Study on the migration and transformation characteristics of arsenic and antimony in the rhizosphere of plants grown in zinc smelting waste residue

Background: In this paper, the physical and chemical characteristics of the rhizospheres of serrulata and Lolium perenne grown in indigenous Zn-smelting waste residue were studied, and the effects of the rhizosphere on the migration and transformation of arsenic and antimony in waste residue were investigated. Results: The results showed that compared with the control waste residue, the pH and Eh of the waste residues from the rhizospheres of the six plants decreased significantly ( P <0.05), and the dissolved organic carbon (DOC) content increased significantly ( P <0.05). The peak strength of hydroxyl groups, aliphatic groups, aromatic groups, calcite and quartz in the rhizosphere waste residue decreased to different degrees compared with that of the control waste residue, and the peak strength of carbohydrates/organosilicates increased. Overall, the contents of arsenic and antimony in the rhizosphere waste residues of the 6 plants were lower than those of the control waste residue, and the contents of arsenic(III) and antimony(III) were significantly higher than those of arsenic(V) and antimony(V), respectively ( P <0.05). The proportions of residual arsenic and antimony in the rhizosphere waste residues were the highest, and the proportion was significantly higher than that of control waste residue, and the proportions of exchangeable state, the aluminum-bound state and the calcium-bound state were lower than that in the control waste residue. The contents of arsenic and antimony in the tissues of the six plants followed the order of roots > leaves > stems, and the enrichment coefficients of arsenic and antimony in different plants were low. Conclusion: After 7 years of phytoremediation, the content and bioavailability of arsenic and antimony in the waste residue were significantly reduced, and the enrichment of arsenic and antimony in plant tissue was also low. It is speculated that the migration of arsenic and antimony in waste residue may increase in the early stage of phytoremediation and gradually stabilize in the later stage.

properties, toxicological properties, and geochemical behaviors, as well as metallic properties and nonmetallic properties [1]. Various arsenic and antimony pollutants have been shown to be toxic and carcinogenic to humans and organisms and to cause diseases of the liver, skin, and respiratory and cardiovascular systems [2,3]; furthermore, these compounds have been listed as priority pollutants by the EPA and EU [4]. The toxicity of arsenic and antimony in the environment depends on their morphology, and the toxicity of arsenic and antimony in general follows the order organic arsenic and antimony < pentavalent arsenic and antimony < trivalent arsenic and antimony [5,6]. Arsenic and antimony often co-occur in sulfur-rich ores [7], so cocontamination of arsenic and antimony is common in mining areas [8]. In recent years, the copollution of arsenic and antimony in the environment has received extensive attention from many scholars and has become a research hotspot in the field of environmental science [9, 10; 11].
Mineral mining, smelting and other human activities are important ways to accelerate arsenic and antimony transport into the environment. Northwestern Guizhou is a typical indigenous zinc smelting area in China. Indigenous zinc smelting slag is a highly heterogeneous and complicated mixture (including smelting slag, ceramic zinc smelting pot fragments, coal cinder, products of the incomplete combustion of coal, etc.) characterized by a poor physical structure, a poor nutrient content, alkaline pH, a variety of high-content heavy metals and low biological activity [12]. Under natural conditions, almost nothing grows in waste residue. Phytoremediation is generally considered an effective and low-consumption method to clean soil [13]. The use of plants for the restoration of waste residue yards and abandoned land is of great significance to effectively control the diffusion of heavy metals to the surrounding environment and to realize the restoration of polluted environments and damaged ecosystems.
As the rhizosphere microenvironment is the first to respond to plant pollution stress, it has become an important microdomain for studying the interactions among plants, soil and microorganisms due to its special physical, chemical and biological characteristics [14]. The rhizosphere microenvironment of plants plays an important role in the biogeochemical processes of heavy metals in waste residues [15], the physicochemical properties of tailings and microbial activity [16]. However, studies on the physical and chemical properties of rhizosphere residues under different plants and the migration and transformation characteristics of arsenic and antimony are rarely reported. Therefore, in this paper, six different types of plants that had been growing healthily for 7 years in an indigenous zinc smelting waste residue yard in northwestern Guizhou, China, were studied to explore the effects of the changes in the physical and chemical characteristics of the waste residue induced by plant activity on the migration and transformation characteristics of arsenic and antimony in the waste residue; the results provide a scientific basis for the phytoremediation of sites polluted with arsenic and antimony compounds.

Study area description
The study area is located in Qunfa village, Houchang town, Weining County, Guizhou Province, China (26°41 '14 "N, 104°43' 45" E). The average annual temperature is 10 °C, the average annual precipitation is 890 mm, the average annual sunshine hours are 1800 h, the frost-free period is 180 d, and the average altitude is 2200 m, belonging to the subtropical monsoon climate. In 2012, the research group carried out an ecological restoration project at an indigenous zinc smelting waste residue yard in the research area. The engineering measures taken were as follows: first, the waste residue yard was leveled; then, the natural mineral heavy metal passivator (containing 23% calcium and 17% silicon) and organic improvers (manure, plant litter and moss) was broadcasted to improve the physical and chemical properties of the waste dump. Finally, Lolium perenne, Broussonetia papyrifera, Cryptomeria fortunei, Arundo donax and Robinia pseudoacacia were selected as the pioneer plants to carry out ecological restoration. As the plant characteristics of the waste residue improved, Photinia serrulata and other native plants also heavily populated the waste residue.

Indigenous zinc smelting process
The main operating sequence of indigenous zinc smelting is ore dressing → crushing → furnace loading → heating and smelting → zinc extraction → dissolution and casting. First, after mechanical crushing, the zinc ore and coal are sieved at a certain mixing proportion, followed by blending. The material is loaded into a ceramic jar, and the gap between the jar and jar is filled with coal. Finally, the coal is burned, and the zinc ore in in the smelting pot is melted. The difference in the boiling points of lead and zinc enables zinc refining. During the smelting process, the temperature in the reaction section of the smelting tank can be as high as 1200°C or above, while the temperature at the top of the smelting tank is only approximately 800°C. In this way, the volatile gases from zinc smelting can be condensed, and crude zinc can be extracted.
The zinc ores used in smelting mainly include zinc sulfide ores (ZnS) and zinc carbonate ores (ZnCO 3 ).

Sample collection
The Zn-smelting waste residue was divided into two different types: (1) control waste residue area, where no vegetation was present, and (2) vegetated area, where six plant species Lolium perenne, Broussonetia papyrifera, Cryptomeria fortunei, Arundo donax and Robinia pseudoacacia were grown for seven years. The spacing of each plant was 1 m and 3 m, respectively. The vegetated area was divided into five quadrats, in December 2018, samples were collected from five quadrats in the study area. Samples of rhizosphere and non-rhizosphere waste residues of Lolium perenne, Broussonetia papyrifera, Cryptomeria fortunei, Arundo donax and Robinia pseudoacacia of similar age and growth were collected within each quadrat. Control waste residue samples and plant tissue samples were also collected. No-rhizosphere slag was randomly collected at a fixed depth of 0-20 cm at a distance greater than 50 cm from the rhizosphere of each plant species in order to avoid the impact of the root. The rhizosphere fraction of the waste slag was obtained from the root surface (0-5 mm, strong adherence to the root) after gentle shaking of the root by hand. Control samples of waste residue were randomly collected from the surface (0-20 cm) of the control waste residue area. Plant rhizosphere slag samples were collected according to Riley [17,18] by the shake-off method. After the collected waste residue samples were brought back to the laboratory, the excess material was removed, and samples of the same plant species from different sites were mixed evenly and air-dried to constant weight at room temperature. After collection and transportation to the laboratory, Broussonetia papyrifera, Cryptomeria fortunei, Arundo donax, Robinia pseudoacacia and Photinia serrulata samples were divided into roots, stems and leaves, and Lolium perenne samples were divided into roots and leaves.

Sample analysis
The pH and Eh were determined by the potentiometric method (waste residue:water = 1:2.5).
Referring to the method of Bolan et al. [19], DOC was extracted and determined by a TOC-2000 instrument (Shanghai Yuanxi Company). Free ferric oxide and aluminum were extracted with sodium bisulfite and determined by spectrophotometry [20]. Then, the contents of arsenic(V) and antimony(V) extracted by citric acid were determined by subtracting the contents of antimony(III) and arsenic(III) from the contents of arsenic and antimony, respectively [21, 22,23]. The classification method for the continuous extraction of phosphorus was adapted to divide the arsenic and antimony fractions into the exchangeable state, aluminum-bound state, iron-bound state, calcium-bound state and residual state [24,25]. The arsenic and antimony contents in the extract were determined by ICP-AES (Shimadzu 9820, Japan).
Determination of arsenic and antimony in plant tissues: Plant samples (0.5 g) were accurately weighed, and 3 mL of high-purity nitric acid was added for digestion. The contents of arsenic and antimony in the digestion solution were determined by hydride generation and atomic fluorescence spectrometry (AFS-8530, Beijing Haiguang).
Blank and parallel samples were prepared for each step of the above experiment, and the measurement results were averaged.

Data processing
All statistical analyses were carried out using SPSS for Windows (Version 25.0). The results are expressed as the means ± standard deviation. All data were analyzed using one-way ANOVA to assess the differences among treatments. The differences between the means were determined using the Duncan multiple range test and were considered as significant at p<0.05. Principal component analysis (SD) was performed by Canoco 5. And the graphics were drawn using Origin 2018.

Results
Influence of phytoremediation on physicochemical properties of waste residue The physical and chemical properties of the waste residue are shown in Table 1. After phytoremediation, there was no significant difference in pH value between the rhizosphere waste residue of Cryptomeria fortunei and the control waste residue (P>0.05). The pH value of the rhizosphere waste residues of other plants was significantly lower (P<0.05) than that of the control waste residue, and the pH values of the rhizosphere waste residues of the six plants were 0.02-0.42 lower than that of the control waste residue. The waste residue Eh ranged from 147.55-211.00 mV; the Eh of Lolium perenne rhizosphere waste residue was significantly higher (P<0.05) than that of the nonrhizosphere waste residue and control waste residue. The Eh of other plant rhizosphere waste residues (P<0.05) was significantly lower than that of the control waste residue and lower than that of nonrhizosphere waste residue; among the results, the Eh of Cryptomeria fortunei rhizosphere waste residue was the lowest, and that of Lolium perenne rhizosphere waste residue was the highest. The DOC content in the slag ranged from 6.14 to 20.77 mg·L -1 . Among the 6 kinds of plant rhizosphere slag, the DOC content was significantly (P<0.05) higher than that of the control slag; the value in the plant rhizosphere slag was increased by 12.15-14.63 mg·L -1 . The DOC content of the rhizosphere waste residue was significantly (P<0.05) higher than that of the nonrhizosphere waste residue. The DOC contents in the 6 kinds of plant rhizosphere slag, from high to low, followed the order Broussonetia papyrifera > Lolium perenne > Robinia pseudoacacia > Arundo donax > Photinia serrulata > Photinia serrulata.
The content of free iron oxide in the waste residue varied from 773.90 to 1917.17 mg·kg -1 . The content of free iron oxide in waste residues from the rhizospheres of Cryptomeria fortunei, Photinia serrulataand Lolium perennewas significantly higher than that of the control waste residue (P<0.05), and the content of free iron oxide in waste residue from the rhizosphere of Broussonetia papyrifera was significantly lower than that of the control waste residue (P<0.05). The content of free iron oxide in waste residues from the rhizospheres of Arundo donax and Robinia pseudoacacia was not significantly different from that of the control waste residue (P>0.05). The content of free iron oxide in the rhizosphere waste residues of Broussonetia papyrifera and Robinia pseudoacacia was significantly lower than that of nonrhizosphere waste residue (P<0.05), and the content of free iron oxide in the rhizosphere waste residues of Photinia serrulata and Lolium perenne was significantly higher than that of nonrhizosphere waste residue (P<0.05). The content of free iron oxide in Lolium perenne rhizosphere waste residue was the highest. The content of free aluminum oxide varied from 5053.72 to 7550.33 mg·kg -1 . The content of free aluminum oxide in the rhizosphere waste residue of Broussonetia papyrifera was significantly higher than that of the control waste residue (P<0.05), the content of free aluminum oxide in the rhizosphere waste residues of Arundo donax and Robinia pseudoacacia was significantly lower than that of the control waste residue (P<0.05), and the content of free aluminum oxide in the rhizosphere waste residues of Cryptomeria fortunei, Photinia serrulata and Lolium perenne was not significantly different from that of the control waste residue (P>0.05).
The content of free alumina in the rhizosphere waste residues of Broussonetia papyrifera and Robinia pseudoacacia was significantly higher than that of nonrhizosphere waste residue (P<0.05), and the content of free alumina in the rhizosphere waste residues of Arundo donax and Robinia pseudoacacia was significantly lower than that of nonrhizosphere waste residue (P<0.05). The content of free alumina in the rhizosphere waste residue of Arundo donax was the lowest, while the content of free alumina in the rhizosphere waste residue of Broussonetia papyrifera was the highest.   Physical and chemical properties of waste residue and principal component analysis of arsenic and antimony In general, pH, Eh, free ferric oxide and aluminum showed a strong positive correlation with arsenic and antimony and a negative correlation with DOC.
The two-dimensional sequence diagram showed that the control waste residue samples were distributed on the positive axis of PCA2, which was positively correlated with arsenic, antimony, Eh and free iron and aluminum oxides in the waste residue and negatively correlated with DOC, indicating that the control waste residue contained higher contents of arsenic and antimony in various forms but very low contents of DOC. Plant root waste residue samples were mainly distributed along the negative axis of PCA2, which was positively correlated with DOC in waste residue, indicating that plant root waste residue had a lower arsenic and antimony content and higher DOC content.

Discussion
Influence of phytoremediation on the physicochemical properties of waste residue Studies have shown that a decrease in pH value in the root environment is mainly attributed to the unbalanced absorption of positive and negative ions by plant roots, which results in the release of H + to neutralize the unbalanced charge of the plant roots [29,30]. Studies have also shown that the acidification of root exudates, litter and their degradation products and the release of CO 2 by root respiration during plant responses to environmental stress may also cause a decrease in pH value [31,32]. The acidification of the plant rhizosphere will accelerate the weathering of waste residue and affect the migration and transformation of arsenic and antimony in the waste residue, as well as the biological availability of these elements. The PCA analysis also showed that pH is positively correlated with arsenic and antimony.
Under the action of phytoremediation, the respiration of plant roots and microorganisms increased the partial pressure of CO 2 in the waste residue and resulted in a decrease in the Eh value of the rhizosphere waste residue. For terrestrial plants, rhizosphere respiration makes antimony(III) in slag and thus improves the activity and toxicity of arsenic and antimony in soils [33]. The PCA analysis indicated that Eh was significantly positively related to arsenic and antimony.  [35,36].
Therefore, the surface adsorption of iron and aluminum oxides an important mechanism for the fixation of arsenic and antimony in waste residues. The PZC of iron oxide is approximately 7; when the pH of the adsorption system is close to or more than the iron oxide PZC, the electronegativity of iron oxide will increase; the same charges of arsenic, antimony, and oxygen anion are mutually exclusive, leading to a decrease in the adsorption rate [37]. As a result, a lower pH of waste residue will increase the iron oxide adsorption of arsenic and antimony. The PZCs of various oxides of aluminum are approximately pH = 6, and these compounds present a net negative charge in neutral and alkaline environments. Therefore, the adsorption capacity of aluminum oxides for arsenic and antimony in waste residues is relatively weak. Arai et al. [38] also found the following in their study on the adsorption behavior of arsenic at the aluminum oxide-water interface: at pH = 4.5, the adsorption of aluminum oxide for arsenic in three days was almost 100%, while at pH = 7.8, only 46% of the arsenic was absorbed in approximately one year. The PCA analysis also showed that free iron and aluminum oxide were positively correlated with arsenic and antimony, indicating that free iron and aluminum oxide might be related to the fixation of arsenic and antimony in waste residue.
Influence of phytoremediation on the chemical structure of waste residue There are abundant active groups in waste residue, such as hydroxyl, carboxyl, phenolic hydroxyl, alcohol hydroxyl and other functional groups with acidic and substitution capacity, which can promote the adsorption of heavy metals by waste residue through ion exchange and complexation and play an important role in regulating the effectiveness of certain nutrients or heavy metals in the medium [39].
Wen et al. [40] also showed that two functional groups, hydroxyl and carboxyl groups, have a high affinity for heavy metals. Plant litter can increase the content of organic matter in soil; organic matter is mainly composed of polysaccharides, proteins, cellulose and lignin [41]. These substances are the main sources of carbohydrates, aliphatic groups and aromatic groups [42]. Furthermore, peptidoglycan, lipoprotein, cell proteins and other substances in soil microbial residues also increase the corresponding characteristic peak strength [43]. However, in this study, the peak strength of aliphatic and aromatic groups in rhizosphere waste residue was decreased, which may be related to the weathering and degradation of unburned coal remaining in the waste residue under the action of plants and microorganisms [44]. The characteristic functional groups in coal mainly include hydroxyl, phenolic hydroxyl, alcohol hydroxyl, carbonyl, amino, methoxy, aliphatic, aromatic and other groups [45].

Effects of phytoremediation on the valence characteristics of arsenic and antimony in waste residue
The oxidation rate of arsenic sulfide decreases significantly with decreasing pH value in the plant rhizosphere [50], and the oxidation rate of antimony(III) to antimony(V) decreases with increasing pH value [51]. Effects of phytoremediation on the occurrence of arsenic and antimony in waste residue Research has shown that the total amount of a certain heavy metal in the soil cannot be used to truly evaluate its environmental behavior and ecological effect and that the morphological content and proportion of heavy metal in the soil are the key factors that determine its impact on the environment and the surrounding ecosystem [55]. Exchangeable heavy metals are the most active in the environment and are easily absorbed, leached or transformed by organisms, with high bioavailability.
The ratio of exchangeable arsenic and antimony in the rhizosphere residues was lower than that of the control residues, which may be because the rhizosphere secretions and microbial activities of plants increased the availability of arsenic and antimony, promoted the absorption of arsenic and antimony, or activated arsenic and antimony that had migrated due to rainfall leaching. Research has shown [56,57] that microbial degradation can significantly increase the humic acid content in lignite.
Saada et al. [58] showed that the adsorption of humic acid on kaolinite in advance could provide new adsorption sites for arsenic, promote the adsorption of arsenic by kaolinite and reduce the activity of arsenic. A study by Guo et al. [59] showed that coal-based humic acid could significantly reduce the exchangeable arsenic content in soil, thus significantly reducing the available arsenic content in soil and promoting the growth of Chinese cabbage to some extent. Sun et al. [60] also showed that activation of weathered coal can reduce the mobility of arsenic in soil, increase the residual arsenic in soil, and reduce the toxic effect of arsenic on plants. After phytoremediation, the habitat of the waste residue was improved, and the microbial degradation of coal was enhanced, which may also be the reason for the decrease in the proportion of exchangeable arsenic and the increase in the proportion of residual arsenic. The decrease in free alumina content (Table 1) resulted in a decrease in the amount of arsenic and antimony adsorbed on alumina, which resulted in a decrease in the proportion of aluminum-bound arsenic and antimony in the plant rhizosphere waste residue. In neutral and alkaline soils, the precipitation of arsenic, antimony and Ca 2+ is an important process controlling the availability of arsenic and antimony in soil [36]. After phytoremediation, precipitates of arsenic, antimony and Ca 2+ may dissolve due to the decrease in pH value, leading to a decrease in the proportion of calcium-bound arsenic and antimony. Residual-state heavy metals generally exist in crystal lattices such as silicate, primary and secondary minerals, which are not easy to release under normal natural conditions, can exhibit long-term stability, and are not easily absorbed by plants. Their contents are mainly affected by mineral composition, rock weathering and soil erosion [61].
Exchangeable, calcium-bound, iron-bound and aluminum-bound arsenic and antimony in waste residue are activated and absorbed by plants or are affected by weathering and leaching, and their contents gradually decrease, which may be the reason for the increase in the proportion of arsenic and antimony in the residual state.

Distribution characteristics of arsenic and antimony in plants
The normal contents of arsenic and antimony in plants are < 1 mg·kg -1 and < 2.2 mg·kg -1 , respectively [62]. The contents of arsenic and antimony in the 6 plants grown on the waste residue mostly did not exceed these standards, and the enrichment coefficients of arsenic and antimony in the 6 plants were small, indicating that the enrichment capacity of arsenic and antimony was low.
Arsenic and phosphorus are homologous elements, and the chemical properties of phosphate and pentavalent arsenates are similar. Thus, in higher plants, arsenates and phosphates share the same absorption transporter [63]. The reason why some plants can adapt to soil with high arsenic contamination is often because the absorption of arsenates is reduced by regulating the expression of phosphate transporters [64]. The six plants grown on waste residues in this study may inhibit the absorption of arsenic through this mechanism, thus avoiding physiological toxicity. When Pratas et al. [65] investigated trees and herbs growing on soil contaminated with antimony (the average concentration of antimony was 663 mg·kg -1 ), they found that the content of antimony in the stems was less than 5 mg·kg -1 . Hammel et al.
[66] studied 19 crops planted in the soil of the abandoned mining area and found that although the background value of antimony in the local soil was very high (up to 500 mg·kg -1 ), the content of antimony in the locally produced rice storage tissue was only 0.09 mg·kg -1 , and the maximum content of antimony in the stems and leaves was 0.34 and 2.2 mg·kg -1 , respectively. These studies show that plants have a strong ability to reject antimony, thus increasing their tolerance to antimony, and this study has similar results.
Plants growing on waste residue containing heavy metal elements will absorb the heavy metals by store heavy metals in their roots or will collect the heavy metals in the rhizosphere, reducing the flow via transfer and diffusion to the ground, thereby reducing harm to photosynthesis, respiration, and the reproductive system to maintain normal plant growth [67]. The highest arsenic and antimony content in the roots of the six plants growing on the waste residue may be related to this.
Although the enrichment capacity of arsenic and antimony is much weaker than that of hyperaccumulator, the response of these 6 kinds of plants to the higher levels of arsenic and antimony in the waste residue with respect to natural growth suggests that plants grown in the extreme habitat of waste residue have strong ecological adaptability; this adaptability is manifested in plants via mechanisms that block or reduce the absorption of high contents of arsenic and antimony from the growth environment to protect themselves from biological poisoning.

Conclusions
In this study, we found that after 7 years of phytoremediation, the pH and Eh of the rhizosphere waste     Physical and chemical properties of waste residue and principal component analysis of arsenic and antimony. F1-5 and f1-5 represent exchangeable state, aluminum-bound state, iron-bound state, calcium-bound state and residual state arsenic and antimony respectively.