- Email: [email protected]
Spatial and temporal evolution of Cu–Zn mine tailings during dewatering
Applied Geochemistry 26 (2011) 1832–1842
Contents lists available at ScienceDirect
Applied Geochemistry journal homepage: www.elsevier.com/locate/apgeochem
Spatial and temporal evolution of Cu–Zn mine tailings during dewatering Barbara L. Sherriff a,⇑, D. Jared Etcheverry a, Nikolay V. Sidenko a,b, Jamie Van Gulck c a
Department of Geological Sciences, University of Manitoba, Winnipeg, MB, Canada R3T 2N2 Stantec Consulting Ltd., Winnipeg, Manitoba, Canada R3C 3R6 c Arktis Solutions Inc., 117 Loutitt St., Yellowknife, NT, Canada X1A 3M2 b
a r t i c l e
i n f o
Article history: Available online 17 June 2011
a b s t r a c t The Ruttan Cu–Zn mine produced about 50 mT of fine-grained tailings over 30 a. Since the closure of the mine in 2002, the tailings have been systematically dewatered through trenches draining into the open pit and underground workings. This study evaluated the evolution of tailings that were underwater until 2002, and also tailings that had been exposed to oxidizing conditions for more than 20 a. Acid generation is dominated by the oxidization of the abundant pyrite and pyrrhotite comprising 25 wt.% of tailings with Zn being mobilized from the sphalerite remaining after beneficiation of the ore. Little Cu is being mobilized partly due to the armouring of remnant chalcopyrite by primary quartz, and also by preferential absorption of Cu rather than Zn on secondary Fe oxy-hydroxides. A very fine grained fraction of unoxidized sulfides is a likely cause of an initial pulse of acid generation, and metal and SO2 4 release occurring at the onset of dewatering. Metals are temporarily attenuated from waters associated with the tailings, either absorbed on Fe oxy-hydroxide precipitates or as evaporite hydroxy sulfate minerals at the surface of the tailings. While some secondary phases are stable, evaporites are only temporary metal sinks as they redissolve in wet weather conditions. Trace amounts of calcite provide little buffering capacity resulting in rapid acidification of pore and surface water. The pH of pore water and shallow groundwater decreased first to the Al oxy-hydroxide buffer at 4.5 and then stabilized at values of 2–3 being controlled predominantly by the dissolution of solid Fe oxy-hydroxides. The metal contents of the ground and surface water are still increasing but the Ruttan Lake reservoir that receives drainage water from the tailings is maintaining a constant composition. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction The mining of base metals produces large quantities of finegrained sulfide tailings. If the tailings are exposed to the atmosphere, sulfide oxidation can produce acid (Blowes et al., 2003a). If not neutralized by carbonates, this acid will release metals and can contaminate ground and surface water (e.g. Blowes and Ptacek, 1994; Johnson et al., 2000; Blowes et al., 2003b). Although the basic reactions have been well studied both in the field (e.g. Petrunic and Al, 2005; Peretyazhko et al., 2009) and the laboratory (e.g. Liu et al., 2008), each mine site is unique in terms of geology, climate, environmental conditions and the proportions of acid producing and neutralizing minerals. The dewatering of the tailings at Ruttan Cu–Zn Mine, Manitoba, Canada, after closure in 2002, provided an opportunity to observe potentially acid generating tailings during the first years of dewatering and oxidation in a subarctic-climate. The objectives of this study were to observe changes in mineralogy and geochemistry during oxidation of the previously submerged tailings, to predict ⇑ Corresponding author. E-mail address: [email protected] (B.L. Sherriff). 0883-2927/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.apgeochem.2011.06.007
their evolution from characterization of already oxidized tailings and to determine the geochemical evolution of Ruttan Lake. Selected analytical data are presented here. Full analytical results are available in Etcheverry (2008). 2. Study site Ruttan Mine-site is located approximately 950 km north of Winnipeg, 30 km east of the community of Leaf Rapids, Manitoba (Fig. 1). From 1973 until 2002, Cu and Zn ore were extracted from volcaniclastic and siliciclastic sequences of the Rusty Lake Greenstone Belt first by Sherritt Gordon Mines Ltd. and after 1987 by Hudson Bay Mining and Smelting Co. Ltd. (HBM&S) (Stantec Consulting Ltd., 2003). The massive-sulfide ore lenses are hosted by tholeiitic, felsic and intermediate volcanic and volcaniclastic rocks. Metamorphism to upper greenschist and lower amphibolite facies is characterized by the presence of cordierite (Mg2Al4Si5O18), almandine ðFe2þ andalusite (Al2SiO5), sillimanite 3 Al2 ðSiO4 Þ3 Þ, (Al2SiO5), biotite (K(Mg,Fe)3[AlSi3O10(OH,F)2), staurolite ((Fe,Mg)2Al9(Si,Al)4O20(O,OH)4), anthophylite (Mg7(Si8O22)(OH)2), and talc (Mg3Si4O10(OH)2). Other gangue minerals include muscovite
B.L. Sherriff et al. / Applied Geochemistry 26 (2011) 1832–1842
1833
Fig. 1. The location of Ruttan Mine and mine site plan showing tailings management cells and drainage trenches from these cells to Ruttan Lake.
(KAl3Si3O10(OH,F)2), chlorite (KAl3Si3O10(OH,F)2), augite ((Ca,Na)(Mg,Fe,Al,Ti)(Si,Al)2O6), quartz (SiO2), anhydrite (CaSO4) and the carbonate minerals, siderite (FeCO3), dolomite (CaMg(CO3)2), ankerite (Ca(Fe,Mg,Mn)(CO3)2 and calcite (CaCO3). Primary sulfide phases consist of pyrite (FeS2), pyrrhotite (Fe(1x)S), chalcopyrite (FeCuS2), sphalerite (ZnS), and galena (PbS) with minor tetrahedrite ðCu9 Fe2þ 3 Sb4 S13 Þ (Barrie et al., 2005). Flotation refinement of the Ruttan ore removed sphalerite and chalcopyrite, producing over 50 million tonnes of fine-grained sulfide-rich tailings which were discharged into Cell 1 and Cell 2 of the tailings impoundment. Cell 3 acted as a polishing pond before the water was discharged into Ruttan Lake (Fig. 1). Most of these tailings remained submerged after deposition but the northern portion of Cell 2 had been exposed to the atmosphere and oxidized for about 20 a (Frank Bloodworth, HBM&S, pers. comm.). During operation of the mine, lime was added both to Ruttan Lake and at a spillway constructed at the outlet to Brehaut Lake (Stantec Consulting Ltd., 2003). In the final stages of remedial activity by HBM&S, this weir was destroyed and the outlet sealed. To improve the long term stability of the tailings dams, tailings ponds are being systematically dewatered through a network of drainage trenches about 3–4 m deep and 3 m wide (Figs. 2 and 3). The drainage water is being directed into Ruttan Lake. The entire onsite watershed including Ruttan Lake, the mill pond and drainage from the waste rock piles was redirected and now drains into the open pit and then to the underground workings (Fig. 2). The drainage of the watershed surrounding the mine was rerouted so that natural surface runoff skirts the site.
The subarctic climate at Leaf Rapids shows annual temperatures from 47 °C to 35 °C with snow cover from October to May and the groundwater often frozen from October to July. This results in a very short period each year for reactions to occur in the tailings at a significant rate and for the groundwater to drain through the trenches.
3. Methods 3.1. Sample collection Sampling locations for solid tailings (RTP2, RTP4) were chosen to be representative of tailings that had been oxidizing over a period of about 20 a (Cell 2) and of those that were water saturated and unoxidized until 2003 (Cell 3) (Fig. 2). In 2004, solid tailings were collected at 10 cm intervals from pits to a depth of <1 m, and by auger beneath the pit at 0.5 m intervals. Samples for thin sections were collected by pressing 4 6 1 cm Al boxes into the tailings and removing them with their contents. Vertical sections of tailings were collected in PVC pipes (30 cm long with 7.6 cm I.D) sealed with PVC end-caps. Precipitates were collected from the bottom of ponds and evaporites from desiccation cracks on the tailings surface. All solid samples were frozen at the end of each field day. Surface water was collected periodically between 2003 and 2009 from several sites including a stream (RW1) and trench (RW2) draining Cell 3, from the stream draining into Ruttan Lake (RW4), from the west side of Ruttan Lake (RW8 and RW9) and from
1834
B.L. Sherriff et al. / Applied Geochemistry 26 (2011) 1832–1842
Fig. 2. Ruttan Mine site after closure in 2002 showing elevation relative to the surface of Ruttan Lake (D), and the drainage plan after mine closure. Solid (RTP), surface water (RW) and groundwater (RBH) sampling sites are shown.
the outlet from Ruttan Lake to the open pit (RW3) (Fig. 2). The Lake samples were collected to study the dilution effect of the Lake and its evolution as the tailings were dewatered. A total of 12 monitoring wells were installed in Tailings Cells 1 and 3 in 2003 and 2004 (Etcheverry, 2008). By 2009, many of the wells had dried up, been broken or buried by the movement of the tailings or become contaminated with surface water or sediments, therefore, longitudinal studies are presented on only three wells, RBH4, RBH5 and RBH7. The locations of these wells and surface sampling sites are shown in Fig. 2. The wells (RBH 4, 5, 6) were constructed with solid 50.8 mm diameter PVC casing (2.70, 1.09, 1.46 m in length) and No. 10 factory slot well screens (0.40, 2.54, 1.98 m) to the borehole base. The annulus was backfilled with clean silica sand from the base to just above the solid/screen join and with granular bentonite to the surface. Groundwater was obtained after each well had been purged. All water samples were passed through 0.45 lm filters and divided into two aliquots. The one for cation analysis was acidified with 0.01 mL of 5.6 M HNO3 per 1 mL of sample whereas the other for anion analyses was left unacidified. Eh and pH were measured immediately with an AP-62 pH/Eh meter using a Thermo Orion ORP-97-87 Ag/AgCl electrode (for Eh) and an Accumet 13-620AP50 Ag/AgCl electrode (for pH). 3.2. Laboratory methods Samples collected for mineralogy were impregnated with epoxy, before being made into polished thin sections which were prepared without the use of water to preserve soluble mineral phases. These samples were analyzed under transmitted and reflected optical microscopy and with a Cambridge Instruments Scanning Electron Microscope (SEM) equipped with an Energy
Dispersive X-ray Spectrometer (EDX) for elemental analysis. Mineral phases in evaporites were manually separated under a binocular microscope for analysis by SEM and by powder X-ray diffraction (XRPD), using a Philips PW1710 XRPD system and MDI Datascan/Jade data processing software. Precipitates, which formed in unacidified water samples stored at room temperature for 2 months, were filtered, dried, and analysed by XRPD and SEM. Bulk samples of tailings obtained from pits and auger holes were dissolved in HF/HNO3 before geochemical analysis at the University of Manitoba using a Varian Liberty 200 Inductively Coupled Plasma Optical Emission Spectrometer (ICP-OES) for cations referenced to ICP standard solutions (SCP Science). Samples of tailings underwent a sequential extraction process as outlined in Table 1 with a sequence of water soluble, adsorbed/exchangeable/carbonate, amorphous Fe oxy-hydroxide and crystalline Fe oxide fractions. The residual sulfide/silicate phase was inferred from a comparison of sequential extraction results with the bulk chemistry. Pore water was extracted from 10 cm intervals of frozen cores of tailings. The samples were thawed in a N2-filled glove bag, placed in the sample chamber of a press and compressed with a hydraulic press operating at 5 tons of pressure. The resultant pore water was passed through a 0.45 lm filter, separated into aliquots and preserved as described for the field samples. Cations in acidified water samples were measured in 2003 and 2004 using a Varian Liberty 200 ICP-OES at the University of Manitoba, and from 2005 by Inductively Coupled Plasma Mass Spectrometry (ICP-MS) at Envirotest (now ALS) Laboratories, Winnipeg. Anions in samples from 2003 and 2004 were analyzed in the Department of Microbiology, University of Manitoba using a Dionex™ ion-chromatography system and in samples from 2005 by ICPMS at Envirotest Laboratories.
1835
B.L. Sherriff et al. / Applied Geochemistry 26 (2011) 1832–1842
Fig. 3. Dewatering trench cutting tailings Cell 1 in (a) 2003 (b) 2005.
Table 1 Sequential extraction procedures. Target phase
Reagent
Method
Reference
Water soluble fraction Adsorbed/ exchangeable/ carbonate
Deionized H2O
Dold (2003) Hall et al. (1996)
Amorphous Fe oxyhydroxide
0.2 M NH4-oxalate 0.1 M (NH2)2C2O4
Crystalline Fe oxide
1.0 M NH2OHHCl 25% CH3COOH
1.0 g of sample into 20 mL deionized H2O, vortex, shake for 1 h. Centrifuge, remove supernatant, preserve with 5.6 M HNO3 and analyze Take the residue from the previous leach and add 20 mL 1.0 M CH3COONa; to pH 5.0 with HCl. Vortex for 5–10 s and place in shaker for 6 h. Centrifuge for 10 min at 2800 rpm and decant supernatant into a labeled test-tube and preserve with 5.6 M HNO3. Rinse residue with 5 mL of water, vortex and centrifuge again: add supernatant rinses to the test-tube. Carry out a second 20 mL 1.0 M CH3COONa leach of the residue, repeating steps 2 and 3 add supernatant to the test-tube Take the residue from the previous leach and add 20 mL NH4-oxalate; to pH 3 with 0.1 M oxalic acid. Vortex, shake on horizontal shaker for 1 h in dark at room temp. Centrifuge the sample for 10 min, and decant the supernatant into a storage container. Acidify the supernatant with HNO3. Wash the sample with deionized water, shake to mix and centrifuge for 30 min. Add rinse water to sample Take the residue from the previous leach and add 20 mL of 1.0 M NH2OHHCl in 25% CH3COOH, vortex from 5–10 s. Place in water oven at 90 °C for 3 h (tightly capped), vortex every 20 min. Centrifuge for 10 min and decant supernatant fluid into a labeled test-tube. Carry out a second 1.0 M NH2OHHCl leach of the residue but heat for only 1.5 h, add supernatant from this step to that of the last, preserve with 5.6 M HNO3 and analyze
1.0 M CH3COONa 0.25 M HCl
4. Results and discussion 4.1. Site observations As dewatering of Ruttan tailings Cells 1 and 3 progressed from 2003, the surface area of unoxidized tailings exposed to the atmosphere increased; however, high rainfall between April and October resulted in very wet conditions reducing the visible effects of dewatering in 2005.
Dold (2003)
Hall et al. (1996)
In 2003, when the drainage trenches were first excavated the surface of the tailings in Cell 1 and Cell 3 was dark gray and completely unoxidized (Fig. 3a). Patchy surface oxidation in 2004 was limited to the upper 1 cm of the tailings and constrained to areas of higher elevation and to 1–2 cm halos along vertical and horizontal cracks. Below this, the tailings were unoxidized. By 2005, the entire surface of Cell 1 was oxidized, with the oxidized layer being up to 15 cm thick (Fig. 3b) and the oxidation much more pervasive. Sidenko and Sherriff (2005) showed that Fe precipitates from
1836
B.L. Sherriff et al. / Applied Geochemistry 26 (2011) 1832–1842
surface water at Ruttan formed schwertmannite ðFe3þ 16 O16 ðOHÞ12 ðSO4Þ2 Þ and jarosite (KFe3(OH)6(SO4)2). These precipitates caused the immediate change in color of Cell 3 surface from gray in 2004 to brown by 2005. A layer of white evaporite minerals that covered the surface of the tailings in 2003 and 2004 was not present in 2005 due to the wet conditions. 4.2. Mineralogy During beneficiation, ore is crushed to the optimal size for extraction of the metals. This process also produces a much finer fraction ‘‘rock flour’’. About 50% of the unoxidized tailings taken from Cell 3 in 2004 were found by optical microscopy to consist of a dark-gray groundmass, which was comprised of very finegrained sulfides and silicates (Fig. 4a). This fine-grained groundmass was not observed in the oxidized samples from Cell 2. This is probably due to the very fine grained sulfide and aluminosilicate minerals, which would be very reactive due to high surface areas, dissolving rapidly and releasing metals such as Fe, Zn, Cu and Al. Optical and SEM examinations showed that the coarser fraction of the unoxidized tailings contained about 50% sulfides (60% pyrite, 35% pyrrhotite, 5% chalcopyrite and sphalerite) and 50% gangue (quartz, muscovite, biotite, chlorite and pyroxene). Sulfide grains, about 1–2 lm in diameter, showed little sign of corrosion (0.1 lm deep) with thin (0.2 lm) dark brown rims (Fig. 4b). Gangue minerals were partially coated with dark-brown precipitates. No carbonate minerals were found in the tailings by optical or electron microscopy despite being listed in the ore mineralogy. The grain size was much coarser in the oxidized zone of the tailings of Cell 2 than in Cell 3, with sulfide minerals up to 50 lm in diameter (Fig. 4c). This grain size distribution is probably due to the relative proximity of the discharge pipe and rapid settling of coarse grains as has been found in Central Manitoba Au Mine tailings (Sherriff et al., 2007). Pyrrhotite grains from Cell 2 were highly corroded and completely surrounded by thick rims of Fe oxyhydroxides (Fig. 5a) with original grain boundaries being indiscernible. The alteration of sphalerite grains was less pronounced, with moderate corrosion and thin rims of Fe oxy-hydroxides (Fig. 5b). Pyrite remained unaltered except for thin rims of Fe oxy-hydroxides. The only chalcopyrite found in the oxidized tailings was disseminated within quartz and therefore armoured from oxidation processes (Fig. 5c). White and yellow evaporites form as thin crusts on the surface of the tailings, or as cumulates within fissures and cracks. Evaporites were shown by SEM EDX to contain Fe, Zn, Al and S. Melanterite (FeSO47H2O), rozenite (FeSO44H2O) and halotrichite (FeAl2(SO4)422H2O) were identified by XRPD. Zinc and Cu can partially replace Fe in these minerals (Jambor et al., 2000; Peterson, 2003; Ballirano et al., 2003). During dry periods, saline pore water evaporates from the tailings surface forming these hydroxy-sulfates. The evaporite minerals are soluble and will redissolve in times of high rainfall or during freshet releasing the metals to the surface water. Lepidocrocite (FeOOH) was identified by XRD in precipitates from unacidified filtered groundwater stored in the laboratory for a month. 4.3. Geochemistry Average metal concentrations in the unoxidized tailings (n = 22) were 21.0 (3.3) wt.% Fe, 0.16 (0.11) wt.% Zn and 0.06 (0.02) wt.% Cu compared to 26.6 (3.3) wt.% Fe, 0.17 (0.01) wt.% Zn, and 0.05 (0.006) wt.% Cu in oxidized tailings (n = 10). In the oxidized tailings for both years, Fe, Cu and Zn are depleted near the surface but below 100 cm depth there was no significant variation (Fig. 6).
Fig. 4. (a) Photomicrograph showing fine grained background in unoxidized tailings in Cell 3, (b) SEM image of fine grained unoxidized tailings from Cell 3, and (c) SEM image of coarser grained oxidized tailings from Cell 2.
The results from sequential extraction showed that about 3=4 of the total Fe was found in the residual phase (68–83%) and about a 1= 4 as Fe oxy-hydroxides (17–31%), with 0.5% being in water soluble, and <0.4% in exchangeable phases (Table 2). Zinc was more mobile with 20–39% of the total being water soluble and 4–8% exchangeable (Table 2). The remaining Zn is divided between Feoxy-hydroxide phases (16–32%) and residual (30–57%). Copper was found primarily in the Fe oxy-hydroxide (46–51%) and residual phases (13–41%). The remaining Cu is mobile in exchangeable (6–10%) and water soluble (13–28%) fractions (Table 2). Pore waters extracted from tailings in Cells 2 and 3 have some similarities, with low pH and high total dissolved solids near the
B.L. Sherriff et al. / Applied Geochemistry 26 (2011) 1832–1842
Fig. 5. SEM backscatter images of (a) a corroded pyrrhotite (po) grain with a zoned rim of precipitated secondary Fe oxy-hydroxide minerals, (b) a corroded grain of sphalerite (sph), and (c) chalcopyrite (cp) inside quartz. Note the relatively unaltered pyrite (py) in each image.
surface where the tailings are oxidized and have a more neutral pH and lower dissolved solids with depth. In general, pH values were lower and dissolved metal concentrations higher in 2005 than 2004 (Figs. 7 and 8). In Cell 3 in 2004, the pore water of the unoxidized tailings had a pH of 4.4 in the top 10 cm of the tailings which increased to 7.3 by 30 cm depth (Fig. 7). Dissolved metals follow pH, with concentrations of 18,000 mg/L Fe, 9600 mg/L Zn and 260 mg/L Al in the top 10 cm decreasing steadily with depth until most metals were below the detection limits (<0.01 mg/L Fe, <0.02 mg/L Al, <0.005 mg/L Zn) at 30 cm (Fig. 7). Due to a high water table, Cell 3 tailings could only be collected to a depth of 60 cm in 2005.
1837
Average values of dissolved metals in pore water increased compared to 2004 with Fe and Al having the highest values at 15 cm depth: 47,000 mg/L Fe, 9100 mg/L Al. Zinc concentration near the surface decreased from 9600 mg/L to 840 mg/L between 2004 and 2005, with a maximum value of 2600 mg/L at 35 cm depth (Fig. 7). From 2004 to 2005, the overall values of pH had decreased by 2–3 pH units to 2.3 at the surface and 4.8 at 30 cm. Concentration of Cu in pore water was less than 2 mg/L in both 2004 and 2005. In Cell 2 in 2004, pore water extracted from the upper 70 cm of oxidized tailings was characterized by pH values of 1.8 at 10 cm depth which increased to 7.1 at 75 cm depth (Fig. 9). Maximum concentrations of Fe (65,000 mg/L), Zn (9800 mg/L), Cu (1100 mg/L) and Al (29,000 mg/L) occurred at 20 cm depth despite the pH being slightly higher than at 10 cm. Below 1.2 m where the tailings are unoxidized, concentrations of Cu and Zn were just above detection limits of 0.0004 mg/L Cu and 0.005 mg/L Zn. The evolution from 2004 to 2005 in Cell 2 differed from that of Cell 3 with the pH increasing from the surface to a near neutral value of 6.9 at 300 cm depth (Fig. 9). The maximum value for Zn (2600 mg/L) was at 35 cm. There were high concentrations of Fe in pore water down to 150 cm with values from 17,000 to 47,000 mg/L Fe. Aluminum concentration decreased from 3700 mg/L at the surface to 500 mg/L by 25 cm depth. Some of this change may have been due to the saturation of the previously dry tailings by the high rainfall in the summer of 2005. Groundwater from the deepest well (RBH4) in unoxidized tailings had an alkaline pH of 9.6 in 2004, which decreased to 7.1 by 2007 (Table 3). The concentration of SO2 decreased from 4 2430 mg/L in 2004 to 1310 mg/L in 2007 and 1500 mg/L in 2009. Copper, Zn and Al concentrations remained below 1 mg/L. The concentration of Fe increased from 0.02 mg/L in 2004 to 9.71 mg/L by 2009. Wells RBH5 and RBH7 sampled shallower groundwater than RBH4. The pH of water from these wells has stabilized at 4.2 ± 0.1 since 2007 after values in 2004 of 7.5 (RBH7) and 6.9 (RBH5). There was a lower pH of 3.5 in both wells in the extremely wet conditions of September 2005 probably due to deeper permeation of acidic surface water (Table 3). The reductions of 2–3 pH units from 2004 to 2005 correspond to those for pore water at 0.7–1 m depth for Cell 2 (Fig. 8). The concentration of Cu in wells RBH5 and RBH7 remained below 0.05 mg/L from 2004 to 2009. The concentration of Zn increased from 2004 to 2009 but was still less than 5 mg/L in RBH5 although for RBH7 there were higher values of 106 mg/L in 2005 and 155 mg/L in 2009 (Table 3). Between 2004 and 2009, Fe concentration increased by up to four orders of magnitude from 1.09 mg/L to 7290 mg/L (RBH5) and from 2.6 mg/L to 11,900 mg/L (RBH7). Sulfate concentrations also increased significantly during the same period from 1770 mg/L to 31,300 mg/L (RBH5) and 1670 mg/L to 33,100 mg/L (RBH7). The initial high values of metals and SO2 4 in the groundwater could have been caused by the disturbance of the tailings due to construction of the trench in 2002–2003 or to the exposure and oxidation of the fine grained material from unoxidized tailings. High metal values in groundwater in 2009 may be due to the formation of extensive gullies at the side of the trenches, which are exposing unoxidized tailings to air causing acidification and release of metals (Fig. 9). The dramatic effect of erosion surfaces on the geochemistry of tailings was shown in a study of Central Manitoba Au mine tailings (Sherriff et al., 2009). The evolution of the runoff from the tailings is shown in the composition of the water draining the tailings from a stream at site (RW1) and a major E–W trench at site (RW2) (Fig. 2). The surface drainage at RW1 has become progressively more acidic with the pH decreasing from 3.1 in 2003 to 2.6 by 2009 (Table 4). There was a progressive increase from 2003 to 2007 of 3190 to 9030 mg/L SO2 4 , 200 to 1240 mg/L Fe, 0.5 to 34.4 mg/L Cu but then
1838
B.L. Sherriff et al. / Applied Geochemistry 26 (2011) 1832–1842
Fig. 6. Concentrations of Fe, Zn and Cu with depth in the oxidized tailings of Cell 2 in 2004.
Table 2 Results of sequential extraction from solid samples of the tailings; weighted means as percentages of total concentration. Year
Water soluble
Exchangeable
unoxidized unoxidized oxidized oxidized
2004 2005 2004 2005
0.5 0.5 0.5 0.5
0.4 0.2 0.2 0.3
unoxidized unoxidized oxidized oxidized
2004 2005 2004 2005
23 25 39 20
Copper Cell 3, unoxidized Cell 3, unoxidized Cell 2, oxidized Cell 2, oxidized
2004 2005 2004 2005
13 21 28 13
Iron Cell 3, Cell 3, Cell 2, Cell 2, Zinc Cell 3, Cell 3, Cell 2, Cell 2,
Crystalline Fe-hydroxide
Residual
6 5 5 6
11 23 19 25
83 72 76 68
4 4 8 5
2 2 3 3
14 22 22 29
57 48 30 44
10 6 9 7
13 10 12 9
23 39 39 40
41 24 13 31
a slight decrease to 5000 mg/L SO2 4 , 770 mg/L Fe and 29.7 mg/L in 2009. Zinc increased from 14.9 to 300 mg/L from 2003 to 2004 but decreased progressively to 64.3 mg/L in 2009. It appears that the drainage from the tailings surface sampled at RW1 had reached a steady composition by 2007 and improved in quality by 2009. These results are based on one set of samples per year, and there may have been a dilution effect in 2009 due to high precipitation. Metal and SO2 4 concentrations in water from the trench at RW2 also show an increase with time but it is much less systematic than for RW1 (Table 4). There was an initial pulse of SO2 4 (9890 mg/L), Cu (23 mg/L) and Zn (228 mg/L) in 2003 which decreased to approximately 5000 mg/L SO2 4 and less than 5 mg/L Cu by 2004. The highest measured values of 1220 mg/L Fe and 247 mg/L Zn occurred during the rain event in June 2004. Although the pH seems to have stabilized at about 2.6, SO2 and metal values increased 4 again in 2009 to 12,100 mg/L SO2 4 , 3700 mg/L Fe, 147 mg/L Zn and 6.7 mg/L Cu (Table 4). In June 2004, water was taken from the trench in tailings Cell 3 during a precipitation event (RW2 June 2004 rain) with a second sample from the same location 24 h after the rain ceased (RW2 June after rain) (Table 4). Iron, Zn, Cu, Mg and SO2 4 concentrations
Amorphous Fe-hydroxide
were higher during precipitation, while Na, K and Ca were higher after the rain ceased. Ponding of water on the surface of the tailings during rain events allows pools of low pH, metal laden water, that were previously isolated on the surface, to join and flow into the drainage trenches. Also evaporite material on the surface of the tailings, which can be dissolved during rain events, would cause an increase in dissolved metals and SO2 4 . Alkali metal concentrations became more dilute during precipitation as they behave as conservative components (Table 4). Since the inception of the Ruttan Mine, Ruttan Lake has been influenced by contaminated water run-off from the mine site as it was classified as part of the tailings management impoundment. Water entering the Lake from the tailings impoundment (RW4) had a stable pH of 2.9 ± 0.1 until 2007 but this decreased to 2.6 in 2009 (Table 5). During this time there has been a considerable variation in the concentration of SO2 (1660–9610 mg/L), Cu 4 (1.6–16.7 mg/L), Fe (125–2410 mg/L) and Zn (29.3–110 mg/L). The highest values are all from 2009, indicating a significant decrease in the quality of water entering Ruttan Lake. The buffering and dilution capacity of Ruttan Lake is shown by a comparison of the composition of the input water at RW4 with the water from
B.L. Sherriff et al. / Applied Geochemistry 26 (2011) 1832–1842
1839
Fig. 7. Cell 3 pore water composition with depth. The solid line represents values from 2004 and the dashed line 2005. All concentrations are in mg/L. The horizontal lines with triangles represent the water table: solid line for 2004, dashed line for 2005.
Fig. 8. Cell 2 pore water composition with depth; the solid line for 2004 and the dashed line for 2005. The water table was below the section in 2004 and at 4 m depth in 2005.
the west side of the Lake at RW8 and RW9 (Table 5). In June 2004, the surface water composition at all three sites are similar, but in September 2005 when the concentrations of SO2 4 , Cu, Fe and Zn
at RW4 were at least double the 2004 values, they stayed constant at RW8 and RW9 (Table 5). At the outlet from Ruttan Lake toward the open pit (RW3), SO2 increased from 2003 to 2007 4
1840
B.L. Sherriff et al. / Applied Geochemistry 26 (2011) 1832–1842 3þ
AlðOHÞ3ðsÞ þ 3Hþ ¼ Al
Fig. 9. Gullies forming at the side of trenches on the tailings in Cell 1 exposing unoxidized tailings.
(2520–3890 mg/L) and pH decreased from 2.9 to 2.7, however, Fe and Zn decreased from 878 to 245 mg/L and 169 to 38 mg/L, respectively, whereas Cu has stayed constant at 3.2 mg/L. A comparison of water from June to October 2004 indicates an increase in concentration of Fe, Cu, Zn and SO2 4 at RW4 and RW3 during the hot dry summer of 2004. 4.4. Mobility of metals The unoxidized tailings have average concentrations of 2200 ppm Zn and 530 ppm Cu, which can be mobilized by oxidation from sphalerite and chalcopyrite, respectively. The remnant sphalerite oxidizes and produces up to 10,000 mg/L Zn in nearsurface pore water but Cu concentration is at least a factor of 10 lower than for Zn. The major acid generating minerals, pyrite and pyrrhotite, can release high concentrations of Fe and SO2 to the pore water 4 (Blowes et al., 2003a). Although carbonate minerals were described in the original ore mineralogy, they are of insufficient abundance in the tailings to be visible in thin sections or to maintain a neutral pH, as in Eq. (1), once acid generation has commenced.
CaCO3 þ Hþ ¼ Ca2þ þ HCO3
ð1Þ
Aluminum liberated by the weathering of aluminosilicate gangue minerals, and Fe3+ will react with water to form Al and Fe oxy-hydroxide secondary phases. A decrease in pH below 6 will cause dissolution of these phases as shown in Eqs. (2) and (3) buffering the water at a pH of 4.3 for Al(OH)3 or 2–2.5 for Fe oxyhydroxides (Blowes and Jambor, 1990; Johnson et al., 2000; Blowes et al., 2003a; Sherriff et al., 2009).
þ 3H2 O
ð2Þ
FeOOHðsÞ þ 3Hþ ¼ Fe3þ þ 2H2 O
ð3Þ
These three buffers (Eqs. (1)–(3)) are controlling the chemistry of the pore, ground and surface water at Ruttan. In 2004, pore and groundwater at 0.8 m below grade generally had near neutral pH, where it is buffered by the dissolution of calcite (Figs. 7 and 8, Table 4). By 2005, the groundwater and pore water at about 1 m depth have stabilized at a pH of 4–4.5. This indicates the complete depletion of carbonate capacity and control of pH by the dissolution of Al(OH)3. Breakthrough of this acidic front correlates with increases in concentration of Fe, Al and Zn in solution. While Al hydroxides dissolve, Fe(II) can be oxidized and precipitated in the form of Fe(III) oxy-hydroxides and hydroxy-sulfates at pH below 4.5. Most surface water including Ruttan Lake has stabilized at pH 2–3 due to the Fe oxy-hydroxide buffers. Sequential extraction data showed that up to 50% of the total Fe is present in the form of Fe-oxide/hydroxide minerals in near surface samples of tailings. Pore water is in direct contact with sulfide grains and acts as the source of water for oxidation reactions and a conduit for metals and protons released from these oxidation reactions. In 2004, maximum concentrations of metals and SO2 were measured in the 4 upper 10 cm of the unoxidized tailings and rapidly decreased over the next 20 cm depth. Maximum concentrations were higher in 2005 and occurred between 20 and 30 cm depth corresponding to observations that oxidation was starting along fissures in the upper 10 cm of Cell 2 in 2004, whereas it was more pervasive and down to 20 cm depth in 2005. Once liberated from sphalerite, Zn is very mobile as a major secondary fraction of Zn was shown by sequential extraction to be in a water soluble form. Only a minor amount was found to be bound to Fe oxy-hydroxides due to its limited adsorption at pH values below 6 (Dzombak and Morel, 1990). The initial high values of metals and SO2 4 in groundwater could have been due to the construction of the trench in 2002–2003 or to the oxidation of the very fine grained material from unoxidized tailings. The increase in metal values in groundwater in 2009 may be due to the formation of extensive gullies at the side of the trenches, which are exposing unoxidized tailings to air causing oxidation with the accompanying acidification and release of metals (Fig. 9). Ruttan Lake acts as a reservoir for both contaminated water from the tailings and meteoric water which could dilute the water coming off the Ruttan tailings. In 2004 and 2005, the composition of the Lake was found to be more or less homogeneous from the east discharge (RW4) to the west shore (RW8 and RW9) and the pH remained constant at 2.8. By 2009, the pH of the water draining from the Lake to the open pit (RW3) had decreased slightly to 2.6 but the SO2 4 and metal content had not increased despite a considerable increase of the input to the Lake (RW4). As the unoxidized tailings in Cell 3 continue to de-water and the water table drops, O2 will continue to diffuse deeper into the tailings creating a deepening zone of oxidation. Modeling shows that sulfide-rich tailings with a deep water table can oxidize for centuries (Blowes and Jambor, 1990). Blowes et al. (2003a) defined a model of the geochemical evolution of tailings which illustrates the location and movement of metals as they oxidize. A comparison of the conditions at Ruttan with their conceptual model suggests that the tailings were still in the early to moderate stage of oxidation in 2005. In both Cells 2 and 3, there is a minor depletion in solid metal concentrations near the surface and a peak in dissolved metal concentrations just below the surface. Although there is a zone of lower metal content of the pore water near the surface of Cell 2 in 2005, it is probably due to dilution from heavy rainfall
1841
B.L. Sherriff et al. / Applied Geochemistry 26 (2011) 1832–1842 Table 3 Composition of Ruttan groundwater (mg/L). Depth to top of screen
RBH4
RBH4
RBH4
RBH5
RBH5
RBH5
RBH5
RBH7
RBH7
RBH7
RBH7
Date
September2004
2.6 m July2007
July2009
September2004
1.1 m September2005
July2007
July2009
September2004
0.8 m September2005
July2007
July2009
pH SO4 Cu Fe Zn Mn Al Ca Mg Na K Cd Co Cr Ni
9.6 2430 <0.05 0.02 <0.10 nd 0.22 523 2.63 83.9 43.8 <0.50 <0.05 nd <0.10
7.1 1310 0.0055 1.36 0.078 0.14 0.65 457 5.16 74.1 43.0 <0.0002 0.003 <0.001 0.005
7.0 1500 0.0062 9.71 0.155 0.018 0.554 460 1.34 92.0 57.8 <0.0002 0.002 0.002 0.007
6.9 1770 <0.05 1.09 0.18 0.43 <0.05 570 10.4 68.3 45.6 <0.50 <0.05 nd <0.10
3.5 3790 0.0086 803 1.26 3.79 1.76 447 176 67.5 124 0.001 0.001 <0.001 0.012
4.1 4460 0.0125 840 4.05 10.1 0.65 375 427 53.1 98.1 <0.0002 0.002 <0.001 0.002
4.2 31,300 0.0205 7290 1.34 36.8 3.12 191 1660 36.5 62.6 <0.0002 <0.001 0.003 0.005
7.5 1690 <0.05 2.60 1.26 0.57 <0.05 540 8.18 57.49 38.4 <0.50 <0.05 nd <0.10
3.5 11,300 0.0119 3770 106 15.1 8.6 269 614 46.9 97.0 0.005 0.006 <0.001 0.022
4.3 5710 0.0293 1670 0.784 12 1.95 320 404 37.5 125.0 0.006 0.004 <0.001 0.003
4.2 33,100 0.0261 11,900 155 40.1 199 147 1730 18.9 58.9 0.012 0.045 0.027 0.049
Table 4 Tailings’ surface water composition (mg/L). RW1
pH SO4 Cu Fe Zn Mn Al Ca Mg Na K Cd Co Cr Ni
RW2
September 2003
June 2004
July 2007
July 2009
September 2003
June 2004 rain
June after rain
October 2004
July 2007
July 2009
3.1 3190 0.5 200 14.9 2.6 39.9 573 76 86.2 50.6 0.02 0.07 0.01 0.08
2.9 6880 27.3 1170 300 12.3 491 236 443 8.8 8.6 0.51 0.86 nd 0.64
2.7 9030 34.4 1240 169 11.4 552 207 419 6.1 3.16 0.31 1.80 0.06 0.65
2.6 5000 29.7 770 64.3 4.6 350 111 270 3.7 2.01 0.15 0.83 0.08 0.29
2.8 9890 23.0 598 228 15.8 641 297 535 7.6 10.6 0.31 1.68 0.03 0.81
3.4 4800 4.7 1220 247 13.8 145 335 393 44.3 24.3 0.31 0.08 nd 0.46
3.5 3960 1.7 844 139 9.6 90.9 414 261 66.4 39.6 0.16 0.07 nd 0.31
2.5 5030 <0.05 415 2.48 nd 0.5 514 106 81.7 60.8 0.00 0.00 nd 0.00
2.8 5070 0.3 1580 35.1 7.0 50.9 413 214 61.6 51.6 0.02 0.11 0.02 0.07
2.6 12,100 6.7 3700 147 10.8 457 248 622 24.6 20.8 0.23 0.66 0.13 0.30
RW1: Stream. RW2: Trench.
Table 5 Lake water composition (mg/L) 2004 and 2005. RW3
pH SO4 Cu Fe Zn Mn Al Ca Mg Na K Cd Co Cr Ni
RW4
RW8
RW9
June 2004
October 2004
September 2005
July 2007
July 2009
September 2003
June 2004
October 2004
September 2005
July 2007
July 2009
June 2004
September 2005
June 2004
September 2005
2.9 2520 3.2 878 63.2 6.5 103 294 286 23.8 9.3 0.08 0.04 nd 0.17
3.2 11,000 21.7 1510 169 nd 529 312 492 26.1 15.4 0.34 1.12 nd 0.57
2.8 2460 4.4 161 50.1 3.4 101 205 152 13.2 4.96 0.11 0.26 0.03 0.18
2.7 3890 3.2 245 38 3.8 115 180 147 12.2 4.18 0.06 0.25 0.03 0.15
2.6 2040 2.2 222 21.8 2.6 85 137 112 8.9 2.54 0.04 0.16 0.03 0.09
2.8 4970 1.6 125 34.6 3.3 65.1 381 123 20.5 16.5 0.05 0.17 0.02 0.16
2.9 1660 1.7 129 29.3 3.0 60.1 286 122 15.9 10.2 0.04 0.09 nd 0.12
3.1 7920 3.1 1140 81.8 nd 162 374 273 48.2 27.9 0.11 0.29 nd 0.20
2.8 3290 3.4 470 50.4 3.7 127 210 185 14.3 4.88 0.09 0.26 0.05 0.16
2.8 6260 11.0 1520 78.3 8.4 228 352 279 45.4 36.6 0.11 0.54 0.03 0.25
2.6 9610 16.7 2410 110 8.9 406 205 484 17.5 14.9 0.20 0.71 0.11 0.31
2.7 1700 1.4 93 28.4 nd 52.7 273 109 17.1 9.4 0.04 0.09 nd 0.00
2.9 1470 1.4 80.1 20.3 nd 47.8 153 88 11.5 3.69 0.04 0.11 0.02 0.08
2.9 1990 1.5 108 28.4 nd 55.3 286 114 17.3 9.5 0.04 0.10 nd 0.00
2.8 2140 2.3 131 29.6 nd 70.2 216 128 15.2 4.29 0.06 0.17 0.03 0.12
RW3: Outflow from Ruttan Lake to open pit. RW4: Inflow to Ruttan Lake from tailings.
1842
B.L. Sherriff et al. / Applied Geochemistry 26 (2011) 1832–1842
rather than a reduction in the reactivity of sulfides. Pore water metal concentrations peak within the first meter of all the tailings, suggesting a juvenile oxidation stage. This is not surprising for Cell 3, as these tailings were very recently introduced to oxidizing conditions. As the tailings are greater than 10 m thick they will be oxidizing and releasing metals for many years. The concentration of metals in Ruttan Lake may increase and pH levels decrease as oxidation of the tailings proceeds, however, the constant flow into the open pit and partial recharge by meteoric waters has a stabilizing effect so that the Lake is unlikely to become as acidic as the ponds on the tailings surface. 5. Conclusions 1. The very fine grained solids which had accumulated in Cell 3, the polishing pond, were immediately oxidized on dewatering possibly producing an initial pulse of sulfates and some metals in the trench and input water into Ruttan Lake. Iron mobilized in these reactions was oxidized to Fe(III) and deposited as Fe oxy-hydroxides surrounding mineral grains. 2. The geochemical reactions occurring during dewatering are dominated by the abundant pyrite and pyrrhotite but with Zn being mobilized from the sphalerite remaining after beneficiation of the ore. Little Cu is being mobilized partly due to the armouring of chalcopyrite by quartz and also by preferential absorption of Cu on secondary Fe minerals. 3. On closure in 2002, the alkaline pH of the tailings water was buffered by dissolution of the scarce natural carbonates and lime added during processing. Within 2 years of mine closure, an acidic front with pH 4.5, buffered by the dissolution of Al(OH)3, had broken through and reached the water table in most of the tailings. The surface runoff and Ruttan Lake water has stabilized at pH values of 2.5 due to the dissolution of Fe oxy-hydroxides. 4. Acidic metal-rich surface and shallow pore water evaporates to form Zn and Fe hydroxy-sulfate evaporite minerals. Metals that are temporarily attenuated on the surface of the tailings are released during periods of rain or snow melt. 5. The composition of Ruttan Lake and the discharge to the open pit seems to have stabilized despite orders of magnitude increases in the loadings of SO2 4 and metals from the tailings.
Acknowledgements Funding for this project was obtained from the Manitoba Sustainable Development Fund and from the Natural Science and Engineering Research Council of Canada. We acknowledge the assistance of Dr. Kathleen Londry, Dr. Tricia Stadnyk, Stephanie Simpson and the late Greg Morden and Sergio Meijia for field and laboratory assistance. We especially acknowledge the assistance of Ben Edirmanasinghe and Ernie Armitt from the Manitoba Department of Energy and mines, and the late Frank Bloodworth, HBM&S for site access and assistance. We acknowledge
anonymous reviewers for thorough helpful reviews of an earlier version of this manuscript. References Ballirano, P., Bellatreccia, F., Grubessi, O., 2003. New crystal-chemical and structural data of dietrichite, ideally ZnAl2(SO4)422H2O, a member of the halotrichite group. Eur. J. Mineral. 15, 1043–1049. Barrie, C.T., Taylor, C., Ames, D.E., 2005. Geology and metal contents of the Ruttan volcanogenic massive sulfide deposit, northern Manitoba, Canada. Mineral. Deposita 39, 795–812. Blowes, D.W., Jambor, J.L., 1990. The pore-water geochemistry and the mineralogy of the vadose zone must be some sulfide tailings, Waite Amulet, Quebec, Canada. Appl. Geochem. 5, 327–346. Blowes, D.W., Ptacek, C.J., 1994. Acid-neutralization mechanisms in inactive mine tailings. In: Jambor, J.L., Blowes, D.W. (Eds.), Mineral. Assoc. Can. Short Course Handbook on Environmental Geochemistry of Sulfide Mine Wastes, vol. 22, pp. 271–292. Blowes, D.W., Ptacek, C.J., Jambor, J.L., Weisner, C.G., 2003a. The geochemistry of acid mine drainage. In: Sherwood Lollar, B. (Ed.), Environmental Geochemistry. Holland, H.D., Turekian, K.K. (Exec. Eds.), Treatise on Geochemistry, vol. 9. Elsevier, Amsterdam, pp. 149–204. Blowes, D.W., Ptacek, C.J., Jambor, J.L., 2003b. Mill tailings: hydrolgeology and geochemistry. In: Jambor, J.L., Blowes, D.W., Richie, A.J.M. (Eds.), Environmental Aspects of Mine Wastes. Mineral. Assoc. Can. Short Course Series 31, pp. 95– 116. Dold, B., 2003. Speciation of the most soluble phases in a sequential extraction procedure adapted for geochemical studies of copper sulfide mine waste. J. Geochem. Explor. 80, 55–68. Dzombak, D.A., Morel, F.M.M., 1990. Surface Complexation Modeling: Hydrous Ferric Oxide. Wiley-Interscience, New York. Etcheverry, D.J., 2008. Spatial and Temporal Variations in the Ruttan Mine Tailings, Leaf Rapids, Manitoba, Canada. M.Sc. Thesis, Univ. Manitoba. Hall, G.E.M., Gauthier, G., Pelchat, J.C., Pelchat, P., Vaive, J.E., 1996. Application of a sequential extraction scheme to ten geological certified reference materials for the determination of 20 elements. J. Anal. Atom. Spectrom. 11, 787–796. Jambor, J.L., Nordstrom, D.K., Alpers, C.N., 2000. Metal-sulfate salts from sulfide mineral oxidation. In: Alpers, C.N., Jambor, J.L., Nordstrom, D.K. (Eds.), Sulfate Minerals—Crystallography, Geochemistry, and Environmental Significance. Reviews in Mineralogy and Geochemistry, vol. 40, pp. 303–350. Johnson, R.H., Blowes, D.W., Robertson, W.D., Jambor, J.L., 2000. The hydrochemistry of the Nickel Rim mine tailings impoundment, Sudbury, Ontario. J. Contam. Hydrol. 41, 49–80. Liu, R., Wolfe, A., Dzombak, D., Stewart, B., Capo, R., 2008. Comparison of dissolution under oxic acid drainage conditions for eight sedimentary and hydrothermal pyrite samples. Environ. Geol. 56, 171–182. Peretyazhko, T., Zachara, J.M., Boily, J.-F., Xia, Y., Gassman, P.L., Arey, B.W., Burgos, W.D., 2009. Mineralogical transformations controlling acid mine drainage chemistry. Chem. Geol. 262, 169–178. Peterson, R.C., 2003. The relationship between Cu content and distortion in the atomic structure of melanterite from the Richmond mine, Iron Mountain, California. Can. Mineral. 41, 937–949. Petrunic, B.M., Al, T.A., 2005. Mineral/water interactions in tailings from a tungsten mine, Mount Pleasant, New Brunswick. Geochem. Cosmochim. Acta 69, 2469– 2483. Sherriff, B.L., Ferguson, I., Gupton, M.W., VanGulck, J.F., Sidenko, N.V., Priscu, C., Marco Pérez-Flores, M., Gómez-Treviño, E., 2009. A geophysical and geotechnical study to determine the hydrological regime of the Central Manitoba gold mine tailings deposit. Can. Geotech. J. 46, 69–80. Sherriff, B.L., Sidenko, N.V., Salzsauler, K.A., 2007. Differential settling and geochemical evolution of tailings’ surface water at the Central Manitoba Gold Mine. Appl. Geochem. 22, 342–356. Sidenko, N.V., Sherriff, B.L., 2005. The attenuation of Ni, Zn, and Cu, by secondary Fe phases of different crystallinity from surface and groundwater of two sulfide mine tailings in Manitoba, Canada. Appl. Geochem. 20, 1180–1194. Stantec Consulting Limited, 2003. Hudson Bay Mining and Smelting First Environmental Effects Monitoring Study and Design. A Report to Hudson Bay Mining and Smelting Company Ltd. by Stantec Consulting Limited, Guelph, Ontario, Canada.
Contents lists available at ScienceDirect
Applied Geochemistry journal homepage: www.elsevier.com/locate/apgeochem
Spatial and temporal evolution of Cu–Zn mine tailings during dewatering Barbara L. Sherriff a,⇑, D. Jared Etcheverry a, Nikolay V. Sidenko a,b, Jamie Van Gulck c a
Department of Geological Sciences, University of Manitoba, Winnipeg, MB, Canada R3T 2N2 Stantec Consulting Ltd., Winnipeg, Manitoba, Canada R3C 3R6 c Arktis Solutions Inc., 117 Loutitt St., Yellowknife, NT, Canada X1A 3M2 b
a r t i c l e
i n f o
Article history: Available online 17 June 2011
a b s t r a c t The Ruttan Cu–Zn mine produced about 50 mT of fine-grained tailings over 30 a. Since the closure of the mine in 2002, the tailings have been systematically dewatered through trenches draining into the open pit and underground workings. This study evaluated the evolution of tailings that were underwater until 2002, and also tailings that had been exposed to oxidizing conditions for more than 20 a. Acid generation is dominated by the oxidization of the abundant pyrite and pyrrhotite comprising 25 wt.% of tailings with Zn being mobilized from the sphalerite remaining after beneficiation of the ore. Little Cu is being mobilized partly due to the armouring of remnant chalcopyrite by primary quartz, and also by preferential absorption of Cu rather than Zn on secondary Fe oxy-hydroxides. A very fine grained fraction of unoxidized sulfides is a likely cause of an initial pulse of acid generation, and metal and SO2 4 release occurring at the onset of dewatering. Metals are temporarily attenuated from waters associated with the tailings, either absorbed on Fe oxy-hydroxide precipitates or as evaporite hydroxy sulfate minerals at the surface of the tailings. While some secondary phases are stable, evaporites are only temporary metal sinks as they redissolve in wet weather conditions. Trace amounts of calcite provide little buffering capacity resulting in rapid acidification of pore and surface water. The pH of pore water and shallow groundwater decreased first to the Al oxy-hydroxide buffer at 4.5 and then stabilized at values of 2–3 being controlled predominantly by the dissolution of solid Fe oxy-hydroxides. The metal contents of the ground and surface water are still increasing but the Ruttan Lake reservoir that receives drainage water from the tailings is maintaining a constant composition. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction The mining of base metals produces large quantities of finegrained sulfide tailings. If the tailings are exposed to the atmosphere, sulfide oxidation can produce acid (Blowes et al., 2003a). If not neutralized by carbonates, this acid will release metals and can contaminate ground and surface water (e.g. Blowes and Ptacek, 1994; Johnson et al., 2000; Blowes et al., 2003b). Although the basic reactions have been well studied both in the field (e.g. Petrunic and Al, 2005; Peretyazhko et al., 2009) and the laboratory (e.g. Liu et al., 2008), each mine site is unique in terms of geology, climate, environmental conditions and the proportions of acid producing and neutralizing minerals. The dewatering of the tailings at Ruttan Cu–Zn Mine, Manitoba, Canada, after closure in 2002, provided an opportunity to observe potentially acid generating tailings during the first years of dewatering and oxidation in a subarctic-climate. The objectives of this study were to observe changes in mineralogy and geochemistry during oxidation of the previously submerged tailings, to predict ⇑ Corresponding author. E-mail address: [email protected] (B.L. Sherriff). 0883-2927/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.apgeochem.2011.06.007
their evolution from characterization of already oxidized tailings and to determine the geochemical evolution of Ruttan Lake. Selected analytical data are presented here. Full analytical results are available in Etcheverry (2008). 2. Study site Ruttan Mine-site is located approximately 950 km north of Winnipeg, 30 km east of the community of Leaf Rapids, Manitoba (Fig. 1). From 1973 until 2002, Cu and Zn ore were extracted from volcaniclastic and siliciclastic sequences of the Rusty Lake Greenstone Belt first by Sherritt Gordon Mines Ltd. and after 1987 by Hudson Bay Mining and Smelting Co. Ltd. (HBM&S) (Stantec Consulting Ltd., 2003). The massive-sulfide ore lenses are hosted by tholeiitic, felsic and intermediate volcanic and volcaniclastic rocks. Metamorphism to upper greenschist and lower amphibolite facies is characterized by the presence of cordierite (Mg2Al4Si5O18), almandine ðFe2þ andalusite (Al2SiO5), sillimanite 3 Al2 ðSiO4 Þ3 Þ, (Al2SiO5), biotite (K(Mg,Fe)3[AlSi3O10(OH,F)2), staurolite ((Fe,Mg)2Al9(Si,Al)4O20(O,OH)4), anthophylite (Mg7(Si8O22)(OH)2), and talc (Mg3Si4O10(OH)2). Other gangue minerals include muscovite
B.L. Sherriff et al. / Applied Geochemistry 26 (2011) 1832–1842
1833
Fig. 1. The location of Ruttan Mine and mine site plan showing tailings management cells and drainage trenches from these cells to Ruttan Lake.
(KAl3Si3O10(OH,F)2), chlorite (KAl3Si3O10(OH,F)2), augite ((Ca,Na)(Mg,Fe,Al,Ti)(Si,Al)2O6), quartz (SiO2), anhydrite (CaSO4) and the carbonate minerals, siderite (FeCO3), dolomite (CaMg(CO3)2), ankerite (Ca(Fe,Mg,Mn)(CO3)2 and calcite (CaCO3). Primary sulfide phases consist of pyrite (FeS2), pyrrhotite (Fe(1x)S), chalcopyrite (FeCuS2), sphalerite (ZnS), and galena (PbS) with minor tetrahedrite ðCu9 Fe2þ 3 Sb4 S13 Þ (Barrie et al., 2005). Flotation refinement of the Ruttan ore removed sphalerite and chalcopyrite, producing over 50 million tonnes of fine-grained sulfide-rich tailings which were discharged into Cell 1 and Cell 2 of the tailings impoundment. Cell 3 acted as a polishing pond before the water was discharged into Ruttan Lake (Fig. 1). Most of these tailings remained submerged after deposition but the northern portion of Cell 2 had been exposed to the atmosphere and oxidized for about 20 a (Frank Bloodworth, HBM&S, pers. comm.). During operation of the mine, lime was added both to Ruttan Lake and at a spillway constructed at the outlet to Brehaut Lake (Stantec Consulting Ltd., 2003). In the final stages of remedial activity by HBM&S, this weir was destroyed and the outlet sealed. To improve the long term stability of the tailings dams, tailings ponds are being systematically dewatered through a network of drainage trenches about 3–4 m deep and 3 m wide (Figs. 2 and 3). The drainage water is being directed into Ruttan Lake. The entire onsite watershed including Ruttan Lake, the mill pond and drainage from the waste rock piles was redirected and now drains into the open pit and then to the underground workings (Fig. 2). The drainage of the watershed surrounding the mine was rerouted so that natural surface runoff skirts the site.
The subarctic climate at Leaf Rapids shows annual temperatures from 47 °C to 35 °C with snow cover from October to May and the groundwater often frozen from October to July. This results in a very short period each year for reactions to occur in the tailings at a significant rate and for the groundwater to drain through the trenches.
3. Methods 3.1. Sample collection Sampling locations for solid tailings (RTP2, RTP4) were chosen to be representative of tailings that had been oxidizing over a period of about 20 a (Cell 2) and of those that were water saturated and unoxidized until 2003 (Cell 3) (Fig. 2). In 2004, solid tailings were collected at 10 cm intervals from pits to a depth of <1 m, and by auger beneath the pit at 0.5 m intervals. Samples for thin sections were collected by pressing 4 6 1 cm Al boxes into the tailings and removing them with their contents. Vertical sections of tailings were collected in PVC pipes (30 cm long with 7.6 cm I.D) sealed with PVC end-caps. Precipitates were collected from the bottom of ponds and evaporites from desiccation cracks on the tailings surface. All solid samples were frozen at the end of each field day. Surface water was collected periodically between 2003 and 2009 from several sites including a stream (RW1) and trench (RW2) draining Cell 3, from the stream draining into Ruttan Lake (RW4), from the west side of Ruttan Lake (RW8 and RW9) and from
1834
B.L. Sherriff et al. / Applied Geochemistry 26 (2011) 1832–1842
Fig. 2. Ruttan Mine site after closure in 2002 showing elevation relative to the surface of Ruttan Lake (D), and the drainage plan after mine closure. Solid (RTP), surface water (RW) and groundwater (RBH) sampling sites are shown.
the outlet from Ruttan Lake to the open pit (RW3) (Fig. 2). The Lake samples were collected to study the dilution effect of the Lake and its evolution as the tailings were dewatered. A total of 12 monitoring wells were installed in Tailings Cells 1 and 3 in 2003 and 2004 (Etcheverry, 2008). By 2009, many of the wells had dried up, been broken or buried by the movement of the tailings or become contaminated with surface water or sediments, therefore, longitudinal studies are presented on only three wells, RBH4, RBH5 and RBH7. The locations of these wells and surface sampling sites are shown in Fig. 2. The wells (RBH 4, 5, 6) were constructed with solid 50.8 mm diameter PVC casing (2.70, 1.09, 1.46 m in length) and No. 10 factory slot well screens (0.40, 2.54, 1.98 m) to the borehole base. The annulus was backfilled with clean silica sand from the base to just above the solid/screen join and with granular bentonite to the surface. Groundwater was obtained after each well had been purged. All water samples were passed through 0.45 lm filters and divided into two aliquots. The one for cation analysis was acidified with 0.01 mL of 5.6 M HNO3 per 1 mL of sample whereas the other for anion analyses was left unacidified. Eh and pH were measured immediately with an AP-62 pH/Eh meter using a Thermo Orion ORP-97-87 Ag/AgCl electrode (for Eh) and an Accumet 13-620AP50 Ag/AgCl electrode (for pH). 3.2. Laboratory methods Samples collected for mineralogy were impregnated with epoxy, before being made into polished thin sections which were prepared without the use of water to preserve soluble mineral phases. These samples were analyzed under transmitted and reflected optical microscopy and with a Cambridge Instruments Scanning Electron Microscope (SEM) equipped with an Energy
Dispersive X-ray Spectrometer (EDX) for elemental analysis. Mineral phases in evaporites were manually separated under a binocular microscope for analysis by SEM and by powder X-ray diffraction (XRPD), using a Philips PW1710 XRPD system and MDI Datascan/Jade data processing software. Precipitates, which formed in unacidified water samples stored at room temperature for 2 months, were filtered, dried, and analysed by XRPD and SEM. Bulk samples of tailings obtained from pits and auger holes were dissolved in HF/HNO3 before geochemical analysis at the University of Manitoba using a Varian Liberty 200 Inductively Coupled Plasma Optical Emission Spectrometer (ICP-OES) for cations referenced to ICP standard solutions (SCP Science). Samples of tailings underwent a sequential extraction process as outlined in Table 1 with a sequence of water soluble, adsorbed/exchangeable/carbonate, amorphous Fe oxy-hydroxide and crystalline Fe oxide fractions. The residual sulfide/silicate phase was inferred from a comparison of sequential extraction results with the bulk chemistry. Pore water was extracted from 10 cm intervals of frozen cores of tailings. The samples were thawed in a N2-filled glove bag, placed in the sample chamber of a press and compressed with a hydraulic press operating at 5 tons of pressure. The resultant pore water was passed through a 0.45 lm filter, separated into aliquots and preserved as described for the field samples. Cations in acidified water samples were measured in 2003 and 2004 using a Varian Liberty 200 ICP-OES at the University of Manitoba, and from 2005 by Inductively Coupled Plasma Mass Spectrometry (ICP-MS) at Envirotest (now ALS) Laboratories, Winnipeg. Anions in samples from 2003 and 2004 were analyzed in the Department of Microbiology, University of Manitoba using a Dionex™ ion-chromatography system and in samples from 2005 by ICPMS at Envirotest Laboratories.
1835
B.L. Sherriff et al. / Applied Geochemistry 26 (2011) 1832–1842
Fig. 3. Dewatering trench cutting tailings Cell 1 in (a) 2003 (b) 2005.
Table 1 Sequential extraction procedures. Target phase
Reagent
Method
Reference
Water soluble fraction Adsorbed/ exchangeable/ carbonate
Deionized H2O
Dold (2003) Hall et al. (1996)
Amorphous Fe oxyhydroxide
0.2 M NH4-oxalate 0.1 M (NH2)2C2O4
Crystalline Fe oxide
1.0 M NH2OHHCl 25% CH3COOH
1.0 g of sample into 20 mL deionized H2O, vortex, shake for 1 h. Centrifuge, remove supernatant, preserve with 5.6 M HNO3 and analyze Take the residue from the previous leach and add 20 mL 1.0 M CH3COONa; to pH 5.0 with HCl. Vortex for 5–10 s and place in shaker for 6 h. Centrifuge for 10 min at 2800 rpm and decant supernatant into a labeled test-tube and preserve with 5.6 M HNO3. Rinse residue with 5 mL of water, vortex and centrifuge again: add supernatant rinses to the test-tube. Carry out a second 20 mL 1.0 M CH3COONa leach of the residue, repeating steps 2 and 3 add supernatant to the test-tube Take the residue from the previous leach and add 20 mL NH4-oxalate; to pH 3 with 0.1 M oxalic acid. Vortex, shake on horizontal shaker for 1 h in dark at room temp. Centrifuge the sample for 10 min, and decant the supernatant into a storage container. Acidify the supernatant with HNO3. Wash the sample with deionized water, shake to mix and centrifuge for 30 min. Add rinse water to sample Take the residue from the previous leach and add 20 mL of 1.0 M NH2OHHCl in 25% CH3COOH, vortex from 5–10 s. Place in water oven at 90 °C for 3 h (tightly capped), vortex every 20 min. Centrifuge for 10 min and decant supernatant fluid into a labeled test-tube. Carry out a second 1.0 M NH2OHHCl leach of the residue but heat for only 1.5 h, add supernatant from this step to that of the last, preserve with 5.6 M HNO3 and analyze
1.0 M CH3COONa 0.25 M HCl
4. Results and discussion 4.1. Site observations As dewatering of Ruttan tailings Cells 1 and 3 progressed from 2003, the surface area of unoxidized tailings exposed to the atmosphere increased; however, high rainfall between April and October resulted in very wet conditions reducing the visible effects of dewatering in 2005.
Dold (2003)
Hall et al. (1996)
In 2003, when the drainage trenches were first excavated the surface of the tailings in Cell 1 and Cell 3 was dark gray and completely unoxidized (Fig. 3a). Patchy surface oxidation in 2004 was limited to the upper 1 cm of the tailings and constrained to areas of higher elevation and to 1–2 cm halos along vertical and horizontal cracks. Below this, the tailings were unoxidized. By 2005, the entire surface of Cell 1 was oxidized, with the oxidized layer being up to 15 cm thick (Fig. 3b) and the oxidation much more pervasive. Sidenko and Sherriff (2005) showed that Fe precipitates from
1836
B.L. Sherriff et al. / Applied Geochemistry 26 (2011) 1832–1842
surface water at Ruttan formed schwertmannite ðFe3þ 16 O16 ðOHÞ12 ðSO4Þ2 Þ and jarosite (KFe3(OH)6(SO4)2). These precipitates caused the immediate change in color of Cell 3 surface from gray in 2004 to brown by 2005. A layer of white evaporite minerals that covered the surface of the tailings in 2003 and 2004 was not present in 2005 due to the wet conditions. 4.2. Mineralogy During beneficiation, ore is crushed to the optimal size for extraction of the metals. This process also produces a much finer fraction ‘‘rock flour’’. About 50% of the unoxidized tailings taken from Cell 3 in 2004 were found by optical microscopy to consist of a dark-gray groundmass, which was comprised of very finegrained sulfides and silicates (Fig. 4a). This fine-grained groundmass was not observed in the oxidized samples from Cell 2. This is probably due to the very fine grained sulfide and aluminosilicate minerals, which would be very reactive due to high surface areas, dissolving rapidly and releasing metals such as Fe, Zn, Cu and Al. Optical and SEM examinations showed that the coarser fraction of the unoxidized tailings contained about 50% sulfides (60% pyrite, 35% pyrrhotite, 5% chalcopyrite and sphalerite) and 50% gangue (quartz, muscovite, biotite, chlorite and pyroxene). Sulfide grains, about 1–2 lm in diameter, showed little sign of corrosion (0.1 lm deep) with thin (0.2 lm) dark brown rims (Fig. 4b). Gangue minerals were partially coated with dark-brown precipitates. No carbonate minerals were found in the tailings by optical or electron microscopy despite being listed in the ore mineralogy. The grain size was much coarser in the oxidized zone of the tailings of Cell 2 than in Cell 3, with sulfide minerals up to 50 lm in diameter (Fig. 4c). This grain size distribution is probably due to the relative proximity of the discharge pipe and rapid settling of coarse grains as has been found in Central Manitoba Au Mine tailings (Sherriff et al., 2007). Pyrrhotite grains from Cell 2 were highly corroded and completely surrounded by thick rims of Fe oxyhydroxides (Fig. 5a) with original grain boundaries being indiscernible. The alteration of sphalerite grains was less pronounced, with moderate corrosion and thin rims of Fe oxy-hydroxides (Fig. 5b). Pyrite remained unaltered except for thin rims of Fe oxy-hydroxides. The only chalcopyrite found in the oxidized tailings was disseminated within quartz and therefore armoured from oxidation processes (Fig. 5c). White and yellow evaporites form as thin crusts on the surface of the tailings, or as cumulates within fissures and cracks. Evaporites were shown by SEM EDX to contain Fe, Zn, Al and S. Melanterite (FeSO47H2O), rozenite (FeSO44H2O) and halotrichite (FeAl2(SO4)422H2O) were identified by XRPD. Zinc and Cu can partially replace Fe in these minerals (Jambor et al., 2000; Peterson, 2003; Ballirano et al., 2003). During dry periods, saline pore water evaporates from the tailings surface forming these hydroxy-sulfates. The evaporite minerals are soluble and will redissolve in times of high rainfall or during freshet releasing the metals to the surface water. Lepidocrocite (FeOOH) was identified by XRD in precipitates from unacidified filtered groundwater stored in the laboratory for a month. 4.3. Geochemistry Average metal concentrations in the unoxidized tailings (n = 22) were 21.0 (3.3) wt.% Fe, 0.16 (0.11) wt.% Zn and 0.06 (0.02) wt.% Cu compared to 26.6 (3.3) wt.% Fe, 0.17 (0.01) wt.% Zn, and 0.05 (0.006) wt.% Cu in oxidized tailings (n = 10). In the oxidized tailings for both years, Fe, Cu and Zn are depleted near the surface but below 100 cm depth there was no significant variation (Fig. 6).
Fig. 4. (a) Photomicrograph showing fine grained background in unoxidized tailings in Cell 3, (b) SEM image of fine grained unoxidized tailings from Cell 3, and (c) SEM image of coarser grained oxidized tailings from Cell 2.
The results from sequential extraction showed that about 3=4 of the total Fe was found in the residual phase (68–83%) and about a 1= 4 as Fe oxy-hydroxides (17–31%), with 0.5% being in water soluble, and <0.4% in exchangeable phases (Table 2). Zinc was more mobile with 20–39% of the total being water soluble and 4–8% exchangeable (Table 2). The remaining Zn is divided between Feoxy-hydroxide phases (16–32%) and residual (30–57%). Copper was found primarily in the Fe oxy-hydroxide (46–51%) and residual phases (13–41%). The remaining Cu is mobile in exchangeable (6–10%) and water soluble (13–28%) fractions (Table 2). Pore waters extracted from tailings in Cells 2 and 3 have some similarities, with low pH and high total dissolved solids near the
B.L. Sherriff et al. / Applied Geochemistry 26 (2011) 1832–1842
Fig. 5. SEM backscatter images of (a) a corroded pyrrhotite (po) grain with a zoned rim of precipitated secondary Fe oxy-hydroxide minerals, (b) a corroded grain of sphalerite (sph), and (c) chalcopyrite (cp) inside quartz. Note the relatively unaltered pyrite (py) in each image.
surface where the tailings are oxidized and have a more neutral pH and lower dissolved solids with depth. In general, pH values were lower and dissolved metal concentrations higher in 2005 than 2004 (Figs. 7 and 8). In Cell 3 in 2004, the pore water of the unoxidized tailings had a pH of 4.4 in the top 10 cm of the tailings which increased to 7.3 by 30 cm depth (Fig. 7). Dissolved metals follow pH, with concentrations of 18,000 mg/L Fe, 9600 mg/L Zn and 260 mg/L Al in the top 10 cm decreasing steadily with depth until most metals were below the detection limits (<0.01 mg/L Fe, <0.02 mg/L Al, <0.005 mg/L Zn) at 30 cm (Fig. 7). Due to a high water table, Cell 3 tailings could only be collected to a depth of 60 cm in 2005.
1837
Average values of dissolved metals in pore water increased compared to 2004 with Fe and Al having the highest values at 15 cm depth: 47,000 mg/L Fe, 9100 mg/L Al. Zinc concentration near the surface decreased from 9600 mg/L to 840 mg/L between 2004 and 2005, with a maximum value of 2600 mg/L at 35 cm depth (Fig. 7). From 2004 to 2005, the overall values of pH had decreased by 2–3 pH units to 2.3 at the surface and 4.8 at 30 cm. Concentration of Cu in pore water was less than 2 mg/L in both 2004 and 2005. In Cell 2 in 2004, pore water extracted from the upper 70 cm of oxidized tailings was characterized by pH values of 1.8 at 10 cm depth which increased to 7.1 at 75 cm depth (Fig. 9). Maximum concentrations of Fe (65,000 mg/L), Zn (9800 mg/L), Cu (1100 mg/L) and Al (29,000 mg/L) occurred at 20 cm depth despite the pH being slightly higher than at 10 cm. Below 1.2 m where the tailings are unoxidized, concentrations of Cu and Zn were just above detection limits of 0.0004 mg/L Cu and 0.005 mg/L Zn. The evolution from 2004 to 2005 in Cell 2 differed from that of Cell 3 with the pH increasing from the surface to a near neutral value of 6.9 at 300 cm depth (Fig. 9). The maximum value for Zn (2600 mg/L) was at 35 cm. There were high concentrations of Fe in pore water down to 150 cm with values from 17,000 to 47,000 mg/L Fe. Aluminum concentration decreased from 3700 mg/L at the surface to 500 mg/L by 25 cm depth. Some of this change may have been due to the saturation of the previously dry tailings by the high rainfall in the summer of 2005. Groundwater from the deepest well (RBH4) in unoxidized tailings had an alkaline pH of 9.6 in 2004, which decreased to 7.1 by 2007 (Table 3). The concentration of SO2 decreased from 4 2430 mg/L in 2004 to 1310 mg/L in 2007 and 1500 mg/L in 2009. Copper, Zn and Al concentrations remained below 1 mg/L. The concentration of Fe increased from 0.02 mg/L in 2004 to 9.71 mg/L by 2009. Wells RBH5 and RBH7 sampled shallower groundwater than RBH4. The pH of water from these wells has stabilized at 4.2 ± 0.1 since 2007 after values in 2004 of 7.5 (RBH7) and 6.9 (RBH5). There was a lower pH of 3.5 in both wells in the extremely wet conditions of September 2005 probably due to deeper permeation of acidic surface water (Table 3). The reductions of 2–3 pH units from 2004 to 2005 correspond to those for pore water at 0.7–1 m depth for Cell 2 (Fig. 8). The concentration of Cu in wells RBH5 and RBH7 remained below 0.05 mg/L from 2004 to 2009. The concentration of Zn increased from 2004 to 2009 but was still less than 5 mg/L in RBH5 although for RBH7 there were higher values of 106 mg/L in 2005 and 155 mg/L in 2009 (Table 3). Between 2004 and 2009, Fe concentration increased by up to four orders of magnitude from 1.09 mg/L to 7290 mg/L (RBH5) and from 2.6 mg/L to 11,900 mg/L (RBH7). Sulfate concentrations also increased significantly during the same period from 1770 mg/L to 31,300 mg/L (RBH5) and 1670 mg/L to 33,100 mg/L (RBH7). The initial high values of metals and SO2 4 in the groundwater could have been caused by the disturbance of the tailings due to construction of the trench in 2002–2003 or to the exposure and oxidation of the fine grained material from unoxidized tailings. High metal values in groundwater in 2009 may be due to the formation of extensive gullies at the side of the trenches, which are exposing unoxidized tailings to air causing acidification and release of metals (Fig. 9). The dramatic effect of erosion surfaces on the geochemistry of tailings was shown in a study of Central Manitoba Au mine tailings (Sherriff et al., 2009). The evolution of the runoff from the tailings is shown in the composition of the water draining the tailings from a stream at site (RW1) and a major E–W trench at site (RW2) (Fig. 2). The surface drainage at RW1 has become progressively more acidic with the pH decreasing from 3.1 in 2003 to 2.6 by 2009 (Table 4). There was a progressive increase from 2003 to 2007 of 3190 to 9030 mg/L SO2 4 , 200 to 1240 mg/L Fe, 0.5 to 34.4 mg/L Cu but then
1838
B.L. Sherriff et al. / Applied Geochemistry 26 (2011) 1832–1842
Fig. 6. Concentrations of Fe, Zn and Cu with depth in the oxidized tailings of Cell 2 in 2004.
Table 2 Results of sequential extraction from solid samples of the tailings; weighted means as percentages of total concentration. Year
Water soluble
Exchangeable
unoxidized unoxidized oxidized oxidized
2004 2005 2004 2005
0.5 0.5 0.5 0.5
0.4 0.2 0.2 0.3
unoxidized unoxidized oxidized oxidized
2004 2005 2004 2005
23 25 39 20
Copper Cell 3, unoxidized Cell 3, unoxidized Cell 2, oxidized Cell 2, oxidized
2004 2005 2004 2005
13 21 28 13
Iron Cell 3, Cell 3, Cell 2, Cell 2, Zinc Cell 3, Cell 3, Cell 2, Cell 2,
Crystalline Fe-hydroxide
Residual
6 5 5 6
11 23 19 25
83 72 76 68
4 4 8 5
2 2 3 3
14 22 22 29
57 48 30 44
10 6 9 7
13 10 12 9
23 39 39 40
41 24 13 31
a slight decrease to 5000 mg/L SO2 4 , 770 mg/L Fe and 29.7 mg/L in 2009. Zinc increased from 14.9 to 300 mg/L from 2003 to 2004 but decreased progressively to 64.3 mg/L in 2009. It appears that the drainage from the tailings surface sampled at RW1 had reached a steady composition by 2007 and improved in quality by 2009. These results are based on one set of samples per year, and there may have been a dilution effect in 2009 due to high precipitation. Metal and SO2 4 concentrations in water from the trench at RW2 also show an increase with time but it is much less systematic than for RW1 (Table 4). There was an initial pulse of SO2 4 (9890 mg/L), Cu (23 mg/L) and Zn (228 mg/L) in 2003 which decreased to approximately 5000 mg/L SO2 4 and less than 5 mg/L Cu by 2004. The highest measured values of 1220 mg/L Fe and 247 mg/L Zn occurred during the rain event in June 2004. Although the pH seems to have stabilized at about 2.6, SO2 and metal values increased 4 again in 2009 to 12,100 mg/L SO2 4 , 3700 mg/L Fe, 147 mg/L Zn and 6.7 mg/L Cu (Table 4). In June 2004, water was taken from the trench in tailings Cell 3 during a precipitation event (RW2 June 2004 rain) with a second sample from the same location 24 h after the rain ceased (RW2 June after rain) (Table 4). Iron, Zn, Cu, Mg and SO2 4 concentrations
Amorphous Fe-hydroxide
were higher during precipitation, while Na, K and Ca were higher after the rain ceased. Ponding of water on the surface of the tailings during rain events allows pools of low pH, metal laden water, that were previously isolated on the surface, to join and flow into the drainage trenches. Also evaporite material on the surface of the tailings, which can be dissolved during rain events, would cause an increase in dissolved metals and SO2 4 . Alkali metal concentrations became more dilute during precipitation as they behave as conservative components (Table 4). Since the inception of the Ruttan Mine, Ruttan Lake has been influenced by contaminated water run-off from the mine site as it was classified as part of the tailings management impoundment. Water entering the Lake from the tailings impoundment (RW4) had a stable pH of 2.9 ± 0.1 until 2007 but this decreased to 2.6 in 2009 (Table 5). During this time there has been a considerable variation in the concentration of SO2 (1660–9610 mg/L), Cu 4 (1.6–16.7 mg/L), Fe (125–2410 mg/L) and Zn (29.3–110 mg/L). The highest values are all from 2009, indicating a significant decrease in the quality of water entering Ruttan Lake. The buffering and dilution capacity of Ruttan Lake is shown by a comparison of the composition of the input water at RW4 with the water from
B.L. Sherriff et al. / Applied Geochemistry 26 (2011) 1832–1842
1839
Fig. 7. Cell 3 pore water composition with depth. The solid line represents values from 2004 and the dashed line 2005. All concentrations are in mg/L. The horizontal lines with triangles represent the water table: solid line for 2004, dashed line for 2005.
Fig. 8. Cell 2 pore water composition with depth; the solid line for 2004 and the dashed line for 2005. The water table was below the section in 2004 and at 4 m depth in 2005.
the west side of the Lake at RW8 and RW9 (Table 5). In June 2004, the surface water composition at all three sites are similar, but in September 2005 when the concentrations of SO2 4 , Cu, Fe and Zn
at RW4 were at least double the 2004 values, they stayed constant at RW8 and RW9 (Table 5). At the outlet from Ruttan Lake toward the open pit (RW3), SO2 increased from 2003 to 2007 4
1840
B.L. Sherriff et al. / Applied Geochemistry 26 (2011) 1832–1842 3þ
AlðOHÞ3ðsÞ þ 3Hþ ¼ Al
Fig. 9. Gullies forming at the side of trenches on the tailings in Cell 1 exposing unoxidized tailings.
(2520–3890 mg/L) and pH decreased from 2.9 to 2.7, however, Fe and Zn decreased from 878 to 245 mg/L and 169 to 38 mg/L, respectively, whereas Cu has stayed constant at 3.2 mg/L. A comparison of water from June to October 2004 indicates an increase in concentration of Fe, Cu, Zn and SO2 4 at RW4 and RW3 during the hot dry summer of 2004. 4.4. Mobility of metals The unoxidized tailings have average concentrations of 2200 ppm Zn and 530 ppm Cu, which can be mobilized by oxidation from sphalerite and chalcopyrite, respectively. The remnant sphalerite oxidizes and produces up to 10,000 mg/L Zn in nearsurface pore water but Cu concentration is at least a factor of 10 lower than for Zn. The major acid generating minerals, pyrite and pyrrhotite, can release high concentrations of Fe and SO2 to the pore water 4 (Blowes et al., 2003a). Although carbonate minerals were described in the original ore mineralogy, they are of insufficient abundance in the tailings to be visible in thin sections or to maintain a neutral pH, as in Eq. (1), once acid generation has commenced.
CaCO3 þ Hþ ¼ Ca2þ þ HCO3
ð1Þ
Aluminum liberated by the weathering of aluminosilicate gangue minerals, and Fe3+ will react with water to form Al and Fe oxy-hydroxide secondary phases. A decrease in pH below 6 will cause dissolution of these phases as shown in Eqs. (2) and (3) buffering the water at a pH of 4.3 for Al(OH)3 or 2–2.5 for Fe oxyhydroxides (Blowes and Jambor, 1990; Johnson et al., 2000; Blowes et al., 2003a; Sherriff et al., 2009).
þ 3H2 O
ð2Þ
FeOOHðsÞ þ 3Hþ ¼ Fe3þ þ 2H2 O
ð3Þ
These three buffers (Eqs. (1)–(3)) are controlling the chemistry of the pore, ground and surface water at Ruttan. In 2004, pore and groundwater at 0.8 m below grade generally had near neutral pH, where it is buffered by the dissolution of calcite (Figs. 7 and 8, Table 4). By 2005, the groundwater and pore water at about 1 m depth have stabilized at a pH of 4–4.5. This indicates the complete depletion of carbonate capacity and control of pH by the dissolution of Al(OH)3. Breakthrough of this acidic front correlates with increases in concentration of Fe, Al and Zn in solution. While Al hydroxides dissolve, Fe(II) can be oxidized and precipitated in the form of Fe(III) oxy-hydroxides and hydroxy-sulfates at pH below 4.5. Most surface water including Ruttan Lake has stabilized at pH 2–3 due to the Fe oxy-hydroxide buffers. Sequential extraction data showed that up to 50% of the total Fe is present in the form of Fe-oxide/hydroxide minerals in near surface samples of tailings. Pore water is in direct contact with sulfide grains and acts as the source of water for oxidation reactions and a conduit for metals and protons released from these oxidation reactions. In 2004, maximum concentrations of metals and SO2 were measured in the 4 upper 10 cm of the unoxidized tailings and rapidly decreased over the next 20 cm depth. Maximum concentrations were higher in 2005 and occurred between 20 and 30 cm depth corresponding to observations that oxidation was starting along fissures in the upper 10 cm of Cell 2 in 2004, whereas it was more pervasive and down to 20 cm depth in 2005. Once liberated from sphalerite, Zn is very mobile as a major secondary fraction of Zn was shown by sequential extraction to be in a water soluble form. Only a minor amount was found to be bound to Fe oxy-hydroxides due to its limited adsorption at pH values below 6 (Dzombak and Morel, 1990). The initial high values of metals and SO2 4 in groundwater could have been due to the construction of the trench in 2002–2003 or to the oxidation of the very fine grained material from unoxidized tailings. The increase in metal values in groundwater in 2009 may be due to the formation of extensive gullies at the side of the trenches, which are exposing unoxidized tailings to air causing oxidation with the accompanying acidification and release of metals (Fig. 9). Ruttan Lake acts as a reservoir for both contaminated water from the tailings and meteoric water which could dilute the water coming off the Ruttan tailings. In 2004 and 2005, the composition of the Lake was found to be more or less homogeneous from the east discharge (RW4) to the west shore (RW8 and RW9) and the pH remained constant at 2.8. By 2009, the pH of the water draining from the Lake to the open pit (RW3) had decreased slightly to 2.6 but the SO2 4 and metal content had not increased despite a considerable increase of the input to the Lake (RW4). As the unoxidized tailings in Cell 3 continue to de-water and the water table drops, O2 will continue to diffuse deeper into the tailings creating a deepening zone of oxidation. Modeling shows that sulfide-rich tailings with a deep water table can oxidize for centuries (Blowes and Jambor, 1990). Blowes et al. (2003a) defined a model of the geochemical evolution of tailings which illustrates the location and movement of metals as they oxidize. A comparison of the conditions at Ruttan with their conceptual model suggests that the tailings were still in the early to moderate stage of oxidation in 2005. In both Cells 2 and 3, there is a minor depletion in solid metal concentrations near the surface and a peak in dissolved metal concentrations just below the surface. Although there is a zone of lower metal content of the pore water near the surface of Cell 2 in 2005, it is probably due to dilution from heavy rainfall
1841
B.L. Sherriff et al. / Applied Geochemistry 26 (2011) 1832–1842 Table 3 Composition of Ruttan groundwater (mg/L). Depth to top of screen
RBH4
RBH4
RBH4
RBH5
RBH5
RBH5
RBH5
RBH7
RBH7
RBH7
RBH7
Date
September2004
2.6 m July2007
July2009
September2004
1.1 m September2005
July2007
July2009
September2004
0.8 m September2005
July2007
July2009
pH SO4 Cu Fe Zn Mn Al Ca Mg Na K Cd Co Cr Ni
9.6 2430 <0.05 0.02 <0.10 nd 0.22 523 2.63 83.9 43.8 <0.50 <0.05 nd <0.10
7.1 1310 0.0055 1.36 0.078 0.14 0.65 457 5.16 74.1 43.0 <0.0002 0.003 <0.001 0.005
7.0 1500 0.0062 9.71 0.155 0.018 0.554 460 1.34 92.0 57.8 <0.0002 0.002 0.002 0.007
6.9 1770 <0.05 1.09 0.18 0.43 <0.05 570 10.4 68.3 45.6 <0.50 <0.05 nd <0.10
3.5 3790 0.0086 803 1.26 3.79 1.76 447 176 67.5 124 0.001 0.001 <0.001 0.012
4.1 4460 0.0125 840 4.05 10.1 0.65 375 427 53.1 98.1 <0.0002 0.002 <0.001 0.002
4.2 31,300 0.0205 7290 1.34 36.8 3.12 191 1660 36.5 62.6 <0.0002 <0.001 0.003 0.005
7.5 1690 <0.05 2.60 1.26 0.57 <0.05 540 8.18 57.49 38.4 <0.50 <0.05 nd <0.10
3.5 11,300 0.0119 3770 106 15.1 8.6 269 614 46.9 97.0 0.005 0.006 <0.001 0.022
4.3 5710 0.0293 1670 0.784 12 1.95 320 404 37.5 125.0 0.006 0.004 <0.001 0.003
4.2 33,100 0.0261 11,900 155 40.1 199 147 1730 18.9 58.9 0.012 0.045 0.027 0.049
Table 4 Tailings’ surface water composition (mg/L). RW1
pH SO4 Cu Fe Zn Mn Al Ca Mg Na K Cd Co Cr Ni
RW2
September 2003
June 2004
July 2007
July 2009
September 2003
June 2004 rain
June after rain
October 2004
July 2007
July 2009
3.1 3190 0.5 200 14.9 2.6 39.9 573 76 86.2 50.6 0.02 0.07 0.01 0.08
2.9 6880 27.3 1170 300 12.3 491 236 443 8.8 8.6 0.51 0.86 nd 0.64
2.7 9030 34.4 1240 169 11.4 552 207 419 6.1 3.16 0.31 1.80 0.06 0.65
2.6 5000 29.7 770 64.3 4.6 350 111 270 3.7 2.01 0.15 0.83 0.08 0.29
2.8 9890 23.0 598 228 15.8 641 297 535 7.6 10.6 0.31 1.68 0.03 0.81
3.4 4800 4.7 1220 247 13.8 145 335 393 44.3 24.3 0.31 0.08 nd 0.46
3.5 3960 1.7 844 139 9.6 90.9 414 261 66.4 39.6 0.16 0.07 nd 0.31
2.5 5030 <0.05 415 2.48 nd 0.5 514 106 81.7 60.8 0.00 0.00 nd 0.00
2.8 5070 0.3 1580 35.1 7.0 50.9 413 214 61.6 51.6 0.02 0.11 0.02 0.07
2.6 12,100 6.7 3700 147 10.8 457 248 622 24.6 20.8 0.23 0.66 0.13 0.30
RW1: Stream. RW2: Trench.
Table 5 Lake water composition (mg/L) 2004 and 2005. RW3
pH SO4 Cu Fe Zn Mn Al Ca Mg Na K Cd Co Cr Ni
RW4
RW8
RW9
June 2004
October 2004
September 2005
July 2007
July 2009
September 2003
June 2004
October 2004
September 2005
July 2007
July 2009
June 2004
September 2005
June 2004
September 2005
2.9 2520 3.2 878 63.2 6.5 103 294 286 23.8 9.3 0.08 0.04 nd 0.17
3.2 11,000 21.7 1510 169 nd 529 312 492 26.1 15.4 0.34 1.12 nd 0.57
2.8 2460 4.4 161 50.1 3.4 101 205 152 13.2 4.96 0.11 0.26 0.03 0.18
2.7 3890 3.2 245 38 3.8 115 180 147 12.2 4.18 0.06 0.25 0.03 0.15
2.6 2040 2.2 222 21.8 2.6 85 137 112 8.9 2.54 0.04 0.16 0.03 0.09
2.8 4970 1.6 125 34.6 3.3 65.1 381 123 20.5 16.5 0.05 0.17 0.02 0.16
2.9 1660 1.7 129 29.3 3.0 60.1 286 122 15.9 10.2 0.04 0.09 nd 0.12
3.1 7920 3.1 1140 81.8 nd 162 374 273 48.2 27.9 0.11 0.29 nd 0.20
2.8 3290 3.4 470 50.4 3.7 127 210 185 14.3 4.88 0.09 0.26 0.05 0.16
2.8 6260 11.0 1520 78.3 8.4 228 352 279 45.4 36.6 0.11 0.54 0.03 0.25
2.6 9610 16.7 2410 110 8.9 406 205 484 17.5 14.9 0.20 0.71 0.11 0.31
2.7 1700 1.4 93 28.4 nd 52.7 273 109 17.1 9.4 0.04 0.09 nd 0.00
2.9 1470 1.4 80.1 20.3 nd 47.8 153 88 11.5 3.69 0.04 0.11 0.02 0.08
2.9 1990 1.5 108 28.4 nd 55.3 286 114 17.3 9.5 0.04 0.10 nd 0.00
2.8 2140 2.3 131 29.6 nd 70.2 216 128 15.2 4.29 0.06 0.17 0.03 0.12
RW3: Outflow from Ruttan Lake to open pit. RW4: Inflow to Ruttan Lake from tailings.
1842
B.L. Sherriff et al. / Applied Geochemistry 26 (2011) 1832–1842
rather than a reduction in the reactivity of sulfides. Pore water metal concentrations peak within the first meter of all the tailings, suggesting a juvenile oxidation stage. This is not surprising for Cell 3, as these tailings were very recently introduced to oxidizing conditions. As the tailings are greater than 10 m thick they will be oxidizing and releasing metals for many years. The concentration of metals in Ruttan Lake may increase and pH levels decrease as oxidation of the tailings proceeds, however, the constant flow into the open pit and partial recharge by meteoric waters has a stabilizing effect so that the Lake is unlikely to become as acidic as the ponds on the tailings surface. 5. Conclusions 1. The very fine grained solids which had accumulated in Cell 3, the polishing pond, were immediately oxidized on dewatering possibly producing an initial pulse of sulfates and some metals in the trench and input water into Ruttan Lake. Iron mobilized in these reactions was oxidized to Fe(III) and deposited as Fe oxy-hydroxides surrounding mineral grains. 2. The geochemical reactions occurring during dewatering are dominated by the abundant pyrite and pyrrhotite but with Zn being mobilized from the sphalerite remaining after beneficiation of the ore. Little Cu is being mobilized partly due to the armouring of chalcopyrite by quartz and also by preferential absorption of Cu on secondary Fe minerals. 3. On closure in 2002, the alkaline pH of the tailings water was buffered by dissolution of the scarce natural carbonates and lime added during processing. Within 2 years of mine closure, an acidic front with pH 4.5, buffered by the dissolution of Al(OH)3, had broken through and reached the water table in most of the tailings. The surface runoff and Ruttan Lake water has stabilized at pH values of 2.5 due to the dissolution of Fe oxy-hydroxides. 4. Acidic metal-rich surface and shallow pore water evaporates to form Zn and Fe hydroxy-sulfate evaporite minerals. Metals that are temporarily attenuated on the surface of the tailings are released during periods of rain or snow melt. 5. The composition of Ruttan Lake and the discharge to the open pit seems to have stabilized despite orders of magnitude increases in the loadings of SO2 4 and metals from the tailings.
Acknowledgements Funding for this project was obtained from the Manitoba Sustainable Development Fund and from the Natural Science and Engineering Research Council of Canada. We acknowledge the assistance of Dr. Kathleen Londry, Dr. Tricia Stadnyk, Stephanie Simpson and the late Greg Morden and Sergio Meijia for field and laboratory assistance. We especially acknowledge the assistance of Ben Edirmanasinghe and Ernie Armitt from the Manitoba Department of Energy and mines, and the late Frank Bloodworth, HBM&S for site access and assistance. We acknowledge
anonymous reviewers for thorough helpful reviews of an earlier version of this manuscript. References Ballirano, P., Bellatreccia, F., Grubessi, O., 2003. New crystal-chemical and structural data of dietrichite, ideally ZnAl2(SO4)422H2O, a member of the halotrichite group. Eur. J. Mineral. 15, 1043–1049. Barrie, C.T., Taylor, C., Ames, D.E., 2005. Geology and metal contents of the Ruttan volcanogenic massive sulfide deposit, northern Manitoba, Canada. Mineral. Deposita 39, 795–812. Blowes, D.W., Jambor, J.L., 1990. The pore-water geochemistry and the mineralogy of the vadose zone must be some sulfide tailings, Waite Amulet, Quebec, Canada. Appl. Geochem. 5, 327–346. Blowes, D.W., Ptacek, C.J., 1994. Acid-neutralization mechanisms in inactive mine tailings. In: Jambor, J.L., Blowes, D.W. (Eds.), Mineral. Assoc. Can. Short Course Handbook on Environmental Geochemistry of Sulfide Mine Wastes, vol. 22, pp. 271–292. Blowes, D.W., Ptacek, C.J., Jambor, J.L., Weisner, C.G., 2003a. The geochemistry of acid mine drainage. In: Sherwood Lollar, B. (Ed.), Environmental Geochemistry. Holland, H.D., Turekian, K.K. (Exec. Eds.), Treatise on Geochemistry, vol. 9. Elsevier, Amsterdam, pp. 149–204. Blowes, D.W., Ptacek, C.J., Jambor, J.L., 2003b. Mill tailings: hydrolgeology and geochemistry. In: Jambor, J.L., Blowes, D.W., Richie, A.J.M. (Eds.), Environmental Aspects of Mine Wastes. Mineral. Assoc. Can. Short Course Series 31, pp. 95– 116. Dold, B., 2003. Speciation of the most soluble phases in a sequential extraction procedure adapted for geochemical studies of copper sulfide mine waste. J. Geochem. Explor. 80, 55–68. Dzombak, D.A., Morel, F.M.M., 1990. Surface Complexation Modeling: Hydrous Ferric Oxide. Wiley-Interscience, New York. Etcheverry, D.J., 2008. Spatial and Temporal Variations in the Ruttan Mine Tailings, Leaf Rapids, Manitoba, Canada. M.Sc. Thesis, Univ. Manitoba. Hall, G.E.M., Gauthier, G., Pelchat, J.C., Pelchat, P., Vaive, J.E., 1996. Application of a sequential extraction scheme to ten geological certified reference materials for the determination of 20 elements. J. Anal. Atom. Spectrom. 11, 787–796. Jambor, J.L., Nordstrom, D.K., Alpers, C.N., 2000. Metal-sulfate salts from sulfide mineral oxidation. In: Alpers, C.N., Jambor, J.L., Nordstrom, D.K. (Eds.), Sulfate Minerals—Crystallography, Geochemistry, and Environmental Significance. Reviews in Mineralogy and Geochemistry, vol. 40, pp. 303–350. Johnson, R.H., Blowes, D.W., Robertson, W.D., Jambor, J.L., 2000. The hydrochemistry of the Nickel Rim mine tailings impoundment, Sudbury, Ontario. J. Contam. Hydrol. 41, 49–80. Liu, R., Wolfe, A., Dzombak, D., Stewart, B., Capo, R., 2008. Comparison of dissolution under oxic acid drainage conditions for eight sedimentary and hydrothermal pyrite samples. Environ. Geol. 56, 171–182. Peretyazhko, T., Zachara, J.M., Boily, J.-F., Xia, Y., Gassman, P.L., Arey, B.W., Burgos, W.D., 2009. Mineralogical transformations controlling acid mine drainage chemistry. Chem. Geol. 262, 169–178. Peterson, R.C., 2003. The relationship between Cu content and distortion in the atomic structure of melanterite from the Richmond mine, Iron Mountain, California. Can. Mineral. 41, 937–949. Petrunic, B.M., Al, T.A., 2005. Mineral/water interactions in tailings from a tungsten mine, Mount Pleasant, New Brunswick. Geochem. Cosmochim. Acta 69, 2469– 2483. Sherriff, B.L., Ferguson, I., Gupton, M.W., VanGulck, J.F., Sidenko, N.V., Priscu, C., Marco Pérez-Flores, M., Gómez-Treviño, E., 2009. A geophysical and geotechnical study to determine the hydrological regime of the Central Manitoba gold mine tailings deposit. Can. Geotech. J. 46, 69–80. Sherriff, B.L., Sidenko, N.V., Salzsauler, K.A., 2007. Differential settling and geochemical evolution of tailings’ surface water at the Central Manitoba Gold Mine. Appl. Geochem. 22, 342–356. Sidenko, N.V., Sherriff, B.L., 2005. The attenuation of Ni, Zn, and Cu, by secondary Fe phases of different crystallinity from surface and groundwater of two sulfide mine tailings in Manitoba, Canada. Appl. Geochem. 20, 1180–1194. Stantec Consulting Limited, 2003. Hudson Bay Mining and Smelting First Environmental Effects Monitoring Study and Design. A Report to Hudson Bay Mining and Smelting Company Ltd. by Stantec Consulting Limited, Guelph, Ontario, Canada.