Appendix 1. Comparison of the concentrations of 51 elements determined by multiple analytical techniques in 8 Argonne Premium Coal samples.

Techniques.--The techniques include inductively coupled argon plasma-atomic emission spectrometry (ICAP-AES) used in both Denver (D) and Reston (R) labs, inductively coupled argon plasma-mass spectrometry (ICAP-MS), flame (FAAS) and graphite furnace (GFAAS) atomic absorption spectrometry, hydride-generation atomic absorption spectrometry (HGAAS), direct-current arc atomic emission spectrography (DCAES), wavelength-dispersive (WDXRF) and energy-dispersive (EDXRF) X-ray fluorescence spectrometry, and instrumental neutron activation analysis (INAA).

Material analyzed.--Whole coal (WC) and ash were analyzed. Concentrations originally reported on an ash basis were converted to a whole-coal basis, and concentrations originally reported as an oxide were converted to an elemental basis for ease of comparison. The ash yield was also reported.

Location.--Materials were analyzed in U.S. Geological Survey laboratories in Denver (D), Colo., and Reston (R), Va.

Concentrations.--The concentrations are reported as weight percent and parts per million (ppm), which is equivalent to micrograms per gram.

Type styles and parentheses.--Average values in parentheses indicate that only one value was reported. Techniques in bold type are high-precision techniques; concentrations (except outliers) determined by these techniques are also in bold type. Techniques in regular type are less precise than those in bold; concentrations determined by the low-precision techniques are also in regular type.

Outliers.--Concentrations in braces are considered outliers for the statistical analys is used to determine the recommended values or the best average (see table 6 and appendix 2). Single outliers were determined by the Grubbs test (J.K. Taylor, 1987, Quality assurance of chemical measurements, Chelsea, Mich., Lewis Publishers, Inc., 328 p.). Multiple outliers for a given sample technique were determined by the 40 percent rule (see paper by Palmer and Klizas, this volume). Values with one significant figure were treated as outliers.

[ Contents ]

Created by the EERT WWW Staff.