Assessing the Influence of Reacting Pyrite and ... - ACS Publications

Geology Department, Eastern Washington University,. Mail Stop 70, Cheney, Washington 99004. The relative abundance of minerals that react to generate...
0 downloads 0 Views 168KB Size
Download PDF
Environ. Sci. Technol. 1999, 33, 3347-3353

Assessing the Influence of Reacting Pyrite and Carbonate Minerals on the Geochemistry of Drainage in the Coeur d’Alene Mining District LAURIE S. BALISTRIERI* U.S. Geological Survey, School of Oceanography, University of Washington, Box 357940, Seattle, Washington 98195 STEPHEN E. BOX AND ARTHUR A. BOOKSTROM U.S. Geological Survey, West 904 Riverside Avenue, Room 202, Spokane, Washington 99201 MOHAMMED IKRAMUDDIN Geology Department, Eastern Washington University, Mail Stop 70, Cheney, Washington 99004

The relative abundance of minerals that react to generate or consume acid in mineralized areas is critical in determining the quality of water draining from such areas. This work examines the fundamental reactions that influence the pH and composition of drainage from mine adits and tailings piles. We construct triangle diagrams that predict stoichiometric relationships between concentrations of dissolved SO4, dissolved Ca and Mg, and either alkalinity or acidity by considering reactions involving the oxidation of pyrite, dissolution of carbonate minerals, and precipitation of iron oxide and iron hydroxysulfate minerals. Drainage data from the Coeur d’Alene mining district are used to test our stoichiometric approach. Comparisons between theoretical predictions and drainage data indicate that the range of pH values in the mining district is due to reacting pyrite to carbonate mineral ratios that range from near 0/1 to 1/1. Calcite and ankerite are the dominant carbonate minerals that buffer the acid produced during pyrite oxidation and ferrihydrite or schwertmannite precipitation.

Introduction Drainage from undisturbed mineralized areas, mine workings, and tailings piles can range from acidic to basic and have highly variable concentrations and compositions of metals (1-4). Understanding the key factors that influence the geochemistry of waters in these areas is critical for minimizing impacts on water quality during the development of new mines and in the remediation of past mining activities. Studies of diverse ore deposit types in the Colorado Mineral Belt indicate that the primary factors affecting the pH and composition of such drainage are the geology and geochemistry of the ore deposit and its surrounding host rocks, physical exposure of the deposit to weathering processes, and geochemical processes such as oxidation, dissolution, precipitation, and adsorption (2). * Corresponding author phone: (206)543-8966; fax: (206)543-6073; e-mail: [email protected]. 10.1021/es980823c CCC: $18.00 Published on Web 08/17/1999

 1999 American Chemical Society

Our study focuses on drainage from adits and tailings piles in the Coeur d’Alene (CdA) mining district. Water quality in certain areas of the CdA River basin appears to be affected by drainage from tailings deposits and mine adits (5, 6). Current remediation activities within the basin require information on point sources of metals to surface and groundwater and the processes controlling metal distributions, transport, and fate. Our discussion of drainage in the CdA mining district examines the geochemical characteristics of the waters and key parameters and processes that influence their composition. First, we compare pH and metal characteristics of drainage in the CdA basin with drainage from wellcharacterized deposits in the Colorado Mineral Belt. Second, we compare relationships between the sum of dissolved Ca and Mg concentrations, SO4 concentrations, and alkalinity or acidity in drainage to stoichiometric relationships between these variables derived from reactions describing the oxidation of pyrite, dissolution of buffering carbonate minerals, and precipitation of secondary minerals. Our goal is to provide a theoretical foundation to empirical determinations of the net acid production potential of mineralized rock and mining wastes.

Study Area The CdA mining district is located in northern Idaho and western Montana in the northwestern United States (see Figure 1). Total production records indicate that this district ranks as one of the world’s largest producers of Ag and one of the United States’ major producers of Pb and Zn. Ore deposits are steeply dipping Pb-Zn-Ag veins occurring in Precambrian rocks of the Belt Supergroup (7-13). These host rocks are primarily quartzite and argillite with lesser amounts of interbedded carbonate and carbonate bearing rocks. The principal economic minerals are galena (PbS), sphalerite (ZnS), and argentiferous tetrahedrite [(Cu,Fe)12Sb4S13]. Most veins contain small amounts of chalcopyrite (CuFeS2). Pyrite (FeS2) is ubiquitous, but variable in abundance, in the veins. The predominant gangue minerals are siderite (FeCO3) and quartz (SiO2). The absolute and relative abundance of sulfide and gangue minerals varies significantly between different vein systems. Mineralogical work done at several mines in the district indicate that wall rocks around veins are concentrically zoned with respect to three carbonate mineralsssiderite (FeCO3), ankerite [CaFe(CO3)2], and calcite (CaCO3) (14). The carbonate mineral closest to the veins is siderite. The outer edge of the siderite zone grades to a much wider zone of ankerite either in an assemblage of ankerite and siderite or an assemblage of ankerite and calcite. Wall rocks typically contain 10-15% carbonate but can have as much as 62% carbonate. Initial ore separation methods were not very efficient. Jig tailings produced before 1915 were highly enriched in metals, especially Zn. Development of more efficient flotation methods between 1915 and 1925 resulted in tailings with lower metal concentrations. One of the tailings piles we sampled consists of jig tailings only (Success), two have jig tailings covered by flotation tailings (Bunker Hill and Interstate-Callahan), and one consists solely of flotation tailings (Rex). Most tailings were deposited directly into the CdA River and its tributaries before environmental regulations required the installation of tailings ponds in 1968. Flood events subsequently have redistributed metal enriched tailings throughout the channel of the South Fork and main stem of the CdA River and its flood plain. Our sample from VOL. 33, NO. 19, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3347

FIGURE 1. Maps indicating the location of the Coeur d’Alene (CdA) mining district and our sample sites. mine tailings near Osburn seeps from a flood plain covered by fluvially deposited jig tailings.

Methods Water draining from adits and beneath tailings piles in the mining district was collected during August 1996, November 1996, and June 1997. The locations of these sites are illustrated in Figure 1. Drainage from adits generally was sampled within 10 m of the mine portal. Drainage from the base of tailings piles was sampled within a meter of the seep source. Strict protocols were followed during sampling and handling of waters to ensure quality and comparison of data collected at different times. Detailed procedures and the complete analytical data set are summarized in Balistrieri et al. (15). Field measurements of water temperature, pH, conductivity, and dissolved O2 concentrations were made upon arrival at each site using Orion conductivity, temperature and pH meters and electrodes, and CHEMETS colorimetric test kits for O2. A 1 L water sample was then collected after rinsing the bottle three times with sample. Aliquots for dissolved metals, Fe(II), major anions (Cl and SO4), and alkalinity were taken from the bottle using a syringe and filtered through 0.45 µm nylon disposable filters into acid cleaned high-density polyethylene (HDPE) bottles for metals and Fe(II) or unwashed HDPE bottles for anions and alkalinity. Samples to be analyzed for metals were preserved by adding 1 drop of redistilled concentrated nitric acid per 10 mL of sample. Samples to be analyzed for Fe(II) were preserved by adding 5 drops of redistilled concentrated hydrochloric acid per 10 mL of sample and covering the bottle with foil. Samples to be analyzed for anions were kept cold in an ice chest during transport to the laboratory. Alkalinity was determined in the field by Gran titration (16). Major and trace ions were determined after returning to the laboratory. Fe(II) was analyzed within 1 week using Ferrozine. K and Na concentrations were determined by flame atomic absorption spectroscopy. Ca and Mg concentrations were determined by inductively coupled plasma-atomic emission spectroscopy. The remaining metal and metalloid concentrations were determined by inductively coupled plasma-mass spectroscopy. Ion chromatography was used to determine concentrations of major anions. Reference standards, duplicates, and acid and filtering blanks were analyzed with each batch of samples. Agreement between concentrations in reference standards and accepted values 3348

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 19, 1999

generally was excellent. Triplicate samples from one adit indicated that variations in metal concentrations were less than 5%. Acid and filtering blanks contained negligible concentrations of all elements.

Results and Discussion The characteristics of pH and dissolved metal concentrations for drainage from adits and tailings piles in the CdA mining district are illustrated on a plot developed by Ficklin et al. (1) (see Figure 2). This plot shows the sum of dissolved base metal (Cd, Co, Cu, Ni, Pb, and Zn) concentrations as a function of pH. The lowest pH of all adit waters (pH ) 2.72) was measured at the Kellogg Tunnel. The remaining adit waters throughout the district fall on a relatively straight line within the neutral to basic category. They have a pH range of 6.64-7.75 and a median pH of 7.32. Except for the upper seep at the Success tailings pile in June 1997 (pH 7.13), all waters seeping from tailings piles are acidic with a pH range of 4.6-6.62. The striking point about the dissolved base metal concentrations of drainage in this district is the wide range. Near neutral adit drainage has dissolved metal concentrations that vary by 3-4 orders of magnitude, while metal concentrations from all measured drainage vary by 4-5 orders of magnitude. The sum of dissolved base metal concentrations is dominated (>90%) by Zn concentrations in drainage from all tailings piles and most adits. However, adit drainage from deposits rich in chalcopyrite has different metal signatures. Dissolved base metal concentrations in water from the Silver Dollar adit contain 79-82% Zn, 9% Cu, and 8% Ni, while drainage in the Hooper Tunnel and Snowstorm no. 3 adit is dominated by Cu (67-97% Cu, 2-18% Zn, 3-9% Ni, and 3-5% Co). Studies in the Colorado Mineral Belt (CMB) examined the relationship between mine drainage composition and mineralogy of ore deposits and host rocks (2). One of the key parameters that influenced composition was the relative abundance of pyrite and carbonate minerals. Their adit data and relative abundance of pyrite and carbonate minerals in deposits and host rocks are compared with drainage data from the CdA mining district in Figure 2. This comparison indicates that the wide ranging pH and metal concentrations found in the CdA mining district likely are due to variable amounts of pyrite and carbonate minerals in the individual deposits and host rocks. The dominance of acid generating minerals (e.g., pyrite) results in acid pH values and the

FeS2 + 14Fe3+ + 8H2O f 15Fe2+ + 2SO42- + 16H+ (1) The Fe(II) that is produced in reaction 1 can oxidize in the presence of O2 to Fe(III):

Fe2+ + 0.25O2 + H+ f Fe3+ + 0.5H2O

FIGURE 2. Relationship between pH and the sum of dissolved base metal (Cd, Co, Cu, Ni, Pb, and Zn) concentrations in waters draining adits and tailings piles in the Coeur d’Alene (CdA) mining district and the Colorado Mineral Belt (CMB). The CMB data are from Plumlee et al. (2). mobilization of metals into solution, whereas significant amounts of acid neutralizing minerals (e.g., calcite) increase the pH and possibly result in the precipitation of oxides and hydroxysulfates that scavenge metals from solution. Thus, low pH values and high dissolved metal concentrations are found in drainage from deposits with high pyrite-to-carbonate ratios. Deposits that are high in pyrite and high in carbonate minerals produce drainage that has high dissolved metal concentrations and near neutral pH values. The lowest dissolved metal concentrations are found in near neutral drainage from deposits that contain low pyrite-to-carbonate ratios. This diagram also indicates that mine drainage is not necessarily acidic and that near neutral mine drainage can have high metal concentrations (2). Comparisons in Figure 2 are based on empirical observations of drainage composition and the mineralogy of the deposits and host rocks. In reality, the composition of drainage is a function of the amount of reacting pyrite and carbonate minerals, not the absolute abundance of the minerals. We have developed a more quantitative approach for assessing the relative abundance of reacting pyrite and carbonate minerals on drainage composition. This new approach compares relationships between selected dissolved ion concentrations and alkalinity or acidity in drainage with stoichiometric relationships between these variables predicted from chemical reactions. The predictions are derived from reactions describing the oxidation of pyrite, precipitation of hydrous Fe oxide or Fe hydroxysulfate minerals, and dissolution of carbonate minerals and from the definitions of alkalinity and acidity. It is well-known that a major source of acid to waters in mineralized areas is the oxidation of pyrite. Although the overall oxidant that drives pyrite oxidation is O2 from the atmosphere, dissolved Fe(III) appears to be the primary oxidant that attacks the pyrite surface to form Fe(II), SO4, and protons (17, 18):

(2)

This reaction is the rate determining step and is usually catalyzed by autotrophic bacteria (19). The Fe(III) produced in reaction 2 can either further oxidize pyrite (reaction 1) or hydrolyze and then precipitate as hydrous Fe oxide (goethite [RFeOOH] or ferrihydrite [∼Fe5OH8‚4H2O]) or as Fe hydroxysulfate minerals (jarosite [KFe3(SO4)2(OH)6] or schwertmannite [Fe8O8(OH)6SO4]) depending on kinetic factors, pH, and concentrations of Fe(III), SO4, and bicarbonate (20, 21). Bigham et al. (21) indicates that although goethite and jarosite ultimately control the concentration of Fe(III) in mine drainage, the metastable phases of ferrihydrite or schwertmannite are frequently found in such systems. Both mineralogical and thermodynamic data suggest that jarosite controls the activity of dissolved Fe(III) in drainage with pH values less than about 2.7, schwertmannite for pH values from about 2.8 to 4.5, mixtures of schwertmannite and ferrihydrite for pH values between 4.5 and 6.5, and ferrihydrite for higher pH values. Given the range of pH values of drainage in the CdA mining district, jarosite or schwertmannite may control the concentrations of dissolved Fe(III) in drainage from the Kellogg Tunnel and either mixtures of schwertmannite and ferrihydrite or predominantly ferrihydrite control dissolved Fe(III) concentrations in drainage from tailings piles and the remaining adits. The precipitation of Fe(III) as jarosite or as schwertmannite consumes SO4 and produces protons

K+ + 3Fe3+ + 2SO42- + 6H2O T

KFe3(SO4)2(OH)6 + 6H+ (3)

8Fe3+ + SO42- + 14H2O T Fe8O8(OH)6 SO4 + 22H+

(4)

whereas the precipitation of ferrihydrite [hereafter denoted by the formula Fe(OH)3] only produces protons:

Fe3+ + 3H2O T Fe(OH)3 + 3H+

(5)

The principal reactions that generate base, in sequential order based on reactivity, are the dissolution of carbonate minerals (calcite [CaCO3], dolomite [CaMg(CO3)2], or ankerite [CaFe(CO3)2]), oxides and hydroxides of Al and Fe, and aluminosilicates (feldspars, chlorite, and muscovite) (22, 23). The most important acid neutralizing reactions are the dissolution of carbonate minerals. The carbonate minerals of importance in the CdA mining district are calcite, ankerite, and siderite. Several analyses of the composition of these minerals for the district have been done (7, 14) and are summarized in Table 1. It is clear that the phases are not pure and analyses are limited to only a few locations in the district. The calculations that follow assume that the average compositions of carbonate minerals at these locations are representative of the entire district. The dissolution of these minerals results in the release of divalent metal (Ca2+, Fe2+, Mg2+, and Mn2+) and bicarbonate (HCO3-) ions to solution. The dissolution reactions for the CdA carbonates are written for calcite

(Ca0.94Fe0.041Mg0.003Mn0.018)CO3 + H+ f 0.94Ca2+ + 0.041Fe2+ + 0.003Mg2+ + 0.018Mn2+ + HCO3- (6) VOL. 33, NO. 19, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3349

TABLE 1. Composition of Carbonate Minerals from the Coeur d’Alene Mining District mineral calcite

mine av std dev

ankerite

siderite

Ca moles

Fe moles

Mg moles

Mn moles

CO3 moles

ref

0.94 0.01 1.07 1.04 1.07 1.05 1.02 1.05 0.02 0.024 0.006 0.02-0.06 0.023 0.014

0.041 0.008 0.70 0.68 0.49 0.63 0.56 0.61 0.08 0.83 0.87 0.75-0.85 0.84 0.03

0.003 0.002 0.13 0.17 0.38 0.22 0.32 0.24 0.09 0.07 0.04 0.06-0.10 0.06 0.02

0.018 0.002 0.097 0.100 0.068 0.098 0.098 0.093 0.013 0.077 0.078 0.08-0.10 0.082 0.006

1

14

2 2 2 2 2 2

14 7 7 7 7

1 1 1 1

14 7 7

Lucky Friday (n ) 8) Lucky Friday Sunshine Coeur d’Alene (n ) 2) Merger Bunker Hill Lucky Friday (n ) 8) Sunshine (n ) 7) Bunker Hill (only ranges given)

av std dev

av std dev

for ankerite

Ca1.05(Fe0.61Mg0.24Mn0.093)(CO3)2 + 2H+ f 1.05Ca2+ + 0.61Fe2+ + 0.24Mg2+ + 0.093Mn2+ + 2HCO3- (7) and for siderite

(Ca0.023Fe0.84Mg0.06Mn0.082)CO3 + H+ f 0.023Ca2+ + 0.84Fe2+ + 0.06Mg2+ + 0.082Mn2+ + HCO3- (8) The Fe(II) that is released during dissolution of the carbonate minerals can oxidize and precipitate either as iron oxide or iron hydroxysulfate minerals. Alkalinity [Alk] is the acid neutralizing capacity of water (16). For most waters alkalinity can be defined in terms of the carbonate ions (HCO3- and CO32-) as the concentrations of other proton acceptors (e.g., NH3, HS-, or B(OH)4-) are negligible:

[Alk] ) [HCO3-] + 2[CO32-] + [OH-] - [H+]

(9)

All adit waters, except the Kellogg Tunnel, and all tailings seeps collected in the CdA mining district contain alkalinity with values ranging from 0.09 to 3.68 mM. Acidity [Acy] is the base neutralizing capacity of water (16). Acidity is measured by titrating water with a strong base to a designated pH, typically pH 8.2 or 8.3 (24, 25). For mine drainage waters, reactions that contribute to acidity during a base titration to pH 8.3 include the neutralization of strong and weak acids (H+, HSO4-, H2CO30), the hydrolysis of metal ions (Fe3+, Al3+), and the oxidation and hydrolysis of redox sensitive metals (Fe2+, Mn2+) (26). Thus, acidity is defined as

[Acy] ) [H+] + [HSO4-] + [H2CO30] + 3[Al3+] + 3[Fe3+] + 2[Fe2+] + 2[Mn2+] (10) Because acidity was not measured by titration for the drainage waters from the CdA mining district, we calculated [Acy] from eq 10. We used measured dissolved concentrations of Al, Fe(III), Fe(II), and Mn(II) to include the free metal ion as well as its solution complexes. Concentrations of H+, HSO4-, and H2CO30 were determined from solution speciation calculations done for the drainage waters using MINTEQA2 (27). The calculated [Acy] range from 0.02 to 41 mM. These calculations assume that titrations of drainage samples would result in complete oxidation and precipitation of Fe(II) and Mn(II). Furthermore, the calculations assume that filtration 3350

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 19, 1999

results in the accurate partitioning of Fe(III) between dissolved and particulate or colloidal forms. If particulate or colloidal Fe(III) passes through the filter and is considered to be dissolved, then the calculated acidity would be too high. These assumptions are only critical for drainage from the Kellogg Tunnel where 64% of the calculated acidity is due to Fe species. Because of their higher pH values, carbonic acid generally is the primary contributor to the calculated acidities of the remaining drainage samples. The above reactions and definitions can be used to predict the relative amounts of pyrite and carbonate minerals that react to form drainage. Concentrations of dissolved SO4 are the net result of oxidation of all sulfide minerals and the precipitation of secondary sulfate minerals such as Fe hydroxysulfates. Dissolved Ca and Mg concentrations represent the net concentration of Ca and Mg produced during the dissolution of carbonate minerals and Ca and Mg consumed during the precipitation of secondary minerals. Thermodynamic speciation calculations using MINTEQA2 indicate that all drainage waters in the CdA mining district are undersaturated with respect to secondary Ca and Mg minerals. Thus, it is assumed that Ca and Mg concentrations are not affected by the precipitation of secondary minerals. Alkalinity is the net result of protons consumed and bicarbonate produced during the dissolution of carbonate minerals and protons generated during pyrite oxidation and precipitation of iron oxide or iron hydroxysulfate minerals. Acidity accounts for the strong acids and metals generated during pyrite oxidation, for the consumption of protons and production of metals during carbonate mineral dissolution, and for protons generated during the precipitation of iron oxides or iron hydroxysulfate minerals. The stoichiometric relationships between the sum of dissolved Ca and Mg, dissolved SO4, and alkalinity or acidity are derived from the reactions and definitions given above. We consider the oxidation of pyrite, dissolution of average CdA calcite, ankerite, or siderite, and either no oxidation of Fe(II) produced from the previous reactions or oxidation of Fe(II) with subsequent precipitation of ferrihydrite, schwertmannite, or jarosite. The stoichiometry for the various cases is determined by combining the appropriate reactions given above. For example, the reaction describing the oxidation of pyrite and precipitation of ferrihydrite is determined by combining reaction 1 + 15 × reaction 2 + reaction 5:

FeS2 + 3.75O2 + 3.5H2O f Fe(OH)3 + 2SO42- + 4H+ (11) The dissolution of average CdA calcite and oxidation and precipitation of released Fe(II) as ferrihydrite is determined

FIGURE 3. Predicted stoichiometric relationships between normalized moles of dissolved SO4, the sum of dissolved Ca and Mg, and alkalinity or acidity for theoretical ratios of reacting pyrite to average Coeur d’Alene (CdA) carbonates [calcite (solid lines), ankerite (dashed lines), or siderite (long dashed, three short dashed lines)] considering either no Fe(II) oxidation (a and c) or precipitation of ferrihydrite or schwertmannite for drainage with alkalinity and pH > 4.6 (b) or precipitation of ferrihydrite, schwertmannite, or jarosite for drainage with acidity and pH 2.72-7.75 (d). The theoretical ratios of reacting pyrite to carbonate minerals are the symbols and fractions on the diagrams. For clarity in diagrams b and d, the regular font is for the cases where ferrihydrite precipitates (solid triangles), the underlined ratios are for the cases where schwertmannite precipitates (open circles), and the bold and italicized ratios are for the cases where jarosite precipitates (solid squares). See text for derivations. by combining reaction 6 + 0.041 × reaction 2 + 0.041 × reaction 5:

(Ca0.94Fe0.041Mg0.003Mn0.018)CO3 + 0.918H+ + 0.1H2O + 0.01O2 f 0.041Fe(OH)3 + 0.94Ca2+ + 0.003Mg2+ +

0.018Mn2+ + HCO3- (12)

These reactions are used to predict relationships between dissolved Ca and Mg, dissolved SO4, and alkalinity (or acidity) for variable ratios of reacting pyrite to carbonate minerals for the case where Fe(II) oxidizes and precipitates as ferrihydrite. Assuming comparable oxidation and dissolution rates, a reacting pyrite to carbonate mineral ratio of 1/4 would yield 2 mol of dissolved SO4 and 4 mol of protons for every mol of oxidized pyrite, while 4 mol of average CdA calcite would consume 3.672 mol of protons and produce 3.76 mol of dissolved Ca, 0.012 mol of dissolved Mg ions, 0.072 mol of dissolved Mn ions, and 4 mol of dissolved bicarbonate ions. Thus, 3.672 mol of alkalinity (i.e., 4 mol of bicarbonate - 0.328 mol of protons; eq 9) and 0.616 mol of acidity (i.e., 0.328 mol of protons + 0.288 mol of Mn; eq 10) are generated for this system. The stoichiometries for other cases are summarized in Table 2. The moles of dissolved Ca and Mg, dissolved SO4, and alkalinity (or acidity) resulting from the calculations for variable ratios of reacting pyrite to average CdA carbonate minerals are normalized to 100% and plotted on triangle diagrams (Figure 3). These diagrams indicate how variations in the reacting solid phase ratios and either no precipitation

or precipitation of various Fe minerals influence solution compositions. Drainage that leaches rocks with relatively very low reacting pyrite to calcite or ankerite ratios is predicted to have relatively lower dissolved SO4 concentrations, higher alkalinity, and lower acidity. This type of drainage plots near the top of the alkalinity diagrams (see Figure 3a,b) and near the bottom of the acidity diagrams (see Figure 3c,d). Drainage that leaches rocks with relatively low reacting pyrite-tosiderite ratios also plots near the top of the alkalinity diagrams (see Figure 3a,b) and near the top of the acidity diagrams (see Figure 3c,d). This water has relatively more acidity compared to dissolved Ca and Mg or SO4. Waters draining rocks with higher reacting pyrite to calcite or ankerite ratios are predicted to be relatively higher in dissolved SO4, lower in alkalinity, and higher in acidity. These waters plot near the bottom of the alkalinity diagrams (see Figure 3a,b) and closer to the top of the acidity diagrams (see Figures 3c,d). Drainage from rocks with higher reacting pyrite-to-siderite ratios also have relatively higher dissolved SO4 and lower alkalinity and plot near the bottom of the alkalinity diagrams (see Figure 3a,b). This drainage plots near the top of the acidity diagram and in a distinctly different location from the other carbonates (see Figure 3c,d). The locations of the specific carbonate minerals are uniquely defined in the diagrams by the relative amount of Ca and Mg that is released during their dissolution. In addition, there are clear differences in the relationships if Fe(II) does not oxidize (see Figure 3a,c) or if it oxidizes and precipitates as iron oxide or iron hydroxysulfate minerals (see Figure 3b,d). Jarosite is not considered in the alkalinity diagram (see Figure 3b), because VOL. 33, NO. 19, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3351

TABLE 2. Stoichiometric Coefficients for Reactions Describing Pyrite Oxidation and Dissolution of Average Coeur d’Alene (CdA) Carbonate Minerals Considering No Precipitation (ppt) or Precipitation of Ferrihydrite (F), Schwertmannite (S), or Jarosite (J)a no ppt

ppt F

ppt S

ppt J

SO4 Fe(II) Fe(III) H+

2 15 14 16

Oxidation of 1 mol of Pyrite 2 1.875 0 0 0 0 4 3.75

1.333 0 0 3

Ca Mg SO4 Fe(II) Mn H+ HCO3-

Dissolution of 1 mol of Average CdA Calcite 0.94 0.94 0.94 0.003 0.003 0.003 0 0 -0.005 0.041 0 0 0.018 0.018 0.018 -1 -0.918 -0.928 1 1 1

0.94 0.003 -0.027 0 0.018 -0.959 1

Ca Mg SO4 Fe(II) Mn H+ HCO3-

Dissolution of 1 mol of Average CdA Ankerite 1.05 1.05 1.05 0.24 0.24 0.24 0 0 -0.077 0.61 0 0 0.093 0.093 0.093 -2 -0.77 -0.93 2 2 2

1.05 0.24 -0.406 0 0.093 -1.39 2

Ca Mg SO4 Fe(II) Mn H+ HCO3-

Dissolution of 1 mol of Average CdA Siderite 0.023 0.023 0.023 0.06 0.06 0.06 0 0 -0.104 0.84 0 0 0.082 0.082 0.082 -1 0.67 0.46 1 1 1

0.023 0.06 -0.556 0 0.082 -0.165 1

a See text for example calculation. Positive numbers indicate production of species and negative numbers indicate uptake of species.

it is not expected to precipitate in the pH range depicted in the diagram (pH > 4.6). These diagrams do not distinguish between the precipitation of ferrihydrite and schwertmannite because differences in the moles of protons and SO4 produced per mole of pyrite oxidized or carbonate mineral dissolved are too small to differentiate between these phases (see Table 2). Successful predictions of the exact ratios of reacting pyrite to carbonate minerals from drainage data require an understanding of the particular Fe phase that is precipitating. This point is demonstrated by the different locations of the same pyrite/ carbonate ratios for the precipitation of ferrihydrite and schwertmannite phases (see Figure 3b,d). However, an estimate of the ratios can be made because the locations for the reacting mineral ratios are nearly the same for the precipitation of ferrihydrite and schwertmannite. The next step is to use the predictions of the triangle diagrams to interpret drainage data from the CdA mining district. Dissolved SO4 concentrations in drainage from the district represent the oxidation of various sulfide minerals. However, only a portion of these minerals (i.e., pyrite) produce acidity during oxidation. To predict how the relative ratios of acid generating and acid consuming minerals vary throughout the district, we correct dissolved SO4 concentrations in drainage from the district for contributions from the major sulfide minerals that do not produce acidity during oxidation (i.e., sphalerite, galena, and chalcopyrite) by assuming that all dissolved Zn, Pb, and Cu concentrations are derived from the oxidation of these minerals. The 3352

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 19, 1999

FIGURE 4. Comparison of predicted (lines) and observed (symbols) relationships between normalized moles of dissolved SO4, the sum of dissolved Ca and Mg, and alkalinity (a) or acidity (b). Predictions consider the oxidation of pyrite, dissolution of average Coeur d’Alene (CdA) carbonates [calcite (solid lines), ankerite (dashed lines), or siderite (long dashed, three short dashed lines)], and precipitation of ferrihydrite or schwertmannite. Drainage from the CdA district is grouped by pH. The pH bands coincide with the mineralogical and thermodynamic studies of Fe minerals done by Bigham et al. (21). dissolved concentrations of SO4, Zn, Pb, and Cu are converted to a molar basis. Then dissolved SO4 derived from pyrite is calculated as the difference between dissolved SO4 and the sum of dissolved Zn, Pb, and Cu concentrations. In addition, because reactions that contribute Al to acidity (e.g., buffering by Al oxides and aluminosilicate minerals) are not considered in our theoretical approach, we correct the calculated acidity values for the Al contribution. This correction is less than 2% except for drainage from the Kellogg Tunnel (5%) and the Interstate-Callahan tailings pile (12-20%). Finally, the corrected dissolved SO4, the sum of dissolved Ca and Mg, and alkalinity (or corrected acidity) data are converted to a molar basis, normalized to 100%, and plotted on the triangle diagrams (see Figure 4). Near neutral to basic drainage has relatively moderate to high alkalinity, low to moderate acidity, and low to moderate dissolved SO4 concentrations indicating that these waters drain mineralized areas that have either very low reacting pyrite-to-carbonate ratios (i.e., near the top of the alkalinity

diagram and the bottom of the acidity diagram) or progressively increasing ratios of reacting pyrite to carbonate minerals. The relative concentrations of dissolved Ca and Mg in these waters suggest that, depending on location within the district, either calcite or ankerite appears to buffer the waters. The acidity diagram shows a clearer distinction between drainage that is buffered by calcite or ankerite. The reacting pyrite to calcite or ankerite ratios for these waters range from about 1/16 to 2/3. In contrast, acid waters from the Kellogg Tunnel and most tailings piles have little to no alkalinity, moderate to high acidity, and higher concentrations of dissolved SO4 suggesting that these waters drain mineralized areas that contain relatively higher ratios of reacting pyrite to carbonate minerals, i.e., up to about 1/1 for the Kellogg Tunnel drainage. The major buffering minerals for these waters also appear to be calcite and ankerite. The location of CdA drainage within the diagrams indicates that Fe(II) oxidizes and precipitates. Drainage data with pH > 4.5 are consistent with the precipitation of ferrihydrite or mixtures of ferrihydrite and schwertmannite. The location of acid drainage from the Kellogg Tunnel suggests that schwertmannite rather than jarosite is the important Fe hydroxysulfate mineral controlling the activity of dissolved Fe(III). These triangle diagrams provide a valuable stoichiometric tool for understanding the key parameters and fundamental reactions that determine the pH of drainage from adits and mining wastes. Our work indicates that the wide ranging pH values of drainage in the CdA mining district result from a combination of reactions involving pyrite oxidation, calcite or ankerite dissolution, and precipitation of ferrihydrite or schwertmannite. The reacting pyrite-to-carbonate ratios are estimated to range from near 0/1 to about 1/1.

Acknowledgments Yvonne Isaak, Kevin Myers, Glen Spinelli, and Jennifer Tonkin helped in collecting and processing samples in the field. Joel Leventhal, Katherine Walton-Day, Jennifer Tonkin, and four anonymous reviewers provided helpful comments on earlier drafts of this manuscript. Additional insights were gained from discussions with James Murray, Kathleen Smith, and colleagues at the 4th International Symposium on Environmental Chemistry where this research was first presented.

Literature Cited (1) Ficklin, W. H.; Plumlee, G. S.; Smith, K. S.; McHugh, J. B. In Water-Rock Interaction; Kharaka, Y. K., Maest, A. S., Eds.; Balkema: Rotterdam, 1992; pp 381-384. (2) Plumlee, G. S.; Smith, K. S.; Ficklin, W. H.; Briggs, P. H.; McHugh, J. B. In Proceedings: Planning, Rehabilitation, and Treatment of Disturbed Lands; Montana State University Reclamation Research Unit: Billings, 1993; Publication 9301, Vol. 1, pp 176186.

(3) Plumlee, G. S.; Smith, K. S.; Mosier, E. L.; Ficklin, W. H.; Montour, M.; Briggs, P. H.; Meier A. In Proceedings: Summitville Forum ’95; Posey, H. H., Pendleton, J. A., Van Zyl, D., Eds.; Colorado Geological Survey: Denver, 1995; Special Publication 38, pp 23-34. (4) Kelley, K. D.; Taylor, C. D. Appl. Geochem. 1997, 12, 397-409. (5) McCulley, Frick, and Gillman, Inc. Final Report; Upstream surface water sampling program, fall 1991 low flow event; South Fork Coeur d’Alene River basin above the Bunker Hill Superfund site, 1992. (6) Paulson, A. J. Appl. Geochem. 1997, 12, 447-464. (7) Fryklund, V. C., Jr. USGS Professional Paper 1964, 445. (8) Hobbs, S. W.; Griggs, A. B.; Wallace, R. E.; Campbell, A. B. USGS Professional Paper 1965, 478. (9) Zartman, R. E.; Stacey, J. S. Econ. Geol. 1971, 66, 849-860. (10) Bennett, E. H.; Venkatakrishnan, R. Econ. Geol. 1982, 77, 18511866. (11) Reid, R. R. In The Belt; Hobbs, S. W., Ed.; Montana Bureau of Mines and Geology: Butte, 1984; Special Publication 90, pp 49-51. (12) Leach, D. L.; Landis, G. P.; Hofstra, A. H. Geology 1988, 16, 122125. (13) Criss, R. E.; Fleck, R. J. Appl. Geochem. 1990, 5, 1-15. (14) Gitlin, E. C. PhD. Dissertation, University of Washington, 1986. (15) Balistrieri, L. S.; Bookstrom, A. A.; Box, S. E.; Ikramuddin, M. USGS Open-File Report 1998, 98-127. (16) Stumm, W.; Morgan J. J. Aquatic Chemistry; Wiley: New York, 1996. (17) Taylor, B. E.; Wheeler, M. C.; Nordstrom, D. K. Geochim. Cosmochim. Acta 1984, 48, 2669-2678. (18) Reedy, B. J.; Beattie, J. K.; Lowson, R. T. Geochim. Cosmochim. Acta 1991, 55, 1609-1614. (19) Singer, P. C.; Stumm, W. Science 1970, 167, 1121-1123. (20) Nordstrom, D. K.; Jenne, E. A.; Ball, J. W. In Chemical Modeling in Aqueous Systems; Jenne, E. A., Ed.; ACS Symposium Series 93, American Chemical Society: Washington, DC, 1979; pp 51-79. (21) Bigham, J. M.; Schwertmann, U.; Traina, S. J.; Winland, R. L.; Wolf, M. Geochim. Cosmochim. Acta 1996, 60, 2111-2121. (22) Blowes, D. W.; Ptacek, C. J. In Short Course Handbook on Environmental Geochemistry of Sulfide Mine-Wastes; Jambor, J. L., Blowes, D. W., Eds.; Mineralogical Association of Canada: Waterloo, Ontario, 1994; pp 271-292. (23) Sherlock, E. J.; Lawrence, R. W.; Poulin, R. Environ. Geol. 1995, 25, 43-54. (24) U.S. EPA Methods for Chemical Analysis of Water and Wastes; EPA-625-/6-74-003; U.S. Environmental Protection Agency: Washington, DC, 1974. (25) APHA Standard Methods for the Examination of Water and Wastewater; American Public Health Association: Washington, DC, 1976. (26) Langmuir, D. Aqueous Environmental Chemistry; Prentice Hall: NJ, 1997. (27) Allison, J. D.; Brown, D. S.; Novo-Gradac, K. J. MINTEQA2/ PRODEFA2, A geochemical assessment model for environmental systems: Version 3.0 User’s manual; EPA: Athens, GA, 1991.

Received for review August 13, 1998. Revised manuscript received July 8, 1999. Accepted July 12, 1999. ES980823C

VOL. 33, NO. 19, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3353