Depolymerisation using Size Exclusion Chromatography

Lignin, the second most abundant polymer in nature, has a significant role to play in the move from petrochemical refinery to biorefinery. However, before that can occur an economically and environmentally sustainable method must be found to depolymerise the recalcitrant polymer into useful monomers. Using steam exploded, hydrolysis lignin from sugar cane bagasse this project investigated the size changes seen in lignin in response to specific variables. Variables investigated are pH, time of UV light exposure, titanium dioxide photocatalyst type (P25, PF2 and Cyc1) and concentration, as well as iron, manganese and hydrogen peroxide concentration. A high through-put, experimental design approach was developed in conjunction with a new aqueous (1:1 H2O:methanol) size exclusion chromatography (SEC) method using a short format (10 i.d. x 100 mm), methacrylate column (PL Rapide ™ Aqua L). High salt (0.2 mol/L NaCl and 0.2 mol/L KSCN, pH 7.3) was used to minimise non-size based separation mechanisms, such as, hydrophobic and stationary phase interactions and aggregation of lignin. The resultant data was subjected to multiple regression analysis and models were constructed to quantitate the effect of each variable studied. Creating models represents a user friendly way of interpreting large amounts of data and can be used whether a traditional matrix or experimental design approach has been taken. Lignin depolymerisation chemistry is particularly suited to this approach as large data sets often result which are difficult to interpret. This work is the first known attempt to model size changes in lignin. Six models were constructed for each of four systems: a no-catalyst system and three containing titanium dioxide catalysts- P25 (Degussa), PF2 (Degussa) and Cyc1 (Chengyin). Each of the six models applies to a different molecular weight range. The interaction between iron and hydrogen was found to be synergistic for PF2 and the interaction between catalyst concentration and iron was found to be antagonistic for Cyc1. For comparison between models the relative success of each model is expresse