XANES Spectroscopic Characterization of Selected Elements in Deep

The authors also acknowledge Element Analysis Corp., Lexington, KY, for the use of their ... Demir, I.; Harvey, R. D.; Ruch, R. R.; Steele, J. D.; Ho,...
0 downloads 0 Views 783KB Size
Download PDF
Energy & Fuels 1997, 11, 691-701

691

XANES Spectroscopic Characterization of Selected Elements in Deep-Cleaned Fractions of Kentucky No. 9 Coal Frank E. Huggins,*,† Srikanta Srikantapura,†,‡ B. K. Parekh,‡ Lori Blanchard,§ and J. David Robertson‡,§ Department of Chemical and Materials Engineering, Center for Applied Energy Research, and Department of Chemistry, University of Kentucky, Lexington, Kentucky 40506 Received July 29, 1996. Revised Manuscript Received January 27, 1997X

The mode of occurrence of various elements in Kentucky No. 9 coal has been examined by XAFS spectroscopic characterization of the elements in float and tailings fractions generated by different flotation tests and chemical leaching methods on the finely ground coal. PIXE spectroscopy was used to determine the elemental concentrations in the same fractions. In this investigation, the elements examined include the lithophile elements, K, Ca, Ti, V, Cr, and Mn, and the predominantly chalcophile elements, Fe and As. The lithophile elements, except for K, exist in two distinct forms: an organically associated form that contributes as much as 50% of the occurrence of Ti, V, and Cr in the coal and a mineralogical form consisting of either illite (K, Ti, V, Cr) or calcite (Ca, Mn). In contrast, the two chalcophile elements examined are associated almost exclusively with pyrite and its oxidation products. Evidence is presented for arsenic as arsenate being incorporated in the major pyrite oxidation product, jarosite. There is a useful synergy in such studies because the more fractions examined with XAFS spectroscopy the better the mode of occurrence is determined, and conversely, the better the mode of occurrence is determined, the better the behavior of elements in flotation and leaching tests can be explained.

Introduction Cleaning of pulverized (-200 mesh) coal is a potentially cost-effective strategy for removing mineral matter and associated minor and trace elements from coal prior to combustion, thereby minimizing the release of pollutants to the atmosphere during combustion. Of special concern are those trace elements designated as potential hazardous air pollutants (HAPs) in the 1990 Amendments to the Clean Air Act.1 Various physical and chemical cleaning methods have been examined with regard to their efficiency for removal of trace elements from coal. These methods have varied from conventional gravity (float/sink) separation to column flotation to chemical extraction methods.2-8 Although the mode of occurrence is recognized to be a key factor in determining the behavior of a given element with respect to a specific cleaning method, no study, as far * Corresponding author address: University of Kentucky, 533 South Limestone Street, Suite 111, Lexington, KY 40506-0043. Phone: (606) 257-4045. FAX: (606) 257-7215. E-mail: [email protected]. † Department of Chemical and Materials Engineering. ‡ Center for Applied Energy Research. § Department of Chemistry. X Abstract published in Advance ACS Abstracts, March 15, 1997. (1) Amendments to the Clean Air Act. U.S. Public Law 101-549; U.S. Government Printing Office: Washington, DC, Nov. 15, 1990; 314 pp. (2) Norton, G. A.; Markuszewski, R. Coal Prep. 1989, 7, 56-68. (3) Akers, D.; Dospoy, R. Fuel Process. Technol. 1994, 39, 73-86. (4) Conzemius, R. J.; Chriswell, C. D.; Junk, G. A. Fuel Process. Technol. 1988, 19, 95-106. (5) Garcia, A. B.; Martinez-Tarazona, M. R. Fuel 1993, 72, 329335. (6) Nowak, M. A.; McLean, V. L. Prepr. Pap.sAm. Chem. Soc., Div. Fuel Chem. 1993, 38 (1), 301-310. (7) Demir, I.; Harvey, R. D.; Ruch, R. R.; Steele, J. D.; Ho, K. K. Prepr. Pap.sAm. Chem. Soc., Div. Fuel Chem. 1994, 39 (2), 530-536. (8) Norton, G. A.; Araghi, H. G.; Markuszewski, R. Prepr. Pap.sAm. Chem. Soc., Div. Fuel Chem. 1985, 30 (2), 58-65.

S0887-0624(96)00118-1 CCC: $14.00

as we are aware, has attempted to include such information. The study that has perhaps come closest is the SEM-AIA investigation of the association of major minerals and elements and its use for predicting cleanabilities of coals carried out by Straszheim and Markuszewski.9 The mode of occurrence is basically a description of how an element occurs in coal. The importance of this information with respect to understanding the behavior of a specific element in coal utilization has often been emphasized (see, for example, recent publications by Swaine,10 Finkelman,11 and Davidson12 ). As summarized in Figure 1 and discussed in more detail elsewhere,13,14 elements in coal can be found in organic association,15 bound to the organic structure by either covalent or ionic bonds, or as ions in the moisture associated with coal, or in inorganic (mineral) association, in which the element may be “captured” by a major mineral or it may form its own discrete accessory (9) Straszheim, W. E.; Markuszewski, R. Energy Fuels 1990, 4, 748754. (10) Swaine, D. J. Trace Elements in Coal; Butterworth: Oxford, U.K.,1990. (11) Finkelman, R. B. Fuel Process. Technol. 1994, 39, 21-34. (12) Davidson, R. M. Trace Elements in Coal. IEA Perspectives Report, IEAPER/21; IEA Coal Research: London, 1996. (13) Huggins, F. E.; Huffman, G. P.; Helble, J. J. In Proceedings, Second International Conference on Managing Hazardous Air Pollutants; EPRI Report TR-104295, Electric Power Research Institute: Palo Alto, CA, 1993; pp II-113-II-133. (14) Huggins, F. E.; Huffman, G. P. In Mineral Spectroscopy: A Tribute to Roger G. Burns; Dyar, M. D., McCammon, C., Schaefer, M. W., Eds.; Special Publication No. 5, The Geochemical Society: Houston, TX, 1996; pp 133-151. (15) In this paper, the term organic association is used merely to indicate that the element occurs with the organic matrix or macerals; it is not meant to imply any specific bonding mechanism between the element and the matrix.

© 1997 American Chemical Society

692 Energy & Fuels, Vol. 11, No. 3, 1997

Figure 1. Classification scheme for elemental modes of occurrence in coal.

mineral. Clearly, where an element is located in Figure 1 will be most important as to how it responds to a particular physical or chemical cleaning method. In this study, a finely ground sample of Kentucky No. 9 coal from Webster County, KY, has been subjected to a combination of physical cleaning and chemical leaching methods. Each generated fraction was then characterized using particle-induced X-ray emission (PIXE) spectroscopy to determine the bulk chemical composition and X-ray absorption fine structure (XAFS) spectroscopy to determine information relevant to the elemental mode of occurrence for selected elements. Inclusion of the mode of occurrence information explains much of the variable efficiency exhibited with respect to the cleaning process by different elements. Conversely, the efficiency of column flotation for removal of mineral matter from clean coal provides samples in which the elements exist almost entirely in organic association. Such samples have proven to be invaluable for generating the XAFS spectral signatures for elements in organic association. Experimental Section (a) Coal Cleaning Studies. A 50 kg sample of as-mined Kentucky No. 9 coal from the Illinois basin region in western Kentucky was obtained from the Big Rivers Utility Co. The coal, originally consisting of 1 in. lumps, was dried and crushed to 28 mesh. A representative kilogram sample of the 28 mesh coal was then ground to 80% passing 60 mesh, from which two head samples of 150 g each were prepared and submitted for analysis. Samples were then prepared for gravity separation testing and release analysis, the results of which are described elsewhere.16,17 A representative sample was then ground to 90% passing 325 mesh to constitute the fine coal sample. Aliquots of this sample were subjected to separation by flotation in a Denver cell and also in a 2 in. diameter “Ken-Flote” column18 into float and tailings fractions. The float fraction from the Denver cell was then used as feed for hydrothermal leaching tests that were conducted in a 1 L autoclave at 110 °C and 300 psi. Four (16) Srikantapura, S. Removal of Trace Elements from Coal by Advanced Physical and Chemical Techniques. M.Sc. Thesis, University of Kentucky, 1996. (17) Parekh, B. K.; Srikantapura, S. Unpublished data, paper in preparation. (18) Parekh, B. K.; Stotts, W. F.; Groppo, J. G. In Processing and Utilization of High-Sulfur Coals III, Coal Science and Technology; Elsevier: Amsterdam, 1990; Vol. 16, pp 197-208.

Huggins et al.

Figure 2. Scheme for physical and chemical cleaning tests performed on the as-mined sample of Kentucky No. 9 coal to generate float, tailings, and leached fractions. samples were prepared: samples were exposed for 15 and 60 min to either a 10% acid (H3PO4) or a 10% basic (NaOH) medium. The overall sample scheme and sample labeling notation are shown in Figure 2. Sufficient samples of each fraction were obtained for proximate analysis, chemical analysis by PIXE spectroscopy, and element speciation analysis by XAFS spectroscopy. (b) PIXE Spectroscopic Analysis. Each sample generated in the experimental scheme shown in Figure 2 was first dried at 50 °C overnight and then pressed into a 1 mm × 19 mm pellet for elemental analysis by particle-induced X-ray emission (PIXE) spectroscopy. To enhance the sensitivity of PIXE for both low- and high-Z elements, a dual-energy irradiation is performed on each sample with the X-ray detector in two positions for data collection. During the highenergy (2.1 MeV) irradiation, the detector is in a close-in position with a thick absorber, and during the low-energy (1.6 MeV) irradiation, the detector is in a backed-out position with no absorber. Variable charge collection at these two energies/ positions allows for spectrum balance and flexibility in the analysis on either the high- or low-energy ends of the X-ray spectrum. Protons enter the target chamber by passing through a 0.30 mil Kapton window and the X-rays exit through a 0.1 mil Mylar window that is at 45° relative to the beam. The beam, which is at an angle of 23° relative to the sample surface, is swept over the target to irradiate a 16 mm diameter area. The sample chamber is flushed with helium at atmospheric pressure to reduce sample heating and charging and each sample is irradiated for 15 min. Data analysis is performed using a modified version of the GUPIX19 PC-based software package. In order that the accuracy and precision of the measurements may be assessed, the results for the PIXE analysis of seven samples of the NIST 1635a Subbituminous Coal standard reference material are presented in Table 1. As can be seen from this table, the typical experimental error is less than (10% in most instances. Further details of the technique, including examples of PIXE spectra, are presented elsewhere.20 (c) XAFS Spectroscopic Analysis. XAFS spectroscopy of selected elements in the samples generated according to the scheme shown in Figure 2 was carried out either at the National Synchrotron Light Source (NSLS) in Brookhaven National Laboratory, NY, or at the Stanford Synchrotron Radiation Laboratory (SSRL) at Stanford University, CA. Similar experimental procedures were used at both of these facilities; these procedures have been described in detail in previous publications.14,21-23 The XAFS spectra from elements (19) Maxwell, J. A.; Campbell, J. L.; Teesdale, W. J. Nucl. Instrum. Meth. 1989, B43, 218-230. (20) Wong, A. S.; Robertson, J. D. J. Coal Qual. 1993, 12 (4), 146150. (21) Huggins, F. E.; Shah, N.; Zhao, J.; Lu, F.; Huffman, G. P. Energy Fuel 1993, 7, 482-489.

Elements in Kentucky No. 9 Coal

Energy & Fuels, Vol. 11, No. 3, 1997 693

Table 1. Results for NIST SRM 1635a Subbituminous Coal element

PIXE resulta

certified valueb

sodium magnesium aluminum silicon sulfur potassium calcium titanium chromium manganese iron nickel copper zinc selenium bromine strontium

0.21 ( 0.02% 760 ( 71 ppm 0.30 ( 0.02% 0.60 ( 0.03% 0.39 ( 0.02% 113 ( 11 ppm 0.69 ( 0.02% 213 ( 8 ppm 1.2 ( 0.9 ppm 19.8 ( 1.2 ppm 0.23 ( 0.01% below LOD 3.1 ( 0.2 ppm 4.6 ( 0.9 ppm below LOD 1.3 ( 0.3 ppm 129 ( 9 ppm

(0.24 ( 0.02%) (1040 ( 130 ppm) (0.29 ( 0.03%) (0.59 ( 0.05%) 0.33 ( 0.03% (96 ( 16 ppm) (0.54 ( 0.03%) (202 ( 6 ppm) 2.5 ( 0.3 ppm 21.4 ( 1.5 ppm 0.239 ( 0.005% 1.74 ( 0.10 ppm 3.6 ( 0.3 ppm 4.7 ( 0.5 ppm 0.9 ( 0.3 ppm (1.4 ( 0.4 ppm) (121 ( 19 ppm)

a Average and standard deviation of the analysis of seven samples. LOD ) limit of detection. b Values in parentheses are the recommended or consensus values.

in coal were collected in fluorescence geometry using a 13element germanium detector developed principally for trace element investigation.24 Spectra were obtained typically from 100 eV below the K-edge of the element under investigation to as much as 500 eV above the edge. As is usually done, the XAFS spectra were subdivided into separate X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) regions. In many instances, particularly when the concentration of the element was less than about 50 ppm, the EXAFS signal was too weak to be reliably interpreted and interpretation of the mode of occurrence of an element was based exclusively on the XANES region. The XANES spectra shown in this paper have been prepared from the raw XAFS spectroscopic data in the usual manner: 25,26 the spectra are first normalized to the edge step and any slope in the data is removed by fitting spline functions to the pre-edge and post-edge regions and subtracting the pre-edge background extended beneath the overall spectrum. Each spectrum shows a zero point of energy that is defined by the position of the same absorption edge in a standard material, the spectrum of which is collected at the same time as the spectrum from the coal. For the elements discussed in this paper, these zero points are defined in standard materials as follows: potassium (3608.4 eV, KCl), calcium (4038.1, calcite CaCO3), titanium (4966 eV, Ti metal foil), vanadium (5465 eV, V metal foil), chromium (5989 eV, Cr in stainless steel), manganese (6539 eV, Mn metal foil), arsenic (11867 eV, As2O3). Inflection points (first-derivative spectral maxima) were used for all elemental standards, except that for arsenic, which has a prominent white line that is sufficiently sharp to be used directly as a calibration point. The analysis of the XANES data for most elements is based largely on qualitative comparison of the spectra of the coal fractions with those for the element in various mineral and chemical forms. For most elements, there is a wealth of published XANES data for common mineralogical occurrences and chemical standards. In particular, the reference data we have generated for most of the spectral data shown in this (22) Huffman, G. P.; Huggins, F. E.; Shah, N.; Zhao, J. Fuel Process. Technol. 1994, 39, 47-62. (23) Huggins, F. E.; Huffman, G. P. Int. J. Coal Geol. 1996, 31, 3153. (24) Cramer, S. P.; Tench, O.; Yocum, N.; George, G. N. Nucl. Instrum. Meth. 1988, A266, 586-591. (25) Koningsberger, D C.; Prins, R. X-ray Absorption. Principles, Applications, Techniques of EXAFS, SEXAFS, and XANES; J Wiley & Sons: New York, 1988. (26) Eisenberger, P.; Kincaid, B. M. Science 1978, 200, 1441-1447. Lee, P. A.; Citrin, P. H.; Eisenberger, P.; Kincaid, B. M. Rev. Mod. Phys. 1981, 53, 769-808.

paper has been summarized in a number of recent papers.14,21-23 Other references were consulted for additional information on minerals, both of a general nature,27 and specifically for Ti,28,29 V,29-31 Cr,32 and Mn.33 A number of additional procedures, such as spectral differentiation and comparison of spectra with those of different but geochemically similar elements, have been used to enhance the validity of the spectral interpretation in the course of this investigation. Semiquantitative estimation of the relative contributions of different elemental forms in an XANES spectrum, including uncertainty estimation, is based on simulation of the experimental spectra by weighted summation of spectra for specific standards. A detailed example of this procedure is presented below for manganese. For arsenic, a more precise method could be used because the two major arsenic forms give rise to well-resolved peaks that can be fitted according to a calibrated least-squares method.22,34 57 Fe Mo¨ssbauer spectroscopy, rather than Fe XAFS spectroscopy, was used to examine the different forms of iron in some of the fractions shown in Figure 2 because this method is more definitive for elucidating details regarding the iron minerals present in coal. A Mo¨ssbauer spectrometer, consisting of a constant-acceleration drive (Halder, GmbH) interfaced and synchronized to the opening of channels in a MCA add-in board in a 286DX PC, was used to generate and record the spectra. Mirror image spectra of iron in the coal fractions were collected over 1024 channels representing the velocity interval from -4 to +4 mm/s and return. The source of γ-rays was 57Co in a palladium foil held in a plastic holder at one end of the Mo¨ssbauer drive; the original activity of the source obtained from DuPont Radiochemicals (formerly New England Nuclear) was 50 mCi. The spectra of the coal fractions were calibrated using the four inner lines in the Mo¨ssbauer spectrum of metallic iron that was collected simultaneously from a second source and spectrometer setup at the opposite end of the drive. Analysis of the Mo¨ssbauer spectral data follows that described previously35,36 and involves least-squares fitting of the spectrum envelope as the sum of quadrupole doublets or individual peaks based on a Lorentzian peak shape. Isomer shift and quadrupole splitting values derived from the peak positions are then used to identify the iron-bearing minerals or phases that give rise to specific absorption features in the spectrum. Pyritic-sulfur values were derived from the areas under the peaks attributed to pyrite, following the method described earlier.35,36

Results (a) Analytical Data. Results from proximate and other coal characterization tests on the head sample of the Kentucky No. 9 coal are summarized in Table 2. On the basis of the moisture content and volatile matter (27) Brown, G. E., Jr.; Calas, G.; Waychunas, G. A.; Petiau, J. In Spectroscopic Methods in Mineralogy and Geology; Hawthorne, F. C., Ed.; Reviews in Mineralogy, 18; Mineralogical Society of America: Washington, DC, 1988; Chapter 11, pp 431-512. (28) Waychunas, G. A. Am. Mineral. 1987, 72, 89-101. (29) Wong, J.; Maylotte, D. H.; Lytle, F. W.; Greegor, R. B.; St. Peters, R. L. In EXAFS and Near Edge Structure; Bianconi, A., Incoccia, L., Stipcich, S., Eds.; Springer Series in Chemical Physics, 27; Springer-Verlag: Berlin, 1983; pp 206-209. (30) Maylotte, D. H.; Wong, J.; St. Peters, R. L.; Lytle, F. W.; Greegor, R. B. Science 1981, 214, 554-556. (31) Maylotte, D. H.; Wong, J.; St. Peters, R. L.; Lytle, F. W.; Greegor, R. B. In Proceedings, International Conference on Coal Science; International Energy Agency, Eds.; Verlag Gluckauf Gmbh: Essen, Germany, 1981; pp 756-761. (32) Calas, G.; Manceau, A.; Novikoff, A.; Boukili, H. Bull. Mineral. 1984, 107, 755-766. (33) Manceau, A.; Gorshkov, A. I.; Drits, V. A. Am. Mineral. 1992, 77, 1133-1143. (34) Huggins, F. E.; et al. Work in progress. (35) Huffman, G. P.; Huggins, F. E.; Dunmyre, G. Fuel 1978, 57, 241-250. (36) Huggins, F. E.; Huffman, G. P. In Analytical Methods for Coal and Coal Products; Karr, C., Jr., Ed.; Academic Press: New York, 1979; Vol. III, Chapter 50, 471-523.

694 Energy & Fuels, Vol. 11, No. 3, 1997

Huggins et al.

Table 2. Analytical Data for Kentucky No. 9 Head Sample analysis

as received

moisture, wt % ash, wt % volatile matter, wt % fixed carbon, wt % calorific value (BTU/lb)

dry, ash-free

10.27 18.29 29.84 41.60

s s 41.77 58.23

10 304

14 423

total sulfur, wt % pyritic sulfur, wt % sulfate sulfur, wt % organic sulfur, wt %

2.96 1.60 0.10 1.26

Table 3. PIXE Analytical Data for Head, Float and Tailings Samples from Denver Cell and “Ken-Flote” Column Flotation Tests sample

wt fraction, wt % ash content, wt % sulfur, wt % elements, ppm Na Mg Al Si S Cl K Ca Ti V Cr Mn Fe Ni Cu Zn Ga As Br Sr

head

DC-1

CF3-1

CF3-2

100.00 18.29 2.96

72.33 5.69 2.25

DC-2 27.67 50.88 5.49

83.17 3.56 2.94

16.83 84.20 6.29

5010 5392 19463 45258 29095 3632 4181 2597 1249 159 40 37 18173 21 8 47 7 10 22 41

2523 2381 8479 15197 22446 1113 1418 1592 931 59 31 43 13465 25 17 20 5 14 26 21

s 10006 57138 129760 54905 s 10019 34524 2430 187 84 491 78278 67 47 154 11 100 27 75

4233 3456 11705 17462 19432 1529 1135 1107 711 36 19 21 8612 19 12 13 3 10 12 15

5852 20743 103127 222744 62873 244 15962 41068 2797 133 98 401 55967 55 44 108 10 50 12 23

release, the coal is estimated to be of high volatile B bituminous rank. Coal characterization data and PIXE analysis data on elemental concentrations are summarized for the different float and tailings fractions in Table 3. The two flotation methods are both efficient at separating the inorganic matter from organic matter, but the column flotation test is definitely the better of the two methods. This is also illustrated in Figure 3, which provides a direct comparison of the two methods. In general, the float/tailings difference indicators in this figure are longer for the column flotation method than for the Denver cell method for most of the elements. It is also instructive to examine the tailings/float concentration ratio for individual elements (or equivalently comparing the length of the indicators in Figure 3) to that for the overall ash. The elements that exceed or equal the overall ash ratio are Ca and Mn; those that are slightly less than the overall ash ratio are Al, Si, K, and Zn in both tests and Fe along with As in the case of the Denver cell test. Not surprisingly, sulfur shows a small value for the ratio because over 40% of the element is organically bound (q.v. Table 2). However, there is a group of elements consisting of Ti, V, Cr, and Ni that also have relatively short indicators, suggesting that significant fractions of these elements might also be in organic association. The mass balances for individual elements in these flotation experiments are mostly acceptable; however,

the balances for a number of elements are much larger than can be attributed to experimental errors in the PIXE analysis (q.v. Table 1). The most significant problem would appear to be segregation of the heavy minerals (pyrite, calcite) in the sampling as the poorest mass balances and largest variations between the flotation experiments are exhibited by calcium, manganese, and iron. In particular, the calculated feed composition for the column flotation experiment has significantly lesser amounts of these elements and somewhat more Al and Si (derived from excess clays?) than that for the Denver cell flotation test. Repeat PIXE analysis of different aliquots of the head sample confirmed that the elements (Ca, Mn, Fe) associated with the heavy minerals did exhibit the greatest variation among the analyses. (b) XAFS Spectroscopic Data. Various elements (Ti, V, Cr, Mn, As) have been examined by XAFS spectroscopy in a number of the fractions of Kentucky No. 9 coal generated according to the scheme shown in Figure 2. In addition, Mo¨ssbauer spectroscopy was used to examine the partitioning of iron forms between float and tailings fractions. Each element will be discussed individually. (i) Titanium. The titanium XANES spectra of the float and tailings fractions are shown in Figure 4. It is readily apparent that the Ti XANES of the float fractions are quite distinct from those of the tailings fractions, regardless of the method of separation. A similar, although perhaps less clear-cut, difference in the XANES spectra was noted previously by Maylotte et al.31 for gravity-separated fractions of Kentucky No. 9 coal. Based upon comparison of the XANES spectra of the fractions with those of mineralogical standards,23,28 the Ti XANES spectra of the tailings fractions are most similar to that for Ti present in the clay mineral, illite. However, the XANES spectra of the float fractions cannot be identified as arising from a specific Ti mineral or combination of minerals; further, any possible contribution of one of the TiO2 polymorphs or Ti/illite to these float XANES spectra cannot exceed 20%. This XANES signature for the float fractions of the Kentucky No. 9 coal is similar to those noted for a fraction of an Illinois No. 6 coal cleaned by selective oil agglomeration and a whole coal spectrum of the subbituminous Wyodak-Anderson coal from the Argonne Premium coal program.23 As shown in Figure 5, the XANES spectrum of the float fraction appears unaltered by treatment with acid (10% phosphoric acid) but is different after treatment with alkali (10% NaOH). However, the concentration of Ti in this fraction is not greatly lowered (Table 4) compared to the acid-leached fraction. We interpret these observations as indicating that the organically associated Ti is present predominantly in a single form that is unaffected by the acid treatment, but which is altered, although not leached, by the NaOH treatment. Hence, there appears to be two distinct Ti forms in the Kentucky No. 9 coal: Ti/ illite and an organically associated form. Since the Ti XAFS spectra of the float and tailings fractions are quite different, it is clear that the tailings fraction contains predominantly Ti as illite, whereas the float fraction contains Ti in organic association. However, we estimate conservatively that it would be possible for up to 20% of the second Ti form to go

Elements in Kentucky No. 9 Coal

Energy & Fuels, Vol. 11, No. 3, 1997 695

Figure 3. Comparison of the relative efficiency of the segregation of elements in Kentucky No. 9 coal between float and tailings fractions generated in the Denver cell test and in a column flotation test. The top and bottom levels of the indicator for each element indicate its concentrations in the tailings and float fractions, respectively.

Figure 4. Comparison of Ti XANES spectra from float and tailings fractions separated from Kentucky No. 9 coal in three different flotation tests.

undetected in the XANES spectra of a fraction dominated by the presence of the other form. Hence, the results summarized in Table 5 for the distribution of Ti in the two forms between the float and tailings fractions are conservative estimates of the efficiency of the separation of these two Ti forms in these flotation processes. (ii) Vanadium. As was observed for titanium, the XANES spectra of vanadium in the float and tailings fractions of the Kentucky No. 9 are significantly different (Figure 6) and indicative of two distinct forms of vanadium present in the Kentucky No. 9 coal. Very similar XANES spectra and results were reported in 198129-31 on heavy-liquid separated float/sink fractions from a vanadium-enriched horizon of the Kentucky No. 9, in which the vanadium content was as high as 1850 ppm. In particular, the variation of the height of the pre-edge feature at about 5 eV should be noted since this feature correlates27,29,30 strongly with the oxidation state of vanadium and indicates that the vanadium

Figure 5. Comparison of Ti XANES spectra from acid- and base-leached samples of the Denver cell float fraction.

oxidation state is higher in the float fractions than in the tailings fractions. The spectra of the tailings fractions is similar to that reported for V3+ in roscoelite,29,30 a vanadium-rich variety of muscovite. Illite has a similar structure to that of muscovite, and it is likely that the XANES spectrum of vanadium in illite would also be quite similar. On the basis of an analysis of the EXAFS region, Wong et al.29,30 proposed that the vanadium form in the Kentucky No. 9 float fraction consisted of V4+ in a distorted octahedral environment of oxygen anions not unlike that in V2O4. Hence, as summarized in Table 5 for the distribution of vanadium forms between the float and tailings fractions, the results are similar to that found for titanium. (iii) Chromium. As shown in Figure 7, the differences between the chromium XANES spectra of float

696 Energy & Fuels, Vol. 11, No. 3, 1997

Huggins et al.

Table 4. Trace Element Concentrations in Leached Samples sample elements, ppm

DC-1 (feed)

AL-1 15 min

AL-2 60 min

BL-1 15 min

BL-2 60 min

Na Mg Al Si S Cl K Ca Ti V Cr Mn Fe Ni Cu Zn Ga As Br Sr

2523 2381 8479 15197 22446 1113 1418 1592 931 59 31 43 13465 25 17 20 5 14 26 21

2718 1922 7064 13098 24590 1167 1085 479 764 35 22 s 5797 18 9 11 3 4 11 18

s 1493 6100 11934 24769 1161 898 411 723 35 28 s 5617 19 11 7 2 7 18 18

72009 s 7287 13321 22982 710 1348 2918 723 23 26 33 10645 29 33 13 s 9 9 18

79887 s 5061 10297 22711 478 1150 2518 710 s 31 29 10428 24 26 10 3 7 6 12

Figure 7. Comparison of Cr XANES spectra and first derivative Cr XANES spectra from float and tailings fractions separated from Kentucky No. 9 coal in the Denver cell flotation tests.

Table 5. Estimated Percentages of Elemental Forms for Selected Lithophile Elements in Kentucky No. 9 Coal Fractions from XAFS Spectroscopy Denver cell

column flotation

form

DC-1

DC-2

CF3-1

CF3-2

% Ti as illite % Ti as org. assoc. %V as illite %V as org. assoc. %Cr as illite %Cr as org. assoc. %Mn as illite %Mn as calcite %Mn as acid insol.a %Mn as base insol.a