| OCR Text |
Show reasonably refractory minerals yet are more abundant in most coals include aluminum and silicon. All three of these generally can be used to give an estimate of overall mass loss. The analysis of mineral matter evolution depends critically on an accurate and statistically complete analysis of overall mass loss. Therefore, these are discussed in some detail here. The determination of the concentration of aluminum, silicon, and titanium are performed independently. Therefore, the fractional mass losses calculated from these three analyses are also statistically independent. This independence can substantially improve the precision with which fractional mass loss measurements can be determined. When combined with the measured mass fed to and collected from the reactor, there is a total of four statistically independent measurements of mass loss. In general, this decreases the uncertainty with which the mass loss is determined by a factor of 10 or greater. Mineral species experience several types of transformations during combustion. As particles approach 100 percent burnout, even the most refractory species experience changes in environment which may affect the analysis described above. For example, they may be released from the receding particle surface or fuse with other molten mineral species. These processes will affect the calculation of overall burnout and the release of inorganic elements if they result in preferential loss of the mineral species. That is, the calculation of overall mass loss will be inaccurate if the mineral species passes through the filter or becomes detached from the coal/char particle and is collected somewhere other than on the filter. Two provisions are made for the detection of errors due to preferential loss of elemental species being used as tracers during combustion. The first is by selection of tracers which occur in significantly different modes in the coal. For example, silicon and aluminum typically occur in the form of silicate grains in the coal matrix, with particle sizes in the range of 10 j.£m. Titanium, on the other hand, is very homogeneously distributed through the coal and typically has a particle size in the range of 0.2 j.£m or less. Mineral species with such large difference in size and mode of occurrence in the coal will not likely be sensitive to the same types of transformations. Burnout analyses which agree closely when based on several different species indicate that no transformation of the species has occurred. If systematic errors do occur in the calculation of residual mass based on a tracer, they will tend to overestimate the residual mass of the sample. That is, there are no mechanisms which will systematically add mass of a given element to the sample. Selectively decreasing the mass of a given tracer element in the sample will increase the estimation of the residual mass of the coal/char, as is seen in Equation l. The amounts of coal fed and char collected in the reactor are also used as an 4 |
