A major goal of the U.S. military is to use solid oxide fuel cells (SOFCs) as a power source during military silent watch missions. Fueling these SOFCs with JP-8, the military's primary fuel type, would be ideal. However, the organosulfur compounds present quickly poison the expensive fuel cell and reformer components, drastically limiting the lifetime of the SOFC. Furthermore, current desulfurization technology is unable to produce ultra-low sulfur content JP-8 fuel required for solid oxide fuel cells (SOFCs).
The current desulfurization technique for fossil fuels is hydrodesulfurization, which is a high temperature, high pressure process that must be done at a refinery. While this process is adequate for achieving low sulfur content in lighter fuels such as gasoline, it is ineffective at removing the bulkier organosulfur compounds found in JP-8. Adsorptive desulfurization is a promising alternative to hydrodesulfurization, as it has the potential to be a portable, on-site process performed on JP-8 stocks already in the field without needing additional hydrogen as is the case with hydrodesulfurization. These adsorptive materials are high surface area sorbents with pore sizes large enough to accommodate sulfur contaminants, and typically these areloaded with d- or f-block metals in ionic, metallic, or oxide form to serve as active sites.
Hierarchical mesoporous monoliths, synthesized using Zeolite Y and Al-SBA-15 with agarose as a templating agent, were explored as possible adsorptive desulfurization materials. Al-SBA-15 monolith was found to be more effective than the Zeolite Y monolith, which is attributed to the difference in pore size between Zeolite Y and Al-SBA-15. However, there was essentially no difference in capacity between the monolith form of Al-SBA-15 and its bulk powder form. Therefore, hierarchical monoliths were not further explored and experiments on bulk Al-SBA-15 were continued.
Three types of SBA-15 were investigated: pure silica SBA-15, aluminoslicate SBA-15 (Al-SBA-15), and aminopropyl-functionalized SBA-15 (APS-SBA-15). Various metal ions were loaded into the frameworks and their loading procedures were evaluated. Ag+ was found to have the highest adsorption capacity, and loading via wet impregnation was beneficial compared to ion exchange procedures. Ultimately, 18 wt.% Ag-Al-SBA-15 was found to have an adsorption capacity of 31.47 mgS/g in JP-8. However, Ag-MCM-41 compared to silver loaded SBA-15 and its derivatives was found to be more effective.
MCM-41, aluminosilicate MCM-41 (Al-MCM-41), and MCM-41 nanoparticles (MSN) were next explored as higher surface area alternatives to SBA-15 and Al-SBA-15. Ag loaded Al-MCM-41 displayed a high desulfurization capacity, but poor reproducibility after regeneration. While silver loaded MCM-41 displayed a slight decrease in capacity, it had a highly reproducible regenerability of ~ 70%. Silver-loaded MSN displayed a four-fold greater performance towards JP-8 fuel over previously reported sorbents, whereas MCM-41 displayed a three-fold greater capacity than previous reports. Silver-impregnated MSN and MCM-41 were found to have saturation adsorption capacities for JP-8 of 32.6 mgS/g and 25.4 mgS/g, respectively. MSN also displayed a high capacity for the sterically hindered 4,-6-dimethyldibenzothiophenes along with a breakthrough capacity of 0.98 mgS/g at 10ppmwS, which is twice that of other published materials.
In an effort to produce novel sorbents with both high surface areas and active surfaces, we produced a series of silica-zirconia mesoporous materials. Three long-chain primary alkylamines were explored: octylamine, dodecylamine, and hexadecylamine, as well as a variety of Si:Zr ratios. These frameworks were loaded with Ag via wet impregnation and tested with model fuels. The optimum material, Ag-DDA-15, produced using dodecylamine and a Si:Zr ratio of 15:1 and 12 wt.% Ag, displayed almost as high of a desulfurization capacity as Ag-MSN, but with a far superior silver efficiency. Ag-DDA-15 also shows better regenerability than Ag-MSN, maintaining 80% of its capacity on the second cycle, compared to 70% with Ag-MSN.