Autothermal fast pyrolysis (AFP), a variation of fast pyrolysis (FP) admitting a small amount of oxygen to provide process heat, has notable merit as a biomass-to-biofuels conversion process. As a result of heat transfer and product collection advantages over standard non-oxidative FP, it has the potential to generate a higher quality product in a more economically competitive manner. Initial investigation and process development efforts, first led by Kwang Ho Kim, and Joseph Polin, respectively, at the Bioeconomy Institute, generated many further questions about the process. One notable question was “where does the energy come from to support autothermal pyrolysis” – to which the obvious answer is exothermic reactions, but beyond that is not well understood. This work explored the chemistry underlying autothermal (partial oxidative) pyrolysis, as distinguished from standard non-oxidative pyrolysis of whole biomass. A critical literature review was carried out to develop a theoretical mechanistic framework which was then applied to a process base case, and experimentally tested.

Key findings of the literature review included reaction mechanisms for the oxidation of: lignin interunit linkages, lignin monomers (and their functionalities), cellulose dimers and monomers, and hemicellulose units and functionalities. As discussed in the cellulose oxidation section, oxidation could occur by means of assisting glycosidic bond hydrolysis (either at a chain end (unzipping) or mid-chain (cracking)), effectively increasing levoglucosan yield, or by oxidation of ring functionalities. If cellulose’s substituents were to measurably react with Reactive Oxygen Species (ROS), the C6 primary alcohol would be the likely candidate, oxidizing to a C6 aldehyde or carboxylic acid, yet theoretically possible for ring-hydroxyls to oxidize.

Similarly to celluloses, hemicellulose might be oxidized by four means; polymer-end-wise chain scission initiation (primary peeling), mid-chain scission, end-chain unit degradation (secondary peeling), or side-chain oxidation. Because of its branched and heterogeneous nature, and tendencies for decomposition of monomeric units following complete depolymerization during non-oxidative pyrolysis, fewer hemicellulose hexoses and pentoses would likely be recovered during oxidative pyrolysis.

Lignin, also structurally diverse, has many possible routes for oxidation. From linkage studies, it is apparent that oxidation of the β- or γ-hydroxyl (in the case of a β-O-4’ linkage), or the α-hydroxyl (for α-O-4’ linkages) greatly weakens ether linkages, making susceptible to cleavage. Lignin’s phenolic substituents are prone to oxidation to aldehydes, carboxylic acids and ketones. Those side chains with reactive double bonds could be oxidatively cleaved or encourage a concerted decomposition reaction. Because products of oxidation can be further oxidized themselves, care must be taken in extrapolating out composition trends to scaled-operation. Even considering these routes which would effect a change in product composition, the most significant effects might come simply due to improved reaction conditions (heat transfer, heating rate, and ventilation (due to greater gas production)).

Experimental work identified reactor limitations, and explored partial oxidation of a number of model compounds, representative of cellulose, hemicellulose, as well as lignin monomers and linkages. It is important to note that the findings of the micropyrolyzer studies are not directly applicable to continuous reactor chemistry due to the fundamentally different hydrodynamics and heat transfer. Additionally, biopolymer characteristics and interaction effects are not accounted for in the monomer and dimer model compound studies, as would be seen with whole biomass.