The Chemical Analysis of Argonne Premium Coal Samples

Edited by Curtis A. Palmer
U.S. Geological Survey Bulletin 2144

Determination of Major and Trace Elements in Eight Argonne Premium Coal Samples (Ash and Whole Coal) by X-Ray Fluorescence Spectrometry

By John R. Evans, George A. Sellers, Robert G. Johnson, Davison V. Vivit, and Judy Kent

ABSTRACT

X-ray fluorescence (XRF) spectrometric methods were used in the analysis of eight Argonne Premium Coal samples. Trace elements (Cr, Ni, Cu, Zn, Rb, Sr, Y, Zr, Nb, Ba, La, and Ce) in both coal ash and whole coal were determined by energy-dispersive X-ray fluorescence spectrometry. Major elements (Na, Mg, Al, Si, P, S, K, Ca, Ti, Mn, and Fe) in coal ash and trace elements (Cl and P) in whole coal were determined by wavelength-dispersive X-ray fluorescence spectrometry. The experimental XRF methods and procedures used to determine these major and trace elements are described.

INTRODUCTION

Energy-dispersive X-ray fluorescence (EDXRF) spectrometry and wavelength-dispersive X-ray fluorescence (WDXRF) spectrometry are used routinely in the determination of major and trace elements in silicate rocks (Norrish and Hutton, 1969; Johnson, 1984); however, the analysis of whole coals by XRF spectrometric techniques is more difficult because of the problem of the very light coal matrix and the scarcity of reliable coal standards. Because coal ash is more similar to silicate matrix rocks, EDXRF and WDXRF techniques developed for silicates can be used for the determination of major and trace elements in coal ash samples.

The rapidity, sensitivity, accuracy, and precision of X-ray fluorescence spectrometric methods are well documented for a wide range of geologic materials (Rose and others, 1963; Norrish and Hutton, 1969; Johnson, 1984; Johnson and others, 1986; Johnson and Fleming, 1987; Evans and Jackson, 1989). Analysis of whole coal by XRF spectrometric techniques has also proven to be successful in many studies (Kuhn and others, 1975; Johnson and others, 1989). Therefore, the determinations of major and minor elements in eight Argonne Premium Coal samples by XRF spectrometric techniques contributed an important part of the geochemical data base compiled for these materials. This study was not intended to include interpretations of the differences of behavior between various coal ranks of the samples studied.

EXPERIMENTAL

Coal Ash-EDXRF

All the Argonne Premium Coal samples were first ashed at 525°C. This is a lower temperature than prescribed by the American Society for Testing and Materials (ASTM, 1996) method (750°C); however, our method eliminated all combustible material while retaining the same or higher concentrations of volatile material. Sample preparation of coal ash samples followed procedures described in other publications (Johnson, 1984; Johnson and others, 1986; Evans and Jackson, 1989). A Kevex 700 EDXRF spectrometer with a Kevex 8000 analyzer was used to fluoresce coal ash samples powdered to approximately 100 mesh. These powders were pressed into cups made of Mylar film (6.35 mm) pulled tightly over an aluminum ring with a Teflon collar. The resultant surface appears to be planar.

Appropriate secondary targets were used (table 1). Each sample was fluoresced, and intensity measurements were determined after making background and spectral overlap corrections. The ratio of the analyte line intensity to the secondary target Compton scatter intensity was used in determining elemental concentrations. The Compton ratio method corrects for matrix effects, particle size variations, packing density variations, heterogeneity effects, instrumental fluctuations, and other sources of error inherent in EDXRF determinations. Trace-element concentrations for coal ash samples were determined from calibration graphs that were constructed by plotting intensity ratio versus the known concentrations for a selected set of standard reference materials (Abbey, 1983).

Whole Coal-EDXRF

Whole-coal samples were prepared by using procedures similar to those described for EDXRF analyses of the coal ash (Johnson, 1984; Johnson and others, 1986; Evans and Jackson, 1989). All intensity measurements were made on a Kevex 700 spectrometer with a Kevex 7000 analyzer. Each whole-coal sample was fluoresced using a secondary target (table 1). Corrections for background interferences and spectral line overlaps were made before integration of the analyte line intensity. Trace elements in whole coal samples were determined by EDXRF by using interelement influence coefficients calculated from fundamental parameters (Johnson and Fleming, 1987). Characterizations of the coal samples by other analytical techniques must be made before trace-element determinations can be obtained with this method. Even though carbon, hydrogen, nitrogen, and oxygen constitute the largest percentage of the whole coal, these elements have very little bearing on absorption and enhancement effects. Major-element concentrations, as determined from the coal ash (see the next section, 'Coal Ash-WDXRF'), identify the most important influences on absorption and enhancement effects necessary to generate accurate interelement influence coefficients from the fundamental parameters algorithm.

The complexities of the fundamental parameters algorithms used in this study are beyond the scope of this paper. Detailed explanations of all equations and variables inherent in the matrix correction procedures were given by Sherman (1959), Rousseau (1984a,b), and Johnson and Fleming (1987).

The lack of a sufficient number of whole-coal standards and the ultimate degradation over time of these standards are major difficulties involved in the characterization of coals. For these reasons, it is not possible to construct routine calibration graphs of standard reference materials for elements of interest; therefore, we must use the fundamental parameters algorithm. Using this algorithm allows the investigator to make accurate trace-element determinations in whole coal with as few as one well-characterized standard. In this study, the National Institute of Standards and Technology (NIST), formerly the National Bureau of Standards (NBS), whole-coal reference materials, NIST 1632a and NIST 1632b (NBS, 1978a, 1985) were used to calculate pure element intensities.

Coal Ash-WDXRF

The fusion method was used to produce glass disks of coal ash samples (Johnson and others, 1989). This method eliminates the need for matrix correction routines, since the significant dilution of the sample by the flux corrects for heterogeneity effects, particle size variations, and other sources of error from instrumental fluctuations. A sample/flux ratio is chosen to yield linear calibration curves over the range of concentrations found in both samples and standards, without the use of a heavy absorber, such as La₂O₃ . This sample/flux ratio is needed because of the low final concentrations of sample components in the sample/flux mix.

A 1:9 dilution of sample to flux is obtained by mixing 0.600 g of the coal ash with 5.400 g of a 2:1 mixture of lithium tetraborate to lithium metaborate. This mixture is carefully transferred to a platinum-gold crucible, and three drops of a 15 percent hydrobromic acid solution are added as a wetting agent. An automatic Claisse fluxer is used to heat/mix the sample to temperatures reaching 1,200°C for approximately 20 minutes. After the sample cools to room temperature, a thin glass disk with a planar analytical surface is produced, which is adequate for WDXRF analysis. Elemental intensity measurements are made on a Diano XRD-8300 wavelength-dispersive X-ray fluorescence spectrometer. Standards used in the construction of calibration graphs were silicate matrix materials selected from those tabulated by Abbey (1983). Because the coal ash matrix closely resembles silicate materials in composition, calibration graphs obtained from silicate standard reference materials are reliable for major-element determinations in coal ash. Standards are prepared for WDXRF analysis in a manner identical to that described above. Calibration graphs were constructed by plotting the analyte intensity with the known concentration for a selected set of standard reference materials for each element of interest. The intensities for the major elements in the coal ash samples were then used in the individual calibration graphs.

A set of synthetic silicate standards was spiked with sulfur before fusion because the chemical matrix of typical silicate rock standards does not have sulfur concentrations similar to those in the coal ash matrix. Because some sulfur is volatilized during fusion, a portion of the fused standard was analyzed by a LECO sulfur analyzer to determine the actual sulfur concentration in the standard. The sulfur determinations of the standards were used to prepare calibration graphs like those described above for the silicate matrix materials tabulated by Abbey (1983).

Whole Coal-WDXRF (Determination of Cl and P)

Briquettes of the whole-coal samples were produced by mixing 0.500 g of the coal with 0.500 g of microgranular cellulose for 10 minutes on a shaker mill and subsequently pressing the mixture against a fibrous cellulose backing at 276 MPa for approximately 30 seconds (Johnson and others, 1989).

The difficulties experienced in the analysis of whole coals by EDXRF also apply for WDXRF. Reliable whole-coal standard reference materials are scarce. Because these standards are not commercially available, synthetic standards as well as coal samples characterized by other laboratories were used. Only three NIST coal standards were used in this study: NIST 1633, 1633a, and 1635 (NBS, 1975, 1979, and 1978b). Spiked graphite samples with varying concentrations of chlorine and phosphorus served as the synthetic whole-coal standards. All standards were prepared identically to those for the whole-coal samples. Intensity measurements for chlorine and phosphorus were made on a Diano XRD-8300 wavelength-dispersive X-ray fluorescence spectrometer.

Calibration graphs were constructed by plotting the analyte intensity versus the known concentration for a set of standards. The intensities for chlorine and phosphorus in the whole-coal samples were then used to calculate chlorine and phosphorus concentrations from the regression curves.

RESULTS AND DISCUSSION

In this study, eight Argonne Premium Coal samples were analyzed by EDXRF and WDXRF spectrometry. Determinations of major oxides in coal ash are detailed in table 2; trace elements in coal ash in table 3; chlorine and phosphorus oxide in whole coal in table 4; and trace elements in whole coal in table 5. The precision and accuracy for the analysis of coal ash samples by EDXRF and WDXRF closely approximate the precision and accuracy for the analysis of silicates. A study by Johnson and others (1989) estimated an average relative difference of ±2 to ±5 percent for WDXRF determinations of major elements (Na, Mg, Al, Si, P, S, K, Ca, Ti, Mn, and Fe) in coal ash samples. Trace-element determinations (Cr, Ni, Cu, Zn, Rb, Sr, Y, Zr, Nb, Ba, La, and Ce) by EDXRF for silicate rocks were estimated to have an accuracy of <±5 percent for the ratio-calibration graph method (Johnson, 1984). This level of accuracy is also expected for EDXRF trace-element determinations on coal ash.

The precision and accuracy of the EDXRF and WDXRF analyses of whole-coal samples were more difficult to estimate, since a wide range of acceptable standards was not available. However, Johnson and others (1989) estimated the average relative difference for chlorine and phosphorus oxide determinations on whole coals to be ±10 percent. Trace-element (Cr, Ni, Cu, Zn, Rb, Sr, Y, Zr, Nb, Ba, La, and Ce) determinations on the whole coal by EDXRF generally show close agreement (±10 percent) between replicate samples. A wide variance was noted, however, when the whole-coal trace-element results were compared with the results obtained on the coal ash. Further investigation is needed to evaluate more clearly the accuracy of the matrix correction method for whole coals.

REFERENCES

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

American Society for Testing and Materials (ASTM), 1996, ASTM Designation D 3174-93, Standard test method for ash in the analysis sample of coal and coke from coal: 1996 Annual Book of ASTM Standards, v. 05.05, Gaseous fuels; Coal and coke, p. 291-294.

Evans, J.R., and Jackson, J.C., 1989, Determination of tin in silicate rocks by energy-dispersive X-ray fluorescence spectrometry: X-Ray Spectrometry, v. 18, p. 139-141.

Johnson, R.G., 1984, Trace element analysis of silicates by means of energy-dispersive X-ray spectrometry: X-Ray Spectrometry, v. 13, no. 2, p. 64-68.

Johnson, R.G., and Fleming, S.L., II, 1987, Energy-dispersive X-ray fluorescence analysis of massive sulfides using fundamental influence coefficients: X-Ray Spectrometry, v. 16, p. 167-170.

Johnson, R.G., Palmer, C.A., Dennen, K.O., and Hearn, P.P., 1986, Energy-dispersive X-ray fluorescence analysis of trace elements in carbonate rocks: Applied Spectroscopy, v. 40, no. 1, p. 76-79.

Johnson, R.G., Sellers, G.A., and Fleming, S.L., II, 1989, The determination of major and minor elements in coal ash and of chlorine and phosphorus in whole coal by X-ray fluorescence spectrometry, in Golightly, D.W., and Simon, F.O., eds., Methods for sampling and inorganic analysis of coal: U.S. Geological Survey Bulletin 1823, p. 35-39.

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.

National Bureau of Standards, 1975, National Bureau of Standards certificate of analysis, standard reference material 1633, trace elements in coal fly ash: Washington, D.C., National Bureau of Standards, 2 p.

------978a, National Bureau of Standards certificate of analysis, standard reference material 1632a, trace elements in coal (bituminous): Washington, D.C., National Bureau of Standards, 2 p.

------1978b, National Bureau of Standards certificate of analysis, standard reference material 1635, trace elements in coal (subbituminous): Washington, D.C., National Bureau of Standards, 2 p.

------1979, National Bureau of Standards certificate of analysis, standard reference material 1633a, trace elements in coal fly ash: Washington, D.C., National Bureau of Standards, 2 p.

------1985, National Bureau of Standards certificate of analysis, standard reference material 1632b, trace elements in coal (bituminous): Gaithersburg, Md., National Bureau of Standards, 5 p.

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.

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.

Rousseau, R.M., 1984a, Fundamental algorithm between concentration and intensity in XRF analysis. 1-Theory: X-Ray Spectrometry, v. 13, no. 3, p. 115-120.

------1984b, Fundamental algorithm between concentration and intensity in XRF analysis. 2-Practical application: X-Ray Spectrometry, v. 13, no. 3, p. 121-125.

Sherman, J., 1959, Research note-Simplification of a formula in the correlation of fluorescent X-ray intensities from mixtures: Spectrochimical Acta, v. 15, no. 6, p. 466-470.

[ Contents ]

Created by the EERT WWW Staff.