Environ. Sci. Technol. 2004, 38, 1561-1568
Effect of Remineralization on Heavy-Metal Leaching from Cement-Stabilized/Solidified Waste M O H A M M A D Z . I S L A M , †,‡ L I O N E L J . J . C A T A L A N , * ,† A N D ERNEST K. YANFUL‡ Department of Chemical Engineering, Lakehead University, 955 Oliver Road, Thunder Bay, Ontario, Canada P7B 5E1, and Department of Civil and Environmental Engineering, The University of Western Ontario, London, Ontario, Canada N6A 5B9
Crushed samples of stabilized/solidified (s/s) waste were leached at constant leachate pH in the pH range 4-7 with nitric acid solutions to evaluate the influence of remineralization on metal release. The s/s waste consisted of synthetic heavy-metal sludge containing 0.1 mol L-1 copper nitrate, 0.1 mol L-1 zinc nitrate, and 0.1 mol L-1 lead nitrate mixed with ordinary Portland cement. Unleached and leached particles were characterized by scanning electron microscopy and energy-dispersive X-ray spectrometry. Two consecutive leaching fronts advancing from the surface of the particles toward the center were identified: the first front was associated with the dissolution of portlandite and partial reaction of the calcium silicate hydrate gel, while the second front was associated with the dissolution of calcium-aluminum hydroxy sulfates such as ettringite and monosulfate. At pH 4 and 5, a remineralization zone rich in heavy metals formed immediately behind the second leaching front. The shell extending from the remineralization zone to the surface of the particles was depleted in calcium, sulfate, and heavy metals. As a result of remineralization, heavy-metal releases to the leachate were reduced by factors ranging between 3.2 and 6.2 at pH 4 and between 74 and 193 at pH 5. At pH 6 and 7, remineralization of Pb and Zn occurred further behind the second leaching front and closer to the surface of the particles. The amount of heavy-metal release depended on both the leachate pH and the remineralization factor.
Introduction Stabilization/solidification (s/s) processes are widely applied for treating a variety of hazardous waste streams, including solutions, sludges, slurries, contaminated soils, dusts, and other particulate matter (1). In general, s/s refers to the mixing of additives with the waste to entrap or encapsulate the contaminants within a solidified matrix, and thereby reduce the rate of contaminant migration to the environment. Many s/s processes have been developed, including chemical processes (cement-based, pozzolan-based, lime-based, and phosphate-based), physical processes (macroencapsulation and microencapsulation), and thermal processes (thermoplastic polymer encapsulation and vitrification) (1). This work * Corresponding author phone: (807) 343-8573; fax: (807) 3438928; e-mail:
[email protected]. † Lakehead University. ‡ The University of Western Ontario. 10.1021/es034659r CCC: $27.50 Published on Web 01/15/2004
2004 American Chemical Society
deals with the leaching behavior of heavy-metal sludges treated by the cement-based s/s processes. A review of s/s applications at Superfund sites indicates that heavy-metal concentrations in untreated waste typically range from 50 to 70000 mg/kg, with concentrations as high as 424000 mg/kg for lead and 170000 mg/kg for cadmium (2). The mobility of contaminants in the treated waste is typically evaluated by means of leaching tests. Monitoring programs for industrial s/s processes usually include regulatory static batch leaching tests, such as the toxicity characteristic leaching procedure (TCLP) (3). In these tests, the s/s waste is crushed to a specific maximum grain size and contacted with a leaching fluid for several hours at a specified liquid-to-solid ratio. Organic and inorganic acid solutions, as well as buffer solutions, are commonly used as leaching fluids. Several investigations of cement mortar specimens exposed to acid or neutral leaching fluids (4-6) have described the formation of an outer layer of altered material having depleted calcium content and reduced pH. Calcium depletion in the outer layer has been attributed to a decrease in the Ca/Si ratio of the calcium silicate hydrate gel (abbreviated as CSH) and to the dissolution of other cement hydration products such as portlandite (Ca(OH)2), monosulfate ([Ca2(Al,Fe)(OH)6]2(SO4)‚xH2O), ettringite (Ca6Al2(SO4)3(OH)12‚26H2O), and calcite (CaCO3) for carbonated samples (6). Remineralization zones located between the outer layer and the core of the specimens have also been reported and characterized as enriched in iron hydroxide (4) or associated with the reprecipitation of secondary monosulfate, ettringite, and calcite (6). Several leaching models have been developed to describe the reaction of cement hydration products in s/s waste contacted by an acid leachant (7-12). Model predictions indicate that heavy metals dissolved in the outer layer can diffuse inward and reprecipitate as hydroxides (9, 10), thus forming a remineralization zone that may provide a sink for heavy metals and limit their release to the leachate. In this work, we demonstrate the existence of a remineralization zone enriched in heavy metals through examination of leached s/s waste samples by combined scanning electron microscopy and energy-dispersive X-ray spectrometry (SEM/EDX), and we attempt to quantify the effect of remineralization on limiting the release of heavy metals during static leaching tests carried out at constant pH in the pH range 4-7.
Materials and Methods Sample Preparation. Heavy-metal sludge was prepared by adding NaOH to a solution of 0.1 mol L-1 copper nitrate, 0.1 mol L-1 zinc nitrate, and 0.1 mol L-1 lead nitrate to reach an end point pH of 9.0. Ultrapure water (Barnstead Nanopure Diamond) was used to prepare all solutions. Heavy-metal concentrations in the sludge were within the range of concentrations encountered in real wastes that are treated by s/s (2, 13). The weight ratio of wet sludge to ordinary Portland cement (OPC) in the s/s waste was 0.4. The required amounts of OPC and sludge were placed in a clean plastic bowl and thoroughly mixed with a plastic spatula. The mixtures were then poured in three layers into PVC molds (7.6 cm diameter × 12.7 cm height) lined with Parafilm. For better consolidation, each layer was rodded with a rounded end plastic tamping rod at a rate of 25 strokes per layer. To eliminate air bubbles, the mold walls were tapped with a mallet. The samples were then cured in an ESPEC ESL-3CA environmental chamber at 20 °C and 99% humidity for 28 days. The cured s/s waste was crushed and ground by mortar VOL. 38, NO. 5, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1561
FIGURE 1. Backscattered electron image of unleached s/s waste showing (a) gel, (b) portlandite, (c) unhydrated Ca3SiO5, and (d) aluminate phase. and pestle. Particles having sizes between 425 and 850 µm were separated using sieves and a CSC Meinzer II sieve shaker for use in the leaching tests. This range of particle sizes allowed viewing the entire particle cross-sections under the scanning electron microscope operated at the lowest available magnification. The selected value of 0.4 for the sludge-to-cement ratio is consistent with previous studies by Park (14) and Poon et al. (15), who used ratios of 0.3 and 0.5, respectively. Attempts at preparing s/s waste samples with higher heavy-metal concentrations (0.2, 0.4, and 0.6 mol L-1) and a sludge-tocement ratio of 0.5 were unsuccessful, as the resulting mixture did not gain strength. Control samples with no addition of heavy metals and a water-to-cement ratio of 0.4 were also prepared. Leaching Tests at Constant pH. Leaching tests were carried out by mixing 30 g of s/s waste with 600 mL of ultrapure water in a three-neck round glass flask. A constant pH was then established and maintained throughout the test duration by controlling the addition of 5 mol L-1 HNO3 solution to the leaching flask using a Metrohm Titrino 785 automatic titrator equipped with a Metrohm Solitrode 6.0228.000 combined glass electrode. The tests reported in this paper were carried out at pH values of 4, 5, 6, and 7. The pH variation during tests was typically less than (0.1 unit. To prevent carbonation, the leaching vessel headspace was filled with nitrogen gas that was continuously bubbled in the leaching solution. Each test ran for about 150 h, and the acid consumption at the end of the tests reached 16.0, 13.3, 12.7, and 12.6 equiv/kg of s/s waste at pH 4, 5, 6, and 7, respectively. The liquid-to-solid ratio decreased from 20:1 at the start of the tests to approximately 13:1 by the end of the tests due to the combined effects of evaporation, leachate sample collection, and acid addition. The initial liquid-to-solid ratio was selected to be the same as in the regulatory TCLP test (3). The slurry in the leaching flask was continuously agitated using a Caframo BDC 1850 stirrer operated at 270 rpm to ensure that all the s/s waste particles were suspended. The stirrer shaft and paddles were made of plastic material to prevent contact of the leaching solution with any metal parts. A constant temperature of 25 ( 0.5 °C was maintained by immersing the leaching flask in a regulated temperature bath filled with water. The leachate was sampled every 3 h during the first 12 h of the tests and every 8 h afterward. Leachate samples were collected using disposable plastic syringes whose tips were connected to a 0.64 cm diameter plastic tube immersed in the leachate. Leachate samples were filtered through 0.22 µm Whatman Nuclepore esters of cellulose membranes. The 1562
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 5, 2004
FIGURE 2. Backscattered electron images of s/s waste particles leached at (a, top) an early stage of leaching and (b, bottom) an intermediate stage of leaching.
FIGURE 3. Backscattered electron image of a partially reacted Ca3SiO5 grain after leaching at pH 7 showing (a) reacted Ca3SiO5, (b) unreacted Ca3SiO5, and (c) nonreactive aluminate phase. filtrates were acidified with 0.04 mL of 12 mol L-1 hydrochloric acid solution and analyzed by inductively coupled plasmaatomic emission spectroscopy (ICP-AES) for metals and major elements. The measurements of solute concentrations in leachate samples were used to calculate the amount of solutes leached from the solids as a function of time (see Appendix S-1, Supporting Information). The variation in liquid-to-solid ratio during the tests was taken into consideration for these calculations.
FIGURE 4. Backscattered electron images of s/s waste particles at the end of the leaching tests at (a, top left) pH 4, (b, top right) pH 5, (c, bottom left) pH 6, and (d, bottom right) pH 7. Leaching times were 120 h at pH 4, 144 h at pH 5, 147 h at pH 6, and 142 h at pH 7. EDX analyses of the gel were carried out along the line segments denoted AB. SEM/EDX Analyses. The solids recovered from the leaching tests were air-dried at 40 °C for 48 h. Thin sections of both unleached and leached s/s waste were prepared by impregnating the particles with epoxy resin and polishing them with kerosene to obtain a smooth grain section without dissolving water-soluble minerals. The thin sections were coated with carbon using an EDWARDS AUTO306 carbon coater and examined with a JEOL JSM 5900 scanning electron microscope. Elemental compositions within the s/s waste particles were measured by EDX using an Oxford Link ISIS system calibrated using a series of standards. Garnet was used for Mg, Al, and Si, orthoclase for K and Na, wollastonite for Ca, barite for S, and the respective pure metals for Cu, Fe, Pb, and Zn. Each EDX spectrum was collected for 200 live seconds of accumulated count duration. The detection limits for Ca, Cu, Pb, Si, and Zn (calculated as twice the standard deviations of EDX measurements) were typically 0.08, 0.07, 0.10, 0.06, and 0.08 wt %, respectively. Because leaching increased the porosity of the s/s waste due to partial dissolution of the cement matrix, the sum of the measured elemental concentrations was sometimes significantly lower than 100 wt %. Reported elemental concentrations were normalized to 100 wt % to represent the concentrations in the solid.
Results and Discussion Leaching of Cement Phases. Figure 1 is a typical backscattered electron image of unleached s/s waste showing the presence of portlandite (Ca(OH)2), a gel-like hydration
product referred to as “gel” in this paper, unhydrated tricalcium silicate (Ca3SiO5), and an aluminate phase having the average elemental composition Ca2.65Al1.00Si0.59Fe0.48Mg0.43. EDX analyses were carried out at different locations in several unleached s/s waste particles to determine the elemental composition of these phases. Nine to twelve analyses were carried out for each phase. The elemental composition of the gel (Table S-1, Supporting Information) indicates that it is a mixture of calcium silicate hydrate (CSH) and calciumaluminum hydroxy sulfates such as ettringite and monosulfate. Heavy metals were only detected in the gel (Cu ) 0.311 ( 0.069 wt %; Pb ) 1.00 ( 0.22 wt %; Zn ) 0.314 ( 0.083 wt %) and in portlandite (Cu ) 0.213 ( 0.052 wt %; Pb ) 0.70 ( 0.23 wt %; Zn ) 0.32 ( 0.12 wt %). Note that the standard deviations given for each element reflect the spatial variability of the gel and portlandite compositions within the s/s waste rather than measurement uncertainty. The control samples devoid of heavy-metal addition had a microstructure similar to that of s/s waste samples and contained the same solid phases with similar elemental compositions (Table S-2, Supporting Information). Figure 2 shows backscattered electron images of s/s waste particles at two successive stages of leaching at pH 4. Figure 2a, which corresponds to an early stage of leaching and an acidity consumption of 3.2 equiv/kg, shows a first leaching front separating the core of the particle from a darker (less dense) outer layer denoted shell 1. The composition of the core was not significantly different from that of the unleached material. However, portlandite was absent from shell 1, and VOL. 38, NO. 5, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1563
SEM/EDX analyses of the gel in the core and in shell 1 (Table S-3, Supporting Information) show that the Ca/Si mass ratio was significantly lower in shell 1 (2.77 ( 0.26) than in the core (3.96 ( 0.26). Hence, the advancement of this first leaching front is associated with complete dissolution of portlandite and partial reaction of the gel. At an intermediate stage of leaching, when the acidity consumption reached 11.2 equiv/kg, a second leaching front became visible near the surface of the particles (Figure 2b). The composition of the gel in the inner part of the particles (Table S-4, Supporting Information) was statistically indistinguishable from that of the gel in shell 1 measured at the earlier stage of leaching (Table S-3, Supporting Information). Therefore, the inner part of the particles at the intermediate stage of leaching is also denoted shell 1. The outer layer is denoted shell 2, and its average composition (Table S-4, Supporting Information) shows that it is largely depleted in sulfur (0.292 ( 0.049 0.338 ( 0.020 0.179 ( 0.028 >0.251 ( 0.033 0.212 ( 0.014
4.10 ( 0.44 >3.21 ( 0.37 6.17 ( 0.31 102.41 ( 16.05 >73.87 ( 9.60 193.27 ( 12.74
( >d0.423 ( 0.049 0.403 ( 0.020 0.181 ( 0.028 >0.255 ( 0.033 0.213 ( 0.014 0.387b
0.041c
a Based on 30 g of s/s waste. The initial concentration of Cu, Zn, and Pb in the s/s waste was 25.1 mmol kg-1 for each metal. b Based on the average metal concentration in shell 2. c Standard error considering the propagation of errors when using concentration averages in calculations. d The “>” sign indicates that some EDX analyses of Zn concentration in shell 2 were below the detection limit. These measurements were taken to be equal to the detection limit for calculating the average Zn concentration in shell 2, and the amount leached from shell 2 is therefore underestimated.
FIGURE 10. Releases of Cu, Pb, and Zn in the leachates versus time. Releases at pH 4 are read on the right vertical axis; releases at pH 5-7 are read on the left vertical axis. mation). The average thickness of shell 2 and the particle size distribution were then used to calculate the volume of shell 2 at each leachate pH. Next, the total amount of metals leached from shell 2 was determined by considering the dissolution of portlandite and the changes in gel composition between the start and the end of leaching (Tables S-5 to S-8, Supporting Information). More details on these calculations are provided in Appendix S-2, Supporting Information. Table 1 compares the amounts of metals leached from shell 2, released to the leachate, and remineralized on the basis of 30 g of s/s waste leached at pH 4 and 5 at the end of the tests. Larger amounts of metals were remineralized at pH 4 than at pH 5; however, remineralization had a larger limiting effect on metal release to the leachate at pH 5. This
was ascertained by calculating a remineralization factor, fr, defined for each metal as the ratio of the total amount leached from shell 2 to the amount released to the leachate. With this definition, fr would be unity in the absence of metal remineralization, and fr would be infinite if all the metals that leached from shell 2 reprecipitated in the remineralization zone and none reported to the leachate. The value of fr can also be interpreted as the factor by which metal release is decreased as a result of remineralization. The values of fr ranged from >3.2 ( 0.4 to 6.2 ( 0.3 at pH 4, and from >74 ( 10 to 193 ( 13 at pH 5. Lead was most effectively remineralized, followed by copper, and then by zinc. The numerical values of the remineralization factor are specific to the leaching conditions of the tests and are not directly applicable to s/s waste placed in a natural environment. In typical land disposal situations, the leachate is constantly renewed due to the flow of groundwater around the s/s waste or through cracks and fractures in the material. Moreover, the leachate pH will remain alkaline for thousands of years because the cement matrix has a high alkalinity (13, 17). As groundwater flows around or through the s/s waste, consecutive leaching fronts associated with the dissolution of portlandite, CSH gel, and calcium-aluminum hydroxy sulfates will form and advance through the material adjacent to the flow paths. Heavy metals will not be significantly mobilized for a long period of time because the pore water pH will remain alkaline. Once the alkalinity of the s/s waste is sufficiently depleted so that the leachate becomes neutral or slightly acidic, the pore water pH will progressively decrease in the material next to the flow paths, and heavy metals will start to leach in this region. A fraction of the leached metals will diffuse toward the leachate contained in the flow paths. The remainder of the leached metals, however, will diffuse away from the flow paths toward regions of the material where dissolved metal concentrations in pore water are low. Because the pore water pH remains alkaline at some distance from the flow paths, these metals will precipitate and form a remineralization zone. With time, the remineralization zone will grow and continue to move away from the flow paths. Therefore, the effect of remineralization in a natural environment is to limit the release of heavy metals to groundwater flowing through or around the s/s waste.
Acknowledgments This work was funded by the Natural Sciences and Engineering Research Council of Canada (NSERC). We thank A. Hammond, A. Mackenzie, and A. Raitsakas for their assistance with the preparation of thin sections for SEM and with ICP analyses.
Supporting Information Available Two appendices, eight tables, and five figures showing that the remineralization of heavy metals in cement-stabilized VOL. 38, NO. 5, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1567
waste exposed to leaching solutions at pH 4-7 reduces the release of toxic metals to the leachate. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Conner, J. R.; Hoeffner, S. L. Crit. Rev. Environ. Sci. Technol. 1998, 28 (4), 397-462. (2) EPA. Solidification/Stabilization Use at Superfund Sites; EPA/ 542-R-00-010; Office of Solid Waste and Emergency Response, Technology Innovation Office, U.S. Environmental Protection Agency: Washington, DC, 2000. (3) Test Methods for Evaluating Solid Wastes, Physical/Chemical Methods; Office of Solid Waste, U.S. Environmental Protection Agency: Washington, DC, SW-846 on-line, http://www.epa.gov/ epaoswer/hazwaste/test/main.htm (accessed November 2003). (4) Chandra, S. Cem. Concr. Res. 1988, 18, 193-203. (5) Cheng, K. Y.; Bishop, P. L. J. Hazard. Mater. 1990, 24, 213-224. (6) Faucon, P.; Adenot, F.; Jorda, M.; Cabrillac, R. Mater. Struct. 1997, 30, 480-485. (7) Baker, P. G.; Bishop, P. L. J. Hazard. Mater. 1997, 52, 311333. (8) Barna, R.; Sanchez, F.; Moszkowicz, P.; Me´hu, J. J. Hazard. Mater. 1997, 52, 287-310.
1568
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 5, 2004
(9) Batchelor, B. Water Sci. Technol. 1992, 26 (1-2), 107-115. (10) Batchelor, B. Environ. Sci. Technol. 1998, 32, 1721-1726. (11) Hinsenveld, M.; Bishop, P. L. In Stabilization and Solidification of Hazardous, Radioactive, and Mixed Wastes; Gillian, T. M., Wiles, C. C., Eds.; ASTM STP 1240; American Society for Testing Materials: West Conshohocken, PA, 1996; Vol. 3, pp 528-539. (12) Park, J.-Y.; Batchelor, B. Water Res. 2002, 36, 156-166. (13) Catalan L. J. J.; Merlie`re, E.; Chezick, C. J. Hazard. Mater. 2002, B94, 63-88. (14) Park, C. K., Cem. Concr. Res. 2000, 30, 429-435. (15) Poon C. S.; Lio, K. W. Waste Manage. 1997, 17 (1), 15-23. (16) Visual MINTEQ Ver 2.0, http://www.lwr.kth.se/english/OurSoftware/vminteq/ index.htm (accessed November 2003). (17) Catalan L. J. J.; Wetteskind, K. L. In Waste Management and the Environment; Almorza, D., et al., Eds.; Wessex Institute of Technology Press: Southampton, U.K., 2002; pp 383-397.
Received for review June 26, 2003. Revised manuscript received November 27, 2003. Accepted December 5, 2003. ES034659R
