The I _geo is a quantitative indicator introduced by Müller in 1969 (Muller, 1969), it has been widely used to assess the extent of sediment and soil pollution. The calculation equation of I _geo is as follows (Eq. 1): I g e o = log ⁡   2 ( C n / 1.5 B n ) (1)

C _n is the measured concentration of each assessed heavy metal element, B _n is the geochemical background value of this element found in the soil ( Supplementary Table S4), and factor 1.5 is the natural fluctuation of heavy metal content during diagenesis. I _geo was divided into seven classes (Class0-Class6) (Supplementary Table S2; the highest (class 6) reflects a richness over 100 times higher than the background values (Förstner et al., 1990).

RAC assesses the availability of heavy metals in soil and sediments by calculating the ratio of exchangeable and carbonate factions to the total metals (Xiao et al., 2014). The calculation equation of RAC is as follows (Eq. 2): R A C = F 1 + F 2 T M × 100 (2) Where F1 and F2 represent the concentration of Fraction 1(water soluble and ion exchangeable metals) and Fraction 2 (bound to carbonates metals) of each metal, respectively, TM is the total concentration obtained from the metal (Liu et al., 2008). The grading standard of RAC is shown in Supplementary Table S3.

The concentrations of HMs in mining waste were shown in Figure 2 and listed in Supplementary Table S5. The pH of flotation tailing (Z1) was 3.61, the concentrations of As, Cd, Zn, Pb, Cu, Cr and Ni in Z1 were 943.39, 0.99, 58.25, 39.80, 11.07, 38.25, and 3.52 mg kg⁻¹. The floatation tailing was acidic because flotation was carried out under acidic conditions. The pH value of heap leaching tailings (Z2-Z6) was 2.73–8.58, and the mean concentrations of As, Cd, Zn, Pb, Cu, Cr, and Ni were 1171.03, 2.12, 310.32, 299.38, 306.56, 85.22, and 12.8 mg kg⁻¹, respectively. During oxidation heap leaching of low-grade raw ore, lime was used as a heap leaching agent to adjust the pH during heap building, which was the reason for the alkaline pH of Z3, Z4, and Z5. The concentrations of As, Cd, Zn, Pb, Cu, Cr, and Ni in sewage sludge were 7527.74, 5.43, 290.31, 159.41, 162.01, 43.80, and 9.18 mg kg⁻¹, respectively, and the pH was 2.86, which is strongly acidic.

The maximum concentrations of Cd, Zn, Pb, Cu, Cr, and Ni in tailings and sewage sludge appear in heap leaching tailing (Z2) and the maximum concentration of As appeared was sewage sludge (Z7), which was significantly higher than those in all soil samples. The characteristics of the high range of heavy metals in tailings sand and the easy migration of fine particles make it the leading cause of heavy metal pollution in the surrounding environment of the mining area (Cui et al., 2014).

The concentrations of HMs in soil were shown in Figure 2 and listed in Supplementary Table S6. The highest range of all heavy metals was found in sample S2 inside the gold mine, likely because it was located near the sewage sedimentation basin. The concentrations of As was significantly high in S2 (1907.75 mg kg⁻¹), 13.6 times higher than the risk-controlling values of the Risk Control Standard for Soil Contamination of Development Land (GB36600-2018) (Ministry of Ecology and Environment and State Administration for Market Regulation, 2018b). The concentration of As in sample S1 (268.47 mg kg⁻¹), located 50 m east of the sewage sedimentation basin, far exceeded the risk-controlling values of GB36600-2018 (Ministry of Ecology and Environment and State Administration for Market Regulation, 2018a) as well. This indicates that mining activities had significantly increased arsenic concentrations in the mining area (Liu et al., 2013).

According to the multi-purpose regional geochemical survey in Anhui Province (Jia, 2007), the farmland soils in the study area did not belong to mineralized areas (Table 1). The concentrations of As, Pb, Cd, Cu, and Zn in sample S3 exceeded their corresponding background values (Chun-Hui et al., 2018), indicating that the accumulation of heavy metals in soil has reached a severe level (Wei et al., 2006; Cheng et al., 2009; Xiao et al., 2017). The concentration of As in sample S3 was extremely high (322.56 mg kg⁻¹), which exceeded the risk-controlling values of the Risk Control Standard for Soil Contamination of Agricultural Land (GB 15618-2018), and Cd exceeded the risk-screening values of GB 15618-2018. Gold usually occurs in fine-grained natural gold and lattice gold in the significant As-bearing minerals pyrite and arsenopyrite (Gieré et al., 2003), so the ore-bearing particles from the gold mine is a major As Source. This may be why the arsenic content in farmland adjacent to gold mines is exceptionally high. The pH values of S3 and S4 in farmland soil were 4.18 and 4.60 (acidic), respectively, which may also be caused by weathering and oxidation of pyrite (FeS₂). As a result, in gold mining areas rich in As-bearing minerals, significant accumulation of arsenic in the soil (and consequently the food chain) may occur as mining activities proceed (Zhu et al., 2008; Cheng et al., 2009; Islam et al., 2013).

The concentrations of As, Zn, Pb, and Cu at point S3 near the gold mine are higher than those at point S4, indicating that the heavy metal content in soil may be related to the location. Generally, the closer farmland soil is to the tailings storage, the more severe the contamination is. (Liu et al., 2013). Through the above analysis, the arsenic pollution in the soil near the gold mine has been severe enough to be unsuitable for agricultural production, and the cadmium pollution is also worthy of attention.

Generally, elements that exhibit a high degree of correlation may have the same origin and, under certain physiochemical circumstances, exhibit similar behaviors during transformation and migration (Zhang et al., 2019). The Spearman’s correlation coefficients for the heavy metal contents of tailings, sludge and soil are shown in Figure 3; Supplementary Table S7. A significantly positive correlation (p < 0.01) was found between Cd-As (0.927) and Cr-Pb (0.764). Meanwhile, five elemental pairs, Cd-Cu (0.709), Cd-Zn (0.655), Pb-Cu (0.691), As-Cu (0.673), Zn-Ni (0.682), had a low positive correlation (p < 0.05). Moreover, the content of heavy metals in soil is closely related to the content of tailing and sludge. As the gold ore grade is generally not high and there are many co- and associated elements, the amount of tailings produced by mining activities is quite large, and the polluting heavy metals are released into the environment through the tailings (Liao, 2005).

The research showed that the content of As and Cd in tailings and sewage sludge is extremely high, and the enrichment of As and Cd in the soil is profound. Tailings contain solid particles, especially the tailings produced in the beneficiation process; the particle size is mostly 0.17–0.07 mm. Due to their small particle size and low gravity, tailings can easily migrate under the action of hydraulic and wind power, causing heavy metal contamination to the surrounding farmland soil (Santibañez et al., 2012). Zhou et al. obtained a conclusion consistent with the research results of this paper through theoretical simulation calculation (Zhou, 2009).

The XRF semi-quantitative results of each element in the samples showed that the samples contained similar elements, mainly Si, Al, Fe, and K (Supplementary Table S8). The contents of S, Na, Mg, and Ca are not high, and As, Cr, Ni, Cu, Pb, and Zn are extremely low. The highest concentrations of As are present in Z2, Z7, and S2, which is consistent with the highest iron and sulfur content in all samples (Supplementary Tables S5, S6). The high content of Fe and S in the samples may be related to Fe-O, sulfide minerals, and their weathering products. Thus, the observed correlation between As and Fe/S concentrations indicates that the Fe-O and sulfide minerals may play an essential role in regulating arsenic retention in soil (Kim et al., 2014). The highest cadmium content in soil samples appeared in S2, coinciding with the highest pH level and CaO, MnO, and Al₂O₃ analyzed by XRF in the soil samples. Soil pH directly affects the current form and adsorption amount of Cd in soil. The amount of H⁺ is inversely proportional to the pH value and the negative charge attached to the soil surface. The increase of negative charge on the soil’s surface benefits the adsorption of Cd²⁺on the earth’s surface, resulting in increased Cd content in the soil. Meanwhile, the CaO content has a synergistic effect with soil pH; the pH value increases with the increase of CaO content in the soil, increasing the Cd adsorption capacity (Tahervand and Jalali, 2016). Ran et al. confirmed that soils with rich clay content and high pH are usually conducive to the stability of heavy metals (Xiao et al., 2017). There is a positive correlation between the range of Al₂O₃ and Cd in soil because the higher the content of Al in soil, the more clay minerals, indicating that the clay minerals in the soil also affect the enrichment of Cd (Yang, 2021). Soil Mn, mainly in oxides and hydroxides, can adsorb the element Cd and affect its morphology, rendering it inactive (Baize et al., 2009). Therefore, soil Al₂O₃ and MnO are positively correlated with Cd, too.

The mineralogical composition of tailings, sewage sludge, and soils were characterized using XRD. According to the XRD analysis patterns (Figure 5), the dominant minerals found in the soil samples are quartz (SiO₂) and muscovite (Kal₂(AlSi₃0₁₀) (OH)₂). In addition, epidote (Ca₂Fe³⁺Al₂ [SiO₄][Si₂O₇]O(OH)) was observed in the S1, epidote, microcline (K [AlSi₃O₈]), and kaolinite (Al₂Si₂O₅(OH)₄) were observed in the S3, microcline was observed in the S4. The tailings and sewage sludge samples have a similar mineral composition to the soil samples. In addition, calcite and kaolinite were observed in the Z2 and epidote in the Z5. Unlike the soil samples, microcline was not observed in tailings and sewage sludge samples.

The inorganic mineral component quartz is the most abundant in the samples, namely silica (SiO₂), the most stable silicate, belongs to the oxide mineral of the trigonal system, and its adsorption capacity for metal ions is weak (Jia and Sun, 2001). Muscovite and kaolinite belong to clay minerals with layered aluminosilicate structures, with a negative charge on the surface, big specific surface area, and strong adsorption characteristics to the metal cations (Cheng and Reinhard, 2006; Sun et al., 2014). Microcline is one of the feldspar minerals, the most widely distributed minerals in the crust, belonging to the tectosilicate. Its prominent pore structure mineral characteristics are conducive to the adsorption of heavy metals (Li, 2007). Calcite is the most common type of natural carbonate mineral in the earth’s crust; it is often used as an adsorbent for environmental pollution remediation because of its large specific surface area and good adsorption capacity (Zhu, 2016). Epidote is an island silicate mineral, and the surface is mainly exposed to iron, aluminium, and other metal ions. In an aqueous solution, they combine with hydroxide and have a good adsorption capacity for metal ions (Jia and Sun, 2001).

We observed the micro-morphology of the samples by scanning electron microscope (SEM), Supplementary Figure S1 shows that the microscopic morphology of the tailings and sewage sludge surfaces were corroded, and many irregular and, uneven intermixtures of big and small particles we found attached to the surface, while the soil samples were irregular, with heterogeneous soil particles, significant agglomeration and a coarse surface with considerable weathering and erosion. After grinding, flotation, heap leaching and other processes, the surface of flotation tailings and heap leaching tailings are destroyed, and small particles are accumulated and adsorbed on large particles in the form of layered and spherical inclusions. At high multiples, the surface of particles is observed to be uneven and eroded. The irregular surface morphology indicates that the morphology and structure of these contaminated tailings and soils are complex (Lenthe et al., 2018).

The SEM-EDX element map of particles (Figure 6), shows that the areas with high arsenic content high overlap with Si, O, and Fe in samples. Inclusions of As-bearing minerals are irregularly shaped, probably a secondary As-bearing minerals—scorodite (FeAsO₄ 2H₂O), a common weathering product of arsenic-bearing ore deposits which were often observed in mine waste sites and arsenic-contaminated soil (Dove and Rimstidt, 1985; Smedley and Kinniburgh, 2002). There are more than 300 types of arsenic-containing minerals are known, such as sulfides, oxides, arsenates, and arsenides, and arsenic-containing minerals in nature usually exist in the form of sulfides, including arsenopyrite (FeAsS), realgar (As₄S₄), and orpiment (As₂S₂) (Adriano, 2001; Jia and Demopoulos, 2005). Au is usually associated with arsenopyrites (FeAsS) and sulfides and a large amount of As-rich mine wastes were produced with the mining and beneficiation of gold ore. The mobilization and migration of As from these wastes into the environment can result from oxidation of arsenopyrites and sulfides. (Leblanc et al., 1996; Dudka and Adriano, 1997; Langmuir et al., 1999; Majzlan et al., 2014). Arsenopyrite (FeAsS) is transformed by weathering into As-rich pyrite (FeS₂), which subsequently forms realgar (As₄S₄) and orpiment (As₂S₂), eventually creating an oxidized phase like scorodite (Nordstrom and Parks, 1987; Smedley and Kinniburgh, 2002; Corkhill and Vaughan, 2009a). Kim and Batchelor also observed that As and pyrite (FeS₂) interact in the anoxic environment to form realgar (As₄S₄) and orpiment (As₂S₂) (Kim and Batchelor, 2009b). The oxidation of FeAsS occurs through several oxidation state changes of As and S by biological/non-biological reactions (Corkhill and Vaughan, 2009b). The chemical reactions are simplified as follows: 3 F e A s S + 3 H 2 O → A s 2 S 3 + H 3 A s O 3 + 3   F e 2 + + 3   H + + 9 e ― F e A s S + 10 H 2 O → F e A s O 4 · 2 H 2 O + S O 2 − + 16 H + + 14 e ―

We have found As-bearing minerals in the mine waste and soil samples based on the SEM-EDS analysis result. More importantly, the arsenic-mineral phases observed in the soil are matched with the weathering and oxidation products of tailings, which seemed to be Fe-(oxy)hydroxides absorbed by co-precipitated As, indicating that arsenic-containing particles in soils may partly from the mining wastes through atmospheric transport, rainwater leaching, overland runoff, and other means (Lee et al., 2020).

We analyzed all of the samples by X-ray photoelectron spectroscopy (XPS), which provides details on the oxidation states of As and Cd and their chemical forms on the surfaces of minerals (Kim and Batchelor, 2009a). The XPS narrow-scan spectra of As (3d) and Cd (3d) are illustrated in Figure 7; Supplementary Figure S2, and their corresponding peak parameters are summarized in Supplementary Tables S9, S10. Except for Z3, Z4, S3, and S4, the deconvolution of the As (3d3/2, 3d5/2) spectrum suggests two types of As-containing phases in the samples: 1) As (III)-O (43.8–45.17 eV), which was arsenite, and 2) As(V)-O (45.49–46.62 eV), which was arsenate. Meanwhile, the major peak of the As 3d spectrum of the farmland soil (S3, S4) and heap leaching tailings (Z3, Z4), which were located at 45.49–46.36 eV, was attributed to As(V)-O and did not present any reduced forms of As, for instance As(III)-O or As sulfide.

In conjunction with SEM observations, most As-bearing phases are embedded in Fe, O, and Si compounds, and we speculated that it might be secondary minerals oxidized from As-bearing minerals—scorodite. The deconvoluted peaks of the As spectrum, which revealed scorodite’s Fe-As(V)-O binding energies (45.49–46.62 eV), further confirmed the presence of scorodite (Tabelin et al., 2020).

Arsenic is a variable element that typically has four oxidations, including As^−III (arsine), As⁰ (metallic arsenic), As^III (As₂O₃, arsenite), and As^V (As₂O₅, arsenate). The different oxidation states of arsenic determine its toxicity and crystal-chemical reactivity (Sharma and Sohn, 2009; Zhao et al., 2010; Gupta, 2011). Among them, As^V is the most stable form because it can be firmly adsorbed on the surfaces of Fe, Al, and Mn oxides, so it is less mobile and toxic than As^III in the environment (Li et al., 2017). The analysis in 3.4.2 shows that the arsenic in the ore is mainly in the form of metallic arsenide, while the arsenic in the tailings is mostly in the pentavalent oxidation state, indicating that during the mining, flotation and heap leaching process, most of the arsenic in the ore is released and oxidized, arsenic and Fe, O, Si compounds were connected and entered the tailings. At the same time, due to the open-pit accumulation of the tailings pond, no vegetation cover, and no wind and shower protection measures, the mineral particles in the tailings enter the soil near the mine through atmospheric transport, rainwater leaching and overland runoff, resulting in arsenic enrichment (Liao, 2005).

In most soil and mine wastes, As is absorbed to or co-precipitated with Fe (hydr) oxides, clay minerals, and organic matter under standard oxidizing conditions (Welch and Stollenwerk, 2003; Wang and Mulligan, 2006), and may be directly substituted into the crystal structure of a limited number of secondary minerals (Savage et al., 2005) and precipitated as secondary As minerals (Majzlan et al., 2014). Arsenic in soil usually coexists in oxidized and reduced states, and arsenite in acidic and oxidized soil has been found (Masscheleyn et al., 1991), which explains the presence of trivalent arsenic in some samples.

As shown in Supplementary Figure S2, two strong absorption peaks appear around 405.6 ev and 412.5 ev, which are attributed to Cd 3d5/2 and Cd 3d3/2 respectively. Both peaks are asymmetric and after fitting, the binding energies of the Cd 3d5/2 and Cd 3d3/2 peaks are located at 405.37–405.77 ev and 411.85–412.66 ev, respectively. The XPS standard spectrum of Cd combined with the relevant literature shows that Cd is present in the form of CdCO₃ (Zhang et al., 2015).

As, Cd, Zn, and Pb residual states are also more than 60%. This indicates that the heavy metal elements in the soil of the study area are mainly contained in the mineral lattice in a residual state, which is very stable, and may be caused by the entry of mining wastewater and mineral dust into the soil.

Results in Supplementary Table S11; Figure 8A indicate that soil was practically uncontaminated by Cr and Ni (I_geo ≤0). For Zn, the I_geo value was <0 (which means practically uncontaminated) except for S2 (0<I_geo<1, which means uncontaminated to moderately contaminated). The I_geo values ranged from −0.09 to 2.03 for Cu. The most heavily polluted sites appear in S2 (2.03, which means moderately to heavily contaminated, the lightest polluted sites appear in S4 (−0.09, which means practically uncontaminated), and S1 and S3 fell in the range of 0–1 (uncontaminated to moderately contaminated). The I_geo levels of Pb at each site were essentially the same as of Cu; the most heavily polluted sites appear in S2 (1.48, which means Moderately contaminated), too.

The I_geo for As was the highest, followed by Cd. The As I_geo values ranging from 1.44–6.70 (moderately contaminated to extremely contaminated), the highest grade of the geo-accumulation index in site S2, reflected a 100-fold enrichment above background values, indicating that As pollution was severe in soil (Ho et al., 2010). For Cd, farmland soil (S3 and S4) was found to have similar levels of contamination (less than 1, which means practically uncontaminated). In contrast, the soil in the mine areas fell in moderately contaminated (S1) and heavily contaminated (S2), respectively.

In farmland soil, the As pollution was severe, but the pollution degree of As decreased significantly with the increased distance from the mining area. Therefore, it can be speculated that introducing As-containing mineral particles into farmland soil by mining wastewater discharge, atmospheric dust fall, and tailings leaching may cause arsenic accumulation in farmland soil near the mining area, which confirms the speculation of 3.1.1.

We calculated the RACs of tailings and sludge to assess the environmental risk of hazardous metals, which is critical for safety control and reutilization of tailings (Supplementary Table S12; Figure 8). According to RACs, Cr poses no risk (RAC <1%) to the environment, while As, Pb and Ni pose no risk (RAC <1%) and low risk (1% ≤ RAC < 10%). Cd and Zn pose a medium risk in Z2 (10%≤RAC <30%), while there is no risk (RAC <1%) to low risk in other tailings (1% ≤ RAC <10%). Cu poses a medium risk in Z1 and Z5 (10% ≤ RAC <30%) and low risk in other tailings (1% ≤RAC <10%). Arsenic and Cr, Pb, and Ni pose low or no risk in tailings, while Cd, Zn, and Cu pose a medium risk, which may limit the utilization of tailings. So, before the transportation, disposal, and utilization of tailings, a safety assessment should be carried out carefully, and corresponding measures should be taken to reduce their harmfulness.

Cr in the soil poses no risk (RAC <1%), while As, Zn, Pb, Cu, and Ni pose no risk (RAC <1%) and low risk (1% ≤RAC <10%) to the environment. Cd poses a relatively high risk, with a medium risk (10% ≤RAC <30%) at S2 and high risk (30% ≤ RAC < 50%) at S4, whereas a low risk at S1 and S3 (1% ≤ RAC < 10%). The As content in the soil was extremely high, but the RACs were low. Nevertheless, its environmental impact could be significant due to its high arsenic concentration and mobility (Kim et al., 2014).