As previously discussed in Chapter 5, iron-oxidizing bacteria can play a critical role in determining the rate of pyrite oxidation. The kinetics of acid formation are dependent on the availability of oxygen, the surface area of exposed pyrite, the activity of iron-oxidizing bacteria, and the chemical characteristics of the influent water. Thiobacillus ferrooxidans is generally regarded as the principal iron-oxidizing bacterium involved in pyrite oxidation (Leathen et al., 1953; Kleinmann and Crerar, 1979). Acidification progresses in a three-stage sequence dependent upon the activity of T. ferrooxidans and solution Eh and pH (Kleinmann et al., 1981). The bacteria serve as a reaction catalyst.

During the first stage of this process, fine-grained pyrite can be directly oxidized by T. ferrooxidans or can be abiotically oxidized by air, with equal amounts of acidity produced by the oxidation of sulfide to sulfate and by the hydrolysis of Fe³⁺. During this stage, it is possible to forestall acidification by adding alkalinity (lime, limestone, etc.) to the reaction system; if alkalinity exceeds acidity, the only major downstream effect is an increase in sulfate concentration. As the pH declines, abiotic oxidation of Fe³⁺ slows as much as 100-fold for each pH unit, and T. ferrooxidans takes on its primary role of oxidizing Fe²⁺. This transition stage is referred to as stage 2. Stage 3 begins when the decreased rate of Fe(OH)₃ precipitation results in increased Fe³⁺ activity. The Fe³⁺ rapidly oxidizes the pyrite, producing Fe²⁺ that is then oxidized by the bacteria to Fe³⁺. This cyclical reaction series greatly accelerates the rate of acid generation at many mine sites.

The importance of oxygen availability is illustrated in Fig. 15.1. At oxygen concentrations above 14%, the rate of abiotic pyrite oxidation is essentially equivalent to what is observed when the bacteria are present and active. However, at lower oxygen levels, the bacteria assume greater importance. For example, at 1% oxygen, the rate of reaction is over seven times faster if the bacteria are active (Hammack and Watzlaf, 1990).

Figure 15.1 Rates of pyrite oxidation with and without iron-oxidizing bacteria in small columns maintained at different oxygen partial pressures (Hammack and Watzlaf, 1990).

Thus, bacterial catalysis is probably not critical in well-ventilated sections of active underground mines but could be very important once the mine is abandoned. In unconsolidated, highly permeable mine spoil, oxygen concentrations often exceed 15%, but in coal refuse piles and low-permeability, shaly spoil, oxygen concentrations decrease dramatically with depth, effectively limiting abiotic pyrite oxidation to a thin near-surface layer less than a meter thick (Erickson et al., 1982; Guo and Cravotta, 1996).

In mine environments where bacterial activity determines the rate of acid generation, inhibition of these bacteria can prevent acidification or greatly reduce the acidity that is produced. This possibility was first considered in 1953 but was rejected as probably impractical (Leathen, 1976). Unsuccessful attempts followed in the 1960's (Barnes and Romberger, 1968; Shearer et al., 1970), followed by successful efforts over a decade later (Kleinmann, 1979; Kleinmann and Erickson, 1982).

The biological literature contains numerous studies of T. ferrooxidans that indicate a vulnerability to

certain metals (such as mercury, tellurium, and molybdenum), thiocyanate, organic acids, anionic surfactants, and food preservatives (Schnaitman et al., 1969; Tuovinen et al., 1971; Imai et al., 1975; Dugan, 1975; Tuttle et al., 1977; Onysko et al., 1984; Sobolewski, 1993). Only anionic surfactants and a food preservative have warranted field testing, and only the surfactant approach has been shown to be cost effective (Kleinmann and Erickson, 1982; Watzlaf, 1986).

Anionic surfactants are commonly regarded as good cleansers but poor bactericides (Walters, 1965), as opposed to the germicidal cationic surfactants commonly used in hospitals. However, it has been shown that anionic surfactants are markedly more inhibitory at low pH (Dychdala, 1968). The various means by which surfactants affect microorganisms have been summarized by Hugo (1965), who concluded that alteration of the semipermeable properties of the cytoplasmic membrane is the most typical mode of inhibition. T. ferrooxidans possesses a multilayered cell wall, which allows it to maintain an approximately neutral internal pH despite the extremely acid environment in which it lives (Howard and Lundgren, 1970; Adapoe and Silver, 1975; Langworthy, 1978; Ingledew, 1982). At low concentrations, it appears that anionic surfactants induce seepage of H⁺ into the cell, which slows Fe²⁺ oxidation by decreasing the activity of pH-sensitive enzymes. Higher concentrations of the surfactant kill the bacteria, presumably by causing permanent damage to these enzymes and the membrane material (Hotchkiss, 1946; Lundgren et al., 1974; Dugan, 1975; Kleinmann, 1979).

Figure 15.2 shows the effect of three anionic surfactants on acid generation in the laboratory. Sodium lauryl sulfate (SLS) was the most effective, killing T. ferrooxidans at 25 mg/L; alkyl benzene sulfonate (ABS) and alpha olefin sulfonate (AOS) required somewhat higher concentrations.

Based on these laboratory results, field tests were initially conducted using SLS as the inhibitory agent. Five hundred and fifty gallons of a 30% solution of SLS was diluted 175:1 with water and applied with a hydroseeder to an 11 ac (4.4 ha) inactive coal refuse pile. Water quality improved after a 3 month lag period, with a 60% decrease in acidity, sulfate and manganese and a 90% decrease in iron concentrations (Kleinmann and Erickson, 1981, 1983). The lag period was presumably caused by the time required for infiltration to flow through the old pile, but stored acidity in the form of sulfate salts was probably also a factor (see Chapter 1). The magnitude of the improvement was greater than any change observed over the prior 10 years, and seasonal effects can be discounted, since in previous years the contaminant concentrations peaked in winter.

Figure 15.3 illustrated the results of a similar application to an active refuse pile in northern West Virginia. Runoff water quality improved dramatically within a month of the SLS application. Acidity, sulfate, and iron concentrations were reduced by more than 95% and remained low for about 4 months after treatment.

Effluent concentrations of surfactant were extremely low at both sites. Except for one measurement of 0.6 mg/L shortly after application of the SLS to the inactive pile, SLS concentrations were consistently less than 0.1 mg/L; no SLS was detected in the stream at the discharge point of either treatment plant.

As a result of these tests, mining companies began to apply surfactant solutions at active refuse and coal storage piles. Initially, SLS was used, but it was subsequently found that ABS, which is a common ingredient in many laundry detergents, was more cost effective. At such sites, the surfactant solution is either diluted and applied to the pile periodically (3-4 times a year) or sprayed onto fresh refuse just before it is added to the pile (Rastogi, 1996). Either approach avoids the problem of bacterial reestablishment, which can be observed occurring 120 days after initial treatment in Figure 15.3.

It should be noted that the U.S. EPA regulates the use of bactericides under the Federal Insecticide, Fungicide and Rodenticide Act (FEFRA). Therefore, only those surfactant or bactericide formulations that are registered under FIFRA can legally be used to avoid or abate acid generation.

Periodic or continuous application of surfactant solution can be appropriate at an active site where new pyritic material is always being added, but cannot be continued after the pile is completed and revegetated. The surfactant would be adsorbed by the soil and never reach the pyritic material. If pyrite oxidation is allowed to proceed unchecked beneath the soil cover, the roots of the vegetation will be exposed to acidic water and will wither away; soil erosion will soon follow.

To counteract this problem, slow-release surfactant pellets that could be applied before the soil cover were developed (Kleinmann, 1982). The surfactant migrates from the interior of the pellet to the pellet surface to replace the surfactant that is dissolved each time the pellet gets wet. Initial formulations used a rubber matrix and lasted about 2 years. Subsequent formulations made of polyethylene and developed for commercial applications have release lifetimes of 6 or more years (Splittorf and Rastogi, 1995), assuming that a soil cover is applied and revegetation commences soon after the surfactant application. The release lifetime is intended to allow time for vegetation and normal soil bacteria to become established in the topsoil layer, consuming the oxygen that would otherwise fuel pyrite oxidation.

Based on a recent revisit to a ten year old field test, it appears that this premise is valid (Splittorf and Rastogi, 1995). The 1984 field test was conducted by the Ohio Department of Natural Resources using an early slow-release formulation: SLS in a rubber matrix. SLS was applied in solution (at the rate of 225 kg/ha) and as pellets (575 kg/ha, containing 16-28% SLS) to a 1.0 ha section of a regraded refuse pile. An adjoining 0.9 ha section received no surfactant and served as a control. Both areas were then covered with 15.2 to 20.3 cm of topsoil, fertilized, limed, seeded and mulched.

Water quality, vegetative cover and activity of normal soil bacteria were all markedly improved by the bactericidal treatment (Sobek et al., 1990). These improvements were sustained long beyond the 2 year lifetime of the pellets. Five years after reclamation, biomass production was nine times greater (2915 kg/ha vs. 315 kg/ha) and acidity in the vadose zone was 80% lower in the treated area than in the control plot. After 10 years, 35-40% of the control area is barren and eroding; the treated area has no significant erosion and 4,118 kg/ha of biomass, including extensive volunteer vegetation (Splittorf and Rastogi, 1995).

The only published results of the first field trial of the ABS polyethylene pellets indicated a similar level of success after 3 years. Acidity was reduced by 96%, sulfate concentration was reduced by 91%, and total iron concentration was reduced by 97%. Biomass production totaled 1,604 kg/ha in the surfactant-treated area, compared to 1,033 kg/ha on the control area (Sobek et al., 1990). Subsequent field applications have no control area for comparison but appear to be effectively preventing or delaying acidification of selectively handled mine spoil and improving water quality by 82-90% at active coal refuse areas (Parisi et al., 1994).

Surfactants have proven to be cost effective when applied directly to highly pyritic material. Since the surfactant must reach the pyritic material to be effective, sites that have been reclaimed with topsoil or backfilled with non-pyritic overburden are generally not appropriate for surfactant treatment. The surfactant will absorb onto the intervening material and never reach the pyrite.

Another factor to be considered is the response time before a surfactant application has an effect on site effluent water quality. Response time is determined by groundwater flow-through time and stored acidity. Groundwater flow-through time is the time it takes for rainwater to infiltrate through the mine material and emerge in a spring or seep. This can vary tremendously from site to site but improvement in water quality at the discharge point cannot occur faster than groundwater flows through the material.

Acidity can be stored as metal sulfate salts that dissolve over time, or ponded as acidic water on the old mine floor or in a refuse area. Such stored acidity retards or masks the effect of reduced acid production. The presence of substantial amounts of stored sulfate salts is indicated if high flows are accompanied by high concentrations of contaminants. The presence of acidic groundwater can sometimes be surmised by a steady acid load in the base flow during dry periods and only slight changes in the acid load during higher flows. If the stored acidity is significant, it could be several months before a decrease in acid production is reflected by more than a gradual improvement in water quality. If faster results are desired, alkalinity must be added to the site in sufficient quantities to neutralize the acidity.

Application rates are largely site-specific, and are heavily dependent on the adsorptive capacity of the material being treated. Laboratory tests have been developed to evaluate this aspect (Kleinmann and Erickson, 1983). However, if one is trying to evaluate the potential cost-effectiveness of applying a bactericidal treatment, consider setting up pilot-scale field tests in well-washed, plastic 55 gallon drums. Rastogi et al., (1995) have shown that small test piles do not accurately simulate larger sites, presumably due to the fact that oxygen concentrations are unrealistically high in the small test piles.

Both hydroseeder and road-watering trucks have been used by mine operators for application of the surfactant solution. With either machine, it is important to add the water to the tank at the bottom (using a hose) to avoid excessive foaming. For the same reason, agitation should be avoided as much as possible. Care should be taken when dealing with the concentrated surfactant; gloves should be worn to avoid "dishpan hands," and contact with the skin and eyes should be avoided. Finally, it is recommended that the tank and hoses be thoroughly rinsed of all surfactant after the application is completed, as the concentrated solution can cause pitting if allowed to remain.

Controlled-release pellets should be applied along with a surfactant solution to pyritic material just before the material is covered with soil or clay. A hydroseeder can be used. The surfactant solution satisfies the adsorptive capacity of the pyritic material; without it, the surfactant released by the pellets will reach very little of the reactive material. Alkalinity can be applied along with LAS (an important advantage over SLS) if needed to neutralize already-formed acidity.

Researchers periodically evaluate alternative bactericides for potential cost effectiveness or extended duration of treatment relative to anionic surfactants. Recently, thiocyanate was investigated in the laboratory; unfortunately, it proved to be effective for only 30-60 days. Apparently, it was converted to ammonia (Sobelewski, 1993).

An earlier laboratory study, evaluating alternative antibiotics and antibacterial agents, identified nitrapyrine as an effective inhibitory agent. Nitrapyrine is the active agent in a commercially-available nitrification inhibitor. However, the nitrapyrine did not appear to be as cost effective as the less expensive anionic surfactants (Sherrard et al., 1990).

Researchers are beginning to evaluate the role that bacteria play in catalyzing acid generation in underground mines and whether it is possible to control their activity in such an environment. Due to the high oxygen concentrations in an active, ventilated mine, the bacteria may be less critical in such an environment; however, they are probably very important in unflooded or partially flooded abandoned underground mines. Inhibitors applied in solution would not be appropriate in such an environment. An inhibitory gas or vapor that can permeate into rock fractures, or a fine mist or aerosol of one of the established inhibitors, may be better. Alternatively, potentially inhibitory compounds with low vapor pressure could be effective; acrolein (2-propanol) for example, kills thiobacilli, though the possibility of groundwater contamination may preclude its use. Another option is to alter the environment so as to ecologically favor bacteria other than T. ferrooxidans.

At present, anionic surfactants are the only appropriate means to directly inhibit the iron-oxidizing bacteria that catalyze acid generation. Surfactants have proven to be cost effective when applied directly to highly pyritic material; they delay or prevent acidification and reduce pyrite oxidation rates 60-95%. They are most effective when applied to compacted or fine-grained material, where most of the pyrite oxidation is occurring near the site of treatment. Dilute surfactant solution can be applied to exposed pyritic material 3 to 4 times a year or can be applied to fresh waste material as it is being transported for disposal. Controlled release of surfactants is appropriately used just before the pyritic material is covered with non-pyritic material; such an application has been shown to have long term benefits with respect to water quality and revegetation.