Investigating the Formation and Utilization of Humins as a By-product of the Levulinic Acid Synthesis

Abstract

The replacement of fossil raw materials is of great interest to the chemical industry. For this purpose, the synthesis of platform chemicals from biomass enables the production of industrially relevant chemicals while avoiding fossil raw material sources. One of the most important platform chemicals is levulinic acid (LA), which has so far been produced industrially using the Biofine process via an acidic reaction using sulfuric acid. However, this process uses catalysts that are harmful to the environment, as well as reactors and produces larger amounts of humins as a byproduct. The formation of these solids causes a significant reduction in the manufacturing efficiency of the sought-after platform chemical, especially as the further utilization of these humins is difficult due to their insoluble nature and unknown structure. A more detailed investigation and optimization of the LA synthesis, also regarding the formation, structure and further utilization of the humins produced, can ensure a more efficient and environmentally friendly production of the platform chemical. First, the formation process and structure of the humins was examined in more detail. 28 different humins were produced through a variation of the converted biomass substrate (fructose, glucose, xylose, sucrose), the used solvent (water, 1:1 mixture of water/ethanol, 1:1 mixture of water/dimethyl sulfoxide (DMSO)) and the used acidic catalyst (sulfuric acid, para-toluene sulfonic acid (pTSA), acetic acid (AA)). The solid analysis of the humins indicated a primarily furan ring-based structure, which resulted from the polymerization of 5-hydroxymethylfurfural (5-HMF) or furfural with themselves, other intermediates in the reaction solution and also the organic solvents. In addition, three recurring fragments of the larger humin structure were also identified. The humins produced were then converted via selective catalytic oxidation (SCO) with the polyoxometalate H8[PV5Mo7O40] (HPA-5) to form low-molecular carboxylic acids in order to determine the viability of the produced humin structures for further valorization. The main products obtained from the SCO were CO2, formic acid (FA) and AA. In general, the greatest influence was exerted by the choice of solvent. The substrate conversion in the ethanolic solution resulted in a reduced humin yield, with a minimum humin yield of about 5 wt%. These humins also showed the best conversion rates via SCO, reaching conversionrates of up to 65 %. A comparison of solids analysis before and after SCO indicated a preferred conversion of aliphatic ether and ester bonds. The combination of AA as a catalyst and the water/ethanol solution produced the lowest humin yield of about 5 wt%, with the produced humin, while also possessing the best conversion rate of 65 %, and therefore shows great potential for the efficient production of LA. The optimization of the LA synthesis was then addressed. The first step was to find an alternative to the conventionally used sulfuric acid. When comparing five different acidic catalysts (sulfuric acid, FA, H3[PMo12O40] (HPMo), H3[PW12O40] (HPW), H3[SiW12O40] (HSiW)) in an acidic conversion of fructose, the highest LA yield of about 61 mol% was achieved using the polyoxometalate HSiW. Using HSiW as an acidic catalyst, an optimization of the reaction conditions of the LA synthesis was then carried out via a design of experiment (DoE) reaction plan using a Box-Behnken design (used parameters: reaction temperature (T), reaction time (t), proportion of the organic solvent acetone (wt%(acetone)), substrate concentration (c(fructose)), which aimed at a minimized humin yield and a maximized LA yield. Here, the humin yield depended primarily on the solvent composition and temperature, while the LA yield was primarily influenced by the solvent composition and the substrate concentration. Within the parameters used (T = 140-180 °C, t = 1-5 h, wt%(acetone) = 0-80 %, c(fructose) = 0.1-0.4 mol*L-1), it was determined by interpolation of the synthesis results that high humin yields (over 40 wt%) were achieved by maximizing all parameters used. High LA yields (over 60 mol%), however, were achieved by maximizing either temperature or time while minimizing the remaining three parameters. Three parameter combinations were determined that focused on different optimization aspects. The most industrially relevant set of parameters was then successfully applied to conversions of the carbohydrates glucose, xylose, sucrose and cellobiose in order to successfully confirm the transferability of the results. Finally, based on the knowledge gained so far, a cyclic LA synthesis was designed in which FA is used as a catalyst for the production of LA and is then recovered via SCO from the humins formed as a byproduct. First, one of the previously optimized parameter sets (T = 180 °C, t = 1 h, wt%(acetone) = 0 %, c(fructose) = 0,1 mol*L-1) was adapted in a two-stage process for the usage of FA instead of HSiW in a larger reactor system. Using the modified reaction conditions, a humin fraction was formed, which was subsequently converted using SCO. Optimized conditions, using H5[PV2Mo10O40] (HPA-2) as a catalyst with the addition of methanol to suppress CO2 formation, were used here, producing a high yield of FA (about 28 carbon wt%) as the main product. The SCO catalyst could afterwards be separated and recycled using membrane filtration. The FA produced was then separated and concentrated by distillation, resulting in solution containing 87.00 wt % water, 11.65 wt% FA, 0.73 wt% succinic acid, 0.49 wt% AA and 0.13 wt% methyl acetate. However, a reaction solution whose composition was similar to the FA fraction obtained could be successfully used for a new LA synthesis, whereby the additional product residues had no visible influence on the structure of the humin produced. The cycle was thus successfully closed and the basis for a more environmentally friendly and efficient synthesis of LA was created.