Ionic Liquids as Solvents for Catalytic Conversion of Lignocellulosic Feedstocks
- Author(s): Dee, Sean Joseph
- Advisor(s): Bell, Alexis T
- et al.
The deconstruction and upgrading of lignocellulosic biomass dissolved in ionic liquids was studied as a potential alternative route to products traditionally synthesized from petroleum. While domestic biomass is a cheaper, lower carbon emission, alternative feedstock to petroleum, its utilization requires the selective deconstruction of the biopolymer to monomeric sugars and upgrading of the sugars to higher value products. Since biomass is soluble in ionic liquids, there is the opportunity to do both the deconstruction and secondary upgrading using "one-pot" homogeneous catalysis.
The primary focus of this work was to understand the kinetics of both biomass deconstruction and secondary sugar chemistry in ionic liquids. Biomass is a complex collection of molecules that consists of three primary components, cellulose, hemicellulose, and lignin. Since cellulose is the primary component, accounting for roughly 45 wt% of the raw biomass on a dry basis, initial studies aimed to understand the hydrolysis of dissolved cellulose to its sugar residue glucose. Using microcrystalline Avicel cellulose as a model, the rate laws and activation energies of cellulose hydrolysis and glucose dehydration were determined in the ionic liquid 1-butyl-3-methylimidazolium chloride ([Bmim][Cl]). No evidence of oligosaccharides was observed, suggesting that hydrolysis occurs preferentially at chain ends and is irreversible. Gradually adding water to the reaction solution, so as not to precipitate cellulose but also limit the secondary dehydration of the resulting glucose to 5-hydroxymethyl furfural (5-HMF), significantly increased glucose yield and limited production of degradation products (humins). Several mechanisms were proposed to explain the effects of water, and possible routes to humin formation.
While understanding the reactivity of model compounds is important to the development of biomass conversion technologies, it is critical to understand how the components of biomass react in their native form. An investigation was carried out to compare the reactivity of cellulose and hemicellulose model compounds to both pretreated and miscanthus grass in 1-ethyl-3-methylimidazolium chloride ([Emim][Cl]). Activation energies of model compounds were compared with the native component in raw biomass. Significant rate decreases in hydrolysis of the cellulosic and hemicellulosic components in raw biomass compared to Avicel and Xylan from Birchwood were attributed to the interaction of lignin with the biopolymers in raw biomass. However, reaction of two pretreated substrates with varying degrees of delignification showed that the presence of lignin did not have a detrimental effect on hydrolysis, but instead suggested that breaking the raw biomass macrostructures is the key to improving hydrolysis of the biopolymers. Gradual water addition strategies further improved saccharine yield, but left the cellulosic component incompletely hydrolyzed. This unhydrolyzed cellulosic component could be further converted by varying the temperature, acid concentration, or performing a second hydrolysis on the reactor residue.
After demonstrating the ability to generate sugars in high yield from miscanthus, we investigated the selective conversion of glucose to 5-HMF in ionic liquids using metal chlorides. Chromium chloride, CrCl2, has been proposed to isomerize glucose to fructose, which is readily dehydrated to 5-HMF in imidazolium chloride ionic liquids without an added catalyst. We began by studying the kinetics of fructose dehydration in [Emim][Cl] and investigating the effects of CrCl2 on the dehydration to 5-HMF. Then the kinetics of glucose isomerization to fructose using CrCl2 were characterized and compared to the rate and activation energy for fructose dehydration. Using the data for fructose, a model for the kinetics of fructose dehydration and isomerization in the presence of CrCl2 was developed. When the model was applied to glucose, it failed to describe the large conversion of glucose and small yields of 5-HMF in the initial reaction period. The accuracy of the model could be improved by including an intermediate in the glucose to fructose dehydration, which highlighted the need to characterize intermediates and products formed during dehydration using more detailed spectroscopic techniques than chromatography.
Finally, we conducted in-situ 13C NMR experiments to characterize the intermediates and products formed and to understand the effect of the ionic liquid solvent on glucose dehydration to 5-HMF catalyzed by metal chlorides in [Emim][Cl]. Compared to H2O, glucose dissolved in [Emim][Cl] exhibited higher equilibrium concentrations of the furanose and acyclic isomers of glucose. These isomers were observed to undergo dehydration more rapidly than the pyranose isomers of glucose. The rate of anomerization was also found to be faster in [Emim][Cl], a process that may facilitate ring opening in several proposed mechanisms for glucose dehydration. In situ catalytic studies were conducted using WCl6, which concluded that fructose is not formed as the reactive intermediate, but rather glucose first undergoes partial dehydration, before it is transformed from its unreactive aldose form to the more reactive ketose form. Using these observations, combined with studies of glucose dehydration catalyzed by H2SO4 several mechanisms were proposed to explain the progressive dehydration of glucose to 5-HMF using metal chlorides.