Ria Abdilla

The ever increasing global consumption of fossil resources, environmental concerns related to high CO2 emissions and diminishing petroleum reserves, have boosted research on renewable feedstocks. In the last two decades, the “biobased economy” concept has been proposed as an alternative to the current fossil based economy. It involves the production of heat, power, transportation fuels and chemicals from renewable resources like biomass with a minimum of waste, resource optimization and low greenhouse gas (GHG) emissions. With an estimated annual global production of 170x10^9 ton, biomass is considered as the only green carbon resource available in high enough quantities to (partly) substitute fossil resources. One of the important elements in a biobased economy is the “biorefinery concept”. It involves separation of the biomass into fractions that are converted using dedicated biological, (bio)-chemical, physical and/or thermal processing. It aims to convert low-value biomass into high value (industrial) intermediates and final products (e.g. fuels, chemicals, material, energy) using sustainable and integrated processes. There are strong analogies between a biorefinery and conventional crude oil refinery regarding operation (not in feeds) and nowadays, biorefineries are strongly advocated.

Although the main driver for the transition from fossil based refineries to biorefineries is the need for sustainable heat, power and transportation fuels, the production of biobased chemicals is also of high interest. Biobased chemicals with a high derivatization potential due to the presence of multiple functional groups are often addressed as platform chemicals. In the last two decades, the conversion of lignocellulosic (woody) biomass has received major interest. Lignocellulosic biomass contains about 30-60% cellulose and 20-40% hemicellulose, which are biopolymers containing C6 and C5 sugar units. A wide variety of platform chemicals can be obtained from sugars and among them, 5-hydroxymethylfurfural (HMF) is of our particular interest.

HMF has been identified as one of the top platform chemicals due to its high versatility. HMF contains both an aldehyde and an alcohol group and is typically obtained by the acid-catalyzed conversion of C6 sugars in water. Unfortunately, selectivity is an issue and significant amounts of byproducts such as levulinic acid (LA), formic acid and insolubles known as humins are formed. The commercial scale production of HMF is still in a stage of infancy, due to the high cost of production, associated with low yields, and difficulties with product separation. In water, D-fructose is the C6 sugar feed of choice as it gives a maximum HMF yield of around 55 mol%, whereas for D-glucose, the yield under similar conditions is less than 10 mol%. However, D-glucose is preferred in terms of feedstock price. Oligo-and polysaccharides, as well as lignocellulosic biomass, are potentially interesting feedstocks for HMF and could offer substantial economic and environmental benefits. As such, the identification of alternative cost-efficient feeds for HMF synthesis is of high interest. An ideal feed should be inexpensive, relatively abundant and require minimum or even no pre-treatment before conversion. In this thesis, experimental studies on the use of two alternative sugar sources namely pyrolytic sugars and thick juice, for the synthesis of platform chemicals via chemo-catalytic routes are described.

Pyrolytic sugar is the commonly used name for the water soluble fraction of pyrolysis liquids, which are produced from biomass by a thermal conversion route known as fast pyrolysis (450-600 °C, 1-2s, in the absence of oxygen). Fast pyrolysis is as a promising and versatile technology to depolymerize and concentrate sugars from lignocellulosic biomass. Pyrolysis liquids contain low molecular weight sugars like 1,6-anhydro-β-D-glucopyranose (commonly known as levoglucosan), 1,6-anhydro-β-D-glucose (cellobiosan) and sugar oligomers. Levoglucosan (LG) in particular, is an interesting source for glucose (GLC). Thick juice is an intermediate sugar stream in a conventional sugar beet plant and contains approximately 60-70% of sucrose (SUC), a dimeric sugar of fructose (FRC) and GLC.

The main objective of the research described in this thesis is the development of synthetic methodologies to convert both alternative feeds to biobased platform chemicals, with an emphasis on HMF. Experimental and kinetic modeling studies were performed, and the effects of process conditions and catalysts on HMF yield were investigated in detail. In some cases, the studies were performed on model components to reduce complexity. Chapter 2 to 4 report studies using pyrolytic sugars as the feed while Chapter 5 and 6 report studies using thick juice.

In Chapter 2, experimental and kinetic modeling studies on the conversion of LG, the major sugar in pyrolytic sugars, to GLC are described. As before mentioned, GLC is an important building block sugar which can be further converted to biobased chemicals, among others LA and HMF. Experiments were conducted in an aqueous-acidic medium using two Brønsted acids (sulfuric acid and acetic acid) as the catalysts under a wide range of conditions in a batch reactor (glass ampoule). The effects of the initial LG loading (0.1–1 M), sulfuric and acetic acid concentrations (0.05–0.5 M and 0.5–1 M, respectively), and reaction temperature (80–200 °C) were determined. The highest GLC yields were obtained using sulfuric acid (98 mol%), whereas the yields were considerably lower for acetic acid (maximum 90 mol%) due to the formation of byproducts such as insoluble polymers (humins). Kinetic parameters were determined using a MATLAB optimization routine. A good agreement between experimental and model was obtained when assuming that the reaction is first order with respect to LG. The activation energies were 123.4 kJ mol-1 and 120.9 kJ mol-1 for sulfuric and acetic acid, respectively. The obtained kinetic parameters were then used to determine the optimum hydrolysis conditions for to convert LG into GLC.

The research described in Chapter 2 was extended to the use of a solid catalyst (Amberlyst 16) instead of sulfuric and acetic acid. Systematic studies in batch and in a continuous fixed bed reactor were performed, and the results are given in Chapter 3. Mass transfer limitations in this heterogenous catalytic system were carefully examined and accounted for in the kinetic modeling. In the batch setup, the effects of the reaction temperature (352–388 K), initial LG intake (100–1000 mol m-3), catalyst loading (1–5 wt%), and stirring rate (250-1000 rpm) on GLC yield were determined. The highest GLC yield was 98.5 mol% (at 388 K, 5 wt% Amberlyst 16, CLG,0 = 500 mol m-3 at 500 rpm stirring rate and 60 min). From the batch data, relevant kinetic parameters were determined using a first order approach including diffusion limitations of LG inside the Amberlyst 16 particles. The activation energy was found to be 132.3 ± 10.1 kJ mol-1. A good agreement between experiments and kinetic model was obtained, and the kinetic model was successfully applied to model the performance of the continuous setup. At a steady state in the continuous experiments, LG conversion (73 mol%) and GLC selectivity were in line with the kinetic model obtained in the batch reactor. Catalyst stability appears to be good, as shown by the experiments in the continuous packed bed reactor for up to 30 h times on stream.

During the processing of pyrolytic sugars (e.g., hydrolysis and subsequent conversion processes), solid formation occurs to a significant extent. These insoluble polymer byproducts are known as humins or biochar. Chapter 4 provides an experimental study on the (catalytic) pyrolysis and hydrotreatment of such humins with the aim to improve the techno-economic feasibility of pyrolytic sugar valorizations. For simplification, a model humin was made from a representative pyrolytic sugar by heating a sample under atmospheric pressure at 130 °C, for 22 h in the absence of an acid catalyst followed by water extraction at 100 °C for 5 h. The insoluble residue obtained after water extraction was used as the model humin for further studies (termed as PS-humin). PS-humin was characterized by elemental analysis, GPC, TGA, HPLC, GC-MS, FT-IR, and NMR. In contrast to typical humins obtained from GLC and FRC conversions, PS-humin is soluble in typical organic solvents (DMSO, THF, IPA), allowing characterization by NMR and GPC. From the analyses, it appears that the humin is oligomeric in nature (Mw of about 900 g mol-1), consisting of sugar and furanic fragments linked with among others (substituted) aliphatic, ester units and, in addition, phenolic fragments with methoxyl groups. PS-humin was used as a feed for catalytic pyrolysis and catalytic liquefaction experiments. Pyrolysis of the humin was carried out in an mg scale PTV-GC-MS unit at 550 °C in the absence and presence of a H-ZSM-5-50 catalyst. Analyses show that low molecular weight aromatics (benzene-toluene-xylene-naphtalene-ethylbenzene, BTXNE) were formed in 4.5 wt % yield on feed intake. The catalytic liquefaction reaction was studied in a batch autoclave reactor at 350 °C, for 4 h using isopropanol as both the solvent and hydrogen donor and a Pt/CeO2 (4.43 wt% Pt) catalyst. At optimized conditions, 80 wt% conversion of the humin feed to a product oil was achieved. The product oil from the reaction was characterized in detail using advanced analytical techniques (e.g., GCxGC-FID) and showed that the product oil contains high amounts of phenolics and aromatics (ca. 22 % based on GC detectables in product oil). These findings will have implications for the techno-economic viability of pyrolysis oil biorefineries, and show that potentially high value products can be obtained from humin type byproducts.

In Chapter 5, an experimental study on the conversion of thick juice (a crude SUC-rich intermediate in sugar refining) to HMF using sulfuric acid as the catalyst in a liquid-liquid system is described. The reactions were carried out in a batch reactor and in a continuous slug flow microreactor. Water was used as the reaction medium while methyl isobutyl ketone (MIBK) and 2-methyltetrahydrofuran (MTHF) were used as the organic extraction solvent. Addition of salts to the system to improve the partitioning of HMF was also considered. High selectivity and yields of > 90% (FRC based) for HMF were achieved in a biphasic reactor setup at 150 °C using thick juice as the feed with H2SO4 as catalyst and MTHF as the bioderived extraction solvent. At these conditions, the conversion of GLC obtained by SUC inversion was limited to < 10 mol %, allowing its recovery for further use. Interestingly, comparative experiments with purified SUC led to only 84 mol% HMF selectivity at > 95 mol% FRC conversion, showing that the use of the crude feedstock is of high interest and opens a new avenue for more cost-effective HMF production.

In the last chapter (Chapter 6), a comprehensive study on the conversion of thick juice to HMF is provided. The objective of this study was to identify the origin of the better performance of thick juice compared to (pure) SUC for HMF synthesis. The experiments were carried out in a batch reactor using aqueous SUC solutions with added contaminants typically present in thick juice such as salts and organic acids. The experimental concentration-time profiles were modeled using kinetic expressions for the individual reactions to quantify the observations and to conclude which of the contaminants in the thick juice are responsible for the better performance of this feed for HMF synthesis. It was found that the higher HMF selectivity obtained when using thick juice as the feed compared to pure SUC is mainly due to the presence of sulfate ions, which in combination with sulfuric acid result in a reduction of among others the rate of subsequent reactions of HMF to LA and humins.