A. Aqueous and Non-aqueous Non-pyrolytic Conversion

Note that the transformations described here, which take place in an aqueous or non-aqueous liquid medium that may or may not react with the biomass feedstock, are termed nonpyrolytic processes, as compared to pyrolysis pro­cesses in which the biomass feedstocks are directly heated.

The direct conversion of cellulosic materials to liquids by heating in aqueous systems has been known for more than 100 years (с/. Ostermann, Bishop, and Rosson, 1980). Pure cellulose is liquefied at 300°C and below at a pressure of 19.3 MPa in less than 1 h with or without added sodium carbonate catalyst up to a concentration of 0.8% in the medium (Molton et al, 1981). A wide range of aliphatic and aromatic alcohols, phenols, hydrocarbons, substituted furans, and alicyclic compounds is formed. The presence or absence of a gaseous atmosphere of carbon monoxide had no effect on the results, which contrasts somewhat to the results obtained from demonstration of a similar conversion system (PERC process) described later. The experimental results support a degradation mechanism that forms acetone, acrolein, and acetoin intermediates, which recondense under alkaline conditions to yield the ob­served products. Model compound experiments indicated that the ketones and furans are formed by aldol condensations and Michael reactions of carbonyl intermediates. The formation of aromatic hydrocarbons and phenol from these molecules under these conditions appears to involve a variant of the aldol condensation. In other experiments at temperatures of 268 to 407°C, pressures of 21 to 35 MPa, and reaction times of 20 or 60 min, the average liquid yield was 25 wt % of the cellulose converted to acetone-soluble oil (Miller, Molten, and Russell, 1981). About 10 wt % charcoal was produced along with gaseous by-products. The heating of poplar wood chips in water alone at about 330°C in autoclaves affords acetone-soluble liquid oils containing about 20 to 35% oxygen at yields up to about 50 wt % of the feedstock (Boocock et al, 1985, 1987). Little or no char is produced, and the physical breakdown of the chips is believed to occur by water absorption, swelling, and disruption and liquefaction of the matrix, whereupon the absorbed water is regenerated. When poplar chips are used, about one-half the oil is phenolic and one-fourth is phenol itself. The phenol yield is 6.5 wt % on a dry wood basis or 25 wt % based on the lignin content. The relatively high yield of phenol was suppressed under alkaline conditions. It was concluded that the steam formed in the process at the self-generated pressure of about 15.9 MPa was responsible for disruption of the chips. In subsequent experiments with steam injection at 350°C into a downflow reactor, the oil yields from poplar wood chips were in excess of 40%. The oils softened and flowed just above 100°C and their oxygen content was in the low 20% range.

Another interesting catalytic liquefaction method involves the reaction of biomass-water slurries (LBL process) or biomass-recycle oil slurries (PERC process) with sodium carbonate and carbon monoxide gas at elevated tempera­ture and pressure to form heavy liquid fuels. Biomass and the combustible fraction of wastes have been converted at weight yields of 40 to 60% at temperatures of 250 to 425°C and pressures of 10 to 28 MPa. Lower viscosity products are generally obtained at higher reaction temperatures and solid or semisolid products are obtained when the reaction temperature is below 300°C. However, the high nitrogen and oxygen contents and the boiling characteristics and high viscosity range of the liquid products make it difficult to classify them as petroleum substitutes. They would have to be upgraded by other processes. The original PERC process consisted of a sequence of steps: drying and grinding wood chips to a fine powder, mixing the powder with recycled product oil (10% wood powder to 90% recycle oil), blending the mixture with water containing sodium carbonate, and treating the slurry with synthesis gas at about 27,579 kPa and 370°C. The modified LBL process consists of partially hydrolyzing the wood in dilute sulfuric acid and treating the water slurry containing dissolved sugars and about 20% solids with synthesis gas and sodium carbonate at 27,579 kPa and 370°C on a once-through basis. The resulting oil product yield is about 1 bbl/400 kg of chips and is approximately equivalent to No. 6 grade boiler fuel. It contains about 50% phenolics, 18% high-boiling alcohols, 18% hydrocarbons, and 10% water.

The evaluation of pressurized wood-slurry liquefaction by the LBL process of wood-water slurries was performed mainly in small-scale equipment (Ergun, 1981; Davis, 1983). Oils similar to those produced by the PERC process were obtained, but at lower yields. In the late 1970s and early 1980s, the PERC process was evaluated in a PDU (Thigpen and Berry, 1982). The purpose of this work was to demonstrate the direct, continuous, thermochemical liquefac­tion of biomass. In this process, a synthesis gas mixture is reacted with wood slurried in recycled oil in the presence of 5% sodium carbonate at pressures of 20.7 MPa and temperatures of about 270°C for 1 to 1.5 h. It was suggested by the researchers who developed this process in the laboratory that carbon monoxide reacts with sodium carbonate in the presence of water to form sodium formate, which in turn deoxygenates the biomass feedstock to yield oil.

Study of the mechanism of this complex reduction-liquefaction process led to the suggestion that part of the mechanism involves formate production from carbonate, dehydration of the vicinal hydroxyl groups in the cellulosic feed to carbonyl compounds via enols, reduction of the carbonyl group to an alcohol by formate and water, and regeneration of formate (Appell et al, 1975). The following reactions were suggested:

Na2C03 + H20 + CO —» 2HC02Na + C02 QH10O5 + HC02Na -» C6H10O4 + NaHCOj (Wood) (Oil)

NaHC03 + CO —» HC02Na + C02 HC02Na + H20 NaHC03 + H2 H2 + C6H10O5 * C6H10O4 + H20.

The approximate stoichiometry of the process developed from the data (Thig­pen and Berry, 1982) corresponded to

C6H9840414 + 3.55CO + 2.14H2—» 0.877C6H762O066 + 3.99CO, + 0.22 CO

+ 4.36H2.

(Wood) (Oil)

In view of the complex nature of the reactants and products, it is likely that a complete understanding of all of the chemical reactions that occur in the PERC process will not be developed unless detailed mechanistic studies are carried out.

In the PDU tests of the PERC process, a crude product oil comparable to No. 6 fuel oil was produced in barrel quantities at yields of about 53 wt % of the feedstock. It had a heating value of up to 34.5 MJ/kg, a specific gravity of 1.1, a viscosity of 0.20 Pa-s at 99°C, and an oxygen content of 12.3%. The distillate from the crude oil had a heating value of 40.4 MJ/kg, a viscosity of 0.01 Pa-s, and an oxygen content of 6.2%. Its characteristics were similar to those of No. 2 fuel oil. These results were obtained from the longest sustained run with Douglas fir; 4988 kg of crude oil was obtained from 9953 kg of feedstock. These data suggest that the PERC process yields what much of the research on the direct liquefaction of biomass has been unable to achieve—a one-step, direct liquefaction process using woody feedstocks that yields a crude

product oil similar to a petroleum fuel oil. Unfortunately, operation of the PDU was terminated before several key questions could be answered. Is sodium carbonate necessary? Is synthesis gas necessary, and if so, how much? If one or both of these reactants is eliminated, what is the effect on the crude product oil’s composition and yield? Eventually, these questions will be resolved, if and when the process is developed further. But it seems evident from the PDU data that the PERC process is capable of overcoming some of the problems encountered in other direct biomass liquefaction processes.