3. Testing Plans, Design, and Protocol

At one end of the design continuum, the danger in any cookbook design approach is that it may not properly address conditions other than those that were originally used to develop the standardized design criteria. For example, the treatment of low pH water containing dissolved aluminum is especially problematic and outside the original U.S. Bureau of Mines design criteria, which addresses the issue by restricting the application of anoxic limestone drains (ALDs). A precise and reliable aluminum design guideline has yet to be developed for ALDs and probably should not be considered due to the complexity of aluminum chemistry. While iron can be precipitated aerobically as ferric hydroxide or anaerobically as a sulfide or carbonate, the list of aluminum mineral species found in nature (and thereby possible in a passive treatment system) is extensive. If heavy metals are present (for instance, in MIW from metal mines), then the design challenges multiply rapidly. Cookbook approaches become futile for passive treatment design when the effect of varying anionic concentrations and water temperature are added to these complications.

At the other end of the design spectrum, mining, chemical, and other industries have long used a phased design process. The concept is simple: start small, learn from failures, and build on successes until the data required to properly design a full-scale treatment system are available. These data significantly reduce the risks of the full-scale system failure or less than optimum performance. Wildeman, et al. (1993) propose a design protocol that includes laboratory-, bench-, and pilot-scale phases. This approach has been used at dozens of MIW sites.

A phased-approach design project typically begins in the laboratory with static tests, then graduating to final testing phases (bench and pilot) performed at the site on the actual MIW. Bench-scale testing determines if the treatment technology is a viable solution for the MIW and narrows initial design variables for the field pilot test. A proper bench-scale test reduces the duration of the more costly field pilot test. Field pilot test duration can range from days to years, depending on the nature of the technology. Depending on the equipment and personnel needed, significant costs may be incurred during the field pilot tests (mostly for sampling and analysis); however, these costs may be significantly less than active treatment pilot tests. This section describes a phased approach, with suggested test phases:

1. Identify and select solid or liquid substrateEither (a) a chemical which reacts or (b) a solid surface or (c) an electron donor. candidates (see Section 3.1).
2. Measure bulk substrate properties (see Section 3.2).
3. Conduct bench-scale quick paste tests (see Section 3.3.1).
4. Conduct bench-scale four-week bottle testing (see Section 3.3.2).
5. Conduct small-scale (barrel-scale) three-month or longer test with flow (see Section 3.4).
6. Conduct pilot-scale field test (see Section 3.5).

Detailed descriptions of testing phase activities are provided in the following sections.

The previous description assumes a solid substrate system; one variant on a BCR is the active addition of a liquid organic substrate (see text box below). In this type of BCR, the media might simply be gravel or local trap rock with an initial inoculum of organisms, as the liquid organic feed rate is subject to control.

Liquid substrates may be selected for design reasons that include the following:

• to start up a BCR
• to stimulate microbial activity in an upset BCR
• if space constraints limit the size of the BCR (such as in urbanized areas)
• if a low-cost liquid-phase substrate is available

Liquid substrates can be locally available waste materials, such as dairy whey, or commercially available products.

Selection of a liquid phase substrate or a combination of liquid substrates is typically based on the following factors: the ease with which the substrate can be degraded by SRBsulfate-reducing bacteria, cost, and ease of handling. Typically, liquid-phase substrates are simple organic compounds that are more easily degraded by SRBs than solid-phase substrates thereby increasing reaction rates within the BCR and reducing the size of the BCR required for a given flow and influent composition. Acids such as citric and acetic, or salts of the conjugate bases, may be selected for liquid-phase substrate BCR. Common antifreeze components ethylene and propylene glycol could also be tested. In addition, low molecular weight alcohols such as ethanol and methanol are candidates for liquid-phase substrate applications (see Appendix B.8 Leviathan Mine Case Study). Low molecular weight acids, such as those found in the microbial citric acid cycle, are typically more easily degradable than alcohols to various classes of micro-organisms. As a result, organic acids may be selected to support BCR start-up. Liquid substrate BCRs usually require more maintenance than solid substrate BCRs due to the need to store and replenish the liquid substrate. Control of the liquid substrate feed rate is important, although concerns over the potential for clogging of liquid substrate BCR from rapid growth of biomass are offset with adequate design and constraint of the substrate loadingMass of something per time entering a volume (volumetric loading rate) or flowing into an area (areal loading rate). rate. Liquid substrates can also be used to decrease the start-up time of a solid substrate BCR with careful selection of the substrate type. An initial substrate screen can identify local sources for materials.

Selection of a liquid-phase substrate may also depend on whether the BCR design targets a diverse community of microorganisms, such as wood (cellulose) degraders, or targets growth of a specific microbial population (as is typical for a BCR with no solid-phase substrate). This design decision is often based on the composition and ratios of the dominant electron acceptors in the BCR influent and the available space for the BCR. The potential for variability in the chemical composition of the BCR influent may also influence the selection of a liquid-phase substrate and associated feed equipment (see Appendix B.8 Leviathan Mine Case Study).

In general, numerous liquid-phase substrates could also be used in a BCR (such as alcohol). Liquid substrates can also be locally available waste materials, such as dairy whey, or commercially available products. Commercially available products used for BCR have included food-grade acetic acid, citric acid, methanol, ethanol, alcohol, calcium magnesium acetate, ethylene, propylene glycol, and sodium benzoate.

Resources for finding locally available substrate include the following:

• Search general internet search engines, electronic business pages, and phone books.
• Base searches on distance from site (25 to 50 miles).
• Search based on the key words for various materials presented in Table 3-2 below:

Some materials are not likely to be locally available but may be cost-competitive. For example, the chitin used in BCRs to date is Chitorem^® from JRW Bioremediation. When vendors are identified, contact them to determine what products are available, quantities in stock, rough unit prices, whether they have loading machinery (large orders), and who generally buys their materials (the competition). For composted materials, be aware that the quality of the compost stream from local governments may vary significantly over time as citizens vary what they provide for collection.

1. First determine the weight of the empty bucket (tare weight). Then add the material.
2. To establish individual material density, the bucket should be filled with the material being tested to the 20 liter mark and compacted by stepping on the material with the weight of a typical person (90 kg). Apply the same level of compaction to all materials being tested so that the values derived are comparable. After the first compaction, top off the bucket to the 20 liter mark and weigh on the scale. Gently move or jostle the bucket to assure the reading on the dial is repeatable (approximately).
3. Record the tare weight of the bucket, the calibration volume (20 liters in this case), and the gross weight for subsequent calculations. Repeat as needed for all materials. If some materials tend to stick to the sides of the bucket, rinse the bucket with fresh water and dry with a towel.
4. Hay or straw bales are a special case. These are typically weighed intact directly on the scale. They may need to be stood on end and carefully balanced in the center of the scale pad. The bale volume is determined by direct measurement (L x W x H). If possible, several straw bales should be measured to obtain an average bale weight. An example worksheet for the calculations is provided in Table 3-2.
5. To establish a mixed substrate density, begin by using the proportions provided for the various organic and inorganic materials, then weigh out enough of each to prepare at least 25 to 30 liters of mixture. First fill the 20-liter bucket to the 16 liter level (80% full) with the most abundant individual material in the mixture, such as the wood chips. Obtain the weight of this volume and then weigh out the remainder of the mixture components proportionately.
6. Mix the components together thoroughly until relative homogeneity is achieved. It may be necessary to “shred” the hay or straw to short stem lengths for acceptable mixing for this small amount of volume. Measure the weight of the mixture. Using the same compaction practice as previously described, measure the volume of the mixture so that the bulk density and the “shrinkage” values can be estimated.
7. Shrinkage is estimated by dividing the volume of the mixture by the sum of the volumes of the individual components. For example, if 24 liters of the mixture (compacted) resulted from the combination of 16 liters of wood chips, 3 liters of manure, 5 liters of seashells, 3 liters of grape skins, and 6 liters of hay, the shrinkage factor would be:

24 / (16+3+5+3+6) = 0.73 or 73%

3.3.1.1 Procedure

1. Prepare a paste consisting of 1 part (by volume or weight) solid substrate material (for example: wood chips, sawdust, manure, or hay) and 2 parts distilled water. Mix thoroughly and allow to stand for one hour. The paste can be mixed in sandwich-sized heavy-duty plastic bags, which can be subsequently discarded. Be sure to label the bags with an indelible marker.
2. Usually about 50 grams of solid material and 100 ml of distilled or de-ionized water will suffice for this test. If the solid material absorbs water excessively, additional moisture may be needed. If necessary add water, record how much is added, and wait another hour before taking measurements as described below.
3. Test the paste for pH, conductivity, and ORP. To obtain a measurement, collect the paste into a corner of the plastic bag and carefully insert the pH/conductivity/ORP probe(s) into the paste so as to avoid damage to the probe. Make sure the tip of the probe is under water or in good contact with the paste.
4. For best results, measure the pH values of the samples first, then measure the conductivity of the pastes, followed by the ORP values. The probes should be rinsed with distilled water after each measurement. If an ORP probe is not available, the pH and conductivity values will suffice. Record all values. Photograph the individual substrate materials close up with a digital camera (equipped with a macro lens if possible) with a ruler or scale in the photo.

3.3.1.2 Data Interpretation

To be a good stand-alone substrate candidate, the paste should ideally have:

• Near neutral pH. This value demonstrates that the substrate alkalinity is able to dissolve into the distilled or de-ionized water.
• Conductivity value of at least 2000 µmhos/cm. This value indicates the substrate alkalinity is able to dissolve into the distilled or de-ionized water and that ion exchange for short-term adsorptive removal is possible.
• ORP value of less than 200 mV. To convert ORP readings taken under these conditions to EhThe redox potential is the tendency of a compound to gain an electron. This is most often measured as the voltage required to prevent electrons to transfer between the measured sample and a standard reference electrode. For Eh, that standard reference, defined as zero volts, is H2 → 2 H+ + 2 e- at a specified standard condition., simply add 200 mV to the ORP voltage (see "Measuring ORP" from YSI Environmentalhttp://www.ysi.com/media/pdfs/T608-Measuring-ORP-on-YSI-6-Series-Sondes-Tips-Cautions-and-Limitations.pdf). This value shows that biological activity has used the oxygen that was in the water, confirming organisms and short-term food are present. An ORP meter reading should be corrected to Eh, since this hydrogen-referenced value is standard when describing redox in biology.

Dozens of test bottles with duplicates can be monitored during the course of the proof of principle test. This extensive monitoring is the key advantage to a proof of principle test using culture bottles. By using small amounts of substrate and MIW, and keeping to a simple experimental design, the test can assess a greater number of removal possibilities and better indicate appropriate substrate and aqueous conditions. Because of cost and time considerations, test only those substrate that are most likely to work based on experience and material availability.

Optimization of the substrate inoculum for maximum contaminant removal is the final step, and data to support this step is developed through dynamic bench scale testing further described in Section 3.4. Success in the proof of principle test is defined by a measurable (typically 50 percent) decrease in contaminant concentration over the course of the experiment, rather than a decrease compared to a water quality standard.

3.3.2.1 MIW Collection Procedures

Begin the Phase 2 bottle testing process by obtaining at least 8 liters of the mine drainage water to be treated. During the collection process, observe the following procedures:

• Collect unpreserved, whole (unfiltered) MIW.
• Check and record pH and other field parameters.
• Rinse each sample collection container once with a small amount of MIW.
• Transport the sampled raw MIW for proof of principle testing immediately to the lab treatability testing facility and keep on ice or refrigerated.

• Fill appropriate sample collection bottles, add preservatives as appropriate, and ship samples to the analytical laboratory for testing.

The MIW sample collection bottles used should filled completely and capped to prevent oxygen intrusion. Include appropriate duplicates of the MIW so that intermediate and confirmatory tests can be performed during the course of the experiment. Assess preliminary results of the analyses after two weeks of testing, and weekly thereafter as results are available.

3.3.2.2 Bottle Setup Methods

From the eight liters of MIW, two 250 mL aliquots should be taken and properly preserved. These aliquots are set aside to be available for analysis at a future date if necessary. Two additional 250 mL aliquots should be analyzed for pH, Eh conductivity, alkalinity, and a full suite of metals and anions, including sulfate. A Hach kit can be used to measure sulfide (which is expected to be low in typical MIW, for instance, less than 0.1 mg/L sulfide).

The quantity of MIW used in proof of principle tests are project specific, but at a minimum should provide sufficient MIW volume to allow for collection of sample aliquots throughout the test or at the end of the test, including laboratory analysis of metals and other parameters. Proof of principle bottles should contain substrate (at as-received moisture) weighed into the bottle. The unfiltered MIW should be added at the rate of about 250 ml for experiments with 100 g of substrate plus 30 g of inoculum. Minor amounts of distilled water can be added to minimize headA specific measurement of water pressure above a geodetic datum. It is usually measured as a water surface elevation expressed in units of length. space.

1. Shake the bottle to mix the inoculum with the substrate materials.
2. Before capping and sealing, measure the pH, Eh and conductivity of the solutions.
3. After measurements, all BCR test bottles may be purged with nitrogen or argon gas for approximately 1 minute to remove dissolved oxygen – this is optional.
4. The samples should be incubated at room temperature, 25 to 30° C.
5. For BCR tests, the electron acceptors are NO₃ (denitrifiers), SO₄^2- (sulfate reducers) and CO₂ (methanogens). These compounds produce gases of N₂, H₂S, and CH₄, respectively (some more than others). Therefore gas appearance is expected but does not confirm sulfate reduction.
6. The progress of sulfate reduction can be followed by observing the color of the sediment in the bottles, which changes from brown to black. Test for free sulfide using a Hach kit.
7. The Eh (redox) of the solutions also indicates that sulfate reduction is probably occurring; sulfate reduction is expected to occur in reducing conditionsA system in which the gain of electrons is energetically favored due to a low reduction potential. of less than -150 to -200 mV.

3.3.2.3 Observation Schedule

Every seven days, inspect the anaerobic bottles for bacterial activity:  gas evolution, bubbles, color changes, and the smell of rotten eggs (sulfide). In addition:

1. At 14, 21, and 28 days, test for sulfide, metal of concern, Eh, and pH of accessible culture solutions.
2. Between 21 and 28 days, photograph representative culture bottles.
3. At 28 days the experiment is considered finished, unless no significant microbial activity has taken place. Measure Eh, pH, and conductivity of all the solutions.
4. Based on Eh, pH and appearance, the ‘best’ of the BCR bottles should be analyzed for the same suite of parameters as the original suite.

3.3.2.4 Interpretation of Results

The bottle tests demonstrate the production of alkalinity and the amount of sulfate reduction and sulfide produced by the substrate mixes that are tested. An ideal result is significant sulfate reduction (100 mg/L or more) and sulfide production (5 mg/L or more), pH near neutral, and Eh below -200 mV. The presence of sulfide should be accompanied by very low concentrations of dissolved metals other than manganese, which is not well removed in a sulfate reducing system (see Appendix A.1.4).

Bench testing (dynamic), not less than three months should include the following:

• The test should show how the variability in site conditions can affect the outcome of the test.
• The test should be long enough to demonstrate sulfate reduction and distinguish it from adsorptionNon-covalent bonding of a chemical to a solid surface..
• The test is a small-scale field or in-house test using site water.
• Test flow conditions (for instance, continuous or pulsed) likely to be encountered at the site.
• Use caution when transporting MIW as precipitation may occur, thus altering the chemistry of the water during transport.
• The test may use MIW surrogates in some cases when pretreatment is expected to produce a specific MIW condition but cannot be implemented, or when the site is inaccessible during the period of testing (for instance due to high elevation or winter weather).
• Test the media from the proof of principle test under flow through conditions.
• Test posttreatment methods after the BCR testing unit(s)
• Test the acid, metal, and hydraulic loading range expected in the field.
• Test acceptable flow through characteristics of the substrate.
• Identify any pretreatment that may be required to avoid substrate plugging.

The ideal result of the bench-scale testing is to determine the best combination of substrate mixture and loading rate deemed to yield acceptable water quality. The results should include an estimate of the sulfate reduction rate per volume of substrate based on observed sulfate removal. The observed metal removal efficiency (MRE) should also be determined. It is important that the rate of sulfate reduction is evaluated for an extended time to determine the long-term rate as opposed to the initial rate. Eger and Wagner (2001) noted in a barrel-scale study that the observed rate of sulfate reduction was initially as high as 300 mmol/m³/d but declined to 20 mmol/m³/d, and that rates were highly variable in part due to seasonal operation.

To determine preliminary Khydraulic conductivity (hydraulic conductivity) values of the substrate, a falling head permeability test may be conducted.

In order to minimize earth work, pilot-scale tests frequently are bermed instead of completely in-ground and may require pumps to feed MIW to the system. Because the range of sites at which MIW is found, this section does not specify design and operation. Instead, see descriptions of pilot-scale systems in the literature (Thomas and Romanek 2002; Reisinger, R.W. and J.J. Gusek 1998; Gusek et al. 2000; Reisman et al. 2008; Gusek et al. 2008; Faulkner et al. 2007; Gusek and Schueck 2004).

The duration of the pilot-scale test in the field should be sufficient to demonstrate treatment effect during any anticipated variations in MIW character. A duration should be set that encompass the seasonal and other variability as described in Section 2.1 to the extent practical. In many cases, at least a year is required to see the full range of variability in MIW at a site and also to monitor performance at the coldest and hottest conditions.

Data collection during the pilot test should include the following:

• Initial substrate bulk density and volatile suspended solids. In addition the substrate could be tested for cation exchange capacity (CEC), reflecting adsorption capacity, and cellulosic content by sequential extraction as described in Hall (2000).
• Water quality parameters of interest for the influent and the effluent. For many sites these parameters include pH, dissolved oxygen (DO), Eh, conductivity, alkalinity, acidity, temperature, ammonia, nitrate/nitrite, sulfate, sulfide, and concentration of metals of concern.
• In situ Eh within the reactor column, DOdissolved oxygen, and pH
• Ending vertical profile of Eh and pH, reflecting degree of exhaustion of alkalinity and organic, substrate bulk density and volatile suspended solids. If initial substrate was tested for CECcationic exchange capacity and cellulose, also test final content.

If the field pilot study does not meet the necessary discharge standards, another treatment technology should be considered or additional pre- or posttreatment added. In most cases pretreatment at pilot-scale will be part of the study along with posttreatment methods to oxidize the BCR effluent. It is also important to determine the sludge characteristics, assuming sludge is produced, during this phase. Will the sludge be hazardous or nonhazardous? Can the treatment sludge be disposed of on the mine site? Sludge management and organic substrate replacement may comprise the principle operating costs of a BCR treatment system.

At the conclusion of field pilot testing three equally important aspects affect full-scale passive treatment system design: biology, MIW chemistry, and precipitate retention. The bench and pilot test results should have yielded the conditions necessary to establish the proper chemistry or dominant geo-ecosystem in a given treatment unit to develop stable chemical precipitates. However, constructing an ideal biochemical environment is a wasted effort if the metal precipitates formed are flushed out of the system because of inefficient retention. Among other factors, this aspect of a proper system design is influenced by the grain-size distribution and compacted density of organic substrates, the settling and flocculating characteristics of the precipitates, and the retention times of the settling units.