Rhett Brewer and Dr. William Hecker, Chemical Engineering
To study the effects of the carbon structure on coal reactivity during the late stages of combustion, it was necessary to be able to know what effect mineral content had on the reactivity. Two coals, dietz (high mineral content) and pitt 8 (low mineral content), were chosen for the experiment so that their behavior during the late stages of combustion could be compared.
To avoid the effects that different particle sizes have on coal combustion kinetics, the pitt 8 sample was prepared by sifting pulverized pitt 8 coal and leaving only particles between 63 and 75 microns across. Coal is a heterogeneous substance that contains both volatile and solid structures. To eliminate the volatile matter, and allow us to examine only the reactivity of the solid carbon lattice, the sample was passed through a flat flame burner (FFB) to burn off the coal’s volatile fraction. This technique successfully separates the volatiles from the carbon lattice because the volatiles burn so much faster than the more stable carbon lattice.
Determining how long to run the coal sample through the FFB was not a trivial problem. The FFB passes the sample, fluidized in nitrogen, through a 4 cm long methane flame, a predetermined length of space, and into a collection probe where it is quenched with nitrogen. The length of the space between the flame and the collection probe, as well as the flow of the injection nitrogen, determines how long the coal sample is burned (dictating the fraction of volatiles and carbon structure burned). The difficulty was to find the gas flow and the probe height settings that would drive off almost all of the volatiles and burn as little of the carbon structure as possible.
In general, pitt 9 contains about 40% weight volatiles. To determine the settings necessary to burn about 40% of the coal in the FFB (hence burning off the volatiles) a series of chars (devolatilized coal samples) were made in the FFB with the same gas flow settings, but different probe heights.
ICP, a method that traces the amount of titanium before and after the sample is burned in the FFB, is normally used to detern-tine the mass release of the sample in the FFB. This method is very time consuming, while alternate methods of determining the mass loss, both the ASTM and one using a thermogravametric analyzer (TGA), are much quicker but less accurate. To improve the accuracy of the method using the TGA (the quickest, easiest method), an experiment was run to make a correlation between the mass loss determined by the ICP method and the TGA method. Two different char samples were made at each of three different probe heights. The mass loss for each sample (six in all) was determined using each of the three methods twice (ICP, TGA, and ASTM). The mass loss measured using the TGA and the ASTM. However, because the ICP method is so time consuming, I have not yet had time to perform the tests. It is important to note that the mass loss reported using the TGA method was very consistent for a given sample, indicating that the method can be relied upon for reproducible results.
Figures 1 and 2 were also used to pick the FFB setting that would produce fully devolatilized char that would be suitable for the study of the carbon structure. The graphs show that the mass loss for the extra 10 cm between 4.5 cm and 15 cm was small compared to the mass loss experienced by the particle between 0 cm and 4.5 cm. This indicates that between 0 cm and 4.5 cm the coal is quickly devolatilizing, while after 4.5 cm it is slowly oxidizing the carbon structure. Using this knowledge, it was decided to use a height of 6 cm to make the char for the study. Six cm was chosen because of limitations of the FFB that would make it difficult to accurately reproduce char samples made at less than a 6 cm probe height. When making a char sample it is essential that the char is fully devolatilized, not too much of the carbon lattice is burned, and the char can be reproduced. Char made with a 6 cm probe height satisfies these conditions.
This concludes the current status of the project. To complete the project, the same process described above must be followed to determine the settings for making the dietz char sample. Both the dietz and pitt 8 samples must also be made. Another obstacle to the project’s completion is that the reactor that is to be used to heat treat the particles is currently being repaired. Once this is done, 0%, 50%, and 90% burned out char samples must be prepared, some heat treated, and kinetic data collected for each one. This data will be compared to study the effect of heat treatment and the changing carbon structure on combustion kinetics during the late stages.
This project has been an amazing experience. It has allowed me to expand my capacity to analyze and solve difficult technical problems. Thank you for your support.