460629 Catalytic Hydrotreatment of Pyrolysis Oil Model Compounds in a Batch Reactor

Wednesday, November 16, 2016: 3:40 PM
Union Square 19 & 20 (Hilton San Francisco Union Square)
LiLu Funkenbusch, Chemical Engineering, Michigan Technological University, Houghton, MI, Michael Mullins, Department of Chemical Engineering, Michigan Technological University, Houghton, MI and Louise Olsson, Chemical Engineering, Chalmers University of Technology, Göteborg, Sweden

Bio-oil generated from the pyrolysis of lignocellulosic feedstocks is a complex mixture of highly oxygenated hydrocarbons (>30wt% oxygen), and is poorly miscible with conventional fossil fuels. Therefore, catalytic hydrodeoxygenation (HDO) is necessary to lower its oxygen content and to produce a compatible “drop-in” transportation fuel. We conducted batch HDO reactor studies using representative phenolic compounds found in pyrolysis oil including anisole, m-cresol and phenol, both individually and in mixtures. These tests were conducted at pressures (50 bar) and temperatures (250 to 350oC) close to industrial conditions to obtain high conversion of the initial reactants. Experiments were conducted with finely powdered catalysts in a well-mixed Parr reactor to reduce mass transfer effects, and samples were collected regularly until completion. Instead of traditional sulfided HDO catalysts, the studies employed platinum on alumina (Pt/Al2O3) and palladium over carbon (Pd/C). Utilizing reaction mechanisms and pathways proposed in our previous studies, the data from the batch reactor experiments was used to fit the kinetic rate constants and Arrhenius parameters for the primary reactants and intermediate compounds. The analytical results show that many of the reactants proceeded through similar intermediates; therefore, a hybrid, lumped-parameter model was created and compared to the data. Higher temperatures led to shorter intermediate lifetimes, and more secondary reactions. While platinum on alumina was primarily focus, tests run with palladium on carbon exhibited less rearrangement of methyl groups. The reaction model developed may then be used to predict product yields, hydrogen consumption, energy requirements, and greenhouse gas (GHG) emissions. This capability is necessary for a more accurate Life Cycle Assessment (LCA) of this biofuel route, and to better assess the economic viability and sustainability of the process. Catalyst coking at longer test times was an issue, and thermal desorption tests followed by mass spectroscopy were performed to determine the rate and extent of the coking process.

Extended Abstract: File Not Uploaded