Lecture No. 24. Special Topics II.

Acid Sulfate Soils (Continued)

In the last lecture we discussed three approaches to remediating acid sulfate soils. Here is a fourth.

4. Special Plantings

Some workers have proposed planting more acid-tolerant species. In fact, trees are much more acid-tolerant than grass; whereas grass prefers alkaline or neutral soils, trees grow better in acid soils. Left to themselves, trees tend to lower the soil pH to about 4.5. Torbert et al. (1988) noted that in Virginia 70-90% of the reclaimed coal mine land would eventually be covered with forest. In spite of this, reclamation guidelines emphasize the establishment of herbaceous ground cover. In establishing such a cover — primarily grass and clover — workers may be delaying reforestation by decades. But coal mining companies do not get their money back under Public Law 95-87 until the reclaimed land has supported herbaceous cover for 5 years, so they have every incentive to make the soil alkaline and plant grass and clover.

Other workers have noted the functions of mycorrhizal associations (associations between roots and certain symbiotic fungi) in protecting plants from excessive metal uptake, thus in a sense making the plants more acid-tolerant. Norland (1993) pointed out that the best soil pH for mycorrhizae is between 4.5 and 5.5. Lawrey (1977) noted the paucity of fungal genera and the poor nutrient cycling in mine soils. He also recalled earlier observations that tree species that harbored mycorrhizal fungi and shrubs that harbored nitrogen-fixing fungi such as Frankia were most successful in colonizing anthracite "black waste" areas. In other words, plants that harbor their own symbiotic soil microorganisms may make good candidates for reclamation. Interestingly, the mycorrhizal fungus Pisolithus tinctorius is able to coexist with Thiobacillus ferrooxidans and still protect its symbiont, Virginia Pine.

Good soil conditions for mycorrhizae are pH between 4.5 and 5.5; a sufficient supply of water but not waterlogged conditions; good soil aeration; and low available phosphorus coupled with a supply of insoluble phosphorus that the mycorrhizae can deliver to the roots.

Another type of special planting is the use of green manure, i.e., crops grown for their nitrogen-fixing properties and for their value as soil amendments when they are plowed into the soil. Clover, beans, various other legumes, and summer squash have been used as green manure.

Finally, in areas where soil acidity is complicated by high heavy-metal concentrations, it may be possible to cultivate hyperaccumulator plants. Alyssum and Thlaspi species are very good hyperaccumulators of nickel and zinc. Some Thlaspi species from areas of zinc-lead mineralization hyperaccumulate zinc, cadmium, and lead. A manganese hyperaccumulator was found growing at an old copper mining site in the Grand Canyon.

5. More Exotic Approaches

Several authors have advocated unusual approaches to remediating acid sulfate soils. These include

• Electrokinetic soil cleaning. A DC voltage is imposed on the soil by sinking an array of cathodes and anodes connected to a voltage supply. If conditions are right, metal ions gradually migrate to the cathodes. The possible loss of beneficial metals is balanced against the need to remove contaminant metals.
• Addition of zeolites to the soil. Zeolites have high ion exchange capacities and the ability to sequester metals like lead.
• Soil washing. Contaminant metals are removed ex situ by removing soil, washing it with solutions of chelating agents, and then replacing the soil. Needless to say, it is expensive and is damaging to the soil. It may be especially useful, however, in "brownfields," i.e., old industrial sites that need to be decontaminated before new construction can take place. If the value of the land is high and the soil productivity is not important (because it will be paved over anyway), soil washing may be appropriate. Soil washing, using surfactants or other treatment agents instead of chelating agents, can be useful in removing organic contaminants from soil.

Conclusion

We have several arrows in our quiver: liming, organic treatment, soil isolation, special plantings, and other techniques. A given situation may call for one of these approaches or for a mixture of approaches. As usual, prevention is cheaper than cure, and mining practices designed to prevent the development of acid soils usually will be cheaper and more effective than remediation.

Volcanigenic Massive Sulfide Deposits

These are sulfide ore deposits that form in deep-sea environments. They are found in and around the following environments:

1. Mid-ocean ridges (spreading centers) and associated with mid-ocean ridge basalts (MORB);
2. Subduction zones (convergent areas);
3. Probably in back-arc basins (spreading centers);
4. Forearc areas of subduction zones;
5. In tensional zones undersea, such as areas of continental rifting;
6. Maybe around intra-plate hot spots such as the Hawaiian hot spot. (This is conjectural; none have been found so far in these environments.

If we are going to mine them, they must be located on or near dry land. Thus deposits that form under the sea must eventually be uplifted.

For a viable massive sulfide deposit, boiling cannot occur; therefore they form at depth. If the temperature is 225¡ C, depth must be > 180 m. If T = 300¡ C, depth > 900 m. The familiar "black smokers" are squirting out under pressure, but they are not boiling. The "smoke" is micron-sized amorphous FeS.

Examples of the types of volcanigenic massive sulfides:

1. Cyprus type (ophiolite complexes)
2. Kuroko deposits ("Black Ore — Japanese)
3. None known
4. Besshi-type (rare)
5. SEDEX (SEDimentary EXhalative) deposits such as the Sullivan deposit in British Columbia, hosted in shale in the Prichard Formation of the lower Belt Supergroup.
6. None known

The Sullivan deposit was extremely rich. Some drill cores ran to 28% combined lead and zinc. It was so rich the ore actually caught fire. Parts of the mine had to be walled off with steel doors so that the rest could be worked. The burning sulfide ore self-smelted, and molten lead ran out under the doors in some spots.

Here is an example of a type of massive sulfide: the Kuroko type.

Kuroko Deposits

These formed under seawater, but as a result of felsic pyroclastic domes. In effect, these are submarine calderas. In Japan, these formed during the Miocene. The ore deposits formed in layers. Listed from top to bottom, these layers were

Copper increases downward in these deposits.

At time zero, a Kuroko deposit starts as a rhyolite volcanic dome with a water circulation cell driven by magmatic heat. Barium in the fluids has reacted with seawater sulfate to form a layer of barite.

The barite insulates the sulfides that are forming in the dome. Copper sulfides grow at the expense of the other sulfides, such as sphalerite and galena.

If the dome blows again, we may find big deposits of fragmented ore nearby. Another cycle of ore formation may happen.

Massive sulfide deposits that form in or on ocean crust end up on dry land through obduction. Those that form on submerged continental crust can be raised later by thrust faulting and other common processes of compressional tectonics.

In any case, seawater seeps into oceanic or continental crust and is heated by magma. Massive sulfide deposits form as a result of some aqueous chemistry that can only happen under high pressures, since the water would boil off under surface pressures. For example, iron (II) reduces sulfate at temperatures above 150¡ C:

Fe²⁺ + SO₄^2– ¨ Fe₃O₄ + H₂S + some FeS₂

The source of the iron (II) is silicates such as chlorite or serpentine. Magnetite forms, some pyrite forms, and hydrogen sulfide is liberated.

With depth, temperatures increase, salinity of water increases, and we have a hot brine. This brine can move metals like copper around as chloride complexes.

At these elevated temperatures, (commonly 350¡ C and up) iron (II) can also reduce water (liberating hydrogen), and carbon dioxide (liberating methane).

When this water reaches a submarine vent, it is very salty, strongly reduced, slightly acidic, and replete with dissolved sulfide and with chloride complexes of metals. It cools suddenly, but it is still under pressure. Sulfides form spontaneously.

What all massive sulfide deposits have in common is extremely high concentrations of sulfides, including pyrite. Most ore deposits of other types consist of altered country rock with veins rich in sulfide, or with disseminated sulfide throughout the rock. Ore grades of a few percent are considered rich. But massive sulfide deposits have layers that are pure or nearly pure sulfide minerals, while other layers are rock with extremely rich disseminated sulfides. Grades of several tens of a percent can occur in some layers.

One might expect that exposure of massive sulfide deposits — even the waste rock from these deposits — might be the source of fairly concentrated acid formation. That happens to be the case. An example of this is the massive sulfide deposit at Iron Mountain, Shasta County, California.

The Iron Mountain Deposit

This deposit is just west of the Sacramento River, near Redding, about 50 miles south of Mount Shasta. It is the site of several underground mines, now abandoned. The water draining these mines is characterized by metal concentrations in the grams per liter, very high sulfate, and negative pH values. The Richmond Mine, Hornet Mine, and other mines at Iron Mountain were studied extensively by D. Kirk Nordstrom of the USGS. The pH of the drainage water ranged as low as –3.4! (I should point out that water at pH –3.4 does not really contain 10^3.4 M (or 2500 M) sulfuric acid. The sulfuric acid concentration is merely high enough so that the water volume is significantly reduced — 1 L of solution contains much less than 1 L of water — and so the activity coefficients of ionic species become greater than 1.)

In past discussions of pyrite oxidation chemistry I have tended to write the reactions involved with all products being ionized except iron and aluminum oxyhydroxides. I have mentioned that the second ionization of sulfuric acid is suppressed below pH 2, and that Fe³⁺ becomes soluble below about pH 2.5. But the picture under extremely acidic conditions is much different and somewhat bizarre.

As pH drops, and under suitably oxidizing conditions, where total dissolved iron is at 0.1 M and total sulfur is 0.2 M, goethite (a -FeOOH) is the most stable solid iron (III) compound down to about pH 2. Jarosite then becomes more stable until about pH 1.5, when iron sulfate complexes become very soluble. FeSO₄⁺ predominates to about pH 0.5, and then FeHSO₄²⁺ takes over. The exact pH boundaries are variable, and if the jarosite or goethite is poorly crystallized they may differ over two or more pH units.

In De Re Metallica, the first systematic textbook on mining technology, Agricola commented on various poisonous "efflorescences" that were known to form on the walls of sulfide-ore mines. Many of these were sulfate minerals such as jarosite and alunite, with or without acid inclusions. Under more acidic conditions, the K⁺ or Na⁺ in these minerals may be replaced by NH₄⁺ or H₃O⁺. Under extremely acidic conditions such as those in weathered massive sulfide deposits, alunogen (Al₂(SO₄)₃ á 17 H₂O) becomes the most stable aluminum mineral (below pH 0) and increasingly acidic iron (III) sulfate minerals form, such as roemerite and rhomboclase; rhomboclase is essentially a crystalline form of sulfuric acid with some ferric sulfate.

The water draining from the mines in Iron Mountain mixes (fortunately) with cleaner ground and surface water, but the result is still a very acidic creek. This very acidic, very metal-rich water eventually feeds into the Sacramento River just upstream of the drinking-water intake for the City of Redding. Fortunately the discharge of the Sacramento River is fairly large at this point, and the acid water, which enters the river on the west bank, does not mix sufficiently with the river to affect the water being drawn out on the east bank.

The Iron Mountain deposit is proving to be very difficult to deal with. The mine tunnels and shafts are in rock that is quite permeable, and the supply of groundwater is much more plentiful than it would be in a desert environment. To stop the formation of the concentrated acid solutions it would be necessary to keep water out of the mines, or to slow down the movement of water through the mines. Since the acid solutions are corrosive to dissolve concrete, granite, and most other materials, stopping the flow is a problem. The best solution might be to mine the water as a source of copper, of zinc, and possibly of sulfuric acid. The value of the extracted materials might pay for the ongoing work of protecting the river from the mine drainage.

Prediction of Acid Drainage

There are a few methods used to model the behavior of ore bodies and the surrounding host rock and to predict the acid drainage that would be produced once the ore is mined.

B. Acid-Base Accounting

This approach looks at the total potential acid production from a mine site, including the host rock, the waste rock, tailings, etc. The basic assumption is that geochemical reactions are the major factors controlling the acid-forming process in mine waste materials. The processes are considered mainly to be (a) oxidation of sulfide minerals, and (b) dissolution of alkali carbonates, dissolution of exchangeable bases, and weathering of silicates.

The components of the acid-base account are

• the acid potential,
• the neutralization potential, and
• the saturated paste pH.

The acid potential is obtained from the total sulfur or the pyritic sulfur. The latter is better, since most metal sulfides, sulfates such as gypsum, and organic sulfur compounds are not acid generators. All acid is assumed to be produced by the reaction

2 FeS₂ + 15/2 O₂ + 7 H₂O ¨ 2 Fe(OH)₃ + 4 H₂SO₄

Acid-generating potential is expressed as kilograms of sulfuric acid per (metric) ton of rock. By the above equation, one mol of pyritic sulfur produces one mol of sulfuric acid, so 32.066 g sulfur produces 98.08 g sulfuric acid, or 1 g sulfur produces 3.059 g acid. Switching units to kilograms,

(kg acid) = 3.059 (kg S)

(kg acid/ton) = 3.059 (kg S/ton)

% S = kg S/100 kg rock

Therefore

Kilograms per ton = kilograms per 1000 kg rock = 10« percent S

and so

kg acid/ton = 3.059 « 10 « (% S)

kg acid/ton = 30.6 (% S)

Acid neutralization capacity (ANC) is a measure of the ability of the rock to neutralize acid from sulfide oxidation. It normally is measured by heating samples with standardized sulfuric or hydrochloric acid, then back-titrating to determine acid remaining. Units are either kg calcium carbonate per ton or kg sulfuric acid neutralized per ton.

Net acid producing potential (NAPP) is the difference between these two measures:

NAPP = 30.6 « %S – ANC (kg H₂SO₄/T)

If NAPP is positive, the rock is expected to produce acid drainage. If it is negative, the rock has excess acid neutralizing potential.

Finally, a saturated paste of the ground rock is made and pH is measured. If the pH of this paste is less than 4, the rock is considered acidic regardless of the acid-base accounting.

Let's look at this method and see what holes we can kick in it.

1. It assumes that pyrite is the only source of acid. This is true sometimes, but there are other sources. For example, there may be iron (II) minerals present other than pyrite, and iron (II) oxidation produced 2 mols of acid per mol iron (II):

2 Fe²⁺ + 1/2 O₂ + 5 H₂O ¨ 2 Fe(OH)₃ + 4 H⁺

What sources of iron (II) might be present?

(a) If the rock is somewhat oxidized, there may be melanterite, FeSO₄ á 7 H₂O. Each mol of melanterite produces 2 mols of acid.

(b) If vivianite (Fe₃(PO₄)₂á8 H₂O)is present, each mol of vivianite will produce 3 mols of acid at pH 7:

2 Fe₃(PO₄)₂ á 8 H₂O + 3/2 O₂ ¨ 6 Fe (OH)₃ + 2 HPO₄^2– + 2 H₂PO₄^– + 6 H⁺ +H₂O

(The phosphate takes up half of the acid produced by iron oxidation.)

(c) If chalcopyrite is present, the iron (II) will produce 2 mols of acid. (The sulfur, being in the –2 state, produces sulfate but no acid.)

2 CuFeS₂ + 17/2 O₂ + 5 H₂O ¨ 4 SO₄^2– + 2 Cu²⁺ + 2 Fe(OH)₃ + 4 H⁺

(d) If arsenopyrite is present, oxidation of iron (II) produces a strong acid, and the weak arsenous acid is also produced.

FeAsS + 3 O₂ + 4 H₂O ¨ Fe(OH)₃ + SO₄^2– + H₃AsO₃ + 2 H⁺

Other minerals containing iron (II), if present and if redox-active, can also produce acid.

2. It assumes 1 mol carbonate will neutralize 2 mols acid, no matter what kind of carbonate it is.

(a) Siderite has no net acid neutralizing capacity, since the carbonate is just adequate to neutralize the acid produced in oxidizing the iron.

2 FeCO₃ + 1/2 O₂ + 3 H₂O ¨ 2 Fe(OH)₃ + 2 CO_2Ð

(b) Ankerite (CaFe(CO₃)₂) will have a reduced acid neutralizing capacity because of the iron present.

(c) Dolomite reacts slowly with acids. Depending on the pH and the contact time, it may not react fast enough to neutralize acid completely if groundwater flow is fairly rapid.

(d) The manganese-containing carbonates (rhodochrosite, MnCO₃, and kutnahorite, CaMn(CO₃)₂), may produce acid at high E_h by oxidation of manganese, and they also may, like dolomite, react too slowly to neutralize acid.

MnCO₃ + 1/2 O₂ ¨ MnO₂ + CO_2Ð

CaMn(CO₃)₂ + 1/2 O₂ ¨ MnO₂ + CO₂ + Ca²⁺ + CO₃^2–

3. It assumes that carbonate is the only acid-neutralizing or acid-consuming species. Yet we know of others.

(a) Phosphate neutralizes acids. (See vivianite.)

(b) Minerals containing copper (I), such as chalcocite (Cu₂S), consume acid in the oxidation of Cu (I) to Cu (II).

(c) If reaction times are reasonably long, feldspars and micas react with acids to form clay minerals.

(d) If alkaline groundwater is flowing in from outside the mine, it may neutralize some of the acid produced by oxidation.

We see that the acid-base accounting method has its drawbacks. Still, if pyrite is the principal source of acid and limestone is the principal potential neutralizer, the method works reasonably well. And the experimental method used to test acid neutralization capacity may detect the positive and negative contributions of phosphates and the contributions of carbonates other than calcite. (It probably will not detect the contributions of silicates.)

In a paper presented at the 1991 Randol Gold Forum conference, Miller et al. made the following recommendations (Miller, S.D., J.J. Jeffery, and J.W.C. Wong, 1991, "Use and misuse of the acid-base account for 'AMD' prediction." Randol Gold Forum, 1991, 69-75.):

1. In planning a mine and assessing the net acid producing potential of its rocks, at least 150 drill cores must be taken.
2. Using the analytical data from the drill cores, percent total sulfur should be plotted against ANC. The lines for NAPP = +20 and NAPP = –20 should be drawn on this plot.
3. If NAPP for most of the plotted points is less than –20, the rock may be regarded as non-acid-forming.
4. If NAPP for most or all of the points is greater than +20, the rock is acid-forming.
5. If NAPP for most points is between +20 and –20, more detailed study is required.

The first recommendation probably gave some of the listeners heart attacks. In the United States, Federal and State agencies require as much as one drill hole per 8 acres, or as little as one hole per 160 acres.

The other recommendations are eminently sensible. What the authors are saying is that the acid-base accounting method is good at spotting highly acid-producing rock and rock that has no chance of producing acid drainage. In between these classes of rock is a class that is not so easy to assess.

The first overhead shows a plot of total % sulfur vs. acid neutralization capacity for some mining waste samples from various coal mines. The next one shows a similar plot for waste samples from various gold mines, and the third is a similar plot for some base metal mines. The fourth overhead is a plot of % S vs. ANC for waste rock sample from a single gold mine. The amount of scatter, with many samples being highly acid-forming and others being non-acid-forming, illustrates why Miller, Jeffery, and Wong recommended taking such a large number of samples.

Now we will look at a more complex method of assessment.

B. Geochemical Modeling

These software models rely on measured or predicted groundwater movement, mass transfer rates, and the thermodynamic (and sometimes kinetic) properties of minerals and dissolved species. There are a number of "forward models" that simulate the reactions of multiple minerals and other chemical species: PHREEQE, MINTEQA2, WATEQF, and WATEQF4 are four of them. There are also "inverse models" that calculate net geochemical mass transfer once the initial and final water properties are known (or predicted). One model of this type is BALANCE. One or more forward models are commonly used with an inverse model.

Inverse models are not constrained by thermodynamics, so they may predict impossible reactions. The forward models also have their limitations. PHREEQE, for example, cannot predict removal of ions from solution by precipitation unless reversible reactions are specified.

Models in general have no common sense. They can do a lot of calculations in a short time, but scientists trained in geochemistry and hydrogeology have to ride herd on them and judge each step to see whether it makes sense. A good paper on this subject is Bird, David A., W. Berry Lyons, and Glenn C. Miller, 1994, "An assessment of hydrogeochemical computer codes applied to modeling post-mining pit water geochemistry," Tailings and Mine Waste '94, Proceedings of the First International Conference on Tailings and Mine Waste (Balkema, Rotterdam), 31-40.

To previous lecture: Lecture No 23. Special Topics I

To next lecture: Lecture No 25. Special Topics III

Back to list of lecture notes