Combustion synthesis of fullerenes and fullerenic nanostructures
Massachusetts Institute of Technology. Dept. of Chemical Engineering.
Jack B. Howard.
MetadataShow full item record
Fullerenes are molecules comprised entirely of sp²-bonded carbon atoms arranged in pentagonal and hexagonal rings to form a hollow, closed-cage structure. Buckyballs, a subset which contains C₆₀ and C₇₀, are single-shell molecules while fullerenic nanostructures can contain many shells and over 300 carbon atoms. Both fullerenes and nanostructures have an array of applications in a wide variety of fields, including medical and consumer products. Fullerenes were discovered in 1985 and were first isolated from the products of a laminar low-pressure premixed benzene/oxygen/argon flame operating at fuel-rich conditions in 1991. Flame studies indicated that fullerene yields depend on operating parameters such as temperature, pressure, residence time, and equivalence ratio. High-resolution transmission electron microscopy (HRTEM) showed that the soot contains nanostructures, including onions and nanotubes. Although flame conditions for forming fullerenes have been identified, the process has not been optimized and many flame environments of potential interest are unstudied. Mechanistic characteristics of fullerene formation remain poorly understood and cost estimation of large-scale production has not been performed. Accordingly, this work focused on: 1) studying fullerene formation in diffusion and premixed flames under new conditions to identify optimal parameters; 2) investigating the reaction of fullerenes with soot; 3) positively identifying C₆₀ molecules in HRTEM by tethering them to carbon black; and 4) providing a cost estimation for industrial fullerenic soot production.(cont.) Samples of condensable material from laminar low-pressure benzene/argon/oxygen diffusion flames were collected and analyzed by high-performance liquid chroma- tography (HPLC) and HRTEM. The highest concentration of fullerenes in a flame was always detected just above the height where the fuel is consumed. The percentage of fullerenes in condensable material increases with decreasing pressure and the fullerene content of flames with similar cold gas velocities shows a strong dependence on length. A shorter flame, resulting from higher dilution or lower pressure, favors the formation of fullerenes rather than soot, exhibited by the lower amount of soot and precursors in such flames. This indicates a stronger correlation of fullerene consumption to soot levels than of fullerene formation to precursor concentration. The maximum flame temperature seems to be of minor importance in formation. The overall highest amount of fullerenes was found for a surprisingly high dilution of fuel with argon. The HRTEM analysis showed an increase of the curvature of the carbon layers, and hence increased fullerenic character, with increasing distance from the burner up to the point of maximum fullerene concentration, after which it decreases, consistent with the HPLC analysis. The soot shows highly ordered regions that appear to have been cells of fullerenic nanostructure formation. The samples also included fullerenic nanostructures such as tubes and spheroids including highly-ordered multilayered or onion-like structures. Studies of turbulent-like benzene/oxygen/argon diffusion flames showed that these flames produce fullerenes over a wider range of heights than laminar flames but with lower yields.(cont.) No discernible trend could be detected in the data and the fullerene results were not easily reproducible indicating that such flames are not suitable for fullerene formation. Soot samples were also collected from a well-characterized laminar premixed benzene/oxygen/argon flat flame under new conditions and analyzed by HPLC and HRTEM. Flame studies using secondary injections of benzene or acetylene show that two-stage flames are unsuitable for fullerene production. It seems that secondary fuel has an adverse effect on the formation of fullerenes and creates conditions that are similar to the early stages of a single-stage flame prior to soot formation. This means that fuel must go through the combustion process to form fullerenes and that they cannot be formed simply by organic pyrolysis. Additionally, fullerene data collected in this study show significantly higher yields than in a previous study and the absence of a concentration drop-off. The coexistence of fullerenes and soot does not support but also does not rule out that fullerenes are consumed by soot, as was suggested by diffusion flame data. Given the discrepancy in the data, fullerene consumption was studied in experiments involving pure fullerenes being sublimated into a passing argon gas stream. This gas stream then passed through a carbon black bed. As the fullerenes passed through the bed, a certain percentage reacted with the surface of the particles and the non-reacted material was collected downstream. Experiments at different temperatures indicate that fullerenes are indeed consumed by soot particles but that the consumption is quite slow.(cont.) The rate coefficient obtained resembles those seen for surface diffusion controlled reactions or for heterogeneous reactions. Extrapolation of the reaction coefficient to flame conditions would indicated that this type of fullerene consumption is not nearly enough to explain the consumption observed in fullerene-forming flames, meaning that fullerenes are consumed by other mechanisms. HRTEM analysis of carbon black with and without tethered fullerenes shows that fullerenes can in fact be observed in TEM micrographs. In this experiment, functionalized C₆₀ molecules were attached to the surface of carbon black particles with a chemical tether. The resulting compound was analyzed by HRTEM and compared with similar analysis of untreated carbon black. The post-treatment carbon black not only has an order of magnitude greater concentration of apparent fullerene structures but size distribution data shows a significant peak at the C₆₀ diameter for the treated sample whereas no peak is observed for the untreated sample. This indicates that the fullerenes have indeed been attached to the particle surface and that they can definitively be seen in images produced from HRTEM. Lastly, a model was built to estimate the cost of the large scale production of fullerenic soot. This model was based on current carbon black technology and takes into account operating parameters specific for fullerene production. Sensitivity analyses performed on the model indicate that soot yield and fuel price are the most important factors in determining production cost while electricity costs are minimally important.(cont.) It was seen that operating pressure and equipment lifetime are negligible in the final cost. Overall, combustion holds immense promise to be a much cheaper and more efficient alternative to the current method of commercial fullerene production.
Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Chemical Engineering, 2002.Includes bibliographical references (leaves 123-132).
DepartmentMassachusetts Institute of Technology. Department of Chemical Engineering
Massachusetts Institute of Technology