Liquid Phase Reforming of Wood Flour to Hydrogen

Wednesday, November 10, 2010: 8:30 AM
Alta Room (Marriott Downtown)
Timothy D. Davis, Sean C. Emerson, Tianli Zhu, Rhonda R. Willigan, Ying She and Thomas Henry Vanderspurt, United Technologies Research Center, East Hartford, CT


Growing concern over the availability, cost, and environmental impact of traditional petroleum energy sources, has emphasized the need to develop alternative energy technologies that reduce our dependence on non-renewable resources.  Lignocellulosic biomass is one such low cost source that is renewable and available in significant quantities. Globally, over 224 billion dry tones of biomass are produced through photosynthesis each year, roughly 10 times the world's energy consumption1. Currently, significant levels of research are in the field of converting this sustainable feedstock into many forms of energy, including syngas and several forms of liquid fuels.

Hydrogen is another possible product from biomass. H2 is useful for numerous applications ranging from fuel cells, chemical processing, or direct consumption as a fuel. The challenge that remains is to provide a low cost, efficient route for its conversion to hydrogen.  For such a technology to be economically feasible, a proposed system must be simple, using raw biomass feedstock, have a high selectivity for H2 production, and not produce toxic byproducts.  Significant work has been reported in the literature for liquid phase reforming (LPR) of preprocessed biomass derived feed stocks and related oxygenates like ethylene glycol. Solutions of glycerol, glucose or sorbitol have been reformed to H2, CO2, and low molecular weight alkanes over Pt, Ni-Sn catalysts2,3 and other4 catalysts.  This work concentrates on the production of H2 directly from raw biomass, without the need for intermediate processing steps.

In 2009 we reported on the high efficiency Liquid Phase Reforming of yellow poplar (Liriodendron tulipifera) saw dust over Pt-Re/ceria zirconia catalyst pellets. The striking observation was that the effluent was water white and substantially free of organic matter except for a trace amounts of formic and possibly acetic acid5. Thus not only were the hemi-cellulose and cellulose reformed but the lignin as well. This catalyst was selective towards hydrogen co-producing relatively little methane and higher alkanes.  It was also robust, but in practice this catalyst would be expensive. Furthermore the high surface area–large pore cubic ceria-zirconia support was not readily available as an item of commerce. Thus one of the next steps before proceeding to a continuous process converting wood to hydrogen was the identification and kinetic characterization of a less expensive, more readily available catalyst. 

Materials and Methods

The biomass materials and surrogates used in this work include ethylene glycol and 200 mesh wood flour obtained from P. J. Murphy.  The catalyst starting materials was Raney Ni obtained from W. R. Grace. Kinetic experiments were performed in a custom build 0.5” OD inch flow reactor constructed of Inconel 625. The system was heated by means of an infrared furnace with an Isco pump, Model 500D, supplying the feed to the reactor. A catalytic bed bed was placed in the reactor, supported by ceria stabilized zirconia beads. Pressure was controlled by means of a Tescom back pressure regulator equipped with an ER3000 electronic pressure controller. Gas analysis was performed by means of dual Varian 3800 GCs, for hydrocarbon and permanent gas analysis.

Results and Discussion

Kinetic studies on the liquid phase reforming of a biomass surrogate, ethylene glycol, have been carried out in alkaline solutions over several base metal catalysts. Several bases and concentrations have been studied to investigate the effect of basicity and concentration on the system. Although not required for the LPR, the alkalis are for the hydrolysis of the biomass. Of the metals, the study was quickly down selected to Raney Nickel, as it was concluded that the other catalysts tested did not have sufficient C-C bond breaking ability at low temperatures to warrant further study.  The complete conversion of wood flour to gas and water white carbonate solution was demonstrated with Raney Ni was confirmed in batch reactor studies.

Flow reactor studies of the LPR of 5 wt.% ethylene glycol over an 8g bed of Raney nickel were performed. Selectivity and yield results are given in Figures 1 and 2. As expected, Raney nickel was capable of completely reforming ethylene glycol to gaseous species at 310 °C in water. In addition, a high level of methanation was evident, an inherent characteristic of Ni reforming catalysts. Increasing amounts of K2CO3 slightly decreased the methanation activity as well as reduced the overall yields. However, increasing concentrations of KOH in the system did not detrimentally affect the yield but dramatically decreased the methanation activity, resulting in significantly higher selectivity's. This reduction of methanation activity may be due to homogeneously catalyzed water gas shift reactions, driving any formed carbon monoxide to carbon dioxide before methane can be produced. CO2 methanation requires higher temperatures to methanate in any significant quantity.

Figures 1 and 2- LPR of 5 wt.% Ethylene Glycol over Raney Nickel, 310 °C, 120 atm.

However, using high levels of hydroxide to improve the selectivity of nickel based catalysts creates a system level energetic issue. The KOH is converted to K2CO3 in situ, and would require regeneration.  As an alternative, a study was performed of doping the Raney nickel catalyst to promoters to drive down the methanation activity. As shown in Figures 3, a small addition of a promoter was sufficient to drive down the methanation activity without significantly decreasing overall reforming activity. This result enables the use of K2CO3 as the base for the wood hydrolysis, which has already been demonstrated as sufficient for the process5.

Figures 3- LPR of 2.5 wt.% Ethylene Glycol over Promoted Raney Nickel, 310 °C, 120 atm.

Work is ongoing to understand the LPR mechanism. A combination of experimentation and atomistic modeling will help explain the mechanism.  This may lead to catalyst modification enabling faster kinetics and tailoring the selectivity to meet the plant system efficiency targets. Furthermore, the surrogate results will be compared to those obtained from directly pumping and reacting raw biomass into the reactor.

In order to bring the technology one step closer to realization, a prototype power plant will be constructed to demonstrate the feasibility of the system for processing of raw biomass. The system is targeted to produce hydrogen sufficient to power a 1 kWe Fuel Cell by liquid phase reforming of wood flour. As of the time of this abstract, a high pressure slurry pump has been acquired and validated for pumping the feedstock at the desired conditions. Work is ongoing to determine the kinetics and optimal conditions for the wood hydrolysis. The system will use a promoted base metal catalyst for reforming and a palladium membrane for the separation of pure hydrogen from the reformate.


This work moves the concept of hydrogen production from raw biomass as describe in DOE contract DE-FG36-05GO15042 “A Novel Slurry-Based Biomass Reforming Process” a significant step closer to realization. 

Extended Abstract: File Not Uploaded
See more of this Session: Renewable Hydrogen Production I
See more of this Group/Topical: Topical 8: Hydrogen Production and Storage