METHODS FOR SAMPLING AND INORGANIC ANALYSIS OF COAL
U.S. Geological Survey Bulletin 1823

Edited by D.W. Golightly and F.O. Simon

THE DETERMINATION OF MAJOR AND MINOR ELEMENTS IN COAL ASH AND OF CHLORINE AND PHOSPHORUS IN WHOLE COAL BY X-RAY FLUORESCENCE SPECTROMETRY

By R.G. Johnson, G.A. Sellers, and S.L. Fleming, II

Abstract

Methods for the X-ray fluorescence spectrometric analysis of 11 elements in coal ash (Na, Mg, Al, Si, P, S, K, Ca, Ti, Mn, and Fe) and of two elements in whole coal (Cl and P) are described. Coal-ash samples are fused with lithium tetraborate to produce glass disks; whole-coal samples are briquetted with a cellulose binder prior to analysis. Calibration for all elements in both whole coal and coal ash is accomplished by using a simple linear regression of X-ray intensity versus concentration. Silicate rock standards are used for calibration in coal ash analysis, and whole-coal analyses are based on coal standards from the National Bureau of Standards (NBS), on coals analyzed by independent methods, and on synthetic standards. Precision, expressed as the relative standard deviation, is 2 to 5 percent for coal ash and 10 percent for determinations of elements in the whole coal.

INTRODUCTION

X-ray fluorescence (XRF) spectrometry is a rapid technique for accurate and precise elemental analysis of solid and liquid specimens. As such, it is suited to the determination of certain important elements that occur in coal and has been used extensively for the analysis of both whole coal and coal ash (Kuhn and others, 1975; Giauque and others, 1979; Mills and others, 1981; Wheeler, 1983). In practice, most XRF spectrometer systems are not able to detect elements having atomic numbers lower than 11, the atomic number of sodium. Thus, concentrations of the most abundant elements in the coal matrix (C, H, N, and O) must be determined by other methods.

In general, the accuracy of XRF spectrometric determinations of inorganic elements depends on concentration, particle size, matrix effects, surface roughness, the quality of standard materials, and other related factors. The accuracy of analysis usually falls between 2 and 10 percent (relative) depending on how well these factors have been controlled. The lower limit of determination is as low as 1 m g/g for elements having atomic numbers greater than 26, the atomic number of iron, and ranges to several hundred micrograms per gram for aluminum and silicon and to several thousand micrograms per gram for sodium. The upper limits for quantitative XRF spectrometric analysis are determined by the availability of standard reference materials.

This section describes procedures for the analysis of 2 minor elements (Cl and P) in whole-coal samples, and 11 major elements (Na, Mg, Al, Si, P, S, K, Ca, Ti, Mn, and Fe) in the corresponding high-temperature (525° C) coal ash.

EXPERIMENTAL METHODS AND PROCEDURES

Equipment and Supplies

The instruments and reagents required in the XRF spectrometric methods are listed here.

1. For the preparation of a coal-ash fusion disk the following items are required.

2. For the preparation of whole-coal briquettes the following items are required.

3. Standard reference materials can be obtained from

National Bureau of Standards (NBS)
Office of Standard Reference Materials
Room B311, Chemistry Building
Gaithersburg, MD 20899

Coal Ash

Because coal ash represents the dominant mineral portion of coal, XRF spectrometric analysis of this material is very similar to the XRF analysis of silicate rocks (Rose and others, 1963; Norrish and Hutton, 1969). Samples are prepared by fusion with lithium tetraborate flux, and silicate rock standards are used to construct calibration curves for each element. Extensive matrix correction techniques are not required because of the almost eightfold dilution of the sample by the flux. Fusion is the preferred method of preparation because it eliminates particle-size effects and mineralogical heterogeneity and thus allows for the highest precision possible.

A sample is prepared by thoroughly mixing 600 mg of coal ash with 5.400 g of lithium tetraborate in a Pt-Au crucible (95 percent Pt and 5 percent Au). To this mixture is added 5 drops of a 15-percent hydrobromic acid solution, which acts as a nonwetting agent during the fusion. Samples are fused for 20 min by a commercial fluxer (Claisse^R) that produces a temperature in the range of 1,100 to 1,200° C. Alternatively, samples may be fused for 1 h at 1,100° C by a U.S. Geological Survey "in-furnace" device (Taggart and Wahlberg, 1980a), and finally cast in Pt-Au molds (Taggart and Wahlberg, 1980b) to produce glass disks. Caution is advised here. Preliminary evidence indicates that the higher temperature effected by the Meeker burner in the Claisse^R fluxer may volatilize sodium from certain samples during the fusion process.

A Diano^R model XRD-8300 wavelength-dispersive XRF (WDXRF) spectrometer is used to make all intensity measurements; instrumental parameters and operating conditions are listed in table 11.

Calibration of the instrument consists of plotting X-ray intensity versus concentration of the standard reference materials for each element and calculating slope and intercept through simple linear regression analysis. Silicate rock standards are used for every element except sulfur because of their availability and similarity to the composition of most coal ashes. The ranges of concentration (Abbey, 1983) covered by these reference materials for 10 elements (excluding sulfur) are listed in table 12. Results from a typical calibration for iron are shown in table 13 in the form of accepted and calculated concentrations generated by the simple linear regression model.

Because many coal ashes contain much higher levels of sulfur than do the available reference materials, the preparation of separate standards for use in determining sulfur is necessary. To prepare these standards, a typical rock standard is spiked with sulfur and diluted in steps with the unspiked material to produce a set of secondary standards covering the intended concentration range. The mixture is fused with lithium tetraborate flux, and sulfur in the product is determined by an independent method (LECO^R sulfur analyzer) because some sulfur is lost during the fusion process.

The sulfur reference standards used for this work were produced by spiking U.S. Geological Survey basalt BIR-1 with ignited calcium sulfate. Serial dilutions of this mixture with BIR-1 were made to cover the concentration range of 0.1 to 10 percent sulfur. Samples were prepared in duplicate, and one of each was crushed, ground, and analyzed by a LECO^R sulfur analyzer. Calibration for sulfur is accomplished in the same way as the calibrations for other elements.

The XRF spectrometric determinations of major elements in coal ash compare well with other analytical methods. Because the chemistry (or at least the range of composition of the major elements) of coal ash is similar to that of silicate rocks, the accuracy is approximately the same. This accuracy is estimated to be ±2 to 5 percent (relative error) depending on the element and the concentrationn level.

Whole Coal

One advantage of analyzing whole coal is that volatile elements normally lost in ashing or fusion can be determined. The analysis of whole coal may include practically any element, at concentrations approaching 1 m g/g for many. One problem, however, is that calibration standards for XRF spectrometric analysis of whole coal must themselves be whole coals. Only a few coal standards exist, and these are certified for only a few elements. Thus, additional reference materials must be established by analysis of existing coals by other methods or by preparation of synthetic standards.

Specimen preparation normally involves grinding a coal sample with a binder, followed by compression to form a briquette. A low weight-ratio of sample to binder is used because the XRF spectrometric analysis of whole coal usually involves the determination of certain trace elements for which the best possible detection limits are desired. This section describes procedures used in the determination of phosphorus and chlorine in whole coal. Similar procedures could be used for the determination of other elements.

Air-dried coal samples (80 mesh) to be prepared for XRF analysis are further dried for three hours at 105° C. Then, 0.500 g of coal is mixed with an equal portion of microgranular cellulose that serves as a binder. The mixture is transferred to a 26-mL polystyrene vial containing two 6-mm-diameter polycarbonate beads (to aid mixing) and is shaken on a mixer-mill for 10 min to ensure homogeneity. This mixture is pressed against an equal volume of fibrous cellulose (to provide a strong backing) in a hydraulic press at 276 MPa (40,000 psi) for 30 s to produce a 2.54-cm-diameter briquette.

Measurements are made by a Diano^R model XRD-8300 WDXRF spectrometer; instrumental operating conditions are listed in table 14.

Standards used for the determination of chlorine and phosphorus in whole coal include three NBS reference materials, coals that have been analyzed by other methods, and synthetic standards. The NBS coal standards include NBS 1633, 1633a, and 1635. NBS 1633 is no longer available, and NBS 1632b has only recently become available. The NBS coals are used in determinations of both chlorine and phosphorus; concentration values are taken from Germani and others (1980) and from Ondov and others (1975) because the concentrations of these elements are not certified by NBS.

Standards for the determination of chlorine in whole coal consist of a set of coals from three power plants. These coals were analyzed by instrumental neutron activation analysis (R.B. Finkelman, written commun., 1980) for 18 elements, including chlorine. A second set of five coals was the subject of an American Society for Testing and Materials (ASTM) cooperative program investigating XRF analysis of whole coals (ASTM D.05.29.02); although the coal samples were not intended to be distributed as standards, they were extensively analyzed by a variety of methods for the 10 major elements and phosphorus.

Finally, when the available standard reference material is not sufficient to meet analytical needs, it is possible to produce synthetic standards. The use of synthetic standards is the least desirable alternative because of the uncertainty in the composition of the final product due to preparation error and the considerable time and effort required to produce such material.

Phosphorus standards can be produced by thoroughly mixing a sample with a known concentration of phosphorus (for example, NBS 120a, a phosphate rock standard) or a pure compound (Na₃PO₄) with spectrographically pure graphite and then diluting this mixture in steps to give a series of samples that define the concentration range needed. These samples are then mixed with cellulose and briquetted as described previously.

Likewise, chlorine standards are prepared by spiking graphite with a substance such as sodium chloride and then diluting the mixture with graphite in steps to produce a series of standards. Standards for other elements can be prepared in a similar fashion.

Calibration for each element is accomplished by plotting concentration and intensity for all the available standards and calculating slope and intercept by a simple linear regression analysis. Results from a typical calibration for phosphorus are shown in table 15, where accepted and calculated concentrations are given. (The standards Syn-C through Syn-H were prepared synthetically, as described in the text. RM-120 through RM-124 were analyzed by other methods. The value for NBS 1632 is from Ondov and others (1975), and the value for NBS 1632a is from Germani and others (1980).) Although mathematical correction for absorption and enhancement by the other matrix elements may improve results (because minimal dilution is used), in practice, such corrections are difficult to accomplish because of the lack of well-characterized standards. Most matrix correction algorithms require that concentrations for all major and minor elements for the standards be known.

Precision is determined by the sample preparation, instrument stability, and counting time and is generally in the 1- to 2-percent range. Because of the lack of appropriate standard reference materials and because of the uncertainty associated with those materials that are being used as reference standards in the calibrations, assessment of the accuracy of the chlorine and phosphorus determinations is difficult. In addition, there are no other methods for determining either of these two elements now in routine use at the U.S. Geological Survey to provide values for comparison. However, based on the few primary standards available, we estimate the relative precision of the method to be approximately 10 percent.

CONCLUSIONS

Methods based on XRF spectrometry are capable of determining 11 inorganic elements in coal, providing relative precision and accuracy of 2 to 5 percent for the determination of sulfur and major oxides in coal ash, and 10 percent for the determinations of chlorine and phosphorus in whole coal. The most important factor limiting the accuracy of analysis is the restricted availability of standard reference materials. Another significant factor that affects accuracy is that many coal samples not stored in a cool, dry, inert atmosphere decompose over time and, consequently, are subject to changes in composition. In addition, other elements, such as Mg, Al, Si, K, Ca, Ti, Mn, Fe, Cu, Ni, Zn, Rb, Sr, Y, Zr, Nb, and Ba have been determined in whole coal by methods quite similar to those previously described.

REFERENCES

Abbey, Sydney, 1983, Studies in "standard samples" of silicate rocks and minerals, 1969-1982: Geological Survey Canada Paper 83-15, 114 p.

Germani, M.S., Gokmen, Inci, Sigleo, A.C., Kowalczyk, G.S., Olmez, llhan, Small, A.M., Anderson, D.L., Failey, M.P., Gulovali, M.C., Choquette, C.E., Lepel, E.A., Gordon, G.E., and Zoller, W.H., 1980, Concentration of elements in the National Bureau of Standards' bituminous and subbituminous coal standard reference materials: Analytical Chemistry, v. 52, no. 2, p. 240-245.

Giauque, R.D., Garrett, R.B., and Goda, L.Y., 1979, Determination of trace elements in light element matrices by X-ray fluorescence spectrometry with incoherent scattered radiation as an internal standard: Analytical Chemistry, v. 51, no. 4, p. 511-519.

Kuhn, J.K., Harfst, W.F., and Shimp, N.F., 1975, X-ray fluorescence analysis of whole coal, in Babu, S.P., ed., Trace elements in fuel: Washington, D.C., American Chemical Society, p. 66-73.

Mills, J.C., Turner, K.E., Roller, P.W., and Belcher, C.B., 1981, Direct determination of trace elements in coal: X-Ray Spectrometry, v. 10, no. 3, p. 131-137.

Norrish, K., and Hutton, J.T., 1969, An accurate X-ray spectrographic method for the analysis of a wide range of geological samples: Geochimica et Cosmochimica Acta, v. 33, no. 4, p. 431-453.

Ondov, J.M., Zoller, W.H., Olmez, Ilhan, Aras, N.K., Gordon, G.E., Rancitelli, L.A., Abel, K.H., Filby, R.H., Shah, K.R., and Ragaini, R.C., 1975, Elemental concentrations in the National Bureau of Standards' environmental coal and fly ash reference materials: Analytical Chemistry, v. 47, no. 7, p. 1102-1109.

Rose, H.J. Jr., Adler, Isidore, and Flanagan, F.J., 1963, X-ray fluorescence analysis of light elements in rocks and minerals: Applied Spectroscopy, v. 17, no. 4, p. 81-85.

Taggart, J.E. Jr., and Wahlberg, J.S., 1980a, A new in-muffle automatic fluxer design for casting glass disks for X-ray fluorescence analysis: Conference of the Federation of Analytical Chemistry and Spectroscopy Societies, 7th, Philadelphia, paper 327a.

_______ 1980b, New mold design for casting fused samples, in Barrett, C.S., Leyden, D.E., Newkirk, J.B., Predecki, P.K., and Ruud, C.O., eds., Advances in X-ray analysis, v. 23: New York, Plenum Press, p. 257-261.

Wheeler, B.D., 1983, Chemical analysis of coal by energy-dispersive X-ray fluorescence utilizing artificial standards, in Hubbard, C.R., Barrett, C.S., Predecki, P.K., and Leyden, D.E., eds., Advances in X-ray analysis, v. 26: New York, Plenum Press, p. 457-466.

[ Contents | Coal Quality CD ]

Created by the EERT WWW Staff.