Lecture No. 22. Pit Lakes III

We will work out one more example calculation using activity coefficients.

Example: Find the solubility of calcium sulfate (gypsum) at 25¡ C.

The equilibrium is

CaSO₄ á 2H₂O ó Ca²⁺ + SO₄^2– + 2 H₂O

The water does not enter into the equilibrium constant. The solubility product constant is

{Ca²⁺} {SO₄^2–} = 2.45« 10^–5.

Besides being in equilibrium with the solid, the ions are also in equilibrium with a molecular complex:

Ca²⁺ + SO₄^2– ó CaSO₄

The equilibrium constant for this reaction is

At saturation, since the product of the calcium and sulfate activities is 2.45« 10^–5, the activity of the molecular complex will be

{CaSO₄} = 204.2« 2.45« 10^–5 = 5.003« 10^–3 M

We will have to allow for this species when we have finished calculating the concentrations of the ionic species.

First Approximation:

Use the concentrations as activities, i.e., assume the activity coefficients are approximately one.

[Ca²⁺] [SO₄^2–] = 2.45« 10^–5

[Ca²⁺] = [SO₄^2–] = 4.9497« 10^–3

Second Approximation:

Calculate ionic strength based on these activities.

I = ¸ [4.9497« 10^–3 (+2)² + 4.9497« 10^–3 (–2)²] = 4 « 4.9497« 10^–3

I = 1.9799« 10^–2

At this ionic strength we can use the extended Debye-HŸckel equation.

At 25¡ C, A = 0.5085 and B = 0.3281«10⁸. The values of a_i (« 10^-8) are 4.92 for Ca²⁺ and 5.16 for SO₄^2–.

Now we can calculate new concentrations for the ions and then a new ionic strength.

g _calcium [Ca²⁺] g _sulfate [SO₄^2–] = 2.45« 10^–5

Since [Ca²⁺] = [SO₄^2–],

[Ca²⁺] = [SO₄^2–] = 8.4517« 10^–3 M

Third Approximation:

I = 4« 8.4517« 10^–3 = 3.3807« 10^–2

Fourth Approximation:

I = 4« 9.5804« 10^–3 = 3.8322« 10^–2

I^1/2 = 1.9576« 10^–1

Fifth Approximation:

I = 4« 9.98945« 10^–5 = 3.95797« 10^-2

I^1/2 = 1.9895« 10^–1

Sixth Approximation:

I = 4« 9.979« 10^–3 = 3.9916« 10^–2

I^1/2 = 1.9979« 10^–1

Seventh Approximation:

I = 4« 1.0001« 10^–2 = 4.0004« 10^–2

I^1/2 = 0.20001

This is good to 4 significant figures, which is good enough for most purposes. At saturation, the concentration is 1.0007 M as calcium and sulfate ions.

But recall that there is also 5.003« 10^–3 M calcium sulfate dissolved as the neutral complex. This is also part of gypsum's solubility. The solubility of gypsum at 25¡ C is the sum of these, or 1.501« 10^–2 M.

The following table lists values for A and B used in the extended Debye-HŸckel and Truesdell-Jones calculations.

Note: The values of B are usually listed as the above values multiplied by 10⁸, and the values of a_i given in the last lecture are usually listed as the values I gave times 10^–8. Since B and a_i are always multiplied together in activity calculations, these factors cancel out.

Saturation Index

At saturation, the activity product of a dissolved ionic compound is equal to the solubility product. If the solution is undersaturated in the compound, the activity product will be less than K_sp, and if the solution is supersaturated, the product will be greater than K_sp. The Saturation Index is a parameter that indicates the degree of saturation of a solution. It is defined by the equation

For example, for calcium chloride the saturation index is given by

If a solution is exactly saturated, its saturation index is zero. A value greater than zero indicates supersaturation, and one less than zero indicates undersaturation.

Case Studies

(The following two case studies were presented in class using a slide presentation.)

1. Summer Camp Pit, Getchell, Nevada

Site description

The Summer Camp Pit is located at the Getchell Mine in Humboldt County, Nevada (41¡ 13¢ N, 117¡ 16¢ W). A former open-pit mine, the Summer Camp Pit was at the time of this study filled with water to a depth of about 18 m. It received precipitation and surface drainage, but primarily water pumped from an underground gold mine. The water in this lake was high in sulfates and mildly acidic; the acidity was partly moderated by carbonate rock in contact with the lake. The ore body is high in orpiment (As₂S₃) and realgar (As₄S₄) but low in metal sulfides other than pyrite. As a result, the lake water was low in metals other than iron, sodium, calcium, and magnesium, but uncommonly high in arsenic, typically containing several parts per million. The drainage water pumped from the underground mine contained small amounts of nitrate and nitrite, residues from explosives used in the mine. The water temperature measured in June 1996 ranged from 9¡ C at the bottom to 20¡ C at the surface; measurements in November 1996 ranged from 10¡ C at the bottom to 7¡ C at the surface. Dissolved oxygen levels measured in November 1996 decreased sharply from 7 - 9 mg/l near the surface to < 1 mg/l below 8 m.

Sample collection.

Sediment samples and depth integrated water samples were collected from the anoxic zone of the lake, between the depths of nine and sixteen meters, on Nov. 5, 1996. Three 20 L carboys (Nalgene ‰ ) were purged with N₂ gas for 5 min prior to being filled with sample. Carboys were fitted with caps containing three ports that allowed for continuous N₂ purging during sample collection. Water and sediment were retrieved through sterile silicone tubing (0.25 in I.D.) connected to a peristaltic pump. The tubing was taped to the cable of an oxygen/temperature probe (YSI Corp., Yellow Springs, OH), that had been marked at 0.5 m intervals, and connected to a YSI model 58 Dissolved Oxygen Meter. The probe provided sufficient weight to get the tubing to the lake bottom and allowed for the simultaneous collection of limnological data. Carboy # 1 contained bottom sediments and water, while carboys nos. 2 and 3 contained the depth-integrated water samples. Once the carboys were filled the pump was shut off and the inlet port capped. Each sample was purged for an additional 5 min with N₂ and the vessels were sealed with slight positive nitrogen pressure for transportation back to the laboratory. Samples were kept on ice during transportation (22 h).

Microcosm setup

Microcosms were contained in 5 L Nalgene ¨ polyethylene bottles. In order to minimize variation between microcosms, each microcosm was established with water and sediment from each of the three carboys. Initially 750 ml of sediment slurry from carboy #1 was added to each microcosm, followed by the addition of 1000 mL of water from carboy #2. The appropriate quantity of each amendment was added and each microcosm was then brought to a final volume of 3250 mL with water from carboy #3. Each of the 20 L carboys was continuously mixed while water was being transferred. All transfers were done under a constant stream of N₂ gas to maintain anoxic conditions. Microcosms were numbered 1 through 11, with microcosm 11 (MC11) having no added amendments. Each of the other microcosms received 0.5 g/L NaNO₃, 0.025 g/L KH₂PO₄, and 0.025 g/L K₂HPO₄ as bacterial nutrients. MC2 through MC6 received respectively 365, 1825, 3650, 7300, and 3650 mg/L organic carbon as waste potato peel. MC7 through MC10 received respectively 500, 2500, 5000, and 10,000 mg/L organic carbon as composted manure. In order to study bacterial activity in the absence of sulfate reduction, excess sodium molybdate (4.7 g/L) was added to MC6 as a competitive inhibitor of sulfate metabolism.

Container Used For Microcosms.

Microcosms were passed into an anaerobic glove box (Coy Laboratories, Grass Lake, MI) containing an atmosphere consisting of 88% nitrogen, 5% carbon dioxide, and 7% hydrogen. The glove box was equipped with an alarm to indicate any oxygen contamination. All sampling and microbiological processing were done under this atmosphere, with the exceptions of the sampling at 150 and 210 days. A methylene blue anaerobic indicator strip was placed inside the cap of each microcosm to indicate redox state. Between sampling intervals the microcosms were sealed, removed from the anaerobe chamber, and stored in an incubator at 15¡ C; this temperature was used because it approximated the average temperature in the lake.

Bacterial enumeration

Sulfate-reducing bacteria were enumerated using a five tube Most Probable Number (MPN) method

Analytical methods

All reagents used were ACS Analytical Reagent grade except as noted. Hydrochloric and nitric acids were Trace Metal grade. All water samples were filtered through 0.20 m m membrane filters. Samples for metal and metalloid determination were acidified to <pH 2 with nitric acid. After sulfide was found to affect arsenic solubility, dissolved arsenic was determined in unacidified samples. Solution pH was measured using a glass combination electrode. Metals and metalloids were determined by inductively coupled argon-plasma emission spectroscopy (ICAPES). Arsenic (III) was determined by hydride generation atomic absorption spectroscopy (HG-AAS). Anions were determined by ion chromatography. Non-purgeable organic carbon was determined using a Shimadzu TOC-5000A analyzer. Soluble sulfides and ammonia nitrogen were determined immediately upon sampling using a Hach DREL/2000 field kit. Total organic carbon in the composted manure was determined using a Coulometrics CO₂ Coulometer system, and in the potato waste it was determined using a Fisons Carlo Erba EA1110.

For ICAPES analysis, the solid organic amendments were digested as follows: Samples (0.5 g) of dried, ground material were sealed in polytetrafluoroethylene vessels with 5 mL aliquots of aqua regia and heated in a microwave oven. Digests were then cooled, diluted to 50 mL volumes, centrifuged, filtered through 0.2 m m membrane filters, acidified with 200 m L nitric acid each, and stored in clean polyethylene bottles for later analysis.

Organic Amendments

Potato waste and composted cattle manure were obtained locally by Getchell Mining Corporation. The amendments contained respectively 50.4% and 24.0% organic carbon (dry basis).

Results and Discussion

In microcosms to which organic amendments had been added, the nitrate concentrations dropped from initial values near 90 mg N/L to near zero within 31 days. At the same time, nitrite concentrations increased from zero to maxima near 30 mg N/L and then decreased to near zero within 60 days. This pattern is consistent with either denitrification or dissimilatory nitrate reduction by bacteria.

The concentration of ammonia nitrogen gradually increased in MC5 and MC10 from about 20 mg/L to about 50 mg/L. In MC2, MC4, MC6, MC8, and MC11 it remained approximately constant at 5 – 10 mg/L. In MC1 it dropped within five days from 7 mg/L to zero. In MC7 it remained constant at about 5 mg/L for the first two months, then decreased to 1 mg/L thereafter. MC9 fluctuated throughout the study between 15 mg/L and 40 mg/L. There were at least three sources and one sink of ammonia N in the systems. Likely sources included dissolution of soluble ammonia from the organic amendments (especially the composted manure), ammonification of organic nitrogen, and dissimilatory nitrate reduction; the most likely sink was nitrogen immobilization by bacteria. These multiple ammonia pathways probably explain the lack of any readily apparent trend in ammonia N concentrations in the microcosms.

The lake water was initially very low in dissolved oxygen, and care was taken to avoid oxygen uptake during transfer to the microcosms. Based on the appearance of the methylene blue indicator strips, all microcosms except 4, 5 and 6 remained anoxic during processing. The addition of higher amounts of potato wastes had the effect of slightly oxidizing these microcosms and retarding the onset of sulfate reduction. The addition of nitrate as a bacterial nutrient had the effect of elevating the redox potentials of the microcosms and preventing sulfate reduction for several weeks until the nitrate and nitrite were reduced. Many heterotrophic bacteria can utilize nitrate and/or nitrite as a terminal electron acceptor when oxygen becomes limited. In environments where both nitrate and sulfate are available, nitrate reduction is the predominant terminal electron accepting process. In pit lake remediation, where promotion of bacterial sulfate reduction is the objective, a better choice of nitrogen source would be an ammonium salt, which would provide reduced nitrogen rather than nitrate.

The water as obtained from the lake was at pH 6.1. The pH values increased in most cases to about 7, but in MC4 and MC5 the pH decreased to values below 6. MC4 eventually recovered to about pH 6, but the pH in MC5 remained low (5.2 at 210 days). The decrease in pH may have been caused by the release of organic acids from the potato waste in these microcosms. In MC11, the control, after 150 days the pH fell sharply and was at pH 4.95 after 210 days. This seems to have been caused by oxygen diffusing into the microcosm, with resultant iron oxidation and acid generation.

The concentrations of SRB, sulfate, and sulfide in all microcosms at the start of the experiment were approximately equal (1 x 10² bacteria/mL, MPN estimated; 1200 mg/L; and < 1 m g/L, respectively). In MC1, to which no organic nutrient was added, the concentration of these analytes remained essentially constant over the 210 days of the experiment. In MC6, which was amended with both a carbon source and molybdate to inhibit SRB, there was no change in sulfate or sulfide concentrations, and SRB numbers dropped to < 2/mL (data not shown). Following a variable lag period, densities of SRB in microcosms 3, 4, 8, and 9 climbed to concentrations that varied between 2 x 10⁴ and 3 x 10⁵ /mL (Fig. 2). The rise in SRB concentrations was followed by a decrease in sulfate and an increase in sulfide (Fig. 2).

Iron behaved as expected in an anoxic system subjected to an initial addition of nitrate followed by a gradual drop in redox potential. The redox potential for the Fe(III)/ Fe(II) couple is higher than that for the NO₂^-/NO₃^- couple under standard conditions, i.e., 1 molal concentrations of all reactants. However, at pH 6 the nitrate couple’s potential (0.408 V) is considerably higher than that of the iron couple, (0.173 V for 1 mM Fe²⁺).

The dissolved iron concentration in lake water was initially over 100 mg/L due to Fe(II) in solution. Within three weeks the dissolved iron concentration dropped to near zero in most microcosms as Fe(II) was oxidized by nitrate to Fe(III), which would precipitate as oxyhydroxides at the prevailing pH. Iron then increased again at 30 to 60 days, as nitrates and nitrites became depleted and the Fe(III) was reduced again to Fe(II). Finally, as sulfides appeared in solution, Fe(II) precipitated as sulfides and the dissolved iron concentration dropped again. MC5 and MC11 were notable exceptions to this general rule.

As in most other microcosms, the iron in MC5 was resolubilized between 31 and 60 days, apparently by reduction to Fe(II). However, at 210 days the iron concentration was still over 100 mg/L, although it appeared to be dropping at this point. The delay in precipitation of iron sulfides in MC5 was likely caused by slower sulfate reduction in this microcosm compared to several others (see Fig. 2). The iron and sulfide concentrations reflect a slow approach to equilibrium that may be related to the low pH and the high level of organic matter in the microcosm.

Between 60 and 150 days, the dissolved iron concentration in MC11 dropped from 156 mg/L to well under 1 mg/L (data not shown). During the same period, an orange-brown precipitate appeared on the sides of the bottle containing the microcosm. A nitrogen-carbon dioxide-hydrogen atmosphere was maintained in the microcosm, and nothing had been added to the water and sediment. The position of the precipitate suggests that oxygen had diffused through the walls of the polyethylene bottle. Polyethylene is known to be permeable to gases, and the time interval between 60 and 150 days was the longest time that the microcosms were stored without being purged with fresh reducing gases. It is likely that oxygen diffused into MC11, oxidizing the iron. This is consistent with both the position of the precipitate on the walls of the container and with the decrease in pH from 6.4 at 60 days to 5.6 at 150 days and 4.95 at 210 days, since oxidation of Fe(II) with precipitation of oxyhydroxides generates acid.

Total dissolved arsenic concentrations were initially 1 to 3 mg/L. They increased somewhat during the first five days and increased more slowly through the first 21 days (Fig. 3). This was unexpected, since at the time iron was precipitating from solution as oxyhydroxides in MC1 through MC10, and arsenates normally coprecipitate with iron oxyhydroxides. It is possible that the addition of 25 mM phosphate as a bacterial nutrient caused the displacement of adsorbed arsenic from mineral or organic surfaces, since phosphate and arsenate ions would tend to compete for the same sites.

Total arsenic and As(III) in MC11 dropped from 5.0 mg/L and 1.35 mg/L respectively at 60 days to near zero at 150 days (data not shown). This behavior is also consistent with inward diffusion of oxygen, since As(III) would be oxidized to arsenates, and arsenates tend to coprecipitate with iron oxyhydroxides.

In microcosms where anoxic conditions were established and measurable sulfide appeared in solution, arsenic concentrations fell; arsenic then rose with further increases in dissolved sulfide. This is consistent with the precipitation of arsenic sulfides such as As₂S₃, followed by redissolution of these sulfides to form thioarsenite complexes. In water containing 60 mg/L sulfide at pH 6, the solubility of arsenic is on the order of 5 mg/L.

The determinations of total arsenic in water samples from microcosms were performed by ICAPES, using solutions acidified to < pH 2 with nitric acid, in accordance with USEPA Method 200.7. As(III) was determined on unacidified samples using HG-AAS. At 31 days, for some of the more sulfidic microcosms, the measured As(III) concentrations were larger than the total arsenic measurements. This anomaly appears to have been the result of sample handling. In the highly sulfidic microcosms, As(III) solubility was increased by sulfide through the formation of thioarsenite complexes . In samples being analyzed by ICAPES, the added nitric acid oxidized some of the sulfide and thus destabilized the thioarsenite complexes, precipitating dissolved arsenic. By contrast, the samples analyzed for As(III) by HG-AAS were not acidified. The sulfide concentrations in those samples were undisturbed, and dissolved arsenic levels remained higher. Once this potential nitric acid interference was noted, further ICAPES determinations of As in sulfidic solutions were performed on unacidified samples. (ICAPES determinations of iron and other metals were still performed on acidified samples, in accordance with the USEPA method.)

Between 60 and 150 days, the arsenic levels in MC4 and MC5 dropped to near zero, although conditions in these microcosms (increasing sulfide concentration, moderate pH) were still conducive to the formation of thioarsenite complexes. This drop in dissolved arsenic is possibly due to conversion of the amorphous ferrous sulfides in the sediments (approximate composition FeS) to pyrite (FeS₂). Pyrite normally forms on a time scale of weeks to years but can form more quickly under favorable conditions. Such favorable conditions require the introduction of small amounts of oxygen to generate elemental sulfur by partial oxidation of sulfide:

S^2- + ¸ O₂ + H₂O ¨ S⁰ + 2 OH^-

followed by the reaction

FeS + S⁰ ¨ FeS₂

Oxygen can be introduced into sulfidic sediments by many mechanisms, including (among others) lake overturn, storm events, and bioturbation. In this case slow diffusion of oxygen through the walls of polyethylene containers, discussed above, might have brought about pyritization of sulfides in the microcosms over a period of a few months. Authigenic pyrite is known to be a strong arsenic scavenger that is able to accommodate significant amounts of arsenic in its crystal lattice. Thus, formation of pyrite is commonly accompanied by a decrease in dissolved arsenic.

Subsequent small increases in arsenic concentrations in MC4 and MC5 after 210 days are not as easily explained. The sulfide levels in these microcosms were increasing steeply during this interval and in fact roughly quadrupled between 150 and 210 days. Any arsenic released into solution would be kept in solution as thioarsenites. It is possible that arsenic could have been released by slow-reacting solid phases in the sediment, such as refractory oxides. Pyrite’s high thermodynamic stability makes it unlikely that arsenic would be released from that mineral by reaction with sulfide.

Dissolved organic carbon concentrations were approximately proportional to the amount of organic matter added. There was about twice as much readily-soluble organic carbon in the compost as in the potato waste. (During the first twenty-one days of incubation, an average 12.5% of the organic carbon added to compost-treated microcosms dissolved, while the dissolution rates for carbon in microcosms treated with potato waste averaged 6.3%.) Water samples taken later than 12 days after the start of the experiment from MC8, MC9, and MC10, which contained large amounts of compost, showed an intense dark brown color; these samples also produced a precipitate upon acidification with hydrochloric acid.

The only metals besides iron that were present at significant concentrations in the pit water were calcium, sodium, and magnesium. Concentrations of these metals remained fairly constant, except for increases in magnesium and sodium in samples from microcosms containing large amounts of potato waste or compost (data not shown).

This experiment has demonstrated the possibility, in principle, of remediating water quality in a mildly acidic pit lake containing high levels of sulfate, iron, and arsenic by the addition of inexpensive organic matter. In some of the microcosms, bacterial reduction of sulfate produced sulfide, and after a period of months the bulk of the sulfate, iron, and arsenic in the microcosms were removed from solution as precipitated sulfides. However, it is also evident that there are multiple poorly-understood variables that affect the rates of sulfate reduction, sulfide precipitation, and the eventual formation of stable mineral phases. The amount and type of organic additive had unexpected effects on the above processes as well as on solution pH. Removal of sulfate, iron, and arsenic and neutralization of acids did not improve uniformly with increased organic matter; rather there appears to be an optimum amount of organic matter that affords the fastest remediation. Moderate amounts of waste potato peel appeared to deliver the best combination of sulfate, iron, and arsenic removal.

Judging by these laboratory results, the best results in the field might be anticipated using about 0.2 g/L (200 g/m³) organic carbon in the form of potato waste. This waste, which is about 92% water, contains about 50% carbon on a dry basis, or about 4% C as received from the plant. Since the lake volume is roughly 2 « 10⁵ m³, it would require 4 « 10⁴ kg carbon, or about 1000 tons of wet potato waste. This amount of material, delivered over a period of months, is well within the capacity of the potato processing plant. It would require only about fifty trips by dump trucks with a standard 20 ton capacity, or correspondingly fewer trips by larger trucks.

After the initial treatment, it might be necessary to continue adding small amounts of potato waste to the lake as long as oxidizing waters from the underground mine were being discharged into the lake. However, with establishment of eutrophic conditions, a healthy aquatic algal population might produce enough organic matter on an ongoing basis to consume the added nitrate from the mine drainage water and obviate further waste additions. The tendency of the lake to stratify at times during the year would also tend to maintain a reservoir of organic matter in an anoxic or suboxic bottom layer. Even in much shallower lakes than the Summer Camp Pit, sulfidic bottom sediments rich in organic carbon have been found to act as sinks for sulfates and dissolved metals.

2. The Berkeley Pit, Butte, MT

From the Getchell mine's relatively small Summer Camp Pit we move on to the big leagues. The lake that now occupies the Berkeley Pit at Butte is a well-known example of a highly acidic pit lake containing high levels of metals and sulfate. It has a pH value near 2.3, is high in copper, zinc, iron, and manganese, and contains about 6400 ppm sulfate (or about 5 times the sulfate concentration in the Summer Camp Pit). ). Until now, no practical method has been found to remediate this water. Its high content of acids and metals, its large size, high rates of flow into the lake, and the high potential for future release of its waters into the Clark Fork River system via Silver Bow Creek all lend urgency to finding such a method.

Two low-cost organic additives that would be readily available in Butte were used. One was sewage sludge from the Missoula Sewage Treatment Plant, which is considered reasonably similar to that produced by the Butte-Silver Bow treatment plant. The other was wood compost from the Clark Fork Compost Company, Turah, MT (Table 3-2). Each additive was used at a ratio of two grams organic carbon per kilogram pit water, a level at which organics were successful in stimulating sulfate reduction in the Getchell pit water.

Solutes Present in Berkeley Pit Water.

Chemical Analyses of Organic Additives (Dry Basis). BDL = Below Detection Limit.

Experimental

All reagents were ACS Analytical Reagent Grade except as noted. Hydrochloric and nitric acids were Trace Metal grade. Deionized water was further purified using a Milli-q ¨ system. Ground dolomite rock, which was used in Microcosms 8-11, was Greenacres ¨ Dolomite Lime produced by Greenacres Gypsum and Lime, Inc., Greenacres, WA 99016. It carried a guaranteed analysis of 50% CaCO₃, 40% MgCO₃, 22% Ca, 13% Mg, and Neutralizing Value of 102% as CaCO₃.

Analytical Methods

All aqueous-phase samples were filtered through 0.20 m m filters. Solid samples of organic amendments were digested with aqua regia for ICAPES analysis. Dolomite was digested for determination of acid-soluble metals as follows: 10.00 g dolomite was digested overnight in 200 mL of 1.0 N Trace Metal Grade hydrochloric acid. A 25 mL aliquot of the digest was then filtered through a 0.2 m m membrane filter, treated with 200 m L concentrated nitric acid, and refrigerated at 4¡ C until analyzed.

Metals, metalloids, phosphorus, and sulfur were determined by ICAPES. Non-purgeable organic carbon was determined using a Shimadzu TOC-5000A analyzer. Soluble sulfides were determined by the methylene blue method using a Hach DREL/2000 field kit. Total carbon and nitrogen in sludge and compost samples were determined using a Fisons EA110 CHNS-O system calibrated against a 2,5-bis-(5-tert.-butyl-benzoxazol-2-yl)-thiophene standard. Inorganic carbon in sludge and compost samples was determined using a Coulometrics CO₂ Coulometer system calibrated against a calcium carbonate standard.

Microcosm Design

Water from the Berkeley Pit was obtained from Terence Dwaime of the Montana Bureau of Mines and Geology in 1993 and had been stored since then in polyethylene carboys. The pit lake water mixed with organic additives was stored in 2 L polyethylene bottles equipped with vented polypropylene screw closures. The closures are equipped with silicone rubber gaskets to ensure a gas-tight seal with the bottle and have three openings that can be closed with silicone rubber stoppers; two have barbed hose fittings above and below, and the third is a simple vent tube. This arrangement makes it possible to siphon liquid from the bottle under nitrogen pressure without contaminating the sample, the remaining liquid and solids in the bottle, or the bottle’s headspace with outside air. The microcosms were stored in the dark at 15¡ C.

Liquid samples were transferred by nitrogen displacement into bottles purged beforehand with nitrogen. Whenever samples were taken, one 30 mL aliquot from each sample was acidified to <pH 2 with concentrated nitric acid and another was similarly acidified with concentrated hydrochloric acid. All aliquots were stored at 4¡ C until analyzed. Separate aliquots were used for immediate measurement of pH and sulfide.

Six microcosms were initially started with no additives other than the organic amendments. Microcosms 1 through 3 (MC1 through MC3) were amended with municipal sewage sludge; MC4 through MC6 were amended with wood compost; and MC7 was a control microcosm in which untreated Berkeley Pit water was held under a nitrogen atmosphere. After 84 days the wood-compost microcosms were terminated. Another three microcosms (MC8 through MC10) were started in which the water was treated with 30 g/L of dolomite. After 4 h, each of the three microcosms received 1.5 L of dolomite-treated water and sufficient sewage sludge to supply 3 g organic carbon. The microcosms were then purged with nitrogen, sealed, and incubated at 15¡ C. A control microcosm (MC11), containing Berkeley Pit water treated with 30 g/L dolomite but no organic additives, was also incubated under nitrogen at 15¡ C. The headspace of each microcosm was purged once a week with nitrogen. Samples were taken from microcosms according to the schedule listed in the table.

Sampling Schedule For Microcosms.

Results

In most cases, sludge addition had larger effects on the principal chemical parameters in the water than did compost addition. In the microcosms treated with sludge, pH and dissolved organic carbon increased. Among metals in the sludge-treated microcosms, the concentration of iron increased substantially, presumably from Fe(III) reduction; copper and aluminum decreased, while calcium and magnesium increased. Other metals were essentially unchanged.

In the microcosms treated with compost, pH increased by a small but significant amount, and organic carbon increased slightly. Among metals in compost-treated microcosms, iron increased, then decreased, and copper decreased slightly. Calcium remained at a constant concentration in the compost treatments, whereas it decreased in the control. Magnesium and aluminum did not change relative to the control.

There were no significant changes in the concentrations of sulfate, silicon, or sodium with either additive, and no detectable sulfide was produced in any microcosm.

Because of the poor performance of both amendments when used alone in the pit water, a combination of organic matter and alkali was used next. Dolomite was used to adjust the pH for two reasons:

• It is relatively inexpensive and widely available.
• Dolomite is less subject to "armoring," i.e., formation of an insoluble sulfate coating that prevents complete dissolution. When added to pit water, limestone releases calcium ions that can precipitate with sulfate to form impervious gypsum coatings. Dolomite, on the other hand, releases both calcium and magnesium ions and, whereas calcium sulfate is sparingly soluble, magnesium sulfate is very soluble.

Three new microcosms (MC8, MC9, and MC10) were made up with dolomite and sludge, and a control microcosm containing dolomite (MC11) was also prepared, as described above in the Experimental section. These four microcosms were incubated at 15¡ C. Because of its low content of soluble organic carbon and its very poor performance in the first set of experiments, compost was not tested again.

The dolomite used in these experiments was digested with 1.0 N hydrochloric acid. It was found to be 84.2 % acid soluble, with quartz sand making up the bulk of the acid-insoluble matter. The results of the ICAPES analysis are listed in Table 3-3. Based on the assumption that all acid-soluble calcium, magnesium, and iron were in the form of carbonates, the material was 82.0% dolomite, which was 48.5 mol % CaCO₃, 51.4 mol % MgCO₃, and 0.1% FeCO₃.

Results of Acid Digest of Dolomite

The added dolomite raised the pH of the lake water from about 2.37 to 3.99 within four hours. The pH in the dolomite control microcosm (MC11) continued to increase slowly over the next eight weeks, leveling off eventually at about pH 5.5. In the microcosms treated with a combination of dolomite and sewage sludge, the pH continued to increase to about pH 6.0, at which point it also leveled off. Dissolved organic carbon increased to about 20 mg/L, then leveled off, in marked contrast to the microcosms treated with sludge alone. In those microcosms the final organic carbon concentrations were approaching 300 mg/L after 168 days. The two control microcosms contained very low levels of organic carbon at the beginning, and none was added. Sulfate dropped immediately from 0.64% to about 0.52% upon addition of dolomite. However, no sulfide was detected in any microcosm during the second study.

Metal concentrations behaved quite differently in microcosms treated with dolomite compared to those without dolomite. Copper dropped sharply from an initial concentration of 162 mg/L to about 20 mg/L at 56 days and less than 2 mg/L after 140 days. Iron increased gradually to a maximum concentration of 5-10 mg/L at 84 days, then slowly dropped off again. Zinc slowly dropped to about half of its original concentration. Aluminum dropped to near zero within the first 28 days. Magnesium increased, as expected with addition of dolomite. Calcium and sodium levels did not change much.

Although the sludge-treated water in the first experiment showed an increase in pH and a decrease in aluminum concentration (normally associated with pH increases), there was no evidence that bacterial sulfate reduction was occurring; in fact no sulfide was detected in any samples. The neutralizing effects of the carbonate and phosphate present in the sludge could account for most of the initial pH increase (from 2.3 to 2.8) in the sludge-treated microcosms without dolomite. The initial steep drop in dissolved iron within 7 days was a result of this initial pH increase, as iron (III) precipitated from solution as oxyhydroxides. The subsequent large increase in dissolved iron, coupled with a further increase in pH, is indicative of iron reduction. This does not necessarily prove that iron-reducing bacteria were active, since abiotic reduction is also a possibility. The lack of any sulfate reduction after 168 days might have been due to the pH, which was still too low for most sulfate-reducing bacteria or it might have been related to the high levels of dissolved metals in the water.

Metals behaved much differently in the second experiment, in which dolomite addition brought about an immediate large pH increase. Copper dropped from 162 mg/L nearly to zero. Aluminum also was removed almost completely from solution, while zinc decreased substantially and manganese modestly. Iron was reduced, but there was much less iron in solution than expected based on the high level of iron reduction seen with sludge addition in the first study.

Perhaps the most important difference between the two experiments was the much lower concentration of organic carbon in solution in the second experiment. This low level of dissolved organic carbon may be closely related to the drops in metal concentrations. The decreases in the concentrations of copper and zinc were not strictly the results of increasing pH. The saturation indices of both Cu(OH)₂ and Zn(OH)₂ were consistently below zero throughout the study, implying that their hydroxides would not precipitate. The table below lists the saturation indices for Zn(OH)₂, Cu(OH)₂, and Mn(OH)₂ as well as for gypsum in the microcosms.

If hydroxides were not precipitating these metals, something else must have been doing so. The consistently low levels of dissolved organic carbon may account for this. The same amount of the same sludge that supplied nearly 300 mg/L (25 mM) dissolved organic carbon under strongly acid conditions in the first experiment supplied no more than about 20 mg/L (1.6 mM) at pH 6 in the second study. Yet the near-neutral pH of the second study would normally favor the dissolution of organic acids (e.g., fatty acids or tannic acids) that are commonly present in sewage sludge more than the lower pH of the first study would. However, in the second study these same organic acids may have been precipitating copper, zinc, and, to a lesser extent, manganese (II) and iron (II). The combined concentrations of copper and zinc removed from solution in the second study totaled 6.5 mM. Thus, insoluble metal salts, if they are in fact forming, could easily account for the depressed organic carbon concentrations in the study.

It was pointed out above that the initial large decrease in sulfate concentration in the second experiment was not due to sulfate reduction, since no sulfides were detected. Sulfate precipitation as gypsum was caused by the initial increase in calcium concentration that accompanied the dissolution of dolomite. The calculated saturation index for gypsum was very close to zero in most samples taken in this study. (See table.)

While the combination of dolomite and sludge removed a large portion of the dissolved heavy metals from solution, it did not remove all of them. There were still on the order of 200-250 mg/L zinc and 180 mg/L of manganese. This probably accounts for the failure of SRB to reduce sulfate. Ueki et al. (1991) reported that 1mM (65 mg/L) zinc severely inhibited the activity of SRB; the Berkeley Pit water contains four times as much zinc. The high level of manganese may also be affecting the SRB, but possibly not severely; Ueki et al. also tested manganese and found no SRB inhibition under the conditions of their study.

One might ask what change in the treatment conditions would make it possible to remove the zinc from solution and enable SRB to thrive in the Berkeley Pit’s waters. The most obvious things to try would be increases in either the organic loading or the dolomite addition. The amount of sewage sludge used in this study is near the upper limit for application of organics, especially for a body of water as large as the Berkeley Pit. Since the sludge normally contains a lot of water, two grams carbon per liter of water (i.e., two kilograms per ton) entails adding fifteen to twenty times that weight of wet sludge to the water (thirty to forty kilograms of wet sludge per ton of water). On the other hand, the amount of dolomite used — 3% by weight, or about 1% by volume — is not excessive. This amount of dolomite adjusted the water’s pH from 2.3 to 5.5. The amount required to adjust the pH from 5.5 to 7.0 would amount to a relatively small increase in dolomite. The principal buffers to be titrated by the carbonate addition would be copper, manganese, and zinc. Assuming the neutralization would be accomplished and the copper and zinc removed by dolomite alone, 29 milliequivalents of carbonate per liter of water would be needed to precipitate both metals completely as the hydroxides. This would be about three additional grams of dolomite per liter — an increase of 0.3% by weight over the 3% already added.

Further experiments should be conducted to find an optimum way to apply SRB stimulation to remediating acidic, metal-rich lakes such as the Berkeley Pit. A possibly fruitful approach would be to add enough dolomite or limestone to adjust the water to pH 7 and therefore precipitate most or all of the copper and zinc from solution. This could then be followed by application of organic waste such as sewage sludge. The amount of sludge required should be significantly less than the 2 g/L of organic carbon used in these studies, since the sludge would be needed only to supply nutrients and increase oxygen demand, and not to contribute to neutralizing acid water.

To previous lecture: Lecture No 21. Pit Lakes II

To next lecture: Lecture No 23. Special Topics I

Back to list of lecture notes