Bioaccessibility of potentially toxic elements in mine residue particles

Abstract

Mining companies used to abandon tailing heaps in countryside regions of Mexico and other countries. Mine residues (MRs) contain a high concentration of potentially toxic elements (PTE). The wind can disperse dust particles (<100 μm) and once suspended in the atmosphere, can be ingested or inhaled; this is a common situation in arid climates. Nowadays, there is little information on the risk of exposure to PTEs from particulate matter dispersed by wind. The pseudo-total PTE in bulk and fractionated MR after aqua regia digestion, the inhalable bioaccessibility with Gamble solution (pH = 7.4), and the gastric bioaccessibility with 0.4 M glycine solution at pH 1.5 were determined. As and Pb chemical species were identified by X-ray absorption near-edge structure (XANES) spectroscopy. The highest rate of dispersion was observed with 74–100 μm particles (104 mg m⁻²s⁻¹); in contrast, particles <44 μm had the lowest rate (26 mg m⁻² s⁻¹). The highest pseudo-total As (35 961 mg kg⁻¹), Pb (3326 mg kg⁻¹), Cd (44 mg kg⁻¹) and Zn (up to 4678 mg kg⁻¹) concentration was in the <20 μm particles and As in the 50–74 ¼m (40 236 mg kg⁻¹) particles. The highest concentration of inhaled bioaccessible As (343 mg kg⁻¹) was observed in the <20 μm fraction and the gastric bioaccessible As was 744 mg kg⁻¹, Pb was 1396 mg kg⁻¹, Cd was 19.2 mg kg⁻¹, and Zn was 2048 mg kg⁻¹. The predominant chemical As species was arsenopyrite (92%), while 54% of Pb was in the adsorbed form. Erodible particle matter is a potential risk for humans in case of inhalation or ingestion.

1. Introduction

Mining generates a large amount of waste. Mine residues (MR) have high concentrations of potentially toxic elements (PTE)¹ and fine grains of size less than 200 μm.² The small particle size favors MR dispersion and resuspension in the air.^3-5 Ettler et al.,⁶ Csavina et al.⁷ and Kastury et al.⁸ considered MR and smelting and mining operations potential sources of emission of fugitive particulate matter (PM) with which high concentrations of PTE are associated. Nowadays, air pollution is one of the most serious environmental issues. In Mexico and Latin American countries, non-ferrous metal mining, which has been practiced since the XVI century,⁹ has left huge amounts of residues. Like other countries around the world, the proximity of communities to the MR is a potential health risk for the residents,^6,10 especially through ingesting or inhaling dispersed PM,^1,11,12 particularly accidental dust ingestion may occur in children.¹⁰ When PM is inhaled or ingested, a fraction of the PTE can be solubilized and be absorbed from the tracheobronchial region, lungs, or the digestive tract into the circulatory system¹³ and have adverse effects on health.^12,14

In general, the total or pseudo-total concentration is used to determine the quantity of PTE in soils, MR, air, water, and other possible sites of contamination. However, when PM is inhaled or ingested, a fraction of the PTE total concentration that can dissolve in body fluids and directly affect human health represents a more real parameter of potential human health risk;⁸ in other words, PTE become bioaccessible. For this reason, new techniques such as studying bioaccessibility have been proposed to evaluate exposure to PTE in humans.¹⁰ These analytical procedures to simulate the bioaccessibility are correlated with in vivo toxicological studies using animal models.^8-11 The bioaccessible concentration is the pollutant concentration that solubilizes in simulated body fluids (pulmonary or gastric) to be predictive of human exposure.¹¹ The bioaccessibility approach is a rapid and cheap way to assess exposure to PTE.¹⁰ According to Kastury et al.,⁸ small particles (PM_2.5) may be engulfed by macrophages. However, 90% of the PM₁₀ deposited in the tracheobronchial airways is cleared from the lung within 24 h by mucociliary action and swallowed, so these particles travel through the gastric system. Due to the high cost of in vivo or swine model studies,¹⁰ to simulate the release of metals in the body, a sequential extraction should be mimicked. We may simulate this process using two-stage sequential extraction: inhalable bioaccessibility followed by gastric bioaccessibility.^12,13 Human populations become exposed to PTE when metallic MR are deposited close to urban and suburban areas,¹⁴ either by inhalation or by incidental ingestion of PM,^6,10,15,16, and several studies have linked fine and ultrafine particle exposure to respiratory diseases and mortality.¹⁷ Elemental components may contribute to airborne particulate matter toxicity, mainly due to the generation of reactive oxidant species.¹⁸ From a conservative point of view, small particles are more likely to be suspended in the air and inhaled.⁶ However, particles larger than 10 μm are retained in the conducting airways and do not move inside the gas-exchange region.¹⁹ Consequently, it is relevant to quantify the bioaccessible PTE concentration (inhalable and gastric)¹³ in MR and the quantity of PTE that is dispersed by PM as these pollutants represent a toxic risk for the human population living near areas where mining takes place. Many studies have addressed the availability of PTE in small particles susceptible to suspension in the air from polluted subtracts.¹⁴ In mining towns where many mining companies have operated over time, several tailing heaps had been abandoned uncovered in open areas,^6,10 and the particle matter is suspended in the air by blowing wind.¹⁴ Particles of size smaller than 100 μm are more likely to be suspended.^10,14, Monitoring stations and wind tunnel experiments may be used for tracking the source of PM from these residues.²⁰

For this study, Zimapan, Hidalgo, a small town in central Mexico, was selected as a model system where several factors come together: MR deposited in the open areas, non-compacted residues with fine grain size (<200 μm), semi-arid climate with little rain, and population living close to the residues. The objectives were to measure the erosion rate of three particle sizes (Fraction 1: <20 μm, Fraction 2: 50–74 μm, and Fraction 3: 74–100 μm) of MR through the wind tunnel; to quantify the pseudo-total concentration of As, Pb, Cd, and Zn in the MR susceptible to dispersion by the wind in the size fractions; to determine the bioaccessible concentration of As, Pb, Cd, and Zn in inhaled and gastric phases in the three fractions; and to identify the As and Pb chemical species in particles of MR.

2. Materials and methods

2.1. Study location and sampling

After a field survey, a tailing heap from the Santa Maria mine of the Zimapan mining district, Hidalgo, Mexico, was selected. The tailings are in the southwest (SW) of the urban area (Fig. S1^†). During the dry season, the dominant winds (October to April) blow in a northeasterly (NE) direction. The MR is susceptible to wind dispersion toward the urban area of Zimapan (Fig. S2^†). The predominant wind strength is 3.6 to 5.7 m s⁻¹, although it can reach 8.8 m s⁻¹ in gusts.²¹ By systematic sampling, 40 samples of MR were collected (Fig. S1^†) to determine the pseudo-total and bioaccessible As, Pb, Cd, and Zn concentrations. Moreover, 40 kg of residue were collected for the wind dispersion experiments using a wind tunnel. The MR was taken from the surface layer (0–5 cm), as suggested by Drahota et al.¹⁴ In situ analysis of resistance to penetration, measured with a Dickey-John® penetrometer, was performed at the 40 sampling sites. The samples were dried at room temperature and sieved in an analytical machine (AS 200 basic®) to homogenize. Small particles are more susceptible to suspension as dust in the atmosphere. To evaluate the proportion of small particles in the tailings and the potential erodibility, the samples were separated into three particle sizes: <20 μm (F1), 50–74 μm (F2), and 74–100 μm (F3). The particle sizes were selected based on the available literature,^6,13,17,19 which indicated that particles with aerodynamic diameters <100 μm can easily be suspended by wind and remain suspended for long periods.^10,18 Moreover, Brown et al.¹⁹ demonstrated that particles of sizes up to 100 μm can be deposited in the respiratory (tracheobronchial) and gastric systems.

2.2. Analysis of dispersion and magnetic measurement of MR

For the wind tunnel dispersion experiment, the MR was separated into 50–74 and 74–100 μm fractions. However, particles smaller than 44 μm were used as F1, because no more than 1% of particles smaller than 20 μm was found. Straight-line particle dispersion was analyzed in a wind tunnel with the following dimensions: 0.3 m height, 0.6 m width, and 3.5 m length (Fig. S3^†). With these measurements, a length/height ratio of 11.6 was achieved, a ratio higher than the minimum value (5) suggested by White and Mounla²² for the natural development of a top layer. Separately, three replicates of the mine residue fractions were exposed for 10 min to three wind speeds (2.7, 4.0, and 5.2 m s⁻¹). In total, 27 samples were tested. These three speeds were observed within the typical wind speed range registered for the municipality of Zimapan.²¹ Wind speed was controlled with a metal Taurus® fan. Speed was corroborated inside the tunnel with a wireless anemometer (Davis® model 6250 Vantage Vue). The residue (500 g; initial weight, W_i) was placed at the entrance of the tunnel in front of the fan. At the end of the tunnel, a 2 × 2 m polyethylene bag collected the dispersed MR, which was later weighed (final weight, W_f); t is the time and A is the transversal area. The wind rate erosion (g m⁻² s⁻¹) of the MR²³ was calculated using this equation.

$$\text{WER}=\frac{(W_{\mathrm{i}}-W_{\mathrm{f}})}{t\times A}$$

To explain the differences in dust dispersion between particle sizes, the isothermal remanent magnetization (IRM), volumetric magnetic susceptibility (κ), and specific magnetic susceptibility (χ = κ/ρ) were estimated. κ was obtained with a Bartintong MS3 susceptibility meter with a dual MS2B sensor in a magnetic field of ~80 A m⁻¹, at low and high frequency (0.465 kHz and 4.65 kHz, respectively). Low-frequency susceptibility (κ_lw) is a measure of the concentration of magnetic minerals and represents the sum of all the ferric-ferromagnetic and diamagnetic minerals. High-frequency susceptibility (κ_hg) is proportional to the concentration of ultra-fine magnetic grain size. The frequency-dependent magnetic susceptibility (%) was calculated, with κ_lw and κ_hg, which is useful to quantify the presence of ultra-fine (<0.03 μm) superparamagnetic (SP) ferri-magnetic minerals using the following equation:

$$\%x=\frac{(k_{\text{lw}}-k_{\text{hg}})}{k_{\text{lw}}}\times 100$$

The IRM was obtained with an IM 10–30 ASC Scientific® pulse magnetizer to saturate the samples to 300 mT, 1000 mT, −25 mT, and −300 mT. Remanent magnetization was measured at each step with an ASC Scientific® gyro magnetometer. The IRM was obtained at 1000 mT, which is considered the saturation value of isothermal remanent magnetization (SIRM), and with magnetization −300 mT, it is possible to obtain the indicator of the relative concentration of minerals of low magnetic coercivity, concerning minerals of high magnetic coercivity (S-300).

2.3. Pseudo-total concentration and bioaccessibility of PTEs in MR

The pseudo-total concentration of As, Pb, Cd, and Zn was measured following the ISO²⁴ method: 250 mg of the sample were digested with 6 mL of HNO₃–HCLO₄ (3 : 1 ratio). The samples were placed in a microwave digestion oven at 120 °C for 4 h. After cooling, the digested samples were gauged to 25 mL with de-ionized water and filtered with Whatman No. 42.

The analysis of PTE bioaccessibility was performed in two sequential phases: inhalable (INBA) and gastric (GABA).⁸ The inhaled phase extract was obtained following the Mos methodology.²⁵ The samples were weighed (200 mg) in 50 mL polypropylene tubes and 20 mL of Gamble solution (pH 7.4) was added. The samples were equilibrated for 24 h at 37 ^°C, and then centrifuged at 4000 rpm for 10 min and filtered through Whatman No. 42 filter paper. The gastric phase extract was obtained following method described by Juhasz et al.²⁶ Ten mL of the gastric solution (glycine 0.4 M, pH = 1.5) was added to the solid residue from the previous step. The samples were equilibrated for one hour at 37 °C and filtered with No. 42 Whatman paper. Lead, Cd, Zn, and As were determined by atomic absorption spectrometry (Perkin Elmer®, model analyst 700) in the pseudo-total, inhaled, and gastric phases. To have a handle mass of particles at the laboratory scale²⁷ and assuming a maximum of 200 mg of inhaled particles, a medium S/L (1 : 100) ratio was used for the extraction.¹⁴ The percentage of the pseudo-total content was calculated (the bioaccessible concentration). For Q_c/Q_a, the extraction was performed in triplicate, and a certified reference material was used for instrumental calibration.

2.4. Chemical speciation of As and Pb in MR

Lead and As chemical speciation was determined in the samples of MR that had the highest concentrations of these elements. For F1, eleven samples were selected, while for F2 and F3, eight samples were used. Arsenic and Pb species were identified using X-ray absorption near edge structure (XANES) spectroscopy. Measurements for Pb (L_III-edge: 13 035 eV) at the 10-ID beam line²⁸ and As (K-edge, 11 867 eV) at the 10-BM beam line²⁹ were conducted at the Materials Research Collaborative Access Team (MRCAT) sector of the Advanced Photon Source (APS), Argonne National Laboratory. A double silica crystal monochromator (111) (cooled with liquid nitrogen to select incident photon energies) and a platinum-coated mirror for harmonic rejection were employed for Pb measurements. The samples were pressed into 7 mm pellets and 10 scans were obtained for each sample (—200 to 600 eV) in fluorescence mode using a Lytle detector purged with Ar for Pb and a 4-element Vortex detector for As. Spectral references of metallic Pb (13 035 eV) or sodium arsenate (11 874 eV) were used to calibrate the monochromators and collected congruently with each sample scan for verification during data analysis. The scans were merged and calibrated to obtain the final scan for each sample. The background noise was corrected with a linear fit to the region before the edge step. Spectral heights of the steps were normalized using a posterior band. The XANES region was selected for the study and the spectra were converted into derivatives for analysis. Arsenic and Pb patterns were collected and prepared previously following the procedure described by Juhasz et al.³⁰ The patterns were later analyzed to facilitate the linear combination fit (LCF) of the sample spectra and determine the As and Pb species in the samples of this study. Arsenic, Pb, Cd, and Zn associated with carbonates and iron oxides were measured by 1 M ammonium acetate buffered at pH 5 with acetic acid³⁰ and oxalate extraction procedures.³¹ Geostatistical analysis was carried out using the database constituted by UTM geographical coordinates, and the EPT pseudo-total and bioaccessible concentration data from the 40 sampling points in the fractions of the tailings heap were mapped by kriging using ArcGis 10.4.1.³²

2.5. Statistical analysis

We corroborated that the variables complied with the assumptions of normality (Shapiro–Wilks, α = 0.05) and variance homogeneity (Bartlett, α = 0.05). Some of the data that did not satisfy the assumptions were converted with Box–Cox power transformation and others with logarithm (base 10). The significant differences in pseudo-total, INBA, and GABA of As, Pb, Cd, and Zn concentrations, as well as speciation, were obtained with analysis of variance (ANOVA; α = 0.05) and with the Tukey comparison of means test (α = 0.05). The statistical analyses were carried out using R-Studio software, version 3.5.0 for Windows.

3. Results and discussion

3.1. Rate of wind blow of MR in the wind tunnel and magnetization

At the present, the mining company operates and the tailings are deposited. The average resistance to penetration in the Santa Maria mine tailing was 171 ± 129 kPa. According to Murdock et al.³³ none of the sites had MR compacted since the readings are below 2068 kPa, and therefore, these are susceptible to wind erosion and transport.³⁴ The material from the tailing is fine-grained, and the physical fractionation showed that 33% of it had grain size <100 μm, which was distributed as follows: F3, 14%; F2, 11%, and F1, 8%. This means that by particle size (<100 μm), a third of the residue is erodible and can generate emissions of dust. One-third of the tailing can be suspended in the air during the dry season, increasing the risk of metal exposure for local people. Kim et al.³⁵ and Martin et al.³⁶ observed that up to 45% of the mining residues in Victoria, Australia, and California, USA, were particles ≤100 μm. The results of the present study are in line with other reports, studying the dust dispersion from tailings dam. Kríbek et al.³⁷ found a high rate of dispersion close to a small particle size, whereas the coarse-grained dust particles settle and accumulate near the tailings heap. The wind erosion rate was proportional to the wind speed, as the power function expresses (Table 1).

Table 1.

Wind erosion rate of three particle fractions from the Santa Maria mine tailing exposed to three wind speeds in a wind tunnel^a

Wind speed (m s−1)	Wind erosion rate by sizes (g m−2 s−1)			Equation	R 2
	<44 μm	50–74 μm	74–100 μm
2.7	0.013 Bc	0.022 Bc	0.062 Aa	Y= 2 × 10−5x2.29	0.964
4.0	0.025 Bb	0.033 Bb	0.096 Aa	Y= 1 × 10−5x1.971	0.894
5.2	0.040 Aa	0.049 Aa	0.152 Aa	Y = 0.0086x0.0319	0.919
Average	0.026 B	0.035 AB	0.104 A

^aDifferent capital letters show a statistical difference between particle sizes at each wind velocity, according to the Tukey test (α = 0.05). Different lower-case letters show a statistical difference in the wind erosion rate for each particle size, according to the Tukey test (α = 0.05).

There was a direct relationship between the erosion rate and particle size. The lowest wind erosion rate (26 mg m⁻² s⁻¹) corresponded to the fraction <44 μm, while F3 had the highest erosion rate (104 mg m⁻² s⁻¹). The particle size distribution is an important factor affecting the rate of dispersion.⁶ The magnetic susceptibility and IRM may help explain the erodibility of the MR (Tables 2 and 3). Mass magnetic susceptibility was 68, 59, and 60 μg³ kg⁻¹ for F1, F2, and F3, respectively. Magnetic properties have been associated with PTE in PM from dust collected from cities³⁸ and automobile emissions.³⁹ Particles with low values were dispersed quicker than the other two fractions. In contrast, the interval of IRM was 0.90 to 1.05, and the largest percentage corresponded to smaller particles (Table 3).

Table 2.

Magnetic susceptibility of the particle fractions of Santa Maria mine tailings^a

Size fraction (μm)	k 1	k 2	k 3	Average	χ = κ/ρ (μg3 kg−1)
<44	8.33 × 10−4	8.27 × 10−4	8.27 × 10−4	8.29 × 10−4	6.83
50–74	7.25 × 10−4	7.25 × 10−4	7.39 × 10−4	7.30 × 10−4	5.90
74–100	7.92 × 10−4	8.03 × 10−4	7.96 × 10−4	7.97 × 10−4	6.06

^ak = volumetric magnetic susceptibility and χ = κ/ρ = specific magnetic susceptibility.

Table 3.

Isothermal remanent magnetization (IRM) in the particle fractions of the Santa Maria mine tailing at Zimapan, Hidalgo^a

Size (μm)	Magnetization (mT)	IRM		S-300
74–100	NRM	7.59 × 10−1	A m−1	0.98
	300	6.34 × 101	A m−1
	1000	6.54 × 101	A m−1
	−25	4.14 × 100	A m−1
	−300	5.93 × 101	A m−1
50–74	NRM	2.79 × 10−1	A m−1	0.96
	300	7.11 × 101	A m−1
	1000	7.72 × 101	A m−1
	−25	1.94 × 101	A m−1
	−300	7.43 × 101	A m−1
<44	NRM	7.07 × 10−1	A m−1	1.05
	300	9.86 × 101	A m−1
	1000	9.76 × 101	A m−1
	−25	3.71 × 101	A m−1
	−300	1.02 × 102	A m−1

^aNRM ¼ natural remanent magnetization, A m⁻¹ = ampere per meter, mT = millitesla.

The IRM values may be due to the presence of magnetic particles in the small size fraction,⁴⁰ and it is likely that ferri-magnetic minerals are involved in the PTE content, so mineral composition affects the dispersion rate. Following the finding of Ayoubi et al.,⁴¹ fine size particles may be deduced from the magnetic data, using linear correlation. Sundborg⁴² pointed out that cohesion is critical for particle resistance to erosion. Moreover, Chen et al.⁴³ determined a positive correlation between the cohesion force and magnetic susceptibility. F1 is more cohesive than F2, so the dispersion rate decreases.

3.2. Pseudo-total PTE concentration in three particle sizes of MR

The distribution of sample sites on the tailing and the As, Cd, Pb, and Zn concentrations are presented in individuals maps. The graphic presentation in Fig. 1-4 enables the pseudo-total and bioaccessible concentrations to be compared. Pseudo-total As (in mg kg⁻¹) ranged from 8000 to 43 000, 5 to 66, 1100 to 4500, and 1100 to 3400 for Cd, Pb, and Zn, respectively (Table 4). The highest pseudo-total concentration (Tukey α = 0.05) was observed in the smallest F1 of the MR. This can be related to an increase in the specific area of this particle size. Several studies have shown an inverse relationship between particle size and pseudo-total PTE concentration.^8,44,45 The windblown particles accumulated in dunes around the tailings. No criteria for the assessment of PTE pollution and bioaccessibility (inhalable or gastric) airborne particles are available in Mexico. Therefore, for the sake of comparison, we used regulations for maximum PTE concentration in residential or industrial soil as a reference for Mexico, UE, and the USA. The Mexican regulation (SEMARNAT⁴⁶) establishes maximum permissible limits for As, Pb, and Cd in industrial soil, which are lower than the concentrations observed.

Fig. 1.

Arsenic distribution, under the kriging model, in three particle sizes from the Santa Maria mine tailing at Zimapan, Hidalgo: (a) pseudo total and (b) inhalable and (c) gastric bioaccessibility.

Fig. 4.

Zinc distribution, under the kriging model, in three particle sizes from the Santa Maria mine tailing at Zimapan, Hidalgo: (a) pseudo total and (b) inhalable and (c) gastric bioaccessibility.

Table 4.

Average concentration of pseudo total, inhalable and gastric bioaccessible, carbonate and iron oxides bound in tailing sample fractions compared to some reported concentrations

Particle fraction μm	Analysis	As, mg kg−1	Cd, mg kg−1	Pb, mg kg−1	Zn, mg kg−1
<20	Ptotala	35 961 ± 22 946	44 ± 22	3326 ± 1484	4678 ± 2140
	INBAb	343 ± 85	2.8 ± 1	44 ± 20	9.5 ± 3.3
	GABAc	743 ± 215	19.2 ± 9.5	1396 ± 565	2047 ± 1485
	Carbonatesd	42.9 ± 50	11 ± 3	234 ± 118	263 ± 98
	Fe oxidese	4234 ± 5090	7.7 ± 5.9	333 ± 237	2602 ± 1487
50–74	Ptotal	40 236 ± 23 255	38 ± 18	1956 ± 811	2780 ± 1043
	INBA	281 ± 76	2.0 ± 0.8	30.6 ± 15	6.7 ± 1.8
	GABA	492 ± 177	10.7 ± 4.5	732 ± 399	891 ± 448
	Carbonates	40 ± 27	4.9 ± 2	92 ± 51	188 ± 209
	Fe oxides	1688 ± 1675	7.1 ± 4	243 ± 166	1166 ± 477
74–100	Ptotal	24 419 ± 17 560	25 ± 11	1762 ± 820	2900 ± 1192
	INBA	259 ± 72	1.6 ± 0.8	29 ± 17	6.8 ± 2.6
	GABA	389 ± 145	9.6 ± 3.9	588 ± 277	935 ± 537
	Carbonates	46 ± 40	4.4 ± 2.6	72 ± 47	116 ± 56
	Fe oxides	1380 ± 1974	6.4 ± 3.9	172 ± 112	841 ± 348
Reference	Analysis	As, mg kg−1	Cd, mg kg−1	Pb, mg kg−1	Zn, mg kg−1
SEMARNAT46	Agr/Ind	22/260	37/450	22/260	400/800
CCME47	Soil	12	22	260	360
USEPA48	Soil	0.11	200	400/1100	0.48
Kastury et al.8	Total	31.6	6.9	13 682	4873
	fPUBA	9.1 ± 0.5		122 ± 3.8
Guney et al.11	PUBA	1549	2.3	380	2.4
	GABA	573	—	—
Drahota et al.14	PUBA	471–7905	—	≤0.5	≤8
	GABA	7–795	—	0.5–23	6–3260
Caboche et al.27	PUBAf	—	56.9 ± 3.4	14.5 ± 0.9	75.5 ± 3.5
Pelfrene et al.50	BCR723	—	DLg	7.8 ± 0.6	44.6 ± 0.8
	NIST27106	—	86 ± 0.28	7.9 ± 0.4	23.7 ± 0.1
	NIST1648a	—	45.2 ± 4	9.1 ± 0.9	43.2 ± 0.2

^aPtotal = pseudo total concentration.

^bINBA = inhaled bioaccessibility.

^cGAPA = gastric bioaccessibility.

^dCarbonates = bound to carbonates.

^eFe oxides = bound to iron oxides.

^fPUBA = pulmonary bioaccessibility.

^gDL = at the detection limits.

The pseudo-total As in <20 μm particles was 35 961 ± 22 946 mg kg⁻¹, in F2 it was 40 235 ± 23 255 mg kg⁻¹ and in F3 it was 24 419 ± 17 560 mg kg⁻¹ (Fig. 1a). The As in F1 was 23 times higher than that determined in MR in Quebec, Canada, for <20 μm particles.¹¹ Likewise, it was 1.9 times higher than the As observed in residue particles <10 μm from mining/smelting impacted regions (18 484 mg kg⁻¹) in Victoria, Australia.⁸ In F2 and F3, the pseudo-total As was 159 and 93 times higher than the regulations;^46-48 it was 3.5 and 2.1 times higher, respectively than that reported by Drahota et al.¹⁴ in the <250 μm fraction of the mine residues in Kaňk, Czech Republic (11 456 mg kg⁻¹). Similarly, it was 2.4 and 1.5 times higher, respectively, than that observed in <45 μm MR in Nova Scotia, Canada (16 672 mg kg⁻¹).⁴⁵

The pseudo-total Pb in F1 was 3326 mg kg⁻¹; in F2, it was 1956 mg kg⁻¹ and in F3, it was 2900 mg kg⁻¹ (Fig. 2a). Lead in F1 was 8.7 and 2.5 times higher than that quantified in MR (<20 μm) in Quebec, Canada (380 mg kg⁻¹)¹¹ and in particles (<10 μm) from Victoria, Australia (1302 mg kg⁻¹).⁸ Moreover, it was 8.7 times higher than the concentration (196 mg kg⁻¹) reported by Guney et al.¹¹ in the <160 μm fraction of MR in Quebec, Canada. Lead in the three fractions (2727 mg kg⁻¹) was in the range found in slags of the most polluted sites^6,14 and higher than the limit established by the USEPA⁴⁷ for residential and industrial soils.

Fig. 2.

Lead distribution, under the kriging model, in three particle sizes from the Santa Maria mine tailing at Zimapan, Hidalgo: (a) pseudo total and (b) inhalable and (c) gastric bioaccessibility.

The pseudo-total Cd is in the range (Fig. 3a) proposed for soils in the Mexican regulations (SEMARNAT).⁴⁵ However, Cd in F1 and F2 was twice as high as the 22 mg kg⁻¹ established limit for Cd in industrial use soils in Canada.⁴⁷ Cadmium in F1 was 6.3 times higher than that quantified in MR (<10 μm) in Australia (6.9 mg kg⁻¹)⁸ and four times higher than that found in the <10 μm fraction of slag dust generated by crushing machines in Zambia (10.7 mg kg⁻¹).⁶ In F2 and F3, Cd was five times higher than the pseudo-total (5 mg kg⁻¹) quantified in the <125 μm fraction of dust on a highway in Islamabad, Pakistan.⁴⁹

Fig. 3.

Cadmium distribution, under the kriging model, in three particle sizes from the Santa Maria mine tailing at Zimapan, Hidalgo: (a) pseudo total and (b) inhalable and (c) gastric bioaccessibility.

The pseudo total Zn was 4678 mg kg⁻¹ in F1, 2780 mg kg⁻¹ in F2, and 2900 mg kg⁻¹ in F3 (Fig. 4a). These values are lower than those observed in bulk slags from Zambia.⁶ F1 has three and 1.7-times higher Zn, respectively, than that determined in <20 ¼m particles of MR (1549 mg kg⁻¹) in Quebec, Canada,¹¹ and in <250 μm samples (1738 mg kg⁻¹) in Kaňk, Czech Republic.¹² In F2 and F3, Zn was similar to the concentration in ≤11 μm particles (2698 mg kg⁻¹),¹⁴ but was 1.7 times higher than that in <160 μm particles (1636 mg kg⁻¹) of MR in Quebec.¹¹ In our study, the pseudo-total Zn in the three particle sizes was at least seven times higher than that established by Canadian regulation (360 mg kg⁻¹).⁴⁷ Ettler et al.¹⁰ pointed out the risk of metal release in the case of particle ingestion because the gastric pH would release these toxic elements.

3.3. Bioaccessible PTE concentration in the inhalable fraction (INBA)

The inhalable concentration was lower than 1% for the pseudo-total As, 6.3% for Cd, 1.6% for Pb, and 0.2% for Zn; the bioaccessible concentrations are slightly lower than that reported by other authors.^6,37 Low percentages of bioaccessibility, compared to pseudo-total concentration, may be related to the low solubility of minerals in the tailings.³⁷ The highest bioaccessible PTEs were observed in F1 (Fig. 1b; Table 4). Although authors such as Brown et al.¹⁹ have demonstrated that airborne particles of up to 100 μm can deposit in the tracheobronchial region and gastric systems, only some papers addressed the inhalable and gastric bioaccessible PTEs from tailing heaps. The INBA-As in F1 was 3.5 times lower than that observed in <10 μm MR in Victoria, Australia (approximately 1200 mg kg⁻¹).⁸ The INBA-As in F2 and F3 can be compared only with the study of Guney et al.,¹¹ who observed that INBA-As was 136 times lower in particle samples of <20 μm. These authors used a two hour extraction time for bioaccessible As, Pb, and Zn, while we used 24 h for the Zimapan samples. Some authors have established 24 h as the optimum time for analysis of inhaled and pulmonary bioaccessibility.^8,11,14,27 Although the INBA-As is less than 2% of the pseudo-total, it was 24.5 times higher than the maximum limit established for Canadian soils⁴⁷ and surpassed the maximum permitted by the EPA.⁴⁸

The INBA-Pb was 1.3% of the pseudo-total in the F1, 1.6% in the F2, and 1% in the F3 (Table 4, Fig. 2b). The concentration decreased as follows F1 > F2 > F3. Bioaccessible lead in F1 was lower than the concentration in MR particles <10 μm (7 mg kg⁻¹) in Victoria, Australia,⁸ and in MR particles <20 μm in northern Quebec, Canada.¹¹ Pelfrêne et al.⁵⁰ reported 436 mg kg⁻¹ of bioaccessible Pb in soil particles (<74 μm) of Montana and atmospheric particles (<100 μm) from St. Louis MO, USA.

In the F1, the INBA-Cd was 6.4%, 5.2% in F2, and 6.4% in F3 of the pseudo-total concentration (Fig. 3b). In F1, INBA-Cd was 2.8 times higher than the INBA-Cd in pulverized carbon ash (<10 μm).²⁷ In contrast, the INBA-Cd in F2 and F3 was 2.5 times lower than that determined in household dust (<100 μm) from residences in the US (4.1 mg kg⁻¹).²⁷

The highest INBA-Zn (0.20% of the pseudo-total; Fig. 4b) was 1.8 times higher than the INBA-Zn found in <11 μm MR in Kaňk, Czech Republic.¹⁴ This concentration was also four times higher than that (2.4 mg kg⁻¹) determined in MR (<20 μm) in Northern Quebec, Canada.¹¹ The average INBA-Zn in F2 and F3 (6.7 mg kg⁻¹; 0.20% of the pseudo-total) was 148 times lower than the concentration quantified in Montana soils (991 mg kg⁻¹) and 309 times lower than that determined in atmospheric particles from St. Louis (2073 mg kg⁻¹).⁵⁰ In F3, the concentration (6.8 mg kg⁻¹) has a similar trend.

Unlike analyses in soil, the scrutiny of inhaled bioaccessible PTE in MR samples is relatively recent. We did not find the regulations of the maximum permitted concentration of the inhalable phase of bioaccessible PTE from MR. Consequently, we made some indirect comparisons, only to highlight the remarkably high PTE concentrations from the MR we observed, which are susceptible to wind dispersion.

In the case of inhalation of the dust from MR, the As accessibility in lungs becomes a concern⁸ because it contributes to the exposure of local inhabitants. Martin et al.⁵¹ emphasized that the test of INBA reliably predicts the bioavailability of As and other PTE of low water solubility. Arsenic has been shown to be soluble in rat lungs.⁵² In the tracheobronchial region and lungs, As causes toxic effects that can result in lung disease or lung cancer,⁵³ due to its toxicity. The INBA-Pb, Cd, and Zn are low, relative to the pseudo-total concentrations. However, this does not guarantee that these concentrations do not present an environmental risk since the average pseudo-totals of As, Pb, Cd, and Zn far surpass the limits proposed by international regulations. Therefore, it is important to establish strict laws to regulate inhalable bioaccessible PTEs.

3.4. Gastric bioaccessible concentrations

Arsenic, Pb and Zn bioaccessibility was related to the pseudo-total concentration; the respective equations were y = 544 + 0.005x (R² = 0.27), y = 668 + 0.215x (R² = 0.31) and y = 0.458x (R² = 0.61). Compared to the inhalable phase, the bioaccessibility in the gastric phase increased due to the low pH in the solution.⁵⁴ Gastric bioaccessible As was lower than 2% of the total, that of Cd was ≤43%, that of Pb was ≤1.6%, and that of Zn was ≤43%. The GABA-As, -Cd, -Pb, and -Zn were higher in F1 (Fig. 1c; Table 4). The average GABA-As concentration in F1 samples was 2.1% relative to the pseudo-total content, 1.2% in F2, and 1.6% in F3. The GABA-As in the F1 was 9.3 times lower than that quantified in MR (<10 μm) in Victoria, Australia (approximately 6900 mg kg⁻¹).⁸ Arsenic in F2 was 3.8 times lower than that determined in the <45 μm fraction of mining residue (1890 mg kg⁻¹) in Nova Scotia, Canada.⁴⁵ In F3, it was 1.5 times lower than that observed in MR (<250 μm) in Kaňk, Czech Republic (573 mg kg⁻¹).¹⁴ The average GABA-As of the three fractions (542 ± 234 mg kg⁻¹) was 1.6 times higher than the average (332.2 mg kg⁻¹) of MR from three tailing heaps reported by Moreno et al.⁹ in bulk slags.

The GABA-Pb was 906 ± 550 mg kg⁻¹ (Fig. 2c), which was two times lower than the concentration (3934 mg kg⁻¹) in bulk samples reported by Moreno et al.⁹ The GABA-Pb in F1 was 1396 mg kg⁻¹, 733 mg kg⁻¹ in F2, and 589 mg kg⁻¹ in F3 (Fig. 2b), representing 42%, 37% and 20% of the pseudo-total concentration, respectively. The GABA-Pb in samples of <20 μm particles was 1.8 times higher than that determined in MR particles <10 μm in Victoria, Australia (763 mg kg⁻¹).⁸ Moreover, in F2 and F3, the GABA-Pb was nine times higher than that quantified in slag dust (<250 μm) from foundries in Zambia (approximately 80 mg kg⁻¹).⁶

The GABA-Cd was 19.2 mg kg⁻¹ in F1, which is 44% of the pseudo-total concentration. In 50–74 μm particles, the bioaccessible Cd was 10.7 mg kg⁻¹ with 28% bioaccessibility. Likewise, in F3, 9.6 mg kg⁻¹ was determined, corresponding to 38% of the pseudo-total concentration (Fig. 3c). Cadmium was 1.6 times lower than that of bulk samples of the mining district of Zimapan (20.9 mg kg⁻¹).⁹ The GABA-Cd in F1 was 16 times higher than that determined in the <63 μm dust fraction from urban parks in Nanjing, China (1.2 mg kg⁻¹).⁵⁵ The gastric-bioaccessible Cd in F2 and F3 was approximately eight times higher than that determined in the dust (<250 μm) from a highway in Newcastle, England (1.2 mg kg⁻¹).⁵⁶

The GABA-Zn in F1 was 2048 mg kg⁻¹; 891 mg kg⁻¹ in F2, and 936 mg kg⁻¹ in F3 (Fig. 4c). These data represent 44% of the pseudo-total concentration for the MR F1 and 32% for the other fractions. The average of the three particle sizes was 1292 ± 1083 mg kg⁻¹, which was 2.4 times lower than the 3130 mg kg⁻¹ of bioaccessible Zn determined in bulk samples from Zimapan, Hidalgo.⁹ The bioaccessible Zn found in <20 μm samples was 6.5 times higher than that determined in the dust (<63 μm) collected in city Parks in Nanjing, China (314 mg kg⁻¹).⁵⁵ The Zn in F2 and F3 samples was 1.8 times higher than the GBA in particles <250 μm in MR in Kaňk, Czech Republic (495 mg kg⁻¹).¹⁴

The GABA-As, -Pb, -Cd, and -Zn were high, relative to their corresponding pseudo-total concentration. For Pb, Cd, and Zn, it was around 40% and 2% for As. The GABA-PTE concentration was higher than the respective inhaled-bioaccessible concentrations because of the acid pH of the extracting solution (pH 1.5). Nowadays, gastric availability is not regulated. However, the GABA-Pb was 3.5 times higher than the limit for total Pb in residential land use soil established by the US EPA (400 mg kg⁻¹),⁴⁸ 5.7 times higher than the total Zn in the ground (360 mg kg⁻¹), and 32 times higher than the total As (12 mg kg⁻¹) in the soil allowed by Canadian regulation.⁴⁷ The amount of PTE released in the gastric fluid is higher than that in the lung fluids, due to the change in pH.¹⁷ According to Ettler et al.,¹⁰ in the case of incidental ingestion of 100 mg of dust particles, the mass of PTE released in the gastric fluid would be 54.1, 90.6, 1.96, and 20.48 μg of As, Pb, Cd, and Zn, respectively. Therefore, people living in adjacent areas to the tailings are at potential risk. Metal(loid) release from PM is the basis for health risk assessment associated with air pollution in the neighborhood of mine tailings. The release concentrations in simulated tracheobronchial and gastric fluids may help to characterize non-carcinogenic and carcinogenic risk using machine learning procedures.⁵⁶ Our results of PM analysis, collected from the atmosphere in a mining village,⁶⁰ are in line with the results of accessibility analysis in slag and dust. The dispersion of tailing by wind^12,14,20 and data about bioaccessibility of PTE^6,10 in the case of inhalation and incidental ingestion^33,37 are relevant evidence of the risk for people living near these waste deposits. This information may help the decision maker plan and design villages for people working in mines and facilities for mineral processing, to reduce the risk of exposure.

3.5. Analysis of As and Pb speciation using XANES

Arsenopyrite (FeAsS) was the predominant chemical species (92.5%) in the residues of the Santa Maria mine in the three fractions (Table S1^†), although other chemical species were observed, such as scorodite (Fe³⁺AsO₄·2H₂O)⁵⁷ and adsorbed As (0.7%). There were no differences in As chemical speciation among fractions. The high percentage of arsenopyrite indicates that the Santa María MR was recently deposited and not oxidized.⁵⁸ Moreno et al.⁹ also identified mainly arsenopyrite in the residues of three mine tailings in the study area. They used polarizing light microscopy, scanning electron microscopy, and electron microprobe. Visually, in our study, we corroborated that the MR was not oxidized and was gray. However, the MR will begin to oxidize and the arsenopyrite will become scorodite.⁵⁹ Then, more mobile and toxic As species will gradually form in the MR.⁵¹ Studies in progress of our research group show that the MR from the Santa Maria tailing is dispersed by the wind as dust, forms part of the PM (average size 2–4 μm), and has more toxic As chemical species, for example, As sulfide, As pentoxide, and a lower quantity of As trioxide.⁶⁰ Walker et al.⁶¹ identified As¹⁻, As³⁺ and As⁵⁺ in proportions of 9%, 32%, and 59%, respectively. Similarly, Paktunc et al.⁶² observed that in the XANES spectra of the MR of the Ketza River in Canada, the predominant species was As⁵⁺. Mineral kind, the stage of oxidation, and the environmental conditions determine the chemical species in MR.⁶³

The species observed in the three particle sizes were Pb carbonate (14%), Pb phosphate (11%), plumbojarosite (21%), and significantly more adsorbed Pb (54%; Table S2^†). Moreover, by particle size, a larger percentage of adsorbed Pb is observed in the F1. Possibly, the highest percentage of adsorbed Pb in the smallest particles is related to the larger specific surface area than that of the other two sizes (50–74 and 74–100 μm). Like in our study, Ostergren et al.⁶⁴ observed that in two fractions of MR (<10 and 100 μm) in the area of Leadville, Colorado, USA, the species that dominated was Pb adsorbed to the surfaces of oxyhydroxides. These authors used X-ray absorption fine structure (XAFS) spectroscopy for speciation analysis; in our study, we used XANES. In contrast, in the present study, the principal Pb carrier mineral was galena (plumbous sulfide) in association with anglesite (lead sulfate), cerussite (lead carbonate), and plumbojarosite.⁹ Hayes et al.⁶⁵ determined that plumbojarosite was the main Pb carrier phase in the Klondyke mine located at the outlet of the Aravaipa Canyon in southeastern Arizona. Arsenic and Pb may be adsorbed on calcium carbonate and iron oxides as was shown by the dissolution of these soil fractions, which are in line with other studies⁶² (Table 4). The speciation of these materials is complex and changes in the mineralogical composition; moreover, the species will change during the dissolution of the tracheobronchial, pulmonary, and gastric fluids.¹⁰ The solubility of these species is relatively low but is affected by the pH of the media and the redox conditions, particularly for reduced species.

4. Conclusions

Mine tailings such as in the Santa Maria mine are highly erodible by wind, due to 33% of the particles having a diameter less than 100 μm, leading to the dispersion of PTE associated with these particles. The dominant winds blow in the direction of the town, increasing the risk of exposure. F3 had a higher average erosion rate (104 mg m⁻² s⁻¹) than the <44 μm fraction (26 mg m⁻² s⁻¹), due to higher magnetic susceptibility in the smallest fraction. Small-particle dispersion is a potential risk for human health due to ingestion and inhalation. The results obtained in the wind tunnel experiment may help track and scale in situ experiments to analyze the behavior of particle dispersion. The XANES spectra and the LFC analyses showed that As occurs mainly as arsenopyrite and a small percentage as scorodite. The predominant Pb species were carbonate, phosphate, plumbojarosite, and adsorbed Pb (54%).

High pseudo-total Pb (3326 mg kg⁻¹) and Zn (4678 mg kg⁻¹) concentrations were found in the <20 μm fraction, while the highest pseudo-total As was observed in the <20 μm (35 961 mg kg⁻¹) and 50–74 μm (40 236 mg kg⁻¹) particles. The highest concentrations of INBA-As, -Pb, -Cd and -Zn were detected in the <20 μm fraction: 343 mg As kg⁻¹, 44 mgPb kg⁻¹, 2.8 mg Cd kg⁻¹ and 9.5 mg Zn kg⁻¹. The highest concentrations of gastric-bioaccessible PTE were found in the <20 μm fraction: 744 mg As kg⁻¹, 1396 mg Pb kg⁻¹, 19.2 mg Cd kg⁻¹ and 2048 mg Zn kg⁻¹. The tracheobronchial- and gastric-bioaccessible PTE should be lower than what we found because they are related to inhalation and ingestion of PTEs. They significantly surpass the reference values in industrial-use soil. Therefore, it is urgent to establish regulations to reduce the risk to human health.

Environmental significance.

After a soil field survey in a mining area and experimentation with a wind tunnel, highly erodible particles, by wind, were identified. The pseudo-total, pulmonary, and gastric bioaccessible concentrations of As, Pb, Cd, and Zn were measured. We also identified the chemical species in the small fractions of mine residues by XANES. The atmospheric dispersion of mine residue particles may have direct implications for human health in the neighborhood due to the reactivity of the chemical species contained in the particles.