Introduction and aims

Acid mine drainage (AMD) is the major pollution problem associated with coal and metallic sulphide mining worldwide. The term “AMD” refers to acidic waters originating from the oxidation of pyrite and other sulphides. In the United States, AMD affects more than 19,000 km of rivers and streams (Kleinmann 1989). The main problems associated with AMD are its acidity and high concentration of trace elements (As, Cd, Cu, Co, Mo, Ni, Pb, Se, Zn, etc.) carried in solution.

The sources of AMD include tailings, mine adits, pit lakes or any material containing sulphides that is prone to oxidation. One of the most important sources of AMD is the waste rock piles, which usually are very heterogeneous in grain size, mineralogy and geochemistry. This makes very difficult to study the generation of AMD from them, since the water geochemistry is affected by all these waste characteristics (Pantelis and Ritchie 1991; Ritchie 1994; Nordstrom and Alpers 1999a; Lefebvre et al. 2001).

Extreme acid waters, with a pH below 1, have occurred in several mines such as Genna Luas (Sardinia, Italy), Iron Duke (Zimbabwe) and San Telmo (Spain) (Frau 2000; Williams and Smith 2000; Sánchez España et al. 2008, respectively). Trace element concentrations were also high in all these waters. For instance, up to 60 mg/L of Cd, 73 mg/L of As and 10,700 mg/L of Zn in Genna Luas, Italy (Frau 2000); 72 mg/L of As in Iron Duke, Zimbawe (Williams and Smith 2000); and 13.76 mg/L of Cd and 1.1 g/L of Zn in San Telmo, Spain (Sánchez España et al. 2008). But all these values were far from those found in Iron Mountain, California (Nordstrom and Alpers 1999b; Nordstrom et al. 2000), with pH values as low as −3.6, As as high as 850 mg/L, Cd up to 370 mg/L, and Zn up to 49 g/L.

In Spain, the contamination associated with AMD has been documented at several sites such as Galicia or Asturias (e.g. Monterroso and Macías 1998a; Loredo et al. 2005), but it largely occurs in the Iberian Pyrite Belt (IPB), which is one of the most important metallogenic provinces in the world. Most of the mines in this area have been abandoned and are discharging huge quantities of acidic waters and heavy metals into the Tinto and Odiel Rivers (González et al. 2004; Sánchez España et al. 2005; Cánovas et al. 2007). As a result, these rivers are among the most polluted in the world and have gained the attention of many researchers and studies (Fernández-Caliani et al. 1997; van Geen et al. 1997; Elbaz-Poulichet and Dupuy 1999; Hudson-Edwards et al. 1999; Braungardt et al. 2003; Galán et al. 2003). However, there is not enough information about the contribution that the different abandoned mine spoils have on those rivers. This is reasonable if we consider that the headwaters of both rivers receive leachates from huge quantities of mine spoils of the Riotinto mining district. The volume of these wastes has not been assessed, but it is known that at least 6.6 million tons of slags are from the Roman and pre-Roma period (Rothenberg et al. 1990). The intense exploitation of the Riontinto Mines, especially during the nineteenth century, has caused about 1,800 ha of land to be covered with dumps, tailings and open pits, making it difficult to characterize and remediate.

Here, we have investigated the geochemistry of stream waters in the Peña de Hierro abandoned mine site, located in the catchment area of the Tinto River and being the first source of pollution to this river (Fig. 1). The deposit consists of a copper-bearing pyrite ore with variable concentrations of copper, usually comprised between 1 and 4 wt%, and lower concentrations of Zn and Pb (Pinedo Vara 1963; IGME 1982; Tornos 2008). The mine was working between the middle of the nineteenth century and 1966, and it produced about 3 million tons of waste rock and other residues that were dumped around the pit, covering 25 hectares. The characterization of the wastes and the mapping of the dumps have been carried out by Romero et al. (2006a). There are several streams that arise from these waste dumps, and thus, the geochemistry of the waters can be directly linked to the different waste rocks. The aim of this work was to correlate the trace element concentration in stream waters with the different mine-waste dumps, in order to determine the source and factors affecting the genesis of AMD. To achieve these objectives, we studied the geochemistry of the stream waters and the results were compared with the chemical and mineralogical characteristics of the waste dumps.

Fig. 1
figure 1

Location of the Peña de Hierro mine in the Iberian Pyrite Belt (SW Spain)

Full size image

Materials and methods

Waste dumps

Waste rock samples were collected from boreholes drilled into the dumps by the company RIOTINTO S.A.L., at depths between 1.5 and 14.0 m. A representative sample was collected from each borehole, up to a total of 58 samples. Since the dumps consist of different wastes, the mineralogy and chemistry were analysed in the <2 mm fraction. The mineralogical characterization was carried out by X-ray diffraction (XRD) using a Bruker D8 Advance equipment with slit fixed at 12 mm and monochromatic Cu Kα radiation. Random powders were scanned at 40 kV and 30 mA from 3° to 70° 2θ at a speed of 0.3° 2θ/min.

Chemical analyses of major, minor and trace elements were made in the <2 mm fraction of selected samples. Nineteen elements were analysed at Activation Laboratories Ltd (1428 Sandhill Drive; Ancaster, Ontario, Canada) from 3 g of sample ground in agate mortar and sieved at 50 μ. Analysis methods included Instrumental Neutron Activation Analysis (INAA), Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES) and Infra-Red Spectroscopy (IR). The detection limits were 1 mg/kg for As, Co, Cu, Hg, Mn, Mo, Ni and Zn, 0.3 mg/kg for Cd, 2 mg/kg for Cr, 3 mg/kg for Pb, 50 mg/kg for Ba and 0.01% for major elements and S. The results were contrasted against certified materials such as DMMAS-14, G-2, SDC-1, DNC-1, SCO-1, GXR-1, GXR-2, GXR-4 and GXR-6. Errors were always below 10%, except for Cd, Al and Mg, which could be higher.

Stream water

A total of 28 water samples were collected in PVC bottles from the streams arising from the waste dumps during two field trips (Fig. 2): the first one was carried out in winter (March 2002, 12 samples), 2 weeks after a period of continuous rain. The second was carried out in summer (July 2002, 16 samples), after more than 2 months of high temperatures and no precipitation. Water samples were collected from the headwaters of each stream next to the waste dumps, and several metres downstream from the confluence of different watercourses. After collection and filtration to 0.45 μ, samples were acidified below pH 2 when needed and kept at 4°C until chemically analysed.

Fig. 2
figure 2

Distribution of mine-waste dumps in Peña de Hierro and location of streams and sampling stations. Blue circles denote sampling in winter, and red triangles are for summer. Modified after Romero et al. 2006a

Full size image

The pH was measured in the field using a CRISON 507 portable pH meter equipped with a 52-00 electrode and automatic temperature compensation. Conductivity was measured using a CRISON 524 portable conductimeter equipped with a 52-90 cell and automatic temperature compensation. Temperature and Eh were measured (in July 2002) using a CRISON 52-61 platinum electrode coupled with the above-mentioned pH meter. All instruments were calibrated in the field before measurements were taken.

In the laboratory, the sulphate concentration was determined by turbidimetry on a Zuzi 4200/50 UV–vis spectrophotometer. Samples collected from the headwaters and some from the confluence of watercourses were selected to analyse the concentration of major, minor and trace elements (21 samples). All analyses were carried out at Activation Laboratories (http://www.actlabs.com) by Inductively Coupled Plasma Mass Spectroscopy (ICP-MS). The detection limits for the analysed elements were 2 μg/L for Al, 50 μg/L for Ca and Si, 5 μg/L for Fe and Na, 10 μg/L for K, 1 μg/L for Mg, 0.03 μg/L for As, 0.01 μg/L for Cd, 0.005 μg/L for Co, 0.5 μg/L for Cr and Zn, 0.2 μg/L for Cu and 0.1 μg/L for Mn, Mo and Pb. The accuracy and precision of the analytical method was verified against the certified reference waters SLRS-4 for major elements and NIST 1640 for trace elements. Errors were always below 5% except for As and Pb (7.5%) and Ni (10%).

Results and discussion

Waste dumps

Waste dumps consist of a mixture of different materials such as volcanic rocks, gossan, shales, and mining residues such as roasted pyrite ashes or floated pyrite. A detailed description of the wastes is presented by Romero et al. (2006a). According to which of these materials predominates in the dumps, a mapping of the waste has been carried out, with the following being the most important units (Fig. 2, Romero et al. 2006a):

  1. 1.

    Tuff dumps are mainly made up of acid volcanic tuffs, but usually gossan and shales are mixed in the dumps. The mineralogy is mainly composed of quartz, muscovite, pyrite and jarosite (Table 1). Pyrite is common in blocks of volcanic rocks, and jarosite precipitates in the <2 mm fraction after the oxidation of pyrite. When gossan fragments are present, hematite and amorphous Fe oxy-hydroxides occur in the fine fraction of the dumps. The concentration of trace elements in the <2 mm fraction of this unit is low when compared with other residues. For example, As is below 263 mg/kg and Pb is below 628 mg/kg (Table 2). By contrast, the concentration of Co (64 mg/kg) was higher than that found in other materials.

    Table 1 Mineralogy of fraction <2 mm in different types of mine-waste dumps (in %)
    Full size table
    Table 2 Chemical analysis of <2 mm fraction of borehole samples (Romero et al. 2006a)
    Full size table
  2. 2.

    Gossan dumps may also contain remains of tuffs and shales. They have a high content of hematite, goethite and X-ray amorphous iron oxy-hydroxides. The <2 mm fraction of these dumps displays high concentration in several trace elements such as As (538–1,710 mg/kg), Mo (143–331 mg/kg), Pb (318–989 mg/kg) or Cu (90–895 mg/kg). These concentrations can be higher in rock fragments of gossan. For example, Romero et al. (2006a) determined up to 2,330 mg/kg of As, 4,560 mg/kg of Pb, and 749 mg/kg of Mo.

  3. 3.

    Tuff–gossan dumps consist of a mixture of tuff, gossan and shales. Their mineralogical and geochemical characteristics are intermediate between the two units mentioned above.

  4. 4.

    Roasted pyrite ash dumps are characterized by a high content of hematite and X-ray amorphous iron oxides. They are also rich in trace elements such as Pb (2,000–3,460 mg/kg), As (303–971 mg/kg), Mo (81–351 mg/kg), Cu (419–527 mg/kg), Cd (2–3 mg/kg) and Zn (283–536 mg/kg).

Other dumps such as floated pyrite and mix dumps are rich in fine pyrite and may have high concentration of trace elements, but they are less important because their volume is relatively low (Romero et al. 2006a, Fig. 2).

Stream waters

Physical–chemical features

The pH was very acidic in most of the streams, ranging between 0.7 (site C, July) and 3.5 (site Z1, March) (Table 3). Streams Y, B, C and D arising from the tuff dumps displayed the lowest pH values (always <2). Site G, located outside the mining area, showed neutral natural water. Stream Z was not affected by leachates from the waste dumps, but pH ranged between 3.1 and 3.5. The pH values recorded in summer were lower than those of winter in most of the streams.

Table 3 Values of pH, conductivity, Eh, temperature and sulphate contents of water samples
Full size table

Conductivity ranged between 1,000 and 47,000 μS/cm in the acidic streams, being of 130 μS/cm in the neutral stream G. The most acidic streams (Y, B, C, D) recorded the highest conductivity values, which were over or about 20,000 μS/cm. The values recorded in the summer field trip were higher than those of winter in most of the streams. The sulphate concentration, which correlates with the conductivity, ranged between 363 and 33,202 mg/L in the acidic samples, and it was about 6–7 mg/L in the neutral stream.

The Eh in the acidic streams ranged between 333 (site Z0) and 563 mV (site Y), and it was of 202 mV for the neutral water. The Eh increased as water moved downstream.

According to the temperature values, stream Z displayed values of about 24°C, streams Y, A and B were about 28–30°C and the other streams ranged between 20 and 22°C. The temperature values recorded at the confluence of different streams, as well as pH, conductivity and sulphate concentration, showed intermediate values with upstream waters.

Stream water geochemistry

The chemical analyses of major and trace elements are shown in Table 4. Among the major elements, Fe displayed the highest concentration, up to 30 g/L at stream C, and many times it was over 10 g/L, especially in streams arising from the tuff dumps. Aluminium and Mg reached 3,730 and 1,680 mg/L, respectively. Calcium ranged between 21 and 470 mg/L. Sodium and Si were lower than 100 mg/L. Potassium displayed the lowest concentration, being in most of the samples below 3 mg/L.

Table 4 Chemical analyses of water samples
Full size table

Among the minor and trace elements, Mn, Cu and Zn displayed values over 100 mg/L in several samples (Table 4). Arsenic and Mo reached up to 72 and 11.3 mg/L, respectively (site C, July). Cadmium was usually over 0.1 mg/L and reached up to 0.85 mg/L in site F (July). Cobalt usually displayed a few mg/L, and it was over 10 mg/L in the most acidic streams. Among the most immobile elements was Pb, which was always below 0.5 mg/L.

Figure 3 shows the relative concentration of elements in the headwaters of every stream, corresponding to summer sampling. The element concentrations of every sample were normalized with the maximum value found for each element. Thus, a value of 1 for a particular element points out the sample with the maximum concentration for this element. So, Fig. 3 shows that stream C had the lowest pH and the maximum concentration in sulphate, Fe, As, Co, Cu and Mo. High concentrations of these elements were also found in other streams arising from the tuff dumps (B, D, Y), and they correlated negatively (r < −0.7) with pH (Table 5). The maximum concentration of K was recorded in stream C, but the concentrations in the other streams were very low.

Fig. 3
figure 3

Conductivity, pH, and concentrations of different elements normalized with the maximum respective value in streams sampled during July 2002 field trip

Full size image
Table 5 Statistical correlations for stream water chemistry
Full size table

Other elements such as Ca and Mn showed low values in the most acidic streams, and the highest values occurred in streams Z, A and F. These streams flow over the bedrock constituted of shales that are rich in Ca and Mn (Fernández-Caliani and Galán 1991), suggesting that these elements can be leached from them into the stream water.

Cadmium and Zn displayed high values in stream F where pH was also high, and in streams B, C and D. The concentrations of Pb were very low in most of the streams, and the highest values were found in streams C and F.

These results show that the streams in Peña de Hierro are among the most polluted in the world (Table 6). For example, the minimum pH and maximum concentration of As (72 mg/L) were similar to those found in other mines such as Genna Luas, Italy, or Iron Duke, Zimbabwe (Frau 2000; Williams and Smith 2000). The concentration of Mo recorded in Peña de Hierro (up to 11.3 mg/L) was extremely high. This value is even higher than the maximum found in Iron Mountain (4.2 mg/L, Nordstrom and Alpers 1999b). A similar value of Mo (10.41 mg/L) has been measured in San Telmo, another mine of the IPB (Sánchez España et al. 2008).

Table 6 Comparison between some of the most polluted waters in the world with the results presented in this study (data in mg/L, except pH)
Full size table

Source of AMD and trace elements

One possible source of acidic waters in Peña de Hierro is the natural acid rock drainage produced by the bedrock, as other authors have suggested (González-Toril et al. 2003; Fernández-Remolar et al. 2004). In this sense, stream Z, which is located in the eastern part of the mine, does not seem to be affected by mine waters (Fig. 2), and its geochemistry is highly affected by elements leached from the bedrock, such as Ca, Na and Mn. Therefore, these waters can represent the natural acid rock drainage in the headwaters of the Tinto River. In any case, and in comparison with the geochemistry of streams arising from these waste dumps, the low content of trace elements in stream Z shows that the mismanagement of mine wastes and their subsequent abandonment have a very negative impact on the environment, since it causes the load of acid and trace elements in stream water to rise by several orders of magnitude.

The high acidity of the streams arising from the tuff dumps (Y, B, C, D) shows that these wastes, which contain pyrite and have no neutralizing minerals, are a very important source of AMD in Peña de Hierro. Although there are other pyrite-containing wastes that may generate AMD, the coarse grain size of the tuff dumps (Romero et al. 2006a) allows the air to enter easily into the dumps, causing the oxidation of pyrite through the following reaction 1 (Ritchie 1994; Lefebvre et al. 2001; Nordstrom and Alpers 1999a):

$$ {\text{FeS}}_{{ 2({\text{s}})}} + 7/ 2 {\text{O}}_{ 2} + {\text{H}}_{ 2} {\text{O}} \to {\text{Fe}}^{ 2+ } + 2 {\text{SO}}_{4}^{2 - } + 2 {\text{H}}^{ + } $$
(1)

The tuff dumps are also an important source of trace elements to streams Y, B, C and D (Fig. 3). Some elements such as Co seem to come from pyrite oxidation, because this element is mainly present in the tuff dumps (Table 2), probably associated to pyrite. By contrast, other elements such as Fe, As and Mo displayed high concentration in gossan wastes, and their occurrence in the water may be influenced by the dissolution of this material. In fact, the Fe/SO 24 ratio is up to more than three times greater than that expected from pyrite dissolution alone (Fig. 4), which suggests that the dissolution of gossan is being accomplished. This process can occur easily if we consider that the gossan is mixed in the tuff dumps and acidic waters with pH < 2 can partially dissolve the Fe oxy-hydroxides and release Fe3+ through the following reaction 2 (Nordstrom and Alpers 1999a). Although we have not measured the Fe2+/Fe3+ ratio, this hypothesis is consistent with the wide occurrence of Fe3+ rich efflorescent sulphates (coquimbite, rhomboclase, magnesiocopiapite) found in the banks of the streams (Romero et al. 2006b). The releasing of Fe3+ may also contribute to lower the pH to values down to 0.7, because it can react with pyrite and oxidize it rapidly through the following reaction 3 (Singer and Stumm 1970; Nordstrom and Alpers 1999a).

$$ {\text{Fe(OH)}}_{{ 3({\text{s}})}} + 3 {\text{H}}^{ + } \to {\text{Fe}}^{ 3+ } + 3 {\text{H}}_{ 2} {\text{O}} $$
(2)
$$ {\text{FeS}}_{{ 2({\text{s}})}} + 1 4 {\text{Fe}}^{ 3+ } + 8 {\text{H}}_{ 2} {\text{O}} \to 1 5 {\text{Fe}}^{ 2+ } + 2 {\text{SO}}_{4}^{2 - } + 1 6 {\text{H}}^{ + } $$
(3)

The dissolution of gossan does not only release iron but also trace elements, particularly As and Mo, which occur widely in these wastes. This explains their very high concentration in the stream water (Table 4). The strong correlation Fe–As–Mo in stream water samples (r ≥ 0.86, Table 5) and wastes with variable concentrations of gossan (r ≥ 0.84, Romero et al. 2006a) support this idea. These results have an important environmental impact because the gossan is stable in an oxidant environment, and a low mobility of trace elements is expected under these conditions. By contrast, when gossan wastes are mixed with high acid potential materials, the acidic leachates can dissolve them and release huge contents of trace elements into the water.

Fig. 4
figure 4

Concentration of Fe versus SO4 2− in water samples. The dotted line shows the Fe/SO4 2− ratio expected from pyrite dissolution

Full size image

Other elements as K, Na and Pb were not correlated with pH, but they displayed high concentrations in stream C (Fig. 3b, d). The low values displayed by K in most of the water samples probably indicate that it is largely retained by the precipitation of secondary jarosite. When the physical–chemical conditions are unfavourable for the precipitation of jarosite (e.g. pH < 1, Dutrizac and Jambor 2000) K is leached into the water. Similarly, other authors have also measured high concentrations of K in waters with pH < 1 in other areas (e.g. Williams and Smith 2000; Frau 2000). In the same way, the release of Na and Pb into stream C suggests that they can be also retained by jarosite when this phase precipitates. Indeed, Na and Pb interexchange with K in the jarosite structure as other authors have previously reported (Dutrizac and Jambor 2000; Acero et al. 2006; Graupner et al. 2007).

Besides the tuff dumps and gossan wastes, the mixed dumps, which are rich in pyrite and the roasted pyrite ashes can be important sources of Cd, Zn and Pb to the water. For example, stream F, which receives leachates from both materials, showed high concentrations of Cd, Zn and Pb (Table 2) despite the fact that the pH was about 2–3, which is higher than the values recorded in other streams. With respect to Cd and Zn, there is a strong correlation between them (r = 0.89), but not with the pH (Table 5). These elements can be easily released into the water due to their high mobility and low pH dependence in very acidic environments (Nordstrom and Alpers 1999a; Plumlee et al. 1999). By contrast, Pb is a very immobile element, and its high concentration in stream F seems to be related to the low Eh values (375 mV), which are unfavourable for the precipitation of secondary phases such as jarosite (Bigham et al. 1996) or any other oxidized phase.

Element mobility

According to the above-discussed considerations, it is plain to see that the release of major and trace elements into the water depends on several factors such as their concentration in the wastes, the pH–Eh environment or the dissolution/precipitation of mineral phases. So, depending on these factors, the mobility of an element may change in different environments.

Considering the waste dumps as the source of elements and the stream water as the environment into which they are released, it is possible to asses their mobility by determining the ratio between their concentration in water versus wastes. So we have calculated the average concentration of major and trace elements in water and divided it by their average concentration in wastes (Table 7). The mobility of the major elements shows the following sequence:

$$ {\text{Mg}},{\text{Ca}},{\text{S}} > {\text{Al}},{\text{Fe}},{\text{Na}} > {\text{K}} $$

The minor and trace elements show the following mobility sequence:

$$ {\text{Mn}} > {\text{Co}},{\text{Zn}},{\text{Cd}} > {\text{Cu}},{\text{Ni}} >> {\text{As}},{\text{Mo}},{\text{Cr}} >> {\text{Pb}} $$

Thus, Na and K are among the most immobile elements, because after their release from primary minerals, they can be largely retained by the precipitation of jarosite. With regard to trace elements, Cd and Zn are highly mobile, while As and Mo are only moderately mobile and Pb is relatively immobile. Similar results were obtained in other AMD-affected areas by McGregor et al. (1998) and Monterroso and Macías (1998b).

Table 7 Relative mobility of some elements determined by means of the ratio “mean value in water”/“mean value in mine-waste dump” (B/A)
Full size table

Conclusions

Mining operations in Peña de Hierro have produced extremely acidic waters containing high concentrations of trace elements such as As, Cd, Cu, Mo, Pb, and Zn. Dumps composed of volcanic tuffs are the main source of AMD, and the partial dissolution of gossan mixed in these dumps is an important source of As and Mo. Roasted pyrite ashes and pyrite wastes are sources of Cd, Pb and Zn, and the bedrock leaches Ca, Na and Mn. Several elements such as Fe, As, Co and Mo are greatly affected by the pH, but others as Pb and Cd–Zn are conditioned by their low or high mobility, respectively.

This study has shown that the geochemistry of stream waters strongly depends on the characteristics of the materials drained. On one hand, the dumping of wastes increased the trace element concentration in streams compared with natural waters, both neutral and acid. On the other hand, the mismanagement of waste rocks when mixing pyrite-rich spoils with others rich in trace elements (e.g. gossan) involved their release through dissolution of iron oxy-hydroxides, increasing the contamination of stream waters.

A selective management of these different mine spoils could be an easy way to control and mitigate the pollution, as it would avoid the leaching of potentially toxic wastes (e.g. gossan) practically without any corrective action, and it would reduce the volume of acid generating wastes. These results can be also applied to the management of waste dumps in new or reopened mines of the IPB.

Characterization studies of waste rocks are necessary for planning accurate remediation actions. Classification/selection of different wastes according to their acid/basic generating potential, concentration in potentially toxic elements and factors affecting their mobility or their harmlessness are necessary to control them. According to the waste potential hazard, specific management criteria can be established, such as insulation, covering, mixing, supervision and temporary control, abandonment, revegetating, etc.