Environ. Sci. Technol. 2003, 37, 189-195
Stabilization of Available Arsenic in Highly Contaminated Mine Tailings Using Iron JU-YONG KIM† AND ALLEN P. DAVIS* Department of Civil and Environmental Engineering and Maryland Water Resources Research Center, University of Maryland, College Park, Maryland 20742 KYOUNG-WOONG KIM Department of Environmental Science and Engineering, Kwangju Institute of Science and Technology, Kwanju 500-712, South Korea
To evaluate the stabilization of available As in contaminated tailings from two abandoned metal mines of South Korea (the Myoungbong and Daduck Mines, 6670 and 56 600 mg/kg total As, respectively), characteristics of the tailings were investigated, and the tailings were treated through precipitation of amorphous iron compounds. Steep decreasing trends of extractable (5% NaOCl) As with increasing initial Fe(III) additions were observed in both treated tailings. In general, the treated tailings had the lowest extractable As concentration at pH 6. Available As, defined as the sum of As concentrations for the first four steps of a sequential extraction, was reduced from 2090 to 428 mg/kg (80% reduction) in the Myoungbong tailings and from 1320 to 395 mg/kg (70% reduction) in the Daduck tailings. As levels in the treated tailings decreased even more after a 1-month dormant period. Adsorption/ coprecipitation tests performed with mixed As(III) and Fe(III) solutions demonstrated dramatically increased As sequestration via interaction with amorphous iron compounds with increasing pH. The bulk of the As appeared to be affiliated with stable Fe precipitates.
Introduction Arsenic, which is a toxic and carcinogenic element, causes serious environmental problems throughout the world. For example, the high natural arsenic content of the drinking water has caused endemic, chronic arsenic poisoning in India, Bangladesh, Inner Mongolia, Argentina, Chile, and Taiwan (1-7). Elevated levels of arsenic can be present in the environment as a result of mineral weathering and dissolution, geothermal activity, and numerous anthropogenic sources including mining and smelting activities, pesticide use, and coal combustion (8). Common routes of exposure to arsenic include ingestion and inhalation of arsenic compounds. Arsenic toxicity, bioavailability, and mobility vary depending on its oxidation state. The most prevalent species of dissolved arsenic are arsenite and arsenate. Arsenate [As(V)] is the predominant As form in well-oxidized waters, while arsenite [As(III)] occurs predominantly in reduced * Corresponding author phone: (301)405-1958; fax: (301)405-2585; e-mail:
[email protected]. † Present address: Kwangju Institute of Science and Technology, Kwanju 500-712, South Korea. 10.1021/es020799+ CCC: $25.00 Published on Web 12/03/2002
2003 American Chemical Society
environments, although because of the relatively slow redox transformations (9), both arsenite and arsenate are often found in either redox environment. Arsenite is listed as 25-60 times more toxic than arsenate and has been reported to be more mobile in the environment (10). Arsenic is commonly found in sulfur minerals along with precious metals. Arsenic species released from sulfur-containing minerals during weathering processes undergo oxidation-reduction, precipitation-dissolution, adsorption-desorption, and organic and biochemical methylation reactions, all of which control the behavior of arsenic in the environment (11). Currently there exists about 1000 abandoned metal mines in South Korea; most of these mines have been left without any management. The tailings from these mines contain several types of toxic contaminants, including As and heavy metals, and the potential is high for deterioration of the ecosystems around these mines. Recently, arsenic and heavy metal contamination of agricultural soils and crops surrounding the mining areas has been identified as one of most serious environmental problems in South Korea (12-14). Inorganic solidification/stabilization processes have been studied to reduce the solubility of As in solid waste, using inorganic binders such as cement, lime, and pozzolanic materials (15-18). Recent investigations of As sorption on amorphous and crystalline iron hydroxides illustrate the potential for As attenuation via interaction with these mineral surfaces (19-23). Sequential extraction of contaminated soils by Dudas (24) indicated that arsenic was primarily associated with crystalline and amorphous oxides of iron. Other studies have implicated amorphous iron and aluminum as preferential adsorbents of arsenic (25) and iron oxides in general for the sorption of arsenite (26). Iron can also form insoluble iron-arsenic compounds such as FeAsO4. Accordingly, the objectives of this study were to examine the release of As from highly contaminated mine tailings using a sequential extraction procedure and to evaluate the stabilization of available As in these tailings using the addition of iron compounds.
Materials and Methods Sample Collection and Preparation. Tailings samples were collected from two abandoned metal mines in South Korea, Myoungbong (Au-Ag) and Daduck (Au-Ag-Pb-Zn) Mines. Genesis of these mines is Au-Ag bearing quartz vein type, and the main sources of As and heavy metals in these mines are arsenopyrite (FeAsS), sphalerite (ZnS), and galena (PbS) (14). Approximately 20 kg of composite tailings sample was collected at each mine in July 2000 using a hand auger. Each composite sample was composed of approximately 100 subsamples taken at surface and 15, 30, and 45 cm depths in the tailing piles. The tailings samples were transferred to plastic-lined canvas bags and transported to the laboratory. The tailings were air-dried, disaggregated, and sieved to a -10 mesh (less than 2 mm). Tailings Characterizations. The pH of the tailings was determined by adding 25 mL of DI water to 5 g of a tailings sample, stirring intermittently for 1 h, and then left standing for 0.5 h. The pH was measured with a pH meter that had been calibrated with standard buffers of 4.01 and 7.00. Organic matter content and sand, silt, and clay size distributions were determined by the University of Maryland Soil Testing Laboratory. Total concentrations of As, Cd, Cu, Pb, and Zn were determined by aqua regia digestion (27) and U.S. EPA method 3050. For aqua regia digestion, 1 mL of HNO3 and 3 mL of HCl were added to 0.25 g of the tailings. Samples were heated VOL. 37, NO. 1, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
189
TABLE 1. Sequential Extraction Scheme for Fractionation of Tailings-Fixed Arsenica
a
step
fraction
extraction procedure
1 2 3 4 5 6 7
soluble adsorbed carbonate soil organic matter easily reducible oxides amorphous oxides crystalline minerals
0.25 M KCl (2 h) 0.1 M Na2HPO4 (pH 8, 20 h) 1 M sodium acetate (pH 5, 1 + 4 h), then 0.1 M Na2HPO4 (pH 8, 20 h) 5% NaOCl (pH 9.5, 0.5 h boil, repeated once) 0.1 M NH2OH (pH 2, 0.5 h) followed by 0.1 M KOH (20 h) 0.25 M NH2OH/HCl (0.5 h at 50°C) followed by 0.1 M KOH (20 h) aqua regia (HCl + HNO3)
Modified from ref 29.
to 70 °C and shaken for 1 h. Afterward, 6 mL of DI water was added to the solution. Toxicity characteristic leaching procedure (TCLP, U.S. EPA Method 1311), synthetic precipitation leaching procedure (SPLP, U.S. EPA Method 1312), 1 N HCl extraction, and NaOCl extraction were conducted to estimate the quantity of available As in the tailings. The TCLP using fluid 1 (diluted glacial acetic acid, pH 4.93) and SPLP were carried out with the following exception: For both TCLP and SPLP, 10 g of tailings rather than 100 g and a corresponding reduced quantity of leaching solution were used. For 1 N HCl extraction, 50 mL of 1 N HCl was added to 10 g of tailings and shaken for 1 h at room temperature. For NaOCl extraction, 10 mL of 5% NaOCl solution adjusted to pH 9.5 was added to 2 g of tailings and boiled for 0.5 h. This procedure was repeated once. All U.S. EPA methods are described in ref 28. Sequential Extraction Analysis. A modified version of the sequential extraction analysis set by Tokunaga et al. (29) was employed to estimate the bonding strength and chemical speciation of As in the tailings. Tokunaga’s method originally consisted of eight fractions: (1) soluble, (2) adsorbed, (3) carbonate, (4) soil organic matter, (5) easily reducible oxides, (6) amorphous oxides, (7) crystalline oxides, and (8) amorphous aluminosilicates. The modified method consisted of only 7 steps, combining steps 7 and 8 to produce an operationally defined crystalline mineral fraction determined by aqua regia digestion. Details of the sequential extraction analysis are presented in Table 1. Extraction results were focused on the weakly bonded As fraction that is assumed to be related to available As. Thus, this analysis was also used to evaluate the efficiency of stabilization for available As after treatment experiments. Treatment of Tailings with Amorphous Iron Precipitates. Amorphous iron precipitates were created to stabilize the available As in the tailings. Twenty milliliters of ferrous or ferric sulfate solution at several different concentrations (20-100 mM) was added to 2 g of tailings; pH was adjusted to the range of 3-6 by NaOH or Ca(OH)2. Following this, the mixture was shaken at 30 rpm for 2 h and centrifuged. Concentrations of As and Pb in supernatant filtered through 0.2 µm membrane filters were measured. NaOCl extraction and sequential extraction were conducted on the treated tailings to examine changes in As availability. Adsorption/Coprecipitation Tests. To examine the fate of As onto precipitated amorphous iron, adsorption/coprecipitation tests were performed with synthetic As and Fe solutions. In these tests, four mixtures [100 mM Fe(III) + 20 mM As, 100 mM Fe(III) + 10 mM As, 20 mM Fe(III) + 20 mM As, and 20 mM Fe(III) + 10 mM As] were made by mixing of 20 mL each of Fe(III) [as Fe2(SO4)3] and As (from As2O3) solutions. The pH of each mixture was adjusted using Ca(OH)2; the mixtures were shaken at 30 rpm for 2 h and centrifuged. Concentrations of As(III) and As(V) in filtered supernatants were measured, and sequential extractions were conducted on the solid phases. Analytical Methods. All reagents used were ACS grade, and all extractions for untreated tailings were performed in 190
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 1, 2003
TABLE 2. pH, Organic Matter Content (OMC), and Particle Size Fractions of Tailings
tailings
pH
OMC (%)
Myoungbong Daduck
3.7 2.5
3.92 0.60
sand fraction (%)
silt fraction (%)
clay fraction (%)
44 91
46 2
10 7
triplicate. Aqueous As concentrations were measured using hydride generation atomic absorption spectrophotometry on a Perkin-Elmer 5100 spectrophotometer. For As redox speciation analysis, the filtered supernatants were allowed to pass through a silica-based anion-exchange cartridge (LC-SAX SPE Tube, Supelco) that retained As(V) (30); As(III) was collected in the effluent solution. Arsenic(V) was extracted from the cartridge using 1.0 M HCl. Cd, Cu, Pb, and Zn concentrations were measured using flame atomic absorption spectrophotometry. All analytical data were assessed for accuracy and precision using a quality control system including duplicates, blanks, and standard samples during the analytical procedure (31). All duplicate analytical measurements demonstrated precision within 5%. Only single extractions were performed for treated tailings because of the number of treatments investigated. Reagent blanks for all extractants were analyzed in parallel with samples and found to have negligible levels of As in all cases.
Results and Discussion Characteristics of Tailings. The pH, organic matter content, and particle size fractions of the tailings are presented in Table 2. The pH values of the tailings from the Myoungbong and Daduck Mines were 3.7 and 2.5, respectively. Under these acidic conditions, As is considered as “moderately mobile” compared with cations such as Cd, Ni, and Zn, which are mobile and easily leachable under acidic conditions (32). The Myoungbong tailings have finer size fractions and greater organic matter content than the Daduck tailings. Arsenic exhibits rather high affinity for soil organic matter (32). Table 3 presents total concentrations of As and some heavy metals in the Myoungbong and Daduck tailings as measured by aqua regia extraction and U.S. EPA Method 3050. In general, the aqua regia digestion extracts from 70 to 90% of the total As in soils (27). The differences between total concentration for aqua regia extraction and U.S. EPA Method 3050 were very low, for example, less than 2.3% for As. Comparing with concentrations in “typical” soils (6 mg/kg As) (33) and tolerable levels of metals (20 mg/kg As, 3 mg/kg Cd, 100 mg/kg Cu, 100 mg/kg Pb, and 300 mg/kg Zn), which are considered as phytotoxically excessive (34), total As concentrations extracted by U.S. EPA Method 3050 were extremely high: 6670 mg/kg in the Myoungbong tailings and 56 600 mg/kg in the Daduck tailings. These exceedingly high concentrations of As apparently result from arsenopyrite, which is the most abundant As-bearing mineral in the both mines (14). The concentrations of all analyzed metals in the
TABLE 3. Total Concentration of As and Heavy Metals (mg/kg) Extracted by Aqua Regia and U.S. EPA Method 3050a mine
digestion
As
Cd
Cu
Pb
Zn
Myoungbong
aqua regia EPA 3050 aqua regia EPA 3050
6 830 ( 510 6 670 ( 100 57 000 ( 900 56 600 ( 800
2.71 ( 0.15 3.71 ( 0.11 8.85 ( 0.31 9.05 ( 0.11
19 ( 7 14 ( 4 267 ( 2 278 ( 6
945 ( 29 913 ( 6 8 720 ( 740 7 880 ( 110
224 ( 37 229 ( 31 847 ( 51 784 ( 28
Daduck a
Average ( standard deviation for triplicate analyses.
TABLE 4. Sequential Extraction Concentrations for As and Fe in the Tailingsa As (mg/kg) extraction step (fraction)
a
Fe (mg/kg)
Myoungbong
Daduck
Myoungbong
Daduck
1 (soluble) 2 (adsorbed) 3 (carbonate) 4 (soil organic matter)
0.7 ( 0.0 407 ( 6 167 ( 2 1 510 ( 30
0.9 ( 0.1 188 ( 7 115 ( 2 1 020 ( 10
7(1 21 ( 1 18 ( 1 20 ( 1
60 ( 2 71 ( 3 38 ( 0 14 ( 1
subtotal
2 090 ( 40
1 320 ( 20
5 (easily reducible oxides) 6 (amorphous oxides) 7 (crystalline minerals)
3 890 ( 110 212 ( 3 398 ( 14
26 200 ( 300 22 700 ( 100 7 330 ( 150
603 ( 6 4 040 ( 270 12 900 ( 800
12 300 ( 100 10 300 ( 600 38 100 ( 1 500
total
6 590 ( 170
57 600 ( 600
17 600 ( 1 100
60 900 ( 3 200
66 ( 4
183 ( 6
Average ( standard deviation for triplicate analyses.
Daduck tailings exceed their respective tolerable levels. In the case of Pb, the measured total concentration was up to 80 times higher. On the other hand, only As and Pb exceeded tolerable levels in the Myoungbong tailings. It is important to note that although sequential extractions attempt to isolate and dissolve one particular geochemical fraction, the efficiency of chemical extractions depends on the affinity and specificity of the extracting chemical for the target phase. Sequential extractions, nonetheless, can provide valuable information regarding general partitioning patterns and estimates of reactive phases within geochemical materials (35). The cumulative As from sequential extractions found from the Myoungbong and Daduck tailings were 6590 ( 170 and 57 600 ( 600 mg/kg, respectively (Table 4). These values represent 99% of that extracted by U.S. EPA Method 3050 for the Myoungbong tailings and 102% for the Daduck tailings. Somewhat different partitioning patterns between the Myoungbong and the Daduck tailings are indicated by the results of sequential extractions for As. The most abundant fraction of As in the Myoungbong tailings was the step 5 (easily reducible oxides) fraction, which was 3890 mg/kg and made up approximately 60% of the total As content. The next most abundant fraction was step 4 (soil organic fraction) at 1510 mg/kg (23% of total As). On the other hand, step 5 (26 200 mg/kg As) and step 6 (22 700 mg/kg As) were the predominant fractions in the Daduck tailings, followed by step 7 (crystalline minerals fraction; 7330 mg/kg As). These three fractions made up 92% of the total As content in the Daduck tailings. In both tailings, the majority of Fe was liberated in the last three fractions, which released up to 99% of the total Fe. Available As, which is considered as weakly bonded and easily releasable, is defined here as resulting from the first four steps of the sequential analysis. The sum of As concentrations from the first four steps was 2090 mg/kg in the Myoungbong tailings and 1320 mg/kg in the Daduck tailings. Considering the total As concentrations, the available As fraction was much higher in the Myoungbong tailings than in the Daduck. Arsenic is expected to exist in various forms in the tailings, originating from the weathering of Asbearing minerals such as arsenopyrite. It can be concluded from the sequential extraction results that the degree of weathering is higher in the Myoungbong tailings than in the
FIGURE 1. Comparison of available As extracted by partial extraction methods and cumulative concentration of As by sequential extraction in mine tailings. Daduck tailings since weathering of sulfide minerals is expected to create more soluble products. Several extraction procedures have been developed to determine available As in geochemical media (36-38). However, correlations between the amount of As extracted by these methods and mobility or phytotoxicity of As are weak. TCLP, SPLP, 1 N HCl extraction, and 5% NaOCl extraction are examined to devise a simple method that can provide information on available As and assess the efficiency of the stabilization treatment procedures. The TCLP is employed for determination of whether soils should be VOL. 37, NO. 1, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
191
TABLE 5. Concentrations of As, Fe, and Pb in Reagents and Supernatant Solutions Collected after 2 h Shaking at pH 6, Adjusted by Ca(OH)2 As (mg/L)
Fe (mg/L)
Pb (mg/L)
reagent
MBa
DDb
Oc
MB
DD
MB
DD
ferrous sulfate (20 mM) ferrous sulfate (60 mM) ferric sulfate (20 mM) ferric sulfate (60 mM)
