Arsenic (As) and cadmium (Cd) are two toxic and hazardous elements that strongly impact human health, agricultural product quality and environmental safety1. Due to the substantial differences in occurrence forms and chemical behaviours, it is difficult to remediate As and Cd together in cocontaminated water or soil2,3. Previous studies have shown that the availability of heavy metals is affected by complex soil variables, such as pH, redox potential (Eh), organic matter, Fe/Mn oxide content, and moisture and tillage management regimes4,5,6,7. Soil moisture affects soil pH and has opposite effects on the availability of As and Cd in soil8. For example, flooding can promote Cd adsorption on the soil surface but increase the available As9. Hence, developing an effective passivator and studying the simultaneous immobilization of As and Cd under different soil water management practices are necessary.

Nano-zero-valent iron (nZVI) is a commonly used passivator for As adsorption in contaminated water and soil remediation due to its strong adsorption capacity and high affinity for As10,11,12. However, it can easily form agglomerates and oxidize after long-term application under aerobic conditions, which may reduce its effectiveness in As remobilization13. In addition, previous works also reported that a single application of nZVI seemed to have a low affinity for contaminated heavy metal cations14, which may limit its application in heavy metal cocontaminated farmland. Hence, most researchers have attempted to increase its effectiveness and achieve the goal of co-immobilize of As and Cd by modification with other materials15,16. Compared with the previously reported modified carrier, sepiolite not only has properties such as alkalinity, porous structure, and high affinity for Cd17, but is also much less expensive (only 1 dollar per kilogram in China). Additionally, it has minimal toxic effects on soil ecology because it is a ubiquitous clay mineral in soil. Therefore, given the favourable characteristics of sepiolite, we believe that utilizing it as a carrier to modify nZVI can alleviate the aggregation and oxidation of nZVI, thereby achieving the goal of simultaneously immobilizing As-Cd in cocontaminated soil.

Soil bacteria constitute a significant active group in soil ecosystems, and their metabolic processes typically play essential roles in the fate of elements, encompassing chemical, physical, and biological aspects. Consequently, the impact of amendments on the soil bacterial community must be considered, irrespective of heavy metal remediation. However, the effect of sepiolite and nZVI on the soil bacterial community is still controversial. Some studies have indicated that sepiolite can inhibit the growth of microorganisms18, while others have discovered that sepiolite has no adverse effect and can even promote or restore the microbial community during remediation19. Like sepiolite, nZVI is a controversial amendment because it can be utilized as a nutrient by microorganisms at a proper dose, while microorganisms20,21 may suffer Fe ecotoxicity at high dosages during remediation22,23,24. Thus, whether sepiolite-modified nZVI can affect microorganisms and the potential impact of sepiolite-modified nZVI on microorganisms during soil remediation remain unknown. Further investigation is warranted to elucidate these aspects.

Our previous work showed that sepiolite-supported nano-zero-valent iron (S-nZVI) is an effective amendment for the simultaneous adsorption of As(III)-Cd(II) in aqueous solution systems25. Considering that less inexpensive, easy-to-synthesize and eco-friendly chemical amendments are applied for As-Cd co-rembolization in soil, we aimed to test whether S-nZVI is also a robust and effective amendment for the remediation of As-Cd cocontaminated soil, which may greatly enhance the application prospects of S-nZVI. Moreover, unlike in aqueous solutions, the ecotoxicity of amendments should also be well evaluated to avoid large adverse effects on soil ecosystems. Hence, the main objectives of this study were to (1) investigate S-nZVI characteristics and how S-nZVI improves soil properties under different water management practices; (2) evaluate the immobilization effect of S-nZVI on the bioavailability and fraction of As-Cd in cocontaminated soil; and (3) assess the underlying impact of S-nZVI on the indigenous soil bacterial community and the abundance of ___domain microorganisms across different soil‒water management conditions.

Materials and methods

Materials. The chemicals used in the present study were purchased from Aladdin Chemical Company, Ltd., China. All the chemicals were of analytical grade and were utilized without further purification. Sepiolite was purchased from Weicheng Mineral Limited Company, China.

Soil sample collection and synthesis of materials. Soil from a former mining-industrial site located in Liuyang, Hunan Province (28°29′ N, 113°88′ E), was cocontaminated with Cd and As, and the collected soil surface layer (0–20 cm depth) was air-dried and sieved through 2 mm mesh. The physicochemical properties of the soil were then measured, and three composite samples were analysed in duplicate. The pH of the soil was 6.03, the total soil organic carbon content was 19.86 g kg−1, and the total Cd and total As contents were 2.40 mg kg−1 and 127.18 mg kg−1, respectively.

Sepiolite was pretreated with 2 M HCl solution (as an ASEP). All the steps of the S-nZVI composite synthesis processes were conducted at room temperature and pressure. Briefly, 8 g of ASEP and 200 ml of 0.24 M FeCl3·6H2O solution were mixed. After 15 min of sonication, the mixture was stirred until it was sufficiently mixed. Subsequently, 200 mL of 0.98 M NaBH4 was added dropwise while stirring. After stirring for 1 h, S-nZVI was produced and magnetically separated from the mixture solution, washed with ethanol and deionized water several times, freeze-dried, and stored under dry conditions10,25.

Incubation experiment. 300 g soil samples included with S-nZVI were placed in 500 mL glass incubator, the incubator was sealed by a rubber cover. The amount of added S-nZVI was set to 1%, 3%, and 5% w/w, namely S-nZVI/soil mass ratio (recorded as SN-1, SN-3 and SN-5, respectively) in the incubation experiment. Then, the soil was stirred thoroughly again and subjected to three different moisture regimes: (a) moderately wet (MW), 70% water-holding capacity; (b) continuous flooding (CF), in which deionized water was added to the beaker to form a 2 cm layer of water over the surface soil (approximately 110% soil water holding capacity); and (c) an alternating wet and dry cycle (AWD), involving a transition from flooding (10 days) to 70% water-holding capacity (5 d) through natural evaporation (3 days), with the treatment repeated twice. All incubation treatments lasted for 36 days. Treatments without S-nZVI were used as the control group (CK), control and all the treatment were set as three replicates. Three sample bottles per treatment were collected at 2, 10, 28, and 36 days after incubation.

Soil properties and As and Cd availability and fraction analysis. The soil pH was measured at a soil‒water ratio of 1:5 (w/v) using a pH meter (PHS-3 C, China) according to the Chinese HJ 962–2018 standard. Soil EC was measured at a soil‒water ratio of 1:5 (w/v) by a soil electrical conductive metre (Huazhi EC-315, China). The available Cd or Fe in the soils was extracted by 0.01 M CaCl2 solution at a 1:10 ratio (w/v) with shaking for 2 h on an orbital shaker (Taicanghaocheng HZQ-F160, China) and determined by inductively coupled plasma optical emission spectroscopy (ICP‒OES Agilent 5110, USA). Available As in the soil was extracted for 2 h by the addition of 0.5 M NaHCO3 solution at a 1:25 ratio (w/v), as determined by atomic fluorescence spectrometry (AFS, Jitian 830, China)26. Free and amorphous iron oxides were extracted by citrate-bicarbonate-dithionite (DCB) and 0.2 mol L−1 ammonium oxalate (pH 3.0-3.2), respectively27, and determined by ICP‒OES. The speciation of soil As or Cd was determined by the modified sequential extraction method by BCR: the acid soluble fraction (F1) was considered bioavailable. In contrast, the reducible (F2), oxidizable (F3), and residual (F4) fractions were ascribed to immobilized As or Cd28. Soil properties and soil types were shown in Table 1.

Table 1 The properties of tested soil.
Full size table

Soil DNA extraction and high-throughput sequencing. Soil bacterial genomic DNA was extracted using an E.Z.N.A. A soil DNA kit (Omega Biotek, Inc., Doraville, GA, USA) was used following the manufacturer’s instructions. The concentration and integrity of the genomic DNA were determined by a NanoDrop 2000 spectrophotometer (Thermo Scientific Inc., USA) and 1% agarose gel electrophoresis. The bacterial 16S rRNA gene (V3-V4 hypervariable region) was amplified with the primer pair 338F (5’-ACTCCTACGGGAGGCAGCAG-3’) and 806R (5’-GGACTACNNGGGTATCTAAT-3’). PCR amplification system (60 µL): 2×Phanta Max Buffer (30 µL), 1.5 µL of dNTP mix (10 mM each), 3 µL of primer pair (10 µmol L-1), 1.5 µL of DNA polymerase, 1 µL of bacterial DNA (< 100 ng), 23 µL of ddH2O (Phanta Max Super-Fidelity DNA Polymerase, ZY505-d2 500 U, Vazyme, China); PCR amplification conditions: 95 ℃ for 5 min; 95 ℃ for 15 s, 55 ℃ for 20 s, 72 ℃ for 30 s, 32 cycles; 72 ℃ for 5 min (Mastercycler® pro S, Eppendorf, Germany). All PCR products were sent to Beijing Allwegene Technology Co., Ltd. (https://www.allwegene.com, Beijing, China) for Illumina MiSeq high-throughput sequencing. The raw high-throughput sequencing data have been uploaded to the NCBI Sequence Read Archive database (SRA, https://www.ncbi.nlm.nih.gov/sra/) and are available under the SRA accession numbers SRR25569156, SRR25569200, and SRR25569600.

Statistical analysis. Statistical analyses were conducted with Origin Pro2021, and the results are presented as the mean ± standard deviation (SD). Pearson correlation and linear regression analyses were used to determine the relationships of the tested indices. Bioinformatic analysis was performed on the online platform Allwegene Cloud (https://www.allwegene.com, Version V2) and Illumina Analysis Pipeline Version 2.6 (Illumina, Inc., USA).

Results and discussion

Characterization of S-nZVI. SEM results showed that the surface structure of nZVI exhibited a chain-like structure with apparent aggregation25,29. After nZVI was loaded onto the sepiolite surface, SEM showed that nZVI was well dispersed on the SEP outer surface and loaded into the pores (or cracks) (Fig. 1A). In addition, the EDS mapping results of S-nZVI also showed that Fe was homogeneously loaded on the SEP, indicating that nZVI was successfully loaded onto the SEP surface. (Fig. S1). The XRD patterns of SEP, nZVI and S-nZVI are presented in Fig. 1B. Peaks at 7.1°, 10.5°, 12.2° and 26–31° represent the characteristic diffraction peaks of SEP, and the peaks at 26.7°, 29.3°, and 43.25° are attributed to SiO2, CaCO3, and MgO, respectively. The nZVI and S-nZVI peaks observed at 2θ = 44.7° might indicate that Fe0 was successfully loaded on SEP, and Fe3O4 (2θ = 35.7◦), Fe2O3 (2θ = 65.4◦) and FeOOH (2θ = 39.09◦) were also present due to the oxidation of Fe0 (Fig. 1B)29,30.

Fig. 1
figure 1

(A) SEM images of sepiolite, nanoscale zero-valent iron (nZVI), and sepiolite-supported nanoscale zero-valent iron (S-nZVI); (B) XRD spectra of materials.

Full size image

Effect of S-nZVI on soil pH and EC. After 36 d of incubation, the soil pH differed among the three water regimes and increased in the general order of CF > AWD > MW (Fig. 2). The variation trend of the SN-1 dose was similar to that of the CK, indicating that a low dose had no significant effect on the pH or EC under the three water regimes throughout the incubation period. In contrast, the SN-3 and SN-5 doses significantly increased the soil pH. A comparison of the variation in soil pH under the different water regimes revealed that the soil pH under SN-3 and SN-5 in the CF treatment continuously increased. The possible reasons can be that (i) Fe0 reacts with H2O and forms OH- (Fe0 + 2H2O→Fe2+ + H2 + OH-), and (ii) alkaline sepiolite can also release OH- in the aqueous soil solution31,32,33. In contrast to the continuously increasing trend in the CF treatment, the pH of the AWD and MW treatments showed a fluctuating increasing trend and even presented a decreasing trend after 10–18 d of incubation (Fig. 2B and C). This means that periodic or continual low soil moisture may hinder the alkalinity of S-nZVI to mitigate the soil pH. In this process, the hydroxylation of iron oxides on the surface of nZVI occurs under aerobic conditions (low soil moisture), as does the hydrolysis of Fe2+ or Fe3+ released by Fe0 dissolution (Fe2++H2O→FeOH++H+), which also leads to an increase in H+ in the soil solution34,35. This may be the reason why the increase in pH was not as high as that in the CF treatment. The effect of S-nZVI on soil pH depends on the alkali materials and their application amount, as well as on the soil moisture and substrate. The variation in EC under the three water regimes is shown in Fig. 2D-F. Compared with that in CK, the EC in the SN-3 and SN-5 dose treatments increased by 60 µs cm-1 under the CF water regimes because S-nZVI could increase the concentration of available iron in the soil solution when nZVI corroded the surface of S-nZVI. Under the MW treatment, soil acidification increased the concentration of dissolved ions, leading to a gradual increase in soil EC with incubation time in both the CK and SN-1 treatments. However, the soil EC in the SN-3 and SN-5 treatment groups decreased and remained stable (Fig. 2F). These results indicated that S-nZVI could adsorb more heavy metal ions (soluble ions) and thus stabilize dissolved ions (Lehmann et al., 2011) in soil36.

Fig. 2
figure 2

Changes in pH and EC in soil under CF (A, D), AWD (B, E), MW (C, F) treatments throughout the incubation period (CK, SN-1, SN-2, SN-3 represents no S-nZVI, 1%, 3% and 5% S-nZVI dose, the same as below).

Full size image

Variation and factors in available and fraction As and Cd. Compared with that under CK, the concentration of available As significant decreased in three water treatment, with the increment of S-nZVI addition, the available As was significantly decreased compared with CK (Fig. 3A and C). For Cd, the decreasing trend was similar to that of As in different water regimes with different S-nZVI dose addition (Fig. 3D and F). Comparing the three water regimes, S-nZVI showed good performance in the co-immobilization of As-Cd in the soil and worked better under the AWD water regime than in the long-term flooded soil or moderately wet soil, indicating that S-nZVI was a suitable amendment for the water regime of natural periodic changes in the soil environment (Table S1 and Table S2). According to the above results, the variation in pH at AWD seems to remain at a suitable pH (6.2-6.8) for minimizing the available As and Cd, as previously reported in paddy soil37. Long-term flooded soil incubation or excessively dry conditions are not suitable for nZVI to form iron oxides and coprecipitate with As. Conversely, periodic redox reactions may promote the oxidation of nZVI in the micropores of sepiolite and decrease the labile As in soil due to the sorption or coprecipitation between As and amorphous iron oxides on the S-nZVI surface by strong inner sphere complexation38. Due to the abundance of Si/Al-OH groups, S-nZVI can also quickly and significantly reduce the availability of Cd through cation exchange for outer-sphere complexation. Free Cd(II) ions can bond with Si-O groups and replace aluminium in the lattice to form a stable phase39,40. In summary, S-nZVI is a robust amendment for the remobilization of As-Cd in periodically flooded soil, and principle water regulation is very important for maximizing its effectiveness. Changes in the availability and fractions of As-Cd (Fig. S2 and Fig. S3.) proved that S-nZVI amendment could significantly decrease the mobility and toxicity of As-Cd in soil under different water regimes and shift As-Cd from a labile form to a more stable fraction.

Pearson’s correlation analysis was conducted (Tables S2, S3, and S4). Soil pH was negatively correlated with available As and Cd under S-nZVI application (p < 0.01). In addition, available As was positively correlated with available Cd (p < 0.01). These results proved that S-nZVI could simultaneously immobilize soil As-Cd by affecting the soil pH. Previous work has shown that some exogenous amendments can increase soil pH and cause an increase in As availability during Cd immobilization3,9,41. However, our results demonstrated that available As was significantly decreased and bound in the F4 fraction (well-crystallized Fe and Al oxides) under S-nZVI addition, suggesting that chemical adsorption by S-nZVI is the main factor in the co-immobilization of As-Cd in soil. First, sepiolite is essential for adsorbing Cd through electrostatic interactions18. In addition, during the synthesis of S-nZVI, Fe0 oxidation produces iron hydroxides FeOOH and Fe2O325, which may even promote complexation of As-Cd with iron hydroxides (or iron oxides). The following equations may describe the surface reactions of As-Cd immobilization by S-nZVI12,25,39,42.

Fig. 3
figure 3

Effects of S-nZVI on the available As and Cd in (A, D) CF, (B, E) AWD, and (C, F) MW three different water treatment during 36 days.

Full size image

Effects of S-nZVI on the soil bacterial community. After 2 d of soil incubation, the Chao1 indices of MW and AWD decreased by an average of 9.53% and 12.26%, respectively, but the Chao1 index of CF increased by an average of 7.72%, and the Shannon index of AWD decreased by 7.02% (operational taxonomic unit, OTU) (Fig. 4), among which the greatest reductions in the Chao1 and Shannon indices occurred under AWD. In this study, increasing the S-nZVI dose and soil incubation time significantly reduced the richness and diversity of soil bacterial populations in As-Cd composite-contaminated soil, especially with 5% S-nZVI addition and CF. Nevertheless, published work has reported that a high dose of nanoscale zero-valent iron (nZVI) has potentially toxic effects on the growth of soil bacteria in heavy metal-contaminated soil45, while a small amount of zero-valent iron nanoparticles has been shown to effectively restore the bacterial community in soils contaminated with heavy metals20,46. The toxicity of nZVI has been shown to negatively impact the microbial community in the subsurface environment, and it persistently accumulates within the ecosystem47,48,49.

The cytotoxicity of nZVI can alter the bacterial community structure in soil23,24. The cluster analysis showed that the soil bacterial communities were distinct according to the soil incubation time, indicating that the soil bacterial communities were similar at the different time points (Fig. S4). The NMDS results showed that the soil bacterial communities clustered after 2 d of soil incubation (OTU level) (blue ellipses) but separated from each other after 36 d of soil incubation (green ellipses) and exhibited significant divergence from the bacterial communities of the original soil (red ellipses), as well as those observed after 2 d of soil incubation (Fig. 5A-C). The same experimental results were also confirmed via PCA and PCoA (Fig. S5). The bacterial communities of soil supplemented with 1%, 3%, and 5% S-nZVI significantly differed with prolonged soil incubation time under MW, CF, and AWD. The autochthonous bacterial communities varied with the soil ecosystem following nZVI exposure49. The negative surface charge of S-nZVI is insufficient to electrostatically repel negatively charged bacteria and reduce contact with nZVI, affecting the uptake of nutrients by microbial cells50,51. The ecotoxicity of S-nZVI was the main reason for the differences in the soil communities, but increased soil incubation time promoted soil bacterial growth. Furthermore, the variation in available Fe also supported this point. After the addition of S-nZVI, available Fe presented a temporary increase at 2 d, followed by a continuous decreasing trend after 2 d (Fig. S6). This trend was even more significant in the 5% SN treatment. Hence, a temporary increase in ecotoxicity is mainly due to the short-term dissolution of nZVI at a high S-nZVI dose. Nevertheless, the risk of Fe toxicity could also be ignored because of the secondary oxidation (or precipitation) of Fe[2+39,46 in the soil after long-term incubation.

Fig. 4
figure 4

S-nZVI and soil incubation time’s effects on richness and diversity of soil bacteria. Note: A and D: Moderate wet (70% water-holding capacity, MW), B and E: Continuously Flooded (CF), C and F: Alternately Wet-Dry (AWD) (The same below). CK1: Primitive soil; *2CK and *36CK (*represented A, B, or C respectively, the same below): The soil without SEP-nZVI addition for 2 d and 36 d; *2SN1, *2SN3, and *2SN5: The soil supplemental with 1%, 3%, and 5% of SEP-nZVI for 2 d; *36SN1, *36SN3, and *36SN5: The soil supplemental with 1%, 3%, and 5% of SEP-nZVI for 36 d.

Full size image

The strong resistance of soil bacterial communities to heavy metal stress is related to the addition of modified materials39,52. After 2 d and 36 d of soil cultivation, the results relative abundances of dominant bacteria shown in Fig. S7 B-C. With the addition of 1%, 3%, and 5% S-nZVI and after 36 d of soil cultivation, the relative abundances of Gemmatimonas and Flavisolibactor decreased by 10.20-21.73% and 0.58-9.20% in the MW, CF, and AWD treatments, respectively, while the relative abundances of Saccharimonadales and Massilia increased by 17.10-29.34% and 0.32-6.51% in the MW and AWD treatments, respectively. The abundance of Bacillus significantly increased in the MW treatment but decreased in the AWD treatment. The stressed microbial populations and quality in contaminated soil are reduced by nZVI treatment53. nZVI can promote or affect the presence and abundance of soil bacterial strains, and lower nZVI concentrations alter bacterial viability54. S-nZVI has selection effects on promoting or inhibiting the dominant soil bacteria with the highest abundance in As-Cd co-contaminated soil55, such as Gemmatimonas and Saccharimonadales. S-nZVI and soil water content can dramatically change the community structure and relative abundance of the original dominant bacteria under different water management practices. These results indicated that decreasing S-nZVI addition and increasing soil water content can reduce the ecotoxicity of S-nZVI to restore soil bacterial growth, significantly regulating the dominant bacterial abundance of As-Cd composite-contaminated soil in the long term under MW, CF, and AWD conditions.

S-nZVI addition and soil water content effectively regulated the viability of the dominant bacteria and their interactions with soil As-Cd. The fundamental physicochemical properties of soil determine the availability of the soil microbial community39. Soil pH is a crucial factor affecting microbial diversity56,57 and is positively correlated with soil bacteria58. Iron ions are an essential biological factor, but a high nZVI content is toxic under iron ion addition59. The relative abundances of identifiable dominant soil bacteria (genus level, relative abundance of TOP 20, total abundance accounting for 52.01%) exhibited a significant interaction with pH, Av-As (available As), and Av-Cd (available Cd) in As-Cd composite-contaminated soil supplemented with 1-5% S-nZVI. Compared to MW, CF and AWD increased the soil water content, thereby facilitating the growth and diffusion of soil bacteria in As-Cd co-contaminated soil (Fig. 6A-C). In addition, the dominant soil bacteria were significantly correlated with Eh, Am-Fe (Amorphous Fe), and Av-Fe (available Fe), but there was no significant correlation between the dominant soil bacteria and soil EC in the MW, CF, and AWD treatments. The effects of nZVI on soil microbial structure and function depend on the dose and species of bacteria involved22,60. Soil bacteria that are well adapted to extreme environments are closely related to the transport functions of various heavy metals61. Soil bacteria play a vital role in transforming nZVI oxidation products, promoting Cd immobilization and increasing soil pH62. This increase in soil water content contributed to the increase in the correlation between the dominant soil bacteria and pH, Av-As, and Av-Cd. The addition of S-nZVI, along with increasing the soil water content, played a crucial role in improving the resistance of the original dominant bacteria in As-Cd composite-contaminated soil to As-Cd.

Fig. 5
figure 5

The difference analysis of soil bacterial population under different water management (A: MW, B: CF, and C: AWD).

Full size image
Fig. 6
figure 6

The correlation analysis of soil dominant bacteria (TOP 20)with soil factors, (A) MW, (B) CF, (C) AWD (note: Av-As: Available As, Av-Cd: Available Cd, Am-Fe: Amorphous Fe, and Av-Fe: Available Fe).

Full size image

Conclusion

Adsorption or surface complexation and precipitate formation on the iron oxyhydroxide (iron oxide) layer of S-nZVI were the possible mechanisms of As-Cd immobilization. The addition of S-nZVI increased the soil pH and EC and promoted the transformation of the acid extraction fraction, oxidation fraction, and reduced fraction of As-Cd into a more stable residue fraction, significantly reducing the available As-Cd under the different soil‒water management conditions of CF, AWD, and MW. However, 16 S rRNA sequencing revealed that S-nZVI can significantly decrease soil bacterial diversity and richness and alter the abundance of dominant bacteria in acidic soil (pH 6.03). S-nZVI enhanced the interaction of high-abundance soil bacteria with pH, Eh, Av-As and Av-Cd during a 36 d soil incubation experiment. These results revealed that controlling the S-nZVI dose at a proper level and increasing the soil water content and incubation time can alleviate S-nZVI ecotoxicity to high-abundance soil bacteria, simultaneously enhancing the S-nZVI remediation capacity and functional bacterial growth vitality. Hence, a medium S-nZVI dose (3% w/w) applied in soil with a normal AWD water regime is can be a good choice for As-Cd co-contamination soil in red soil of Hunan Province.