Environ. Sci. Technol. 2008, 42, 920–926
Heavy Metal Immobilization through Phosphate and Thermal Treatment of Dredged Sediments PETER NDIBA, LISA AXE,* AND THIPNAKARIN BOONFUENG Civil and Environmental Engineering, New Jersey Institute of Technology, University Heights, Newark, New Jersey 07102
Received August 20, 2007. Revised manuscript received November 13, 2007. Accepted November 19, 2007.
Disposal of dredged sediments is expensive and poses a major challenge for harbor dredging projects. Therefore beneficial reuse of these sediments as construction material is highly desirable assuming contaminants such as heavy metals are immobilized and organics are mineralized. In this research, the effect of the addition of 2.5% phosphate, followed by thermal treatment at 700 °C, was investigated for metal contaminants in dredged sediments. Specifically, Zn speciation was evaluated, using X-ray absorption spectroscopy (XAS), by applying principal component analysis (PCA), target transformation (TT), and linear combination fit (LCF) to identify the main phases and their combination from an array of reference compounds. In dredged sediments, Zn was present as smithsonite (67%) and adsorbed to hydrous manganese oxides (18%) and hydrous iron oxides (15%). Phosphate addition resulted in precipitation of hopeite (22%), while calcination induced formation of spinels, gahnite (44%), and franklinite (34%). Although calcination was previously used to agglomerate phosphate phases by sintering, we found that it formed sparingly soluble Zn phases. Results from the U.S. EPA toxicity characteristic leaching procedure (TCLP) confirmed both phosphate addition and calcination reduced leachability of heavy metals with the combined treatment achieving up to an 89% reduction.
Introduction Large quantities of sediments are dredged from harbors and waterways for maintenance and extension of water depths. These sediments are often contaminated with heavy metals and organic toxins arising from harbor, industrial, and urban activities. Historically, dredged materials have been relocated to open waters, rivers, and estuaries or placed in wetlands; however, enactment of stringent ocean disposal criteria in the U.S. has reduced the volume of materials acceptable for open sea disposal (1, 2). With increasing costs for land deposition (2), disposal of dredged sediments poses a major challenge for dredging projects. Treatment and reuse of sediments as an alternative to disposal is highly desirable as it would reduce costs and conserve resources. Many of the existing treatment technologies for reuse involve thermal application for production of cement, lightweight aggregate, bricks, and ceramic and glass tiles (2–4). However, few of these technologies are commercially used because of high costs and the large quantities of sediments involved (4, 5); as such, there is need for further development. * Corresponding author e-mail:
[email protected]; fax: 973– 596–5790. 920
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 3, 2008
The potential for producing construction materials from sediments using phosphate addition, followed by calcination at 700-900 °C, to stabilize heavy metals and mineralize organics was demonstrated recently where metal leachability was significantly reduced (6). Phosphate minerals are widely used to stabilize heavy metals in wastes, contaminated soils, and sediments (7–9). Stabilization of metals is via formation of sparingly soluble phosphate phases (7, 8); however, the mechanism may vary from metal to metal (9). Combined phosphate addition and thermal treatment was first applied to fly ash (10, 11), where the thermal step was used to agglomerate phosphate phases by sintering. Nzihou and Sharrock (11) however observed reduction in Cu leachability after calcination only and hypothesized that Cu was incorporated into the glassy phase. Similar observations have been made for Cu-doped kaolin heated at 900 °C (12), and Cd and Zn sorbed to hydrous iron oxide after calcinations at 900 °C as compared to 600 °C (13). Recently, Shih et al. (14) identified nickel aluminate spinel (NiAl2O4) as the stabilized phase in nickel-laden sludge sintered with kaolinite. Specific to the issue of heavy metal leaching and stabilization is an understanding of speciation to elucidate precipitation/sorption mechanisms. X-ray absorption spectroscopy (XAS) has in the recent past proven to be useful in direct assessment of speciation for amorphous, poorly crystalline, or low-concentration phases observed in soils and sediments (15–17). Because of heterogeneity, however, bulk XAS spectra are weighted averages of the species present (16) and therefore require deciphering. Multishell fitting using ab initio techniques can determine the predominant species but can easily overlook those with weaker backscattering contributions (15). Linear combination fitting (LCF) can identify and quantify up to three main species; however, it is uncertain whether a complete set of species is obtained and whether the set is unique (17, 18). An alternative technique, principle component analysis (PCA), addresses the number of species present which are then identified by target transformation (TT) and quantified by LCF. The objective of this study was to investigate speciation in contaminated sediments treated by phosphate addition followed by calcination, demonstrating treatment effectiveness in immobilizing metals. Zinc speciation was assessed with XAS specifically because of its abundance in contaminated sediments samples from both the U.S. and Europe. On the basis of the phases present, a thermodynamic analysis was conducted to examine whether solubility could predict leaching with the U.S. Environmental Protection Agency (EPA) toxicity characteristic leaching procedure (TCLP) (19).
Materials and Methods Laboratory quality assurance and quality control procedures were based on Standard Methods for the Examination of Water and Wastewater (20). Dredged sediments were obtained from Dampremy, Belgium, courtesy of Solvay Company. Treatment was carried out by addition of phosphoric acid at 2.5% of sediments by dry weight, followed by calcination at 700 °C for 3 h (6). Sediment mineralogy was characterized by X-ray diffraction (XRD) using a Philip’s X’Pert X-ray diffractometer. Total metal concentrations were obtained by HNO3 and HCl acid (aqua regia) digestion (21), while organic matter was estimated by loss on ignition at 445 ( 10 °C for six hours (22). XAS Data Collection and Analyses. X-ray absorption spectroscopy was conducted at beamline X11A of the National Synchrotron Light Source (NSLS), Brookhaven National 10.1021/es072082y CCC: $40.75
2008 American Chemical Society
Published on Web 01/05/2008
TABLE 1. Sediments Total Metal Concentrations and TCLP Results for Dredged and Treated Sediments total metala,b metal leached from in dredged dredged sediments hazardous materials sediments by TCLPb,c regulatory limitsd (mg/kg) (mg/kg) (mg/l)
dredged 5.56 ( 0.03
pH Zn Co Ni Cu Pb Cd
TCLP extract pH and metal concentration (mg/L)
3400 ( 130 30 ( 0.1 105 ( 1.3 120 ( 4.5 878 ( 9.8 11 ( 2.3
740 ( 36 3.1 ( 0.02 9.9 ( 0.24 1.8 ( 0.08 7.0 ( 0.30 3.5 ( 0.08
250 80 25 20 5.0 1.0
f
37 ( 1.8 0.16 ( 0.003 0.50 ( 0.012 0.09 ( 0.004 0.35 ( 0.015 0.18 ( 0.004
2.5% H3PO4 treated
calcined
2.5% H3PO4 and calcined
5.02 ( 0.06
6.04 ( 0.17
5.37 ( 0.15
19.8 ( 0.36 0.15 ( 0.003 0.55 ( 0.004 0.07 ( 0.008 0.15 ( 0.019 0.10 ( 0.001
3.7 ( 0.87 0.16 ( 0.003 0.24 ( 0.006 0.07 ( 0.005 0.25 ( 0.004 0.04 ( 0.001
5.1 ( 0.49 0.14 ( 0.008 0.23 ( 0.009 0.08 ( 0.011 0.22 ( 0.035 0.04 ( 0.009
reduction in leachinge (%)
89 ( 1.3 27 ( 3.5 62 ( 1.5 26 ( 0.2 48 ( 5.5 82 ( 2.3
a EPA Method 3050B (21). b Based on dry weight. c Computed by multiplying the fifth column by the liquid to solid ratio, 20. d RCRA limits (32) for Pb, Cd; CCR limits (46) for Zn, Ni, Co and Cu. e Adjusted to account for loss of 16.8% by weight organic matter content during calcination. f Error refers to 2 × standard error.
Laboratory, where the average electron beam energy was 2.8 GeV with a current of 260–280 mA. All spectra were collected at the Zn K-edge (9,659 eV) using a double-Si (111) crystal monochromator over the energy range of 9630–10 410 eV. Ground sediments were measured in fluorescence mode using a Lytle detector (23) filled with Ar, soller slits, and a z-1 filter. Reference compounds were measured in transmission mode with nitrogen gas in the ion chamber before the sample (I0) and in the transmission chamber (It). Harmonic rejection was achieved by detuning 30% of I0. XAS spectra were analyzed using WinXAS (version 3.1) (24), following standard procedures (25). The background X-ray absorbance was subtracted by fitting a linear polynomial through pre-edge region, and the edge jump was normalized with a zero-order polynomial over 9.759–9.859 keV. The threshold energy (Eo) was determined from the first inflection point in the edge region and was used to convert the spectra from energy to k-space. An advanced spline function was employed to enhance isolated atomic absorption over the range of 2.6–13.9 Å-1 and to convert the data to χ(k) that was then weighted by k3 to enhance higher k-space data. A Bessel window function was used in Fourier transforms to produce the radial structure–functions (RSF), which were then fit with theoretical models generated using FEFF7 (26). The amplitude reduction factor (S02) for spectra collected in fluorescence mode was obtained by fixing first-shell coordination numbers (CN) at crystallographic values and floating all other parameters. Selection of Reference Spectra. The success of PCA, TT, and LCF analyses depends on the availability of an exhaustive set of reference spectra. Because samples in this study were sediments, dredged and treated by phosphate addition and calcination, potential Zn phases in these environments were considered. The selected reference minerals (Supporting Information) included sphalerite (ZnS), zincite (ZnO), smithsonite (ZnCO3), hydrozincite (Zn5(OH)6(CO3)2), hemimorphite (Zn4Si2O7(OH)2 · 2H2O), willemite (Zn2SiO4), gahnite (ZnAl2O4), franklinite (ZnFe2O4), hopeite (Zn3(PO4)2 · 4H2O), and scholzite (CaZn2(PO4)2 · 2H2O). Others selected were ones addressing sorption to ubiquitous heavy metal scavengers in the environment: hydrous iron oxide (HFO), hydrous manganese oxide (HMO), and montmorillonite; important sorbents for Zn (27–29). PCA-TT-LCF Analysis. PCA determination of the number of species requires the number of spectra to exceed the number of species in the sample with a varying species composition (17). Because, in this study, bulk spectra were collected for samples of fixed composition, PCA was used primarily to facilitate TT identification of the likely endmembers of the sample spectra. Sample χ(k)k3 spectra in the range of 3.0–10.0 Å-1 for dredged sediments and 2.7–10.5 Å-1
for treated sediments were loaded into SIXPack PCA program (30) and decomposed mathematically into the minimum number of components necessary to describe variance in the data (31). The likelihood of reference spectra to be endmembers of the sample spectra was determined by target transformation of the reference spectra using the decomposed components and was assessed by a fitting factor R ) Σ(k3χexp - k3χmodel)2/(k3χexp)2 (16). Proportions of the identified member spectra were quantified by LCF starting with the lowest R spectrum for contributions of not less than 10%, the approximate method precision (16). To address differences in spectra collection modes between reference minerals and samples, the amplitude reduction factor (S02) was applied to the reference spectra collected in transmission. Leaching Tests. The U.S. EPA TCLP leaching procedure (19) was carried out to determine the efficacy of the treatment, as well as obtain regulatory information on the treated product (12, 32). The procedure was performed on sediments: dredged, 2.5% phosphate addition, calcination at 700 °C, and phosphate addition followed by calcination. Extractions were carried out with 20:1 liquid to solid ratio acetic acid solution at pH 2.88 ( 0.05 based on the sediment pH (19), in 2 L HDPE bottles tumbled at 30 rpm for 18 h. The pH was measured, and the extract was filtered, acidified, and stored in closed Nalgene bottles under refrigeration at 4 °C (20) until analysis by flame atomic absorption (FAA) spectrometry using AAnaylst 400 spectrometer (33).
Results and Discussions Sediments Characterization. Zinc was the most abundant heavy metal at 3400 mg kg-1 dry sediment (Table 1). The organic matter concentration was 168 g kg-1 dry sediment. XRD patterns showed quartz (R-SiO2) as the predominant mineral in dredged sediments, followed by calcite (CaCO3), hematite (R-Fe2O3), and aluminosilicates ((Na,Ca)AlxSiyOz). After phosphate addition and calcination, quartz (SiO2) and hematite (R-Fe2O3) were predominant, with smaller amounts of calcium phosphate Ca3(PO4)2 and hydroxylapatite (Ca5(PO4)3(OH)). Similar to other studies (34, 35), Zn-bearing minerals were not detected by XRD, probably because of their low concentration (
