Lignin valorization is one of the key challenges for the development of economically sustainable biorefineries and fast pyrolysis holds great potential for converting lignin into high-value products. However, there are still significant challenges to overcome regarding difficulties in continuous feeding and processing of lignin in fluidized bed reactors, due to melting and agglomeration of issues. For the lignin fast pyrolysis experiments carried out in the FLEXI-GREEN FUELS project, the process development unit ‘Python’ operated at the KIT [1] was used to test a different reactor design with an aim to overcome the previously mentioned challenges.
The ‘Python’ unit has a feed capacity of 10 kg h-1 and uses a twin-screw mixing reactor (auger type reactor) to achieve high heating rates by mixing the feedstock with a preheated heat carrier [2]. Solid particles are separated at the exit of the reactor in two serial cyclones, and the main product, fast pyrolysis bio-oil, is recovered by a fractional condensation system, operated at 90°C and 20°C. To reduce plugging issues related to lignin, it was planned within the FLEXI-GREEN FUELS project to optimize the feeding section of the reactor for lignin conversion by reversing the order of feedstock and heat carrier inlet. This way the feedstock drops onto the moving bed of hot steel particles, and it is expected that the mechanical forces induced by the agitation of the bed could potentially break up agglomerates formed. To test this configuration, all necessary hardware changes were conducted at KIT and the experiments were conducted.
The changed setup was tested for an alkali lignin originated from Miscanthus (purchased from Miscancell). During the experiments there was no built up of feed material at the biomass inlet drop pipe and also no significant material deposition on the screws. Two batches of the lignin were processed in the ‘Python’ pilot unit and the overall experiment results were rated successful. Mass balances of the process were assessed and liquid samples could be produced for project partners to conduct further catalytic upgrading.
Additionally, the experimental values were used to establish a model to predict vapor liquid phase equilibria in fractional condensation. The model predicted yields and key compound concentrations (phenol, guaiacol, and water) were in good agreement with experimental results and the validated model was used for a parametric study to investigate the influence of the temperature of the first condensation stage on yield and composition of the produced condensates with the aim to optimize for downstream upgrading. It was found that a condensation temperature of 90°C was a reasonable compromise between target compounds of interest (phenol/guaiacol) and undesired compounds (water).