is listed at the top of
Table 7. This was followed by the publication by van der Ent et al., which also appeared to be a highly co-cited document, as seen in
Table 7.
In the reference co-citation network, the cluster analysis function divided the co-cited references into several clusters according to the degree of the close association between the references. The references that occurred within the same clusters were tightly connected to each other and loosely connected to the other clusters. In this research, the reference co-citation network was divided into 19 co-citation clusters. The cluster results with burst detection are shown in
Figure 9a. The clustering results had a mean
Q value of 0.8828 and a mean
S value of 0.9354, which indicates a high reliability of the results. On the basis of the LLR algorithm and according to the titles of their own citing articles, the reference co-citation clusters were obtained and automatically labeled with CiteSpace with the format “# + number + Label”. As shown in
Figure 9a, we found that in the past 10 years, a total of nine co-citation clusters have occurred. They were arranged according to the cluster size in
Figure 9b,c. These clusters were defined as the active clusters in this study. We were particularly interested in these clusters because they were more likely to represent the research trends of the phytoremediation of HMs, and these clusters were analyzed in detail as follows:
The largest cluster (#0) had 40 members, and the
S value was 0.917. It was labeled as “EDDs deficiency” by LLR, had a mean year of 2001, and the most active citation in this cluster was a review article on the application of field crops in the phytoremediation of metal-contaminated land [
68]. In this review paper, the advantages and disadvantages of crops as HM pollution remediation plants were reviewed by Vamerali et al., and some technical measures that may be beneficial to improving the enrichment ability of crops for HMs were proposed. Tao investigated the effects of ethylenediaminetetraacetic acid (EDTA), triethanolamine (TEA), and citric acid on the extractability of metals from soil and HM uptake and accumulation in
Glycine max L. [
69]. Zaier reported that the exogenous addition of EDTA can significantly enhance shoot HM accumulation in
Brassica napus [
70]. Cluster #0 has the most members and contains a large number of large nodes and a series of nodes with strong citation bursts; thus, cluster #0 is one of the research hotspots in the field of phytoremediation of HMs. This cluster mainly focused on the effects of exogenous chelating agents on the enrichment of HMs in plants.
As the largest active cluster in references co-cited network, cluster (#1) was labeled as “
Arabidopsis thaliana” by LLR, with the mean year of 2006 and an
S value of 0.958. It had 27 members and the most active citer to the cluster was a review article on HMs hyperaccumulator [
46]. Rascio and Navari-Izzo reviewed the process and mechanism of plant hyperaccumulation of HMs. The importance of the regulation and expression of genes (such as members of the
ZIP (
Zinc-regulated transporter Iron-regulated transporter Proteins)
, HMA (
Heavy Metal transporting ATPases)
, MATE (
Multidrug And Toxin Efflu)
, YSL(
Yellow Strip1-Like)
, and MTP (
Metal Transporter Proteins) families) was summarized through comparative physiological and molecular analyses. In this review article by Rascio and Navari-Izzo, two classic enrichment hypotheses of HM uptake, the “elemental defense” and “joint effects” were summarized in detail. The authors also noted that although these hypotheses have been partially confirmed in laboratory experiments, they still need to be further studied to be verified in more plant varieties and more natural environments. Ó Lochlainn’s laboratory research found that tandem duplication and deregulation of HMA4 expression occurred during the process of HM hyperaccumulation in
Arabidopsis halleri and
Noccaea caerulescens. This study demonstrated that the parallel evolutionary pathways may be the basis for these two occurrences of Zn/Cd hyperaccumulation in the Brassicaceae and pointed out that novel cis-regulatory elements help to increase the expression of the HMA4 gene in
N. caerulescens [
71]. Cluster #1 had relatively more members, a later average year, and more nodes with strong citation bursts. It is reasonable to assume that the research field of cluster #1, which represents the phytoremediation mechanism/process of HMs in hyperaccumulators, represents one of the research trends of phytoremediation of HMs.
Cluster #2 was the second-largest active cluster in the reference co-cited network, with a mean year of 2008, an
S value of 0.983, and 24 members. It was labeled as “plant-associated bacteria” by LLR. The most active citation in the cluster was a research article on the promoting effect of endophytes on the uptake of HMs [
72]. The second most active citation in the cluster was a review article about the bacterial and fungal microbiota of hyperaccumulator plants [
73]. Thijs et al. provided insights into the research status of the potential for the typical plant-associated microbiota of hyperaccumulator plants for enhancing metal uptake by plants and discussed the positive impact of microbial-enhanced metal phytoremediation and phytomining. Therefore, cluster #2 represents the trend of research on the plant-associated microbiota-enhanced HM uptake for phytoremediation of HMs.
Cluster #3 was the newest active cluster in the reference co-cited network, with a mean year of 2012, an
S value of 0.977, and 24 members. It was labeled as “ultramafic soil” by LLR. The most active citation in the cluster was a research article on the agromining potential of nickel-hyperaccumulating plants [
74]. The field experiment of Pardo et al. found that nickel hyperaccumulators such as
Alyssum murale and
Leptoplax emarginata both have good application potential for nickel agromining. The second most active citation in cluster #3 was a review article on agromining systems for nickel recovery [
75]. Kidd reviewed cases of the implementation of agromining engineering practices for the restoration of natural metalliferous soils and summarized the positive effects of fertilization regimes, crop selection and cropping patterns, and bioaugmentation in nickel agromining. At the same time, Kidd pointed out that the subsequent development of this technology needs to be improved in the posttreatment of the hyperaccumulator biomass, and the hydrometallurgical process should be considered to replace the pyrometallurgical process [
76]. Therefore, cluster #3 represents the trend of research on agromining for phytoremediation of HMs.
Cluster #4 was the fourth largest active cluster in the reference co-cited network, with a mean year of 2007, an
S value of 0.974, and 24 members. It was labeled as “microbial communities” by LLR. The most active citation to the cluster was a research article about the effect of culturable bacteria on plant growth and HM availability [
77]. Kuffuner et al. found that the bacteria associated with Zn/Cd-accumulating
Salix caprea had a reliable effect on promoting the growth of
S. caprea and the potential to increase the uptake of Zn/Cd. The second active citation in this cluster was a review article about the approaches for enhanced phytoextraction of HMs [
47], in which the authors reviewed metal tolerance and accumulation mechanisms in plants and explored the effects of environmental and genetic factors on HM uptake. Thus, cluster #4 represents the trend of research identifying approaches for enhanced phytoextraction of HMs for phytoremediation.
Clusters #6, #7, and #9 were all focused on HM hyperaccumulators. The most active citation in cluster #6 was a part of the book series “Reviews of Environmental Contamination and Toxicology” that focuses on the phytoextraction of Cd [
78]. Shahid et al. comprehensively reviewed the biogeochemical behavior of Cd in soil–plant systems. This article was also the most active citation in cluster #9. The second most active citation in cluster #6 was a research article about the enhancing effect of an exogenous substance on HM uptake in remediations of HM-contaminated soil [
79]. Dary et al. evaluated the effects of
Bradyrhizobium sp. 750 and HM-resistant PGPR (plant growth promoting rhizobacteria) on the reclamation of multimetal-contaminated soil by
Lupinus luteus. The second active citation in cluster #9 was a research article about the tolerance and accumulation characteristics of
Siegesbeckia orientalis L. on cadmium [
80]. The most active citation in cluster #7 was a review article on hyperaccumulators [
46]. The second most active citation in cluster #7 was a research article on a cadmium hyperaccumulator that forces the cellular sequestration of Cd in
Sedum alfredii [
81]. Thus, clusters #6, #7, and #9 represent the research trends of phytoremediation of HMs by hyperaccumulators.
Cluster #10 was the 11th largest active cluster in the reference co-cited network, with a mean year of 2012, an
S value of 0.907, and 20 members. It was labeled as “castor bean” by LLR. The most active citation in the cluster was a review article on the positive role of plant-associated microbes in the process of phytoextraction of HMs [
82]; this review highlighted the plant–microbe–metal interaction in the phytoextraction process. Pérez-Palacios constructed a double genetically modified symbiotic system to enhance the phytostabilization of copper in legume roots [
83]. Thus, cluster #10 represents the trend of research on the enhancing effect of plant-associated microbes on HM uptake in phytoremediation of HMs pollution.
The 14th largest cluster (#13) had 19 members and an
S value of 0.974. It was labeled as “
Brassica napus” by LLR, phytoextraction by TFIDF, and emerging technology by MI. The most active citation in the cluster was a research article on the enhanced effect of chemical amendments (rhamnolipid, citric acid, and EDDs) on the phytoremediation of HMs [
84]. Zaier et al. reported another study on the effects of EDTA on the phytoextraction of HMs [
70]. Thus, cluster #14 represents the trend of research on the enhancing effect of exogenous substances on HM uptake in phytoremediation.