- Email: [email protected]
Antimony distribution and environmental mobility at an historic antimony smelter site, New Zealand
Environmental Pollution 129 (2004) 257–266 www.elsevier.com/locate/envpol
Antimony distribution and environmental mobility at an historic antimony smelter site, New Zealand N.J. Wilson, D. Craw*, K. Hunter Chemistry and Geology Departments, University of Otago, PO Box 56, Dunedin, New Zealand Received 16 June 2003; accepted 15 October 2003
‘‘Capsule’’: High levels of antimony in primitive smelter soils remain largely immobile on the metre scale.
Abstract A historic antimony smelter site at Endeavour Inlet, New Zealand has smelter residues with up to 17 wt.% antimony. Residues include coarse tailings (cm scale particles, poorly sorted), sand tailings (well sorted) and smelter slag (blocks up to 30 cm across). All of this material has oxidised to some degree over the ca. 100 years since the site was abandoned. Oxidation has resulted in acidification of the residues down to pH 2–5. Smelter slag contains pyrrhotite (FeS) and metallic antimony, and oxidation is restricted to surfaces only. The coarse tailings are the most oxidised, and few sulfide grains persist. Unoxidised sand tailings contain 10–20 vol.% stibnite (Sb2S3) containing up to 5% As, with subordinate arsenopyrite (FeAsS), and minor pyrite (FeS2). The sand tailings are variably oxidised on a scale of 2–10 cm, but original depositional layering is preserved during oxidation and formation of senarmontite (Sb2O3). Oxidation of sand tailings has resulted in localised mobility of both Sb and As on the cm scale, resulting in redistribution of these metalloids with iron oxyhydroxide around sand grain boundaries. Experiments demonstrate that Sb mobility decreases with time on a scale of days. Attenuation of both As and Sb occurs due to adsorption on to iron oxyhydroxides which are formed during oxidation of the smelter residues. There is no detectable loss of Sb or As from the smelter site into the adjacent river, <50 m away, which has elevated Sb (ca. 20 mg/l) and As (ca. 7 mg /l) from mineralised rocks upstream. Despite the high concentrations of Sb and As in the smelter residues, these metalloids are not being released into the environment. # 2003 Elsevier Ltd. All rights reserved. Keywords: Antimony; Arsenic; Oxidation; Adsorption; Stibnite; Arsenopyrite; Senarmontite; Metalloids; Mining
1. Introduction Antimony is increasingly being identified as a toxic environmental pollutant and has been implicated in cancer development (Foy et al., 1978; Gurnani et al., 1994; Gebel, 1997). The principal global sources of anthropogenic antimony pollution are the mining and smelting industries (Adriano, 1986), and very high levels of pollution have been detected around smelter sites in particular (Ragaini et al., 1977; Ainsworth et al., 1990; Baroni et al., 2000; Flynn et al., 2003). However, little is known about the environmental mobility of antimony in such situations (Filella et al., 2002). Baroni et al. (2000) have shown that some plants growing on smelter * Corresponding author. Tel.: +64-3-479-7519; fax: +64-3-4797527. E-mail address: [email protected] (D. Craw). 0269-7491/$ - see front matter # 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.envpol.2003.10.014
residues can accumulate high levels of antimony. Conversely, Flynn et al. (2003) have shown that despite very high levels of soil antimony at mine and smelter sites, antimony bioavailability is low, and the antimony is fixed in the soils. Flynn et al.’s (2003) study, and earlier studies, did not investigate the disposition of the antimony in the mine soils, as they were dealing with well-developed organic-rich soils in which original mine and smelter materials were dispersed. In this study, we address the nature of antimony-bearing material in primitive soils at an historic smelter site that has remained undisturbed and has negligible revegetation after ca. 100 years. This study examines the environmental mobility of antimony in the early stages of soil development at the smelter site, and the processes which fix antimony in the primitive soils. Both arsenic and antimony are chalcophilic group V metalloids that form oxyanions in water, and therefore
258
N.J. Wilson et al. / Environmental Pollution 129 (2004) 257–266
these elements have some similarities in chemical behaviour in the environment. Arsenic is considered to be toxic and carcinogenic in a similar way to antimony (Gebel, 1997) and is currently regarded as one of the most serious inorganic drinking water pollutants in the world (Smedley and Kinniburgh, 2002). Maximum drinking water limits have been set by the World Health Organisation at 10 and 5 mg l1 for arsenic and antimony, respectively, with the latter value provisional until a greater understanding of its toxicity (WHO, 1996). Because of these similarities in environmental chemistry and toxicity between antimony and arsenic, this paper examines environmental arsenic mobility in the smelter soils in conjunction with our study of antimony, to allow direct comparison. The site we use facilitates this comparison because antimony and arsenic are present together at strongly elevated levels.
2.2. Historical mining and smelting Mining for antimony began in the hills above Endeavour Inlet in 1874, following discovery of stibnite bearing veins in 1873 (Johnston, 1993). In 1875, a smelter was erected to process the mined ore, on an alluvial terrace at the base of the hills about 1 km from Endeavour Inlet (Fig. 1). The smelter initially used four cast iron retorts and wood-fired furnaces to melt the ore, which was mixed with scrap iron to produce 90–92% pure antimony (Johnston, 1993). In 1885, the smelter was redeveloped with a large sorting and crushing facility and a more complex smelting process. Concentrated ore was first melted using wood-fired furnaces, then remelted at higher temperatures using coke, before final production of 100% antimony (Johnston, 1993). Further refinements followed, but falling antimony prices saw the operation progressively scaled down, and the site was abandoned by 1907 (Johnston, 1993).
2. General setting 2.1. Geological background The Endeavour Inlet area that is the subject of this study is underlain by schistose Mesozoic sediments that make up most of the South Island of New Zealand (Fig. 1). These schists are locally known as Marlborough Schist, but are genetically related to the more extensive Otago Schist in the south of the South Island where mesothermal antimony deposits occur also (Williams, 1974; Ashley et al., 2003). The Marlborough Schist has uniform mineralogy dominated by quartz, albite, muscovite, chlorite and calcite. The rocks are extensively recrystallised and have a near-pervasive foliation that is generally flat-lying or gently dipping. Late Tertiary tectonics have caused geomorphic rejuvenation, resulting in rugged ridges and steeply incised streams. Principal valleys have been flooded by the sea to make an intricate set of sounds, and Endeavour Inlet (Fig. 1) is a minor bay within this sound complex. The climate is temperate (mean annual temperature ca. 12 C) with ca. 1000 mm/year of rainfall that supports locally lush stands of forest. The foliation of the Marlborough Schist is cut by steeply dipping mesothermal quartz veins that are variably mineralised with pyrite, arsenopyrite, scheelite, gold, and/or stibnite (Williams, 1974). The mines at Endeavour Inlet were developed in a swarm of mineralised veins (typically < 1 m wide) dominated by quartz, stibnite, and arsenopyrite (Pirajno, 1979). The veins form a swarm of sub-parallel structures in a zone approximately 200m wide, which is continuous for at least 5 km north of Endeavour Inlet. Mineralisation was zoned with depth, so that at lower levels arsenic is the dominant hydrothermal element, while at higher levels antimony is more common (Pirajno, 1979).
Fig. 1. Location map of the Endeavour Inlet antimony smelter site, and sites of water analyses in the adjacent river (site numbers refer to analyses in (Table 1). The site is on alluvium in a valley, and approximate groundwater flow directions in that alluvium are indicated, based on topography and surface water flows.
N.J. Wilson et al. / Environmental Pollution 129 (2004) 257–266
Now, approximately 100 years since the smelter was scaled down, no equipment remains at the site and the site is unpopulated. Farms and tourist accommodation dot the shore of the nearby inlet, but the population is small (< 100). Much of the valley upstream of the site has reverted to native forest and is now a natural reserve. The smelter site is marked by an area of ca. 1 ha on which little revegetation has occurred. There are several distinctly different types of material present at the site, and these are discussed separately in this paper. Most of the material at the site consists of coarse tailings from the crushing and sorting plant, mixed with slag from the smelting process. In addition, there is sand rich in sulfide minerals derived from the crushing and concentration plant. Unoxidised sand is grey and this is surrounded by brown oxidised sand (Fig. 2 A, B; 3).
3. Methods 3.1. Field sampling Samples of solid material were collected from the tailings heaps and nearby soil at the downstream of the smelter site (Figs. 1 and 2A, B). The sulfide sand is variably oxidised, so associated oxidised and unoxidised material were sampled. These samples were kept dry to
Fig. 2. Sketch sections (constructed from photographs of shallow excavations) through smelter residues, showing scales and relative distributions of different types of materials. A. Coarse tailings pile at the smelter site, with layers of sand tailings and cobbles of smelter slag. B. Sand tailings at the outwash area downstream of the smelter site (Fig. 1).
259
prevent further oxidation after sampling. Fine grained (sand and finer) stream sediments were sampled from the bed of the main stream that passes the smelter site, at points indicated in Fig. 1. Labplus sterile CS/500 bags were used to collect solid samples. Water samples were collected from the river at the same points as sediment samples (Fig. 1). A set of four water samples upstream and downstream of the smelter site was taken in September when the site had been through a wet winter season and water levels were still high. Water samples were collected in triplicate at each site in acid-washed 500 ml high-density polyethylene (HDPE) bottles. The water samples were clear with no obvious particulate material or brown iron oxyhydroxide colloidal material. Samples were not filtered and chemical preservatives were not used (Pierce and Brown, 1977; Aggett and Kriegman, 1987). Water and soil pH and redox potential were measured in the field using an Oakton WD-35615 meter calibrated with standard pH solutions and Zobell’s solution for redox potential (Nordstrom, 1977). Paste-pH measurements (Sobek et al., 1978) were made at points where insufficient soil moisture was present for direct measurement. 3.2. Analytical methods Water collected from the smelter site was analysed by graphite furnace atomic absorption spectrometry (GFAAS) for As and Sb using a Perkin-Elmer 4100ZL graphite furnace. Pd/Mg(NO3)2 was added to sample aliquots as a matrix modifier to overcome incomplete atomisation (Ouishi et al., 1994). The detection limit for both Sb and As was 5 mg l1. Results are presented in Table 1. Determination of antimony speciation by voltammetry as part of a larger study by Wilson (2002) demonstrated that nearly all dissolved Sb in surface waters is present as the Sb(V) species. Solid samples were dried at 60 C before being ground to < 50 mm using a TEMA rock crusher. Sample powders were treated by nitric/hydrochloric acid digestion (USEPA 200.2) followed by solution analysis for As and Sb by inductively-coupled plasma (ICP) at ALS Chemex, Brisbane, Australia, or Hill Laboratories, Hamilton, New Zealand. Results from the different ICP labs are within 10% of each other. Detection limits were < 1 mg/kg, and all analytical errors, as indicated by replication of analyses and separate analyses of triplicate water samples, are contained within the symbols of diagrams in this study. Results are presented in Table 1. Minerals were identified in thin and polished sections by light microscopy, augmented with X-ray diffraction (XRD) and electron microprobe semiquantitative analysis. XRD analysis was performed on powder smears crushed finely in an agate mortar and pestle, using Cu Ka radiation on a Philips PW 1050 diffractometer.
260
N.J. Wilson et al. / Environmental Pollution 129 (2004) 257–266
Antimony and arsenic element maps of small (0.500.45 mm) areas of polished sections were generated using a Jeol microprobe at 25 kV by automated stepping (1 micron steps) across the area. The resultant images (Fig. 4A–D) are compilations of the results of 225 000 one-second analyses, with darkest patches representing highest concentrations. White matrix or
sparsely coloured patches in these maps reflect zones near to, or below, background levels of ca. 0.5 wt.% Sb or As. 3.3. Experimental study In order to simulate short-term mobility of As and Sb from the solid smelter residues in the environment, three dissolution or leaching experiments were performed. Distilled water was used as a leaching fluid in these simulations. Results of duplicate experiments, as outlined below, have been averaged for presentation in Table 2. To simulate aerobic dissolution, an experiment based on those of Ashley et al. (2003) was set up. Approximately 5 g of solid material were accurately weighed into LDPE plastic jars (60 ml capacity), then 15 ml of distilled water were added and the jars sealed. It was assumed that enough oxygen was present within the jars that any sediment dissolution would proceed via aerobic pathways. These jars were then stored in darkness at ca. 20 C for up to 28 days. Sufficient jars were prepared so that the amount of dissolved arsenic and antimony could be measured after periods of 1, 7, 14 and 28 days, then the jar discarded. Duplicate samples from three different tailings heaps, grey sulfide sand, and brown (oxidised sulfide) sand, were prepared. After sufficient time had elapsed, 10 ml of water from each sample were pipetted into two 5 ml plastic vials. The water in these vials was then analysed for As and Sb by ICP–OES using a Thermo Jarrell Ash AtomScan 25 ICP-spectrometer. Analyses were recalculated to (mg metalloid released)/(kg solid material in experiment) to quantify amounts of metalloids released (Table 2) rather than mineral solubility (cf Ashley et al., 2003).
Fig. 3. Photograph of a thin section through the boundary between grey, least oxidised, sand tailings (within dashed white lines) and brown oxidised sand tailings from the smelter site. Original depositional layering (near-horizontal) is accentuated by iron oxyhydroxide (dark) in the oxidised areas, and to a lesser extent in some layers in unoxidised zones. A coarser sand layer at top has been extensively cemented with iron oxyhydroxide. The layering has undergone some post-depositional deformation, resulting in warps and a sharp step just left of centre. Pale grey material around the tailings is mounting glue. Table 1 Chemical analyses of solids and waters at the smelter site Site
Material
Solid Sb (mg/kg)
Solid As (mg/kg)
River water Sb (mg/l)
River water As (mg/l)
pH
Stream 1 Stream 2 Stream 2 Stream 2 Stream 3 Stream 4 Smelter Smelter Smelter Smelter Smelter Smelter Smelter Smelter Smelter Smelter Smelter Smelter Smelter
Sediment Sediment Sediment Sediment Sediment Sediment Soil Coarse tailings Coarse tailings Coarse tailings Coarse tailings Coarse tailings Coarse tailings Slag Slag Grey sand Grey sand Brown sand Brown sand
150 243 68 38 22 18 80 200 40 100 47 000 34 600 65 000 176 700 45 100 123 500 22 200 132 900 35 100 88 600 25 800
67 190 125 83 14 27 2220 8350 4320 4610 4870 203 65.2 135 36 9120 6110 8550 5630
23.5 14.1 30.4 22.9 24.4 23.8
7.59 6.5 5.49 8.33 8.41 8.64
7.6 7.6 7.6 7.6 7.6 7.6
Eh (mV)
472
4.2
609
4.1 3.1
618 662
2.4 2.6 3.9
690 673 627
Date September April June September September September
261
N.J. Wilson et al. / Environmental Pollution 129 (2004) 257–266
Fig. 4. Microprobe element maps of unoxidised (left) and oxidised (right) sand tailings, for antimony (top) and arsenic (bottom). See text for construction methods. Both A and C are maps of a single area, and B and D are maps of a single separate area. Sulfide grains are readily apparent in A and C (maximum concentration examples arrowed). Ellipse in A encloses an example of secondary antimony precipitated with iron oxyhydroxides around sand grain boundaries. Diffuse distribution of Sb and As in oxidised sand is shown in B and D, with maximum concentrations indicated. Table 2 Results of experimental mobilisation of Sb and As from smelter residues. Experiments were conducted in duplicate, and results averaged Experiment
Aerobic Aerobic Aerobic Aerobic Unsaturated Anaerobic Anaerobic Anaerobic
Time (days)
1 7 14 28 1 7 14
Grey sand (mg /kg)
Brown sand (mg /kg)
Coarse tailings 1 (mg /kg)
Coarse tailings 2 (mg/ kg)
Coarse tailings 3 (mg/ kg)
Sb
As
Sb
As
Sb
As
Sb
As
Sb
As
21.4 7.56 7.71 7.75 15.7 82.1 89 58.2
411 601 778 768 108 335 690 701
19.9 14.6 16.1 15.9 5.46
113 194 214 306 8.48
6.72 4.72 6.06 4.46
2.9 2.77 3.37 2.93
18.4 3.12 3.28 4.14
2.7 1.74 1.71 1.74
24.3 26.3 27.3 28.8
1.9 2.38 2.42 2.31
N/M
A parallel experiment was constructed using grey sulfide sand in order to evaluate anaerobic dissolution of Sb and As. Distilled water was deoxygenated by bubbling it with nitrogen gas for at least 20 min. About 15 g of sample were accurately weighed into 60 ml jars and the jars weighed with the lids on. The jars were then filled to the brim with deoxygenated water, sealed and reweighed to
N/M
determine the amount of water added. These jars were then placed in sealable plastic bags which were filled with deoxygenated water, sealed and left in darkness at ca. 20 C. Sufficient samples were prepared to allow for duplicated aliquots to be analysed after 1, 7 and 14 days. A third experiment was designed to simulate As and Sb mobilisation under unsaturated conditions. Distilled
262
N.J. Wilson et al. / Environmental Pollution 129 (2004) 257–266
water was dripped at a rate of approximately 1 ml s1 through a burette into a funnel in which 5 g of grey or brown sand rested upon a 0.45 mm filter. A Buchner funnel was used so that water would be sucked through the sediment and filter to fall into a clean 25 ml plastic vial. Once 15 ml of water had been collected, the burette was closed and the water collected was analysed for As and Sb by ICP–OES. This experiment was performed in duplicate for both grey and brown sand.
4. Results 4.1. Description and mineralogy Coarse tailings heaps are up to 5 m high and 20 m long, and consist mainly of cm scale pebbles, with some finer grained matrix. Heaps have high porosity and water is not retained, so the coarse tailings are dry except during rain events. Quartz is the most common mineral in the tailings, and up to 30% of the tailings are monominerallic quartz clasts. Other clasts are quartzbearing but include fragments of basement schist. Most clasts have sulfide mineral remnants (stibnite and arsenopyrite) that have been extensively oxidised, and these tailings are a deep brown colour. At least some of the matrix is decomposed sulfidic schist fragments pervaded by iron oxyhydroxide. XRD analysis shows only quartz with minor muscovite in the matrix material. Blocks of slag from the smelter are common on the surfaces of the coarse tailings heaps and intervening flat ground, and are also distributed through the coarse tailings within the heaps. These blocks are up to 30 cm across and irregular in shape. Slag is hard but porous and locally vesicular. Some blocks of slag still retain the shape of the crucibles into which the molten ore was poured. Quartz is still recognisable within some slag samples, as millimetre scale grains. Much of the slag is made up of pyrrhotite and fine grained micaceous material, and locally graphite zones up to 2 cm thick occur. Droplets of antimony metal up to 1 mm across occur distributed throughout the slag. Slag blocks are variably oxidised to deep brown on the surfaces, but fresh metallic minerals are readily seen on broken surfaces. Unoxidised sulfide bearing sandy tailings are grey in hand specimen, and are soft, pliable, and cohesive. Most of this material was water-saturated or at least moist when first exposed. The sand occurs in thin (ca. 20 cm) layers within heaps of coarse tailings, and as an outwash deposit on the alluvial terrace up to 100 m from the smelter site (Fig. 2A, B). The sand is well-sorted with angular clasts of quartz, micas, and sulfides. Layering at the mm to cm scale is preserved locally, although some post-depositional deformation has distorted this layering (Fig. 3). The layering is defined by varying proportions of sulfides and silicates. Stibnite
is the most common sulfide (10–20%), but arsenopyrite grains (2–5%) are scattered throughout. Pyrite grains are rare (< 1%). Evidence for incipient oxidation of this grey sand can be seen in the microprobe map, which shows faint secondary seams of mobilised antimony around grains of silicates (Fig. 4A, ellipse area). These seams follow zones of pigmentation by iron oxyhydroxide along sand grain boundaries. The grey unoxidised tailings are preserved only as irregular remnant patches within the sandy tailings horizons. These patches are surrounded by pale brown oxidised sand, particularly in layers perched in porous sediments (Fig. 2A, B). The least altered horizon lies on flat ground on the smelter site (Fig. 2A), and appears to be commonly water-saturated. Individual depositional layers (mm scale) can be traced from unoxidised to oxidised sand (Fig. 3). Layers in the oxidised sand are defined by varying degrees of pigmentation by brown iron oxyhydroxide. No sulfide minerals remain in oxidised sand, but sulfide grain shapes are preserved indistinctly (Fig. 4A–D). XRD analysis reveals quartz as the dominant mineral, swamping most other lines. A prominent XRD line at 3.2 A˚ in oxidised sand suggests that senarmontite (SbIII 2 O3) is present, but other lines for this mineral are not apparent. No other metallic minerals have been identified. The 3.2 A˚ line attributed to senarmontite may overlap the principal scorodite (FeAsO4.2H2O) line (3.18 A˚), but As concentrations are generally too low in solids examined in this study (Table 1) for these solids to contain scorodite in sufficiently high quantities ( > ca. 5%) to be detected by powder diffraction. Soil on flat ground between piles of coarse tailings consists of mixtures of coarse and fine material washed off the nearby heaps, with some sandy tailings in places. Blocks of slag have also rolled off tailings piles and lie on top of, or partially buried in, this flat soil zone. 4.2. Chemistry of solid materials River sediments have elevated Sb and As concentrations (Table 1) compared to background schist (Fig. 5, Craw, 2002). This reflects the occurrence of numerous mineralised veins in the catchment (Pirajno, 1979). These sediments contain approximately equal amounts of Sb and As (Table 1). In contrast, material from the smelter site is generally richer in antimony than arsenic, and some very high antimony concentrations (> 17 wt.% Sb with ca. 4 wt.% As; Table 1) are apparent (Table 1). Concentrations of both metalloids are highly variable, varying by orders of magnitude within and between samples (Table 1). This variation arises because of the irregular distribution of sulfide mineral grains, as shown by microprobe maps in sand tailings (Fig. 4A–D). Unoxidised sand has irregularly scattered grains of stibnite and arsenopyrite in a largely metalloid-free matrix, and this matrix dominates the tailings volume
N.J. Wilson et al. / Environmental Pollution 129 (2004) 257–266
263
Paste pH for smelter site samples varied between 2.4 and 4.2, while EH ranged from +600 to +690 mV (Table 1). The most acid material is the grey (unoxidised) sulfide-bearing sand (Table 1). Soil samples collected from a site upslope of the smelter site had a paste pH of 7.7 and an EH of +445 mV, similar to river water. 4.4. Experimental results
Fig. 5. EH–pH diagram for As and Sb, compiled from Vink (1996), Ashley et al. (2003), and Craw et al. (2003). Antimony species are indicated in bold, and arsenic species are indicated with plain text. Solid arsenic species are stippled. The senarmontite stability field for the indicated solution concentrations is between hatched lines which separate it from stibnite and dissolved antimony. The approximate range of observed EH–pH conditions at the smelter site is shown with a striped ellipse (left).
(Fig. 4A, C). Stibnite grains have detectable As (up to ca. 5%) as solid solution impurity (Fig. 4A, C), so some of the As concentration variations are related to the irregular distribution of stibnite. Oxidised sand has both metalloids more generally dispersed through the matrix, although locally high concentrations persist around altered sulfide grains (Fig. 4B, D). 4.3. Water chemistry River water was typically circumneutral (pH 7.6) and well oxygenated (EH=+450 mV), with a temperature of 10 C. Arsenic and antimony concentrations in the river running alongside the smelter site exhibited no significant changes from upstream to downstream, with only minor fluctuations occurring. Levels of both metalloids were elevated above recommended trigger values for maintenance of ecosystem health (ANZECC, 2000). Sb concentrations were determined to be about 25 mg l1 (trigger value 9 mg l1), while As concentrations fluctuated around 8 mg l1 (trigger value 1 mg l1).
Aerobic experiments released up to 25 mg/kg Sb from all materials initially, and then the amount released subsequently decreased for most materials (Table 2). One coarse tailings sample showed a small increase in released Sb with time, up to 29 mg/kg (Table 2). Brown (oxidised) sand tailings initially released similar amounts of antimony to the unoxidised grey sands, but the decrease in amount released over longer times was less for this brown sand than for the grey sands (Table 2). The anaerobic experiments caused substantially more antimony release from grey sands than under aerobic conditions, with up to ca. 90 mg/kg Sb going into solution (Table 2). There was a decrease in amount released from this grey sand under anaerobic conditions after 14 days (Table 2). Coarse tailings released less arsenic than antimony under aerobic conditions, with less than 4 mg/kg As going into solution and little variation over time (Table 2). In contrast, the grey (unoxidised) sand tailings released over 400 mg/kg As initially under aerobic conditions, and the amount increased with time to almost 800 mg/kg (Table 2). This grey sand released an almost identical amount of As under anaerobic conditions (Table 2). The brown (oxidised) sand showed a similar pattern of As release to the grey sand under aerobic conditions, but the amounts of As released were lower (ca. 100–300 mg/kg, Table 2). Both arsenic and antimony were instantly liberated upon contact with dripping distilled water in the unsaturated experiment (Table 2). More arsenic than antimony was liberated from both grey and brown sand tailings, as occurred in the longer term experiments described above, although the brown sand released only marginally more As than Sb (Table 2). Over 100 mg/kg As was released from the grey sand in this experiment (Table 2). Substantially more of both Sb and As were released from the grey sand than the brown sand, especially for As (Table 2).
5. Discussion 5.1. Redox–pH conditions Redox conditions and pH are the principal parameters that control speciation of Sb and As in the environment, and these can depicted in an Eh–pH diagram (Fig. 5; Vink, 1996; Ashley et al., 2003; Craw et
264
N.J. Wilson et al. / Environmental Pollution 129 (2004) 257–266
al., 2003). This diagram suggests that antimony should be principally in solid form under all but highly oxidised environments. Sb will precipitate as either a sulfide under reducing conditions (stibnite; Fig. 5), or as an oxide, Sb2O3, (=senarmontite in this study) under moderately oxidised conditions. Arsenic should be in solid form as sulfides under reduced conditions, as arsenopyrite, or in solid solution in stibnite (Fig. 5). On oxidation, arsenic is released into solution as a range of aqueous species (Vink, 1996, not shown in Fig. 5) irrespective of pH. In alkaline environments, arsenic remains in solution over the full range of oxidised conditions (Fig. 5). Scorodite can be precipitated under acid oxidised conditions (Fig. 5) provided sufficient Fe3+ is available (Dove and Rimstidt, 1985), or else scorodite-like adsorption complexes arise when dissolved arsenate is adsorbed on to iron oxyhydroxide (Roddick-Lanzilotta et al., 2002). The observed Eh–pH conditions in this study (Table 1; Fig. 5) suggest that senarmontite and scorodite should be stable minerals in all samples studied. Most mesothermal vein systems and mining sites maintain neutral pH in the environment because of the high acid neutralising capacity (ANC) of the host rocks (Ashley et al., 2003). The schist which hosts the vein system in this study is no exception, and typically has ANC of ca. 5 wt.% CaCO3 equivalent (Craw, 2001). The observed low pH of the smelter residues at Endeavour Inlet clearly reflects the excess of acid generation over the neutralising capacity of the remnants of the host rocks. Pyrite is rare in the materials examined in this study, and oxidation of this will contribute a minor amount of acidification only. Smelter slag contains pyrrhotite, presumably formed after addition of scrap iron to the smelting sulfides, and oxidation of this pyrrhotite can contribute to acidification (Jambor, 1994). However, slag is a minor component of the observed residues, and slag forms hard impermeable lumps which have not oxidised beyond the immediate surface zone. Hence, the contribution of slag to environmental acidification is small. Acidification of the Endeavour Inlet site is more likely to be arising from oxidation of the abundant stibnite, and to a lesser extent arsenopyrite. Probable reactions are (Ashley et al., 2003; Craw et al., 2003): Sb2 S3 ðstibniteÞþ3H2 O þ 6O2 ¼ 3SO42 þ 2Sb2 O3 ðsenarmontiteÞ þ 6H þ
ð1Þ
FeAsS ðarsenopyriteÞ þ 7H2 O ¼ Fe2þ ðaqÞ þ H3 AsO3 ðaqÞ þ 11H þ þ 11e þ SO42 ð2Þ Further acidification results when the limited amounts of ferrous iron are oxidised to iron oxyhydroxide which is dispersed through all the oxidised materials. The measured redox potentials for the smelter residues imply that all the materials are moderately oxidised
(Table 1; Fig. 5). Even the grey sand tailings, which contain up to ca. 15% sulfide minerals, are far removed from the redox range of stability of the sulfides (Fig. 5). The low pH of this material is mainly a result of acidification in situ via reaction 1 above. 5.2. Mobility of Sb and As Total liberated antimony is similar across all samples used in our experiments, and is environmentally significant on a time scale of days (Table 2). Our experiments, including the unsaturated experiments, suggest that antimony is readily leached by water from stibnitebearing material on this short time scale. However, our microprobe maps (Fig. 4A, B) show that much antimony has remained in the immediate vicinity (cm scale) of its liberation from sulfides, on the 100 year time scale. Antimony derived from oxidation of stibnite has been largely fixed with iron oxyhydroxide, presumably by adsorption initially, around silicate grain boundaries in oxidised tailings (Fig. 4B). There has been negligible volume loss from tailings during oxidation on the 100 year time scale, so oxidation appears to result in mainly localised redistribution of antimony (Fig. 3 and 4A, B). The elevated amounts of Sb released in our short-term experiments is in accord with observations of Ashley et al. (2003) who examined environmental dissolution of stibnite and Sb oxides, and found Sb to be highly mobile locally. Their theoretical solubility predictions, calibrated with experiments, suggest that senarmontite should dissolve to yield between 10 and 100 mg/l Sb in solution (Ashley et al., 2003). However, Ashley et al.’s (2003) experiments were conducted in iron-free or lowiron systems, where iron oxyhydroxide was rare or absent, and attenuation by iron oxyhydroxides was limited. In the Endeavour Inlet smelter material, iron oxyhydroxide is widespread, and adsorption of dissolved Sb (as in Fig. 4B) is expected. Antimony is probably adsorbed when dissolved as oxyanions, analogous to arsenic oxyanions which are most strongly adsorbed at low pH (Filella et al., 2002; RoddickLanzilotta et al., 2002; Smedley and Kinniburgh, 2002). The lack of mobility of antimony from the smelter site is confirmed by the September water analyses (Table 1). The four water sample sites (Fig. 1) are within the same reach of the river, with no significant tributaries entering along that reach. Hence, the volume of water in the river is approximately constant through the reach. There is no significant difference between the dissolved antimony contents upstream and downstream of the smelter site, and no evidence for a contaminant plume from the site (Table 1) despite the preceding wet winter. Any antimony being leached from the immediate site of the smelter is being attenuated within the alluvial terrace before reaching the river, a distance of tens of metres at most.
N.J. Wilson et al. / Environmental Pollution 129 (2004) 257–266
Arsenic is more readily mobilised than antimony, and dissolved As of several hundred mg/kg can arise from solids that have < 1% As (Table 2). Dissolved arsenic is rapidly attenuated by adsorption to iron oxyhydroxides and/or formation of scorodite in iron-rich systems (Fig. 4C, D; Fig. 5; Roddick-Lanzilotta et al., 2002; Smedley and Kinniburgh, 2002). Hence, arsenic, like antimony, is apparently immobilised in the immediate vicinity of the smelter site and does not reach the adjacent stream (Table 1).
6. Conclusions Antimony bearing residues (up to 17 wt.% Sb) at an historic smelter site in New Zealand have become variably oxidised over the past 100 years. Coarse (cm scale) tailings are almost fully oxidised as they allow free and rapid movement of rain water through their pore spaces. Smelter slag forms hard massive blocks with only incipient surficial oxidation. Sulfide-rich sand tailings still contain centimetre scale zones that are largely unoxidised. Incursion of an oxidation front into these sand tailings has occurred on the 10 cm scale in sand perched in porous coarse tailings, and on the cm scale in lower-lying, frequently saturated beds. Stibnite (Sb2S3) is the most common sulfide mineral in the smelter residues, and this contains up to 5 wt.% arsenic in solid solution. Arsenopyrite (FeAsS) forms up to 5% of unoxidised sandy tailings, and pyrite is rare. Oxidation of these sulfides has caused acidification of the smelter residues, typically to pH 2–5. Oxidation causes some in situ transformation of stibnite to senarmontite (Sb2O3), and minor (cm scale) mobility of antimony. Likewise, arsenic is released into solution by oxidation processes, but moves only short distances (cm scale). Redeposition of both Sb and As occurs via adsorption on to iron oxyhydroxide which pervades most oxidised smelter residues. There is no detectable Sb, As, or acidity released from the smelter site to a stream < 50 m away. Our observations support recent conclusions of Flynn et al. (2003) that antimony is not readily mobilised into the environment at smelter sites despite high antimony contents of the soils. However, we predict that both antimony and arsenic will be dissolved from the smelter residues if these residues are released into higher pH streams or sea water, as adsorbed metalloids are less tightly bound under high pH conditions (Langmuir, 1997; Roddick-Lanzilotta et al., 2002).
Acknowledgements This study was financed by University of Otago and NZ Foundation for Research, Science, and Technology. Department of Conservation kindly gave permission to
265
conduct the study. Sylvia Sander gave helpful advice on analytical techniques, and Debra Chappell assisted with microprobe mapping. Steve Read and Brent Pooley provided able technical assistance throughout the study. Helpful reviews by two anonymous referees substantially improved the manuscript.
References Adriano, D.C.A., 1986. Trace Elements in the Terrestrial Environment. Springer-Verlag, Heidelberg. Aggett, J., Kriegman, M.R., 1987. Preservation of arsenic(III) and arsenic(V) in samples of sediment interstitial water. Analyst 112, 153–157. Ainsworth, N., Cooke, J.A., Johnson, M.S., 1990. Distribution of antimony in contaminated grassland: I: vegetation and soils. Environmental Pollution 65, 65–77. ANZECC, 2000. Australian and New Zealand Guidelines for Freshwater and Marine Water Quality. Australian and New Zealand Environmental and Conservation Council, Canberra. Ashley, P.M., Craw, D., Graham, B.P., Chappell, D.A., 2003. Environmental mobility of antimony around mesothermal stibnite deposits, New South Wales, Australia and southern New Zealand. Journal of Geochemical Exploration 77, 1–14. Baroni, F., Boscagli, A., Protano, G., Riccobono, F., 2000. Antimony accumulation in Achillea ageratum, Plantago lanceoata, and Silene vulgaris growing in an old Sb-mining area. Environmental Pollution 109, 347–352. Craw, D., 2001. Tectonic controls on gold deposits and their environmental impact, New Zealand. Journal of Geochemical Exploration 73, 43–56. Craw, D., Falconer, D., Youngson, J.H., 2003. Environmental arsenopyrite stability and dissolution: theory, experiment, and field observations. Chemical Geology 199, 71–82. Dove, P.M., Rimstidt, J.D., 1985. The solubility and stability of scorodite, FeAsO4.2H2O. American Mineralogist 70, 838–844. Filella, M., Belzile, N., Chen, Y-W., 2002. Antimony in the environment: a review focused on natural waters. I. Occurrence. EarthScience Reviews 57, 125–176. Flynn, H.C.A., Meharg, A.A., Bowyer, P.K., Paton, G.I., 2003. Antimony bioavailability in mine soils. Environmental Pollution 124, 93–100. Foy, C.A.D., Chaney, R.L., White, M.C.A., 1978. The physiology of metal toxicity in plants. Annual Reviews of Plant Physiology 29, 511–566. Gebel, T., 1997. Arsenic and antimony: comparative approach on mechanistic toxicology. Chemico-Biological Interactions 107, 131–144. Gurnani, N., Sharma, A., Tulukder, G., 1994. Effects of antimony on cellular systems in animals: a review. Nucleus 37, 71–96. Jambor, J.L., 1994. Mineralogy of sulphide-rich tailings and their oxidation products. In: Jambor, J.L., Blowes, D.W. (Eds.), Environmental Geochemistry of Sulphide Mine-wastes. Mineralogical Association of Canada Short Course Handbook 22. pp. 59–102. Johnston, M.R., 1993. Gold in a Tin Dish: the Search for Gold in Marlborough and Eastern Nelson. Nikau Press, Nelson. Langmuir, D., 1997. Aqueous Environmental Geochemistry. Prentice Hall, Upper Saddle River, New Jersey. Nordstrom, D.K., 1977. Thermochemical redox equilibria of ZoBell’s solution. Geochimica et Cosmochimica Acta 41, 1835–1841. Ouishi, K., Yasuda, K., Morishige, Y., Hirokawa, K., 1994. Role of metal-matrix modifier in ashing and beginning of the atomization process in graphite furnace-atomic absorption spectrometry. Fresenius Journal of Analytical Chemistry 348, 195–200. Pierce, F.D., Brown, H.R., 1977. Comparison of inorganic inter-
266
N.J. Wilson et al. / Environmental Pollution 129 (2004) 257–266
ferences in atomic absorption spectrometric determination of arsenic and antimony. Analytical Chemistry 49, 1417–1422. Pirajno, F., 1979. Geology, geochemistry, and mineralisation of the Endeavour Inlet antimony-gold prospect, Marlborough Sounds, New Zealand. NZ Journal of Geology and Geophysics 22, 227–236. Ragaini, R.C.A., Ralston, H.R., Roberts, N., 1977. Environmental trace metal contamination in Kellogg, Idaho, near a lead-smelting complex. Environmental Science and Technology 11, 773–781. Roddick-Lanzilotta, A.J., McQuillan, A.J., Craw, D., 2002. Infrared spectroscopic characterisation of arsenate (V) ion adsorption from mine waters, Macraes mine, New Zealand. Applied Geochemistry 17, 445–454. Smedley, P.L., Kinniburgh, D.G., 2002. A review of the source, behaviour and distribution of arsenic in natural waters. Applied Geochemistry 17, 517–568.
Sobek, A.A., Schuller, W.A., Freeman, J.R., Smith, R.M., 1978. Field and Laboratory Methods Applicable to Overburdens and Minesoils (EPA-600/2-78-054). Environmental Protection Agency, Washington, DCA. Vink, B., 1996. Stability relations of antimony and arsenic compounds in the light of revised and extended Eh-pH diagrams. Chemical Geology 130, 21–30. WHO, 1996. Guidelines for Drinking-water Quality, vol. 2: Health Criteria and Other Supporting Information, second ed. World Health Organisation, Geneva. Williams, G.J., 1974. Economic Geology of New Zealand. Australasian Institute of Mining and Metallurgy Monograph 4. Wilson, N.J. 2002. Elevated Arsenic and Antimony Levels in South Island Mesothermal Mineralized Zones. Unpublished MSc Thesis, University of Otago Library, Dunedin, New Zealand.
Antimony distribution and environmental mobility at an historic antimony smelter site, New Zealand N.J. Wilson, D. Craw*, K. Hunter Chemistry and Geology Departments, University of Otago, PO Box 56, Dunedin, New Zealand Received 16 June 2003; accepted 15 October 2003
‘‘Capsule’’: High levels of antimony in primitive smelter soils remain largely immobile on the metre scale.
Abstract A historic antimony smelter site at Endeavour Inlet, New Zealand has smelter residues with up to 17 wt.% antimony. Residues include coarse tailings (cm scale particles, poorly sorted), sand tailings (well sorted) and smelter slag (blocks up to 30 cm across). All of this material has oxidised to some degree over the ca. 100 years since the site was abandoned. Oxidation has resulted in acidification of the residues down to pH 2–5. Smelter slag contains pyrrhotite (FeS) and metallic antimony, and oxidation is restricted to surfaces only. The coarse tailings are the most oxidised, and few sulfide grains persist. Unoxidised sand tailings contain 10–20 vol.% stibnite (Sb2S3) containing up to 5% As, with subordinate arsenopyrite (FeAsS), and minor pyrite (FeS2). The sand tailings are variably oxidised on a scale of 2–10 cm, but original depositional layering is preserved during oxidation and formation of senarmontite (Sb2O3). Oxidation of sand tailings has resulted in localised mobility of both Sb and As on the cm scale, resulting in redistribution of these metalloids with iron oxyhydroxide around sand grain boundaries. Experiments demonstrate that Sb mobility decreases with time on a scale of days. Attenuation of both As and Sb occurs due to adsorption on to iron oxyhydroxides which are formed during oxidation of the smelter residues. There is no detectable loss of Sb or As from the smelter site into the adjacent river, <50 m away, which has elevated Sb (ca. 20 mg/l) and As (ca. 7 mg /l) from mineralised rocks upstream. Despite the high concentrations of Sb and As in the smelter residues, these metalloids are not being released into the environment. # 2003 Elsevier Ltd. All rights reserved. Keywords: Antimony; Arsenic; Oxidation; Adsorption; Stibnite; Arsenopyrite; Senarmontite; Metalloids; Mining
1. Introduction Antimony is increasingly being identified as a toxic environmental pollutant and has been implicated in cancer development (Foy et al., 1978; Gurnani et al., 1994; Gebel, 1997). The principal global sources of anthropogenic antimony pollution are the mining and smelting industries (Adriano, 1986), and very high levels of pollution have been detected around smelter sites in particular (Ragaini et al., 1977; Ainsworth et al., 1990; Baroni et al., 2000; Flynn et al., 2003). However, little is known about the environmental mobility of antimony in such situations (Filella et al., 2002). Baroni et al. (2000) have shown that some plants growing on smelter * Corresponding author. Tel.: +64-3-479-7519; fax: +64-3-4797527. E-mail address: [email protected] (D. Craw). 0269-7491/$ - see front matter # 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.envpol.2003.10.014
residues can accumulate high levels of antimony. Conversely, Flynn et al. (2003) have shown that despite very high levels of soil antimony at mine and smelter sites, antimony bioavailability is low, and the antimony is fixed in the soils. Flynn et al.’s (2003) study, and earlier studies, did not investigate the disposition of the antimony in the mine soils, as they were dealing with well-developed organic-rich soils in which original mine and smelter materials were dispersed. In this study, we address the nature of antimony-bearing material in primitive soils at an historic smelter site that has remained undisturbed and has negligible revegetation after ca. 100 years. This study examines the environmental mobility of antimony in the early stages of soil development at the smelter site, and the processes which fix antimony in the primitive soils. Both arsenic and antimony are chalcophilic group V metalloids that form oxyanions in water, and therefore
258
N.J. Wilson et al. / Environmental Pollution 129 (2004) 257–266
these elements have some similarities in chemical behaviour in the environment. Arsenic is considered to be toxic and carcinogenic in a similar way to antimony (Gebel, 1997) and is currently regarded as one of the most serious inorganic drinking water pollutants in the world (Smedley and Kinniburgh, 2002). Maximum drinking water limits have been set by the World Health Organisation at 10 and 5 mg l1 for arsenic and antimony, respectively, with the latter value provisional until a greater understanding of its toxicity (WHO, 1996). Because of these similarities in environmental chemistry and toxicity between antimony and arsenic, this paper examines environmental arsenic mobility in the smelter soils in conjunction with our study of antimony, to allow direct comparison. The site we use facilitates this comparison because antimony and arsenic are present together at strongly elevated levels.
2.2. Historical mining and smelting Mining for antimony began in the hills above Endeavour Inlet in 1874, following discovery of stibnite bearing veins in 1873 (Johnston, 1993). In 1875, a smelter was erected to process the mined ore, on an alluvial terrace at the base of the hills about 1 km from Endeavour Inlet (Fig. 1). The smelter initially used four cast iron retorts and wood-fired furnaces to melt the ore, which was mixed with scrap iron to produce 90–92% pure antimony (Johnston, 1993). In 1885, the smelter was redeveloped with a large sorting and crushing facility and a more complex smelting process. Concentrated ore was first melted using wood-fired furnaces, then remelted at higher temperatures using coke, before final production of 100% antimony (Johnston, 1993). Further refinements followed, but falling antimony prices saw the operation progressively scaled down, and the site was abandoned by 1907 (Johnston, 1993).
2. General setting 2.1. Geological background The Endeavour Inlet area that is the subject of this study is underlain by schistose Mesozoic sediments that make up most of the South Island of New Zealand (Fig. 1). These schists are locally known as Marlborough Schist, but are genetically related to the more extensive Otago Schist in the south of the South Island where mesothermal antimony deposits occur also (Williams, 1974; Ashley et al., 2003). The Marlborough Schist has uniform mineralogy dominated by quartz, albite, muscovite, chlorite and calcite. The rocks are extensively recrystallised and have a near-pervasive foliation that is generally flat-lying or gently dipping. Late Tertiary tectonics have caused geomorphic rejuvenation, resulting in rugged ridges and steeply incised streams. Principal valleys have been flooded by the sea to make an intricate set of sounds, and Endeavour Inlet (Fig. 1) is a minor bay within this sound complex. The climate is temperate (mean annual temperature ca. 12 C) with ca. 1000 mm/year of rainfall that supports locally lush stands of forest. The foliation of the Marlborough Schist is cut by steeply dipping mesothermal quartz veins that are variably mineralised with pyrite, arsenopyrite, scheelite, gold, and/or stibnite (Williams, 1974). The mines at Endeavour Inlet were developed in a swarm of mineralised veins (typically < 1 m wide) dominated by quartz, stibnite, and arsenopyrite (Pirajno, 1979). The veins form a swarm of sub-parallel structures in a zone approximately 200m wide, which is continuous for at least 5 km north of Endeavour Inlet. Mineralisation was zoned with depth, so that at lower levels arsenic is the dominant hydrothermal element, while at higher levels antimony is more common (Pirajno, 1979).
Fig. 1. Location map of the Endeavour Inlet antimony smelter site, and sites of water analyses in the adjacent river (site numbers refer to analyses in (Table 1). The site is on alluvium in a valley, and approximate groundwater flow directions in that alluvium are indicated, based on topography and surface water flows.
N.J. Wilson et al. / Environmental Pollution 129 (2004) 257–266
Now, approximately 100 years since the smelter was scaled down, no equipment remains at the site and the site is unpopulated. Farms and tourist accommodation dot the shore of the nearby inlet, but the population is small (< 100). Much of the valley upstream of the site has reverted to native forest and is now a natural reserve. The smelter site is marked by an area of ca. 1 ha on which little revegetation has occurred. There are several distinctly different types of material present at the site, and these are discussed separately in this paper. Most of the material at the site consists of coarse tailings from the crushing and sorting plant, mixed with slag from the smelting process. In addition, there is sand rich in sulfide minerals derived from the crushing and concentration plant. Unoxidised sand is grey and this is surrounded by brown oxidised sand (Fig. 2 A, B; 3).
3. Methods 3.1. Field sampling Samples of solid material were collected from the tailings heaps and nearby soil at the downstream of the smelter site (Figs. 1 and 2A, B). The sulfide sand is variably oxidised, so associated oxidised and unoxidised material were sampled. These samples were kept dry to
Fig. 2. Sketch sections (constructed from photographs of shallow excavations) through smelter residues, showing scales and relative distributions of different types of materials. A. Coarse tailings pile at the smelter site, with layers of sand tailings and cobbles of smelter slag. B. Sand tailings at the outwash area downstream of the smelter site (Fig. 1).
259
prevent further oxidation after sampling. Fine grained (sand and finer) stream sediments were sampled from the bed of the main stream that passes the smelter site, at points indicated in Fig. 1. Labplus sterile CS/500 bags were used to collect solid samples. Water samples were collected from the river at the same points as sediment samples (Fig. 1). A set of four water samples upstream and downstream of the smelter site was taken in September when the site had been through a wet winter season and water levels were still high. Water samples were collected in triplicate at each site in acid-washed 500 ml high-density polyethylene (HDPE) bottles. The water samples were clear with no obvious particulate material or brown iron oxyhydroxide colloidal material. Samples were not filtered and chemical preservatives were not used (Pierce and Brown, 1977; Aggett and Kriegman, 1987). Water and soil pH and redox potential were measured in the field using an Oakton WD-35615 meter calibrated with standard pH solutions and Zobell’s solution for redox potential (Nordstrom, 1977). Paste-pH measurements (Sobek et al., 1978) were made at points where insufficient soil moisture was present for direct measurement. 3.2. Analytical methods Water collected from the smelter site was analysed by graphite furnace atomic absorption spectrometry (GFAAS) for As and Sb using a Perkin-Elmer 4100ZL graphite furnace. Pd/Mg(NO3)2 was added to sample aliquots as a matrix modifier to overcome incomplete atomisation (Ouishi et al., 1994). The detection limit for both Sb and As was 5 mg l1. Results are presented in Table 1. Determination of antimony speciation by voltammetry as part of a larger study by Wilson (2002) demonstrated that nearly all dissolved Sb in surface waters is present as the Sb(V) species. Solid samples were dried at 60 C before being ground to < 50 mm using a TEMA rock crusher. Sample powders were treated by nitric/hydrochloric acid digestion (USEPA 200.2) followed by solution analysis for As and Sb by inductively-coupled plasma (ICP) at ALS Chemex, Brisbane, Australia, or Hill Laboratories, Hamilton, New Zealand. Results from the different ICP labs are within 10% of each other. Detection limits were < 1 mg/kg, and all analytical errors, as indicated by replication of analyses and separate analyses of triplicate water samples, are contained within the symbols of diagrams in this study. Results are presented in Table 1. Minerals were identified in thin and polished sections by light microscopy, augmented with X-ray diffraction (XRD) and electron microprobe semiquantitative analysis. XRD analysis was performed on powder smears crushed finely in an agate mortar and pestle, using Cu Ka radiation on a Philips PW 1050 diffractometer.
260
N.J. Wilson et al. / Environmental Pollution 129 (2004) 257–266
Antimony and arsenic element maps of small (0.500.45 mm) areas of polished sections were generated using a Jeol microprobe at 25 kV by automated stepping (1 micron steps) across the area. The resultant images (Fig. 4A–D) are compilations of the results of 225 000 one-second analyses, with darkest patches representing highest concentrations. White matrix or
sparsely coloured patches in these maps reflect zones near to, or below, background levels of ca. 0.5 wt.% Sb or As. 3.3. Experimental study In order to simulate short-term mobility of As and Sb from the solid smelter residues in the environment, three dissolution or leaching experiments were performed. Distilled water was used as a leaching fluid in these simulations. Results of duplicate experiments, as outlined below, have been averaged for presentation in Table 2. To simulate aerobic dissolution, an experiment based on those of Ashley et al. (2003) was set up. Approximately 5 g of solid material were accurately weighed into LDPE plastic jars (60 ml capacity), then 15 ml of distilled water were added and the jars sealed. It was assumed that enough oxygen was present within the jars that any sediment dissolution would proceed via aerobic pathways. These jars were then stored in darkness at ca. 20 C for up to 28 days. Sufficient jars were prepared so that the amount of dissolved arsenic and antimony could be measured after periods of 1, 7, 14 and 28 days, then the jar discarded. Duplicate samples from three different tailings heaps, grey sulfide sand, and brown (oxidised sulfide) sand, were prepared. After sufficient time had elapsed, 10 ml of water from each sample were pipetted into two 5 ml plastic vials. The water in these vials was then analysed for As and Sb by ICP–OES using a Thermo Jarrell Ash AtomScan 25 ICP-spectrometer. Analyses were recalculated to (mg metalloid released)/(kg solid material in experiment) to quantify amounts of metalloids released (Table 2) rather than mineral solubility (cf Ashley et al., 2003).
Fig. 3. Photograph of a thin section through the boundary between grey, least oxidised, sand tailings (within dashed white lines) and brown oxidised sand tailings from the smelter site. Original depositional layering (near-horizontal) is accentuated by iron oxyhydroxide (dark) in the oxidised areas, and to a lesser extent in some layers in unoxidised zones. A coarser sand layer at top has been extensively cemented with iron oxyhydroxide. The layering has undergone some post-depositional deformation, resulting in warps and a sharp step just left of centre. Pale grey material around the tailings is mounting glue. Table 1 Chemical analyses of solids and waters at the smelter site Site
Material
Solid Sb (mg/kg)
Solid As (mg/kg)
River water Sb (mg/l)
River water As (mg/l)
pH
Stream 1 Stream 2 Stream 2 Stream 2 Stream 3 Stream 4 Smelter Smelter Smelter Smelter Smelter Smelter Smelter Smelter Smelter Smelter Smelter Smelter Smelter
Sediment Sediment Sediment Sediment Sediment Sediment Soil Coarse tailings Coarse tailings Coarse tailings Coarse tailings Coarse tailings Coarse tailings Slag Slag Grey sand Grey sand Brown sand Brown sand
150 243 68 38 22 18 80 200 40 100 47 000 34 600 65 000 176 700 45 100 123 500 22 200 132 900 35 100 88 600 25 800
67 190 125 83 14 27 2220 8350 4320 4610 4870 203 65.2 135 36 9120 6110 8550 5630
23.5 14.1 30.4 22.9 24.4 23.8
7.59 6.5 5.49 8.33 8.41 8.64
7.6 7.6 7.6 7.6 7.6 7.6
Eh (mV)
472
4.2
609
4.1 3.1
618 662
2.4 2.6 3.9
690 673 627
Date September April June September September September
261
N.J. Wilson et al. / Environmental Pollution 129 (2004) 257–266
Fig. 4. Microprobe element maps of unoxidised (left) and oxidised (right) sand tailings, for antimony (top) and arsenic (bottom). See text for construction methods. Both A and C are maps of a single area, and B and D are maps of a single separate area. Sulfide grains are readily apparent in A and C (maximum concentration examples arrowed). Ellipse in A encloses an example of secondary antimony precipitated with iron oxyhydroxides around sand grain boundaries. Diffuse distribution of Sb and As in oxidised sand is shown in B and D, with maximum concentrations indicated. Table 2 Results of experimental mobilisation of Sb and As from smelter residues. Experiments were conducted in duplicate, and results averaged Experiment
Aerobic Aerobic Aerobic Aerobic Unsaturated Anaerobic Anaerobic Anaerobic
Time (days)
1 7 14 28 1 7 14
Grey sand (mg /kg)
Brown sand (mg /kg)
Coarse tailings 1 (mg /kg)
Coarse tailings 2 (mg/ kg)
Coarse tailings 3 (mg/ kg)
Sb
As
Sb
As
Sb
As
Sb
As
Sb
As
21.4 7.56 7.71 7.75 15.7 82.1 89 58.2
411 601 778 768 108 335 690 701
19.9 14.6 16.1 15.9 5.46
113 194 214 306 8.48
6.72 4.72 6.06 4.46
2.9 2.77 3.37 2.93
18.4 3.12 3.28 4.14
2.7 1.74 1.71 1.74
24.3 26.3 27.3 28.8
1.9 2.38 2.42 2.31
N/M
A parallel experiment was constructed using grey sulfide sand in order to evaluate anaerobic dissolution of Sb and As. Distilled water was deoxygenated by bubbling it with nitrogen gas for at least 20 min. About 15 g of sample were accurately weighed into 60 ml jars and the jars weighed with the lids on. The jars were then filled to the brim with deoxygenated water, sealed and reweighed to
N/M
determine the amount of water added. These jars were then placed in sealable plastic bags which were filled with deoxygenated water, sealed and left in darkness at ca. 20 C. Sufficient samples were prepared to allow for duplicated aliquots to be analysed after 1, 7 and 14 days. A third experiment was designed to simulate As and Sb mobilisation under unsaturated conditions. Distilled
262
N.J. Wilson et al. / Environmental Pollution 129 (2004) 257–266
water was dripped at a rate of approximately 1 ml s1 through a burette into a funnel in which 5 g of grey or brown sand rested upon a 0.45 mm filter. A Buchner funnel was used so that water would be sucked through the sediment and filter to fall into a clean 25 ml plastic vial. Once 15 ml of water had been collected, the burette was closed and the water collected was analysed for As and Sb by ICP–OES. This experiment was performed in duplicate for both grey and brown sand.
4. Results 4.1. Description and mineralogy Coarse tailings heaps are up to 5 m high and 20 m long, and consist mainly of cm scale pebbles, with some finer grained matrix. Heaps have high porosity and water is not retained, so the coarse tailings are dry except during rain events. Quartz is the most common mineral in the tailings, and up to 30% of the tailings are monominerallic quartz clasts. Other clasts are quartzbearing but include fragments of basement schist. Most clasts have sulfide mineral remnants (stibnite and arsenopyrite) that have been extensively oxidised, and these tailings are a deep brown colour. At least some of the matrix is decomposed sulfidic schist fragments pervaded by iron oxyhydroxide. XRD analysis shows only quartz with minor muscovite in the matrix material. Blocks of slag from the smelter are common on the surfaces of the coarse tailings heaps and intervening flat ground, and are also distributed through the coarse tailings within the heaps. These blocks are up to 30 cm across and irregular in shape. Slag is hard but porous and locally vesicular. Some blocks of slag still retain the shape of the crucibles into which the molten ore was poured. Quartz is still recognisable within some slag samples, as millimetre scale grains. Much of the slag is made up of pyrrhotite and fine grained micaceous material, and locally graphite zones up to 2 cm thick occur. Droplets of antimony metal up to 1 mm across occur distributed throughout the slag. Slag blocks are variably oxidised to deep brown on the surfaces, but fresh metallic minerals are readily seen on broken surfaces. Unoxidised sulfide bearing sandy tailings are grey in hand specimen, and are soft, pliable, and cohesive. Most of this material was water-saturated or at least moist when first exposed. The sand occurs in thin (ca. 20 cm) layers within heaps of coarse tailings, and as an outwash deposit on the alluvial terrace up to 100 m from the smelter site (Fig. 2A, B). The sand is well-sorted with angular clasts of quartz, micas, and sulfides. Layering at the mm to cm scale is preserved locally, although some post-depositional deformation has distorted this layering (Fig. 3). The layering is defined by varying proportions of sulfides and silicates. Stibnite
is the most common sulfide (10–20%), but arsenopyrite grains (2–5%) are scattered throughout. Pyrite grains are rare (< 1%). Evidence for incipient oxidation of this grey sand can be seen in the microprobe map, which shows faint secondary seams of mobilised antimony around grains of silicates (Fig. 4A, ellipse area). These seams follow zones of pigmentation by iron oxyhydroxide along sand grain boundaries. The grey unoxidised tailings are preserved only as irregular remnant patches within the sandy tailings horizons. These patches are surrounded by pale brown oxidised sand, particularly in layers perched in porous sediments (Fig. 2A, B). The least altered horizon lies on flat ground on the smelter site (Fig. 2A), and appears to be commonly water-saturated. Individual depositional layers (mm scale) can be traced from unoxidised to oxidised sand (Fig. 3). Layers in the oxidised sand are defined by varying degrees of pigmentation by brown iron oxyhydroxide. No sulfide minerals remain in oxidised sand, but sulfide grain shapes are preserved indistinctly (Fig. 4A–D). XRD analysis reveals quartz as the dominant mineral, swamping most other lines. A prominent XRD line at 3.2 A˚ in oxidised sand suggests that senarmontite (SbIII 2 O3) is present, but other lines for this mineral are not apparent. No other metallic minerals have been identified. The 3.2 A˚ line attributed to senarmontite may overlap the principal scorodite (FeAsO4.2H2O) line (3.18 A˚), but As concentrations are generally too low in solids examined in this study (Table 1) for these solids to contain scorodite in sufficiently high quantities ( > ca. 5%) to be detected by powder diffraction. Soil on flat ground between piles of coarse tailings consists of mixtures of coarse and fine material washed off the nearby heaps, with some sandy tailings in places. Blocks of slag have also rolled off tailings piles and lie on top of, or partially buried in, this flat soil zone. 4.2. Chemistry of solid materials River sediments have elevated Sb and As concentrations (Table 1) compared to background schist (Fig. 5, Craw, 2002). This reflects the occurrence of numerous mineralised veins in the catchment (Pirajno, 1979). These sediments contain approximately equal amounts of Sb and As (Table 1). In contrast, material from the smelter site is generally richer in antimony than arsenic, and some very high antimony concentrations (> 17 wt.% Sb with ca. 4 wt.% As; Table 1) are apparent (Table 1). Concentrations of both metalloids are highly variable, varying by orders of magnitude within and between samples (Table 1). This variation arises because of the irregular distribution of sulfide mineral grains, as shown by microprobe maps in sand tailings (Fig. 4A–D). Unoxidised sand has irregularly scattered grains of stibnite and arsenopyrite in a largely metalloid-free matrix, and this matrix dominates the tailings volume
N.J. Wilson et al. / Environmental Pollution 129 (2004) 257–266
263
Paste pH for smelter site samples varied between 2.4 and 4.2, while EH ranged from +600 to +690 mV (Table 1). The most acid material is the grey (unoxidised) sulfide-bearing sand (Table 1). Soil samples collected from a site upslope of the smelter site had a paste pH of 7.7 and an EH of +445 mV, similar to river water. 4.4. Experimental results
Fig. 5. EH–pH diagram for As and Sb, compiled from Vink (1996), Ashley et al. (2003), and Craw et al. (2003). Antimony species are indicated in bold, and arsenic species are indicated with plain text. Solid arsenic species are stippled. The senarmontite stability field for the indicated solution concentrations is between hatched lines which separate it from stibnite and dissolved antimony. The approximate range of observed EH–pH conditions at the smelter site is shown with a striped ellipse (left).
(Fig. 4A, C). Stibnite grains have detectable As (up to ca. 5%) as solid solution impurity (Fig. 4A, C), so some of the As concentration variations are related to the irregular distribution of stibnite. Oxidised sand has both metalloids more generally dispersed through the matrix, although locally high concentrations persist around altered sulfide grains (Fig. 4B, D). 4.3. Water chemistry River water was typically circumneutral (pH 7.6) and well oxygenated (EH=+450 mV), with a temperature of 10 C. Arsenic and antimony concentrations in the river running alongside the smelter site exhibited no significant changes from upstream to downstream, with only minor fluctuations occurring. Levels of both metalloids were elevated above recommended trigger values for maintenance of ecosystem health (ANZECC, 2000). Sb concentrations were determined to be about 25 mg l1 (trigger value 9 mg l1), while As concentrations fluctuated around 8 mg l1 (trigger value 1 mg l1).
Aerobic experiments released up to 25 mg/kg Sb from all materials initially, and then the amount released subsequently decreased for most materials (Table 2). One coarse tailings sample showed a small increase in released Sb with time, up to 29 mg/kg (Table 2). Brown (oxidised) sand tailings initially released similar amounts of antimony to the unoxidised grey sands, but the decrease in amount released over longer times was less for this brown sand than for the grey sands (Table 2). The anaerobic experiments caused substantially more antimony release from grey sands than under aerobic conditions, with up to ca. 90 mg/kg Sb going into solution (Table 2). There was a decrease in amount released from this grey sand under anaerobic conditions after 14 days (Table 2). Coarse tailings released less arsenic than antimony under aerobic conditions, with less than 4 mg/kg As going into solution and little variation over time (Table 2). In contrast, the grey (unoxidised) sand tailings released over 400 mg/kg As initially under aerobic conditions, and the amount increased with time to almost 800 mg/kg (Table 2). This grey sand released an almost identical amount of As under anaerobic conditions (Table 2). The brown (oxidised) sand showed a similar pattern of As release to the grey sand under aerobic conditions, but the amounts of As released were lower (ca. 100–300 mg/kg, Table 2). Both arsenic and antimony were instantly liberated upon contact with dripping distilled water in the unsaturated experiment (Table 2). More arsenic than antimony was liberated from both grey and brown sand tailings, as occurred in the longer term experiments described above, although the brown sand released only marginally more As than Sb (Table 2). Over 100 mg/kg As was released from the grey sand in this experiment (Table 2). Substantially more of both Sb and As were released from the grey sand than the brown sand, especially for As (Table 2).
5. Discussion 5.1. Redox–pH conditions Redox conditions and pH are the principal parameters that control speciation of Sb and As in the environment, and these can depicted in an Eh–pH diagram (Fig. 5; Vink, 1996; Ashley et al., 2003; Craw et
264
N.J. Wilson et al. / Environmental Pollution 129 (2004) 257–266
al., 2003). This diagram suggests that antimony should be principally in solid form under all but highly oxidised environments. Sb will precipitate as either a sulfide under reducing conditions (stibnite; Fig. 5), or as an oxide, Sb2O3, (=senarmontite in this study) under moderately oxidised conditions. Arsenic should be in solid form as sulfides under reduced conditions, as arsenopyrite, or in solid solution in stibnite (Fig. 5). On oxidation, arsenic is released into solution as a range of aqueous species (Vink, 1996, not shown in Fig. 5) irrespective of pH. In alkaline environments, arsenic remains in solution over the full range of oxidised conditions (Fig. 5). Scorodite can be precipitated under acid oxidised conditions (Fig. 5) provided sufficient Fe3+ is available (Dove and Rimstidt, 1985), or else scorodite-like adsorption complexes arise when dissolved arsenate is adsorbed on to iron oxyhydroxide (Roddick-Lanzilotta et al., 2002). The observed Eh–pH conditions in this study (Table 1; Fig. 5) suggest that senarmontite and scorodite should be stable minerals in all samples studied. Most mesothermal vein systems and mining sites maintain neutral pH in the environment because of the high acid neutralising capacity (ANC) of the host rocks (Ashley et al., 2003). The schist which hosts the vein system in this study is no exception, and typically has ANC of ca. 5 wt.% CaCO3 equivalent (Craw, 2001). The observed low pH of the smelter residues at Endeavour Inlet clearly reflects the excess of acid generation over the neutralising capacity of the remnants of the host rocks. Pyrite is rare in the materials examined in this study, and oxidation of this will contribute a minor amount of acidification only. Smelter slag contains pyrrhotite, presumably formed after addition of scrap iron to the smelting sulfides, and oxidation of this pyrrhotite can contribute to acidification (Jambor, 1994). However, slag is a minor component of the observed residues, and slag forms hard impermeable lumps which have not oxidised beyond the immediate surface zone. Hence, the contribution of slag to environmental acidification is small. Acidification of the Endeavour Inlet site is more likely to be arising from oxidation of the abundant stibnite, and to a lesser extent arsenopyrite. Probable reactions are (Ashley et al., 2003; Craw et al., 2003): Sb2 S3 ðstibniteÞþ3H2 O þ 6O2 ¼ 3SO42 þ 2Sb2 O3 ðsenarmontiteÞ þ 6H þ
ð1Þ
FeAsS ðarsenopyriteÞ þ 7H2 O ¼ Fe2þ ðaqÞ þ H3 AsO3 ðaqÞ þ 11H þ þ 11e þ SO42 ð2Þ Further acidification results when the limited amounts of ferrous iron are oxidised to iron oxyhydroxide which is dispersed through all the oxidised materials. The measured redox potentials for the smelter residues imply that all the materials are moderately oxidised
(Table 1; Fig. 5). Even the grey sand tailings, which contain up to ca. 15% sulfide minerals, are far removed from the redox range of stability of the sulfides (Fig. 5). The low pH of this material is mainly a result of acidification in situ via reaction 1 above. 5.2. Mobility of Sb and As Total liberated antimony is similar across all samples used in our experiments, and is environmentally significant on a time scale of days (Table 2). Our experiments, including the unsaturated experiments, suggest that antimony is readily leached by water from stibnitebearing material on this short time scale. However, our microprobe maps (Fig. 4A, B) show that much antimony has remained in the immediate vicinity (cm scale) of its liberation from sulfides, on the 100 year time scale. Antimony derived from oxidation of stibnite has been largely fixed with iron oxyhydroxide, presumably by adsorption initially, around silicate grain boundaries in oxidised tailings (Fig. 4B). There has been negligible volume loss from tailings during oxidation on the 100 year time scale, so oxidation appears to result in mainly localised redistribution of antimony (Fig. 3 and 4A, B). The elevated amounts of Sb released in our short-term experiments is in accord with observations of Ashley et al. (2003) who examined environmental dissolution of stibnite and Sb oxides, and found Sb to be highly mobile locally. Their theoretical solubility predictions, calibrated with experiments, suggest that senarmontite should dissolve to yield between 10 and 100 mg/l Sb in solution (Ashley et al., 2003). However, Ashley et al.’s (2003) experiments were conducted in iron-free or lowiron systems, where iron oxyhydroxide was rare or absent, and attenuation by iron oxyhydroxides was limited. In the Endeavour Inlet smelter material, iron oxyhydroxide is widespread, and adsorption of dissolved Sb (as in Fig. 4B) is expected. Antimony is probably adsorbed when dissolved as oxyanions, analogous to arsenic oxyanions which are most strongly adsorbed at low pH (Filella et al., 2002; RoddickLanzilotta et al., 2002; Smedley and Kinniburgh, 2002). The lack of mobility of antimony from the smelter site is confirmed by the September water analyses (Table 1). The four water sample sites (Fig. 1) are within the same reach of the river, with no significant tributaries entering along that reach. Hence, the volume of water in the river is approximately constant through the reach. There is no significant difference between the dissolved antimony contents upstream and downstream of the smelter site, and no evidence for a contaminant plume from the site (Table 1) despite the preceding wet winter. Any antimony being leached from the immediate site of the smelter is being attenuated within the alluvial terrace before reaching the river, a distance of tens of metres at most.
N.J. Wilson et al. / Environmental Pollution 129 (2004) 257–266
Arsenic is more readily mobilised than antimony, and dissolved As of several hundred mg/kg can arise from solids that have < 1% As (Table 2). Dissolved arsenic is rapidly attenuated by adsorption to iron oxyhydroxides and/or formation of scorodite in iron-rich systems (Fig. 4C, D; Fig. 5; Roddick-Lanzilotta et al., 2002; Smedley and Kinniburgh, 2002). Hence, arsenic, like antimony, is apparently immobilised in the immediate vicinity of the smelter site and does not reach the adjacent stream (Table 1).
6. Conclusions Antimony bearing residues (up to 17 wt.% Sb) at an historic smelter site in New Zealand have become variably oxidised over the past 100 years. Coarse (cm scale) tailings are almost fully oxidised as they allow free and rapid movement of rain water through their pore spaces. Smelter slag forms hard massive blocks with only incipient surficial oxidation. Sulfide-rich sand tailings still contain centimetre scale zones that are largely unoxidised. Incursion of an oxidation front into these sand tailings has occurred on the 10 cm scale in sand perched in porous coarse tailings, and on the cm scale in lower-lying, frequently saturated beds. Stibnite (Sb2S3) is the most common sulfide mineral in the smelter residues, and this contains up to 5 wt.% arsenic in solid solution. Arsenopyrite (FeAsS) forms up to 5% of unoxidised sandy tailings, and pyrite is rare. Oxidation of these sulfides has caused acidification of the smelter residues, typically to pH 2–5. Oxidation causes some in situ transformation of stibnite to senarmontite (Sb2O3), and minor (cm scale) mobility of antimony. Likewise, arsenic is released into solution by oxidation processes, but moves only short distances (cm scale). Redeposition of both Sb and As occurs via adsorption on to iron oxyhydroxide which pervades most oxidised smelter residues. There is no detectable Sb, As, or acidity released from the smelter site to a stream < 50 m away. Our observations support recent conclusions of Flynn et al. (2003) that antimony is not readily mobilised into the environment at smelter sites despite high antimony contents of the soils. However, we predict that both antimony and arsenic will be dissolved from the smelter residues if these residues are released into higher pH streams or sea water, as adsorbed metalloids are less tightly bound under high pH conditions (Langmuir, 1997; Roddick-Lanzilotta et al., 2002).
Acknowledgements This study was financed by University of Otago and NZ Foundation for Research, Science, and Technology. Department of Conservation kindly gave permission to
265
conduct the study. Sylvia Sander gave helpful advice on analytical techniques, and Debra Chappell assisted with microprobe mapping. Steve Read and Brent Pooley provided able technical assistance throughout the study. Helpful reviews by two anonymous referees substantially improved the manuscript.
References Adriano, D.C.A., 1986. Trace Elements in the Terrestrial Environment. Springer-Verlag, Heidelberg. Aggett, J., Kriegman, M.R., 1987. Preservation of arsenic(III) and arsenic(V) in samples of sediment interstitial water. Analyst 112, 153–157. Ainsworth, N., Cooke, J.A., Johnson, M.S., 1990. Distribution of antimony in contaminated grassland: I: vegetation and soils. Environmental Pollution 65, 65–77. ANZECC, 2000. Australian and New Zealand Guidelines for Freshwater and Marine Water Quality. Australian and New Zealand Environmental and Conservation Council, Canberra. Ashley, P.M., Craw, D., Graham, B.P., Chappell, D.A., 2003. Environmental mobility of antimony around mesothermal stibnite deposits, New South Wales, Australia and southern New Zealand. Journal of Geochemical Exploration 77, 1–14. Baroni, F., Boscagli, A., Protano, G., Riccobono, F., 2000. Antimony accumulation in Achillea ageratum, Plantago lanceoata, and Silene vulgaris growing in an old Sb-mining area. Environmental Pollution 109, 347–352. Craw, D., 2001. Tectonic controls on gold deposits and their environmental impact, New Zealand. Journal of Geochemical Exploration 73, 43–56. Craw, D., Falconer, D., Youngson, J.H., 2003. Environmental arsenopyrite stability and dissolution: theory, experiment, and field observations. Chemical Geology 199, 71–82. Dove, P.M., Rimstidt, J.D., 1985. The solubility and stability of scorodite, FeAsO4.2H2O. American Mineralogist 70, 838–844. Filella, M., Belzile, N., Chen, Y-W., 2002. Antimony in the environment: a review focused on natural waters. I. Occurrence. EarthScience Reviews 57, 125–176. Flynn, H.C.A., Meharg, A.A., Bowyer, P.K., Paton, G.I., 2003. Antimony bioavailability in mine soils. Environmental Pollution 124, 93–100. Foy, C.A.D., Chaney, R.L., White, M.C.A., 1978. The physiology of metal toxicity in plants. Annual Reviews of Plant Physiology 29, 511–566. Gebel, T., 1997. Arsenic and antimony: comparative approach on mechanistic toxicology. Chemico-Biological Interactions 107, 131–144. Gurnani, N., Sharma, A., Tulukder, G., 1994. Effects of antimony on cellular systems in animals: a review. Nucleus 37, 71–96. Jambor, J.L., 1994. Mineralogy of sulphide-rich tailings and their oxidation products. In: Jambor, J.L., Blowes, D.W. (Eds.), Environmental Geochemistry of Sulphide Mine-wastes. Mineralogical Association of Canada Short Course Handbook 22. pp. 59–102. Johnston, M.R., 1993. Gold in a Tin Dish: the Search for Gold in Marlborough and Eastern Nelson. Nikau Press, Nelson. Langmuir, D., 1997. Aqueous Environmental Geochemistry. Prentice Hall, Upper Saddle River, New Jersey. Nordstrom, D.K., 1977. Thermochemical redox equilibria of ZoBell’s solution. Geochimica et Cosmochimica Acta 41, 1835–1841. Ouishi, K., Yasuda, K., Morishige, Y., Hirokawa, K., 1994. Role of metal-matrix modifier in ashing and beginning of the atomization process in graphite furnace-atomic absorption spectrometry. Fresenius Journal of Analytical Chemistry 348, 195–200. Pierce, F.D., Brown, H.R., 1977. Comparison of inorganic inter-
266
N.J. Wilson et al. / Environmental Pollution 129 (2004) 257–266
ferences in atomic absorption spectrometric determination of arsenic and antimony. Analytical Chemistry 49, 1417–1422. Pirajno, F., 1979. Geology, geochemistry, and mineralisation of the Endeavour Inlet antimony-gold prospect, Marlborough Sounds, New Zealand. NZ Journal of Geology and Geophysics 22, 227–236. Ragaini, R.C.A., Ralston, H.R., Roberts, N., 1977. Environmental trace metal contamination in Kellogg, Idaho, near a lead-smelting complex. Environmental Science and Technology 11, 773–781. Roddick-Lanzilotta, A.J., McQuillan, A.J., Craw, D., 2002. Infrared spectroscopic characterisation of arsenate (V) ion adsorption from mine waters, Macraes mine, New Zealand. Applied Geochemistry 17, 445–454. Smedley, P.L., Kinniburgh, D.G., 2002. A review of the source, behaviour and distribution of arsenic in natural waters. Applied Geochemistry 17, 517–568.
Sobek, A.A., Schuller, W.A., Freeman, J.R., Smith, R.M., 1978. Field and Laboratory Methods Applicable to Overburdens and Minesoils (EPA-600/2-78-054). Environmental Protection Agency, Washington, DCA. Vink, B., 1996. Stability relations of antimony and arsenic compounds in the light of revised and extended Eh-pH diagrams. Chemical Geology 130, 21–30. WHO, 1996. Guidelines for Drinking-water Quality, vol. 2: Health Criteria and Other Supporting Information, second ed. World Health Organisation, Geneva. Williams, G.J., 1974. Economic Geology of New Zealand. Australasian Institute of Mining and Metallurgy Monograph 4. Wilson, N.J. 2002. Elevated Arsenic and Antimony Levels in South Island Mesothermal Mineralized Zones. Unpublished MSc Thesis, University of Otago Library, Dunedin, New Zealand.