Determination of Arsenic Species Distribution in Arsenide Tailings and Leakage Using Geochemical and Geophysical Methods

Abstract

This study describes the distribution of arsenic mobile species in the tailings of Cu–Co–Ni–arsenide using the sequential extraction and determining the contents of arsenate (As^V) and arsenite (As^III). The object of this study is the tailings ponds of the Tuvakobalt plant, which contains waste from the hydrometallurgical arsenide ore processing of the Khovu-Aksy deposit (Republic of Tuva, Russia). A procedure of sequential extraction for arsenic was applied, and it includes the extraction of the following forms: water-soluble, potentially water-soluble and exchangeable, easily sorbed on the surface of carbonates, associated with Fe/Mn oxides/hydroxides, associated with easily oxidized minerals, and accounted for by non-oxidized arsenic minerals. This procedure, which takes into account the peculiarities of the physical and chemical composition of the waste, was supplemented by the analytical determination of the arsenite and arsenate content by using the methods of inductively coupled plasma atomic emission spectrometry (ICP-AES) combined with the hydride generation technique (HG-ICP-AES). The content of the most mobile forms of arsenic, which are water-soluble, potentially water-soluble, and exchangeable species, is equal to 56% of the total arsenic content, 23% and 33% of which are arsenite and arsenate, respectively. Unlike arsenic, the mobile forms of metals have been determined in small quantities. The largest proportion of water-soluble and exchangeable forms is formed by Mg, Ca, and Sr at 11, 9.4, and 20%, respectively (residual and redeposited carbonates). The proportion of water-soluble forms of other metals (Cu, Zn, Co, and Ni) is < 1% or 0. The main part of the metals is adsorbed on the surface of Fe and Mn hydroxides, enclosed in easily and hardly oxidized minerals. In addition to geochemical studies, the presence of leaks from the tailing ponds into ground waters was determined by using electrical resistivity tomography. The data obtained indicate a high environmental hazard of tailings and the possibility of water-soluble and highly toxic arsenic compounds entering ground waters and aquifers.

1. Introduction

The importance of conducting a comprehensive monitoring of the state of a tailings dam during long-term storage of waste is due to the purpose of environmental assessment of the state of the surrounding area and the study of the state of engineering structures, as well as the possibility of further extraction of useful components.

Arsenic is a ubiquitous metalloid found in various species in groundwater, soil, and food, and has traditionally been considered one of the most common potentially toxic elements [1,2,3]. The toxicity of arsenic depends on its specific chemical form and oxidation state. The most toxic species of arsenic in the environment are inorganic forms: arsenite (As^III) and arsenate (As^V), which are prevalent in soils, deposits, and waters of various natures. Less toxic species are represented by organic forms: monomethylarsonate and dimethylarsonate, which are present in trace quantities in seafood. Non-toxic species are arsenocholine, arsenobetaine, and arsenosuclear, the main metabolites of arsenic in marine animals [4]. Under environmental conditions, arsenic is characterized by different ranges of arsenic content in groundwater as a result of water–rock interactions and the gradual entry of arsenic into aquifers and soil. Anthropogenic sources also contribute substantially to pollution, mainly attributable to waste and atmospheric emissions from the mining and processing industries and the energy sector; these emissions result in the accumulation of arsenic-containing waste and pollution of soils and sediments near pollution sources [5,6]. Furthermore, at the same time, the environmental risks of mine tailings are of particular concern [7,8,9,10,11]. A lot of research is devoted to finding effective methods for purifying natural waters from arsenic. These methods include the adsorption technique [12,13]; use of a nanofiltration membrane [14]; biological oxidation followed by As^V removal with Fe and Mn dioxide [15,16]; and with the help of modelling various adsorption processes [17,18,19,20,21].

Element speciation is fundamental to understanding the toxicity and mobility of chemical elements and is an important addition to the knowledge of the total content of elements, including arsenic. In geochemistry, physicochemical modeling, which is widespread, is used to calculate the content of various forms based on the known chemical composition and physicochemical characteristics of the system under consideration. However, the applicability of this approach critically depends on the availability of reliable thermodynamic data and timely updating of databases. In the case of arsenic, the thermodynamic properties of metal arsenates are well characterized, but metal arsenites are poorly studied, and thermodynamic data for arsenides and arsenic sulfides are practically not studied [22,23].

Alternative approaches are the determination of arsenic forms using a sequential extraction approach, which allows the extraction of a set of chemical forms with similar physicochemical characteristics, or the direct use of complex analytical methods that make it possible to directly determine specific chemical forms of arsenic [24]. Most of the work on the determination of arsenic species using the step leaching approach was focused primarily on contaminated soils and deposits and mining waste [25,26,27,28,29,30]. However, at the same time, there is little work that is directly devoted to the use of step leaching for As-rich waste, which has high arsenic contents and characteristic As-containing minerals [31,32]. The key feature is the choice of leaching conditions and reagents, taking into account such parameters as geochemical parameters (pH, Eh, total composition, the ratio of Fe, As, and S in the tailings), the age of the tailings, and mineral transformation (oxidation, dissolution, and redeposition) under storage conditions.

In addition, there are chemical analysis methods for determining the forms of arsenic in solutions, including high-performance liquid chromatography and capillary electrophoresis in combination with inductively coupled plasma-mass spectrometry (HPLC-ICP-MC/CE-ICP-MS) [24]. Techniques have also been described for the non-chromatographic determination of the arsenic forms arsenite (As^III) and arsenate (As^V) using the methods of atomic emission spectrometry with hydride generation technique (HG-ICP-AES) [33,34].

One of the objects on the territory of Eastern Siberia of particular interest is the tailings ponds of the Tuvakobalt plant, which is located in the village Khovu-Aksy (Chedi-Kholsky Kozhuun, Republic of Tuva, Russia). These tailings are recognized as the most environmentally hazardous facilities in the Republic of Tuva and pose a threat of interregional environmental pollution. The tailings have also been repeatedly mentioned in reports on the state of the environment and the media. According to previous studies, more than 2 million m³ of waste from the Cu–Ni–Co hydrometallurgical ore processing containing underexploited arsenides was stored in tailings ponds, which causes the high arsenic content of up to 2.5% in the tailings [35]. One of the key tasks for comprehensive monitoring of tailings status is selection of optimal and correct methods. In particular, it is necessary to determine the arsenic species and evaluate the potential mobility of arsenic.

The purpose of this work was: (i) to develop a procedure for sequential extraction of arsenic in relation to As-rich tailings of Cu–Co–Ni arsenide ores; (ii) to determine the ratio of As^V and As^III compounds in the upper horizons of the tailings; (iii) to estimate of metals (Cu, Zn, Co, and Ni) mobility; and (iv) to trace the migration paths of the surface waters interacting with the tailings into underground horizons using electrical resistivity tomography.

2. Study Object

The Khovu-Aksy deposit of cobalt–nickel arsenide ores is located in the village of Khovu-Aksy of the Republic of Tyva (Russia) (Figure 1). The deposit had been developed by the Tuvakobalt plant since 1970 until it was mothballed in 1991 [36].

Primary ore mineralization led to the formation of carbonate mineral veins that cut the Silurian skarns [37]. The principal ore minerals within the vein system correspond to two paragenetic stages: skutterudite–safflorite and skutterudite–niccolite–rammelsbergite. The minor ore minerals include galena, sphalerite, argentite, native silver, native arsenic, and native bismuth. Small quantities of tennantite, chalcopyrite, bornite, and pyrite are formed in a subsequent stage (the sulfide–tetrahedrite stage).

Ore processing technology was developed by the Gipronickel Institute (St. Petersburg, Russia). Hydrometallurgical ore processing was carried out by autoclave ammonia-carbonate leaching without preliminary enrichment and subsequent purification of technological solutions from arsenic. After the ore was crushed and ground, autoclave leaching was carried out with a solution of carbonate and ammonium hydrate (100–110 g/L) with the introduction of 0.5 m³ of air per 1 kg of ore and heating to 90 °C and a pressure of 14–15 atm. for 2–3 h.

At the second stage, technological solutions were purified from soluble arsenic by precipitation of sparingly soluble compounds of arsenic oxyanions and magnesium/ammonium (Mg(NH₄)AsO₄ ∗ H₂O, Mg₃(AsO₄)₂ ∗ nH₂O) with the addition of caustic magnesite (MgO) or sorption of arsenic oxyanions on the surface of magnesite. After this procedure, according to the standards, the residual arsenic content in the solution should not exceed 0.5 g/L. Further, excess carbon dioxide was passed through the ammonia solution of Co, Ni, and Cu to precipitate them as carbonates. The resulting precipitate was separated from the solution and sent for further processing [35].

Ore dressing wastes in the first years of operation of the plant were transported to a remote isolated site and buried in trenches. Subsequently, five pond reservoirs (tailings ponds) were equipped and were the locations to which waste and arsenic-containing waste were pumped through pipelines. The tailings ponds are rectangular reservoirs with an area of ~250∙100 m²; they are diked with excavated soils and equipped with an anti-filtration screen made of polyethylene film laid on a sandy underlying layer. After the slurry arrived and the solids settled, the waste was dehydrated, and the water returned to the mill for further use. The residual water evaporated naturally. Temporary water bodies with snow and rain supplies also appeared on the surface of the ponds. The tailings in pond 1 were covered by a soil layer of 10–20 cm and initially were seeded with scattered perennial grasses. Currently, vegetation covers the surface of the former pond site and grows well. Pond 2 was reclaimed after pond 1, and vegetation only partially covers its surface. The exposed surfaces of the ponds were a source of soil pollution in the surrounding area used for grazing [35,38].

Previous studies demonstrated the transformation of waste over 20 years of waste storage, the data obtained indicate the occurrence of active processes of metal (Cu, Co, and Ni) precipitation and their transition to poorly soluble forms, whereas As remained in highly movable forms [39].

3. Materials and Methods

3.1. Field Sampling

During the field work, solid tailings samples weighing ~1 kg of each were collected from small pits (50 cm) in tailings ponds. On the surface of each pond, small pits distributed evenly over the entire area were dug, such that the collected samples characterize the substance of the entire pond. It was important to characterize the surface horizons of the ponds because the upper layers are the most susceptible to oxidation and destruction of the substance. Therefore, in ponds 1 and 2, waste tailings mixed with the soil layer entered the samples. Then, the samples were pooled to obtain representative material for research. The sample weighing ~10 kg (KhA_tot) was packed on site in a sealed plastic bag.

Water samples were taken from small surface reservoirs formed during the influx of seasonal precipitation (snow and rain). Water was placed into plastic bottles with a volume of 0.5 L after the bottles had been pre-rinsed three times with water at the sampling site. Water samples were filtered through a 0.45 μm filter and conserved with an ultra-pure HNO₃. The values of pH and Eh were measured in situ using the Expert 001 hY/ion-meter (JSC “Ekoniks-Expert”, Moscow, Russia).

3.2. Geophysical Investigation

Electrical resistivity tomography was used to detect underground leaks and the entry of solutions into drinking water horizons. Electrotomography profiles were constructed along the sides of ponds 1, 2, and 3 (600 m) and ponds 4 and 5 (360 m). The measurements were carried out using the Skala 48 (SibER 48, LLC “KB Electrometry”, Novosibirsk, Russia), a piece of multielectrode electrical survey equipment. The standard electrode penetration is ~10 cm with a spacing of 5 m. A prerequisite for each measurement is to check the quality of grounding; the grounding resistance did not exceed 1 kΩ for any of the profiles. A standard Schlumberger array, which sequentially alternates 48 AMNB electrodes for 529 measurements, was used.

3.3. Laboratory Procedures and Analyses

3.3.1. Preparation of Solid

After thorough mixing and quartering, the sample was divided into several parts. The first part was intended to obtain a water extract. The second set was dried at room temperature for 48 h, homogenized, and powdered by abrasion in an agate mortar for bulk analysis. The third set was divided into six sub-samples for a sequential extraction procedure.

3.3.2. Water Extracts

The water extracts were prepared by stirring 10 g of solid in 100 mL of distilled water for 24 h at room temperature. We used the standard procedure for obtaining aqueous extract from the waste material, according to the EPA Method 1213 [40]. In the literature, there is no consensus on the time of agitation; it ranges from a few hours [41] to 24 h [42] to seven days [43]. However, all soluble components release into solution within 24 h in most cases.

3.3.3. Sequential Extraction Scheme

The algorithm of the stepwise leaching procedure itself was chosen based on the general geochemical parameters of the tailings: composition, pH of the media, high content of carbonate minerals, and low mobility of cations. At the same time, it was important to consider increased content of magnesium compounds in tailings and arsenic oxyanions adsorbed on them/co-precipitated with it.

It was decided to use a combined six-stage sequential extraction procedure of As leaching (

Table 1. Sequential extraction scheme.

Steps	Extractable Components/Forms	Extractant and Conditions	Possible Mechanism
F1	Water-soluble, arsenic forms	H2O (V = 50 mL) t 24 h	Dissolution in water
F2	Arsenic forms sorbed on the surface of magnesium minerals and deposited as magnesium arsenates/arsenites	0.1 M NaH2PO4 (pH = 8, V = 40 mL) t 24 h	Anionic exchange of ion phosphate for arsenate and arsenite
F3	Arsenic forms sorbed on the surface of carbonates	0.1 M CH3COOH (V = 40 mL) pH 4 t 24 h	Dissolution of carbonate minerals with separation into solution As
F4	Arsenic forms associated with iron hydroxides	2 M NH2OH • HCl in 0.1 M CH3COOH solution (pH = 2, V = 40 mL), 12 h in a water bath (T = 80 °C)	Reduction of Fe oxyhydroxides
F5	Forms of arsenic associated with oxidizable minerals and organic tailings	H2O2(conc) (V = 25 mL, T = 80 °C)	Oxidation of organic tailings and easily oxidizable minerals
F6	Arsenic in the composition of hardly soluble minerals	Decomposition with a mixture of H2O2 and HNO3 by heating in a water bath (V = 25 mL)

) in combination with a non-chromatographic determination of arsenic species, arsenite (As^III) and arsenate (As^V).

This scheme is a hybrid of several previously used sequential extraction procedures developed and tested on oxidized soils and mineral processing waste, including BCR and CIEMAT schemes [44,45,46,47,48]. The total of the As contents recovered by the various reagents differs slightly from the results obtained by full acid microwave decomposition.

Samples weighing 1 g were placed in 50 mL centrifuge tubes, then the necessary volumes of extraction reagents were added successively and placed on a vibration table for the required time. After each extraction step, the suspension was centrifuged for 20 min at 5000 rpm. The supernatant was decanted and, if necessary, acidified with high-purity HNO₃; then, immediately after preparation, the total elemental composition in the solutions was determined, including the total arsenic content and the content of As^III.

The water extracts (F1) were divided into three parts: the first part was preserved with ultrapure nitric acid in a ratio of 0.25 mL of acid per 10 mL of solution for determination of the elemental composition. The second part was prepared for analysis of As^III. The third aliquot was reserved for anion analyses. Solutions of fractions 2–4 were divided into two parts for bulk analysis and determination of As^III. In solutions of fractions 5 and 6, after the oxidation of the substance, only the total composition was analyzed. All resulting solutions were analyzed immediately after procedures.

The pH of the solutions (solutions F1-F5) was controlled using a laboratory pH meter Anion-4100 (Infraspak-Analyte, Novosibirsk, Russia) equipped with a combined ESK-10602 electrode.

3.4. Electrotomography Processing

Data preprocessing and filtering were performed using the SibER Tools program, which is designed to create and edit files with a description of installations, as well as data preprocessing, layout, and exportation to common formats (IPI2Win, Res2dInv, and Res3dInv). Further processing and data inversion were performed in Res2dInv (Geotomo Corp., Houston, TX, USA), and the processing of a three-dimensional data array was carried out in Res3dInv. The visualization of the sections was performed using the Golden Software program Surfer. It was performed for the 3D images using the Golden Software program Voxler. It was performed for the final layout of the drawings using the CorelDRAW Graphics Suite.

3.5. Laboratory Analyses

3.5.1. Solid

The contents of the major oxides in the bulk solid samples were determined via X-ray fluorescence (XRF) analysis by using 3-g sample aliquots at the Analytical Center of the Institute of Geology and Mineralogy, SB RAS. The elemental compositions of the solid materials were determined using inductively coupled plasma atomic mass spectrometry via the ELAN-9000 DRC-e, (Perkin Elmer, Shelton, CT, USA) (“PLASMA” Company, Tomsk, Russia).

3.5.2. Solutions

The cationic composition of solutions was determined by inductively coupled plasma atomic emission spectrometry (ICP-AES) and microelements by ICP-MS.

An ICP-AES instrument iCap 6000 Duo (Thermo Scientific, Waltham, MA, USA) in a standard configuration and with an alternative hydride generation system was applied. The working parameters of ICP-AES were the following: the power supply was 1150 W; the nebulizer flow rate was 0.75 L∙min⁻¹; the flow rate for the sample was 10 mL∙min⁻¹; the flow rate for the reducing agent was 2.7 mL∙min⁻¹; and the flow rate for the acid was 1.25 mL∙min⁻¹. The data acquisition and processing were performed by iTEVA software (Software for ICap ICP-AES spectrometers. Cat. No. 8499 400 30001, Thermo Scientific, Waltham, MA, USA).

An Agilent 8800 mass spectrometer (ICP-QQQ, Agilent Technologies, Santa-Clara, CA, USA) in a standard configuration and using a reaction–collision cell with helium and oxygen was used to determine the elements by ICP-MS technique. High-purity argon (99.95%) was used as a plasma-forming, transporting, and cooling gas. A solution of Li, Mg, Co, Y, and Tl in 2 % nitric acid with a concentration of 1 μg/L for each determined element was used for the adjustment. All analytes were measured in He mode, Hydrogen mode, and MS/MS O₂ mode. A Sc, Y, and Tl internal standard solution (15 µg/L) was mixed online with the sample via the standard ISTD mixing T-connector. All measurements were conducted in three replicates (n = 3) for each element. The relative standard deviation did not exceed 13% for all measurements.

3.5.3. As Species

Determination of arsenic (III) content in solutions was carried out by atomic emission spectrometry with hydride generation technique (HG-ICP-AES) according to the guide (Thermo Scientific, Waltham, MA, USA) [49]. The content of arsenic (V) was determined as the difference between the total content and the content of As (III).

3.5.4. Anions

The anionic composition of the solutions was determined by capillary electrophoresis using the 105-M Kapel system (Lumex, St. Petersburg, Russia). Quantitative analysis was performed using external calibration and processing of results in software for capillary electrophoresis Elforan (Version 4.2.4, Lumex, St. Petersburg, Russia). The relative standard deviation was less than ± 15% in the concentration range of 0.1 to 200 mg/L.

4. Results of Geochemical and Geophysical Studies

4.1. Solid

Composition of the tailings solid has been previously described in detail [35]. The combined sample for experiments, on the whole, corresponded to the average composition of the tailings and contained more aluminosilicate components and slightly less metals and metalloids (

Table 2. Composition of solid tailings (sample KhA_tot): SiO₂–LOI in % and Cr–Sr in ppm.

Component	Content	Element	Content
SiO2	44	Cr	61
TiO2	0.60	Co	420
Al2O3	12	Ni	370
Fe2O3	8.0	Cu	750
MnO	0.26	Zn	290
MgO	3.4	As	3900
CaO	15	Ag	8.4
Na2O	0.42	Cd	0.78
K2O	2.3	Sn	1.7
P2O5	0.23	Sb	35
BaO	0.05	Pb	30
SO3	3.5	Bi	19
LOI	10	Sr	100

). It should be noted that our task was to characterize the storage surface, which is more prone to oxidation and removal of ore elements and, at the same time, to the inflow of natural material.

4.2. Water Extract

The water extract is a subalkaline solution with low salinity (TDS = 0.2 g/L) (

Table 3. Composition of water extract (sample KhA_totWE): Eh in mV, SO₄²⁻–Sb in mg/L, and Mn–Sr in µg/L.

Component	Content	Element	Content
pH	8.07	Mn	3.1
Eh	410	Co	16
SO42−	44	Ni	5.6
Cl−	0.4	Cr	0.15
NO3−	9.8	Cu	4.3
HCO3−	105	Zn	3.2
Ca2+	11	Pb	0.053
Mg2+	29	Ag	0.005
Na+	0.75	Cd	0.018
K+	1.5	Sn	0.065
Fe	0.022	Hg	0.024
Al	0.011	Ba	26
As	26	Sr	40
Sb	0.21

). In anionic composition, hydrocarbonate (HCO₃⁻) and sulfate (SO₄²⁻) are major contributors to TDS; the major cation is Mg²⁺. The concentration of As (26 mg/L) is almost the same as the major cation (Mg²⁺). The concentrations of metals (Cu, Zn, Co, Ni, and Fe) in the water extract are low compared to metalloids.

4.3. Arsenic Species

As a result of the sequential extraction procedure, the arsenic content was estimated in the fractions of the leaching solutions to determine the percentage of different arsenic species. The content of element species (C) in the solid was calculated using the formula:

C(mg/kg) = C(mg/L) ∗ V(L)/m(kg)

(1)

where V is the volume of solution in liters (L), and m is the sample weight (kg).

Table 4. Arsenic content in solutions (mg/L) and solid tailings (mg/kg).

	Fraction	F1	F2	F3	F4	F5	F6	Total Content
Solutions	AsIII	3.1 ± 0.5	13 ± 1	7.6 ± 0.3	10 ± 1	-	-	-
	AsV	5.0 ± 0.9	20 ± 5	10 ± 1	8 ± 3	-	-	-
	Astotal	8.1 ± 0.8	33 ± 5	18 ± 1	18 ± 3	33 ± 0.5	12 ± 1	-
Solid	AsIII	160 ± 25	530 ± 60	300 ± 20	400 ± 40	-	-	1400 ± 140 (For F1-F4)
	AsV	240 ± 50	770 ± 160	420 ± 45	320 ± 130	-	-	1750 ± 220 (For F1-F4)
	Astotal	400 ± 40	1300 ± 150	720 ± 40	720 ± 120	810 ± 20	330 ± 20	4300 ± 600

shows the results of the analysis of the content of arsenic in various fractions, both in solutions of stepwise leaching and in terms of solid tailings.

The distribution of arsenic species by fraction in the solid, shown in Table 4, was evaluated according to the results:

(F1)

In total, 9.3% of the total arsenic is represented by water-soluble forms, specifically—3.7% of As^V and 5.6 % of As^III (for example, in Mg, Ca, ammonia, arsenate, and arsenite according to its solubility product);

(F2)

in total, 30% is represented by bound/co-precipitated with magnesium, i.e., potentially water-soluble forms, 18% of As^V and 12% of As^III (corresponding to Mg and Ca arsenates and arsenites remaining in the precipitate according to their solubility product);

(F3)

in total, 17% is represented by adsorbed on the surface of carbonates (calcite and dolomite) 9.8% of As^V and 7% of As^III;

(F4)

in total, 17% is associated with oxides/hydroxides of iron/manganese, of which 9.3% is As^III, 7.5% is As^V;

(F5)

in total, 19% is associated with easily oxidized minerals (residual arsenides and isomorphic admixture in sulfides) and organic material;

(F6)

and in total, only 7.7% is accounted for by non-oxidized arsenic minerals (for example, As-crystallohydrates, As^III oxides).

Thus, the percentage of the most mobile forms of arsenic is 56% of the total arsenic in tailings materials, represented by water-soluble forms (fraction F1), potentially water-soluble (fraction F2), and adsorbed on the surface of carbonates (fraction F3), of which 23% is As^III and 33% is As^V (

Table 5. Distribution of arsenic species in leach fraction solutions. %.

Fraction	F1	F2	F3	F4	F5	F6	Share
AsIII	3.7	12	7.0	9.3	-	-	32 (For F1-F4)
AsV	5.6	18	9.8	7.5	-	-	41 (For F1-F4)
Astotal	9.3	30	17	17	19	7.7	100

).

The experiment and the developed approach to sequential extraction showed its applicability for assessing the content of mobile forms of arsenic in the arsenide tailings, combined with an analytical arsenic speciation. It is shown that the content of directly water-soluble forms of 9.3% is limited by the solubility of the corresponding compounds of magnesium arsenates and arsenites. As previously shown for Khovu-Aksy tailings, a single water extraction does not allow the complete extraction of all potentially water-soluble forms of arsenic. Therefore, the logic of using a phosphate buffer for the extraction of all potentially water-soluble forms of arsenic in this work is confirmed. This made it possible to estimate the total content of potentially water-soluble forms of arsenic (which is 39% of the total As) due to competitive anion exchange. Thus, release arsenic into solutions is limited by the presence of a heterogeneous “magnesium arsenate-arsenite buffer”.

At the same time, the content of arsenic forms sorbed on carbonate minerals is 17%. These forms can also be considered as potentially mobile arsenic. Given that when the pH of the medium changes, arsenic associated with magnesium is washed out, these forms associated with carbonates are also capable of further migration and transport. It is also necessary to consider the different toxicity of As^V and As^III contained in potentially water-soluble and mobile forms. As a result, it was shown that the percentage of water-soluble forms of arsenic (F1-F2) in the tailings is 39% of the total arsenic content, and the percentage of potentially mobile forms of arsenic (F1-F2-F3) is already 56%. Thus, the content of arsenic in water during the interaction of seasonal precipitation (rain and snow) will be several orders of magnitude higher than the WHO-recommended Permissible Limit of arsenic concentration in groundwater (10 µg/L) [2]. Moreover, there are significant contents of As^III, as a more toxic form of arsenic relative to As^V.

In addition to carrying out the sequential extraction procedure, the obtained results were compared with the results of the content of the components in a single aqueous extract prepared in a ratio of 1:10 substance–water (

Table 6. Comparison of the content of water-soluble forms.

Component	Water Extract	F1 Fraction	F2 Fraction
Solutions, mg/L	26 ± 3	8.1 ± 0.8	33 ± 5
Solid sample, mg/kg	260 ± 30	400 ± 40	1300 ± 150

). According to the results obtained, it can be concluded that for these tailings, the use of single water extracts to assess the content of water-soluble forms of arsenic is not correct because it gives a significant underestimation. The share of water-soluble forms according to a single water extract is 6% of the total content, whereas, according to the data of stepwise leaching, the share of potential water-soluble forms (F1-F2) is already 39%. The difference is almost six times.

4.4. Forms of Metals

During the experiment, the concentration of main cations and metals in the leaching fractions were also determined (

Table 7. Content of cations and metals in leach fraction solutions: Mg and Ca in mg/L and Sr–Pb in μg/L.

	Fraction	F1	F2	F3	F4	F5	F6
Elements
Mg		6 ± 1.5	35 ± 7	20 ± 7	125 ± 25	330 ± 50	280 ± 20
Ca		30 ± 15	60 ± 30	100 ± 50	1500 ± 130	290 ± 80	280 ± 60
Sr		50 ± 3	220 ± 10	220 ± 90	1500 ± 130	620 ± 60	280 ± 10
Cd		<10	<10	<10	90 ± 15	60 ± 10	140 ± 20
Co		20 ± 6	390 ± 80	1000 ± 400	7600 ± 740	5400 ± 680	1700 ± 300
Cu		10 ± 6	160 ± 30	300 ± 140	10,000 ± 1000	11,000 ± 800	10,000 ± 1600
Ni		8 ± 2	180 ± 50	370 ± 110	5000 ± 670	10,000 ± 1600	10,000 ± 1500
Zn		<10	440 ± 110	320 ± 100	5300 ± 660	6000 ± 1000	4200 ± 700
Ba		<15	20 ± 6	25 ± 15	1300 ± 400	1600 ± 200	960 ± 200
Cr		<25	<75	<75	250 ± 90	900 ± 130	1200 ± 100
Mn		<15	85 ± 20	490 ± 360	23,000 ± 3000	19,000 ± 4000	14,000 ± 1500
Fe		<300	<300	<300	1400 ± 500	590,000 ± 70,000	760,000 ± 70,000
Sb		30 ± 3	70 ± 8	40 ± 4	170 ± 20	25 ± 5	290 ± 30
Pb		<10	<20	<20	400 ± 200	620 ± 180	250 ± 150

and

Table 8. Distribution of elements in the solid tailings by mobility fractions (in numerator): Mg and Ca in g/kg, Sr–Pb in mg/kg, and percentage of each fraction (in denominator), %.

	Fraction	F1	F2	F3	F4	F5	F6	Total Content
Elements
Mg		0.30 ± 0.075	1.4 ± 0.28	0.79 ± 0.28	5.0 ± 1.0	8.1 ± 1.2	7.7 ± 0.55	23 ± 1.7
		1.3	6.1	3.4	22	35	33
Ca		1.5 ± 0.74	2.4 ± 1.2	3.9 ± 1.9	60 ± 5.2	7.1 ± 1.9	7.7 ± 1.5	83 ± 13
		1.8	2.9	4.7	72	8.6	9.3
Sr		2.5 ± 0.150	9.0 ± 0.4	8.7 ± 3.5	60 ± 5.2	15 ± 1.5	7.7 ± 0.27	100 ± 11
		2.5	9.0	8.7	60	15	7.7
Cd		<0.50	<0.40	<0.40	3.6 ± 0.60	1.5 ± 0.25	3.5 ± 0.50	8.6 ± 1.4
		0	0	0	42	17	41
Co		0.99 ± 0.30	16 ± 3.3	40 ± 16	300 ± 30	130 ± 17	43 ± 7.5	540 ± 74
		0.18	3.0	7.4	56	24	7.9
Cu		0.49 ± 0.30	6.4 ± 1.2	12 ± 5.6	440 ± 40	270 ± 20	250 ± 4.4	1000 ± 100
		0.05	0.64	1.2	44	27	25
Ni		0.40 ± 0.099	7.3 ± 2.0	15.0 ± 4.3	200 ± 27	240 ± 39	250 ± 38	720 ± 110
		0.06	1.0	2.1	28	33	35
Zn		<0.50	18 ± 4.5	13 ± 3.9	210 ± 26	150 ± 24	105 ± 18	500 ± 77
		0	3.6	2.6	42	30	21
Ba		<0.75	0.81 ± 0.24	0.99 ± 0.59	52 ± 16	39 ± 4.9	24 ± 5.0	120 ± 27
		0	0.7	0.8	44	33	21
Cr		<1.2	<1.0	<3.0	10 ± 3.6	23 ± 3.2	30 ± 2.5	62 ± 9.0
		0	0	0	16	37	48
Mn		<0.75	3.5 ± 0.8	19 ± 14	920 ± 12	470 ± 98	350 ± 37.5	1800 ± 270
		0	0.2	1.1	52	27	20
Fe		<1.5	<1.2	<1.2	56 ± 20	15,000 ± 1700	19,000 ± 1700	34,000 ± 3500
		0	0	0	0.16	44	56
Sb		1.5 ± 0.15	2.9 ± 0.33	1.6 ± 0.16	6.8 ± 0.8	0.61 ± 0.12	7.3 ± 0.75	20 ± 2.3
		7.1	14	7.6	32	2.9	35
Pb		<0.50	<0.80	<0.80	16.0 ± 8.0	15 ± 4.4	6.2 ± 3.7	38 ± 17
		0	0	0	42	39	16

). The table does not show alkali metals due to the use of buffer solutions with sodium salts during sequential extraction. Based on general considerations, the following forms of metals are extracted during leaching:

• F1—water-soluble;
• F2 and F3—exchangeable;
• F4—sorbed on the surface of oxides/hydroxides of iron and manganese;
• F5—associated with easily oxidized minerals, as well as being associated with organic material;
• F6—associated with sparingly soluble minerals.

We observe a different result from that of arsenic. The mobile forms of metals have been determined in small quantities. The largest proportion of water-soluble and exchangeable species (F1-F3) is formed by Mg, Ca, and Sr: 11, 9.4, and 20%, respectively (residual and redeposited carbonates) (Table 8). The proportion of water-soluble forms of other metals is < 1% or 0. Note that antimony, like arsenic, exhibits a significantly higher mobility than metals, the percentage of its water-soluble forms is 7.1%, and that of the exchangeable forms is 22%.

The main part of the metals is concentrated in fractions F4-F6, which contains forms sorbed on the surface of Fe and Mn hydroxides, enclosed in easily oxidized minerals, associated with organic material, and in the composition of hardly oxidized minerals. These results confirm the revealed features of the behavior of As and metals in Khovu-Aksy tailings: immobilization of metals due to sorption on newly formed phases and in the composition of poorly soluble minerals, as well as the high mobility of As due to its being in water-soluble and sorbed forms [39].

4.5. Composition of Surface Reservoirs

The forms of occurrence of elements in the solid tailings determine the composition of solutions in reservoirs, which are formed on the surface of the ponds during interactions with seasonal precipitation. Reservoirs in all ponds have approximately the same composition. Water is a subalkaline–alkaline solution with oxidizing conditions of the sulfate–calcium–magnesium type (

Table 9. Composition of water in the surface reservoirs SO₄²⁻–As in mg/L and Fe–Sb in µg/L.

	Samples	P1–1	P1–2	P3–1	P3–2	P4–1	P4–2	P5
Components
pH		8.47	8.24	7.63	8.88	8.49	7.78	8.55
Eh, mV		420	385	370	460	472	340	496
SO42−		140	56	114	115	23	81	74
NO3−		27	34	70	70	23	52	160
HCO3−		70	160	140	140	110	120	87
NH4+		1.1	0.8	1.3	1.4	0.5	0.5	1.0
Ca2+		38	17	12	12	7.7	23	9.3
Mg2+		23	21	52	54	20	23	58
Na+		0.50	0.46	3.6	3.8	0.91	0.83	1.4
K+		2.2	2.0	2.9	3.1	0.68	0.78	1.4
As		12	9.3	4.6	4.8	13	12	7.2
Fe		52	23	15	20	11	39	46
Mn		0.67	2.6	12	12	2.5	12	14
Co		6.7	6.0	7.0	6.1	4.7	16	9.9
Ni		4.8	3.0	3.9	2.9	4.9	12	7.2
Cu		6.1	2.6	0.82	0.76	0.85	1.0	1.3
Zn		2.0	0.42	1.4	0.61	0.91	0.96	1.2
Sb		62	60	41	40	45	37	54

). At low concentrations of metals, the content of arsenic in water is 5–13 mg/L. Considering that a third of the arsenic in the tailings is represented by As^III, it can be argued that the solutions in reservoirs are highly toxic waters.

4.6. Inner Zonality of Pond Dumps

The sections of electrical resistivity tomography demonstrate the alternation of low- and high-conductivity areas in the subsurface space (Figure 2). Zones with low resistivity indicate the presence of flooded lenses, which gradually expand downward from the surface to a depth of more than 40 m. It is likely that incoming seasonal precipitation gradually seeps down and, when interacting with the waste tailings, leaches easily soluble components from it, including arsenic.

As can be seen from the sections, low-resistivity zones tend to expand towards the bottom of the valley in the direction of litho- and hydroflows. Thus, components from the tailings penetrate into groundwater and further: into the horizons of drinking waters. Given the results of stepwise leaching and determination of arsenic species (even on the surface of the tailings, the percentage of As^III as the most toxic form is 32% of the total arsenic content), it can be argued that groundwater is affected by tailings. We emphasize that the existence of conditions for arsenic oxidation in underground horizons is unlikely. Under these conditions, it is possible to predict the stability of As^III and its distribution with groundwater.

5. Conclusions

The performed sequential extraction procedure made it possible to obtain extended data on the content of arsenic and its species in arsenide waste. The chosen scheme took into account the features of the physical and chemical composition of the waste and was supplemented by the analytical determination of the arsenic species by using the methods of inductively coupled plasma atomic emission spectrometry (ICP-AES) combined with hydride generation technique (HG-ICP-AES). During the study, the extraction of the following forms of arsenic and metals was carried out: water-soluble, potentially water-soluble and exchangeable, easily sorbed on the surface of carbonates, associated with Fe/Mn oxides/hydroxides, associated with easily oxidized minerals, and accounted for by non-oxidized arsenic minerals. The percentage of the most mobile forms of arsenic is 56% of the total arsenic in tailings materials, represented by water-soluble forms (fraction F1), potentially water-soluble (fraction F2), and adsorbed on the surface of carbonates (fraction F3). Arsenate is the predominant species of mobile arsenic in tailings (33% of As^V). The share of arsenite is lower (23% of As^III); however, given its high toxicity, it can be argued that Khovu-Aksy tailings are dangerous when the substance is washed out by seasonal precipitation. Metals (Co, Ni, Cu, Zn, and Pb), unlike metalloids (As and Sb), are contained in inert, poorly soluble forms. The main quantity of metals was determined in fractions F4-F6, into which metals adsorbed on the surface of oxides enclosed in easily oxidized minerals, associated with organic tailings, and in the composition of hardly oxidized minerals were extracted.

Geoelectrical zonality of the subsurface space in close proximity to the ponds indicates the existence of highly conductive areas, which expand to the bottom of the valley. The formation of such zones is associated with the infiltration of surface waters through the substance of the tailings into underground horizons. The presence of a large amount of As^III in tailings and its migration with streams creates a great danger for the population when contaminated groundwater enters drinking water sources.