(89c) Exploring Diverse Lignin Populations: Complex Structural Representations and Fast Pyrolysis Modeling

Authors: 
Yanez, A. J., Northwestern University
Broadbelt, L. J., Northwestern University

Lignin is the least valorized component of the three major structural polymers of the plant cell wall, even though its constituent phenylpropanoid monomers could serve as chemical intermediates and its lower oxygen content relative to cellulose and hemicellulose makes it an attractive feedstock. To prepare biomass for subsequent upgrading, conversion strategies such as fast pyrolysis deconstruct the structural polymers into a complex mixture of reduced molecular weight species. Efforts to develop kinetic models capable of predicting complex product distributions resulting from pyrolysis have been more successful for cellulose than for lignin, in part because of the much greater structural heterogeneity and complexity of lignin. Lignin is a polydisperse polymer with no regular repeating structure, comprised of three monomers connected by fourteen reported bond types, forming a hyperbranched topology; moreover the relative abundance of monomers and bonds, as well as the molecular weight distribution, varies depending on the biomass source. Canonical representations of lignin’s structure, such as Freudenberg’s spruce lignin model, are single representations guided by experimental observations, whereas more recent approaches propose instead a diverse population of simplified linear structures whose average properties match experimentally measurable bulk properties. We present a stochastic method of producing libraries of diverse and complex structural representations of lignin that is generally applicable to any biomass source. The four experimental quantities considered as targets are the monomer, bond, and molecular weight distributions, as well as the average branching coefficient. The method is demonstrated by generating libraries of wheat straw lignin, a ‘grass-type’ lignin that can serve as a representative model for agricultural residues and energy grasses. Statistical agreement between the libraries and the experimental targets is verified, the dyadic bonding patterns between monomers are explored, and predicted product slates resulting from applying a fast pyrolysis mechanism to the libraries of structures are presented.