Environ. Sci. Technol. 2006, 40, 6438-6443
Use of Caustic Magnesia To Remove Cadmium, Nickel, and Cobalt from Water in Passive Treatment Systems: Column Experiments TOBIAS STEFAN RO ¨ TTING,* JORDI CAMA, AND CARLOS AYORA Institute of Earth Sciences Jaume Almera, CSIC, Lluis Sole´ i Sabarı´s s/n, 08028 Barcelona, Spain JOSE-LUIS CORTINA AND JOAN DE PABLO Department of Chemical Engineering, Technical University of Catalonia (UPC), Diagonal 647, 08028 Barcelona, Spain
In the present study caustic magnesia obtained from calcination of magnesium carbonate was tested in column experiments as an alternative material for passive remediation systems to remove divalent metals. Caustic magnesia reacts with water to form magnesium hydroxide, which dissolves increasing the pH to values higher than 8.5. At these pH values, cadmium is precipitated as otavite and to a minor amount as a hydroxide. Cobalt and nickel are precipitated as hydroxides which form isostructural solids with brucite. Thus, metal concentrations as high as 75 mg/L in the inflowing water are depleted to values below 10 µg/L. Magnesia dissolution is sufficiently fast to treat flows as high as 0.5 m3/m2‚day. For reactive grain size of 2-4 mm, the column efficiency ends due to coating of the grains by precipitates, especially when iron and aluminum are present in the solution.
Introduction Metal pollution from acid mine drainage (AMD) and industrial activities poses a major threat to ecosystems. Of particular concern are abandoned mining and smelting sites where sustained pollutant releases can persist for decades following the cessation of industrial activity. Thus, in many industrial and mining areas with presence of poly metal-sulfides (Fe, Cu, Cd, Zn, Ni, Co), nonadmissible levels of these metals are found in surface and groundwater. Particularly, very low levels of Ni and Cd are required due to their properties as active carcinogenic agents with a maximum admissible concentration (MAC) of 20 µg/L for Ni and 5 µg/L for Cd (1, 2). The toxicity of Co is usually minor but can be high in the case of radioactive wastes in conjunction with Ni. The conventional solution to the metal pollution in water consists of treatment plants where chemicals are added to the water in controlled doses. In the case of groundwater, prior pumping is necessary. This is expensive and requires continuous maintenance. An alternative is passive treatment whereby filters are constructed in the path of acidic drainage in the form of reducing and alkalinity producing systems * Corresponding author phone: +34 934 095 410; fax: +34 934 110 012; e-mail:
[email protected]. 6438
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 20, 2006
(RAPS) and permeable reactive barriers (PRB) (3, 4). These filters are simple biogeochemically based systems which improve the water quality by means of reactions such as calcite dissolution, sulfate reduction, etc. These methods make use of gravitational gradients and require very low maintenance. The materials needed for such systems should be as follows: (1) reactive so as to form an insoluble solid to retain metals; (2) permeable in order to allow the water to percolate, (3) durable, and (4) available and inexpensive so as to be competitive. In passive systems to treat AMD, such as RAPS and PRB, organic matter with calcite is used as filling material (5, 6). Due to two main reasons they only remove part of the metal load from groundwater. First, sulfate reduction is a slow process, and the water needs a long residence time within the barrier for the reduction to take place (7). This would require uneconomically thick barriers for fluxes higher than 0.2 m3/m2‚day commonly found in alluvial aquifers. Second, due to the high Ca concentration in AMD, equilibrium with calcite is reached at pH between 6 and 7. As shown in Figure 1, this pH is not high enough for divalent metal hydroxides and particularly Cd, Ni, and Co to precipitate. High pH values can also be achieved with lime dissolution. However, the solubility of metal hydroxides increases again for very high pH values (Figure 1 (8, 9)). The addition of lime in excess, such as it would occur if it were used in a passive treatment, may raise pH up to 12. These high pH values make the treatment inefficient and release harmful solutions to the environment. Magnesium oxide dissolution, in contrast, can buffer solutions between pH 8.5 and 10.5 and has been recently used to remove Ni and Co from sulfate leach liquors (10). Caustic magnesia is an inexpensive product where MgO is the main component. Large amounts of caustic magnesia are produced from calcination of magnesium carbonate, and the applications of caustic magnesia in mineral processing, hydrometallurgy, and environmental problems are increasing in the last years (11, 12). In an earlier paper, the suitability of caustic magnesia to retain Zn, Cu, Pb, and Mn in permeable reactive barriers was demonstrated (13). The present paper is complementary and investigates the removal of Cd, Ni, and Co in passive remediation systems by means of column experiments filled with caustic magnesia. First, monocomponent solutions of Cd, Ni, and Co were used as inflow waters to understand the retention processes. Then, since AMD usually contains aluminum and iron in variable concentrations, the behavior of Cd, Ni, and Co in sulfate and acidic solutions containing Al, Fe(II), and Fe(III) was used as a more realistic proxy.
Materials and Methods The commercial product Magna L manufactured by Magnesitas Navarras S.A., Spain, was selected in accordance with the grain size and permeability criteria. It is mainly made up of MgO (76 ( 2%), with minor amounts of CaO (11 ( 2%), SiO2 (5 ( 1%), MgCO3 (0.9 ( 0.3%), and CaCO3 (0.6 ( 0.2%) (13). For the experiments, caustic magnesia was sieved to grain sizes ranging from 4 to 2 mm. Column experiments were designed to measure the reactivity of caustic magnesia under flow conditions similar to those expected in alluvial aquifers. Solutions were administered continually through Teflon tubing connected to a Gilson Minipuls peristaltic pump (inlet fluxes shown in Table 1). Tracer tests were conducted using acetone and a Hewlett-Packard HP53 spectrophotometer with flow-through 10.1021/es061092g CCC: $33.50
2006 American Chemical Society Published on Web 09/09/2006
FIGURE 1. Variation of metal concentration in equilibrium with the hydroxides. Thermodynamic data from the Wateq4f database (8), except for Co(OH)2 (9). The vertical rectangular fields indicate the pH values expected after passive treatment of acid water with calcite and MgO.
TABLE 1. Summary of the Physical and Chemical Characteristics of the Different Column Experiments experiment length section solid pore volume porosity flow Darcian flux residence time pH Cd(II) Ni(II) Co(II) Al(III) Fe(II) SO42-
mm mm2 g mL mL/min m3/m2year min mg/L mg/L mg/L mg/L mg/L mg/L
1.1 50 314 22 6.1 0.39 0.06 100 100 2.0 75
1.2 50 314 22 6.2 0.4 0.06 100 100 2.0
1.3 50 314 23 6.1 0.39 0.06 100 100 2.0
75 75 545
600
600
2 80 480 53 16.9 0.44 0.1 110 170 2.6 15 6 6 100 710
3 80 480 52 17.6 0.46 0.1 110 175 2.5 15 6 8 100 240
cuvette. Measured pore volumes and porosities for the different columns are given in Table 1. The inflowing solutions were prepared by dissolving the metal sulfates in deionized H2O. In order to obtain a sufficient solid amount for mineralogical determination after a reasonable time span, the Cd, Ni, and Co concentrations of the inlet solution ranged between 10 and 75 mg/L, depending on the experiment, more than 1 order of magnitude higher than those commonly reported for AMD (4, 14). All pH measurements were made routinely in a sealed flow-through cell placed at the exit of the column. Column effluent samples were filtered through 0.45 µm and acidified with HNO3 to pH lower than 1. Flow rates were determined gravimetrically. Most elements were analyzed with ICP-AES. Analytical detection limits were (in mg/L) as follows: 0.25 for Mg, 0.20 for Ca, 0.03 for Fe, and 0.02 for Al. Analyses of Cd, Ni, and Co were performed with ICP-MS, and the detection limit was 0.5 µg/L. After the end of the experiments, the substrate of the columns was observed under a JEOL 3400 scanning electron microscopy with an energy dispersive system (SEM-EDS). Moreover, the mineral phases present in the column fillings were identified with a BRUKER D5005 X-ray diffractometer (XRD), with Cu LR radiation.
Results Metal Retention from Monocomponent Solutions. Three solutions at pH 2 containing 75 mg/L of Cd, Ni, and Co,
respectively, were used as inflow solutions for three columns filled with caustic magnesia. Other conditions are detailed in Table 1, experiments 1.1, 1.2, and 1.3. The experiments lasted for 11, 18, and 12 months for Cd, Ni, and Co, respectively. The evolution of Cd, Ni, and Co concentration and pH of the outflowing solution, normalized to pore volumes, are represented in Figure 2. In the three experiments, the initial pH values around 12 correlated with high Ca concentrations (not represented) and were attributed to the dissolution of a minor amount of lime present in the caustic magnesia sample. After lime exhaustion, the Ca concentration dropped to the inflowing limit, the Mg concentration increased to values varying between 40 and 110 mg/L, and the pH remained around 9. This is consistent with the dissolution of brucite, Mg(OH)2, a surface layer which results from the hydration of magnesia grains (13). Magnesium concentration, however, remains far from the values expected for brucite saturation (4500 mg/L), indicating that the equilibrium with this mineral was not achieved. In the case of Cd, for pH higher than 8.5, concentrations between 5 and 20 µg/L were measured in the outlet solution, close to regulatory limits. Some occasional decrease in pH is also correlated with higher Cd concentrations. After 2000 pore volumes, pH decreased to values lower than 8. Mass balance considerations show that only 6 wt % of the magnesia was consumed at that point. The pH decrease is accompanied by a significant increase in the Cd content in the outlet solution. In some samples, Cd concentration in the outlet exceeds the inflowing value (75 mg/L) due to the dissolution of the Cd precipitates previously formed in the pores of the column. Ni showed similar behavior. For pH higher than 8, the concentration measured in the outlet was always below the detection limit (0.5 µg/L). The pH of the outlet dropped after 6000 pore volumes and after 18% of the initial magnesia was consumed. Then, Ni concentration started to increase. A similar pattern is also shown by Co, which concentration in the outlet remained below 100 µg/L for pH values higher than 8. When pH decreased to values close to 7, Co concentration also increased exceeding the inlet value. The extrapolation of these experiments (2000 pore volumes) to a field scale treatment suggests that, from the reactivity point of view, a 1 m thick system could be operational during several years for a water flux as high as 0.5 m3/m2‚day. Since the pore solution did not reach equilibrium with brucite, higher pH and more efficient precipitation of metals is expected for longer residence times, as will be shown below. At the end of experiment 1.1., several powder diffractograms of the precipitates collected in different points of the column clearly confirmed the presence of otavite (CdCO3) as a major Cd phase. In one case, several peaks of the Cd(OH)2 solid phase were also identified. SEM-EDS observations of the precipitates from the same experiment 1.1 showed the ubiquitous presence of Cd-O-C-bearing rhombohedra, which were interpreted as otavite (Figure 3A). Otavite has been interpreted to control the Cd concentration in aquifers and mine pit lakes (14, 15). It has also been observed to precipitate in columns of apatite to remediate heavy metal polluted water (16). For experiments 1.2 and 1.3, SEM-EDS observations reveal Co-Mg-O (Figure 3B) and Ni-Mg-O (not shown) bearing platy crystals formed throughout the brucite matrix substrate. Powder diffractograms of the solid residue confirmed the presence of theophrastite (Ni(OH)2) and β-Co(OH)2, respectively. These two phases have a crystalline structure very similar to brucite, which is the major component of the column, and are only distinguishable by double peaks at some diffraction angles. VOL. 40, NO. 20, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6439
FIGURE 2. Evolution of pH (right-hand scale), Cd (A, experiment 1.1), Ni (B, experiment 1.2), and Co (C, experiment 1.3) concentration (left-hand scale) of the outlet solution in column experiments with monocomponent inlet solutions.
FIGURE 3. (A) Rhombohedral crystals of otavite, CdCO3, precipitated from a cadmium-sulfate acid water (experiment 1.1). (B) Rosy aggregate of Co-Mg-O-bearing platy crystals formed on the surface of brucite grains from a cobalt-sulfate acid water (experiment 1.3). The geochemical behavior of the experiments with monocomponent solutions is consistent with the solubility data available for the main solids found ((8, 9, 17, 18) see Supporting Information for speciation details). Thus, for the conditions of the experiment, cadmium is stable in solution for pH below 6.5. For the range of pH controlled by caustic magnesia dissolution, the more stable phase is otavite, in agreement with observations. The range of otavite stability increases with the amount of CO2 dissolved. Finally, Cd(OH)2 is only stable for high pH values, and its occasional presence in the column could be due to its early precipitation during lime dissolution. Nickel and cobalt speciation with pH is simpler. Ni(OH)2 and Co(OH)2 precipitate for pH values higher than 8, in complete agreement with observations. Metal Retention from an Al(III)-Bearing Acid Solution. The behavior of Cd, Ni, and Co in a SO4-rich solution of pH 2.6 containing Al (100 mg/L), Cd (15 mg/L), Ni (6 mg/L), and Co (6 mg/L) was investigated. The rest of the experimental conditions are detailed in Table 1, experiment 2. The experiment lasted for 10 months, and the evolution of the Al, Cd, Ni, and Co concentration and of the pH in the outlet solution is represented in Figure 4A. The pH started with values close to 10.7, due to lime dissolution, and decreased soon to values close to 9.5, indicating the exhaustion of lime and the brucite control. As 6440
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 20, 2006
expected from the longer residence time (170 min), brucite dissolution reached equilibrium, as indicated by the pH values close to 10 and the Mg concentrations varying between 140 and 200 mg/L (Figure 4B), which remained stable for 85 days (700 pore volumes). Then, the pH dropped to 4 and remained constant for more than 210 days (1800 pore volumes). This is attributed to a sudden decrease in the dissolution of brucite, as also pointed out by a decrease in the Mg concentration of the outlet solution to below 50 mg/L (Figure 4B). As in the case of monocomponent columns, Cd, Ni, and Co concentrations in the outlet are closely correlated with pH. Thus, for pH exceeding 8.5, Cd concentrations between 10 and 20 µg/L were measured in the outlet solution, and a clear increase up to the input value of 15 mg/L was observed when the outlet solution reached pH values around 4 (Figure 4A). Likely, the outlet solution showed Ni concentrations below the detection limit at pH higher than 8.5, whereas for pH values below 4.0 they increased up to the input value of 6 mg/L. Finally, for the case of Co, outlet concentrations between 100 and 200 µg/L were measured at high pH, with an increase up to the input value of 6 mg/L when the outlet solution reached acidic pH values. Aluminum also showed an evolution similar to the rest of metals. At high pH, Al concentrations ranged between 50 to less than 20 µg/L
FIGURE 4. Evolution of (A) Cd, Co, Ni, Al and (B) Mg concentrations of the outlet solution in column experiment 2. Concentration of the inlet (mg/L): Al ) 100, Cd ) 15, Ni ) 6, and Co ) 6; pH ) 2.6. The evolution of pH is plotted in the right-hand scale in both figures.
FIGURE 5. (A) Aggregate of platy crystals of (Mg,Co,Ni)(OH)2 growing on microcrystalline aggregate of an unidentified aluminum oxihydroxide (experiment 2). (B) Roselike aggregate of goethite platy crystals (experiment 3).
FIGURE 6. Variation with pH of (A) Al and (B) Fe measured in the outlet solution of experiments 2 and 3, respectively. Metal concentrations in equilibrium with selected solid phases are also plotted with respect to pH. Thermodynamic data from the Wateq4f database (8), except for bayerite (19). (detection limit), and when pH dropped they increased to values close to or slightly higher than the input concentration (100 mg/L). After the experiment, the column was dismantled, and part of the solid was analyzed. SEM-EDS observations showed a crust with a noncrystalline appearance, covering the brucite surface, and mainly formed by Al and O (Figure 5A). Cd-, Ni-, and Co-bearing solids were always found as microplates covering the noncrystalline Al-rich crust. In addition to major constituents of the column matrix (periclase, brucite, quartz, and calcite), several new precipitated phases were identified by XRD: bayerite [Al(OH)3], Mg2Al(OH)7, (Mg0.667Al0.333)(OH)2 (CO3)0.167(H2O)0.5, theophrastite [Ni(OH)2] and β-Co(OH)2, and Co6Al2CO3(OH)16‚4H2O. No Cd phases have been identified, probably due to the small amount of precipitate in the samples analyzed. Unlike the columns of monocomponent solutions, the life of the aluminum column as an active system to retain metals is between six and ten times shorter. This is attributed to the coating of the magnesia grains which prevents further
dissolution. It has to be noted that the subsequent drop in pH and metal release is very abrupt, and no signal of when passivation is going to occur can be anticipated. Mass balance calculations point out that only 3 wt % caustic magnesia was dissolved prior to coating. The geochemical performance of Cd, Ni, and Co in the column is comparable to that of the monocomponent solutions described above: similar variation of metal concentration with pH and precipitated phases. However, as brucite dissolves and pH increases, Al precipitates in higher amount than the divalent metals. Despite the cryptocrystalline appearance of Al-precipitates, the low concentration of Al dissolved at high pH is not consistent with equilibrium with amorphous Al(OH)3 but with a less soluble Al(OH)3 phase (gibbsite or bayerite (19) in Figure 6A). No thermodynamic data are available for the rest of Al-bearing solids identified by XRD. Metal Retention from a Fe(II)-Bearing Acid Solution. A SO4-rich solution of pH 2.5 containing Fe(II) (100 mg/L), Cd (15 mg/L), Ni (6 mg/L), and Co (8 mg/L) was used as an VOL. 40, NO. 20, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6441
FIGURE 7. Evolution of (A) Cd, Co, Ni, Fe and (B) Mg concentrations of the outlet solution in column experiment 3. Concentration of the inlet (mg/L): Fe(II) ) 100, Cd ) 15, Ni ) 6, and Co ) 8; pH ) 2.5. The evolution of pH is plotted in the right-hand scale in both figures. inflow water. The rest of the experimental conditions are detailed in Table 1, experiment 3. The experiment lasted for 10 months, and the evolution of the Fe, Cd, Ni, and Co concentration and of the pH in the outlet solution is represented in Figure 7A. As in previously described experiments, the pH started with values close to 11 due to lime dissolution and then decreased to values close to 9.5, indicating the exhaustion of lime and brucite control. Similarly to the aluminum case, brucite dissolved sufficiently fast to reach equilibrium (residence time of 175 min) and was able to maintain the pH stable for 130 days (1200 pore volumes). Then, the pH dropped to 5.5 for 30 days (250 pore volumes) due to the temporary decrease of pH to 2.0 in the input solution. As the input pH was restored to 2.5, it also rose up to 7.5 in the outlet and remained constant for the rest of the experiment (1000 pore volumes). As expected, Fe, Cd, Ni, and Co concentrations in the outlet are closely correlated with pH (Figure 7A). For pH exceeding 8.5 the outlet had metal concentrations below detection limits, with some occasional values as high as 15, 10, 90, and 240 µg/L, for Cd, Ni, Co, and Fe, respectively. For pH values near 7.0, the metal concentrations usually remained below detection limits, although higher values occurred more often and reached peaks as high as a few mg/L. XRD patterns of solid samples collected from the pores after the end of the experiment revealed peaks characteristic of Fe(III) minerals, such as goethite [R-FeOOH], lepidocrocite [γ-FeOOH], coalingite [Mg10Fe2CO3(OH)24‚2H2O], and an Fe(II)-Fe(III) carbonate hydroxide [Fe6(OH)12CO3]. SEM-EDS observations showed a layer of Fe-O platy crystals interpreted as goethite (Figure 5B). Also covering the brucite surface, platy crystals of Ni-O and Co-O were observed as well as more occasional otavite rhombohedra. The geochemical behavior of Cd, Ni, and Co in this column is similar to that described for the previous columns. The removal of Fe, however, is less efficient than Al. When pH increases in the column, Fe(II) oxidation to Fe(III) is enhanced (20), and precipitation of very insoluble Fe(III) phases takes place. The concentration of Fe in the outlet, however, is much higher than that expected from equilibrium with Fe(III) phases (amorphous Fe(OH)3 as the most soluble one is represented in Figure 6B). This is interpreted as the oxidation of Fe(II) is limited by the availability of oxygen inside the column, and part of the Fe is released as Fe(II) in the outlet. With respect to the Al column, however, the complete coating and passivation of the caustic magnesia takes longer for the same input and flow conditions, probably because the coating of Fe(III) precipitates appears more crystalline and less compact than in the case of Al (Figure 5B). Nevertheless, as for the Al column, only 5 wt % of caustic magnesia was dissolved prior to coating. The last period of the column shows some differences with respect to Al. The amount of Fe release is significantly lower and the pH higher than the input. This is interpreted as caustic magnesia is still 6442
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 20, 2006
dissolving despite the coating, as also indicated by the high concentration of Mg in the outlet (Figure 7B). These column experiments demonstrate that caustic magnesia can abate dissolved Cd, Ni, and Co from concentrations much higher than those commonly found in polluted waters down to a few µg/L. This fact complements its efficiency to precipitate Zn, Cu, Pb, and Mn, demonstrated in a previous work (13). As expected from the obtained pH, between 8.5 and 10, caustic magnesia also precipitates Al and Fe(III), which usually accompany divalent metals in acidic solutions. The efficiency in precipitating Fe(II) is limited, however, by the availability of oxygen. After exhaustion, Cd and Co (and possibly Ni) concentration in some samples exceeds the inflowing value (75 mg/L) due to the dissolution of the solids previously formed. Therefore, the remediation system must be dismantled once the first symptoms of exhaustion appear. Intensive monitoring is necessary because loss of reactivity is abrupt, and no signals preceding passivation are apparent. The reactivity of caustic magnesia is high enough for the fluxes expected in permeable aquifers. Thus, as shown in the case of Al and Fe columns, the pore solution is practically in equilibrium with brucite during the active life of the column, for fluxes higher than 0.5 m3/m2‚day. Despite the high metal concentrations treated, the durability of the reactive system extends for 2000-6000 pore volumes, which decreases to 700 pore volumes if a high concentration of Al is present in the solution. Therefore, our results improve those obtained by Komnitsas et al. (21) who evaluated limestone and red mud as reactive media for permeable reactive barriers for Cd, Ni, and Co, among others. There, column experiments achieved satisfactory removal conditions for only up to 30 pore volumes, but Fe concentration was much higher (1500 mg/L). The price of commercial caustic magnesia is about ten times that of limestone, but its excellent removal capacity for divalent metals may justify this cost, especially if Al and Fe are removed in a prior limestone treatment step. The extrapolation of the results to the durability of field systems (>1 m thick) suggests that the treatment may remain active during several years if no high amounts of Fe and Al are present. At field scale, however, additional effects may have to be taken in account and experimentally tested (e.g. preferential flow, clogging, compaction of the substrate by self-loading or overburden (22, 23)). For the fluxes, concentrations, and grain size (2-4 mm) used in the experiments, the system becomes inactive due to coating when most of magnesia has not reacted. Ongoing work is oriented to employ finer grain size MgO and disperse it in a coarse and porous inert matrix acting as a support.
Acknowledgments We are grateful to Magnesitas Navarras SA for supplying the samples and to M. Marsal (UPC), M. Caban ˜ as, and J. Elvira (CSIC) for technical assistance. This work was funded by the
Spanish Government projects REN-2003-09590-C04 and CGL2005-08019-C04-03HID.
Supporting Information Available Speciation diagrams of Cd, Ni, and Co. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Official Journal of the European Community. Directive 98/78/ EC, 1998; control: 0378-6978, 40. (2) USEPA. National primary drinking water standards; EPA 816F-01-007; 2001. http://www.epa.gov/safewater/. (3) Hedin, R. S.; Nairn, R. W.; Kleinmann, R. L. P. Passive Treatment of Coal Mine Drainage; U.S. Bureau of Mines Information Circular 9389; U.S. Department of the Interior, Bureau of Mines: Pittsburgh, PA, 1994. (4) Blowes, D. W.; Ptacek, C. J.; Benner, S. G.; McRae, C. W. T.; Bennett, T. A.; Puls, R. W. Treatment of inorganic contaminants using permeable reactive barriers. J. Contam. Hydrol. 2000, 45 (1-2), 123-137. (5) Younger, P. L.; Banwart, S. A; Hedin, R. S. Mine Water: Hydrology, Pollution, Remediation; Kluwer Academic: Dordrecht, 2002. (6) Benner, S. G.; Blowes, D. W.; Ptacek, C. J. A full scale porous reactive wall for prevention of acid mine drainage. Ground Water Monit. Restorat. 1997, 17, 99-107. (7) Benner, S. G.; Blowes, D. W.; Ptacek, C. J.; Mayer, K. U. Rates of sulfate reduction and metal sulfide precipitation in a permeable reactive barrier. Appl. Geochem. 2002, 17, 301-320. (8) Ball, J. W.; Nordstrom, D. K.; Zachmann, D. W. WATEQ4F - A personal computer fortran translation of the Geochemical model WATEQ2 with revised database; U.S. Geological Survey Open File Report 87-50; 1987. (9) Baes, C. F., Jr.; Mesmer, R. E. The hydrolysis of cations; John Wiley and Sons: New York, 1976. (10) Oustadakis, P.; Agatzini-Leonardou, S.; Tsakiridis, P. E. Nickel and cobalt precipitation from sulphate leach liquor using MgO as neutralizing agent. Miner. Eng., doi:10.1016/j.mineng. 2005.11.006 (in press). (11) Huang, H.-H.; Liu, Q. Bench-scale chemical treatability study of the Berkeley Pit water. In Emerging Technologies in Hazardous Waste Management V; Tedder, D. W., Pohland, F. G., Eds.; ACS Symp. Ser. 607; Oxford University Press: Washington, DC, 1995; pp 196-209.
(12) Garcia, M. A.; Chimenos, J. M.; Ferna´ndez, A. I.; Miralles, L.; Segarra, M.; Espiell, F. Low-grade MgO used to stabilize heavy metals in highly contaminated soils. Chemosphere 2004, 56, 481-491. (13) Cortina, J. L.; Holtermann, I.; de Pablo, J.; Cama, J.; Ayora, C. Passive in-situ remediation of metal-polluted water with caustic magnesia. Environ. Sci. Technol. 2003, 37, 1971-1977. (14) Eary, L. E. Geochemical and equilibrium trends in mine pit lakes. Appl. Geochem. 1999, 14, 963-987. (15) Stipp, S. L. S.; Parks, G. A.; Nordstrom, D. K.; Leckie, J. O. Solubility product constant and thermodynamic properties for synthetic otavite, CdCO3(s), and aqueous association constants for the Cd(II)-CO2-H2O system. Geochim. Cosmochim. Acta 1993, 57 (12), 2699-2713. (16) Chen, X.; Wright, J. V.; Conca, J. L.; Peurrung, L. M. Evaluation of heavy metal remediation using mineral apatite. Water, Air, Soil Pollut. 1997, 98, 57-78. (17) Hummel, W. Nagra/PSI thermochemical database update: data selection of nickel; PSI Note AN-44-00-12; 2000. (18) Knacke, O.; Kubaschewski, O.; Hesselmann, K. Thermochemical properties of inorganic substances, 2nd ed.; Springer-Verlag: Berlin, 1991. (19) Hemingway, B. S.; Robie, R. A.; Kittrick, J. A. Revised values for the Gibbs free-energy of formation of Al(OH)4-aq, diaspore, boehmite and bayerite at 298.15K and 1 bar, thermodynamic properties of kaolinite to 800 K and 1 bar, and heats of solution of several gibbsite samples. Geochim. Cosmochim. Acta 1978, 42, 1533-1543. (20) Singer, P. C.; Stumm, W. Acidic mine drainage. The rate determining step. Science 1970, 167, 1121-1124. (21) Komnitsas, K.; Bartzas, G.; Paspaliaris, I. Efficiency of limestone and red mud barriers: laboratory column studies. Miner. Eng. 2004, 17, 183-194. (22) Amos, P. W.; Younger, P. L. Substrate characterisation for a subsurface reactive barrier to treat colliery spoil leachate. Water Res. 2003, 37, 108-120. (23) PIRAMID Consortium. Engineering guidelines for the passive remediation of acidic and/or metalliferous mine drainage and similar wastewaters; University of Newcastle Upon Tyne: Newcastle Upon Tyne, 2003. http://www.ncl.ac.uk/piramid/.
Received for review May 8, 2006. Revised manuscript received July 31, 2006. Accepted August 14, 2006. ES061092G
VOL. 40, NO. 20, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6443
