Discipline: Technology and Engineering
Subcategory: Chemistry (not Biochemistry)
Hannah Smith - University of Illinois-Champaign
Co-Author(s): Nathaniel Eagan and George Huber, University of Wisconsin-Madison, WI
Biomass is a promising alternative source for fuels but is currently difficult to convert to diesel and jet fuels in high yields at the industrial scale. Sorbitol, which can be produced from cellulose in high yields, can be converted to mono-oxygenate (MO) fuel precursors through aqueous-phase hydrodeoxygenation (APHDO). APHDO proceeds through a complex network of acid- and metal-catalyzed reactions which remove oxygen atoms and may also cleave C-C bonds. This occurs until MOs are produced and ultimately deoxygenated to alkanes. The intermediate MO species in the pathway can be coupled to create renewable distillate fuels. The metal and acid activity of the catalyst must be balanced to achieve optimal yields of the desirable MO compounds rather than higher oxygenates or alkanes. Additionally, the pathway can proceed through either a dehydration route, which produces C4+ products, or a retro-aldol condensation route, which produces C3- products. The goal of this research was to show that metal identity and acidity in WOx-ZrO2 systems can be modulated to adjust MO yields. Catalysts were synthesized by incipient wetness impregnation on WOx-ZrO2 (yWZ) supports (where y is WO3 wt%). This included Pd/18WZ, Pd-Ag/9WZ, Rh/18WZ, Rh/9WZ, and Pd/ZrP as a reference support. Reactions were run in a high pressure continuous flow fixed-bed reactor with aqueous sorbitol solution and hydrogen gas. Reaction products were analyzed at several time points using gas and liquid chromatography. The weight hourly space velocity (WHSV) was decreased half way through each reaction so the product distribution near complete deoxygenation could be observed. At high space velocities, higher oxygenate species formed earlier in the pathway are more prevalent. Decreasing space velocity causes these species to further deoxygenate to alkanes, C1 gases, and MOs. It also appears that catalysts with Pd have a higher C5+ selectivity than Rh, suggesting that Pd sends more sorbitol through initial dehydration vs. retro-aldol condensation. The 9WZ support appears to be more active for sorbitol conversion than the 18/WZ support, which could be due to stronger acid sites. There appears to be no visible trend in MO yield; however, this is likely due to the WHSV and reaction conditions not being optimized to produce the maximum intermediate MO species for each catalyst. This work adds to our knowledge of catalyst design and shows which properties are important for catalyzing specific pathways. Future work will test more metals and support preparation methods to further explore the way these properties influence catalyst performance.
References: Kim, Y. T.; Dumesic, J. A.; Huber, G. W., Aqueous-phase hydrodeoxygenation of sorbitol: A comparative study of Pt/Zr phosphate and Pt ReOx/C. Journal of Catalysis 2013, 304, 72-85.
Funder Acknowledgement(s): This study was funded by the University of Wisconsin-Madison EFRI-REM program sponsored by the National Science Foundation (EFRI-1240268). Funding was also obtained from The University of Wisconsin-Madison Graduate School and National Science Foundation through the University of Wisconsin-Madison Materials Research; Science and Engineering Center (DMR-0520527) and Nanoscale Science and Engineering Center (DMR-0425880). I would like to thank George Huber, Nat Eagan, and everyone in the lab for hosting me this summer. I would also like to thank Andrew Greenberg and Jennifer Weber for facilitating the REU program.
Faculty Advisor: George Huber, gwhuber@wisc.edu
Role: I synthesized the 18WOx-ZrO2 and ZrP supports and prepared all of the catalysts. I ran the reactions and collected 75% of the samples. I analyzed and interpreted the gas chromatography data with the guidance of my graduate student, Nathaniel Eagan. My graduate student operated the equipment for X-ray Diffraction and Temperature-Programmed Reduction.