Passive treatment of tailings seepage waters in former settling pools – Enonkoski case study

Marja Liisa Räisänen, Geological Survey of Finland, P.O.Box 1237, FI-70211 Kuopio, Finland marja-liisa.raisanen(at)gtk.fi

Introduction of the mine site

The Enonkoski Ni-Cu mine is located in the Laukunkangas area 14 km southwest of the Enonkoski municipality centre and about 30 km north of the city of Savonlinna in eastern Finland. The mine is also named the Laukunkangas mine, after the place-name. The mine was opened in 1984 and shut down in 1994. At present, the mine site includes a tailings facility with former water treatment pools and building area. The mill and concentrator was demolished when the mine was closed. The repair workshop and building area are hired by a private company and in use as a bus repair shop.

Geology, ore mineralogy and chemistry

The Laukunkangas Ni-Cu deposit is located in a Svecocarelian synorogenic plutonite. Supracrustal gneisses and various intrusive rocks form the geological setting of the region. The Laukunkangas intrusion consists of a differentiation series of an olivine tholeiitic suite ranging from peridotites to quartz diorites. It is crosscut by numerous silicic pegmatite veins and mafic diabase dykes. The intrusion is embedded in a metasedimentary mica gneiss and veined gneiss environment. The wall-rock mica gneiss is magmatic and brecciated along the contact zone of the ore deposit. Numerous graphite gneiss, black schist, and graphite fragments within the intrusion are very heterogeneous, strongly folded and brecciated. (Grundström 1985 and 1986)

Sulphides have accumulated in the ultramafic rocks at eastern end of the Laukunkangas intrusion. The Ni-Cu deposit was subdivided into four independent orebodies: 1) the main orebody, high-grade ore in olivine norite and peridotite hostrocks, 2) the slope orebody, low-grade ore at the northern contact, mainly in norite, 3) the deep orebody, low-grade ore at the southern contact and a weak lens-shaped dissemination in the middle of the intrusion, and 4) the Leo orebody, an offset ore type in mica gneiss, trondhjemite and black schist north of the intrusion (Grundström 1986).

The main ore minerals of all ore types were pyrrhotite (FeS), pentlandite ((FeNi)₉S₈) and chalcopyrite (CuFeS₂) (Grundström 1985). Accessory minerals were millerite (NiS), violarite (FeNi₂S₄), gersdorffite (NiAsS), pyrite (FeS₂), magnetite (FeFe₂O₄), ilmenite (FeTiO₃), rutile and anatase (TiO₂), chromite (FeCr₂O₄), graphite (C), and sporadically sphalerite [(Zn, Fe)S] and molybdenite (MoS₂) (Grundström 1985, Alopaeus et al. 1986, Pöyry & Isomäki 1996). Veins of nickel arsenides (nickeline, NiAs) occured in the contact zone and in the ore associated with ultramafic rocks (Grundström 1985).

The ore extracted (about 6.7 Mt) consisted of 0.75% Ni and 0.22% Cu (Pöyry & Isomäki 1996). The nickel values for 100% sulphides ranged from 1.9% (off set ore) to 4.5% (main ore) (Grundström 1985). The total ore resources of 9.4 Mt had roughly about 0.67% Ni, 0.18% Cu, 0.03% Co and 5.2% S (Pöyry & Isomäki 1996).

Mining and milling processes

The Laukunkangas ore was extracted from an underground mine. The main mining method was sublevel stoping with the waste rock and partly consolidated backfilling. A total of 8.4 Mt of rocks and 6.7 Mt of ores were mined at Enonkoski mine during 1985-1994. At the beginning, the annual production of ore was 0.7 Mt, which increased to 1.0 Mt from 1992-1994. Besides the Laukunkangas ore, the two small satellite ores, Hälvälä (1.4% Ni) and Telkkälä (1.4% Ni), were processed at the mill in 1988-1992. The total mill feed was 7.4 Mt of ore with average Ni content of 0.82% and Cu content of 0.23%. Altogether 0.524 Mt of Ni-Cu concentrates with an average Ni content of 9.2% was produced from 1986-1994. (Pöyry & Isomäki 1996)

Ore refining at the mill started in 1986 after a short test run in November 1985. The ore was crushed underground, and skiphoisted to the surface. The crushed ore was ground in two stages, first in one primary lump mill and then in one secondary pebble mill. Later, when the hard norite ore was in production, grinding was conducted in two separate lines, each with a one stage ball mill. (Pöyry & Isomäki 1996)

The refining mainly included the frothing of Ni-Cu bulk concentrates. In the first year of activity, the flotation consisted of three steps, first the frothing of bulk Ni-Cu concentrates, followed by the separation of Ni and Cu concentrates, and finally the separation of Ni bearing pyrrhotite concentrates. The chemicals used in the flotation were mainly Na isobutyl xanthate, pine oil (Oulu 412) and lime. In the first year of activity, also dextrin, sulphuric acid, CMC (carboxymethyl celluloce), and Cu sulphate were used. (Alopaeus et al. 1986, Pöyry & Isomäki 1996)

Tailings facility and characterization

The main mine waste facility on the ground is the tailings impoundment, which covers 44 hectares. About 6.6 million tons of tailings were produced at the mill. Waste rocks were backfilled in the underground mine.

The tailings consist of silicates (mainly olivine, pyroxene, amphibole, chlorite, and feldspar), metal oxides (magnetite, ilmenite, rutile, and chromite), graphite and minor sulphides (pyrrhotite, pentlandite, chalcopyrite, and pyrite). The mineralogy is assessed by the mineralogy of the ore types (see Grundström 1985).

According to unpublished drill sample data from the Outokumpu Company, the tailings consist of 9.2% S, 19.8% Fe, 0.23% Ni, 0.06% Cu, 0.03% Zn, 0.01% Co, 0.08% Cr, 2.2 ppm Cd, 20 ppm As, 18 ppm Sb and 48 ppm Pb (Saari et al. 1993). The oxide concentrations of main elemetns are 39.2% SiO₂, 3.4% CaO, 7.0% MgO, 1.30% Na and 0.48% K. The concentration of C (graphite) is about 0.27%. Based on static tests, the tailings are classified as acid generating mine waste. (Saari et al. 1993)

The tailings facility is located in a valley surrounded my moraine hills in the northern part of the basin. The tailings are underlain by glacial till. Presumably, the till material is water permeable and dominated by sand. In the north, the staged conventional dam was constructed with sandy till and covered by waste rock boulders. This main dam separates the impoundment from the first settling pool (Fig. 1). In the south and east, the tailings extend on a bank constructed with sandy till, and in the west, on banks of the road. The banks were originally not categorized as dams. However, tailings pore water will seep through them if the water table in the tailings rises due to precipitation (see Fig.1).

Water management of the tailings facility

During operations process water was pumped into the tailings impoundment, from which the water was pumped via pipe into the first settling pool after clarification. In the pool, water was neutralized with lime before discharging into the second settling pool. The pools are located within a bog depression. There are no dams except the road bank that separates the first pool from the second (Fig. 1). Waters from the first pool flow into the second through a pipe in the dam structure of a road. The overflow from the second pool was drained into the Sortavalanjärvi lake, which discharges to the Haukivesi lake. (Saari et al. 1993)

State of the tailings facility closure

In the closure stage, the water table in the facility was dropped and the tailings were covered by till. The till cover is about 50 cm thick. Pine seedlings were planted in the western part of the impoundment and grass seeds were sown in central, northern and eastern impoundment areas (Fig. 1).

The main gradient of the seepage in the tailigns is toward the north, into the first settling pool (Wetland1). In the south, seeping gradients are toward the Sakastinjoki river, which discharges into Polvijärvi Lake and downstream to the Hiesunjoki River, which eventually discharges into Haukivesi Lake. In west, seepage waters flow via ditches into Särkijärvi lake that discharges to the Haukivesi lake.

The settling pools have naturally evolved into flooded mires, i.e. wetland pools growing mainly cattails, horsetails and reeds (Myllymäki 2006). Leafy trees (birch) and bushes (willow) are growing on the driest humps in the wetlands and the slope fringing moraine hummocks. Open water is covering almost half of Wetland1 in the area bordering the road bank. In Wetland 2, open water is only in the canal excavated through the pool (See Fig. 1). At present, the water level rises and decreases according to annual precipitation and the water volume seeped from the tailings. Unfortunately, there are no data on water table level variation in the tailings, nor on annual seepage water quantities.

Figure 1. The air photo of the closed Laukunkangas mine area, Enonkoski. Light blue arrows show the water flow direction into, and out of the settling pools, and white arrows show the direction of seeping from the tailings facility. The groundwater gradient generally follows surface water flow directions. Airphoto © National Land Survey of Finland.

Enonkoski wetland case study

The Enonkoski case study introduces passive treatment technology for tailings seepage waters. The waters are treated passively in the former settling pools (Fig. 1). The two-segment pools were constructed on a bog. The segments are separated by an embankment of the road that leads to the former mill. Waters flow from the first pool (Wetland1) into the second pool (Wetland2) through a pipe through the base structure of the road. Moraine hummocks surround the pools and naturally embank surface waters in the depression (Saari et al. 1993). Since mine closure, the water level in the pools has fluctuated according to annual precipitation and seepage from the tailings facility. This has promoted spreading of aquatic vegetation, dominatingly cattails (Typha latifolia) and reeds (Phragmites australis), but also water horsetails (Equisetum fluviatile), sedges (Carex acuta) and rush (Juncaceae filiformis) (Myllymäki 2006). Areas growing aquatic plants are rimmed with willow bushes (Salix phylicifolia) and birch trees (Betula Pubescens, Fig. 2). It is assumed that the flooded wetland rich in aquatic plants promotes biogeochemical water purification. Wetlands, especially those that include cattails, are known to maintain pH-redox conditions for biologic sulphate reduction and accordingly promote the precipitation of Fe and other metal sulphides (Sencindiver & Bhumbla 1988, Sengupta 1993).

This case study aims to examine treatment efficiency of the tailings seepage waters in the former settling pools, i.e. wetlands. The quantitative estimates of metal loads and salinity are not included due to lack of monitoring data on annual water flow (volumes) and rainfall. The efficiency of the treatment is estimated by comparing the chemical content of pool and outflow waters with that of seepage waters. Data used in the present study was collected from pool sediments in 2002, and from seepage and pool water in 2002 and 2005 by the Geological Survey of Finland (GTK). The water data was earlier published as part of the M.Sc. thesis by Tiina Myllymäki (2006), but the GTK sediment data are unpublished.

Figure 2. (a) The first (Wetland1) and (b) second wetland pools (Wetland2), of the closed Enonkoski mine area, eastern Finland. Photos © T. Myllymäki.

Materials and methods

Surface waters

Water data was collected in 2002 and 2005 by GTK. Altogether 36 water samples were taken from seepage waters (15 samples), Wetland1 (11 samples), Wetland2 (5 samples), and the recipient ditch (5 samples, See Fig. 3). In 2002, the waters were sampled at the beginning of August, and four times (in June, July, August and October) in 2005 (Myllymäki 2006). In the both years, the sample was first collected in a 0.5 l-volume polyethylene bottle, from which a 100 ml-volume subsample was filtered into a 100 ml-volume bottle. The filtered subsample was preserved by adding 0.5 ml Suprapur® nitric acid in the field. Soluble concentrations of 33 elements were determined with ICP-AES and MS-ICP methods (Myllymäki 2006).

Temperature, pH, redox, electrical conductivity, oxygen concentration and oxygen saturation of the surface waters were determined with WTW field meters in the field. The pH was measured with a WTW pH340i/Set meter, the electrical conductivity (EC) with a Cond340i/Set meter, and oxygen concentration and saturation with an Oxi330/Set meter. The remainder of the 0.5 l-samples were used for the measurements of pH, EC, oxygen concentration and saturation with the mentioned meters in the laboratory the day after sampling. (Myllymäki 2006)

Figure 3. Location of water sampling sites, the closed Enonkoski mine area, eastern Finland. Airphoto © National Land Survey of Finland.

Wetland sediments

Peat sediments were taken with a peat corer in March 2002 (14.3.2002) when the settling pools were covered by ice. Altogether 16 peat samples were taken at five sampling sites (Fig. 4). In Wetland1, subsamples of the topmost sediments at three sampling sites (Wet1_P1, Wet1_P2, Wet1_P3) were taken first from the sediment surface to a depth of 0.3 m, and then from 0.3 m to 0.5 m or 0.6 m. In Wetland2, the topmost peat sediment was only sampled from the sediment surface to a depth of 0.3 m. In both wetlands, the topmost peat sediments consisted of gmixed with secondary inorganic precipitates and weakly decomposed peat. The lower peat sediments were sampled from 0.5 m or 0.6 m to 1 m, and from 1 m to 1.5 m, in the both wetlands. In the first wetland, the lower peat was sampled at two sites (the Wet1_2, Wet1_3) and in the second wetland at one site (Wet2_2, Fig. 4). The lower samples consisted of decomposed peat.

Peat sediment samples were freeze-dried before chemical analyses. Concentrations of acid extractable elements were determined with concentrated nitric acid leach method using ICP-AES technique (US EPA 3051, See Niskavaara 1995). Nitric acid entirely decomposes organic matter, sulphides and salt minerals, and selectively, micas, clay minerals from silicates (Doležal et al. 1968). A 1-M NH₄ acetate extraction buffered at pH 4.5, and a 0.01-M BaCl₂ extraction were used to determine elemental contents of the chemical and physical adsorption fractions. The solid-solution ratio for the acetate extraction was 1:60 (Kumpulainen et al. 2007) and 1:10 for the BaCl₂ extraction (ISO 11260, See Schultz at al. 2004). The acetate extraction is assumed to leach selectively non-crystalline secondary precipitates and elements chemically adsorbed on solid particles, whereas the dilute BaCl₂ solution selectively extracts easily leachable and exchangeable elements (Schultz et al. 2004, Kumpulainen et al. 2007, Heikkinen & Räisänen 2009).

The element distribution in three geochemical fractions (tightly bound, chemical and physical adsorption fraction) was based on the above non-sequential extraction scheme, in which separate subsamples were leached with extractive solutions of increasing effectiveness, assuming that the stronger solutions dissolve also the phases leached with the weaker solutions. Consequently, the amount of each fraction is calculated by subtracting the concentration of the previous (weaker) extraction from the next step (See also Heikkinen & Räisänen 2009).

Figure 4. Location of peat sampling sites in Wetland1 (Wet1_P1, Wet1_P2, Wet1_P3) and Wetland2 (Wet2_P1, Wet2_P2) of the closed Enonkoski mine area, eastern Finland (See also Fig.1). Airphoto © National Land Survey of Finland.

Results and discussion

Physical and chemical properties of surface waters

The mean pH values of the seepage waters measured in the field varied between 5.7 and 6.6, and did not change much when measured after one day in the laboratory (Table 1). The pH of waters in Wetland1 rose slightly from 7.0 at sites close to the seepage area, to 7.4 close to the discharge area. The pH of waters in Wetland2 (7.4) and further downstream in the receiving ditch (7.2) showed little spatial variability. Furthermore, the pH values measured in the laboratory were pretty close to the pH values measured in the field. This suggests that water was constantly neutralized from the pools to the recipient water way.

Overall, seepage waters showed weak oxidation properties. Slightly acidic seepage waters at the Seepage1 site were somewhat more oxidative than seepage waters at the Seepage2 site, with redox means of 260 mV and 74 mV, respectively (Table 1, See also Fig. 4). In the wetland pools, redox values varied between 120 and 160 mV, showing weak oxidative character. The outflow from the second wetland pool into the receiving ditch became slightly reducing. These findings confirm reducing conditions within the wetlands.

The electrical conductivity of the slightly acidic seepage water (pH<6) was on average somewhat greater (140 mS/m) than that of seepage waters with pH>6 (Table 1). In the wetland pools, EC decreased from a mean of 100 mS/m in Wetland1 waters to a mean of 67 mS/m in Wetland2 and the receiving ditch waters. The comparison shows that the conductivity decreased almost by half from the seepage waters to the waters of Wetland2.

The lowest oxygen concentration, 3.4 mg/l with a saturation of about 40%, was measured from waters at the Seepage2 site (Table 1). Waters at the other seepage site (Seepage1) and in the first wetland pool had dissolved oxygen concentrations slightly over 5 mg/l. The concentration increased to about 6 mg/l in waters of the second wetland pool and decreased back to 5 mg/l in the water of the receiving ditch. The oxygen measurements in the laboratory showed small changes, either slight increases or decreases, relative to field measurements.

Table 1. Physical property of the surface waters at the seepage sites (Seepage1, Seepage2), wetland pools (Wetland1_1, Wetland1_2, Wetland2) and receiving ditch (Recipient), of the closed Enonkoski mine area, eastern Finland (see sampling sites in Fig. 2).

Seepage waters were characterized by fairly high concentrations of soluble earth alkaline (Ca, Mg), alkaline metals (Na, K), sulphur, and trace metals (Mn and Ni), whereas concentrations of Cu, Co and Zn were fairly low (Table 2). Slightly acidic seepage waters (Seepage1) had somewhat more Ca, Mg, Na, S, Al, Ni, Co, Zn and Cu than the the near-neutral pH seepage water of Seepage2. In contrast, concentrations of Fe and Mn were somewhat greater in the Seepage2 than Seepage1 waters. Overall, the chemical content of the seepage waters depicts tailings weathering, as well as the remains of chemicals used in ore processing. Alkaline earth, alkaline metals and sulphur originate from lime, Na and/or K bearing xanthates (Na also from explosive chemicals) and sulphuric acid used in the ore processing. Fe, trace metals and part of S are dissolution products of metal sulphide oxidation processes in the shallow tailings (Heikkinen et al. 2009).

The sum of earth alkaline and alkaline metals was about 360 mg/l for seepage waters, 140 mg/l for the outflow from the first wetland pool (Wetland1) and 160 mg/l for the outflow from the second wetland pool (Wetland2) and the receiving ditch (Recipient, See also Table 2). Sulphate concentrations ranged from 810 mg/l in seepage waters, to 260 mg/l in the outflow of the first wetland pool. The outflow of the second wetland pool had somewhat more sulphate, about 300 mg/l, which stayed the same in the receiving ditch.

Concentrations of soluble Fe ranged from 3.3 mg/l in the seepage waters, to 0.07 mg/l in the outflow of the first wetland pool (Table 2). The outflow of the second wetland pool had about 0.05 mg/l Fe, which increased ten-fold (to 0.5 mg/l) in the receiving ditch. Nickel was the major trace metal and its concentration decreased on average nearly two orders of magnitude (i.e. from 2.0 mg/l to 0.08 mg/l) from seepage waters to the outflow of the first wetland pool. However, the Ni concentration rose ten-fold (0.5 mg/l) at the outflow of the second wetland pool. The mean concentration of Mn ranged from about 1.4 mg/l in seepage waters to 0.02 and 0.03 mg/l in the outflows of the wetland pools, but rose to 0.3 mg/l in the receiving ditch. The concentration of Al was greatest in the slightly acidic seepage water, and relatively low in waters of the wetland pools and the receiving ditch. Concentrations of Zn, Cu and Co were much smaller than those of Ni, Mn and Al, but they showed similar trends, being notably higher in the seepage waters than in the outflows of the wetland pools.

Overall, metal concentrations showed a diminishing trend from seepage waters to waters of the first wetland pool. An increasing trend in some trace metal concentrations (Ni, Zn, Cu, Co) was observed from waters of the first to second wetland pools. Additionally, Fe and Mn concentrations increased from the outflow of the second wetland pool to waters of the receiving ditch.

Table 2. Soluble element concentrations in the surface waters at the seepage sites (Seepage1 and 2), wetland pools (Wet1_1, Wet1_2, Wet2) and the receiving ditch (Recipient) of the closed Enonkoski mine area, eastern Finland (see sampling sites in Fig. 2).

Physical and chemical properties of peat sediments in the wetland pools

The pH of the topmost peat sediments was about 6 in the first wetland (Wetland1) and neutral in the second wetland (Wetland2, Table 3). The lower peat sediments of both wetland pools had pH values somewhat over 5. Total concentrations of organic C were lowest, on average 36%, in the topmost peat above the depth of 0.3 m, and highest, 42-47%, below the depth of 0.3 m. The difference between the sediment layers is caused by the accumulation of minerogenic matter in the topmost sediments of the pools. Total concentrations of N and S did not, however, reveal notable accumulation in the topmost sediments of the wetland pools. In the both wetland pools, the topmost peat sediments had lower concentrations of N and S than the peat sediments below a depth of 0.6 m (Table 3).

Table 3. pH and total concentrations of S, organic C and N in the peat sediments of the first (Wetland1) and second (Wetland2) wetland pools of the closed Enonkoski mine area, eastern Finland.

Major acid extractable elements in the peat sediments were S, Fe, Ca and Mg (Table 4). In the first wetland pool, the topmost peat sediment above the depth of 0.3 m had somewhat greater acid extractable concentrations of S (2.1%) than the peat sediment, below the depth of 0.6 m (1.9%). In the second wetland pool, the acid extractable S concentrations were distributed vice versa. The topmost peat sediments had 1.4% acid extractable S, but lower peat sediments had much more, about 4.3%, acid extractable S. The distribution of the acid extractable S concentrations paralleled the total concentrations of S, despite differences in the concentration levels, presumably due to differences in precision and accuracy of the methods for peat sample collection.

The difference in S concentrations between the sediment depths was notable for the acetate and BaCl₂ extractions. The acetate extractable S concentration was lower than the BaCl₂ extractable S concentration even though buffered 1-M acetate solution is assumed to facilitate a more effective extract than un-buffered 0.01-M BaCl₂ solution. This was true for the topmost peat sediments of the first wetland pool, but not for those of the second wetland pool. Obviously, the chloride based salt extraction activates more sulphate exchange than the acetate based extraction. This may indicate that the BaCl₂ extractable S depicts easily leachable and exchangeable SO₄-S concentration better than the acetate extractable S. The acetate extractable S presumably reflects the occurrence of SO₄-S fixed as Fe oxyhydroxides and within organic matter rather than the solely exchangeable SO₄-S adsorbed on solid particles (See Kumpulainen et al. 2007). In the first wetland pool, the topmost peat sediment BaCl₂ extractable S was on average 0.4%, relative to 0.2 % acetate extractable S,  from which the concentration diminished by more than an order of magnitude below the depth of 0.3 m (≤0.02 % BaCl₂ extractable S). In the second wetland pool, in the topmost peat sediment, the acetate extractable S was on average 0.2% and BaCl₂ extractable S 0.05 %. Concentrations of the acetate and BaCl₂ extractable S were small in the lower peat sediments of the both wetland pools (Table 4).

Surface peat sediments of the first wetland pool had on average 2.1% of acid extractable Fe, while lower peat sediments below the depth of 0.6 m contained 1.3% of Fe (Table 4). In the second wetland pool, surface peat sediments contained about 0.7% of acid extractable Fe, while in the lower peat sediments acid extractable Fe was much greater, at about 3.8%. It is notable that the surface peat of the first wetland pool and the lower peats of the both pools had 2:1 S:Fe molar ratios, indicating potential formation of Fe sulphides, whereas in the other samples the acid extractable concentration of Fe was much lower than that of S (Table 4).

Table 4. Concentrated nitric acid extractable (Acid leach), acetate (Ac leach) and dilute BaCl₂ (BaCl₂ leach) extractable concentrations of the essential elements in topmost and lower peat sediments of the wetland pools (Wetland1, Wetland2), the closed Enonkoski mine area, eastern Finland.

The distribution of the acetate extractable Fe concentrations also supported the accumulation of Fe in the topmost rather than underlying peat sediments in the first wetland pool. In the second wetland pool, the acetate extractable Fe concentrations were 0.04% in the topmost, and 0.07% below the depth of 0.6 m. The acetate extractable Fe concentration may indicate the presence of non-crystalline Fe precipitates (Fe hydroxide and mono-sulphide) and Fe bound with organic compounds. In contrast to S, the extractability of Fe in the BaCl₂ leach was pretty small (<0.002%) which parallels the soluble Fe concentrations of the waters.

The acid extractable concentration of Ca ranged from 1.5-1.6% in the topmost peat sediments, to 2.1% in the lower peat sediments of the first wetland pool. In the second wetland pool, the acid extractable Ca concentration of the topmost peat sediments was six-fold (9.4 %) that observed in the first wetland pool, while the concentrations in the lower peat sediments were slightly lower than in Wetland1 (1.4 %). Furthermore, the topmost peat sediments of the first wetland pool had somewhat less acetate extractable Ca (1.3-1.4 %) than its lower peat sediments (1.8 %), whereas the topmost peat sediments of the second wetland pool had nine-fold more acetate extractable Ca (9.0 %) than its lower peat sediments (1.4 %). The BaCl₂ extractable (i.e. easily leachable and exchangeable) concentrations of Ca varied between 0.2 and 0.3% in the both wetland pools. In summary, the majority of the acid extractable Ca was extractable in the acid acetate, which indicates that Ca is mainly bound by non–crystalline, secondary precipitates (sulphate and hydroxide), and acetate extractable Ca bearing organic compounds. The main Ca source was the lime added to the pools to precipitate metals during active mining. However, part of the Ca, and presumably most of the Ca bound by the lower peat sediments, originate from natural nutrient accumulation from run-off during the postglacial period.

The acid extractable concentrations of Mg, K and Na were greater in the surface peat sediments of the first wetland pool than the second (Table 4). Concentrations of acetate and BaCl₂ extractable (i.e. exchangeable, See Fig. 5) Mg, K and Na were distributed similarly. The difference between the acid and acetate extractable concentrations was greater for the topmost peat sediments of the first wetland pool than in the second (Table 4). Due to the partial extraction character of the nitric acid leach, acid extractable concentrations of Mg and K can reflect the presence of mica and clay minerals (biotite, chlorite) in the sediments. Therefore, it is suggested that the topmost peat sediments of the first wetland pool were more enriched with Mg and K bearing silicate minerals than those of the second wetland pool.

Figure 5. Concentrations of the BaCl₂ extractable (i.e. exchangeable) Mg, K and Na (a) in the topmost (0-0.3 m) and (b) lower peat sediments of the first (Wetland1) and second (Wetland2) wetland pools from the closed Enonkoski mine area, eastern Finland.

Furthermore, it is assumed that acetate, and especially BaCl₂, extractable concentrations of Mg, K and Na depict both silicate weathering products and process chemical remnants (K, Na) in the seepage waters discharged from the tailings facility. The concentrations in both extractions showed that the topmost peat sediments of the first wetland pool have trapped somewhat more Mg, Na and K than those of the second wetland pool. The BaCl₂ extractable concentrations of the first wetland pool, however, show that minor exchangeable Mg, K and Na originating from tailings seepage have infiltrated into the lower peat sediments. Additionally, the lower peat sediments of the second wetland pool, which had somewhat more easily leachable and exchangeable Mg, K and Na, may indicate tailings pore water leakage via groundwater into the second wetland pool (see Figs. 1 and 3). Nevertheless, it should be added here that similar to Ca, some Mg and K originates by natural nutrient cycling within peat sediments.

Concentrations of acid extractable trace metals were notably greater in the topmost peat sediments of both wetland pools than in their lower peat sediments (Table 4). The highest average concentrations were measured from the topmost peat sediments of the second wetland pool, and were 490 mg/kg Mn, 1010 mg/kg Ni, and 330 mg/kg Cu. In the first wetland pool, the topmost peat sediments had on average 330 mg/kg Mn, 810 mg/kg Ni and 170 mg/kg Cu. The concentrations of Zn and Co varied between 30 mg/kg and 60 mg/kg. The lower peat sediments had on average 270 mg/kg Mn, 9 mg/kg Ni and 20 mg/kg Cu in the first wetland pool, and 180 mg/kg Mn, 30 mg/kg Ni and 40 mg/kg Cu in the second wetland pool. The concentrations of Zn and Co in the first wetland pool were almost one-tenth of the concentrations in the second wetland pool. The concentration of Cd were below the detection limit (<0.5 mg/kg) in all of the peat samples.

Concentrations of acetate extractable Mn were almost the same as its acid extractable concentrations which indicates that Mn is mainly chemically adsorbed by solid particles (Mn organo-complexes and/or fixed with non-crystalline inorganic precipitates). This agrees with the finding that the easily leachable and exchangeable (BaCl₂ extractable) concentrations of Mn were on average small (<30 mg/kg) in all samples (Table 4).

The topmost peat sediment of the first wetland pool had somewhat lower (300 mg/kg) acetate extractable Ni than that of the second wetland pool (450 mg/kg) (Table 4). The concentration decreased with depth to one-tenth of surface values below 0.3 m, and to one-hundredth in the lower peat sediments below the depth of 0.6 m. In the both wetland pools the average BaCl₂ extractable Ni concentration was low, ranging from 5 to 7 mg/kg in the topmost peat sediments, to about 2 mg/kg in the lower peat sediments. The distribution of Ni in the different extraction fractions shows that more than half (55-65%) of the Ni content was tightly bound with solid particles, with the remainder chemically adsorbed to solid particles. Overall, the concentration of potentially mobile Ni in peat sediments within the wetland pools was small (<10 mg/kg).

In contrast to Ni, concentrations of acetate and BaCl₂ extractable Cu was small in all sediment samples (Table 4). This indicates that Cu was mainly tightly bound with solid particles. Nevertheless, the topmost peat sediments had small quantities of acetate extractable Co and Zn. The concentrations of the BaCl₂ extraction showed that there was no risk of Co and Zn mobilization from peat sediments into the pool waters. Furthermore, the peat sediments had no measureable concentration of acetate and BaCl₂ extractable Cd.

Efficiency of passive water treatment in two wetland pools

The efficiency of the water treatment in the wetland pools was estimated by comparing element concentration means of the seepage waters with those of the receiving ditch waters (recipient). The decrease in the concentrations was about 60% for S, 80-90% for Fe, Mn, Ni, Cu and Co, and about 75% for Zn (Fig. 6a). Aluminium retention was highly efficient, at 99%. From 60% to 70% of Mg and alkaline metals (Na, K), and only 40% of Ca was retained in the wetland pools. The retention of the metals and sulphur of the seepagae waters occurered mainly in the first wetland pool. The retention of Mg, Na, K, S and Mn continued in the second wetland pool, while Ca, Ni, Zn and minor amounts of Cu and Co were remobilized into the waters of the second wetland pool (Fig. 6b). The origin of the remobilized elements is assumed to be linked to dewatering of the tailings rather than seepage. During mine closure water was pumped from the mine shaft and limed to precipitate metals as hydroxides in the second settling pool (U.-P. Mustikkamäki 2005, Pers. Comm., See also Myllymäki 2005). Obviously, the metal hydroxides and remains of lime additive (Ca) stay acetate extractable. The resultant labile compounds could be one explanation for the abundance of Ni and Ca in waters of the second wetland pool compared to the first wetland pool (Fig. 7). For instance, Ni hydroxides are soluble in neutral and slightly acidic conditions.

The mean concentration of soluble Ni (340 µg/l) in waters of the receiving ditch exceeds its environmental quality standard of 21 µg/l (Government Degree 1022/2006 and amendment 868/2010, Finnish legislation). Furthermore, some Fe and Mn is remobilized from the second wetland pool into the water discharging to the receiving ditch (See Table 2). Concentrations of the other trace metals were, however, very small in waters of the receiving ditch.

Figure 6. (a) Average total retention (%) of Ca, Mg, Na, K, Al, S, Fe, Mn, Ni, Cu, Co and Zn in the wetland pools and (b) average retention (%) of sulphur and chalcophile metals in the first and second wetland pools of the closed Enonkoski mine area, eastern Finland. The total retention of an element in the wetland pools is calculated by subtracting the mean concentration in receiving ditch waters from that of seepage waters. In the histogram b, the retention (%) of an element in the first wetland pool (Wetland1) was calculated by subtracting the mean concentration in wetland1 effluent from seepage, and the retention in wetland2 was calculated by subtracting mean concentration in wetland2 effluent from wetland1 effluent.

Figure 7. Distribution of soluble Ni and Ca concentrations from seepage waters (Seepage1 and -2) downstream to waters of the Wetland 1 (Wet1_1, Wet1_2) and Wetland2 (Wet2) and receiving ditch (Recipient), the closed Enonkoski mine area, eastern Finland.

In addition, the treatment rate of the wetland pools was estimated on the bases of the geochemical fractionation of peat sediments. Here the retention percent was estimated by calculating proportion of an element concentration sum of tightly bound and chemically adsorbed fractions from its acid extractable concentration. It is assumed that in addition to the tightly bound fraction, elements bound to the chemical adsorption fraction (i.e. acetate extractable elements) are fairly stable if pH and redox conditions of the pool stay unchanged (See Kumpulainen et al. 2007, Heikkinen & Räisänen 2009).

Results of the geochemical fractionation show that peat sediments of the wetland pools trapped the majority of sulphur and metals from seepage waters of the tailings. From 90 to 100% of S and chalcophile metals (Fe, Mn, Ni, Cu, Co, Zn) were tightly bound and chemically adsorbed by the topmost peat sediments of both wetland pools (Fig. 8). Furthermore, Fe and S were mostly (80-90%) bound to the tightly bound, acid extractable fraction that indicates the formation of Fe sulphides. More than 50% of the other chalcophile metals (Ni, Cu, Zn, Co) were bound to the tightly bound fraction, presumable by sulphides. The rest of the trace metals and most of Mn were chemically adsorbed by solid particles within noncrystalline hydroxide precipitates.

The geochemical distribution of sulphur, Fe and other chalcophile (Mn, Ni, Cu, Co, Zn) metals of the lower peat sediments was similar to that of the topmost peat sediments (Figs. 8-9). Sulphur and Fe were however, entirely (100%) bound to sulphides within the lower peat sediments of both wetland pools, and their distribution to the chemical and physical adsorption fraction was pretty small (≤2%, Fig. 9). Furthermore, the lower peat sediments had small concentrations of the other chalcophile trace metals (Ni, Cu, Co, Zn). The findings suggest that the lower peat sediments mostly represent the natural, background composition of the peat. In contrast to the first wetland pool, the lower peat sediments of the second wetland pool had somewhat higher chalcophile metals concentrations (See also Table 4). This finding may indicate that tailings seepage flows via groundwater into the second wetland pool. This transport pathway could be further investigated by monitoring the groundwater in the till between the wetlands and the tailings impoundment.

The topmost peat sediments of the first wetland pool retained from 80% to 90% of the Ca, Mg and K in seepage waters, whereas the topmost peat of the second wetland pool retained more Ca (97%) and somewhat less Mg (70%) and K (60%, Fig. 8). 35-40% of Na was retained in the topmost sediments. Compared to these findings, the lower peat sediments in both wetland pools trapped less Ca, Mg, K and especially Na to the tightly bound and chemically adsorbed fractions (Fig. 9).

Figure 8. Proportions (%) of element (Ca, Mg, K, Na, S, Fe, Mn, Ni) concentrations bound to the tightly bound, chemically and physically adsorbed fractions of the topmost peat sediments (a) of the first and (b) second wetland pools of the closed Enonkoski mine area, eastern Finland.

Figure 9. Proportions (%) of element (Ca, Mg, K, Na, S, Fe, Mn, Ni) concentrations bound to the tightly bound, chemically and physically adsorbed fractions of the lower peat sediments (a) of the first and (b) second wetland pools of the closed Enonkoski mine area, eastern Finland.

Conclusions

After the closure of the Enonkoski Ni-Cu mine, the former settling pools close to the tailings impoundment have naturally evolved into wetland pools, growing aquatic plants in shallow water. The two-segment settling pools were constructed on a bog surrounded by moraine hummocks. The analysis of surface waters and peat sediments within the pools showed that slightly acidic and almost neutral seepage waters from the tailings impoundment were passively neutralized and treated within the wetland pools. The retention rate for S was 60%, 80-90% for Fe and other chalcophile metals (Mn, Ni, Cu, Co), and 75% for Zn.

On the basis of the geochemical fractionation data of wetland peat sediments, it can be interpreted that the main mechanism for treatment is sulphate and metal reduction, and secondarily the chemical adsorption of metals and sulphate onto solid particles. Overall, the treatment efficiency was somewhat better in the first than in the second wetland pool. The pumping and liming of dewatering waters from the mine shaft during mine closure was obviously the main source for lowered metal retention in the second pool. Labile metal hydroxides have remobilized as the pH of pond water has dropped to neutral, which is especially apparent in the dissolution of Ni. In addition to Ni, the outflow of the second wetland pool contained Fe and Mn. Furthermore, the canal-type design of the second wetland pool may result in faster water removal, which hinders metal sulphide formation and may increase remobilization of metals via sediment erosion.

Performance and suitability of settling pools for passive treatment of tailings seepage waters

In the closed Enonkoski mine area, the treatment of seepage waters from the tailings facility is based on passive retention of metals and sulphur in wetland sediments. The wetlands were originally settling pools constructed on a bog for clarifying tailings waste water. The two-segment settling pools were left to growing aquatic plants and evolve into wetland pools. The water level and volume of the pools fluctuate according to annual precipitation and seepage discharge from the tailings impoundment. The basement (pad) of each pool consists of peat sediments enriched in inorganic matter at shallow depths (0-0.3 m).

The results of the Enonkoski case study showed that the water treatment rate is fairly good. From 60 to 90% of sulphur and chalcophile metals were trapped in wetland peat sediments. Also from 60% to almost 100% of alkaline earth and alkaline metals, with the exception of Na (≤40%), were retained fairly well in the sediments. It is assumed that the main treatment mechanism is the chemical reduction of sulphate and metals and subsequent formation of Fe and trace metal sulphides. The secondary treatment mechanism is the chemical adsorption of metals and sulphate on to solid particles.

The discharge from the second wetland pool into the receiving ditch had elevated concentrations of Ni, Fe, Mn and Ca. This is assumed to be caused by the addition of limed dewatering waters during the mine closure. The subsequent decrease in pH is assumed to have caused remobilization of Ni from sediments into water. In addition, water removal is obviously faster in the second wetland pool than first due to differences in the pool designs. This could increase sediment erosion and remobilize metals.

The Enonkoski settling pool case is a good example of the natural development of wetland pools with fluctuating water levels that promote the re-establishment of aquatic plants and maintain reducing conditions in the substrate. Here one key for passive water treatment is the peat substrate, which naturally generates sulphate reducing bacteria (Walton-Day 2003). Furthermore, rot of aquatic plants and the abundance of cattails support reducing conditions and deliver organic carbon as an energy source to SRB (sulphate reducing bacteria). According Todorova et al. (2005), the optimal conditions for pyrite formation is between pH 6 and 9 and slightly negative Eh values (-3 to -6). Although redox values within the peat sediments were not measured during sampling, it is assumed on the basis of the abundance of Fe and S in the acid extractable fraction, i.e. sulphide fraction that reducing conditions were present within the peat. Even though the wetland pools showed good treamtent efficiency of waters, the remobilization of some metals (Ni, Mn, Fe) in the second wetland pool revealed potential interference caused by liming. The liming quickly elevated water pH and precipitated metals as metal hydroxides that were sensitive to subsequent dissolution as pH decreased. Furthermore, design of a wetland pool can promote fast water removal and sediment erosion that weakens permanence of metal fixation in the substrate of the wetland.

Study materials consist of unpublished sediment data from the wetland pools collected in 2002 by the GTK, and water data collected in 2002 and 2005 published in the MSc thesis of Myllymäki in 2006. Despite some general lack of data (e.g. the small amount of monitoring data, no sediment redox measurements, low grade detection limits for several trace analytes), the general performance of the wetland pools could be estimated. To identify the actual biological treatment mechanisms and controlling factors would require analyses of the bacterial population and secondary precipitates such as gypsum, Fe oxyhydroxides and metal sulphides, as well as in situ measurements of pH and redox from both water and peat sediments across different seasons. In addition, the quantification of metal and sulphur loads into receiving waterways requires water discharge measurements in the seepage area, both wetland pools and the receiving ditch.

References

Alopaeus, E., Grundström, L., Pitkänen, R. & Virtanen, M. 1986. Outokumpu Oy:n Enonkosken kaivos. Vuoriteollisuus 44 (1), 16-23 (in Finnish, English Summary).

Doležal, J., Povondra, P., & Šulcek, Z. 1968. Decomposition techniques in inorganic analysis. London Iliffe books Ltd, London.

Grundström, L. 1985. The Laukunkangas nickel-copper deposit. In: H. Papunen & G. I. Gorbunov (eds.) Nickel-copper deposits of the Baltic Shield and Scandinavian Caledonides. Geological Survey of Finland, Bulletin 333, 240-256.

Grundström, L. 1986. The Enonkoski nickel-copper mine. Excursion guide, C3: Metallogeny and ore deposit in south Finland. 17e Nordiska Geologmötet, Helsinki, 63-70.

Heikkinen, P.M. & Räisänen, M.L. 2009. Heavy metal and As fractionation in sulphide mine tailings – indicators of sulphide oxidation in active tailings impoundments. Applied Geochemistry 24, 1224-1237.

Heikkinen P.M., Räisänen M.L. & Johnson R. H. 2009. Geochemical characterisation of seepage and drainage water quality from two sulphide mine tailings impoundments: acid mine drainage versus neutral mine drainage. Mine Water and the Environment 28, pp. 30-49.

Kumpulainen, S., Carlson, L. & Räisänen, M.L. 2007. Seasonal variations of ochreous precipitates in mine effluents in Finland. Applied Geochemistry 22, 760-777.

Myllymäki, T. 2006. Kuparin, nikkelin ja sinkin pidättyminen luonnollisesti syntyneillä kosteikoilla. Pro gradu-tutkielma. Kuopion yliopisto, ympäristötieteiden laitos. 106 s.

Niskavaara, H. 1995. A comprehensive scheme of analysis for soils, sediments, humus and plant samples using inductively coupled plasma atomic emission spectrometry (ICP-AES). In: S. Autio (ed.) Geological Survey of Finland, Current Research 1993-1994. Geological Survey of Finland, Special Paper 20, 167-175.

Pöyry, H. & Isomäki, O.-P. 1996. Outokumpu Finnmines Oy:n Enonkosken kaivos. Vuoriteollisuus 54 (1), 33-40.

Saari, J., Eerola, M., Karinen, M. & Luhta, U. 1993. Kaivos- ja metallurgisen teollisuuden jätteiden inventointi ja luokittelu (Kuparin, sinkin ja nikkelin perustuotannossa Suomessa syntyvät jätteet). Outokumpu Research Oy. Julkaisematon raportti ORC-T 15.1.1993. 197 s.

Schultz, E., Joutti, A., Räisänen, M.L., Lintinen, P., Martikainen, E. & Lehto, O. 2004. Extractability of metals and ecotoxicity of soils from two old wood impregnation sites in Finland. Science of the Total Environment 326, 71-84.

Sencindiver, J.C. & Bhumbla, D.K. 1988. Effects of cattails (Typha) on metal removal from mine drainage. U.S. Bureaus of Mines Circular, 359-366.

Sengupta, M. 1993. Environmental impacts of mining. Monitoring, Restoration, and Control. Boca Raton, Florid: CRC Press LLC, 494 s.

Todorova, S.G., D.I. Siegel & Costello, A.M. 2005. Microbial Fe(III) reduction in a minerotrophic wetland – geochemical controls and involvement in organic matter decomposition. Applied Geochemistry 20, p. 1120-1130.

Walton-Day, K. 2003. Passive and active treatment of mine drainage. In: Jambor, J.L., Blowes D.W. & Ritchie, A.I.M. (eds.): Environmental aspects of mine wastes. Mineralogical Association of Canada. Short Course Series vol. 31, p. 335-359.