A Novel Slurry-Based Biomass Reforming Process Final Technical Report

This project was focused on developing a catalytic means of producing H2 from raw, ground biomass, such as fast growing poplar trees, willow trees, or switch grass. The use of a renewable, biomass feedstock with minimal processing can enable a carbon neutral means of producing H2 in that the carbon dioxide produced from the process can be used in the environment to produce additional biomass. For economically viable production of H2, the biomass is hydrolyzed and then reformed without any additional purification steps. Any unreacted biomass and other byproduct streams are burned to provide process energy. Thus, the development of a catalyst that can operate in the demanding corrosive environment and presence of potential poisons is vital to this approach. The concept for this project is shown in Figure 1. The initial feed is assumed to be a >5 wt% slurry of ground wood in dilute base, such as potassium carbonate (K2CO3). Base hydrolysis and reforming of the wood is carried out at high but sub-critical pressures and temperatures in the presence of a solid catalyst. A Pd alloy membrane allows the continuous removal of pure , while the retentate, including methane is used as fuel in the plant. The project showed that it is possible to economically produce H2 from woody biomass in a carbon neutral manner. Technoeconomic analyses using HYSYS and the DOE's H2A tool [1] were used to design a 2000 ton day-1 (dry basis) biomass to hydrogen plant with an efficiency of 46% to 56%, depending on the mode of operation and economic assumptions, exceeding the DOE 2012 target of 43%. The cost of producing the hydrogen from such a plant would be in the range of $1/kg H2 to $2/kg H2. By using raw biomass as a feedstock, the cost of producing hydrogen at large biomass consumption rates is more cost effective than steam reforming of hydrocarbons or biomass gasification and can achieve the overall cost goals of the DOE Fuel Cell Technologies Program. The complete conversion of wood to hydrogen, methane, and carbon dioxide was repeatedly demonstrated in batch reactors varying in size from 50 mL to 7.6 L. The different wood sources (e.g., swamp maple, poplar, and commercial wood flour) were converted in the presence of a heterogeneous catalyst and base at relatively low temperatures (e.g., 310 °C) at sub-critical pressures sufficient to maintain the liquid phase. Both precious metal and base metal catalysts were found to be active for the liquid phase hydrolysis and reforming of wood. Pt-based catalysts, particularly Pt-Re, were shown to be more selective toward breaking C-C bonds, resulting in a higher selectivity to hydrogen versus methane. Ni-based catalysts were found to prefer breaking C-O bonds, favoring the production of methane. The project showed that increasing the concentration of base (base to wood ratio) in the presence of Raney Ni catalysts resulted in greater selectivity toward hydrogen but at the expense of increasing the production of undesirable organic acids from the wood, lowering the amount of wood converted to gas. It was shown that by modifying Ni-based catalysts with dopants, it was possible to reduce the base concentration while maintaining the selectivity toward hydrogen and increasing wood conversion to gas versus organic acids. The final stage of the project was the construction and testing of a demonstration unit for H2 production. This continuous flow demonstration unit consisted of wood slurry and potassium carbonate feed pump systems, two reactors for hydrolysis and reforming, and a gas-liquid separation system. The technical challenges associated with unreacted wood fines and Raney Ni catalyst retention limited the demonstration unit to using a fixed bed Raney Ni catalyst form. The lower activity of the larger particle Raney Ni in turn limited the residence time and thus the wood mass flow feed rate to 50 g min-1 for a 1 wt% wood slurry. The project demonstrated continuous H2 yields with unmodified, fixed bed Raney Ni, from 63% to 100% with corresponding H2 selectivities of 6% to 21%, for periods of several hours. The fixed bed form of the Raney Ni exhibited signs of deactivation which requires further study