With the world's energy demands rapidly increasing, it is necessary to look to sources other than fossil fuels, preferably those that minimize greenhouse emissions.  One such renewable source of energy is biomass, which has the added advantage of being a near-term source of hydrogen.  There are three main potential routes to produce hydrogen from biomass thermally.  The first involves gasification of biomass to synthesis gas (CO + H₂) at high temperatures in which air, oxygen, or steam is used as oxidizers.  The second requires that a liquid intermediate (bio-oil) first be produced via fast pyrolysis at relatively moderate temperatures and then reformed.  The third route includes a variety of high-pressure processes known as wet gasification or aqueous-phase reforming that is especially suitable for high-moisture content biomass.  The advantage of the second method is that it allows for the processing of biomass and the production of hydrogen at different locations utilizing low-cost biomass resources and existing infrastructure for hydrogen distribution while saving on high-cost transportation of biomass and hydrogen.  This approach is especially well suited for smaller-scale reforming plants located at hydrogen distribution sites such as filling stations.

The goal of the present work is the development of a process for the distributed reforming of bio-oil.  The focus in earlier studies of bio-oil conversion to hydrogen has been on catalytic steam reforming at high temperatures (~800 ^oC) using a fluid bed reactor, which is not optimal for a small scale distributed operation.  The present research focuses on the conversion of bio-oil to hydrogen in two steps.  The first step is non-catalytic partial oxidation at relatively low temperatures (~650 ^oC) primarily by homogeneous gas phase chemistry.  However, some nonvolatile components of bio-oil present as aerosols may react heterogeneously.  The products from the non-catalytic step are passed over a packed bed of precious metal catalyst where further reforming as well as water gas shift reactions are accomplished.  This two step approach requires significantly lower catalyst loadings than conventional catalytic steam reforming.  To date, catalyst screening experiments have used Engelhard noble metal catalysts. The catalysts used for these experiments were 0.5 % rhodium, ruthenium, platinum, and palladium (all supported on alumina).  Experiments were performed using pure alumina as well.  Both the catalyst type and the effect of oxygen on the residual hydrocarbons and accumulated carbon containing particulates were investigated.  The goal is to reform and selectively oxidize these species and catalyze the water gas shift reaction without catalyzing methanation or oxidation of CO and H₂, thus attaining equilibrium levels of H₂, CO, H₂O, and CO₂ at the exit of the catalyst bed.

Bio-oil (mixed with varied amounts of methanol to reduce the viscosity) or selected bio-oil components are introduced at a measured flow rate through the top of a vertical reactor using an ultrasonic nozzle.  The nozzle creates a fine mist, which allows the bio-oil to flow down the center of the reactor.  Additionally, the fine mist allows for intimate mixing and efficient heat transfer, providing optimal conditions to achieve high conversion at relatively low temperatures.  Generation of the fine mist is especially important for providing good contact between non-volatile bio-oil components and oxygen.  Oxygen and helium are also delivered at the top of the reactor via mass flow meters with the amount of oxygen being varied to maximize the yields of H₂ and CO and the amount of helium being adjusted such that the gas phase residence time in the hot zone remains ~0.3 s.  The reactor effluent is quenched by a flow of 10 SLPM He which serves to sweep the products quickly (~0.03s) to a triple quadrupole Molecular Beam Mass Spectrometer (MBMS) for analysis.  The dilution reduces the potential problems caused by matrix effects associated with the MBMS analysis.  The MBMS serves as a universal detector and allows for real time data collection.  Argon is used as an internal standard in the quantitative analysis of all the major products (CO, CO₂, H₂, H₂O, and benzene) as well as any residual carbon, which is determined by subsequent oxidation of carbon (monitored as CO₂) after shutting of the feed and maintaining the oxygen/helium flow.  The temperature is maintained using a five-zone furnace.  

Catalytic experiments using a 50:50 (weight basis) bio-oil:methanol mixture at 650 ^oC with an oxygen to carbon molar ratio (O:C) of 1.3 have identified Rh as the catalyst that brings the system closest to equilibrium predictions (see Table 1).  Further gas phase experiments showed that the non-catalytic step was optimal at an O:C of 1.7 (650 ^oC).  Table 2 shows a comparison of the product yields and methanol conversion from equilibrium predictions, non-catalytic (gas phase) experiments, and Rh catalyst experiments under these conditions.   60% yield of CO and nearly 28 % yield of H₂ are observed without a catalyst.  With a 0.5 % Rh catalyst inserted into the system, the CO yield decreases to ~53%, while the H2 yield increases to 74 %.  These changes are a result of increased conversion as well as the water gas shift reaction.  

Additional experiments with neat methanol (650 ^oC: r.t. = 0.45 s) were performed and compared to predictions of a detailed kinetic model.  The model is a rule-based mechanism that has been used to predict both oxidation and pyrolysis of hydrocarbons.  The premise behind the model is that three types of free radical reactions, dissociation/recombination, hydrogen abstraction, and b–scission/radical addition, dominate the kinetics.   

Future plans include expanding the temperature range (550-750 ^oC) and varying the O:C ratio at the various temperatures to determine optimal conditions for thermal conversion of bio-oil to synthesis gas before the catalyst stage.  The effects of adding a catalyst to the system and the effect of bio-oil:methanol composition will also be explored further.   The kinetic model will be extended to attempt to predict the gas-phase kinetics of selected components of bio-oil.

Table 1.  Equilibrium predictions compared to experiments in gas phase and over catalysts (O:C = 1.3).  CH3OH is given as conversion (wt. %).  Other carbon containing species are given as a yield (wt. %) based on the carbon entering the system.  H2 and H2O are yields based on the total hydrogen entering the system.     Equil. Gas Phase Alumina Pd Pt Rh Ru CH3OH  100.00 49.13 38.75 82.09 88.20 99.64 75.39 CO 68.70 38.56 36.00 50.87 45.46 61.15 48.35 CO2 23.77 7.20 7.46 9.56 11.55 23.25 9.28 H2 80.30 13.12 9.19 34.14 31.18 83.69 27.56 H2O 13.49 38.46 37.98 40.57 45.31 17.83 40.69 CH4 3.89 5.19 5.33 7.71 8.19 5.90 9.47 Residual Carbon 3.64 0.99 1.55 7.66 10.37 4.38 6.87 Benzene 0.00 1.33 0.95 4.37 2.32 1.43 2.04

                 

Table 2.  Equilibrium predictions compared to experiments in gas phase and over catalysts (O:C = 1.7).  CH3OH is given as conversion (wt. %).  Other carbon containing species are given as a yield (wt. %) based on the carbon entering the system.  H2 and H2O are yields based on the total hydrogen entering the system.     Equil. Gas Phase Rh CH3OH  100.00 87.47 99.98 CO 56.98 60.34 52.91 CO2 41.82 11.31 37.60 H2 72.32 27.71 74.07 H2O 25.76 54.30 31.68 CH4 1.20 6.51 4.99 Residual Carbon 0.00 0.93 1.44 Benzene 0.00 2.67 0.45