Environ. Sci. Technol. 2002, 36, 3165-3169
Direct Determination of Cadmium Speciation in Municipal Solid Waste Fly Ashes by Synchrotron Radiation Induced µ-X-ray Fluorescence and µ-X-ray Absorption Spectroscopy M A R I A C A T E R I N A C A M E R A N I P I N Z A N I , * ,† ANDREA SOMOGYI,‡ ALEXANDRE S. SIMIONOVICI,§ STUART ANSELL,# BRITT-MARIE STEENARI,† AND OLIVER LINDQVIST† Department of Inorganic Environmental Chemistry, Chalmers University of Technology, S-412 96 Go¨teborg, Sweden, Department of Chemistry, University of Antwerp (UIA), B-2610 Antwerp, Belgium, and ID22 and BM29, European Synchrotron Radiation Facility (ESRF), BP 220 F-38043 Grenoble, France
Cadmium is a toxic metal that causes environmental concern in connection with utilization and land filling of ash from combustion of municipal solid waste (MSW). Collecting information about the chemical associations of Cd in ash is fundamental since this affects its solubility and leachability from the ash material. In the work presented here, the content, distribution, and chemical forms of toxic metals especially of Cd on/in individual Municipal Solid Waste (MSW) fly ash particles have been investigated in situ by synchrotron radiation induced µ-X-ray fluorescence and absorption spectrometry. The use of an excitation energy of 27 keV made it possible to detect trace metals, such as Cd, present at ppm levels routinely. Changing the excitation energy in the vicinity of the absorption edge of Cd (26.71 keV), the absorption spectra of this element were measured for the first time in this high energy range in micron-sized spots of individual fly ash particles. The measurements indicated Cd to be preferably concentrated in some small areas (“hot-spots”) with high concentration (up to 200 ppm) rather than in a homogeneous distribution or as a thin coating on the whole particle surface, making the surface-reaction the most probable mechanism of Cd enrichment during MSW combustion processes. Comparisons of XAS spectra of fly ashes and reference compounds showed that in the particles studied Cd is present in the oxidation state +2. Analyses of linear combinations of standard spectra allowed estimating the Cd presence within fly ash particles as an admixture of primarily CdSO4, CdO, and CdCl2 as well as an unidentified compound not included as a standard.
* Corresponding author phone: +46 31 7728210; fax: +46 31 7722853; e-mail:
[email protected]. † Chalmers University of Technology. ‡ University of Antwerp. § ID22, European Synchrotron Radiation Facility. # BM29, European Synchrotron Radiation Facility. 10.1021/es010261o CCC: $22.00 Published on Web 06/07/2002
2002 American Chemical Society
Introduction Incineration of municipal solid waste (MSW) has two main advantages: reducing the waste volume by about 90% and reducing reactivity by destruction of organic compounds. As in combustion of other fuels, the potentially toxic trace metals are concentrated in the fine ash fractions, i.e., the fly ash (1). The content of these metals (i.e. Pb, Ni, Cu, but mainly Cd) makes this residue ecologically harmful. An effective and safe handling of such ash requires a thorough knowledge of its chemical properties, in particular, their potential for dissolution and leaching. The dissolution and transport of metal ions from the ash matrix to soil water are key steps since dissolved ions are available for biological uptake and groundwater contamination. As a consequence, knowledge of the total concentrations of heavy metals in ashes provides only limited information, as it does not show how strongly the metals are bound to ash constituents. Thus, the risk associated with the presence of heavy metals depends primarily on their speciation. The notion of speciation is taken in its broadest sense and includes metal characteristics, such as electronic structure (oxidation state), possible associations with other elements, and chemical state. However, the determination of the in situ-speciation of heavy metals in fly ash particles is a difficult task because of their relatively high dilution and the structural and chemical complexity of the host material. Furthermore, the literature lacks studies of heavy metal within individual fly ash particles, which are of great importance for understanding the long term stability of the toxic fly ash components. Due to its wide range of applications and ability of accumulation, Cd is highly concentrated in many municipal wastes. Cadmium and cadmium compounds are used for example as a stabilizer in poly(vinyl chloride)s, and they were used as electrodes in batteries and other electrochemical cells for many years. This work describes a method for the determination of Cd speciation and its possible quantitation in single MSW fly ash particles. Cd distribution within single particles was investigated by Synchrotron Radiation induced µ-X-ray Fluorescence (µ-SRXRF) spectrometry and its speciation on single spots by Synchrotron Radiation induced µ-X-ray Absorption Spectroscopy (µ-SRXAS). Both techniques can be used in air, and they are usually nondestructive, due to the lower energy deposition compared to charged particle excitation. The high brightness of the third generation synchrotron radiation sources and the development of X-ray focusing optical elements make it possible to create beams of micrometer size with high-intensity making µ-XRF and µ-XAS appropriate tools for the analysis of individual particles with diameters of several tens of micrometers (2, 3). Elemental maps, showing the two-dimensional projection of the threedimensional elemental distribution of a particle, are created by scanning the particle in a regular grid pattern by the X-ray microbeam and detecting the induced fluorescent intensities at each position (xy-µ-XRF scan) (4). µ-XAS spectroscopy was used for the direct determination of the chemical forms of Cd in particular microsized areas of high Cd concentrations, by measuring the absorption of the excitation beam or the intensity of the Cd KR X-ray line as a function of the excitation energy in the vicinity of the K absorption edge (EK ) 26.711 keV) of Cd (-100 eV < E-EK < 400-1000 eV). Since the spectral scan is performed in the vicinity of an X-ray absorption edge of a chosen target element (Cd in our case), the method is therefore element-specific. The elemental specificity precludes interferences from compounds of other elements, and means that little or no sample preparation is VOL. 36, NO. 14, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3165
TABLE 1. Average (Av) and Maximal (Max) Halogens and Heavy Metals Content of the MSW Fly Ash Particles particle no.
particle size (ca., µm)
1 2 3 4 5 6 7 8 9 10 11
32 40 50 60 72 90 100 130 160 195 200
Br (ppm) av max 6.7 12 500 0.4 1400 500 2100 88 2500 1800 400
500 56 3100 18 3000 2900 4600 500 8700 4200 2400
Cl (% w/w) av max 0.55 0 0.49 0 0.62 0.54 0.31 0.57 0.80 0.06 21.5
2.69 0 10.3 0 4.71 8.79 4.05 7.18 4.39 1.45 21.7
required, eliminating the possibility of inadvertently altering the speciation during separation processes. In combination with the penetrating capability of X-rays, it makes µ-XAS ideal for complex or difficult samples, like fly ashes.
Experimental Section Materials. The study has been performed on fly ashes from a Bubbling Fluidized Bed (BFB) combustion unit of 2 × 15 MW fired with MSW. Only textile filter ashes were investigated here. Eleven single particles of different dimensions (varying from ca. 30-200 µm in diameter) were selected, and each of them was glued on a 100 µm diameter quartz capillary before the µ-XRF analysis. The mounted samples were stored in plastic boxes prior to the analysis, to minimize the risk of contamination by dust particles. The dimensions of the fly ash particles analyzed here are given in Table 1. µ-XAS experiments were performed only on the particles presenting high Cd concentrations (three particles) and only on those spots where Cd was predominantly accumulated (four spots totally), as shown by the XRF maps. Methods. Both µ-SRXRF and µ-SRXAS measurements were performed in the first experimental hutch of the ID22 beamline (EH1) of the ESRF. The beamline uses a U42 undulator on a high beta section producing a beam size of (VxH, 95% of the beam): 0.85 × 1.5 mm2 at the sample location in EH1 (42 m distance from the source). A flat Si mirror with two coatings (Pd and Pt strips, respectively) is used in horizontal deflection mode at 0.15° (2.6 mrad) nominal angle in order to filter out the high energy harmonics of the X-ray beam (cutoff energies of the strips are Si: 12 keV, Pd: 24 keV, Pt: 32 keV). A vertical, flat, fixed exit, cam system double crystal monochromator (Kohzu) is used in the 3-30° angular range in order to achieve the energy ranges of 4-37 keV with Si[111] and 7-72 keV with Si[311] crystals, respectively. The (∆E/E) energy resolution of the monochromatic excitation radiation is a few 10-4. For the demagnification of the synchrotron source and creating the microbeam, a compound refractive lens (CRL) consisting of 94 individual Al lenses (5, 6) was employed. A Au knife-edge sample was used to determine the size of the focused beam. The beam size was H × V ) 10*8 µm2 during fluorescence experiments and 12*3 µm2 during the absorption experiments. The horizontal beam size, used for the fluorescence experiments, was determined by a pinhole of 10 µm. The larger vertical beam size of the fluorescence experiments helped to obtain a faster mapping of the Cd distribution within the particle. The sample stage consisted of one rotation and three translation motor-stages: two for scanning the sample in the beam in horizontal and vertical directions and the third one in order to align the sample into the image plane of the focusing device. The minimal stepsizes of the horizontal and vertical translation stages were 1 µm and 0.1 µm, respectively. 3166
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 14, 2002
Zn (ppm) av max 2600 100 100 65 100 96 400 9900 200 400 1000
7900 900 2500 200 800 900 7600 29700 4000 2200 19100
Cd (ppm) av max 43 1.5 5.4 0 10 4.2 16 56 13 20 29
150 9.9 65 0 33 42 74 200 38 72 100
Pb (ppm) av max 500 84 200 51 300 200 500 600 300 500 600
3900 400 7600 200 2000 11800 1100 2200 3100 1400 4700
Cu (ppm) av max 500 47 87 63 200 100 200 800 200 400 600
1600 200 1200 400 700 1100 1300 2300 3100 1700 3600
During the measurements, the intensities of the incoming, focused and transmitted beam were monitored by an ionization chamber and two photodiodes, respectively. The intensity of the focused beam was used for normalizing the measured X-ray intensities of the sample. The characteristic X-ray line intensities of the sample were detected by a Si(Li) detector (GRESHAM Scientific Instruments LTD) of 30 mm2 active area, 5 mm thickness, and 8 µm thick Be window. The X-ray detector was placed at 90° to the incoming linearly polarized X-ray beam in order to decrease the intensity of the scattered lines in the spectra. A videomicroscope in reflection mode and a high-resolution X-ray CCD camera in transmission mode made the alignment of the sample and the micro-probe setup easier. A CANBERRA spectroscopy amplifier, analogue to digital converter (ADC), and 556A acquisition interface module (AIM) were used for collecting the multichannel analyzer (MCA) spectra. The AXIL software package (7, 8) was used for spectrum evaluation. The NISTSRM1832 thin glass and Cu and Au thin metal foils (Goodfellow, UK) were measured in order to estimate the number of incident photons. All the µ-XRF experiments were performed by using monochromatic radiation at an excitation energy of 27 keV. Cd K-edge µ-XAS experiments were performed on pure Cd compounds and fly ash particles as well. The reference compounds used in this study (Cd, CdCl2, CdO, CdSO4, CdS, CdBr2) were chosen due to their probability to be found in the ash material. Each reference material (0.5 mg) was crushed and mixed with 0.3 mg of boron nitride and pressed to form pellets of 1 mm thickness. The reference XAS spectra were recorded in transmission mode. Due to the low Cd concentration and the small dimensions of the fly ash particles, their XAS spectra were recorded in fluorescence mode. Due to the relatively small thickness of the fly ash particles, no self-absorption correction was necessary for these spectra, as evidenced by the good agreement between the fluorescence and transmission XAS spectra integrated for a longer time. All the µ-XAS spectra were obtained by scanning the energy in the 26.65-26.95 keV range in 1 eV steps maintaining the same setup used for the microfluorescence experiments. Each energy scan was repeated 4 times with 2 s live time/energy point. The higher intensity Si [111] reflex was used, yielding a resolution of about 3.5 eV, to be compared with the Cd core-hole width of 7.8 eV. Explanations of the XAS technique are not attempted here. Theoretical details or applications can be found in refs 9 and 10. The Command View (CMDV) data analysis package by S. Ansell (11) was used to fit the data obtained on the fly ashes by linear combinations of the measured reference compounds. Initially, Factor Analysis (FA) using the Singular Value Decomposition method was performed in order to roughly estimate the dominant standards for further fitting the fly
FIGURE 1. Elemental maps of particle 10 of diameter of ca. 200 µm. Pixel size: 10 × 8 µm2, spectrum collection time per pixel: 6 s. Darker tones indicate higher elemental abundance; light tones indicate lower concentrations. See the corresponding concentrations in Table 1.
FIGURE 2. Elemental maps of particle 8 of diameter of ca. 130 µm. Pixel size: 10 × 8 µm2, spectrum collection time per pixel: 6 s. Darker tones indicate higher elemental abundance; light tones indicate lower concentrations. See the corresponding concentrations in Table 1.
FIGURE 3. (a) Correlation between the intensities of Br and Cd in particle 8, points denote the measured values, the line denotes the correlation curve-set calculated from assuming a spherical particle shape containing Br and Cd in homogeneous distribution (16). (b) Correlation between the intensities of Br and Cd in particle 10, points denote the measured values, the line denotes the correlation curve-set calculated from assuming a spherical particle shape containing Br and Cd in homogeneous distribution (16).
ash spectra. All the standards presented in Figure 4, except Cd metal, were used as a basis set for the factor analysis. A linear combination of the three dominant contributions (12) which was obtained from the FA, namely CdSO4, CdCl2, and CdO, was then used to directly fit the fly ashes spectra.
Results and Discussion The concentrations of the major and minor components in each measured voxel of the single fly ash particles were determined from scanning µ-XRF experiments. Using prerecorded elemental sensitivity curves obtained for all elements present in the sample from calibrated concentration standards such as NIST SRM 1832/1833 (thin glass or thin metal foils), and tabulated fluorescence cross-sections, one can directly convert the measured characteristic X-ray line intensities into concentrations. The measured incident flux expressed in units of photons/second/µm2 of sample surface is used as a normalization parameter, rendering the calculated concentrations absolute, as detailed for example in ref 13. Table 1 reports the average and maximum concentrations of some heavy metals and halogens in each particle calculated from the concentration values determined in each voxel of the particles. The large differences between the average and maximum concentrations within individual particles indicate a considerable variation in concentrations with “hot-spots” containing about 10-100 times higher amount of a given element than the average. This finding supports the idea that the trace metals have special affinity to some ash minerals located near the particle surface (14). This can occur as a
FIGURE 4. Absorption spectra of the measured pure Cd compounds. The spectra are vertically offset to allow comparisons and were taken in fluorescence mode. result of the scavenging of Cd enriched submicron particles by larger particles. In addition, a number of variables affect the chemical and physical transformation of elements in the furnace, i.e.: (i) occurrence and distribution of the elements in the input waste, (ii) physical and chemical conditions in the furnace bed, e.g. temperature, redox conditions, chlorine and oxygen content, and (iii) the kinetics in the furnace bed, e.g. residence time and mixing conditions (15). The inhomogeneous distribution of heavy metals among and within individual particles is illustrated, as an example, in Figure 1, VOL. 36, NO. 14, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3167
FIGURE 5. µ-XRF maps of various elements obtained from particle 9 (diameter ca. 160 µm), showing self-absorption of the light elements. Imagine size: 20 × 20 pixels, pixel size: 10 × 8, spectrum collection time per pixel: 6 s. Darker tones indicate a higher elemental abundance; light tones indicate lower concentrations. µ-XAS in the 26.65-26.95 keV range was performed on a (HxV) 3 × 12 µm2 micro-area containing the highest concentration of Cd (“hot spot 9a”). where the spatial metal distributions of particle 10 (ca. 200 µm diameter) are shown. In each of those particles, the enrichment of Cu, Zn, Pb, and Cd (at the 3600, 29700, 11800, and 200 ppm level, respectively) in some well-defined hot-spots relative to the surrounding areas is clearly visible, suggesting that a part of these metals might be arranged in inorganic precipitates. Around the boundaries of those spots, zones with intermediate concentration levels are present, probably caused by diffusion of the heavy metals during the condensation process or by binding to ion exchange sites. The right-hand lower and the left-hand upper quarters of particle 10 show, for example, a zone with a high abundance of Cd, possibly corresponding to the nucleation area. For Cu, Zn, and Fe that area is on the left-hand upper quarter. Some variation in the halogens concentrations can be observed from particle to particle, but their presence in significant quantities indicates that the trace metals may be present in easily soluble salts. The elemental concentration distributions of particles 10, 9, and 8 are indicated in Figures 1, 5, and 2, respectively. The maps show the coexistence of heavy metals (Cd, Cu, Zn, and Pb) and halogens (Cl, Br) in certain small areas of those particles. In Figure 3a,b, the correlation graphs between Cd and Br are shown for particles 8 and 10 as an example. The effect of sample topology and self-absorption was modeled by assuming a SiO2 sphere of 130 and 195 µm diameters, respectively, and containing Br and Cd in homogeneous distribution (16). From the resulting correlation curve-set, plotted as a line, it can be seen that the points are mainly inside or in the vicinity of the correlation curve-set, indicating a homogeneous distribution of (correlation between) Cd and Br. The outlying points correspond to inhomogeneities in the particle matrix or hot-spots. The results obtained here are in agreement with the thermodynamical equilibrium results reported by Verhulst et al. (17) showing that, at standard conditions for waste incineration, Cd, Pb, and Zn volatilize as the corresponding halogens. Table 1 shows also that the volatile heavy metals do not present increased concentration trends with decreasing particle size (particle 11, 7, and 8 have a coarse diameter but still one of the highest concentrations of Zn, Cd, Cu, and Pb). According to previous investigations (18, 19), the smaller ash particles provide a large specific surface area for condensation of vaporized species (surface enrichment mechanism via 3168
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 14, 2002
diffusion-controlled heterogeneous condensation). Therefore, the precise mechanism whereby the larger particles become highly enriched in volatile is not obvious. A possible mechanism acting in parallel to the diffusion/condensation is one of the surface-reaction controlled deposition whereby heavy metals vapors react with highly reactive particle surfaces (14). The fact that some of the depositing material goes also to coarser particles implies that the surfaces of those particles might present a higher chemical or morphological reactivity than the smaller ones. Further work on the influence of ash particle surface chemistry and matrix components on the reactivity toward gaseous metal species is thus required. Despite its low concentration, Cd could be detected by µ-SRXRF in each single particle. Due to the penetration depth of the high energetic primary and characteristic photons (few hundred µm range), the elemental signals originating from the whole excited depth of the particle are detected simultaneously. This makes the µ-XRF maps, shown in Figures 1, 2, and 5, become two-dimensional (2D) projections of the three-dimensional (3D) distribution of these trace elements throughout the fly ash particles. So, from the µ-XRF maps alone, it is not possible to judge whether the spots of accumulation of the heavy metals are situated on the particle surface (i.e. most prone to leaching) or are present at some depth within the particle (where they might be more shielded from chemical attack by water). Other types of analytical X-rays methods, such as µ-tomography, might also be applicable together with XRF to get more detailed information about the microchemistry of individual fly ash particles. To further understand the speciation of Cd in single fly ash particles, the oxidation state and the chemical surroundings of Cd ions were studied by µ-XAS in the areas of the particles showing a higher Cd concentration. µ-XAS spectra of pure Cd compounds such as Cd, CdO, CdSO4, CdS, CdCl2, CdBr2 pellets were measured prior to the single particle analysis, normalized, and reported in Figure 4. Comparisons of XAS spectra for fly ashes and for reference compounds show that in all the particles studied Cd is present in oxidation state +2 instead of metallic form. The valence state was determined by estimating the position of the K edge from the first inflection point of all the spectra, calculated from the zero of the second derivatives. The +2 oxidation state can be explained by the fact that the conditions during combustion and in the flue gas are generally oxidizing. Bulk
TABLE 2. Best Linear Combination of Standards Fitting the Analyzed Spots particle and hot-spot no.
CdSO4 (%)
CdCl2 (%)
CdO (%)
11 7 9a 9b
10 76 70 70
86 24 11 14
4
CdBr2 (%)
CdS (%)
Cd (%)
19 16
measurements showed that the content of unburned matter in the filtered ash studied was only 0.02 wt % on dry matter (20). The µ-XAS spectra of the hot-spots of the particles were also expressed mathematically as a Linear Combination (LC) of XAS fit vectors, using the measured absorption data of the Cd reference compounds. The µ-XAS spectra of one of the two hot-spots (marked with a ring), within particle 9, is shown in Figure 5. The comparison between the linear combinations of the standard spectra and the measured XAS-spectra of the Cd hot-spots allow for estimation of the concentrations of the possible Cd compounds in those spots, e.g. in the case of the spot a in particle 9, Cd is present as an mixture of CdSO4 (70%), CdO (19%), and CdCl2 (11%) with over 90% confidence level in the fitting process. In all the spots analyzed, Cd was always present as a combination of just CdSO4, CdCl2, and CdO. The set of standards we used is by no means exhaustive. However, at the 90% confidence level, the fit is relevant, and it can serve as a suitable example for this method. Probable causes of disagreement between the fit and the measured spectrum include also the high degree of inhomogeneity of the fly ash grains which may cause local distortions. Furthermore, the assumption that the observed spectrum is a linear combination of the spectra from the separate constituents may not be valid over the whole range of energy and experimental conditions (21). Table 2 reports the quantity of each of those standards giving the best fit in each spot. It can be seen that the Cd in MSW fly ashes is mostly present as water-soluble species, (CdCl2 and CdSO4) up to 86% and 76%, respectively. These results confirm what was found by µ-XRF mapping and agree with earlier studies (19, 22). Such metal speciation would make the management of this filter ash problematic with respect to the possible leaching. On the other hand this is not a conclusive identification because the spectra of other Cd sulfates and some other possibly significant compounds, like Cd silicates, were not examined due to the lack of standard materials. Moreover, we cannot exclude the formation of other larger molecules with the same elements, which would closely fit our experimental spectra. Consequently, the result we present is rather a proof-of-principle than a fully quantitative one. This is a semiquantitative method which has its limitations. Further evaluations using a larger set of standards, as well as more fly ash particles are planned during the next experimental campaign.
The use of µ-XRF and µ-XAS provide valuable new information due to the small spots investigated within single fly ash particles and demonstrates the power of the combined techniques as nondestructive, analytical tools.
Acknowledgments The financial support of the Swedish Natural Science Research Council (NFR) for the work of M.C.C.P. is gratefully acknowledged.
Literature Cited (1) Alba, N.; Gasso´, S.; Lacorte, T.; Baldasano, J. M. J. Air Waste Manage. Assoc. 1997, 47, 1170-1179. (2) Jansses, K.; Vincze, L.; Vekemans, B.; Adams, F.; Kno¨chel, A. J. Anal. Atom. Spectrom. 1998, 13, 339-350. (3) Rindby, A.; Engstro¨m, P.; Janssens, K. J. Synchrotron Radiat. 1997, 4, 228-235. (4) Lankosz, M.; Holynska, B.; Pella, P. A. X-ray Spectrom. 1993, 22, 54-57. (5) Snigirev, A.; Kohn, V.; Snigireva, I.; Lengeler, B. Nature 1996, 384, 49-51. (6) Lengeler, B.; Schroer, C.; Tu ¨ mmler, J.; Benner, B.; Richwin, M.; Snigirev, A.; Snigireva, I.; Drakopoulos, M. J. Synchrotron Radiat. 1999, 6, 1153-1167. (7) Van Espen, P.; Janssens, K.; Nobels, J. Chemom. Intell. Lab. Syst. 1986, 1, 109-114. (8) Nullens, H.; Van Espen, P.; Adams, F. X-ray Spectrom. 1979, 8, 104-109. (9) Teo, B. K. EXAFS: Basic Principles and Data Analysis; Inorganic Chemistry Concepts; Springer-Verlag: Berlin, 1986. (10) Koningsberger, D. C.; Prins, R. X-ray Absorption. Principles, Applications, Techniques of EXAFS, SEXAFS and XANES; John Wiley & Sons: New York, 1988; p 673. (11) CMDV data analysis programme, S. Ansell 2001, http:// www.esrf.fr/computing/scientific/catalog/cmdv/welcome.html (accessed October 1999). (12) Hsiao, M. C.; Paul Wang H.; Huang, Y. J.; Yang, Y. W. J. Synchrotron Radiat. 2001, 8, 931-933. (13) Camerani, M. C.; Somogyi, A.; Drakopolos, M.; Steenari, B. M. Spectrochim. Acta B 2001, 56, 1355-1365. (14) Haynes, B. S.; Neville, M.; Richard, J. Q.; Sarofim, A. F. J. Colloid Interface Sci. 1982, 87(1), 267-278. (15) Beleviand, H.; Moench, H. Environ. Sci. Technol. 2000, 34, 25012506. (16) Somogyi, A.; Janssens, K.; Vincze, L.; Vekemans, B.; Rindby, A.; Adams, F. Spectrochim. Acta B 2000, 55, 1039-1049. (17) Verhulst, D.; Buekens, A.; Spencer, P. J.; Eriksson, G. Environ. Sci. Technol. 1996, 30, 50-56. (18) Cahill, C. A.; Newland, L. W. Int. J. Environ. Anal. Chem. 1982, 11, 227-239. (19) Golding, A.; Bigelow, C.; Veneman, L. M. Chemosphere 1992, 24(3), 271-280. (20) Abbas, Z.; Andersson, B.Å.; Steenari, B. M. Presented at the 15th International Conference on Fluidized Bed Combustion, Savannah, Georgia, May 9-13, 1999; Paper 0104. (21) Wasserman, S. R. J. Phys. IV 1997, 7, C2-203-205. (22) Kirby, C. S.; Rimstidt, J. D. Environ. Sci. Technol. 1993, 27, 652660.
Received for review October 15, 2001. Revised manuscript received April 22, 2002. Accepted April 29, 2002. ES010261O
VOL. 36, NO. 14, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3169
