University of Groningen
Experimental and modelling studies on the synthesis of 5-hydroxymethylfurfural from sugars van Putten, Robert-Jan
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record
Publication date: 2015
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA): van Putten, R-J. (2015). Experimental and modelling studies on the synthesis of 5-hydroxymethylfurfural from sugars. [S.n.].
Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license. More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne- amendment.
Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
Download date: 04-10-2021
Experimental and modelling studies on the synthesis of 5-hydroxymethylfurfural from sugars
Robert-Jan van Putten
ISBN 978-94-6259-502-6 ISBN 978-94-6259-504-0 (electronic version)
Experimental and modelling studies on the synthesis of 5-hydroxymethylfurfural from sugars
PhD thesis
to obtain the degree of PhD at the University of Groningen on the authority of the Rector Magnificus Prof. E. Sterken and in accordance with the decision by the College of Deans.
This thesis will be defended in public on
Friday 16 January 2015 at 16.15 hours
by
Robert-Jan van Putten
born on 18 August 1981 in Hilversum
Supervisor Prof. H.J. Heeres
Co-supervisors Dr. E. de Jong Dr. ir. J.C. van der Waal
Assessment committee Prof. A.A. Broekhuis Prof. R. Palkovits Prof. G. Centi
Table of contents
1 Preface 1 1.1 Biomass as source for energy, transportation fuels and materials 1 1.2 This thesis 3 1.3 References 5
2 Hydroxymethylfurfural, a versatile platform chemical made from 7 renewable resources 2.1 Introduction 7 2.2 Nutritional and toxicological aspects of HMF and its derivatives 11 2.2.1 HMF occurrence in our diet 11 2.2.2 Metabolic breakdown of HMF and derivatives 16 2.2.3 Toxicological effects of HMF and its derivatives 18 2.3 Dehydration chemistry 20 2.3.1 Neutral monomeric sugars 20 2.3.2 Disaccharides and polysaccharides 42 2.3.3 Sugar acids 44 2.3.4 Conclusion 46 2.4 Process chemistry 47 2.4.1 HMF formation in single-phase systems 47 2.4.2 HMF formation in biphasic solvent systems 100 2.4.3 HMF formation in ionic liquids 123 2.5 Process technology 155 2.5.1 Introduction 155 2.5.2 Kinetic studies on HMF formation 159 2.5.3 Reactor concepts 201 2.5.4 Separation and purification strategies 206 2.5.5 Pilot scale production of HMF 212 2.5.6 Technoeconomic evaluations of different modes of HMF 216 production 2.6 Relevance of 5-hydroxymethylfurfural as a platform chemical 220 2.6.1 Conversion of HMF to monomers for polymers 220 2.6.2 Fine chemicals 233 2.6.3 HMF as precursor of fuel components 259 2.7 Conclusions 261 2.8 References 264
3 The dehydration of different ketoses and aldoses to 5- 281 hydroxymethylfurfural 3.1 Introduction 282 3.2 Experimental section 285 3.2.1 High-throughput screening 285 3.2.2 Kinetic experiments 285 3.2.3 Determination of the kinetic parameters 286 3.2.4 DFT calculations 286 3.3 Results and discussion 286 3.3.1 High-throughput screening 286 3.3.2 Kinetic study 291 3.3.3 DFT calculations 293 3.3.4 Mechanistic aspect 295 3.4 Conclusions 298 3.5 References 299
4 A comparative study on the reactivity of various ketohexoses to furanics in 301 methanol 4.1 Introduction 302 4.2 Experimental section 305 4.2.1 Chemicals 305 4.2.2 High-throughput experimentation 305 4.2.3 Synthesis of 2-methoxyacetylfuran 306 4.2.4 Chromatographic analysis 306 4.2.5 Experiments with L-[6-13C]sorbose 307 4.3 Results and discussion 307 4.3.1 High-throughput experimentation 308 4.3.2 Mechanistic considerations and 13C labelling experiments with 320 sorbose 4.4 Conclusions 326 4.5 References 328
5 Reactivity studies on the acid-catalysed dehydration of ketohexoses to 5- 329 hydroxymethylfurfural in water 5.1 Introduction 330 5.2 Experimental section 332 5.3 Results and discussion 332 5.3.1 Sugar reactivity 332 5.3.2 HMF yield 334 5.3.3 Product selectivity 337 5.4 Conclusions 338 5.5 References 339
6 Experimental and modelling studies on the solubility of D-arabinose, D- 341 fructose, D-glucose, D-mannose, sucrose and D-xylose in methanol and methanol-water mixtures 6.1 Introduction 342 6.2 Experimental section 345 6.2.1 Chemicals 345 6.2.2 Solubility measurements 345 6.2.3 UNIQUAC modeling 346 6.3 Results and discussion 348 6.3.1 Experimental studies 348 6.3.2 Modelling studies 351 6.3.3 Literature comparison 352 6.4 Conclusions 355 6.5 Symbols 355 6.6 References 356
7 Concluding remarks and recommendations 359
Summary 363
Samenvatting 367
Acknowledgements 371
Publications 373
|1
1 Preface
1.1 Biomass as resource for energy, transportation fuels and biobased chemicals
Our world is completely dependent on fossil resources for materials and energy production. Almost everything we use and consume in everyday life has a significant input of oil, coal and natural gas, including the production of food. The earth’s population is still growing, mainly in developing countries, and combined with economic development and the accompanying increase in consumption in countries like India and China creates an enormous pressure on earth’s resources. This is not a sustainable situation and it is of critical importance to develop alternatives for fossil resources for energy and bulk materials production.
For energy generation, renewables like solar and wind energy are available. For bulk materials production, though, sources of fixed carbon are required. This leads to biomass as the obvious solution, since it is the largest sustainable global source of fixed carbon. There are however major differences in the chemical composition of fossil carbon sources and biomass. Fossil feedstocks generally have a very high carbon and hydrogen content and are very low in heteroatoms like oxygen, sulphur and nitrogen. On the contrary, biomass generally has a very high oxygen content and also a higher nitrogen content than fossil feeds. Especially the amount of oxygen in the molecular structure of the biomass has to be reduced before it can be used in any potential application.
Apart from water, biomass mainly consists of carbohydrates, lignin, fatty acids, lipids and proteins of which the carbohydrates are the most abundant. This makes carbohydrates a very appropriate feed stock for a biobased economy. The conversion of carbohydrates into suitable building block molecules for the petrochemical industry in most cases requires the removal of the majority of the oxygen from the molecular structure. There are three main methods for oxygen removal: (i) removing small, highly oxidised molecules such as CO2, CO, formaldehyde and formic acid; (ii) hydrogenolysis, typically removing water at the cost of hydrogen; and (iii) dehydration. The removal of highly oxygenated compounds like CO2 2| Chapter 1 comes at the cost of carbon loss, a significant disadvantage when the carbohydrate acts as a carbon source. Hydrogenolysis requires at least one molecule of hydrogen for each oxygen atom that is removed. This is only applicable in bulk application when a sustainable source of hydrogen can be used. Dehydration, when possible, is very appealing as it retains all the carbon atoms from the carbohydrate by removing water exclusively.
Avantium Chemicals B.V., founded in 2000, is a spin-off from Royal Dutch Shell and based in Amsterdam, the Netherlands. Its core business revolves around advanced catalysis research, selling both services and high-end equipment to a broad customer base, consisting of some of the world’s largest chemical companies. Advanced catalysis research deals with high-throughput experimentation on a small scale and an infrastructure of experimental design and analytics, which allows the generation of large amounts of data at the cost of relatively little resources (time, material). In 2006 Avantium started the development of its own process for the production of biobased plastics by applying their advanced technology to biomass conversion. This YXY project deals with the production of a polyethylene terephtalate (PET) replacement named polyethylene furanoate (PEF). PET is one of the most used plastics on the planet with a global production of around 65 million tons per year. Replacing it with PEF would result in a significant reduction of non-renewable energy use and thus increasing sustainability.1 PEF is synthesised by polymerising 2,5-furandicarboxylic acid (FDCA), which replaces terephtalic acid used in PET, and ethylene glycol (Scheme 1). PEF is not just favourable compared to PET from an environmental point of view, but it also has strongly improved material properties.2,3 The barrier properties, for instance, are better than for PET and show much lower permeability for most gases, especially CO2 and oxygen. Therefore PEF has solid potential for use in packaging of oxygen sensitive products, like fruit juices and alcoholic beverages, an area in which PET is underrepresented.
Scheme 1: PEF synthesis from FDCA and ethylene glycol Preface |3
FDCA can be formed by oxidation of 5-hydroxymethylfurfural (HMF), a molecule formed through acid-catalysed dehydration of six-carbon sugars, categorised as hexoses, see Scheme 2 for details. This reaction is already known since the 19th century.4,5. Especially the last decade has seen an enormous increase in the development of novel and improved HMF chemistry from the scientific community.
Scheme 2: The formation of FDCA from hexoses through HMF
1.2 This thesis
In this thesis the reactivity of different C6 sugars (hexoses) in the acid-catalysed dehydration to HMF and derivatives was investigated. This was done with two main objectives in mind. The first objective was to gain new insights in the reaction mechanism of the dehydration reaction of hexoses to HMF. Both experimental and theoretical studies were performed on various sugars to establish structure-reactivity relations. The experimental part included high- throughput experimentation and experiments with a 13C-labeled sugar. Kinetic modelling and density functional theory (DFT) calculations were then performed and compared to the experimental results. The second objective was to identify the most suitable hexose for HMF production from a selectivity and activity point of view. This was done by performing experimental studies with a range of hexoses in either water or alcohol solutions using a cheap inorganic Brønsted acid as the catalyst.
The dehydration of six-carbon sugars, like glucose and fructose, yields 5- hydroxymethylfurfural (HMF). In Chapter 2 the state of the art on the synthesis and applications of HMF, as well as its natural occurrence is discussed in detail. The goal of this large overview, unlike the studies performed by others, was to combine all known work on HMF chemistry and to come up with a comprehensive review. In water, the traditional
4| Chapter 1 solvent for sugar chemistry for solubility reasons, HMF yields from fructose are always around 50%, whereas HMF yields from glucose, the preferred feedstock based on availability and price considerations, in general does not exceed 5%. Significant advances in HMF yield and selectivity have been made by applying aprotic polar solvents such as DMSO, and in the last decade, ionic liquids. For fructose, HMF yields over 90% have been reported, and in combination with Lewis acid catalysts HMF yields of 70-90% from glucose have been mentioned. Despite this impressive progress, a number of key questions on sugar conversion to HMF still need to be answered. For smart catalyst selection strategies, for instance, more information is required on the reaction mechanism of the actual dehydration reaction. A number of theories have been proposed, but hard and sound experimental evidence is lacking. Hardly any research has been performed on proposed intermediates and C6-sugar structure- reactivity relations in order to identify key structural properties within the C6 sugars are lacking. Furthermore observation of proposed intermediates with spectroscopic techniques has proven difficult.
To shed light on the reaction mechanism of especially fructose dehydration and why glucose and fructose show such different reactivities towards HMF, we set out to explore especially the structural properties of the sugars in relation to HMF formation. The main focus was to investigate the effect of the relative orientation of the hydroxyl groups on the hexoses on the selectivity of the dehydration reaction to HMF.
In Chapter 3 the acid-catalysed dehydration of three aldose sugars (glucose, galactose and mannose) and three ketose sugars (fructose, sorbose and tagatose) in water is investigated. High-throughput experimentation was combined with kinetic studies and density functional theory (DFT) calculations to study and explain the differences in reactivity between aldoses and ketoses, and between the different ketoses. The consequences of these findings for the reaction mechanism of HMF formation will be discussed.
Exploratory studies on the acid-catalysed dehydration of the four ketohexoses (fructose, sorbose, tagatose and psicose) in methanol using sulphuric acid are described in Chapter 4. High-throughput experimentation was used and significant differences in reactivity between the sugars were observed. Additional experiments with 13C-labelled sorbose were performed to relate the position of the carbons in the sugars to those in the products. The results shed new light on the mechanism of sugar dehydration to HMF. Preface |5
In Chapter 5, the acid-catalysed dehydration of all four possible ketoses (fructose, sorbose, tagatose and psicose) to HMF using sulphuric acid as the catalyst in water is described. The results are compared to those obtained in methanol (Chapter 4) and the DFT calculations presented in Chapter 3.
For the development of an industrially viable HMF process, high substrate concentrations in the reaction mixture are preferred to reduce large solvent recycle streams. Therefore, in Chapter 6, a solubility study of different sugars (fructose, glucose, mannose, xylose, arabinose and sucrose) in methanol and methanol-water mixtures is presented. The studied sugars are all relatively abundant in nature, both in their monomeric or dimeric form as well as in various polysaccharides. The experimental results, obtained with high- throughput experimentation, were modelled using an appropriate thermodynamic (UNIQUAC) model.
In the final chapter (Chapter 7), the major findings of the research presented in this thesis are summarised and recommendations for further research in this challenging research field are provided.
1.3 References (1) Eerhart, A. J. J. E.; Faaij, A. P. C.; Patel, M. K. Energy Environ. Sci. 2012, 5, 6407-6422. (2) De Jong, E.; Dam, M. A.; Sipos, L.; Gruter, G. J. M. In ACS Symp. Ser.; Smith, P. B., Gross, R., Eds., 2012. (3) Burgess, S. K.; Leisen, J. E.; Kraftschik, B. E.; Mubarak, C. R.; Kriegel, R. M.; Koros, W. J. Macromolecules 2014, 47, 1383-1391. (4) Düll, G. Chem. Ztg. 1895, 19, 216. (5) Kiermayer, J. Chem. Ztg. 1895, 19, 1003.
6| Chapter 1
|7
2 Hydroxymethylfurfural, a versatile platform chemical made from renewable resources*
2.1 Introduction
Over the last century, the world has become increasingly dependent on oil as its main source of platform chemicals and energy. Driven largely by the strong economic growth of India and China, demand for oil is expected to increase significantly in the coming years. This growth in demand, combined with diminishing reserves, will require the development of new, sustainable sources for fuels and bulk chemicals. Biomass is the most attractive alternative feedstock, as it is the only widely available carbon source apart from oil and coal. Biomass consists of carbohydrates, lignin, fatty acids, lipids, proteins and others. Carbohydrates in particular show promise, as they form by far the largest natural source of carbon. The main drawback of carbohydrates as a feedstock is the overabundance of oxygen within their molecular structures. There are three main ways of lowering the oxygen content of carbohydrates. The first option is the removal of small, highly oxidised carbon molecules such as CO2, formaldehyde and formic acid. An example is the fermentative conversion of carbohydrates into ethanol, butanol and CO2. The second option is hydrogenolysis, which typically removes oxygen from the molecule by forming water at the expense of one molecule of hydrogen for each oxygen atom. The third option is the removal of water, exemplified by the dehydration of carbohydrates into a wide variety of interesting compounds, especially furans and levulinic acid. One class of dehydration products, the furan compounds, is considered by many to have especially high potential for the production of chemicals and fuels. Bozell recently published an updated evaluation of the US Department of Energy (DOE) top 10 list of bio-based chemicals,1 where furan molecules such as 5-hydroxymethyl-furfural (HMF), furfural and 2,5-furandicarboxylic acid are mentioned in the ‘Top 10 +4’ as additions to the original DOE
* Published as ‘Hydroxymethylfurfural, a versatile chemical made from renewable resources,’ Van Putten, R.-J.; Van der Waal, J. C.; De Jong, E.; Rasrendra, C. B.; Heeres, H. J.; De Vries, J. G., Chem. Rev. 2013, 113, 1499-1597.
8| Chapter 2 list.2 In this review, the authors focus on one particular route, namely the dehydration of hexoses to obtain furan-type platform chemicals, HMF in particular. The first report of sugar dehydration under aqueous acidic conditions dates from 1840, by the Dutchman Mulder.3 He described the formation of formic acid, and what was later found to be levulinic acid4, from sucrose. Another important sugar dehydration product is furfural, the formation of which Döbereiner discovered in the early 19th century by the action of manganese oxide and sulfuric acid on sugars, as reported by Newth.5 The first publications on HMF (1, Scheme 1) synthesis, go back as far as 1895 by Düll6 and by Kiermayer.7 Since then, there has been continued and growing interest in furan derivatives as important compounds in our diet and as feedstocks with great potential for bulk chemicals and fuels production. Especially the last few years have seen an enormous increase in the number of publications on HMF chemistry, as described in Figure 1.
Table 1 provides an overview of the physical and chemical properties of HMF.
160
140
120
100 HMF synthesis publications 80 HMF synthesis patents 60 HMF use publications 40 HMF use patents
Number Number publications of 20
0 1900 1920 1940 1960 1980 2000 2020 Year
Figure 1. The number of publications on HMF per year, as registered by Web of Science
Scheme 1. 5-Hydroxymethylfurfural (HMF)
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|9
Table 1. Chemical and Physical properties of HMF
CAS Registry Number 67-47-0
EC-No 200-654-9
Chemical Abstracts Name 5-(Hydroxymethyl)-2-furancarboxaldehyde
Synonyms 5-(Hydroxymethyl)-2-furaldehyde; HMF; 5-(Hydroxymethyl)-2- furancarbonal; 5-(Hydroxymethyl)-2-furfural; 5-Hydroxymethyl-2- formylfuran; 5-Oxymethylfurfurole; Hydroxymethyl furfuraldehyde
Molecular Formula C6H6O3
Smiles C1=C(OC(=C1)C=O)CO
Molecular Weight 126.11
Description Yellow powder;8 odour of chamomile flowers9
Boiling Point 110 °C at 0.02 mm Hg;9 114-116 °C at 1 hPa8
Melting Point 31.5 °C;9 28-34 °C8
Solubility Freely soluble in water, methanol, ethanol, acetone, ethyl acetate, dimethylformamide; soluble in ether, benzene, chloroform; less soluble in carbon tetrachloride; sparingly soluble in petroleum ether9
Density 1.243 at 25 °C8
Refractive Index 1.5627 at 18 °C9
Flash Point 79 °C - Closed cup8
UV Absorption Maximum 283 nm
10| Chapter 2
The production of furan-type compounds, especially HMF, from carbohydrates has been reviewed regularly. The first was published in 1951 by Newth.5 During the rest of the 20th century a number of reviews appeared.10-12 Lewkowski’s furan chemistry review, published in 2001, provides a comprehensive overview of the history of HMF synthesis and its fields of application13. In 2004 Moreau et al. updated this with a review on furan chemistry14, followed by updates in 2010 by Tong15 and 2011 by Rosatella.16 Recently Stark,17 Ståhlberg,18 Lima19 and Zakrzewska20 reviewed the use of ionic liquids as green and benign solvents for selective sugar dehydration. A broader picture of biomass conversion into useful chemicals by Corma in 200721 also included sugar dehydration and subsequent conversion of the furan products obtained in useful chemicals and polymers. This review addresses both the general mechanistic aspects of the dehydration reaction from the earlier literature and the recent progress in HMF synthesis. Special emphasis is placed on the use of various solvent systems, and on the different production routes reported in the last decade. The impact of the recent progress on the development of economic production routes (catalysis, purification, etc.) will also be discussed, as well as an overview of chemical conversions possible from HMF, leading to a wealth of interesting products. The literature until 23 June 2012 has been taken into account. As can be seen in Scheme 2 a number of important C-6 compounds can be formed through one common intermediate molecule, namely HMF. Alkoxymethylfurfurals (2), 2,5- furandicarboxylic acid (3), 5-hydroxymethylfuroic acid (4), bishydroxymethylfuran (5), 2,5- dimethylfuran (6) and the diether of HMF (7) are furan derivatives with a high potential in fuel and/or polymer applications. Some important non-furanic compunds can also be produced from HMF, namely levulinic acid (8), adipic acid (9), 1,6-hexanediol (10), caprolactam (11) and caprolactone (12). The difficulty of achieving a highly selective process with a high isolated yield has thus far resulted in a relatively high cost price of HMF, restricting its potential as a key platform chemical.22 However, this may change in the near future. More details on the economics of HMF products will be discussed in section 2.5.6.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|11
Scheme 2. HMF as a platform chemical
Concurrent with these chemical developments the natural occurrence and nutritional and toxicological relevance of HMF and its derivatives in our diet have received increased attention. This will be reviewed in the light of their social and economic relevance in the large scale application of bio-based chemicals.
2.2 Nutritional and toxicological aspects of HMF and its derivatives
2.2.1 HMF occurrence in our diet
Ever since mankind started heating their food, furan compounds have been part of the human diet, as HMF is formed during the thermal decomposition of carbohydrates. Nowadays, HMF is a recognised indicator of non-enzymatic browning, and it is often used as an index of deteriorative changes that take place during excessive heating and/or prolonged storage of foods. However, HMF and its derivative 2,5-methoxymethylfurfural (MMF) can also be isolated from different plants, such as Schisandra23 and Cornus officinalis,24 and the marine red algae Laurencia undulata25 (HMF) and Jaborosa magellanica, a member of the Solanacea family26 (MMF). HMF is formed as an intermediate in caramelisation27 and Maillard reactions.28,29 Caramelisation occurs during dry heating and roasting of foods with a high concentration of carbohydrates. It starts at relatively high temperatures and depends on the type of sugar. Caramelisation reactions of fructose start already at 110 ˚C, while other hexoses caramelise
12| Chapter 2 above 160 ˚C. Maillard reactions can already occur at room temperature. The Maillard reaction is named after the French scientist Louis Camille Maillard (1878-1936), who studied the reactions of amino acids and carbohydrates.28,29 In many cases, such as in coffee, the flavour is a result of both Maillard (described in section 2.3.1.1) reactions and caramelisation. These reactions occur during many different food preparation conditions such as baking, smoking and roasting. Therefore, HMF is present in many different food items (a.o. honey, barley, brandy, citrus juices, tomato products, syrup, grape juice, freeze-dried pears, wine, coffee, caramel products, dried fruit, prune juice and bread) and every person is exposed to HMF and some of its derivatives. Contact with HMF can occur by ingestion, inhalation, or skin absorption. Information concerning the human daily dietary exposure is scarce. In fresh foods, the HMF level is close to zero.30 However, it is found to be at a significant level in processed foods. The daily intake of HMF from heat-processed food by the Spanish population was recently assessed based on the HMF levels in Spanish foods previously published.31 A mean HMF intake of 10 mg/day was analysed, which is only ten-fold lower31 than the advised tolerable daily intake of 132 mg person-1day-1. However, there is no consensus yet among the scientific community what the tolerable daily intake should be.31-36 Coffee and bread are the most important food items that contribute to HMF exposure, at nearly 85% of the total ingested amount. Biscuits, breakfast cereals, beer, pasteurised milk and tomato products are also important sources of HMF exposure.31 A median level of 5.26 mg HMF/day was calculated for HMF intake by coffee consumption.37 Ulbricht et al. estimated a human ingestion up to 150 mg HMF/day,38 whereas from a recent paper by Delgado-Andrade et al.39 a mean HMF intake of 5.1 mg/day for Spanish adolescents was calculated. Whatever the value obtained, it exceeds the intake of other food processing contaminants, such as acrylamide and furan, by many orders of magnitude.31 Table 2 gives an overview of the HMF concentrations found in a number of food items. Capuano shows that whatever the formulation and toasting temperature, HMF formation followed a first order kinetic. Moreover HMF formation was highly affected by the residual moisture content of the sample.40 A water activity, defined as the vapour pressure of water in a substance divided by that of pure water at the same temperature, of 0.4 has been considered critical, since it reflects a stage in baking where the temperature of bread crisp begins to rise above 100 °C, which greatly accelerates HMF formation.41 In 40 commercial honey samples from 12 different floral origins 3-deoxyglucosone (3-DG) was detected. The concentrations of 3-DG, a precursor for HMF, ranged from 75.9 to 808.6 mg/kg and were significantly
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|13 higher (up to 100-fold) than those of HMF.42 Also in 12 commercial high fructose corn syrup (HFCS) samples high concentrations of 3-DG (194-730 mg/l) and glucosone (32-401 mg/l) were found.43
Table 2. The occurrence of HMF in different food items
Food Stuff HMF (mg/kg dry mass) Treatment Temperature Time (min) (°C)
Grain Products
Rye40 46.7 25 180
Wheat40 47.0 25 180
Whole Wheat40 17.2 25 180
Corn Flakes44 46.8
Breakfast cereals44,45 6.6 – 241
Cookies (sucrose) 46 430 10 250
Cookies (fructose) 46 215 9 250
Toasted bread47 392 4.2
Fruit products
Boiled Pomgranate530 514-3500
Boiled Grape Juice30 18-200
Plum jam44,45 12-13
Prunes48 237
Dried plums49 2200
Dried pears50 3500
Apple jam51 14.9 63
14| Chapter 2
Bilberry jam51 56.9 60
Must syrup52 3500-11000 75
Processed Grapefruit juice53 15.1
Raisins54 444
Honeys
Multifloral44,45 4.6-42.3
Acacia55 8.4-16.2
Citrus55 8.1-45.2
Malaysian56 3 – 1100
Sugarcane52 100-300
Caramels50 9500
Whole meals
Paella57 21.2 15 (+20) 180 (@100)
Churros57 19.5 5 180
Kid stew with garlic57 37.9 15 (+45) 180 (@120)
Coffees
Roasted coffee45,49 300-1900
Soluble coffees37,45,48,58 93-5130
In addition to HMF, some of its derivatives also occur in foodstuffs. For instance 5- ethoxymethylfurfural (EMF) is regularly observed in alcoholic beverages (Table 3) and 5- acetoxymethylfurfural (AMF) is a regular ingredient in vinegars and related products (Table 3). Masino found a strong correlation of AMF concentration with HMF concentration and with the sugar content. Several vinegars were grouped based on their HMF content.59 Balsamic vinegar of Modena (BVM) showed concentrations ranging between 300 and 3300 mg/l and even higher concentrations of up to 5500 mg/kg were found in traditional balsamic
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|15 vinegar (TBV) samples.60 Other authors reported similar results for BVM.61,62 Most likely AMF can be found in all vinegars with high HMF concentrations as reported in Table 4, but has probably not been looked for by other authors.
Table 3. The occurrence of HMF and EMF in different alcoholic beverages
HMF EMF Reducing sugars Treatment
(mg/l) (mg/l) (g/l)
Madeira wine
Boal63 74.3 10.9 20.9 3 month at 50 °C, 25 years at ambient and pH 3.5
Malvazia63 100.3 13.2 29.6 3 month at 50 °C, rest at ambient and pH 3.4
White wine
Average 13 wines64 Not analysed 0.2 4 years at 5 °C
White Grenache, and 54 0.43 0.5 years at 37 °C Macabeu65
Red Wine
Sweet fortified Black 149 0.0 0.5 years at 37 °C Grenache65
Beer
Beer66 3.3 – 9.2
Tsingtao Beer67 0.4-2.9 pH 4.1
Brandy
Brandy68 20-155
16| Chapter 2
Table 4. The occurrence of HMF in different vinegars
Food Stuff HMF AMF Reducing sugars Treatment
(mg/l) (mg/l) (g/l)
Balsamic vinegar60 300-3200
TBVa,69 1590-3200 trace 350-700
TBVa (gem 6)59 2670 68.4 65.7 pH 2.46
Trebbiano Grape must for 3100 n.d. 354 30 h at 85 °C TBVa,70
Salamino Grape must for 145 n.d. 445 42 h at 85 °C TBVa,70
Balsamic Wine Vinegar71 220-480 n.d. 540 d at pH 2.8
TBVa,72 2900 – 3900 10-71 67 -70 pH 2.4-2.5
a: Traditional Balsamic Vinegar
The occurrence of HMF and its derivatives is not restricted to pure carbohydrate food sources and products thereof. Capuano observed that the HMF formation in sugar-amino acid model systems is much higher than in sugar systems at temperatures above 100°C and low moisture content.40 HMF is the most studied thermal degradation product of sugar, but other furanic congeners (e.g. furoic acid, furfural and acetoxymethylfurfural) could be quantified in traditional balsamic vinegars.59,73 Humans can also be potentially exposed to HMF through pharmaceutical preparations and cigarette smoke.74 In conclusion, HMF and its derivatives are available in wide variety of human food sources. It is more pronounced in processed foods than in fresh products, but it is nearly impossible to avoid its in-take. Estimated daily consumption of HMF and its derivatives is calculated to be below 5–150 mg/per person.28,30-32,67
2.2.2 Metabolic breakdown of HMF and derivatives
The estimated daily intake of HMF is 5–150 mg/per person based on reports by several authors (Janzowski et al.;75 Rufián-Henares;31 Arribas-Lorenzo;37 Delgado-Andrade et al.;39 Ulbricht et al.38). The main degradation product of HMF detected in urine is 5-
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|17 hydroxymethyl-2-furoic acid (4, HMFA), its concentration was in the range of 0 to 100 mg/L with most of the samples around 10 mg/L.76 Jellum, Borresen and Eldjarn determined that fructose solutions used for parenteral nutrition can contain up to 1.2 g/L HMF.77 Patients who obtained these fructose solutions metabolised 50% of the HMF to 4 and FDCA (3, Scheme 3) in urine samples.69 These compounds have also been detected by others.78-80 Additionally, a reaction of a carboxylic acid with an amino acid is possible. Prior, Wu and Gu80 detected an HMFA - amino acid conjugate, e.g. N-5-(hydroxymethyl)-2-furoylglycine (13) at 3.6% of the amount of HMF administered (Scheme 3). Furthermore, Prior et al.80 detected metabolites in urine derived from 13, like 5-((carboxymethyl)carbamoyl)furan-2- carboxylic acid (14, 4.2% of the administered HMF dose, also detected by Godfrey et al.78) and 5-(methylcarbamoyl)furan-2-carboxylic acid (15, 1.8% from the HMF dose). In addition to the above mentioned metabolites, HMF can be bioactivated to 5- (sulfoxy)methyl furfural (16, Scheme 3), through an enzymatic esterification of its hydroxy group by sulfotransferases.81
Scheme 3. The biotransformations of HMF (adapted from Glatt and Sommer;82 D. Jöbstl et al.76)
In conclusion, ingested HMF is converted to a number of oxidised metabolites which are excreted from the human body via urine. This has been confirmed by experiments on rats. In rats orally dosed HMF was eliminated for 95-100% after 24 hours, almost exclusively via the
18| Chapter 2 urine.79 The biochemistry and genetics of the microbial degradation of HMF and furfural has recently been reviewed. The oxidation and/or reduction to the furanic alcohol and acid forms constitute the initial steps of the HMF and furfural degradation pathways.83
2.2.3 Toxicological effects of HMF and its derivatives
Being part of the human diet, the toxicological effects of HMF, its derivatives and metabolites should be considered. HMF was once thought to have only certain negative side- effects,81 but more and more papers have appeared indicating that HMF can also have a positive pharmacological activity such as propelling blood circulation and anti-oxidant activity84-86 and activity against sickle cell disease.87 Ding, Wang and Cai88 and Wang et al.89 recently showed that HMF extracted from wine-processed Fructus corni could inhibit hepatocytes oxidative damage caused by H2O2. It was suggested that the hepatoprotective effects of HMF might be related to resisting apoptosis. It is claimed that this is the first report on the direct morphological protective effects of HMF against chemical liver cell injury in vitro. Sriwilaijaroen et al. showed that Mumefural (18, Scheme 4), the monoester of HMF and citric acid, and related HMF malic and citric acid esters from Japanese apricot fruit juice concentrate show multiple inhibitory effects on pandemic influenza A (H1N1) virus.90
Scheme 4. Mumefural
Up to now, it is not clear whether normal dietary human exposure to HMF represents a potential health risk, though it is known that HMF is cytotoxic at high concentrations, causing irritation to eyes, upper respiratory tract, skin and mucous membranes.32 Both the national toxicity program of the National Institute of Environmental Health Sciences (NIEHS) as well as the European Food Safety Authority (ESFA) journal extensively addressed the toxicity of HMF.32,33 At the time of those reports no positive or negative epidemiological studies or case reports associating HMF with a cancer risk in humans were identified in the available literature. HMF was also found to be inactive in standard genotoxicity tests. The toxicity of HMF has been reviewed by several authors.34,35 Janzowski performed an extensive study to elucidate the toxic potential of HMF by assessing cytotoxicity, growth inhibition,
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|19 mutagenicity, DNA damage and depletion of cellular glutathione were investigated in mammalian cells, genotoxicity was monitored in Salmonella typhimurium.75 The author concluded that HMF does not pose a serious health risk, even though the highest concentrations in specific foods approach the biologically effective concentration range in cell systems. Severin and co-workers investigated the genotoxicity of HMF using the Ames test. They concluded that HMF did not induce any genetic mutation in bacteria whatever the concentration in the Ames test. Furthermore, HMF does not induce clastogenic (disruption or breakages of chromosomes) or aneugenic (abnormal number of chromosomes) effects in the HepG2 cells. In contrast, it induced HepG2 DNA damage at concentrations from 7.87 to 25mM in the comet assay suggesting a weak genotoxic effect of HMF in the HepG2 cells which probably is repaired.91 5-Hydroxymethylfuroic acid, a metabolite of HMF in humans, was not mutagenic.92 However, HMF can be metabolised in vitro by sulfotransferases to a chemically reactive intermediate, 5-sulfoxymethylfurfural (SMF, Scheme 3, 16).81 The intrinsic mutagenic properties of SMF were enhanced by addition of extra chloride ion to the assay medium. In the resulting ester, the sulfate is a good leaving group, thus producing a highly electrophilic benzylic-like carbocation, which could be stabilised by distribution of charges on the furan ring. The subsequent interaction of this reactive intermediate with critical cellular nucleophiles (i.e. DNA 17, RNA and proteins) may result in toxic and mutagenic effects. The HMF derivative 5-chloromethylfurfural (CMF) shows a higher mutagenic and cytotoxic activity in bacteria than the sulfuric acid ester 81. It was recently found that direct parental administration of SMF to mice leads to abundant acute necrosis and proteinaceous casts in the proximal tubules as the dominating toxicological effect.93 Additional research provided evidence for the involvement of organic anion transporters in the renal accumulation of SMF. These transport characteristics could be responsible for the selective damage of renal proximal tubules by this reactive metabolite.94 In contrast, 5- acetoxymethylfurfural (AMF) was neither mutagenic nor cytotoxic under the same experimental conditions used for SMF and CMF.81 Both HMF and SMF are weak intestinal carcinogens in Min/+ mice.95 As was discussed before, HMF is mainly present in heat treated food. Contrary to most test organisms, humans have been exposed to heat treated food for thousands of years. This likely makes it difficult to directly translate the effects of exposure to HMF/SMF observed in rodents into similar effects in human beings. The so-called Maillard Reaction Products (MRP, see also section 2.3.1.1) are a group of compounds found in foods that is closely related to the HMF formation. They have been assessed for toxicological effects. It was found that the consumption of a diet rich in MRP,
20| Chapter 2 exemplified by a high content of HMF (3.8 mg/kg), correlates negatively with protein digestibility. Therefore, the possible effects of an excessive intake of MRP’s during adolescence needs further research, and also long-term effects should be evaluated.96 However, it is unclear what role HMF plays in the observed effect. Lee et al. reported that 3,4-dideoxyglucosone-3-ene (3,4-DGE), a possible intermediate/side product in the HMF formation and present in perennial fluids.97, induces apoptosis in leukocytes and renal tubular epithelial cells and it was recently shown that 3,4- DGE also promotes apoptosis on human peritoneal mesothelial cells.98 In their recent review Abraham and co-workers concluded that in vitro genotoxicity of HMF was positive when the metabolic preconditions for the formation of the reactive metabolite 5-sulfoxymethylfurfural were met. However, so far in vivo genotoxicity was negative. Results obtained in short-term model studies for HMF on the induction of neoplastic changes in the intestinal tract were negative or cannot be reliably interpreted as "carcinogenic". In the only long-term carcinogenicity study in rats and mice no tumors or their precursory stages were induced by HMF. Hence, no relevance for humans concerning carcinogenic and genotoxic effects can be derived. The remaining toxic potential is rather low. Various animal experiments reveal that no adverse effect levels are in the range of 80-100 mg/kg body weight and day. Therefore it was concluded that current safety margins are generally sufficient.35,36
2.3 Dehydration chemistry
In this chapter the dehydration of C-6 sugars and their derivatives including di- and polysaccharides will be discussed with an emphasis on the mechanistic aspects. In addition, an overview will be given of side products that can be formed.
2.3.1 Neutral monomeric sugars
The acid catalysed dehydration of neutral monomeric C5 and C6 carbohydrates has long been known to produce a wide variety of products. The most prevalent products reported are HMF, furfural and levulinic acid, with humins as unwanted polymeric side-products. Within this section the focus will be on the chemistry of HMF formation and to a lesser extent on the formation of levulinic acid from hexoses. It will not discuss the formation of furfural from pentoses, which is excellently described elsewhere.99-101 Furfural from hexoses, however, is discussed briefly as a side product in HMF formation. In this chapter the various reported mechanisms will be reviewed and discussed in section 2.3.1.1 and the formation of by-products in section 2.3.1.2.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|21
2.3.1.1 Mechanistic aspects
Based on experimental data several different routes for the formation of HMF and derivatives from sugars have been proposed, but no definite proof of the dehydration kinetics has been published. In general three main routes have been described. The first and most widely described route is the direct formation of HMF by acid catalysed dehydration of a hexose, in which three water molecules are consecutively removed from the sugar molecule (Scheme 5). The second route is possible in the presence of amino-acids and amines when the hexoses undergo Maillard reactions and finally HMF can be produced by aldol condensation reactions of smaller C3 molecules.
Scheme 5. The general dehydration route of hexoses to HMF
Direct dehydration of hexoses Several mechanisms for the direct formation of HMF by acid-catalysed dehydration of a hexose by elimination of three water molecules have been suggested in the literature. The mechanistic pathways can be divided in two general groups. One group assumes the reaction to proceed via acyclic intermediates5,11,12,102,103 and the other group assumes cyclic intermediates.5,11,12,102-104 The body of experimental evidence to support either of the mechanistic routes is still very small, and a consensus on the actual mechanism has not been reached. It should be noted that almost all the proposed mechanisms so far are based on research in aqueous systems. In recent years a shift towards the use of other solvent systems was observed; here, a different mechanism could be operative. The acyclic pathways (Scheme 6) assume as the rate-limiting step the formation of a linear 1,2-enediol 19,11,12,104 which is widely accepted as the intermediate in the aldose-ketose isomerisation by the so-called Lobry deBruyn-Alberda van Ekenstein (LBAE) transformation.105 This is followed by two consecutive β-dehydrations and a ring closure with a final water elimination to yield HMF.
22| Chapter 2
Scheme 6. The acyclic pathway in the dehydration of hexoses to HMF, as proposed by Anet103 Moreau et al. proposed an acyclic mechanism for the dehydration of fructose, based on the observation of small amounts of glucose and mannose, indicating an isomerisation of fructose through a 1,2-enediol species (19), in which the formation of furfural (20) and hydroxyacetylfuran (HAF, 21) are also explained (Scheme 7).102 The formation of furfural was attributed to a decarbonylation that competes with the final dehydration step to HMF, whereas the formation of HAF was proposed to take place through the consecutive dehydration of a 2,3-enediol species 22, formed through a rearrangement of the 1,2-enediol, also mentioned by Kuster.12 The HMF route in this scheme is in essence identical to the one presented by Anet (Scheme 6).102,103 The schemes by Moreau and by Feather and Harris both mention the formation of 3-deoxy- D-glucosone (3-DG, 23), a keto tautomer of the key intermediate enol 23a formed after the first dehydration step. Efforts by Anet in the early 1960’s led to the identification of many of the intermediates by looking at different hexoses and assumed intermediates.103,106,107 The formation of 23, observed as its (2,4-dinitrophenyl)osazone 24 (Scheme 8), was confirmed in the acid-catalysed formation of HMF from fructose by Anet.103,106,107 Similarly Wolfrom et al. reported the isolation of a methylated phenyl osazone intermediate 25 (Scheme 9) in the acid catalysed decomposition of 2,3,4,6-tetramethylglucoseen-1,2 (26).108 The 2,3,4,6-
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|23 tetramethylglucoseen-1,2 is essentially the tetramethylated pyranose derivative of 23. Interestingly the major product of the decomposition was methyl methoxyfurfural (MMF, 27). Recently Jadhav et al. compared the formation of HMF from fructose and 23 in which HMF formation from 23 was observed to proceed at a significantly higher rate than from fructose, which showed that 23 cannot be excluded as a potential intermediate in the formation of HMF from fructose.109 In addition to these acyclic pathways, several cyclic pathways have been proposed as well. The mechanisms of these cyclic pathways (Scheme 10) all start from the cyclic ketofuranose.
The first step is expected to be the dehydration of the hemiacetal at C2, forming a tertiary carbenium cation. This is then followed by two consecutive β-dehydrations in the ring to form the HMF.5,11,12,104 Recent 13C-NMR studies by Akien et al. in DMSO and other polar solvents are in agreement with this cyclic dehydration pathway by assigning peaks in the NMR spectrum to both intermediates mentioned in Scheme 10.110 Furthermore experiments in the presence of D2O showed that all steps after the first dehydration are irreversible, which was explained by the lack of deuterium incorporation in HMF. This also makes an acyclic pathway highly unlikely, since this would require deuterium incorporation at C-3 of HMF, through intermediate 23 (Scheme 7).
24| Chapter 2
Scheme 7. The formation of HMF, Furfural and hydroxyacetylfuran from Fructose102
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|25
Scheme 8. 3-DG and its (2,4-dinitrophenyl)osazone
Scheme 9. 2,3,4,6-tetramethylglucoseen and its phenyl osazone
Scheme 10. The cyclic pathway in the dehydration of fructose to HMF.
The first dehydration step from a cyclic D-fructofuranose would yield the enol form of 2,5- anhydro-D-mannose, a 2,5-anhydro sugar also referred to as chitose. Furanic compounds have been observed in the acid dehydration of 2,5-Anhydro sugars.103,111 The dehydration of 2,5-anhydro-D-mannose to HMF was observed to proceed slower than when starting from fructose. Based on this observation chitose was excluded as a likely intermediate in the dehydration of fructose to HMF.103,111 Anet cited work in which the yield of HMF from chitose (12%) was much lower than from fructose (20-25%).103 Dekker and Hashizume
26| Chapter 2 reported the formation of HMF from 2,5-anhydro-L-idose (28, Scheme 11), which reacted several hundred times faster than glucose.112 Feather and Harris explain the necessity of the formation of the C1-aldehyde group upon dehydration at C-2 as a driving force for a consecutive β-elimination by mentioning that hydroxymethyl furancarboxylic acid formation from the C1 carboxylic acid counterparts of 2,5-anhydro-hexoses required much harsher conditions than HMF formation from 2,5- anhydro-hexoses.11,113 Recent in situ 13C-NMR studies by Zhang and Weitz using [13C-1]fructose and [13C- 6]fructose confirmed that fructose C-1 forms the carbonyl carbon of HMF and fructose C-6 forms the hyroxymethyl carbon of HMF. This is in agreement with both the acyclic and cyclic mechanisms proposed.114
Scheme 11. 2,5-anhydro-L-idose
Another insight in the mechanism of HMF formation can be found in the different reactivity of fructose compared to that of glucose. In general, fructose is much more reactive and selective towards HMF than glucose. Kuster explains that glucose shows much lower selectivity for HMF formation, due to its more stable ring structure, which hinders its ability to form the acyclic enediol intermediate.12 The cyclic mechanism as postulated above (Scheme 10) requires glucose to isomerise to fructose prior to dehydration to HMF. Another point to take into account is the solvent and temperature dependence of the tautomeric distribution of the different sugars. At room temperature glucose exists almost exclusively in pyranose (6-ring) form when dissolved in water. For fructose in water at room temperature the two pyranose forms are prevalent, although the furanose forms are also present in considerable amounts.115 With regard to solvent dependence of the tautomer distribution it is interesting to consider the significant amount of research on HMF formation in other media than water. This includes both organic solvents and ionic liquids as reaction media. It has, however, been reported in recent NMR studies that under reaction conditions the tautomerisation of fructose is very rapid.110,116 When tautomerisation proceeds at a significantly higher rate than dehydration this is not expected to be an important factor in improving the selectivity of HMF formation.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|27
In DMSO the furanose forms are prevalent, adding up to around 70% at 20 °C,117,118 and this fraction increases with temperature.117,118 An NMR study on fructose dehydration in DMSO by Amarasekara showed that DMSO acts both as a solvent and as a catalyst.117 The results of this work show intermediates in the solution that are consistent with a cyclic dehydration mechanism (Scheme 12). This mechanism is very similar to the one described in Scheme 10, where water is present as the solvent. The difference is the coordination of DMSO on the oxygen atoms of C1 and C2 prior to dehydration.
Scheme 12. The proposed cyclic mechanism for fructose dehydration in DMSO by Amarasekara117 A recent publication by Binder and Raines119 describes the use of lithium halide additives in N,N-dimethylacetamide (DMA) as the solvent. A mechanism was described (Scheme 13) in which the halide functions as both a nucleophile and leaving group. It was observed that using bromide and iodide as additives gave significantly higher yields of HMF than in the case of chloride, which was explained by the consensus that bromide and iodide are both better nucleophiles and better leaving groups than chloride. It must be noted though, that the results presented with KCl as additive were obtained at lower temperature (80 °C) and lower salt content (1.5 % w/w) than those obtained with the bromide and iodide additives (100 °C, 10 % w/w). A kinetic study indicated a first-order dependence of the rate of HMF formation on iodide concentration. The authors did not mention why an acyclic mechanism could not explain the observations. Akien et al. recently suggested, based on 13C-NMR studies, that chloride has a negative influence on the formation of difructose dianhydrides.110
Scheme 13. The cyclic dehydration mechanism for nucleophilic halide catalysed fructose dehydration proposed by Binder et al.119
28| Chapter 2
Apart from fructose, other hexoses have also been reported as substrates in the formation of HMF and derivatives. In particular glucose has been studied to a great extent because compared to fructose it is much cheaper and potentially readily available from lignocellulosic feedstock. Typically the HMF yields from fructose are superior to those of glucose under the same reaction conditions. Isomerisation to fructose appears to be necessary in order to obtain HMF from glucose. Several studies on glucose decomposition in high temperature liquid water in the absence of catalyst report the formation of fructose.120-122 The formation of glucose from fructose was reported to be negligibly low under comparable conditions.122 In high temperature liquid water with a mineral acid104 and organic acid,123 isomerisation of fructose to glucose is also observed, though the amounts of observed glucose are always very small. Recent work on hexose dehydration, catalysed by a three dimensional mesoporous silica Al-TUD-1 at 170 ºC showed that starting from glucose the fructose yield was 16% at 61% conversion, whereas starting from fructose less than 3% glucose was obtained.124 The reported glucose yield was consistently around 3%, irrespective of conversion (20-100%). The same was reported in work by Antal in which reaction time, initial fructose concentration and acid concentration were varied.104 Bicker observed similar results in water in the presence of small amounts of 125 ZnSO4. These results imply that at high temperature in water, fructose is thermodynamically favoured over glucose. Work by Watanabe et al. on glucose reactions in hot compressed water confirms these observations, reporting fructose formation (5-6% after 200 s at 473 K) in the absence of a catalyst.126,127 This was later confirmed for even higher pressures and temperatures.128 HMF is also formed here, which could be explained by some isomerisation from glucose to fructose, followed by fructose dehydration to HMF. No glucose formation was reported upon treatment of fructose under the same conditions. This is consistent with Kabyemela’s work.122 Fructose did, however, yield 21% HMF in the absence of catalyst, which was much higher than when starting form glucose. Using anatase TiO2 (α-
TiO2) as catalyst, fructose and glucose behaved very similar with regard to isomerisation, 126,127 furfural formation and HMF formation. Watanabe claims that the α-TiO2 acts both as acid and base, because it enhances the HMF yield from glucose conversion. The HMF yield from fructose is the same in both the absence and the presence of this catalyst (around 20%), but the conversion is much higher in its presence (95% vs. 40%). Combined with the observation that glucose is formed when reacting fructose in the presence of this catalyst suggests that α-TiO2 behaves exclusively as a basic catalyst that catalyses the glucose- fructose isomerisation. A later publication from the same group using essentially the same
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|29 conditions, but using microwave heating, showed significant improvement in HMF yield by 129 adding α-TiO2 to the fructose reaction mixture. A number of studies on glucose decomposition also reported isomerisation to fructose.120-122 The formation of glucose from fructose, albeit in smaller amounts, was also reported in high temperature liquid water with both mineral acid104 and organic acid.123 Work by Yan et al. on fructose and glucose in DMSO using bifunctional heterogeneous 2- SO4 /ZrO2-Al2O3 type catalysts provided further clues on the requirement of glucose isomerisation to fructose for the formation of HMF.130 In their attempts to find an efficient catalyst for producing HMF directly from glucose, an optimum yield of 48% was obtained 2- using a SO4 /ZrO2-Al2O3 catalyst with a Zr/Al ratio of 1, compared to only 4% yield in the absence of the catalyst. The rate of fructose formation was found to be dependent on the amount of base sites on the catalyst. The highest HMF yield was 72%, obtained from fructose in the absence of catalyst. Adding catalyst showed a decrease in HMF yield with decreasing Zr/Al ratios. This was explained by increasing basicity of the catalyst with decreasing Zr/Al ratio. Interestingly, starting form fructose, the catalyst with the lowest Zr/Al ratio (1:10) showed the same yield (47%) as the best catalyst for glucose dehydration (48%). These results, combined with the observation that glucose does not dehydrate to HMF in DMSO in the absence of catalyst whereas fructose does, strongly indicate that this heterogeneous catalyst behaves exclusively as an isomerisation catalyst for the glucose-fructose isomerisation. The notion that glucose dehydration to HMF involves fructose as an intermediate has been challenged by Jadhav et al. They proposed a pathway through 3-deoxyglucosone (23, Scheme 6) to be dominant, based on the observation that in the dehydration of glucose to HMF in DMA/LiCl both fructose and were observed and that 23 was much more efficiently converted to HMF than fructose, showing a higher rate of HMF formation.109 Ishida and Seri performed a study on homogeneous lanthanide (III) salt catalysed glucose dehydration to HMF.131 They observed a clear, though non-linear relationship between the atomic radii of the lantanide ions and the initial rate of reaction. No change in the NMR spectra of glucose upon addition of the lanthanide ions was observed suggesting that displacement of water as a ligand to the lanthanide by glucose is slow Thus, the authors concluded the relationship between the atomic number (and thus the radius) of the lanthanide ion and the rate is determined by the ease of coordination with glucose. In glucose peritoneal dialysis fluids several glucose degradation products (GDP) have been detected. In addition to HMF (3.3 – 26 µmol/l), 3-DG (41-564 µmol/l) and 3,4-
30| Chapter 2 dideoxyglucosone-3-ene (3,4-DGE, 29) (1.2 - 36 µmol/l) were observed in relatively high amounts.97 These products can be formed from glucose by the removal of one and two water molecules, respectively (Scheme 14). Erixon mentioned that both molecules can be further dehydrated to HMF.132 However, for 3,4-DGE only hydration into 3-DG was shown.133 3,4- DGE has been identified as the most bioreactive GDP in those fluids. The concentrations of 3-DG and 3,4-DGE produced during heat sterilisation decreased when the pH was reduced to about 2.132 An extensive discussion on the role of intermediates between 3-deoxyglucosone and HMF is given by Anet .103.
Scheme 14. The formation of 3-DG and 3,4-DGE from glucose
In 2007 Zhang and co-workers made an important advance in the synthesis of HMF from 134,135 glucose. By using a system of 10 mol% of CrCl2 in ionic liquid (1-ethyl-3- methylimidazolium chloride, [EMIm]Cl) they were able to obatin 70% yield of HMF at 95% conversion. Fructose dehydration in the same system resulted in almost equal HMF yield.
Their proposed mechanism (Scheme 15) ascribes the role of CrCl2 as an isomerisation catalyst for the glucose to fructose isomerisation, followed by fructose dehydration to HMF in the acidic medium.
Scheme 15. The role of chromium in the isomerisation of glucose to fructose, followed by dehydration to HMF, as proposed by Zhang et al.134
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|31
Pidko and co-workers combined X-ray absorption spectroscopy (XAS), density functional theory (DFT) and kinetic experiments to confirm the role of Cr in the isomerisation of glucose to fructose.136,137 The EXAFS and DFT results provided strong clues towards a combination of a mono- and binuclear Cr system, in which ring opening of glucose is catalysed by a mononuclear Cr complex and the actual isomerisation to fructose proceeds through a binuclear Cr complex (Scheme 16). In this respect it is interesting to note that a two metal centre is generally accepted to be the active site of typical enzymes that can isomerise glucose to fructose with high selectivity.138-140,141 The role of ‘free’ Cl- in this system is important, functioning as a proton acceptor and forming a hydrogen-bonding network with the hydroxyl groups on the carbohydrate. A subsequent publication by the same group showed higher activity and selectivity towards HMF formation for Cr3+ than for Cr2+ catalysed reactions.142 This was attributed to the increased Lewis acidity of Cr3+ compared to Cr2+, resulting in a more efficient stabilisation of the negative intermediates formed during the isomerisation of glucose to fructose. An important observation in work by Yong et al. is the different behaviour of fructose and glucose towards changes in reaction time and substrate concentration.143 This suggests a different reaction mechanism for both substrates, which is in accordance with findings by Pidko.136 Up to an initial substrate/ionic liquid ratio of 0.5 an increase in the fructose concentration showed a gradual decrease in HMF yield. For glucose there does not appear to be a decrease in HMF yield upon increase of the initial concentration.143
Hu et al. performed glucose dehydration in [EMIm]BF4 with SnCl4 as the catalyst, obtaining HMF yields over 60%.144 A reaction mechanism was proposed as described in Scheme 17.
32| Chapter 2
Scheme 16. The proposed mechanism for Cr-catalysed isomerisation of glucose to fructose in ionic liquid, as proposed by Pidko et al. Reprinted from Ref.136 with kind permission from Wiley-VCH
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|33
Scheme 17. Mechanism of the SnCl4-catalysed isomerisation and dehydration of glucose to HMF, as proposed by Hu et al.144
Ståhlberg investigated the boric acid-catalysed glucose isomerisation/dehydration to HMF in ionic liquid.145 This work was based on the premise that boric acid forms an anionic borate complex with carbohydrates to facilitate the isomerisation of aldohexoses to ketohexoses.146,147 Based on DFT calculations and deuterium labelling experiments for the boric acid-catalysed isomerisation of glucose to fructose and subsequent dehydration to HMF a mechanism described in Scheme 18 was proposed. In fructose dehydration reactions the presence of >0.2 eq. boric acid, relative to fructose, had a negative effect on the HMF yield, which was explained by the formation of stable fructose-borate complexes. It was also concluded that boric acid acted mainly as an isomerisation catalyst rather than a dehydration catalyst. In the isomerisation a stabilising effect of borate coordination on the open chain form of glucose and the decrease in activation energy on the protonation of the aldehyde on C-1 were calculated. Deuterium labelling on C-2 of glucose yielded HMF with less than 5% deuterium incorporated in the HMF product, corresponding with isomerisation through an enediol mechanism, expelling almost all deuterium into the solution by formation of a ketone at C2. In the case of a hydride shift, 100% would still be incorporated in fructose and assuming a cyclic dehydration mechanism described in Scheme 10, Scheme 12 and Scheme 13 around 50% should still have been present in HMF.
34| Chapter 2
Scheme 18. Proposed mechanism for the boric acid-catalysed isomerisation of glucose to fructose, followed by fructose dehydration to HMF145
The currently available knowledge of the dehydration of fructose does not allow us to decide which of the proposed mechanistic routes is correct. One exception is Amarasekara’s NMR studies on fructose dehydration in DMSO, which clearly showed cyclic intermediates. The authors proposed adducts of the intermediates with DMSO that were not observed in NMR,117 making it inapplicable to dehydration in other solvents, water in particular. It is clear that fructose is much more readily converted to HMF than glucose. This is corroborated by results that show that an efficient isomerisation catalyst is required for efficient glucose conversion to HMF. This information makes a cyclic dehydration mechanism from fructose much more likely than an acyclic dehydration mechanism from an intermediate enediol, as the latter does not explain the large differences in reactivity and selectivity between fructose and glucose in their reactions to HMF, since the enediol would be a common intermediate.
HMF formation via Maillard reactions HMF and its derivatives are often reported to be present in food. In many cases it is believed to be formed by the so-called Maillard reaction. The first step of the Maillard reaction is the reaction of a reducing sugar with an amino acid, forming a so-called Amadori compound (30, Scheme 19a). Removal of the amino acid results in reactive compounds that are subsequently degraded to the important flavour components furfural and HMF (Scheme 19b). Another reaction pathway is the so-called Amadori-rearrangement, which is the starting
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|35 point of several browning reactions (Scheme 19b). The rate of the Maillard reaction and the nature of the products formed are mainly determined by the reaction conditions.148 The combined influence of time, temperature and pH is most relevant for the formation rate in food systems. Other less important factors are the chemical composition (nature of the reactants and type of buffer), water activity, the presence of oxygen, metals and reaction inhibitors (like sulfur dioxide). These factors thus have a high impact on the quality of processed foods. There is general consensus that a pH below 7 is necessary for substantial HMF formation.148
Scheme 19a.
Scheme 19b
Scheme 19. (a) The initial step of the Maillard reaction between glucose and an amino acid (RNH2), including the Amadori rearrangement, and (b) the subsequent formation of HMF.
36| Chapter 2
The Amadori compound 30 is depicted in the Fischer projection, which does not necessarily mean it is mainly in its open form.
Yaylayan and co-workers studied the chemical reactivity of 5-(hydroxymethyl)-2- furaldehyde (HMF) with lysine, glycine, and proline using isotope labeling technique. The formed products were identified using mass spectroscopy. A range of new adducts was discovered.149
HMF formation from reverse Aldol products of sugars
Cämmerer and co-workers investigated C3 sugar degradation products during the Maillard and caramelisation reactions.150 In both reactions traces of HMF (0.3%) were observed. A β- dehydration is claimed to take place on glyceraldehyde (31), leading to the formation of pyruvaldehyde (32), which then reacts with glyceraldehyde to form HMF (Scheme 20). Follow-up research by Murkovic and Bornik showed that reacting pyruvaldehyde with glyceraldehyde can lead to the formation of HMF (1.5 mol%) while combinations of glyceraldehyde and pyruvate resulted in substantial amounts (4.0 mol%l) of 5- hydroxymethylfuroic acid 4.151 Several research groups have reported the formation of glyceraldehyde and pyruvaldehyde in sugar dehydration.104,121,125,152,153 This indicates that a minor part of the detected HMF in those studies could be formed via aldol condensation of C3 retro-aldol degradation products.
Scheme 20. The β-dehydration of glyceraldehyde to pyruvaldehyde, followed by the formation of HMF from pyruvaldehyde and glyceraldehyde.150
2.3.1.2 By-products
An important factor in the synthesis of HMF through sugar dehydration is the occurrence of side reactions. The reported by-products include organic acids, other furans, aromatics, retro- aldol products, fructose dianhydrides154 and polymeric materials, including humins. An overview of these by-products is given in Table 5. Quantities of the different by-products
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|37 depend on feedstock and process conditions and in many cases quantitative data or even qualitative data on by-product formation are absent. It is generally accepted that the formation of levulinic acid and formic acid goes through rehydration of HMF and not directly from sugars. This is based on the general observation that the HMF yield has an optimum and that at prolonged reaction time an increase in levulinic acid yield is observed at the cost of the HMF yield.102,155-159 Recently an in situ 13C- NMR study with 13C labelled fructose on C-1 and C-6 revealed that the carbonyl carbon of HMF is incorporated in formic acid, whereas the hydroxymethyl carbon forms the methyl (C- 5) carbon in levulinic acid.114
Table 5. By-products of hexose dehydration
Furans Organic acids Aromatics Retro-aldol products Others
Furfural Levulinic acid 1,2,4-Trihydroxy- Pyruvaldehyde Fructose benzene dianhydrides 2-Hydroxyacetylfuran Formic acid Dihydroxyacetone Polymeric (21) material Acetic acid Glyceraldehyde HMF dimers (incl. humins) Lactic acid Erythrose
This notion was challenged by Pidko et al. for the formation of LA from glucose in recent computational work- in which a pathway from glucose to LA without HMF as an intermediate was calculated to be the most energetically favourable dehydration route.160,161 A computational study on the conversion of HMF to levulinic acid was published by Wang et al.162 High HMF selectivities are generally observed in the initial stage of the dehydration reaction, which was interpreted by Seri et al. that the by-product formation, especially humins, does not originate from the substrate only.163 Girisuta et al. published kinetic data on the rehydration of HMF to levulinic acid.164 Their research was mainly focussed on levulinic acid production from glucose165,166 and untreated biomass.167 In these studies HMF was observed as an intermediate. Research on autocatalysis in fructose dehydration, through the formation of organic acids during the reaction, by Kuster showed that HMF formation is pH dependent. Water-based
38| Chapter 2 fructose dehydration in the initial absence of catalyst showed HMF formation and a gradual decrease in pH to a value of 3.2.157 Furfural is another important by-product, that was found in particular in reactions performed in supercritical fluids.102,120,122 Until now it is not well understood if this is formed directly from the carbohydrate, from HMF or via another intermediate. The formation of furfural via 3,4-DGE (29, Scheme 21) was proposed by Kallury et al.168 and this pathway was incorporated in the dehydration mechanism proposed by Moreau (Scheme 7). In situ 13C- NMR experiments with 13C labelled fructose on C-1 showed for their specific conditions that C-1 on fructose converted into the carbonyl on HMF, as well as the carbonyl on furfural.114 The formation of furfural was not observed in all cases and appeared to depend on which catalyst was used. The result, however, was not in agreement with the mechanism described in Scheme 21, as this scheme proposes the C-2 of the hexose to convert into the carbonyl group of furfural.
Scheme 21. Furfural formation from 3,4-DGE168
Silberman reported work on the reactions of sugars in the presence of acids,169 which showed that at higher concentrations, glucose and other aldohexoses undergo condensation to form reversion products (mainly β-isomaltose or β-gentobiose), whereas ketoses are more prone to dehydration and subsequent humin formation.169 Glucose has been shown to condensate to anhydroglucoses, which apparently cannot be converted to HMF.121,128,170 Other products formed from glucose and fructose under aqueous sub- and supercritical conditions are erythrose (33), pyruvaldehyde (32), glyceraldehyde (31), dihydroxyacetone (34), lactic acid (35) and glycolaldehyde (36).121,128,171,172 Scheme 22 describes the formation of these compounds, among others, as explained by Aida et al.128,172 A complete list of glucose degradation products formed by sterilisation of glucose containing pharmaceutical solutions is provided by Witowski and Jörres (Table 6).173
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|39
Scheme 22. An overview of the thermal degradation routes of glucose and fructose in sub- and supercritical water,128,172 RA = retroaldol, Dehyd = dehydration, BR = benzylic acid rearrangement, Hydr = hydration
Table 6. Glucose degradation products identified in sterilised pharmaceutical solutions
HMF Formic acid (37)
Acetaldehyde Levulinic acid (8)
Formaldehyde 5-Hydroxymethylfuroic acid (4)
Furfural (20) 2,5-Furandicarboxylic acid (3)
Glyoxal 2-Hydroxyacetylfuran (HAF, 21)
Pyruvaldehyde (32) 3-deoxy-D-glucosone (3-DG, 23)
Valeraldehyde
Soluble and insoluble polymeric materials are another group of by-products. The insoluble polymers are generally referred to as humins. Dumesic et al. report decreased selectivity of fructose dehydration to HMF when controlled amounts of HMF were initially added to the
40| Chapter 2 reaction mixture, implying a possible reaction between HMF and fructose, or a derivative thereof.170 Stability tests of HMF in the absence of sugars under reaction conditions showed only marginal loss of HMF.174 It is generally assumed that the humins are formed from polymeric condensation of HMF with sugars. A similar effect is also observed in furfural production from pentoses where adding furfural to the pentose feed enhances the formation of humin-like materials.100 A recent 13C-NMR study by Akien et al. proposed the formation of humins through 2,6-anhydro-ß-D-fructofuranose (Scheme 23).94 Due to their complex structure and composition, the nature of the oligomeric materials and humins are until now not very well characterised. A recent IR study on humins formed from HMF by Patil and Lund did indicate the presence of the furan and hydroxymethyl groups from HMF and the aromatic ring of benzaldeyde in the humins.175 At the moment the most likely applications for these humins are either as a fuel, particularly to generate heat for the various unit operations of the HMF production, or as compost.
Scheme 23. Proposed pathway for humin formation
Zhang et al. studied the formation of carbonaceous microspheres from HMF formed in the dehydration of fructose, also proposing a mechanism for their formation.176
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|41
2.3.1.3 Computational Studies
In a recent computational study on glucose dehydration, through fructose, to HMF via a cyclic mechanism and subsequent hydration to levulinic acid (Scheme 24) the reaction energies were calculated using Gaussian and density functional theoretical (DFT) methods.177 The Gibbs free energies were calculated at 25 and 175 °C. At 25 °C the solvation effect of water was also taken into account. These calculations all showed an increase in Gibbs free energy for both the isomerisation of glucose to fructose and the first dehydration step of fructose at a combined ΔG298K around 40 kJ/mol. All subsequent steps resulted in a decrease in ΔG, especially the third dehydration (ΔG298K ≈ -92 kJ/mol), attributed to the formation of the stable furan ring, and even more for the rehydration to levulinic acid (8) and formic acid
(37) (ΔG298K ≈ -113 kJ/mol).
Scheme 24. The cyclic dehydration of glucose and subsequent rehydration to levulinic acid177
Calculations were also performed on solvent effects of water and DMSO. Water was found to have a significant stabilising effect on glucose compared to fructose whereas with DMSO this was not the case.177 Apart from solvent effects, these calculations do not explain the significant difference in the ease with which fructose is dehydrated to HMF compared to glucose, because activation energies were not taken into account. The computational calculations are in line with what has already been observed experimentally. The first step is rate determining and all subsequent steps are fast, explaining the difficulty in determining stable intermediates.
42| Chapter 2
Caratzoulas et al. performed a computational study on the dehydration of fructose to HMF at 90 °C.178 The calculations were made for the protonated intermediates. The biggest difference with Assary’s work177 can be found in the intermediate with the highest ΔG (around +210 kJ/mol), which was involved in the second dehydration, where Assary mentioned the first dehydration as the step with the highest increase in ΔG. The difference in calculated ΔG is also quite big at a factor of around 5. The same group recently published work in which the high HMF selectivity of fructose dehydration in DMSO is explained, showing a preferential coordination of DMSO around the HMF molecule, protecting it from side-reactions. Also a specific coordination of DMSO around the fructose molecule was found, preventing reversion and other side reactions.179 In a recent publication Qian reports calculations that indicate a pathway from glucose to HMF that does not involve formation of fructose as an intermediate, but rather through an aldehyde.180 This aldehyde is alike to anhydro-D-mannose, which is presented in its enol form in Scheme 10.
Guan et al. performed DFT calculations on MCl3 catalysed glucose conversion to HMF in 1-butyl-3-methylimidazolium chloride ([BMIm]Cl) ionic liquid.181 Calculations were performed with FeCl3, WCl3, CrCl3 and MoCl3. From these calculations it was concluded that
WCl3 would be a very promising catalyst for this chemistry, however, this was not confirmed by experimental data.
2.3.2 Disaccharides and polysaccharides
Most of the work published in the field of HMF synthesis is focussed on the conversion of monosaccharides. From an economical and environmental point of view, it is also interesting to consider polysaccharides. In addition to the challenges connected to the conversion of monosaccharides, these feedstocks have additional physical properties that complicate their application. The two most important factors to take into account are the reduced solubility in almost all solvents and the presence of glycosidic bonds between the sugar moieties that can be difficult to break. Isotope labeling studies on pyrolysis GC-MS by Perez Locas and Yaylayan focused on HMF formation from sucrose at 250 °C with the purpose of determining the relative contribution of the fructose moiety to the HMF yield.182 By 13C labelling on the fructose moiety, they determined that 90% of the HMF formed originated from the fructose moiety and 10% from the glucose moiety. At 300 °C HMF formation from sucrose, fructose and 3- DG was observed. The HMF yields from both fructose (4.5 fold) and sucrose (2.4 fold) were
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|43 significantly higher than from 3-DG (23). Based on these results 3-DG was ruled out as an intermediate in the formation of HMF from fructose or sucrose. Glucose only generated 0.16 fold HMF relative to 3-DG. Based on these results a mechanism was proposed (Scheme 25).
Scheme 25. A mechanism proposed by Perez Locas and Yaylayan182 for the dehydration of sucrose, fructose and glucose to form HMF (adapted from Perez Locas and Yaylayan182)
Research by Carlini et al. on sucrose and inulin dehydration showed that in an aqueous medium the hydrolysis of these compounds to monosaccharides was faster than the dehydration of fructose.183,184 This is consistent with work by Haworth and Jones, who also showed that the HMF produced from sucrose originates almost entirely from the fructose moiety.185 In the conversion of inulin and Jerusalem artichoke extracts, which are rich in polyfructans, hydrolysis was also observed to be much faster than dehydration.186 Mascal showed that cellulose can be readily converted into furan compounds using a mixture of LiCl in concentrated hydrochloric acid187 and research on La(III)-catalysed cellulose degradation in water by Seri et al. showed the formation of cellobiose (a D-glucose
44| Chapter 2 dimer), glucose, HMF and levulinic acid.188 Based on their research, Seri et al. proposed the reaction pathway as described in Scheme 26.
Scheme 26. The reaction pathway for La(III)-catalysed carbohydrate dehydration188
Research by Girisuta et al. on levulinic acid formation from cellulose under aqueous acidic conditions show HMF as intermediate in the conversion of glucose to levulinic acid (8).189 It was observed that higher temperatures lead to higher HMF amounts, which is consistent with observations by Peng et al. on levulinic acid production from cellulose.190
2.3.3 Sugar acids
The dehydration to furanic compounds is not exclusive to neutral sugars. Even though HMF is not the primary product formed, the dehydration of sugar acids (Scheme 27) is interesting as these substrates may have similarities in their dehydration. They could thus provide further clues towards the mechanisms of neutral sugar dehydration. In addition, the compounds obtained, can be seen as oxidation products from HMF and thereby an alternative route towards the highly interesting furan dicarboxylic acid (FDCA, 3) or 5-formyl furancarboxylic acid (38).
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|45
Scheme 27. An overview of the formation of 2,5-furan-dicarboxylic acid from sugar acids
Several literature references mention the acid-catalysed degradation of galacturonic acid under aqueous conditions. Most of this work deals with the quantification of sugars and studies the decarboxylation reaction via quantification of the released CO2 and furfural from the remaining pentose. The production of 5-formyl-2-furancarboxylic acid (FFCA) from galacturonic acid has been mentioned but no quantification has been reported. In the work of Popoff et al., the reaction of glucuronic acid and galacturonic acid in aqueous acidic media are reported to yield an array of products (Scheme 28), including furfural and 2-furoic acid (39).191 The yields of the majority of these compounds was higher from glucuronic acid than from galacturonic acid.
46| Chapter 2
Scheme 28. Detected products on the acid catalysed reaction of galacturonic acid and glucuronic acid.
Usuki et al. observed that the degradation of glucuronic and galacturonic acids proceeds at lower temperatures (140, 150 and 160 °C) than the degradation of pentoses, namely arabinose, xylose and lyxose (200, 220 and 240 °C).192 The formation of acidic compounds during degradation led to a fast pH decrease. The decomposition of glucuronic acid and galacturonic acid in subcritical water was kinetically analysed. In acidic solution at elevated temperature, hexuronic acids undergo decarboxylation, forming 2-furfuraldehyde and reductic acid (2,3-dihydroxyclopenten-1-one). The reaction is nearly quantitative and has been used as an analytical method for hexuronic acids. Only traces of 5-formyl-2- furancarboxylic acid were observed. In 1876 Fitting and Heinzelman synthesised 2,5-furandicarboxylic acid (FDCA) from mucic acid using concentrated hydrobromic acid, as described in Scheme 29. Several other catalysts and substrates have been tested subsequently. All the reactions required severe conditions (highly concentrated acids, temp > 120 °C, React time > 20h) and all the methods were non-selective with yields below 50%.13
Scheme 29. The synthesis of furandicarboxylic acid from mucic acid
2.3.4 Conclusion
We can conclude that there are still many uncertainties with regard to the mechanism(s) of hexose dehydration to furan compounds. A distinction can be made between mechanisms based on cyclic intermediates and mechanisms based on acyclic intermediates. The
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|47 differences within either group are only marginal and only limited evidence has been provided. The effect of the solvent is an additional factor to be taken into account, as for both DMSO and ionic liquids reaction mechanisms have been proposed in which the solvent plays an active role. It is therefore not inconceivable that the mechanism depends on the solvent. The literature strongly indicates that the first dehydration of fructose is the rate determining step in the formation of HMF, with all subsequent steps proceeding much faster. This makes it almost impossible to determine the reactive intermediates by analysis. A second complicating factor is the inherent reactivity of sugars, facilitating a wide array of side- reactions of the sugar. With regard to HMF formation from glucose, the information at hand indicates a reaction pathway through isomerisation to fructose. This presents an interesting challenge, because the glucose-fructose isomerisation is base-catalysed and the dehydration of fructose is acid- catalysed. In order to overcome this challenge an increasing amount of work has been published concerning bifunctional catalyst systems (vide infra).
2.4 Process chemistry
The development of efficient methods for HMF production from carbohydrates has been on-going for almost a century. Until the 1980’s research in this field almost exclusively focussed on the use of homogeneous acids as the catalyst in water-based reaction media, a solvent common for traditional sugar chemistry. The past three decades have seen a shift from water-based chemistry to alternative solvent systems in order to improve the overall yield of HMF. Based on the solvent system used, HMF synthesis from carbohydrates can roughly be divided into three types of processes: traditional single-phase systems; biphasic systems and ionic liquid-based systems.
2.4.1 HMF formation in single-phase systems
Going back as far as the early 20th century most research on the formation of HMF was performed in aqueous systems with mineral acids as catalysts (mainly sulfuric and hydrochloric acid). Water is an obvious choice of solvent because it dissolves the majority of the sugars in high concentrations unlike most organic solvents. Because of the relatively low HMF yields in aqueous systems, the use of organic solvents has grown since the 1980’s. Solvents like DMSO, DMF, DMA, acetone, acetic acid and methanol have been reported in the literature. The work discussed in this section will be divided in processes from fructose,
48| Chapter 2 glucose and polysaccharide dehydration, respectively. The substrate concentrations are generally described in weight percentage relative to the total weight of the solution and the catalyst concentrations are described in percentages relative to the amount of substrate where homogeneous catalyst loading is related to the molar amount of hexose monomers and the heterogeneous catalyst loading is given in weight percentages relative to the weight of the substrate. Yields, selectivities and conversions are described in molar ratios unless mentioned otherwise. The yield describes the amount of HMF formed relative to the starting amount of hexose monomers, the conversion describes the amount of sugar monomers converted into non-saccharides and the selectivity describes the amount of HMF formed relative to converted substrate.
2.4.1.1 Fructose dehydration in single-phase systems
Research on fructose dehydration in traditional single-phase liquid systems, including supercritical fluids, will be reviewed here. By far the highest HMF yields reported in the literature have come from fructose dehydration. HMF synthesis in aqueous systems will be discussed first, followed by a summary of the work done in organic solvents and aqueous/organic single-phase mixtures.
Fructose dehydration in water based systems This paragraph deals with the dehydration of fructose under aqueous conditions, excluding HMF synthesis in the presence of extracting solvent, which is discussed in section 2.4.2. The basic structure is selected as chronological, describing the evolution of the work in the last half-century, combined with clustering comparable work. In this way the initial work at lower temperatures and pressures is first discussed, followed by the more recent work at higher temperatures and pressures. At the end of the paragraph summarising tables and a brief conclusion are given. Process research on HCl catalysed fructose dehydration in water was published by Kuster and co-workers in 1977.155,156,193 The reactions were performed at 95 ºC under ambient conditions. The best HMF yields were around 30% and the highest selectivity around 60%. The high HCl concentration (0.25-2 M) in combination with the relatively low temperature favoured formation of levulinic acid over HMF, explaining the low yield/selectivity for HMF.155 The addition of polyethylene glycol-600 (PEG-600) as a co-solvent led to an improved dehydration rate and subsequently a reduced rehydration rate. A maximum HMF yield of around 70% at 86% conversion was obtained when using 70% PEG-600 in water.156
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|49
In the same series of publications Kuster and Temmink reported work on aqueous fructose dehydration catalysed by formic acid at 175 °C and 50 bar.157 With a fructose concentration of 0.25 M HMF yields of 50-60% with 80-100% selectivity were reported at pH ≈ 3 at reaction times around 1 h. At a controlled pH of 2.7, 56% HMF yield was reported at 56% conversion after 45 min. An experiment in the absence of catalyst was reported to yield 56% HMF at 70% conversion after 90 min and showed a gradual decrease in pH from 7 to 3.2, which indicated the formation of acids during the reaction.157 There is no reason to assume that starting the reaction under acidic conditions prevented the formation of organic acids that were formed during the initially uncatalysed dehydration of fructose, which makes the high selectivities obtained at pH 2.7 surprising. In some cases the yield was reported to be higher than the conversion, corresponding with selectivities over 100%. An important observation in the work of Kuster and Temmink is the increase in levulinic acid yield with decreasing pH (at pH >1). Following Kuster’s work, Van Dam tested the influence of a number of variables, namely substrate concentration, acidity, metal salts and water content, on the formation of HMF from fructose.158 A maximum HMF yield of around 25% at fructose conversions of 50-60% was obtained with 0.2 M fructose catalysed by 1 M p-toluene sulfonic acid, at 88 ºC after 4-5 h. The addition of 50 vol% of PEG-4000 increased the yield to around 50%. This is in line with the results of Kuster described above for PEG-600.156 Vinke and Van Bekkum later used activated carbon as adsorbent for HMF to obtain 43% yield at 72% conversion in the dehydration of 0.25 M fructose at 90 °C for 7 h.194 A gradual increase in the applied reaction temperature can be observed since the 1980’s. In the mid 1980’s, the Süddeutsche Zucker-Aktiengesellschaft developed a method for preparing HMF from fructose on 10 kg scale.195 From the dehydration of 25 wt% fructose in water, catalysed by 1 wt% oxalic acid at 135-142 °C for 130 min a 34% HMF yield at 61% conversion was claimed prior to isolation, of which 80% could be separated by chromatography. The presence of glucose, polysaccharides and humic material before purification is mentioned. With 1% oxalic acid relative to fructose El Hajj et al. reported an isolated HMF yield of 23% at 145 ºC for 2.5 h.196 Antal and co-workers continued in the field of homogeneous acid catalysed fructose dehydration under aqueous circumstances at even higher temperatures and pressures. The highest HMF yields obtained were around 50% at 95% conversion with 2 mM H2SO4 and 0.05 M fructose concentration at 250 °C and 340 bar.104
50| Chapter 2
In the 1990’s Carlini and co-workers published research on HMF synthesis in aqueous environment using heterogeneous catalysts.183,184,197,198 Very high HMF selectivities (85- 100%) using niobium phosphate based catalysts in batch experiments at 100 ºC were reported.183 These selectivities were, however, only obtained at fructose conversions between 25 and 35%. At higher conversions the selectivities dropped significantly. The highest reported selectivity at 50% conversion was around 60%. By applying intermittent extraction with MIBK, a strategy originally developed by Kuster,199 the selectivities could be kept high (98%) at higher conversions. Comparable results were reported by the same group with heterogeneous titanium and zirconium catalysts.184 Carlini and co-workers claimed HMF selectivities over 85% at fructose conversions around 50% using a cubic zirconium pyrophosphate catalyst at 100 ºC. A heterogeneous γ-titanium phosphate was reported to produce comparable results. Intermittent extraction with MIBK led to 67% HMF yield at 71% conversion. Work by the same group on fructose dehydration under aqueous conditions with vanadyl phosphate catalysts showed comparable results.198 Carniti et al. published work on niobium phosphate and niobic acid.200 Experiments were carried out in a continuous flow reactor under aqueous conditions at temperatures of 90-110 ºC. The niobium phosphate was found to be more active, but both catalysts showed the same trends with regard to selectivity vs. conversion. A 25% HMF yield was obtained at 77% fructose conversion. In a recent publication Carniti et al. observed HMF to be stable in water in the presence of niobic acid catalyst, showing no levulinic acid formation. In the presence of fructose, however, humin formation was observed.201 Seri et al. reported fructose dehydration in water catalysed by lanthanide(III) chlorides.163 With 0.67 mol% of La, Nd, Eu, Dy and Yb compared to the substrate at 140 ºC for 1 h HMF yields between 16 and 19% were reported.The HMF yield for the uncatalysed fructose dehydration was approximately 4%. For DyCl3 experiments the conversion was tracked over time, along with selectivity and yield. These data showed an HMF yield of 24% at around 60% conversion after 2 h. At low conversions (<10%) the HMF selectivity was over 90%. Lower temperatures were also tested, but this caused the reaction to proceed very slowly.
Very recently Deng et al. published work on ZnCl2 catalysed fructose dehydration, but this showed the best results (53% HMF at 97% conversion) in the presence of HCl, which is a 202 result that can also be expected in the absence of ZnCl2. The Research Center of Supercritical Fluid Technology from Tohuku University in Japan produced a number of publications on sugar dehydration at high temperatures and pressures in water (HHW).126-129,172 Fructose degradation was studied in a flow process at 350 ºC and
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|51
400 ºC at 40, 80 and 100 MPa, showing HMF formation in the absence of catalyst. An HMF yield of 7.8% at 67% conversion was obtained at 350 ºC and 100 MPa with a residence time of 0.6 s.172 The main products under these conditions were retroaldol products, namely pyruvaldehyde and a mixture of dihydroxyacetone and glyceraldehyde, which combined yields up to 60%.172 Under these conditions it is possible that HMF is formed through aldol condensation reactions (Scheme 20), given the significant amounts of pyruvaldehyde and glyceraldehyde that are present. HMF yields of around 50% were reported by Li et al. from fructose decomposition experiments in high temperature liquid water at 10 MPa and 180-220 ºC.123 In the absence of catalyst a consistent maximum yield of around 50% was reported, requiring shorter reaction times with increasing temperature. The conversion was typically between 90 and 95%. At 220 ºC a yield of 50% was reached after approximately 10 min, at 200 ºC this was after about 30 min and at 180 ºC after about 80 min, but here the conversion was lower (around 80%).123 The addition of formic acid and acetic acid led to increased reaction rates and the maximum yields in HMF were around 58%. The increase in reaction rate was much more pronounced for formic acid than for acetic acid.123 Work on fructose dehydration in a high temperature (>200 °C) flow process was performed by Tarabanko et al.203 Phosphoric acid was used as the catalyst with an optimum HMF yield of 40%. This could be obtained either with an acid concentration of 0.01 M at 260 °C or with an acid concentration of 0.05 M at 240 °C. The optimum temperature increased with decreasing acid concentration, together with a trend of increasing yield at higher temperature and lower acid concentration. Asghari and Yoshida applied heterogeneous zirconium phosphate catalysts in sub-critical 204 water. Different H3PO4 treatment times and calcination temperatures were tested. Dehydration experiments were performed at 240 °C at 33.5 Bar for 120 s with a fructose concentration of 1 wt% and a catalyst-fructose ratio of 0.5 m/m. With the non-calcined catalysts the HMF yields were consistent around 50% (80% conversion, 59-62% selectvity), with no visible effect of the H3PO4 treatment. In the absence of catalyst only 19% yield (59% conversion) was obtained. Calcination decreased the yield by 5-13%. The negative effect of calcination increased with H3PO4 treatment. Hansen and co-workers reported work on microwave assisted synthesis of HMF from concentrated aqueous fructose.159 Heating a 30% fructose solution in the absence of catalyst yielded only 1% HMF at 5% conversion after 5 min at 160 °C. Increasing the temperature to 190 °C led to around 35% HMF yield at ~70% conversion. The authors point to the presence
52| Chapter 2 of formic acid and levulinic acid as the cause for the enhanced HMF yield under these circumstances. At 1 M HCl concentration a large amount of formic acid (39%) and levulinic acid (30%) were formed together with an HMF yield of 28%. At 0.01 M HCl concentration tests on the influence of temperature and reaction time with 27% fructose solution, showed high temperatures favoured the dehydration reaction to HMF. An HMF yield of 53% at 95% fructose conversion was reported at 200 °C after 60 s. Increasing the initial microwave power from 90 W to 150 W and 300 W had no effect. Tables 7-9 present an overview of the research on fructose dehydration in water, with a focus on the best yields and selectivities reported in each reference. Some general trends can be seen, such as a decrease in selectivity with increasing conversions and low yields in the absence of a catalyst. The decrease in selectivity with increasing conversion corresponds with the occurrence of rehydration to LA and polymerisation reactions of HMF under those reaction conditions. The general consensus is that an increase in fructose concentration favours polymerisation reactions, which has a negative effect on the HMF selectivity. There does not appear to be a clear trend in the results obtained with homogeneous and heterogeneous catalysts. Under similar conditions with similar catalysts a great variation in the reported HMF yields can be observed that cannot be easily explained. This also makes it essentially impossible to make a comparison between heterogeneous and homogeneous catalysts. The fact that HMF reacts further to levulinic acid and formic acid in the presence of acid and water, makes it surprising that some publications claim high selectivities (>90%) at significant conversions (>50%) using Brønsted acid catalysts, where others report selectivities of around 50% at best under similar conditions. An explanation for these divergent results in Tables 7-9 could lie in the challenging analytics, especially for the polymeric by-products. In addition, most of the analytical results have been obtained without the use of internal standards, which may lead to serious errors. Thus the reader is advised to exercise caution, in particular regarding the reports on high selectivities at high conversions. Many authors have drawn the conclusion that in order to increase the selectivity to HMF at higher conversions, its rehydration should be prevented by either stabilisation, for instance with a specific solvent, by removal of water or through its continuous removal from the reaction mixture.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|53
Table 7. Fructose dehydration to HMF in aqueous systems in the absence of catalyst
Fructose Temperature Reaction Yield Conversion Selectivity Reference concentration (°C) Time (%) (%) (%) (wt%) 5 140 1 h 4 - - 163
30 160a 5 min 1 5 28 159
4.5 175 1.5 h 56 72 78 157
30 190a 5 min 36 67 54 159
11 200 30 min 51 89 57 123
2 200a 5 min 13 28 46 129
9 200 5 min 21 40 53 126,127
0.9 350 0.59 sb 8 67 11 172 a: heating by microwave irradiation; b: continuous flow process
54| Chapter 2
Table 8. Fructose dehydration to HMF in aqueous systems, catalysed by homogeneous catalysts
Fructose Catalyst Catalyst Temperature Reaction Time Yield Conversion Selectivity Reference concentration loading (°C) (%) (%) (%) (wt%)
11 Acetic Acid 100 wt% 200 20 min 58 92 64 123
a c 5 AlCl3 50 mol% 120 5 min 50 - - 205
4.5 Formic Acid pH 2.7 175 45 min 56 56 100 157
11 Formic Acid 100 wt% 200 10 min 58 96 61 123
b 1 H2SO4 10 mM 180 600 s 28 80 35 206
a 2 H2SO4 50 wt% 200 5 min 47 97 48 129
9 H2SO4 1 mM 200 5 min 23 93 25 126
b 0.9 H2SO4 4 mol% 250 32 s 50 95 53 104
b 4.5 H3PO4 16 mol% 240 3 min 40 - - 203
b 4.5 H3PO4 4 mol% 260 3 min 40 - - 203
4.5 HCl 320 mol% 90 7 h 43c,f 72 60 194
9 HCl 400 mol% 95 16 min 26 46 57 155
9 HCl 400 mol% 95 24 min 30 62 48 155
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|55
9d HCl 200 mol% 95 1.5 h 68 86 79 156
2 HCl/ZnCl2 100 mol%/ 120 - 53 97 55 202 2381 mol%
27 HCl 1 M 130 5 min 28 99 28 159
30 HCl 0.25 M 180 2.5 - 3 min 25 50 51 170
27 HCl 0.01 M 200a 1 min 53 95 56 159
26 Oxalic acid 1 mol% 145 2,5 23c,e - - 196
25 Oxalic acid 2 mol% 135-142 130 min 34 61 55 195
9d PTSA 200 mol% 88 190 min 50 88 57 158
3.6 PTSA 500 mol% 88 4 h 25 53 47 158
5 YbCl3 0.67mol% 140 1 h 18 34 54 163
5 YbCl3 0.67mol% 140 2 h 24 59 40 163 a: heating by microwave irradiation; b: continuous flow process; c: isolated yield; d: PEG as cosolvent; e: analysis by NMR only; f: in situ HMF adsorbance by activated carbon
56| Chapter 2
Table 9. Fructose dehydration to HMF in aqueous systems, catalysed by heterogeneous catalysts
Fructose Catalyst Catalyst Temperature Reaction Yield Conversion Selectivity Reference concentration loading (°C) Time (%) (%) (%) (wt%)
6 10%-wt AlVOP 35 wt% 80 2 h 58 76 76 198
30 3.75%-wt FeVOP 1.8 wt% 80 1 h 60 71 84 198
6 Cubic ZrP2O7 55 wt% 100 2 h 43 53 81 184
2 Dowex 50wx8-100 100 wt% 150 15 min 73d 90 81 207
2 Dowex 50wx8-100 100 wt% 150 30 min 33 54 62 207
6 -TiP 55 wt% 100 2 h 39 57 69 184
6 -TiP 55 wt% 100 1 h 67c 71 95 184
6 H3PO4-treated 63 wt% 100 0.5 h 28 29 98 197 niobic acid
6 H3PO4-treated 63 wt% 100 2 h 22 61 35 197 niobic acid
5.4 Niobium Phospate Fixed bed (3-4 110 n/a b 25 77 33 200 g)
6 NiP2O7 71 wt% 100 0.5 h 29 29 100 197
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|57
6 NiP2O7 59 wt% 100 3 h 30 51 59 183
2- a 2 SO4 /ZrO2 20 wt% 200 5 min 26 59 44 208
9 TiO2 100 wt% 200 5 min 22 98 22 126,127
9 TiO2 100 wt% 200 5 min 22 98 22 126,127
a 2 ZrO2 50 wt% 200 5 min 31 65 48 129
9 ZrO2 100 wt% 200 5 min 15 90 17 126,127
9 ZrO2 100 wt% 200 5 min 15 90 17 126,127
1 ZrP 50 wt% 240 2 min 49 81 61 204
a 2 α-TiO2 50 wt% 200 5 min 38 84 45 129 a: heating by microwave irradiation; b: continuous flow process, reaction time not applicable; c: HMF extraction with MIBK after 0.5 h and 1 h; d: 70 wt% acetone as co-solvent
58| Chapter 2
Fructose dehydration in organic solvents Research by Kuster,156 Van Dam158 and Qi207 showed the positive effects of the addition of organic solvents on the rate of HMF formation and on the HMF yield through a decrease in the rate of its decomposition/rehydration/condensation. It is thus not surprising that many groups have started investigating non-aqueous solvent systems in order to prevent rehydration of HMF to levulinic acid. Sugars typically have very low solubility in organic solvents, with the exception of polar coordinating solvents, i. e. DMSO and DMF. The work done on these coordinating solvents will be discussed first, followed by work in other organic solvents. The first HMF synthesis in DMSO was reported by Nakamura and Morikawa in 1980 and yielded 90% HMF using a Diaion PK-216 ion exchange resin.209 Shortly after, Szmant and Chundury, reported the use of boron trifluoride etherate as a catalyst in DMSO and obtained yields of >90% HMF.210 The dehydration was tested at different temperatures, catalyst concentrations and fructose concentrations. The reported maximum HMF yields varied between 55% and 99% at reaction times of 0.5-3 h. With 25 mol% boron trifluoride etherate in a 1.4 M fructose solution 99% HMF yield was reported after 45 min at 100 °C. The quantification was performed with UV spectrometry, making it impossible to distinguish between HMF and other furfural or polymeric (humin) derivatives. An interesting trend was found for an experiment in DMSO with 25 mol% boron trifluoride etherate at 100 °C, showing a maximum furanic yield of about 90% after 0.5 h that decreased to 80% after 1 h and subsequently remained stable for at least 2.5 h. DMF, diethylene glycol monoethyl ether, 2-methoxyethanol and 2-ethoxyethanol were also tested as solvents, resulting in significantly lower HMF yields. In DMF the maximum HMF yield was only 55% with 25 mol% boron trifluoride etherate at 100 °C after 1.5 h. Mercadier et al. reported 80% HMF yield in DMF after 5 h at 96 ºC with a Lewatit SPC 108 ion- exchange resin.211 Brown et al. also looked at DMSO as the solvent, reporting quantitative conversion to HMF after 16 h at 100 ºC in the absence of catalyst.212 The yields were estimated by 1H-NMR analysis only. A number of acid and basic catalysts were tested, showing that the basic catalysts inhibited the HMF formation and the acid catalysts enhanced the rate of HMF formation. With 0.1 M
NH4Cl a complete conversion to HMF in 45 min was claimed. Separating HMF from DMSO through distillation was found to be difficult. For this reason other solvents were tested, such as sulfolane, DMF, ethyl acetate, butyl acetate and a collection of alcohols.212 DMF and sulfolane
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|59 showed similar results and isolation problems as DMSO. In ethyl acetate the reaction was much slower and less selective, resulting in an isolated HMF yield of 58% at full conversion after 30 h at 77 °C. Contrary to the reaction in DMSO, DMF and sulfolane significant by-product formation was observed in ethyl acetate, i.e. 5-chloromethylfurfural (CMF) and levulinic acid. Butyl acetate did not show this by-product formation, but the isolated HMF yield was only 31% at incomplete conversion after 16 h at 100 °C. Further work in DMSO by Musau and Munavu213 was based on the knowledge that DMSO facilitates the formation of furans from 1,4-diketones214 and tetrahydrofurans from 1,4-diols.215 A maximum yield of 92% HMF at 150 ºC was reported.213
A comparable yield of 95% was obtained by Seri et al. with 2.5 mol% LaCl3 as catalyst in DMSO after 4 h at 100 ºC.216 Use of DMA and DMF resulted in almost the same yield.216 Use of sulfolane (~50%), 1,4-dioxane and 1-butanol (both ~25%) led to significantly lower yields. Dehydration of fructose and sorbose, another ketose, was tested in DMSO, yielding 93% and 61% HMF respectively after 2 h at 120 °C.216 Wang et al. investigated scandium and lanthanide triflates as catalysts for fructose dehydration in organic solvents.217 In DMSO an HMF yield of 83% at full conversion was obtained with 2 wt% fructose in DMSO in the presence of 10 wt% Sc(OTf)3 after 2 h at 120 °C. Shimizu et al. published an elaborate study on fructose dehydration to HMF in DMSO with a variety of catalysts, including zeolites, ion exchange resins, heteropolyacids and basic heterogeneous catalysts.218 The experiments were carried out with 3 wt% fructose and 6 wt% catalyst in DMSO at 120 °C for 2 h under continuous water removal by mild evacuation. Without additional catalyst a 32% HMF yield at 81% conversion was reported. Very high yields (>90%) were reported for a number of catalysts, such as zeolites, heteropolyacids and acidic resins. Without continuous water removal the yields decreased, except when using powdered Amberlyst 15, which was claimed to yield 100% HMF in either case, even after recycling (three times) and at fructose concentrations as high as 50%. In case of FePW12O40 and H-BEA zeolite the HMF yields decreased to below 50% at 50 wt% fructose concentration. Basic catalysts Al2O3 and MgO were found to inhibit HMF formation, but did convert fructose to unknown products.
Yan et al. performed a study on fructose dehydration in DMSO with a number of ZrO2 based catalysts.130 A reaction with 7.6 wt% fructose at 130 °C for 4 h in the absence of catalyst yielded
60| Chapter 2
72% of HMF. Adding different basic catalysts had a negative impact on the yield, because of their activity in fructose to glucose isomerisation, as was already discussed in section 2.3.1.1.
Recent work using aspartic acid templated TiO2 in DMA-LiCl (10%) by De et al. resulted in 219 82% HMF yield. Dutta et al. applied TiO2 in a number of solvents, obtaining the best yield (54%) in DMSO.220a The same group used hierarchically macro/mesoporous titanium phosphate nanoparticles with aggregated particles of 15–20 nm in DMA-LiCl. Although the large pores were supposed to aid transport, the yield did not improve and HMF was obtained in only 42% yield.185b Continuing work on acetone/water mixtures,207 Qi et al. performed fructose dehydration in acetone-DMSO under otherwise identical conditions.221 With 2 wt% fructose in 70:30 (w/w) acetone-DMSO and DOWEX 50WX8-100 resin as catalyst, 88% HMF yield was obtained at 98% conversion after 20 min at 140 °C. Decrease of the acetone ratio led to a decrease in reaction rate, but the selectivity vs. conversion plots were essentially identical. The catalyst showed only 2% decrease in yield after five recycles. Work was also published with zirconia 2- 208 (ZrO2) and sulfated zirconia (SO4 /ZrO2) as catalyst. Experiments were performed with 2 wt% fructose in acetone-DMSO (70:30 w/w) in the presence of 20 wt% catalyst. When 2- comparing ZrO2 with SO4 /ZrO2, the latter showed higher activity, with HMF yields of 60-66% at 84-91% conversion depending on the calcination temperature. With the regular ZrO2 the HMF yields were below 50% at lower conversion and varied significantly with calcination temperature. The selectivity was consistently around 10% higher for the sulfated catalysts than for the regular catalyst. In the absence of catalyst the reaction also took place, but at a slower rate, with a reported HMF yield of 66% at 85% conversion after 20 min, which is only slightly lower than those obtained after 20 min in the presence of catalyst. Research on the use of ionic liquids as homogeneous catalyst was published by Tong and Li.222 Two types of cations, namely [HMIm]+ and [HNMP]+, were tested in combination with either - - HSO4 or CH3SO3 as anion. The NMP based ionic liquids were found to have stronger Brønsted + - acidity, with [HNMP] [HSO4] being the most acidic. Experiments were done in DMSO, water, ethanol, N, N-dimethylacetamide, carbon tetrachloride and acetonitrile. The HMF yield was reported to be by far the highest in DMSO, in which a yield of 72% at 83% conversion was + - obtained by reacting 7 wt% fructose at 90 ºC for 2 h with 7.5 mol% of [HNMP] [CH3SO3] . With + - [HNMP] [HSO4] an HMF yield of 69% at 99% conversion was obtained.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|61
Recently Qu et al. published 92% HMF yield in DMSO with [BMIm]OH as the catalyst after 8 h at 160 °C.223 As was already mentioned above, Brown reported quantitative HMF yields in DMSO under less severe conditions in the absence of a catalyst. Combined with the other results reported in Table 10 this indicates a negative effect of the ionic liquid additive, since far more severe conditions were applied. Ohara et al. reported 90% HMF yield at 100% fructose conversion in DMF with 100 wt% Amberlyst 15 as the catalyst after 1 h at 100 ºC.224 In 2008 Bao and co-workers published work on fructose dehydration, catalysed by dilute acidic ionic liquids 41 and 42 (Scheme 30) and their immobilised counterparts 43 and 44 in DMSO.225 Ionic liquid 41 is a Brønsted acid and 42 is its Lewis acid derivative. The dehydration was performed in DMSO at temperatures between 80 and 160 ºC under microwave irradiation using 3 wt% fructose in combination with 50 mol% of catalyst. The effects of catalyst loading, temperature, reaction time and recycling of the catalyst were tested. With 200 mol% catalyst loading at 100 °C for 4 min HMF yields of 85% and 88% at full conversion were obtained with 41 and 42 respectively. A decrease of catalyst loading coincided with a decrease in yield and selectivity. Yield and selectivity increased with increasing temperature, an effect that was stronger for the Brønsted acid, which was significantly less active than the Lewis acid in the lower temperature range. This is surprising, since the Lewis acid is expected to convert to the Brønsted acid in the presence of water, which is formed during the dehydration and which is in general present in DMSO. With 50 mol% of either catalyst at 160 ºC HMF yields of around 90% were obtained at full conversion after 4 min. Experiments with both these ionic liquids immobilised on silica (ILIS) were reported, resulting in yields of around 70% HMF at full conversion for both types, compared to around 60% at 90-95% conversion for silica gel supported acids SiO2-SO2Cl and SiO2-SO3H. The selectivities were not significantly different, all between 65% and 70%. Recycle experiments with the ILIS catalysts showed no loss in activity for both immobilised ionic liquids, even after seven cycles. The silica gel supported acids showed a sharp decrease in activity upon recycling, becoming almost inactive after 3 cycles. Zhang et al. investigated the use of polytungstic acid (PTA), encapsulated in MIL-101, as a catalyst for fructose dehydration to HMF.226 MIL-101 is a metal-organic framework with a chromium carboxylate cubic structure. An encapsulated PTA catalyst (PTA/MIL-101, 40 wt%)
62| Chapter 2 was tested with a 9 wt% fructose solution in DMSO and yielded 63% HMF at 82% conversion after 30 min at 130 °C.
Scheme 30. Ionic liquids used by Bao et al.
Experiments with immobilised ionic liquid on silica in DMSO were reported by Sidhpuria et al. 1-(tri-ethoxy-sylil propyl)-3-methyl imidazolium hydrogen sulfate (IL-HSO4) was synthesised and immobilised on silica nanoparticles to form Si-3-IL-HSO4 with a 20 wt% IL loading.227 Reaction conditions were optimised by experimental design, from which a mathematical model was derived. From an 8 wt% fructose solution in the presence of 80 wt% catalyst an HMF yield of 63% at full conversion was obtained after 30 min at 130 °C. In the absence of any catalyst no activity was observed, which is in contrast with findings by Yan,130 Musau213 and Brown.212 Binder and Raines published a collection of results on fructose, glucose and cellulose dehydration to HMF in dimethyl acetamide (DMA) based reaction mixtures.119,228 DMA was applied as solvent, combined with alkali metal salts with halogen counterions. The highest HMF yields were obtained using metal bromides and iodides at a concentration of 10 wt% in DMA. Yields of around 90% were obtained by reacting 10 wt% fructose at 100 ºC in the presence of 6 mol% sulfuric acid for 2-6 h in DMA/MX, with M = Li, Na, K and X = Br, I. The use of LiF as the salt did not yield any HMF. Systems with ionic liquids as additives were also tested, and the results suggested an improvement of the HMF yield at 80 °C, but no comparative example was reported to verify this. Caes and Raines published work on fructose dehydration in sulfolane, catalysed by various halide salts and acids.229 At 100 °C with 11 mol% HBr as catalyst 93% HMF yield was reported. In the presence of 207 mol% LiBr the HMF yield was 78%.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|63
Sanborn patented HMF production from Cornsweet® 90 high fructose corn syrup (HFCS), which contains 77 wt% sugars of which 90% is fructose, in N-methyl pyrrolidinone (NMP) and DMA.230 HFCS is produced as a sweetener by enzymatic isomerisation of glucose to fructose, followed by highly advanced methods of chromatographic separation (Simulated Moving Bed) in which the fructose content is increased and the glucose is recycled for isomerisation. A reaction of 33 wt% HFCS in NMP with 40 wt% Amberlyst 35 at 115 °C was reported to yield 81% HMF at 94% fructose conversion after 5 h. In DMA under similar circumstances, but at 105 °C, the HMF yield was 62% at 75% conversion. Chen et al. adopted a different strategy for avoiding water as main solvent by derivatising fructose in acetone, forming 1,2:4,5-di-O-isopropylidene-β-D-fructopyranose (45) and 2,3:4,5- di-O-isopropylidene-β-D-fructopyranose (46) in order to improve its solubility in organic solvents (Scheme 31).231 A solvent system of 7.5 M water in ethylene glycol dimethlyl ether with
0.5 M of substrate and 5 mM H2SO4 yielded around 70% HMF at 180 ºC. The presence of water in the starting reaction mixture was required in order to generate free fructose, which could subsequently be dehydrated. Partially replacing water with acetone resulted in an increase in reaction rate and HMF yield with increasing acetone content.207 Experiments by Qi et al. with acidic ion-exchange resin (Dowex 50wx8-100) under microwave irradiation yielded over 70% HMF upon using at least 70 wt% acetone.207 In alcohols HMF can be converted to the appropriate ether (RMF) in the presence of an acid catalyst. The first publication on 5-methoxymethyl-2-furfural (MMF) dates from 1927, by Haworth et al., in which it was observed in an attempt to deduce the structure of sucrose.232 It was later also synthesised by Wolfrom et al. from tetramethyl glucosene.108
Scheme 31. Di-O-isopropylidene-β-D-fructopyranose derivatives; 1,2:4,5-di-O-isopropylidene- β-D-fructopyranose (45) and 1,2:4,5-di-O-isopropylidene-β-D-fructopyranose (46)
64| Chapter 2
Dehydration of fructose in various alcohols by Brown et al. led to the formation of the appropriate HMF ether and levulinic acid ester.212 The reactions were performed at 14 wt% fructose in the presence of 100 wt% Amberlyst-15 (dry) for 20 h at 100 ºC. A number of primary and secondary alcohols was tested, from methanol to 2-butanol. In methanol a yield of 43% MMF was reported in combination with 47% methyl levulinate. In ethanol, 55% of EMF and 25% ethyl levulinate were reported and in n-propanol a 19% PMF yield was reported with 69% propyl levulinate. Bicker published work on sugar dehydration in flow.125,206,233 Experiments were performed in water, acetone/water, acetone, methanol and acetic acid as solvents at sub- and supercritical conditions. In water containing 1 wt% fructose and 10 mM H2SO4 28% HMF yield at 80% conversion was obtained at 180 ºC and a residence time of 10 min.233 In acetone/water (9:1 V/V) under comparable circumstances with 2 min residence time an HMF yield of 75% at 98% conversion was obtained.206,233 When methanol was used as the solvent, the main product was 5- methoxymethyl-2-furfural (MMF), which resulted from the acid catalysed etherification of HMF.233 An MMF yield of 78% at 99% conversion was obtained at 240 °C and a residence time of 2 s. At short residence times, especially at lower temperatures significant amounts of HMF were observed and smaller amounts of MMF. The amount of MMF increased with residence time at the cost of the amount of HMF, which indicates that MMF formation goes through HMF. Variation of the pressure between 15 and 35 MPa at 180 °C did not significantly affect the HMF and MMF selectivity. In the absence of sulfuric acid no furan formation was observed. In acetic acid yet another main product was formed, namely 5-acetoxymethyl-2-furfural (AMF), resulting from the esterification of HMF. An AMF yield of 38% was obtained at full fructose conversion at 180 ºC with 2 min residence time, with no HMF present.233 Tarabanko et al. worked on fructose dehydration in several aliphatic alcohols.234 The alkylated forms of HMF and levulinic acid were the major products. Yields of over 60% EMF were reported for initial fructose concentrations of 0.08 M, 0.11 M and 0.15 M in 1.8 M H2SO4 in ethanol at 82.5 °C with reaction times of 30-50 min. The ethyl levulinate yield was always below 10%. Yields of both products decreased with increasing fructose concentration. Comparable experiments were performed in butanol, but at 92 °C. 5-butoxymethyl furfural (BMF) formation was significantly slower than EMF formation, yielding around 60% BMF with 0.15 M initial fructose concentration after 2 h. The butyl levulinate yield was around 30%. No fructose
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|65 conversions were mentioned. Recently more work has been published on fructose dehydration in alcohols.235-238 Several patents have been published by the company Avantium on the production of alkoxymethyl ethers (RMF) and esters (AMF) of HMF.239-245 These patents concern the acid catalysed dehydration of hexose containing starting material in combination with various alcohols239-244 or organic acids and their anhydrides.245 The formation of RMF from hexoses or HMF by using olefins was also patented.244,246 Sanborn patented work on fructose dehydration in ethanol to form EMF230 and in organic acids to form HMF esters.247 The reported yields in the Sanborn patents were low and not clearly quantified. Tables 10-12 provide an overview of the best results obtained in fructose dehydration in organic solvents. The reported HMF yields and selectivities of fructose dehydration in organic solvents are much higher and more consistent than those in water. In DMSO, DMF and DMA yields of >90% have been reported. Even though water is formed during the reaction, apparently rehydration to LA is limited, which suggests that this group of solvents has a stabilising effect on the HMF molecule, most likely through coordination, making the molecule inaccessible to water for rehydration. Compared to dehydration under aqueous conditions the reaction conditions are typically much milder when using organic solvents. A number of publications even showed significant HMF yields in DMSO in the absence of catalyst at relatively low temperatures, which also places question marks in the proposed effectiveness of some catalyst systems researched in DMSO. Fructose dehydration in DMSO has been published by a significant number of researchers, but has shown varying results, making it very difficult to explain these differences, especially if one finds no HMF yield in DMSO and another finds quantitative conversion in DMSO in the absence of catalyst. Furthermore it is of questionable use to research all kinds of metals as catalysts if in the absence of catalyst this reaction is already selective and when certain heterogeneous acids provide good results (Table 12). A major disadvantage of the use of DMSO is its known instability at temperatures over 100 °C, something that could also be a factor in its catalytic activity. Furthermore it is difficult to separate the HMF from the DMSO because of its high affinity for the solvent, requiring either large amounts of extraction solvent or the evaporation of DMSO, both of which are energy demanding processes.
66| Chapter 2
In alcohols (Table 13) HMF was mostly obtained as its alkyl ether. The yield and reaction rate generally decrease with increasing chain length of the alcohol. The yields of furanic products are generally higher than in water and lower than in aprotic solvents like DMSO.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|67
Table 10. Fructose dehydration to HMF in organic solvents in the absence of catalyst
Fructose Solvent Temperature Reaction Yield Conversion (%) Selectivity Reference concentration (°C) Time (%) (%) (wt%)
5 DMSO 100 16 h 100c 100 100 212
3 DMSO 120 2 h 32d 81 40 218
8 DMSO 130 30 min 0 0 0 227
8 DMSO 130 4 h 72 100 72 130
5 DMSO 140 5 min 22 - - 205
34 DMSO 150c 2 h 72b - - 213
21 DMSO 150 - 92 - - 213
2 Acetone:DMSO 180a 20 min 66 85 78 208 (70:30 w/w) a: heating by microwave irradiation; b: isolated yield; c: analysis by 1H-NMR only; d: continuous water evacuation
68| Chapter 2
Table 11. Fructose dehydration to HMF in organic solvents, catalysed by homogeneous catalysts
Substrate Solvent Catalyst Catalyst Temperature Reaction Yield Conversion Selectivity Reference concentration loading (°C) Time (%) (%) (%) (wt%)
3.5 1,4-Dioxane LaCl3 2.5 mol% 100 5 h 27 - - 216
2 1,4-Dioxane Sc(OTf)3 10 wt% 120 2 h 16 86.2 19 217
h 32 2-Ethoxyethanol BF3.OEt2 50 mol% 100 2 h 64 - - 210
h 27 2-Methoxyethanol BF3.OEt2 50 mol% 100 1 h 78 - - 210
1 Acetone:Water (9:1 H2SO4 10 mM 180 2 min 75 98 77 206 V/V)
c 5 Butyl acetate NH4Cl 10 mol% 77 16 h 31 - - 212
h 27 Diethylene glycol BF3.OEt2 50 mol% 100 0.5 h 40 - - 210 monoethyl ether
4 DMA LaCl3 2.5 mol% 100 4 h 92 - - 216
2 DMA Sc(OTf)3 10 wt% 120 2 h 50 98.4 51 217
10 DMA (1.5 wt% KCl) H2SO4 6 mol% 80 2 h 56 - - 119
10 DMA (10 wt% NaBr) H2SO4 6 mol% 100 2 h 93 - - 119
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|69
10 DMA (10 wt% KI) H2SO4 6 mol% 100 5 h 92 - - 119
h 33 DMF BF3.OEt2 25 mol% 100 1.5 h 55 - - 210
4 DMF LaCl3 2.5 mol% 100 4 h 92 - - 216
c 5 DMF NH4Cl 10 mol% 100 1 h 55 100 - 212
3 DMSO 41 200 mol% 100a 4 min 85 100 85 225
3 DMSO 41 10 mol% 100a 4 min 50 79 63 225
3 DMSO 41 50 mol% 160a 4 min 93 100 93 225
3 DMSO 42 50 mol% 100a 5 min 86 - - 225
3 DMSO 42 50 mol% 100 3 h 86 - - 225
3 DMSO 42 200 mol% 100a 4 min 88 100 88 225
3 DMSO 42 10 mol% 100a 4 min 71 98 72 225
3 DMSO 42 50 mol% 160a 4 min 90 100 90 225
a c 5 DMSO AlCl3 50 mol% 140 5 min 69 - - 205
h 34 DMSO BF3.OEt2 25 mol% 40 3 h 55 - - 210
h 19 DMSO BF3.OEt2 25 mol% 100 45 min 99 - - 210
h 20 DMSO BF3.OEt2 100 mol% 100 45 min 99 - - 210
70| Chapter 2
2 DMSO Ho(OTf)3 10 wt% 120 2 h 78 100 78 217
3 DMSO LaCl3 2.5 mol% 100 4 h 95 - - 216
3 DMSO LaCl3 5 mol% 120 2 h 93 - - 216
d 3 DMSO LaCl3 5 mol% 120 2 h 61 - - 216
2 DMSO Nd(OTf)3 10 wt% 120 2 h 64 100 64 217
e 5 DMSO NH4Cl 10 mol% 100 45 min 100 100 100 212
e 5 DMSO NH4HSO4 10 mol% 100 2.5 h 100 100 100 212
+ - 7 DMSO [HNMP] [CH3SO3] 7.5 mol% 90 2 h 72 83 87 222
+ - 7 DMSO [HNMP] [HSO4] 7.5 mol% 90 2 h 69 99 70 222
2 DMSO Sc(OTf)3 10 wt% 90 2 h 66 95 70 217
2 DMSO Sc(OTf)3 10 wt% 120 2 h 83 100 83 217
2 DMSO Sm(OTf)3 10 wt% 120 2 h 73 100 73 217
2 DMSO Yb(OTf)3 10 wt% 120 2 h 80 100 80 217
c 5 Ethyl acetate NH4Cl 10 mol% 77 30 h 58 100 - 212
4 n-Butanol LaCl3 2.5 mol% 100 5 h 24 - - 216
9 NMP FeCl3/Et4NCl 10 mol%/18 90 2 h 82 100 82 248
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|71
mol%
2 PEG-400-water, Sc(OTf)3 10 wt% 120 2 h 13 37 35 217 unknown ratio
6.7 Sulfolane HBr 11 mol% 100 1 h 93 - - 229
7 Sulfolane LaCl3 2.5 mol% 100 5 h 52 - - 216
6.7 Sulfolane LiBr 207 mol% 100 2 h 78 - - 229
8.3 Sulfolane LiCl 425 mol% 90 2 h 67 - - 229
e 5 Sulfolane NH4Cl 10 mol% 100 30 min 100 100 100 212
f 15 MeOCH2CH2OMe H2SO4 1 mol% 180 5 min 72 >98 73 231
(7.5 M H2O)
g 15 MeOCH2CH2Ome H2SO4 1 mol% 180 5 min 68 >98 69 231
(7.5 M H2O) a: heating by microwave irradiation; b: continuous flow process; c: isolated yield; d: Sorbose as substrate; e: analysis by 1H-NMR only; f: 1,2:4,5-di-O-isopropylidene-β-D-fructopyranose as substrate; g: 2,3:4,5-di-O-isopropylidene-β-D-fructopyranose as substrate; h: distinction between different furans could not be made due to the analytical method; i: See Scheme 30 for a description of this catalyst
72| Chapter 2
Table 12. Fructose dehydration to HMF in organic solvents, catalysed by heterogeneous catalysts
Fructose Solvent Catalyst Catalyst Temperature Reaction Yield Conversion Selectivity Reference concentration loading (rel. to (°C) Time (%) (%) (%) (wt%) substrate)
2 Acetone:DMSO Dowex 50WX8-100 100 wt% 150a 20 min 88 98 90 221 (70:30 w/w)
2- d a 2 Acetone:DMSO SO4 /ZrO2 20 wt% 180 5 min 63 84 74 208 (70:30 w/w)
2- c a 2 Acetone:DMSO SO4 /ZrO2 20 wt% 180 5 min 62 88 70 208 (70:30 w/w)
d a 2 Acetone:DMSO ZrO2 20 wt% 180 5 min 24 37 66 208 (70:30 w/w)
c a 2 Acetone:DMSO ZrO2 20 wt% 180 5 min 41 71 57 208 (70:30 w/w)
35 DMAe Amberlyst 35 40 wt% 105 5 h 62 75 83 230
a 5 DMA (10 wt% TiO2 50 wt% 130 2 min 74 - - 219 LiCl)
a 4 DMA (10 wt% TiO2 50 wt% 130 2 min 82 - - 219 LiCl)/[BMIm]Cl (5:1 w/w)
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|73
3 DMF Amberlyst 15 100 wt% 100 1 h 90 100 90 224
26 DMF Lewatit SPC 108 0.61 meq/mol 96 5 h 80 - - 211
3 DMSO Amberlyst 15 6 wt% 120 2 h 92b 100 92 218
3 DMSO Amberlyst 15 6 wt% 120 2 h 76 100 76 218
3 DMSO Amberlyst 15 powder 6 wt% 120 2 h 100b 100 100 218
3 DMSO Amberlyst 15 powder 6 wt% 120 2 h 100 100 100 218
50 DMSO Amberlyst 15 powder 6 wt% 120 2 h 100b 100 100 218
b 3 DMSO Cs2.5H0.5PW12O40 6 wt% 120 2 h 91 100 91 218
8 DMSO Diaion PK-216 10 meq 80 500 min 90 - - 209
b 3 DMSO FePW12O40 6 wt% 120 2 h 97 100 97 218
3 DMSO FePW12O40 6 wt% 120 2 h 49 100 49 218
b 50 DMSO FePW12O40 6 wt% 120 2 h 48 - - 218
b 3 DMSO H3PW12O40 6 wt% 120 2 h 95 100 95 218
3 DMSO H-BEA Zeolite 6 wt% 120 2 h 97b 100 97 218
3 DMSO H-BEA Zeolite 6 wt% 120 2 h 51 100 51 218
50 DMSO H-BEA Zeolite 6 wt% 120 2 h 40b - - 218
74| Chapter 2
a 3 DMSO ILIS-SO2Cl 50 mol% 100 4 min 67 100 67 225
a 3 DMSO ILIS-SO3H 50 mol% 100 4 min 70 100 70 225
3 DMSO Nafion 6 wt% 120 2 h 94b 100 94 218
3 DMSO Nafion 6 wt% 120 2 h 75 100 94 218
9 DMSO PTA/MIL-101 40 wt% 130 30 min 63 82 77 226
8 DMSO Si-3-IL-HSO4 80 wt% 130 30 min 63 100 63 227
a 3 DMSO SiO2-SO2Cl 50 mol% 100 4 min 60 92 65 225
a 3 DMSO SiO2-SO3H 50 mol% 100 4 min 63 95 66 225
2- b 3 DMSO SO4 /ZrO2 6 wt% 120 2 h 92 100 92 218
2- 7.6 DMSO SO4 /ZrO2-Al2O3 20 wt% 130 4 h 57 99 57 130
a 4 DMSO TiO2 50 wt% 140 10 min 54 - - 220
b 3 DMSO WO3/ZrO2 6 wt% 120 2 h 94 100 94 218
8 DMSO Zeolite H-beta (Si/Al = 80 wt% 130 30 min 63 100 63 227 25)
33 NMPe Amberlyst 35 40 wt% 115 5 h 81 94 86 230 a: heating by microwave irradiation; b: continuous water evacuation; c: calcination at 500 °C; d: calcination at 700 °C; e: Cornsweet 90 high fructose syrup as the substrate
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|75
Table 13. Fructose dehydration to HMF derivatives in alcohols and organic acid
Fructose Solvent Catalyst Catalyst T Reaction Conversion HMF HMF RMF RMF R group Reference concentration loading (°C) Time (%) yield selectivity Yield selectivity (wt%) (%) (%) (%) (%)
14 2-BuOH Amberlyst 100 wt% 100 20 h - - - 19b - sec-Butyl 212 15 dry
1 AcOH H2SO4 10 mM 180 15 s 93 7 8 24 26 Acetyl 233
1 AcOH H2SO4 10 mM 180 120 s 98 0 0 37 38 Acetyl 233
14 EtOH Amberlyst 100 wt% 100 20 h - - - 55b - Ethyl 212 15 dry
b 1.8 EtOH H2SO4 1.8 M 82.5 30 min - - - 62 - Ethyl 234
b 2.5 EtOH H2SO4 1.8 M 82.5 35 min - - - 66 - Ethyl 234
b 3.4 EtOH H2SO4 1.8 M 82.5 40 min - - - 60 - Ethyl 234
14 i-PrOH Amberlyst 100 wt% 100 20 h - - - 45b - i-Propyl 212 15 dry
14 MeOH Amberlyst 100 wt% 100 20 h - - - 43b - Methyl 212 15 dry
a 0.6 MeOH H2SO4 10 mM 180 30 s - - - - 82 Methyl 233
76| Chapter 2
a 7 MeOH H2SO4 10 mM 180 30 s - - - - 61 Methyl 233
a 1 MeOH H2SO4 10 mM 240 2 s 99 - - 77 78 Methyl 233
b 3 n-BuOH H2SO4 1.8 M 92 125 min - - - 61 - n-Butyl 234
14 n-BuOH Amberlyst 100 wt% 100 20 h - - - 0b - n-Butyl 212 15 dry
4 n-BuOH LaCl3 2.5 100 5 h - 24 - - - n-Butyl 216 mol%
14 PrOH Amberlyst 100 wt% 100 20 h - - - 19b - Propyl 212 15 dry
a: continuous flow process; b: Significant yield of levulinic acid ester observed
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|77
2.4.1.2 Glucose dehydration in single-phase systems
In this section the glucose dehydration in traditional single-phase systems will be reviewed, along with more recent progress in this area using sub- and supercritical conditions. Publications on HMF synthesis in aqueous systems will be discussed first, followed by a review of the research on reactions in organic solvents and aqueous/organic single-phase mixtures.
Glucose dehydration under aqueous conditions This paragraph describes research performed on glucose dehydration to HMF in water. First the thermal decomposition studies on glucose in the absence of catalyst are described, followed by work using homogeneous and heterogeneous catalyst systems, respectively. Jing and Lü reported significant HMF yields from glucose decomposition experiments in water at 180-220 °C and 100 bar in the absence of catalyst.249 An HMF yield over 30% at around 70% conversion was reported when a 1% glucose solution was heated for 30 min at 220 °C. An interesting observation was that the HMF degradation rates between 180 and 220 °C were substantially lower than the degradation rates of glucose. Work by Aida et al. on glucose decomposition in water at 350 and 400 ºC at pressures of 40, 70 and 80 MPa and at short residence times (<2 s) showed HMF yields under 10%.128 The reported furfural yields increased with residence time and pressure, rather than temperature, up to 12%. Mednick performed aldose dehydration with combinations of weak acids and bases, based on the knowledge that ketoses are much more readily dehydrated to HMF than aldoses and the hypothesis that the keto-enol isomerisation is acid-base-catalysed.105,250 A 20 wt% glucose solution was reacted at 160-190 °C for around 20 min with a warm-up time around 35 min. The main focus was on ammonium salts, phosphoric acid and phosphates. The HMF yields varied between 3 and 23%. The highest yield was obtained with a combination of phosphoric acid and ammonia at an initial pH of approximately 4. Pyridine-phosphoric acid systems with pH ~4.5 were studied for a solution of glucose (20 wt%) in H2O/p-dioxane (1:1 v/v). At 200-230 °C for 30 min with around 35 min warm-up time, HMF yields of ~45% were obtained. In 1964 ammonium sulfate catalysed aldose dehydration was patented by Smith et al.251 Using a 10% glucose mixture at pH 1.65 at 150-155 °C an HMF yield of 31.5% after 50 min was claimed.
78| Chapter 2
Glucose dehydration with 10 wt% sulfuric acid by Li et al. under microwave irradiation at 400W for 1 min yielded 49% HMF.252 In the absence of catalyst the yield was less than 1%. Tyrlik and co-workers have done work on glucose dehydration in aqueous systems under reflux with a variety of metal salts.253-255 The highest HMF yield of 9% was achieved with 15 254 wt% glucose in a 5 M MgCl2 system. Work on glucose dehydration by Seri et al., using lanthanide(III) chlorides in water at 140 ºC for 1h, yielded between 3 and 8% HMF.163 In the absence of catalyst only traces of HMF were observed. The same experiments with galactose and mannose showed comparable yields.163 Y-zeolite catalysed glucose decomposition studies in water by Lourvanij showed the formation of HMF at relatively low yields (<10%).256 In this work a 33% fructose yield was found after 2 min at 160 °C. Watanabe et al. reported work on glucose reactions in hot compressed water at 200 °C with a substrate concentration of 9 wt%.126,127 An HMF yield of 20% at 81% conversion was 126,127 obtained after 5 min in the presence of anatase TiO2. Rutile TiO2 did not contribute to the dehydration to HMF, showing identical results to those obtained in the absence of catalyst, which was 6.6% HMF yield at 20% conversion after 10 min. ZrO2 was tested as a catalyst under the same conditions yielding 13% fructose and only 5.2% HMF at 47% conversion. Later work on TiO2 catalysed dehydration using a 2 wt% glucose solution and microwave heating resulted in an HMF yield of 19% at 64% conversion after 5 min.129 Chareonlimkun and co-workers looked at reactions of glucose with differently prepared 257 TiO2 and ZrO2 catalysts in hot compressed water at 250 °C for 5 min. The work was focussed on the preparation of the catalysts, looking at different calcination temperatures (500-700 °C) and different precursors, namely nitrates and chlorides. The highest reported
HMF yield was 27%, using TiO2 from its chloride precursor, calcined at 500 °C. Increasing the calcination temperature resulted in lower HMF yields, decreasing to 25% at 600 °C calcination and 20% at 700 °C calcination. At 500 °C TiO2 was found to be mainly in the anatase phase, whereas at higher temperatures some rutile phase formation was detected. This is consistent with what Watanabe observed with regard to anatase and rutile phase TiO2 126,127 catalyst, in which rutile TiO2 was much less active. Yields obtained with titania from its nitrate precursor were typically around 5% lower at each calcination temperature than those from titania prepared from its chloride precursor. This is explained by temperature programmed desorption (TPD) results, which showed that TiO2 from TiCl4 had a higher acid site density than TiO2 from TiO(NO3)2, but the acidity of the sites was lower. The highest
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|79
reported HMF yield with ZrO2 as catalyst, calcined at 500 °C, was around 17%. ZrO2 showed the same trends with regard to precursor and calcination temperature as TiO2. The TPD results showed a lower acid site density for ZrO2 than for TiO2. In the absence of catalyst an HMF yield of 10% was reported. Apart from HMF, other products were also formed: furfural, levoglucosan and fructose. The fructose yield was consistently about 3% when either
TiO2 or no catalyst was used. When ZrO2 was used the fructose yield was consistently over 257 5%. Furfural and AHG (anhydroglucose) yields were always around 2%. A TiO2-ZrO2 mixed oxide catalyst was used in glucose dehydration at 250 ºC and 34.5 MPa for 5 min, showing almost 30% HMF yield, based on carbon balance, at around 80% conversion.258 Asghari and Yoshida looked at glucose dehydration catalysed by a heterogeneous ZrP catalyst.204 With 1 wt% glucose and a 1:1 (w/w) catalyst/substrate ratio at 240 °C and 33.5 bar they obtained 23.5% HMF yield (72% conversion) at a residence time of 240 s. Recently a combination of immobilised D-glucose/xylose isomerase and oxalic acid was applied as catalyst system for glucose dehydration in seawater.259 Glucose was first isomerised at 60 °C to fructose until equilibrium, followed by removal of the heterogeneous enzyme and addition of oxalic acid for dehydration of the fructose in the reaction mixture at 140 °C. The HMF was extracted with 2-methyl tetrahydrofuran and the oxalic acid crystallised out, leaving an aqueous glucose solution for potential recycle to the enzyme. An HMF yield of 64% was mentioned, based on the amount of fructose from the isomerisation. No explanation is given for the extremely high fructose yield of 64% reported for glucose isomerisation, while the authors mention a 50:50 equilibrium between glucose and fructose. Tables 14-16 show an overview of results published on glucose dehydration to HMF under aqueous conditions. HMF yields in processes catalysed by heterogeneous metal oxides are generally somewhat higher than those obtained in homogeneous systems. A bifunctional type of catalyst with basic sites for isomerisation and acid sites for dehydration showed the most 258 126,127 promise. Results from Chareonlimkun and Watanabe indicated that ZrO2 functioned mainly as an isomerisation catalyst to form fructose, whereas anatase TiO2 functioned as an acidic catalyst for dehydration to HMF. A ZrO2-TiO2 mixed oxide catalyst appeared to combine these functions to achieve a relatively high HMF yield of 30% under aqueous conditions.
80| Chapter 2
Table 14. Glucose dehydration to HMF in aqueous systems in the absence of catalyst
Glucose Temperature Reaction Yield Conversion Selectivity Reference concentration (°C) Time (%) (%) (%) (wt%)
9 200 10 min 7 31 23 126
2 200a 3 min 3 15 23 129
9 200 5 min 3 21 15 126,127
1 220 30 min 32 71 45 249
9 250 5min 10c 21 51 257
0.9 350d 0.7 sb 2c,f 72 3 128
0.9 350e 0.8 sb 4c,g 83 5 128
0.9 350e 1.6 sb 7c,g 84 8 128 a: heating by microwave irradiation; b: continuous flow process; c: small amounts of fructose were observed; d: at 400 Bar; e: at 800 Bar; f: 14% furfural yield; g: 7-8.5% furfural yield
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|81
Table 15. Glucose dehydration to HMF in aqueous systems, catalysed by homogeneous catalysts
Glucose Catalyst Catalyst Temperature Reaction Yield Conversion Selectivity Reference concentration loading (°C) Time (%) (%) (%) (wt%)
b c 20 (NH4)2HPO4 0.07 M 172-182 20 min 11 - - 250
b c 20 (NH4)2HPO4/H3PO4 0.07 M/ 174-180 20 min 23 - - 250 0.05 M
a 5 AlCl3 50 mol% 120 20 min 40 - - 205
8 DyCl3 0.7 mol% 140 1 h 7 17 41 163
8 DyCl3 0.7 mol% 140 2 h 12 30 40 163
a 2 H2SO4 50 wt% 200 3 min 2 11 23 129
e 9 H2SO4 1 mM 200 5 min 2 32 8 126
10 H2SO4:(NH4)2SO4 pH 1.65 150-155 50 min 32 - - 251
c 20 H3PO4 0.13 M 173-187 20 min 5 - - 250
14 Immobilised 3 wt%/0.1 60/140 2 h/1 h 64g - - 259 isomerase/oxalic acidf M
82| Chapter 2
d 13 MgCl2 593 mol% Reflux 3 h 9 - - 254
1 H3PO4/Nb2O5 1000 wt% 120 3 h 52 92 57 260
b c 20 Pyridine:H3PO4 0.3 M/0.2 200-225 8.5 min 45 - - 250 M a: heating by microwave irradiation; b: warm up time of 0.5 h or more; c: isolated yield; d: 26% yield of humins reported; e: 3% fructose observed; f: two-steps; g: yield based on fructose yield from isomerisation
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|83
Table 16. Glucose dehydration to HMF in aqueous systems, catalysed by heterogeneous catalysts
Glucose Catalyst Catalyst Temperature Reaction Yield Conversion Selectivity Reference concentration loading (°C) Time (%) (%) (%) (wt%) (wt%) (min)
12 HY-zeolite 50 160 3 8 83 10 256
c 9 TiO2 100 250 5 27 39 71 257
c 9 TiO2-ZrO2 100 250 5 29 44 67 258
a 2 ZrO2 50 200 3 10 57 18 129
b 9 ZrO2 100 200 5 5 48 11 126,127
c 9 ZrO2 100 250 5 17 38 46 257
a 2 Α-TiO2 50 200 5 19 64 29 129
c 9 Α-TiO2 100 200 5 20 81 25 126,127
a: heating by microwave irradiation; b: 13% fructose yield; c: 2-2.5% fructose yield
84| Chapter 2
Glucose dehydration in organic solvents The developments in fructose dehydration and the work by Mednick250 on glucose dehydration suggest that lowering the water content is favourable for the HMF yield. For this reason a number of groups has tested organic solvents as reaction media for glucose dehydration. Yan et al. studied HMF formation from fructose (section 2.4.1.1) and glucose in DMSO.130 A typical experiment was performed with 7.6 wt% glucose at 130 ºC. In the absence of catalyst an HMF yield of 4.3% at 94% conversion was found after 4 h. 2- The highest HMF yield of 48% was obtained when SO4 /ZrO2-Al2O3 catalysts with an Al-Zr molar ratio of 1:1 was used at 20 wt% relative to glucose. This was not influenced by an increase in reaction time to 6 and 15 h, showing stability of HMF under the reaction conditions.
Seri et al. performed LaCl3 catalysed glucose dehydration in DMSO at 120 ºC and found an HMF yield of 9.8%.216 Other aldoses, namely galactose and mannose, showed even lower yields of 6.7% and 4.8% respectively.216 Beckerle and Okuda recently performed comparable research with rare earth metals in DMA as the solvent, reporting much higher HMF yields of up to 30% from glucose.261
Hu et al. performed glucose dehydration in DMSO with 10 mol% SnCl4 and reported a 44% HMF yield at 96% conversion with a 9 wt% glucose solution at 100 °C after 3 h.144 The group of Ebitani examined the possibility of a combined isomerisation and dehydration catalyst system for glucose conversion in DMF.224 The isomerisation of 3 wt% glucose catalysed by 1 wt eq. hydrotalcite at 80 °C yielded 40% fructose at 47% conversion after 3 h. When a combination of hydrotalcite and Amberlyst 15 was used HMF formation was observed. The highest HMF yield of 42% at 73% glucose conversion, with no observed fructose, was obtained with 3 wt% glucose in combination with 2 wt eq. hydrotalcite and 1 wt eq. Amberlyst 15 at 80 °C after 9 h. Essentially the same result was obtained at 100 °C after 3 h, but in this case 4% of fructose yield was also observed. When glucose was first reacted for 2.5 h with hydrotalcite at 100 °C before addition of Amberlyst 15 an HMF yield of 45% was obtained at 61% conversion. No HMF formation was reported when either catalyst was tested separately under otherwise comparable conditions. DMA, DMSO, acetonitrile and water were also tested with the same Hydrotalcite/Amberlyst system
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|85 at 100 °C. With DMSO at 80 °C an HMF yield of 25% at 41% conversion was obtained. The results of the other solvents are described in Table 17.224 DMF was found to give the highest HMF yield. In water no HMF was formed, but 23% fructose yield was observed. The authors attributed this to the loss of activity of Amberlyst-15 in water. Addition of a small amount of water (3%) to DMF was claimed to be beneficial to HMF selectivity, anhydroglucose formation, with results comparable to DMSO at 80 °C. Comparable work was recently reported by the same group.262 The Amberlyst-15 catalysed dehydration of glucose in DMF resulted only in 1,6- anhydroglucose formation, of which yields of up to 70% were obtained. Higher temperatures favoured the formation of 1,6-anhydroglucose relative to the formation of HMF.224,263 As was already discussed in section 2.3.1.2, 1,6-anhydroglucose does not dehydrate to HMF.121,128,170 Binder and Raines published significant HMF yields from glucose dehydration in 119,228 DMA/LiBr/CrXn systems (X = Br or Cl). With 10% LiBr and 6% CrXn in DMA at 100 ºC an HMF yield of 80% with in 4-6 h. Essentially the same yields were obtained when using CrCl2, CrCl3 or CrBr3. These HMF yields are by far the highest reported from glucose in systems without ionic liquids. Based on these results, Chen and Lin applied a mixture of LiCl in caprolactam as solvent in combination with a number of catalysts that are known to facilitate the isomerisation/dehydration of glucose to HMF.264 Yields of around 60% at >90% conversion were obtained with CrCl2, SnCl4 and SnCl2 by reacting 10 wt% glucose in caprolactam:LiCl (3:1 n/n) with 6 mol% catalyst at 100 °C for 3 h. In addition to their work in 100% aqueous systems,254 Tyrlik et al. performed research on glucose dehydration in mixed water-alcohol systems.255 The highest combined yield of HMF and HMF alkyl ether of 22% was obtained in a system in which saturated Al2(SO4)3 in water was combined with ethanol (>30 M). Some work was also done in acetonitrile by Yasuda et al. with a composite of silica and MgCl2 as the catalyst, reporting an HMF yield of 70% from glucose, but much lower yields from mannose (10%) and galactose (32%).265 In a mixture of water and ethanol Yang et al. obtained a combined yield of HMF 266 and EMF of 57% with AlCl3 as the catalyst. Lew et al. used a heterogeneous Lewis acid, Sn-beta zeolite, in combination with a heterogeneous acid, Amberlyst 131, in ethanol to obtain an EMF yield of 31%.267
86| Chapter 2
Tables 17 and 18 give an overview of the highest reported HMF yields from glucose in organic solvents. When comparing these results with those in water, it is clear that the yields are generally significantly higher in organic solvents. Especially when dehydration catalysts are combined with isomerisation catalysts in aprotic polar solvents good yields of almost 50% are reported. Use of chromium halides as isomerisation catalyst in HCl/DMA even resulted in yields of around 80% HMF.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|87
Table 17. Glucose dehydration to HMF in organic solvents, catalysed by homogeneous catalysts
Glucose Solvent Catalyst Catalyst Temperature Reaction Yield Conversion Selectivity (%) Reference concentration loading (°C) Time (h) (%) (%) (wt%) (mol%)
10 Caprolactam:LiCl 3:1 (n/n) CrCl2 6 100 3 59 94 62 264
10 Caprolactam:LiCl 3:1 (n/n) SnCl2 6 100 3 55 94 59 264
10 Caprolactam:LiCl 3:1 (n/n) SnCl4 6 100 3 65 98 66 264
10 DMA LnCl3 10 145 2 34 100 34 261
10 DMA (10 wt% LiBr) CrBr3 6 100 6 80 - - 119
10 DMA (10 wt% LiBr) CrCl2 6 100 4 76 - - 119
10 DMA (10 wt% LiBr) CrCl3 6 100 6 79 - - 119
5 DMF GeCl4 10 100 1.25 34 85 40 268
9 DMF H3BO3 80 120 3 7 64 11 145
a 5 DMSO AlCl3 50 140 0.08 52 - - 205
9 DMSO CrCl3 7 100 3 28 79 35 269
88| Chapter 2
5 DMSO GeCl4 10 100 1.25 37 85 43 268
9 DMSO H3BO3 80 120 3 13 35 37 145
3.6 DMSO LaCl3 5 120 2 9.8 - - 216
a 3.6 DMSO LaCl3 5 120 2 6.7 - - 216
b 3.6 DMSO LaCl3 5 120 2 4.8 - - 216
9 DMSO SnCl4 10 100 3 44 96 45 144
c 6 Water:Ethanol Al2(SO4)3 100 Reflux 144 22 - - 255 a: Galactose as the substrate; b: Mannose as the substrate; c: combined yield of HMF and EMF
Table 18. Glucose dehydration to HMF in organic solvents, catalysed by heterogeneous catalysts
Glucose Solvent Catalyst Catalyst Temperature Reaction Yield Conversion Selectivity Reference concentration loading (°C) Time (h) (%) (%) (%) (wt%) (wt%)
3 DMA Hydrotalcite/Amberlyst 15 2:1 w/w 300 100 3 14 97 14 224
3 DMF Hydrotalcite/Amberlyst 15 2:1 w/w 300 80 9 42 73 58 224
3 DMF Hydrotalcite/Amberlyst 15 2:1 w/w 300 100 3 41a 72 57 224
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|89
3 DMF Hydrotalcite/Amberlyst 15 2:1 w/w 300 100 4.5d 45 61 73 224
3 DMF + 3 Hydrotalcite/Amberlyst 15 2:1 w/w 300 100 3 29 45 64 224 vol% water
3 DMSO Hydrotalcite/Amberlyst 15 2:1 w/w 300 80 3 25 41 61 224
3 DMSO Hydrotalcite/Amberlyst 15 2:1 w/w 300 100 3 12b 94 13 224
2- 7.6 DMSO SO4 /ZrO2-Al2O3 (Zr:Al = 1:1 n/n) 20 130 4 48 97 49 130
2- 7.6 DMSO SO4 /ZrO2-Al2O3 (Zr:Al = 1:1 n/n) 20 130 6 48 100 48 130
2- 7.6 DMSO SO4 /ZrO2-Al2O3 (Zr:Al = 1:1 n/n) 20 130 15 48 100 48 130
2- 3.5 DMSO/water SO4 /ZrO2-Al2O3 (Zr:Al = 3:7 n/n) 19 150 4 56 - - 270 (4:1 v/v)
3 MeCN Hydrotalcite/Amberlyst 15 2:1 w/w 300 100 3 10c 88 12 224
3 MeCN + 3 Hydrotalcite/Amberlyst 15 2:1 w/w 300 100 3 28 91 31 224 vol% water a: 4% fructose yield and 10% AHG yield; b: 6% fructose yield; c: 12% AHG yield; d: Amberlyst 15 was added after 2.5 h reaction time; e: galactose; f: mannose
90| Chapter 2
2.4.1.3 The dehydration of di- and trisaccharides, polysaccharides and biomass feedstock in single-phase systems
Disaccharides and trisaccharides Mednick et al. performed experiments with 19 wt% sucrose in a 1:1 (w/w) mixture of water and dioxane containing 5 mol% of pyridine and 3 mol% of H3PO4, relative to the substrate, HMF yields of 44% were obtained at unknown conversion after 32-36 min at 200-230 °C.250 Carlini et al. reported an HMF selectivity of >90% relative to the fructose moiety from niobium phosphate catalysed sucrose dehydration in water at 100 °C for 4 h.183 No conversion of the glucose moiety was observed. When regarding both the glucose and fructose moieties as substrate the HMF yield was 14% at 30% conversion.
Studies on LaCl3 catalysed dehydration of fructose containing di- and trisaccharides at a substrate concentration of 0.20 M in DMSO at 120 °C for 2 h showed that only the fructose moiety was converted to HMF.216 It was also observed that the type of glycosidic bonding between the fructose and other moieties influenced the HMF yields. In sucrose, glucose and fructose are connected through a Glcpα1↔2ßFruf bond, whereas in the isomeric turanose they are connected through a Glcpα1→3Fruf bond. Sucrose to HMF yields of 47% and 12% were reported for sucrose and turanose, respectively. The sugar conversions, selectivities and reaction rates were not mentioned. Ohara et al. applied the hydrotalcite-Amberlyst 15 system already described for the glucose dehydration224 in the sucrose and cellobiose dehydration to HMF.263 A reaction of 0.1 g substrate in 3 ml DMF with 0.1 g of either catalyst at 120 °C for 3 h yielded 54% HMF at 58% conversion from sucrose and 35% HMF at 52% conversion from cellobiose. Tarabanko et al. performed dehydration on 0.2 M sucrose in water with 0.6 M acetic acid in a continuous process at 250-260 °C, reporting a 40% HMF yield with respect to stoichiometry, though it is not entirely clear if this is based on the both the glucose and fructose or solely on the fructose moiety present.203
Polyfructans The Süddeutsche Zucker-Aktiengesellschaft looked at polyfructan, a polymer of fructose that is also called inulin, as starting material for a potential commercial HMF production process.195 From 20 kg chicory roots, which contains about 18% inulin on dry-matter base, in 21 kg aqueous sulfuric acid of pH 1.8 at 140 °C for 2 h an HMF yield of 13% was reported. Next to HMF, fructose (30%) and glucose (3.5%) were obtained.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|91
Carlini et al. performed research on HMF production from inulin with heterogeneous catalysts. From a 6 wt% inulin solution in water at 100 °C an HMF yield of 31% at 47% conversion was obtained with a niobium phosphate catalyst after 3 h.183 Under comparable conditions with γ-Titanium phosphate (γ-TiP) an HMF yield of 42% at 44% conversion was obtained after 1 h.184 Intermittent extraction of HMF by MIBK after 0.5 h improved the yield to 67% HMF at 71% conversion after 1 h. The same experiments were performed with a cubic ZrP2O7 catalyst yielding 35% HMF at 39% conversion after 1 h without intermittent extraction and 70% HMF at 73% conversion after 1 h with intermittent extraction. Experiments with various types of vanadyl phosphate catalysts gave somewhat lower selectivities at comparable conversions.198 271 Wu et al. published work on inulin dehydration in water with pressurised CO2. At 180
°C the effect of CO2 pressure was investigated by reacting 0.1 g inulin in 2 ml water for 3 h.
Experiments were performed with 0, 4, 6, 9 and 11 MPa CO2. The authors observed a beneficial effect of the CO2 on the HMF yield, ascribed to the formation of carbonic acid in combination with water. Without CO2, a maximum HMF yield of around 45% was obtained.
When 6 MPa CO2 was applied an increase in reaction rate and yield, up to a maximum of around 50%, was observed. Data obtained at 160 °C and 200 °C did not show a consistent beneficial effect in yields and reaction rates upon addition of CO2. A significant increase in reaction rate was observed at higher temperatures, with maximum HMF yields of around 50% at 200 °C after 45 min.
Starch, Cellulose and lignocellulosic biomass Mednick et al. performed experiments with 19 wt% corn starch in a 1:1 (w/w) mixture of water and dioxane containing 5 mol% of pyridine and 3 mol% of H3PO4, relative to the substrate. HMF yields of 44% were obtained at unknown conversion after 32-36 min at 200- 230 °C.250 This result is essentially identical to that obtained with sucrose, which is unexpected, since the majority of HMF formed from sucrose is generally believed to originate from the fructose moiety of sucrose. HMF was reported as a degradation product in aqueous dilute acid catalysed cellulose hydrolysis by Mok et al.272 At temperatures over 200 °C a yield of around 5% was reported. Ehara and Saka reported work on cellulose hydrolysis in high temperature liquid water in the absence of catalyst.273,274 At 280 ºC and 40 MPa the highest HMF yield of 12% was obtained after 240 s.274 Yields of around 10% were reported by Sasaki at 25 MPa and 320 to 350 ºC.275
92| Chapter 2
Seri et al. published work on La(III) catalysed cellulose degradation at 250 ºC in water.188
They reported a maximum HMF yield of 19%. Work by Peng et al. on CrCl3 catalysed dehydration of cellulose to levulinic acid in water at 180 °C showed traces of HMF to a maximum of 3%, but did yield 40% levulinic and formic acid.190 Based on the work of Girisuta et al.189 on LA production from cellulose Yin et al. performed cellulose hydrolysis/dehydration under basic, neutral and acidic conditions.276 At 300 °C the highest HMF yield of 21%, based on carbon, was reported at a cellulose loading of 3 wt% at a pH of 3.0 (HCl). The residence time was defined as the actual time at the defined temperature. It took about 40 min to reach 300 °C. At neutral pH the carbon yield of HMF was 10%. At a pH of 11 only traces of HMF were observed.
In a 10 wt% LiCl solution in DMA containing 10 mol% HCl and 25 mol% CrClx, relative to cellulose, Binder and Raines obtained HMF yields of 22% and 33% using CrCl2 and 119,228 CrCl3, respectively. This difference is surprising since with glucose as substrate (section 119,228 2.4.1.2) no difference between CrCl2 and CrCl3 was observed and CrClx is not expected to influence hydrolysis of cellulose in an acidic environment. Comparable work was recently published by Dutta et al.277 As early as 1958, Snyder claimed the formation of HMF by heating oak wood chips, sprayed with 0.6% sulfuric acid, at 286 °C at 69 bar steam pressure. A theoretical HMF yield of 50-80 % was claimed.278
Amarasekara and Ebede examined cellulose degradation in the presence of 0.5 ZnCl2 mol per mol of glucose unit when heating without added solvent.279 A maximum HMF yield of 9% was reported after 150 s at 200 °C. Some patents by Lightner describe processes for the production of heterocyclic compounds, including HMF and furfural, from biomass feedstock in aqueous environment, but no yields were given.280,281 Asghari and Yoshida published work on the conversion of Japanese red pinewood into valuable chemicals under aqueous acidic conditions.282 The highest HMF yields of around 25% related to the mass of the starting material were obtained at 270 °C and pH 2 at autonomous pressure after 2 min. Dedsuksophon et al. published research on the combined hydrolysis, dehydration, aldol condensation and hydrogenation of lignocellulosic biomass in organic media, namely acetone/DMSO.283 Chareonlimkun et al. reported HMF production from lignocellulosic biomass in hot 258 compressed water using TiO2, ZrO2 and mixed oxide TiO2-ZrO2 catalysts. The best yields
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|93
based on carbon balance were obtained at 250 ºC and 34.5 MPa with TiO2-ZrO2 at a Ti/Zr molar ratio of 1/1 after 5 min. From sugar cane bagasse the HMF and furfural yields were 6- 7% and 10% respectively. The yields of HMF and furfural from rice husk were about 3% and 8% and the yields from corn cob were around 9%, respectively. Glucose, fructose, xylose and anhydroglucose (AHG) were also formed in small amounts from all three substrates. AHG and fructose were present in significantly smaller amounts than glucose and xylose. The experiments were also performed with cellulose and xylan. At 70% cellulose conversion the yield of furan and sugar products added up to approximately 28%, half of which consisted of HMF. About 2% of furfural, 4% of glucose and 6% of fructose were also detected. Xylan yielded approximately 27% furfural and 7% xylose under identical conditions.258 Work by Zhao et al. combined chromium salts with heteropolyacid in order to hydrolyse and dehydrate cellulose to HMF. An HMF yield of 36% at 57% conversion was obtained at 150 °C after 2 h with 5 wt% cellulose and 0.015 M catalyst loading in an aqueous system.284 The yield was improved to 53% at 77% conversion by applying a surfactant type catalyst
Cr[(OSO3C12H25)H2PW12O40]3. In the absence of chromium though, no HMF formation was observed. The formation of small amounts of fructose was generally reported from cellulose under high temperature and high pressure conditions.258,282,283 This is in agreement with the work of Kabyemela122 and Bicker,125 who observed fructose formation starting from glucose under comparable conditions. Chitosan, a copolymer of glucosamine and n-acetylglucosamine, was also shown to yield 285 up to 10 wt% HMF in the presence of SnCl4 in water. Tables 19-21 give an overview of the dehydration of di-, oligo- and polysaccharides in water. An overview of the dehydration in organic solvents is provided in Table 22Table 22. Work by Carlini et al. showed that HMF yields from inulin are comparable to those from fructose, indicating that hydrolysis proceeds much faster than dehydration. The HMF yields from cellulose are generally very low (<10%). Various lignocellulosic biomass sources show yields in the same ballpark. The main difficulty with these substrates is their low solubility in most solvents, except some ionic liquids. This is improved by initial hydrolysis, which yields glucose that is still difficult to dehydrate to HMF, as described in section 2.4.1.2.
94| Chapter 2
Table 19. The dehydration of polysaccharides to HMF in aqueous mixtures in the absence of a catalyst
Substrate Saccharide Temperature Reaction Yield Conversion Selectivity Reference loading (°C) Time (%) (%) (%) (wt%)
Inulin 5 160 4 h 40.9 - - 271
Inulin 5 180 2 h 44 - - 271
Inulin 5 200 45 min 41 - - 271
Cellulose 9 250 5 min 4 15 28 257
Cellulose 4 280 4 mina 11.9 - - 274
Cellulose 3 300 0 sb 10 - - 276
Sugar cane bagasse 9 250 5 min 3c - - 257
Cellulose 2 350 8.8 sa 11 99 11 275 a: continuous flow process; b: reaction time excluding 40 min pre-heating time; c: 3.9% furfural yield
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|95
Table 20. The dehydration of polysaccharides to HMF in aqueous mixtures, catalysed by homogeneous catalysts
Substrate Saccharide Solvent Catalyst Catalyst Temperature Reaction Yield (%) Reference loading loading (°C) Time (wt%)
a Cellulose - Water H2SO4 20 mM 215 1 h 5 272
Cellulose 3 Water HCl pH 3.0 300 0 sb 21 276
d Chicory roots 49 Water H2SO4 pH 1.8 140 2 h 13 195
b Corn starch 19 Water:dioxane1:1 Pyridine:H3PO4 5% : 3% n/n 220-226 30 min 44 250 v/v 5:3 n/n
Inulin 5 Water CO2 6 MPa 160 4 h 45 271
Inulin 5 Water CO2 9 MPa 160 4 h 42 271
Inulin 5 Water CO2 4 Mpa 180 2 h 45 271
Inulin 5 Water CO2 6 Mpa 180 2 h 50 271
Inulin 5 Water CO2 11 Mpa 180 2 h 52 271
Inulin 5 Water CO2 6 MPa 200 45 min 53 271
Inulin 5 Water CO2 9 MPa 200 45 min 49 271
96| Chapter 2
d Oak wood Solid none H2SO4 (0.6 wt% - 286 90 s 50-80 278 chips aq)
b Sucrose 19 Water:dioxane1:1 Pyridine:H3PO4 5% : 3% n/n 220-229 30 min 44 250 v/v 5:3 n/n a: semi-batch flow process; b: reaction time excluding >30 min preheating time; c: 30% fructose yield and 3% glucose yield; d: 69 Bar steam
Table 21. The dehydration of polysaccharides to HMF in aqueous mixtures, catalysed by heterogeneous catalysts
Substrate Saccharide Catalyst Catalyst Temperature Reaction Yield Conversion Selectivity Reference loading loading (°C) Time (%) (%) (%) (wt%)
Cassava 2 Sulfonated carbon- 100 wt% 250 1 min 9.0a - - 286 Waste based catalyst
Cassava 2 Sulfonated carbon- 100 wt% 250 1 min 11a - - 286 Wastee based catalyst
g Cellulose 5 Cr[(DS)H2PW12O40]3 15 mM 150 2 h 53 77 68 284
Cellulose 5 Cr[H2PW12O40]3 15 mM 150 2 h 36 57 62 284
Cellulose 5 H3PW12O40 15 mM 150 2 h 0 33 0 284
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|97
b Cellulose 9 TiO2 100 wt% 250 5 min 13 34 37 257
c Cellulose 9 ZrO2 100 wt% 250 5 min 8.3 25 33 257
a Corn Cob 9 TiO2 100 wt% 250 5 min 7.9 - - 257
a Corn Cob 9 TiO2-ZrO2 100 wt% 250 5 min 8.5 - - 257
Inulin 6 Cubic ZrP2O7 55 wt% 100 1 h 35 39 89 184
d Inulin 6 Cubic ZrP2O7 55 wt% 100 1 h 70 73 97 184
Inulin 6 FeVOP 5 wt% 80 2 h 35 42 83 198
Inulin 6 - 55 wt% 100 1 h 42 44 94 184
Ti(PO4)(H2PO4).2H2O
Inulin 6 - 55 wt% 100 1 h 67d 71 95 184
Ti(PO4)(H2PO4).2H2O
Inulin 6 Niobium phosphate 63 wt% 100 3 h 31 47 66 183
a Rice Husk 9 TiO2-ZrO2 100 wt% 250 5 min 3.3 - - 257
Sucrose 13 Niobium phosphate 53% w/w 100 4 h 14f 30 46 183
a Sugarcane 9 TiO2 100 wt% 250 5 min 6.3 - - 257 bagasse
98| Chapter 2
a Sugarcane 9 TiO2-ZrO2 100 wt% 250 5 min 6.6 - - 257 bagasse a: significant furfural yields (2-11%) observed; b: 6% fructose observed in the reaction mixture c: 8% fructose observed in the reaction mixture; d: HMF extraction with MIBK after 0.5 h and 1 h; e: H2O/acetone/DMSO 90:7:3 (w/w) as the solvent; f: The formed glucose was not converted; g: DS = dodecylsulfate, a surfactant
Table 22. The dehydration of polysaccharides to HMF in organic solvents
Substrate Substrate Solvent Catalyst Catalyst loading Temperature Reaction Yield Conversion Selectivity Reference loading (°C) Time (h) (%) (%) (%) (wt%)
Cellobiose 3 DMF Hydrotalcite/Amber 200 wt% 120 3 35 52 67 263 lyst 15 1:1 w/w
Cellulose 4 DMA (10 wt% CrCl2/HCl 25 mol%/10 mol% 140 2 22 - - 119 LiCl)
Cellulose 4 DMA (10 wt% CrCl3/HCl 25 mol%/10 mol% 140 2 33 - - 119 LiCl)
a Melezitose 10 DMSO LaCl3 5 mol% 120 2 8 - - 216
b Raffinose 10 DMSO LaCl3 5 mol% 120 2 22 - - 216
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|99
2- Starch 3 DMSO/water SO4 /ZrO2-Al2O3 6 wt% 150 4 48 - - 270 (4:1 v/v) (Zr:Al = 3:7 n/n)
Sucrose 10 DMA CrCl3/NH4Br 9.5 mol%/ 55 mol% 100 1 87 - - 287
Sucrose 3 DMF Hydrotalcite/Amber 200 wt% 120 3 54 58 93 263 lyst 15 1:1 w/w
Sucrose 7 DMSO LaCl3 5 mol% 120 2 47 - - 216
Sucrose 9 NMP FeCl3/Et4NBr 10 mol%/18 mol% 90 3 40 - - 248
Tapioca 50 H2O/Acetone/ WO3-ZrO2 100 wt% 230 - 22 - - 283 flour DMSO unknown ratio
Turanose 7 DMSO LaCl3 5 mol% 120 2 12 - - 216
a: a trisaccharide containing one fructose and two glucose moieties b: a trisaccharide containing one fructose, one glucose and one galactose moiety
100| Chapter 2
2.4.2 HMF formation in biphasic solvent systems
The low HMF yields sometimes observed simultaneously with high LA yields point to the importance of repressing the HMF rehydration. In addition HMF is assumed to react with sugars and sugar intermediates to form so-called humins. A number of groups have combined the sugar dehydration under aqueous conditions with an in situ extraction of the HMF in an organic phase. By continuously removing HMF, these undesired side reactions can be suppressed to a large extent. The first to apply such a continuous extraction system in HMF synthesis from carbohydrates was Peniston, named as inventor on a patent from 1956.288
2.4.2.1 Fructose dehydration in biphasic solvent systems
Peniston performed HMF synthesis in 0.05 M aqueous sulfuric acid containing 2 wt% fructose, claiming 68% yield after 8 min at 170 °C in the presence of one equivalent of n- butanol.288 The analysis was performed using UV only, without chromatography, which makes the quantification less reliable. Kuster and Van der Steen performed biphasic HMF formation in a continuous stirred tank reactor (CSTR), studying the effects of temperature, fructose concentration, residence time, phosphoric acid concentration and MIBK-water ratio on the yield and selectivity.199 In general an increase in temperature and acid concentration led to an increase in the reaction rate. Longer residence times led to higher HMF yields without significant losses in selectivity. An increase in the MIBK-water ratio led to an increase in both yield and selectivity. The highest HMF yield of 69% at 94% fructose conversion was obtained with 1M fructose and 0.1 M H3PO4 at 190 °C with an MIBK-water ratio value of 7.5 and a residence time of 5 min. The laboratory for organic and agricultural chemistry in Toulouse continued in this direction in the first half of the 1980’s. Mercadier et al. published a three-part study on fructose dehydration in biphasic water-organic systems, catalysed by ion-exchange resins.211,289,290 An aqueous fructose solution of about 25 wt% was reacted in the presence of Lewatit SPC 108 and SPC 118 at 88 °C for 5 h. In the absence of extraction solvent the HMF yield was 10% at 69% conversion, whereas in the presence of 9 eq. MIBK the HMF yield was 28% at 45% conversion.211 Increasing the reaction time to 15 h in the presence of MIBK increased the HMF yield to 56% at 84% conversion. Tests with other extraction solvents under the same conditions for 5 h showed that alkanes, such as heptane, were highly unfavourable with HMF yields below 10% at conversions around 30%.290 The application of
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|101 benzonitrile and 2,2’-dichloroethylether as extraction solvents showed results comparable to MIBK with regard to selectivity, but at higher reaction rates with HMF yields of around 40% at around 65% conversion after 5 h.211 Experiments in which the MIBK:water ratios were varied showed increasing HMF yield and selectivity with increasing amounts of MIBK up to a MIBK:water ratio of 9. In a later publication under comparable conditions yields of >40% at around 80% conversion were reported.291 Gaset et al. also reported work on a DMSO:MIBK biphasic system.292 A 97% HMF yield was reported by reacting a 20 wt% fructose solution in DMSO at 76 °C in flow (213 ml/h) with an MIBK counterflow (1500 ml/h). Rigal and Gaset reported an HMF yield of 74% from aqueous fructose with SPC-108 resin as catalyst after 7 h at 78 °C under continuous extraction by MIBK.186 From sorbose 47% HMF yield was obtained after 12 h. El Hajj used a number of solid acid resins for aqueous fructose dehydration to HMF in the presence of MIBK.196 HMF yields as high as 90% were obtained by dehydrating 20 wt% fructose in water, catalysed by 1 wt% Duolite C20 at 90 ºC for 14 h in presence of 9 equivalents of MIBK. MIBK was refreshed every 2 h, resulting in a total use of 8 l for 100 ml of water containing 25 g of fructose. In the mid 1990’s Moreau and co-workers ventured into zeolite catalysed sugar chemistry.102,293-295 In a water/MIBK (1:5 v/v) system several zeolites were tested: H-Y faujasites and H-mordenite catalysts with different Si/Al ratios, H-beta and H-ZSM5. In general the more active catalysts provided relatively low selectivities at high conversion. Through dealumination higher selectivities were obtained by lowering the activity. A 9 wt% aqueous fructose solution was reacted at 165 ºC in the presence of 29 wt% catalyst relative to fructose. The best result of 69% HMF yield at 76% conversion was obtained after 60 min using H-mordenite with an Si/Al ratio of 11.102,293,294 A recent publication by Ordomsky et al. on zeolite catalysed dehydration of fructose in water and water/MIBK systems showed significantly lower selectivity to HMF at comparable conversion with H-mordenite with an Si/Al ration of 11.7, reporting 42% yield at 64% conversion.296 In the same publication the effect of deactivation of the outer surface of the zeolite on the selectivity was researched. In the absence of extracting solvent the selectivity increased significantly at all conversions, improving the maximum selectivity from 48% to 63% at 56% conversion. This effect was explained by a decrease in the rate of oligomerisation due to steric constraints in the pores of the acid catalyst. Adding 1 eq. of MIBK did not increase the maximum selectivity, but did increase the selectivity at high conversions.
102| Chapter 2
The group of Dumesic performed extensive research on biphasic solvent systems, applying a continuous extraction of an aqueous reaction medium by an organic phase from which HMF was continuously recovered by solvent evaporation.170,297-301 Experiments in 0.25 M HCl at 180 ºC for 2.5 to 3 min with 30 wt% aqueous fructose yielded 26% HMF at 50% conversion in the absence of an extraction.170 In the presence of 2 weight equivalents MIBK the yield was improved to 55% at 75% conversion. The HMF yield was further improved by modification of both phases. From 10 wt% fructose in 1:1 (w/w) water:DMSO an HMF yield of 85% at 95% conversion was obtained with HCl at pH 1 after 4 min at 170 °C in combination with 2 eq. (w/w) 7:3 (w/w) MIBK:2-butanol.300 In 3:7 (w/w) water:DMSO in the absence of catalyst and with 1 eq. (w/w) dichloromethane as extracting solvent, an 87% HMF yield at 100% conversion was obtained at 140 °C after 2 h. To avoid the use of corrosive homogeneous acids, Dumesic et al. tested propylsulfonic acid-functionalised and thiopropyl-modified silicas in order to combine acidity with the promoting effect of DMSO and DMF in one catalyst.299 SBA-15, a mesoporous silica with two-dimensional hexagonal pore structure, and A380, a non-porous silica, were used as supports. The experiments were performed in batch with 30 wt% aqueous fructose, 11 wt% catalyst and 2 eq. (w/w) MIBK/2-butanol (7:3 w/w) at 180 °C for 2 h. The SBA-15 experiments showed no beneficial effect of thiopropyl modification, but the installation of propylsulfonic acid groups improved the activity and selectivity, yielding 49% HMF at 66% conversion after 30 min. The A380 experiments showed no improvement upon propylsulfonic acid functionalisation of the thiopropyl modified precursor. An experiment with Amberlyst 70 resulted in 58% yield at 86% conversion after 10 min. The difference in activity between Taa-SBA-15 and Taa-A380 was explained by the higher acid loading on SBA-15. Recently this group published additional work with SBA-15 as the catalyst.302 In another publication by Dumesic et al. different water/DMSO and water/NMP mixtures were tested with either MIBK or DCM as extraction solvent.301 The experiments were catalysed by the acidic ion exchange resin DIAION® PK216 at 90 °C and 120 °C with a feed concentration of 10 wt%. An 83% HMF yield at 98% conversion was obtained in 4:6 (w/w) water/NMP at 90 °C after 18 h with MIBK as extraction solvent. An increase in selectivity was observed with decreasing fructose concentration, increasing DMSO or NMP content and increasing extraction solvent.170,300,301 A study on different types of organic solvents as extracting agents in combination with an aqueous phase, saturated with salt, was performed by Román-Leshkov and Dumesic, 297 focussing on primary and secondary alcohols (C3-C6), ketones (C3-C6) and cyclic ethers.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|103
Typical experiments were performed with 30 wt% fructose in the aqueous phase, catalysed by HCl (pH 0.6) at 150 °C for 35 min with a Vorg/Vaq value of 3.2. The highest yields were around 69%, with 2-butanone and 2-pentanol achieving this at conversions of 84% and 83% respectively. 2-Butanol is presented as the solvent providing the highest HMF selectivity of 85%, though this is only at 67% conversion, resulting in 57% yield. The results indicate that the highest yields at non-complete conversions are obtained with the C4 and C5 extracting solvents, with the secondary alcohols leading to the highest selectivities and the ketones resulting in the highest conversions. If the reaction system is assumed to be biphasic, then it is surprising that the type of extraction solvent affects the fructose conversion rate, unless this is dependent on the HMF concentration. Several mineral salts were tested for saturation of the aqueous phase in which application of chlorides showed a beneficial effect compared to bromides and sulfates on the HMF yield and selectivity, especially for sodium, potassium and cesium salts. This is in contrast with what Binder and Raines reported with regard to the effect of halides on the selectivity of fructose dehydration to HMF (section 2.3.1.1, Scheme 13).119 Work by Lima et al. in a water-toluene system, catalysed by Al-TUD-1 showed yields of less than 20% HMF from fructose.124 A publication by McNeff an co-workers describes the 303,304 dehydration of a number of carbohydrate feedstocks in a two-phase flow process. TiO2- catalysed fructose dehydration at 200 °C with 1/3 volume equivalent n-Butanol as extraction solvent yielded 18% HMF. An increase in the amount of n-butanol decreased the yield.
Hansen et al. published results with boric acid-catalysed dehydration of fructose in H2O, in which the effects of the catalyst, extraction solvent and salt concentration were studied.305 With a 30 wt% aqueous fructose solution containing 85 mol% boric acid and 4 eq of MIBK as extraction solvent 22% HMF yield at 43% conversion was obtained after 45 min at 150 °C. When the catalyst loading was doubled the yield increased to 28% at 53% conversion. The addition of halide salts to the system resulted in a remarkable improvement of the HMF yield and selectivity. The presence of 0.87 M NaCl improved the yield to 46% at 70% conversion. When THF was used as extraction solvent the yield was improved to 51% at 75% conversion after 75 min under otherwise identical conditions. When halide salts were applied the selectivity improved from just over 50% to around 65-70%. In a publication by Yang et al. very high HMF yields and selectivities were reported when niobic acid treated with phosphoric acid (NA-p) was applied.306 A 6 wt% aqueous solution of fructose was reacted in the presence of 8 wt% NA-p and 1.5 eq (V/V) 2-butanol at 160 °C for 50 min, leading to 89% HMF yield at 90% conversion. This is a surprising result, because the
104| Chapter 2
NA-p is a heterogeneous Brønsted acidic catalyst, which would be expected to also catalyse the rehydration of HMF in the presence of water, since even in the presence of extraction solvent, a significant amount of HMF remains in the aqueous phase. When comparing these data to the state of the art described in Tables 23 and 24, it becomes clear that this is the only publication claiming essentially quantitative yields. Similar experiments were performed with modified hydrated tantalum oxide (TA-p).307 From fructose 90% HMF at 94% conversion was obtained after 100 min at 160 °C.
Fan et al. applied a solid heteropolyacid, Ag3PW12O40, to the dehydration of fructose to HMF. A 23 wt% aqueous mixture with a 2.25 volume ratio of MIBK as extracting solvent was reacted at 120 °C.308 Various catalyst loadings and reaction times were tested with a reported maximum HMF yield of 78% at 83% conversion after 1 h. In this work also significant HMF yields were reported in the absence of catalyst, namely 33% HMF at 47% conversion after 1 h at 120 °C. The levulinic acid yields were also reported, showing increased amounts at higher conversions.
Zhao et al. applied approximately 20 wt% of Cs2.5H0.5PW12O40 as heteropolyacid catalyst in the dehydration of 30 wt% aqueous fructose in the presence of 3 volume equivalents of MIBK, yielding 74% HMF at 78% conversion after 1 h at 115 °C.309 Also experiments were reported with different catalyst loadings, fructose concentrations and at different temperatures and reaction times. Starting with a 10 wt% fructose solution or a 50 wt% fructose solution both gave almost the same result of 71-74% HMF yield at around 85% conversion, which is slightly worse than the result with 30 wt% fructose. Increasing the temperature increased the reaction rate, but at comparable selectivities. Brasholz et al. published work on fructose dehydration in 32% aqueous HCl in flow with an equal flow of dichloromethane as extraction solvent.310 With a 10 wt% fructose solution at 100 °C and a residence time of 1.67 min a CMF yield of 81% was reported. No HMF was observed in this reaction. Tables 23 and 24 provide an overview of the results of fructose dehydration in biphasic solvent systems. Although some other solvents have been tested the majority of the work was done with MIBK and n-BuOH as extraction solvents. It is clear that the selectivity towards HMF formation was increased significantly from the regular water-based systems.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|105
Table 23. Fructose dehydration to HMF in biphasic solvent systems, catalysed by homogeneous catalysts
Fructose Reaction solvent Extraction solvent Org./Aq. Catalyst Catalyst T Reaction Yield Conv. Sel. Ref. conc. phase ratio loading (°C) Time (%) (%) (%) (wt%) (V/V)
g 5 Water MIBK 2 AlCl3 50 mol% 130 5 min 61 - - 205
30 8:2 Water:DMSO (w/w) 7:3 MIBK:2-BuOH (w/w) 2 w/w H2SO4 0.25 M 180 2.5-3 60 85 71 170 min
2 Water n-BuOH 1 H2SO4 45 mol% 170 8 68 - - 288
30 Water (0.87 M NaCl) 7:3 MIBK:2-BuOH (w/w) 4 H3BO3 85 mol% 150 45 min 50 72 70 305
30 Water MIBK 4 H3BO3 85 mol% 150 45 min 22 43 50 305
30 Water MIBK 4 H3BO3 171 mol% 150 45 min 28 53 54 305
30 Water (0.87 M LiCl) MIBK 4 H3BO3 85 mol% 150 45 min 45 69 66 305
30 Water (0.87 M NaCl) MIBK 4 H3BO3 85 mol% 150 45 min 46 70 65 305
30 Water (0.87 M KCl) MIBK 4 H3BO3 85 mol% 150 45 min 44 67 65 305
30 Water (0.44 M MgCl2) MIBK 4 H3BO3 85 mol% 150 45 min 52 81 65 305
30 Water (0.87 M NaCl) THF 4 H3BO3 85 mol% 150 1.25 h 51 75 68 305
106| Chapter 2
a 3.6 Water MIBK 1 H3PO4 10 mol% 180 3 min 25 35 70 199
a 18 Water MIBK 1 H3PO4 10 mol% 180 3 min 25 43 57 199
a 18 Water MIBK 1 H3PO4 5 mol% 180 3 min 19 31 62 199
a 18 Water MIBK 1 H3PO4 50 mol% 180 3 min 51 75 68 199
a 18 Water MIBK 3.5 H3PO4 10 mol% 190 5 min 55 77 71 199
a 18 Water MIBK 5.7 H3PO4 10 mol% 190 5 min 64 89 72 199
a 18 Water MIBK 7.5 H3PO4 10 mol% 190 5 min 69 94 73 199
a 18 Water MIBK 4.6 H3PO4 10 mol% 200 2 min 56 79 71 199
a 18 Water MIBK 6.2 H3PO4 10 mol% 200 2 min 62 85 73 199
a 18 Water MIBK 4 H3PO4 10 mol% 200 3.8 min 62 85 73 199
a 18 Water MIBK 1 H3PO4 5 mol% 213 3 min 48 72 66 199
a 18 Water MIBK 1 H3PO4 10 mol% 220 3 min 52 85 61 199
10 Water MIBK 3 HCl 0.25 M 140 15 minb 74 - - 310
10 1:1 Water:DMSO (w/w) 7:3 MIBK:2-BuOH (w/w) 2 w/w HCl pH 1 170 4 min 85 95 89 300
10 1:1 Water:DMSO (w/w) 7:3 MIBK:2-BuOH (w/w) 2 w/w HCl pH 2 170 8 min 82 95 86 300
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|107
30 Water 7:3 MIBK:2-BuOH (w/w) 2 w/w HCl 0.25 M 180 2.5-3 69 86 80 170 min
30 8:2 Water:DMSO (w/w) 7:3 MIBK:2-BuOH (w/w) 2 w/w HCl 0.25 M 180 2.5-3 71 87 82 170 min
30 7:3 (8:2 Water:DMSO):PVP 7:3 MIBK:2-BuOH (w/w) 1 w/w HCl 0.25 M 180 2.5-3 68 82 83 170 (w/w) min
30 7:3 (8:2 Water:DMSO):PVP 7:3 MIBK:2-BuOH (w/w) 2 w/w HCl 0.25 M 180 2.5-3 76 89 85 170 (w/w) min
50 7:3 (8:2 Water:DMSO):PVP 7:3 MIBK:2-BuOH (w/w) 2 w/w HCl 0.25 M 180 2.5-3 71 92 77 170 (w/w) min
30 Water MIBK 2 w/w HCl 0.25 M 180 2.5-3 55 75 73 170 min
10 3:7 Water:DMSO (w/w) DCM 1 w/w None 140 2 h 87 100 87 300 a: continuous stirred tank reactor (CSTR); b: continuous flow; 6% fructose observed in the reaction mixture c: 8% fructose observed in the reaction mixture; d: HMF extraction with MIBK after 0.5 h and 1 h; e: H2O/acetone/DMSO 90:7:3 (w/w) as the solvent; f: The formed glucose was not converted; g: isolated yield
108| Chapter 2
Table 24. Fructose dehydration to HMF in biphasic solvent systems, catalysed by heterogeneous catalysts
Fructose Reaction solvent Extraction solvent Org./Aq. Catalyst Catalyst T Reaction Yield Conv. Sel. Ref. conc. phase ratio loading (°C) Time (%) (%) (%) (wt%) (V/V)
23 Water MIBK 2.25 Ag3PW12O40 6.7 wt% 120 1 h 75 84 88 308
23 Water MIBK 2.25 Ag3PW12O40 3.3 wt% 120 1 h 78 83 94 308
9 Water Toluene 2.3 Al-TUD-1 67 wt% 170 2 h 20 76 26 124
30 Water 7:3 MIBK:2-BuOH (w/w) 2 (w/w) Amberlyst 70 11 wt% 180 10 min 58 86 67 299
23 Water MIBK 3 Cs2.5H0.5PW12O40 20 wt% 115 1 h 74 78 95 309
10 3:7 Water:DMSO (w/w) MIBK 1 (w/w) Diaion PK216 500 wt% 90 4 h 73 87 84 301
10 1:1 Water:DMSO (w/w) MIBK 1 (w/w) Diaion PK216 500 wt% 90 12 h 73 90 81 301
10 4:6 Water:NMP (w/w) MIBK 1 (w/w) Diaion PK216 500 wt% 90 18 h 83 98 85 301
20 Water MIBKb 9 Duolite C20 1 wt% 90 14 h 89 - - 196
9 Water MIBK 5 H-form Zeolite, 29 wt% 165 30 min 50 54 92 102,293 Si/Al = 11
9 Water MIBK 5 H-form Zeolite, 29 wt% 165 1 h 69 76 91 102,293
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|109
Si/Al = 11
9 Water MIBK 5 H-form Zeolite, 29 wt% 165 1 h 53 65 81 293 Si/Al = 6.9
6 Water MIBK 5 H-form Zeolite, 67 wt% 165 6.5 h 42 64 66 296 Si/Al = 11.7
20 DMSO MIBK 7 IE resin (acidic) Fixed bed. 76 -a 97 - - 292
20 NMP MIBK 7 IE resin (acidic) fixed bed 76 -a 88 - - 292
20 DMF MIBK 7 IE resin (acidic) Fixed bed 76 -a 84 - - 292
30 7:3 (8:2 7:3 MIBK:2-BuOH (w/w) 1 (w/w) IE resin (acidic) 100 wt% 90 8-16 h 54 83 65 170 Water:DMSO):PVP (w/w)
4.8 Water MIBK 11.5 Lewatitt SPC 108 168 meq/l 80 14 h 44 79 56 291
20 Water 2,2'-dichloroethylether 9 Lewatitt SPC 108 61 mol% 88 5 h 42 68 62 211 H+
20 Water Benzonitrile 9 Lewatitt SPC 108 61 mol% 88 5 h 40 63 63 211 H+
20 Water Heptane 9 Lewatitt SPC 108 61 mol% 88 5 h 6 21 28 211
110| Chapter 2
H+
20 Water MIBK 9 Lewatitt SPC 108 61 mol% 88 15 h 56 84 66 211 H+
20 Water MIBK 9 Lewatitt SPC 108 61 mol% 88 5 h 28 45 63 211 H+
4.8 Water MIBK 11.5 Lewatitt SPC 108 168 meq/l 88 3 h 42 81 52 291
20 Water None - Lewatitt SPC 108 61 mol% 88 5 h 10 69 14 211 H+
6 Water 2-Butanol 1.5 NA-p 8 wt% 160 50 min 89 90 99 307
30 Water 7:3 MIBK:2-BuOH (w/w) 2 (w/w) SBA-15 11 wt% 180 2 h 31 59 52 299
- Water MIBK - SPC-108 fixed bed 78 7 ha 74 - - 186
- Water MIBK - SPC-108 Fixed bed 78 12 ha 47 - - 186
30 Water 7:3 MIBK:2-BuOH (w/w) 2 (w/w) Taa-A380 11 wt% 180 2 h 38 62 61 299
30 Water 7:3 MIBK:2-BuOH (w/w) 2 (w/w) Taa-SBA-15 11 wt% 180 30 min 49 66 74 299
6 Water 2-Butanol 1.5 TA-p 8 wt% 160 100 min 90 94 96 307
a 23 Water n-BuOH 0.33 TiO2 Fixed bed 200 3 min 18 - - 303
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|111
a 23 Water n-BuOH 3 TiO2 Fixed bed 200 3 min 11 - - 303
30 Water 7:3 MIBK:2-BuOH (w/w) 2 (w/w) TP-A380 11 wt% 180 2 h 43 67 64 299
30 Water 7:3 MIBK:2-BuOH (w/w) 2 (w/w) TP-SBA-15 11 wt% 180 2 h 32 61 52 299 a: continuous flow; b: continuous extraction
112| Chapter 2
2.4.2.2 Glucose dehydration in biphasic solvent systems
Cope was the first to apply biphasic mixtures in HMF production from glucose, claiming 21-25% isolated yield of HMF by reacting a 50 wt% aqueous glucose solution at 160 °C for around 9 h in the presence of approximately 20 volumetric equivalents of MIBK.311 The only acid present was that formed during the reaction. Rigal and Gaset performed aldose dehydration in a water/MIBK system, catalysed by an acidic ion-exchange resin. HMF yields below 10% were obtained at 78 °C.186. Work on glucose dehydration by the group of Dumesic generally resulted in low HMF yields.300 A 24% HMF yield at 50% conversion was obtained in aqueous HCl (pH 1) from 10 wt% glucose in the presence of 2 weight eq. MIBK/2-butanol (7:3 w/w) at 170 °C after 17 min. Experiments without acid were carried out at 140 °C in water/DMSO 3:7 in combination with an equal weight of dichloromethane resulting in an HMF yield of 30% at 62% conversion after 4.5 h. More recently work has been reported with a bifunctional catalyst 312 system in which AlCl3 was combined with HCl to yield up to 62% HMF. A comparable result was obtained by Yang et al. in a biphasic water-NaCl/THF system with AlCl3 as the catalyst.313 Dumesic et al. also applied solvents derived from lignin as extracting phase.314 In research by Lima et al. in a water-toluene system, catalysed by Al-TUD-1 yields of less than 20% of HMF were obtained from glucose.124 McNeff et al. described the dehydration of a number of carbohydrate feedstocks in a two-phase flow process.303,304 In this process an aqueous substrate mixture and organic extraction solvent were pumped into the reactor from separate reservoirs and reacted at 160-200 °C in the presence of a TiO2 catalyst. n-Butanol and methyl isobutyl ketone (MIBK) were applied as extraction solvents. Different organic/aqueous ratios were tested. The highest HMF yields from glucose were between 25% and 30% at 180 °C with MIBK as extraction solvent (organic:aqueous 10:1) and with aqueous feed containing 23 or 50 wt% glucose. With n-butanol as extraction solvent yields were typically between 6 and 13% from glucose.303 The authors explain the lower yield compared to MIBK systems by too high solubility of water in n-butanol under reaction conditions. Experiments using Lite Corn Syrup, honey and glucose/fructose/sucrose mixtures were also described. At 180 °C McNeff also tested the effect of the addition of homogeneous acid to similar 303 reaction mixtures as described above. Both HCl (0.05-0.15 M) and H3PO4 (0.10 M) were tested for this purpose. A run with just 0.15 M HCl as catalyst yielded 13% HMF from
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|113
glucose. Combining TiO2 with 0.15 M, 0.10 M and 0.05 M HCl or 0.10 M H3PO4 led to yields of 37%, 34%, 26% and 33% respectively. Hansen et al. also used boric acid as catalyst for the glucose dehydration under the same circumstances as applied for fructose, but found that the yields were very low.305 An HMF yield of 14% was obtained at 41% conversion after 5 h at 150 °C in the presence of 50 g/l NaCl and 4 volumetric equivalents of MIBK, compared to 46% HMF at 70% conversion for fructose after 45 min at under identical conditions. In the same system as applied for fructose, Yang et al. reacted glucose, reporting very high HMF yields and selectivities.306 A 6 wt% aqueous solution of glucose, in the presence of 8 wt% NA-p and 1.5 eq (V/V) 2-butanol, yielded 49% HMF at 72% conversion at 160 °C for 110 min. With TA-p as the catalyst 58% HMF yield at 70% conversion was reported after 140 min.307 Similar to the fructose data, the yields and selectivities are much higher than would be expected when comparing to the results of other publications under similar conditions. 308 Fan et al. used Ag3PW12O40 as a catalyst for glucose dehydration to HMF. With 13 wt% catalyst at 130 °C a 76% HMF yield at 90% conversion after 4 h was reported. What is surprising about this result when compared to the work with fructose, yielding 75% HMF and 11% levulinic acid at 95% conversion with the same catalyst loading at 120 °C after 1 h, is a significantly lower levulinic acid yield of 6%. This means that under harsher conditions HMF appeared to be more stable. Nikolla et al. published promising results in glucose dehydration with a bifunctional catalyst system of tin and titanium ß-zeolites in combination with HCl. From a 10 wt% glucose solution in 26 wt% aqueous NaCl solution at pH 1 an optimum HMF yield of 57% at 79% conversion was obtained in the presence of 0.5 wt% Sn-beta and 3 eq. (v/v) of THF as extraction solvent after 70 min at 180 °C. When Ti-beta was used, the yield was 53% at 76% conversion after 105 min.315
Degirmenci et al. prepared a heterogeneous catalyst that mimicked the CrCl2/ionic liquid system applied by Zhao et al.134 by covalently binding 1-(3-triethoxysilylpropyl)-3- 316 methylimidazolium chloride on SBA-15, followed by coordination of CrCl2. A number of solvent systems were researched, with a water:DMSO/2-BuOH:MIBK system giving the highest yield (35%) and selectivity (70%). Leaching of chromium caused deactivation of the catalyst in recycling experiments. Recently some additional work was published on glucose dehydration in aqueous/organic biphasic solvent systems with acidic catalysts.317,318319
114| Chapter 2
Mascal and Nikitin investigated a water-1,2-dichlororethane system for HCl catalysed carbohydrate conversion to 5-chloromethylfurfural (CMF).187 In the presence of LiCl a homogeneous mixture of glucose was reacted for 30 h at 65 °C. Three types of furanic compound were isolated with a combined yield of around 85%: CMF at 71% yield, 2-(2- hydroxyacetyl)furan (HAF) at 7% yield and HMF at 8% yield. CMF can be easily converted to HMF, 5-methylfurfural (MF), 5-ethoxymethylfurfural (EMF) and 2,5-dimethylfuran.187,320 This system was later improved by application of a closed system at 100 °C for 3 h in which 1,2-dichloroethane was refreshed every hour.321 From a 1 wt% glucose solution 81% CMF yield was reported. Brasholz et al. published similar research, in which a 2 wt% glucose 310 solution in 32 wt% HCl(aq) was reacted in flow at 120 °C. With a residence time of 5 min and an equal flow of reaction mixture and 1,2-dichloroethane a CMF yield of 58% was obtained. Table 25 provides an overview of the HMF yields obtained in glucose dehydration reactions in aqueous systems with in situ extraction. The data on glucose dehydration in aqueous mixtures, as described in section 2.4.1.2 and Table 15, are difficult to compare with the presented data on glucose dehydration under in situ extraction because of the small amount of experiments that were performed under comparable conditions. It is clear, though, that even with extracting solvent present the HMF yields from glucose remain low, with the exception of the results obtained with bifunctional catalyst systems.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|115
Table 25. Glucose dehydration to HMF in biphasic solvent systems
Glucose Reaction solvent Extraction solvent Org./Aq. Catalyst Catalyst T Reaction Yield Conv. Sel. Ref. conc. phase ratio loading (°C) Time (%) (%) (%) (wt%) (V/V)
10 3:7 Water:DMSO (w/w) DCM 1 (w/w) None 140 4.5 h 30 62 48 300
23 Water MIBK 2.25 Ag3PW12O40 13 wt% 130 4 h 76 90 85 308
9 Water Toluene 2.3 Al-TUD-1 67 wt% 170 6 h 18 76 23 124
d e 5 Water MIBK 2 AlCl3 50 mol% 130 5 min 43 - - 205
30 Water (0.87 M NaCl) MIBK 4 H3BO3 85 mol% 150 5 h 14 41 34 305
10 1:1 Water:DMSO (w/w) 7:3 MIBK:2-BuOH (w/w) 2 (w/w) HCl (pH 1) 170 17 min 24 50 47 300
6 Water 2-BuOH 1.5 NA-p 8 wt% 160 110 min 49 72 68 306
10 2:8 Water:DMSO 7:3 MIBK:2-BuOH 1 SBA-15- - 150 3 h 35 50 70 316
[PMIm]Cl/CrCl2
10 Water n-BuOH 3 Sn-Beta/HCl 0.5 mol% Sn 160 90 min 20 77 26 315
(pH 1)
10 Water (26 wt% NaCl) n-BuOH 3 Sn-Beta/HCl (pH 0.5 mol% Sn 160 90 min 41 75 55 315
116| Chapter 2
1)
10 Water (26 wt% NaCl) THF 3 Sn-Beta/HCl (pH 0.5 mol% Sn 180 70 min 57 79 72 315 1)
- Water MIBK - SPC-108 Fixed bed 78 22 hc 9 - - 186
-a Water MIBK - SPC-108 Fixed bed 78 16 hc 7 - - 186
-b Water MIBK - SPC-108 Fixed bed 78 16 hc 5 - - 186
6 Water 2-Butanol 1.5 TA-p 8 wt% 160 140 min 58 70 83 306
10 Water (26 wt% NaCl) THF 3 Ti-Beta/HCl (pH 0.5 mol% Ti 180 105 min 53 76 70 315 1)
c 23 Water MIBK 10 TiO2 Fixed bed 180 2 min 29 - - 303
c 50 Water MIBK 10 TiO2 Fixed bed 180 2 min 26 - - 303
c 50 Water (0.15 M HCl) MIBK 10 TiO2/HCl Fixed bed 180 2 min 37 - - 303
c 23 Water n-BuOH 1 TiO2 Fixed bed 200 3 min 13 - - 303
c 23 Water MIBK 10 ZrO2 Fixed bed 180 2 min 21 - - 303
a: mannose as the substrate; b: galactose as the substrate; c:continuous flow; d: microwave irradiation; e: isolated yield;
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|117
2.4.2.3 The dehydration of oligo- and polysaccharides in biphasic solvent systems
Gaset et al. reported almost identical HMF yields of ~70% from inulin and crude polyfructans from Jerusalem artichoke as from fructose in an acidic resin-catalysed water- MIBK system at 78 °C.186 HMF yields from sucrose and raffinose were 41% and 27% respectively. This ratio is consistent with the amount of fructose moieties present in each molecule. Yang et al. applied TA-p catalysts on 6 wt% inulin in water with 1.5 (v/v) 2-butanol as extraction solvent, resulting in 87% HMF yield at 95% conversion at 160 °C after 2.5 h.307 In their studies on fructose and glucose dehydration, the group of Dumesic reported work on di- and polysaccharides.300,301 In water/DMSO (5:5 w/w) in the presence of 2 eq (w/w) MIBK/2-butanol (7:3 w/w) and catalysed by HCl, 10 wt% inulin yielded 75% HMF at 98% conversion after 5 min at 170 °C and pH 1.5.300 In water/DMSO 4:6 (w/w), sucrose yielded 50% HMF at 65% conversion after 5 min, starch yielded 26% HMF at 61% conversion after 11 min and cellobiose yielded 27% HMF at 52% conversion after 10 min at pH 1 Experiments without acid were carried out at 140 °C in water/DMSO 3:7, in combination with dichloromethane as organic phase in equal weight to the aqueous phase. At 100% conversion, the HMF yield from inulin was 70% after 2h. From cellobiose an HMF yield of 38% at 85% conversion after 9.5 h and for starch a 36% HMF yield at 91% conversion after 11 h were reported. In another publication by the same group different water/DMSO and water/NMP mixtures in the presence of acidic ion exchange resin DIAION® PK216 were tested with either MIBK or DCM as extraction solvent.301 The feed concentration was 10 wt%, equal amounts (weight) of the aqueous and organic layers were used and the organic/catalyst ratio was 1.
Experiments with 10 wt% inulin at 90 °C with H2O-NMP (4:6 w/w) and H2O:DMSO (5:5 w/w) yielded 69% and 62% HMF respectively at 100% conversion. At 120 °C in water/DMSO (5:5 w/w) in combination with DCM a 61% HMF yield was obtained at 100% conversion after 5.5 h The same three experiments were performed with sucrose as substrate, yielding around 40% HMF yield, with regard to fructose content, at around 70% conversion. McNeff et al. described the dehydration of a number of carbohydrate feedstocks in a two- 303,304 phase flow process. A 23 wt% sucrose solution catalysed by anatase TiO2 yielded 16% HMF at 180 °C and 3 min residence time under continuous extraction with MIBK, in which the MIBK flow rate was three times as high as that of the substrate solution. A 5% starch
118| Chapter 2 solution yielded 15% HMF at 2 min residence time with ten fold MIBK flow rate. Cellulose dehydration was performed in the same system via a hot extraction of solid cellulose, which meant that hot water was passed through solid cellulose, prior to the entering the reactor. At 270 °C a 30% yields was reported at 87% conversion with a 2 min contact time and five fold MIBK flow. Work on sucrose dehydration by Tarabanko et al. in biphasic systems consisting of an acidic aqueous solution of sucrose in combination with a number of aliphatic alcohols produced the appropriate HMF ethers in yields of 14-25% relative to sucrose.234 The aqueous phase consisted of 0.44 M sucrose, and 4.2 M Na2SO4. The reactions were performed at 102 °C and the organic to aqueous phase ratio was 2. The yields of both the HMF ethers and alkyl levulinates decreased with alkyl chain length. Mascal and Nikitin developed a biphasic system for cellulose conversion to 5- chloromethylfurfural (CMF).187,321,322 A homogeneous mixture of cellulose in concentrated HCl yielded three types of furanic compound with a combined yield of ~85% after 30 h at 65 ºC in the presence of 1,2-dichloroethane: CMF at 71% yield, 2-(2-hydroxyacetyl)furan (HAF) at 8% yield and HMF at 5% yield187. These are essentially equal to the yields reported for glucose, which shows that the rate-limiting step in this process is the glucose dehydration and not cellulose hydrolysis. Sucrose produced comparable, slightly higher combined yields of around 91%. This CMF can be easily converted to HMF, 5-methylfurfural (MF), 5- ethoxymethylfurfural (EMF) and 2,5-dimethylfuran.187 This system was later improved by application of a closed system at 100 °C for 3 h in which 1,2-dichloroethane was refreshed every hour.321 From a 1 wt% substrate solution cellulose yielded 84% CMF, sucrose yielded 90% CMF and corn stover yielded 80% CMF. Brasholz et al. reported 51% CMF yield from 2 wt% sucrose in 32% aqueous HCl at 2.5 min residence time in a flow process at 100 °C, in which dichloromethane was used as extraction solvent in equal flow as the reaction mixture.310 Table 26 describes results obtained in the dehydration of sucrose, raffinose and polysaccharide in biphasic solvent systems. From inulin HMF yields >70% were reported at full conversion. Dumesic et al. obtained HMF yields of around 50% from sucrose, which is approximately the average between that of the typical yields form fructose and glucose under similar conditions, indicating that sucrose behaves as een equimolar mixture of fructose and glucose. Only very little work has been done on polysaccharides made up of glucose monomers, like starch and cellulose. Cellobiose has been used as a model compound for cellulose, but this is only a disaccharide that presents far less challenges, especially with
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|119 regard to solubility. McNeff et al. are the only group to publish cellulose dehydration under continuous extraction, presenting around 30% HMF yield. Starch dehydration has also been reported by McNeff et al., as well as by Dumesic’s group, both under very different reaction conditions, which makes it difficult to compare their results.
120| Chapter 2
Table 26. Di- and polysaccharide dehydration to HMF in biphasic solvent systems
Substrate Substrate Reaction solvent Extraction Org.c/Aq. Catalyst Catalyst T. Reaction Yield Conv. Sel. Ref. loading solvent phase ratio loading (°C) Time (%) (%) (%) (wt%) (V/V)
Cellobiose 3 Water Toluene 2.3 Al-TUD-1 200 wt% 170 6 h 12c 98 12 124
Cellobiose 10 3:7 Water:DMSO DCM 1 w/w None 140 9.5 h 38 85 45 300 (w/w)
b a Cellulose - Water MIBK 5 TiO2 Fixed bed 270 2 min 30 87 34 303
Inulin 10 4:6 Water:NMP MIBK 1 w/w Diaion PK216 500 wt% 90 18 h 69 100 69 300 (w/w)
Inulin 10 5:5 Water:DMSO 7:3 MIBK:2- 2 w/w HCl (pH 1.5) - 170 5 min 75 98 77 300 (w/w) BuOH (w/w)
Inulin 9 Water MIBK 5 H-form Zeolite, 29 wt% 165 1 h 39 44 88 293 Si/Al =11
Inulin 6 Water 2-Butanol 1.5 NA-p 8 wt% 160 140 min 54 86 63 306
Inulin 10 3:7 Water:DMSO DCM 1 w/w None 140 2.5 h 70 100 70 300 (w/w)
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|121
Inulin - Water MIBK - SPC-108 Fixed bed 78 12 ha 67 - - 186
Inulin 6 Water 2-Butanol 1.5 TA-p 8 wt% 160 140 min 87 95 92 307
Jerusalem 9 Water MIBK 5 H-form Zeolite, 29 wt% 165 1 h 57 66 87 293 Artichoke Si/Al =11
Jerusalem - Water MIBK - SPC-108 Fixed bed 78 15 ha 73 - - 186 Artichoke
Jerusalem 6 Water 2-Butanol 1.5 TA-p 8 wt% 160 2 h 79 91 87 307 Artichoke Juice
Raffinose - Water MIBK - SPC-108 Fixed bed 78 10 ha 27 - - 186
Starch 10 3:7 Water:DMSO DCM 1 w/w none 140 11 h 36 91 40 300 (w/w)
a Starch 5 Water MIBK 10 TiO2 Fixed bed 180 2 min 15 - - 303
Sucrose 3 Water Toluene 2.3 Al-TUD-1 200 wt% 170 6 hd 17 100 17 124
Sucrose 10 4:6 Water:DMSO 7:3 MIBK:2- 2 w/w HCl (pH 1) - 170 5 min 50 65 77 300 (w/w) BuOH (w/w)
Sucrose 9 Water MIBK 5 H-form Zeolite, 29 wt% 165 1 h 28 57 49 293
122| Chapter 2
Si/Al =11
Sucrose 10 3:7 Water:DMSO DCM 1 w/w none 140 4.5 h 51 82 62 300 (w/w)
Sucrose - Water MIBK - SPC-108 Fixed bed 78 12 ha 41 - - 186
a Sucrose 23 Water MIBK 3 TiO2 Fixed bed 180 3 min 16 - - 303 a: continuous flow; b: chamber with solid cellulose between pre-heater and reactor; c: 50% glucose and 10% fructose yield; d: 26% glucose and 15% fructose yield
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|123
2.4.3 HMF formation in ionic liquids
2.4.3.1 A definition of ionic liquids
Ionic liquids typically consist of a combination of organic and inorganic ions and are unique compared to conventional solvents due to their extremely low vapour pressure. They cannot be classified as simply molten salts as they have some distinctly different properties, like their low melting point, which is typically below 100 ºC, and relatively low viscosity.323 The first reported ionic liquid stems from 1914.324 A significant number of reviews on ionic liquids has been published.323,325-328 A lot of the early ionic liquids are air and moisture sensitive, greatly limiting their application. An important breakthrough came with the development of a series of air and moisture stable imidazolium type ionic liquids, which were first reported by Wilkes’ group in 1992.327,329 The most common ionic liquids are N,N’-dialkylimidazolium, N-alkylpyridinium, alkylamonium and alkylphosphonium based. A good insight on their synthesis is given in reviews by Welton and Marsh.326,330
2.4.3.2 The dehydration of carbohydrates to HMF in ionic liquids
The last 5-8 years have seen a strong growth in interest towards the use of ionic liquids in biomass conversion. This started with a publication on fructose dehydration by Moreau’s group in 2003.331 Since then a variety of imidazolium-based ionic liquids has been used in the dehydration of especially fructose. However, already twenty years earlier Fayet and Gelas reported the use of pyridinium chloride, a low melting salt, in sugar conversion, obtaining HMF in 70% yield from fructose.332 Zakrzewska et al. reported an elaborate research on the solubility of a number of carbohydrates in a wide range of ionic liquids, divided in six categories. Ionic liquids based on imidazolium, pyridinium, benzotriazonium, pyrrolidinium, alkylammonium and alkyphosphonium cations were researched.333 Recently the synthesis of HMF has been reviewed by Zakrzewska et al.,20 Ståhlberg et al.18 and Lima et al.19
Fructose dehydration in ionic liquids Lansalot-Matras and Moreau performed fructose dehydration in 1-Butyl-3- methylimidazolium ([BMIm]) type ionic liquids both in the presence and absence of DMSO.
124| Chapter 2
In the absence of DMSO as a solvent a 52% HMF yield was obtained from an 8 wt% fructose solution in 1-butyl-3-methylimidazolium tetrafluoroborate ([BMIm]BF4), catalysed by ~3 weight equivalents of Amberlyst 15 at 80 °C after 3 h.331 At reaction times >3 h the yield started decreasing. When a mixture of ([BMIm]BF4) and DMSO was used the yields improved to 87% after 32 h with 2 eq Amberlyst 15. In 2006 Moreau et al. published work on fructose and sucrose dehydration in the Brønsted acidic 1-H-3-imidazolium chloride [HMIm]Cl.334 With a fructose concentration of 23 wt% a yield of 92% HMF at 98% conversion was reported after 45 min at 90 ºC. The sucrose dehydration gave a similar molar yield with respect to the fructose moiety and conversion of the glucose moiety was estimated at around 3%. Zhao et al. used [EMIm]Cl as catalyst in sugar dehydration to HMF. In the absence of catalyst 73% HMF yield was obtained at full conversion by heating 9% fructose in [EMIm]Cl 134 at 120 °C for 3 h. When 6 mol% PtCl2 was used as a catalyst an HMF yield of 83% was obtained at 99% conversion. Further research on fructose dehydration in acidic ionic liquids was done by Hu et al.335 PyHCl and [HMim]Cl experiments led to similar results of 70% HMF yield at 90-95% conversion from a 17 mol% fructose solution at 80 ºC for 1 h. Experiments in ChoCl/citric acid were reported to yield 75% HMF at 92% fructose conversion. Continuous extraction with ethyl acetate improved overall HMF yield to 92% at 98% conversion in ChoCl/citric acid. HMF was found to be stable under reaction conditions in citric acid for 1 h. Fructose dehydration in [EMIm]BF4 with 17 wt% fructose, catalysed by 10 mol% SnCl4 at 100 °C for 3 h yielded 62% HMF at 100% conversion.144 Cao et al. looked into a number of imidazolium based ionic liquids as solvent and catalyst for fructose dehydration.336 In the absence of catalyst 63% HMF yield was obtained at 93% conversion from a 9 wt% fructose solution in [BMIm]Cl after 50 min at 120 °C. When using [HMIm]Cl as the ionic liquid at 120 °C the HMF yield was 7% at 66% conversion and at 100 °C no HMF formation was reported for any of the ionic liquids tested, which conflicts with the work of Moreau334 and Hu,335 who both found significant HMF yields at lower temperatures under otherwise comparable conditions. Additional experiments were performed with sulfuric acid and chromium(III) chloride hexahydrate as catalysts.336 With 10 mol% sulfuric acid 83% and 77% HMF yield at full conversion were obtained after 50 min at 100 °C in [BMIm]Cl and [BeMIm]Cl respectively. With 10 mol% chromium(III) chloride hexahydrate HMF yields of 75% and 71% at full conversion were obtained after 2 h at 120 °C in [BMIm]Cl and 1-Benzyl-3-methylimidazolium chloride ([BeMIm]Cl) respectively.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|125
A combination of [EMIm][HSO4] and MIBK as extracting solvent was used by Lima et al. to obtain 88% HMF yield from 0.67 M fructose after 30 min at 100 °C.337 Replacing the
MIBK with toluene yielded 79% HMF and subsequently replacing the [EMIm][HSO4] with [BMIm]Cl led to a drop in yield to 16% at 23% conversion. The authors concluded that the Brønsted acidity plays an important role in fructose conversion to HMF. Some recent literature on fructose dehdyration in imidazolium based ionic liquids in the absence of an additional catalyst provided high yields in the range of 60-97%.338-342 Also work on fructose dehydration in tetraethylammonium chloride343 and betaine hydrochloride344 was recently published.
GeCl4 catalysed dehydration of carbohydrates to HMF in ionic liquids by Zhang et al. resulted in very high yields of HMF.268 A 5 wt% solution of fructose in [BMIm]Cl, in the presence of 10 mol% GeCl4 relative to the substrate, yielded 92% HMF at full conversion after 5 min at 100 °C. Recycling of the system showed no significant decrease in yields after at least five cycles. In the absence of GeCl4 no HMF was detected under otherwise identical conditions, which is consistent with the observations of Cao et al.336 Recently results were 345 also published with GeCl4 in mixtures of [BMIm]Cl and DMSO or DMF at 25 °C. Generally a higher fraction of ionic liquid led to higher yields, up to 70%. It is not clear from the data if this is due to activity or selectivity, since no conversion data are shown. Fructose dehydration in [BMIm]Cl in combination with various homogeneous and heterogeneous catalysts by Qi et al. led to HMF yields around 80%.346 With 1 weight equivalent of Amberlyst 15 in a 5 wt% fructose mixture yielded 83% HMF at 100% conversion at 80 °C after 10 min. Increasing the initial fructose concentration resulted in a gradual decrease in HMF yield, leading to a 5% drop in HMF yield at a tenfold increase of initial fructose concentration, past which no additional drop was observed. A similar trend was observed by Yong et al.143 The ionic liquid was recycled successfully at least seven times. HMF was shown to be stable under reaction conditions in the absence of fructose and water. Water started having a negative effect on the HMF yields when the initial concentration was higher than 5%. A later publication by the same group reported a 78%
HMF yield by microwave heating of 5 wt% fructose in [BMIm]Cl with 20 mol% CrCl3 at 100 °C for 1 min.347 Furthermore Qi published a study on fructose dehydration in various 2- 348 imidazolium based ionic liquids, catalysed by sulfated zirconia (SO4 /ZrO2). From a 5 2- wt% fructose solution in [BMIm]Cl with 40 wt% SO4 /ZrO2 88% HMF yield was obtained at 96% conversion after 30 min at 100 °C. With ZrO2 in [BMIm]Cl under the same conditions a an HMF yield of 55% at 60% conversion was obtained, leading to the same
126| Chapter 2 selectivity of around 90%. This showed that the reaction rate was improved by using sulfuric acid treated ZrO2. Recently a number of articles have been published in which heterogeneous acid catalysts were applied for fructose dehdyration in ionic liquids,349-351 reporting very high yields of HMF, up to 94%.351 Qi et al. also performed dehydration experiments at room temperature.174 This required the addition of a co-solvent, which decreased the activity of the system, in order to allow proper stirring. Only a small amount of 0.15 molar equivalents of co-solvent was added. Acetone,
DMSO, methanol, ethanol, ethyl acetate and supercritical CO2 (150 Bar at 35 °C) were tested, in experiments with 5 wt% fructose in the presence of 100 wt% Amberlyst 15 and 6 h reaction time. The HMF yields were 78-82% at 89-95% conversion. Yong et al. applied N-heterocyclic carbenes (NHC) as ligands (Scheme 32) for chromium chlorides that were used as catalysts for fructose and glucose dehydration in ionic liquids.143 A solution of 9 wt% fructose in 1-butyl-3-methyl-imidazolium chloride ([BMIm]Cl) was reacted in the presence of 9 mol% of NHC/Cr catalyst at 100 ºC for 6 h. HMF yields over 95% were obtained from fructose for one of the NHC ligands (47, Scheme 32) in combination with either CrCl2 or CrCl3. The best results were obtained with the most bulky types of NHC tested. The authors explain that this effect could be caused by the protection of the metal centre from steric crowding from the ionic liquid by the ligand.
Scheme 32. N-heterocyclic carbene ligands for CrCl2 and CrCl3 used as catalysts for fructose and glucose dehydration.143
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|127
269 Wei et al. published work on fructose dehydration in [BMIm]Cl, catalysed by IrCl3 and
AuCl3·HCl. After 3 h at 80 °C an HMF yield of 63% at 93% conversion and 22% at 50% conversion were obtained in the presence of 7 mol% IrCl3 and AuCl3·HCl respectively. With
10 mol% IrCl3 an HMF yield of 86% was achieved at full conversion after 30 min, however no temperature was mentioned. The IrCl3 catalysed system led to HMF yields of 70% at 100
°C and 65% at 120 °C, at around full conversion. In the case of AuCl3·HCl it was suggested that only HCl plays a role and its additional effect disappears at higher temperatures because of the increased catalytic activity of the ionic liquid. Zhang et al. investigated the use of polytungstic acid (PTA), encapsulated in a metal- organic framework built up from chromium carboxylate (MIL-101), as a catalyst for fructose dehydration to HMF.226 MIL-101 is a metal-organic framework with a chromium carboxylate cubic structure. Various encapsulated PTA (PTA/MIL-101) catalysts were tested and HMF yields as high as 80% at 87% conversion were obtained after 1 h in [EMIm]Cl at 80 °C. The chromium in the MIL-101 structure was found to be essentially inactive in fructose dehydration. Ilgen et al. published work on highly concentrated low melting mixtures of sugars and several hydrogen bond donor and acceptor additives.352 When using a fructose-N,N,N’,N’- tetra-methylurea melt (9:1 w/w) in combination with FeCl3 at 100 ºC for 1 h an 89% HMF yield was obtained. Fructose-ChoCl (4:6 w/w) with 10 mol% p-Toluene sulfonic acid (pTsOH) yielded 67% HMF. Recently a comparable HMF yield 72% was reported using a 353 very high fructose concentration in a ChoCl/CO2 eutactic solvent system. In a 2:1 (w/w) mixture of water and 1-(4-sulfonic acid) butyl-3-methylimidazolium hydrogen sulfate a 92% yield of HMF at 98% conversion was reported in the presence of 9 volumetric equivalents of MIBK after 2 h at 120 °C.354 Avantium patented the formation of HMF ethers and esters from fructose in ionic liquid by adding the appropriate alcohol or carboxylic acid to the ionic liquid.355a In [EMIm]Cl-acetic acid (1:4 w/w) 4 wt% of fructose was reacted in the presence of 4 wt% CrCl2 at 100 °C for 3 h, yielding 22% HMF and 72% AMF. Similarly, Kraus and co-worker were able to obtain EMF in 54% yield from fructose using imidazolium propanesulfonic acids in and ethanol, hexane mixture.289b Tables 27-29 give an overview of the results obtained in ionic-liquid based fructose dehydration. High yields, consistently over 65% have been reported. Some groups even reported yields in excess of 90%. The reaction temperatures were generally between 80-100 °C, which is similar to those used in the experiments in aprotic polar solvents and
128| Chapter 2 significantly lower than those used in aqueous systems. The ionic liquids applied in this field of research are often Brønsted acidic, making them catalytically active in sugar dehydration. It is thus difficult to explain the large differences in the reported yields. A strong point in this system is the apparent stabilising effect of the ionic liquid on HMF, blocking rehydration reactions to, among others, levulinic acid.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|129
Table 27. Fructose dehydration to HMF in ionic liquids without additional catalyst
Fructose Solvent Temperature Reaction Time Yield Conversion Selectivity Reference concentration (°C) (%) (%) (%)
8 wt% [BMIm]BF4-DMSO (5:3 v/v) 80 32 h 36 - - 331
23 wt% / 17 mol% [HMIm]Cl 90 45 min 92 98 94 334
9 wt% [EMIm]Cl 120 3 h 73 100 73 134
9 wt% [EMIm]Cl 120 3 h 78 100 78 145
17 mol% [HMIm]Cl 80 1 h 70 97 73 335
17 mol% PyHCl 80 1 h 69 92 76 335
17 mol% [BMIm]HSO4 80 1 h 57 100 57 335
17 mol% ChoCl/citric acid 80 1 h 79 99 80 335
17 mol% ChoCl/citric acida 80 1 h 92 98 94 335
20 t% ChoCl/CO2 120 1.5 h 72 - - 353
b 9 wt% [EMIm][HSO4] 120 30 min 88 100 88 337
9 wt% [BMIm]Cl 80 3 h 0 4 0 269
130| Chapter 2
9 wt% [BMIm]Cl 100 3 h 28 60 47 269
9 wt% [BMIm]Cl, [HMIm]Cl 100 50 min 0 - 0 336
9 wt% [BMIm]Cl 120 50 min 63 93 68 336
9 wt% [HMIm]Cl 120 50 min 7 66 11 336
9 wt% [BMIm]Cl 140 50 min 60 100 60 336
9 wt% [HMIm]Cl 140 50 min 22 85 26 336
9 wt% [BenzylMIm]Cl 140 50 min 53 100 53 336
9 wt% [BMIm]Cl 120 3 h 48 100 48 269
5 wt% [BMIm]Cl 100 5 min 0 - 0 268
c 40 wt% H2O/ [SBMIm][HSO4] 2:1 (w/w) 120 2 h 92 98 94 354 a: continuous HMF extraction by ethyl acetate; b: 2.3 volume equivalents of MIBK as extraction solvent; c: 9 volume equivalents of MIBK as extraction solvent
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|131
Table 28. Fructose dehydration to HMF in ionic liquids in the presence of homogeneous catalysts
Fructose Solvent Catalyst Catalyst Temperature Reaction Yield Conversion Selectivity Reference concentration loading (°C) Time (%) (%) (%) (wt%) (mol%)
9 [BMIm]Cl AuCl3·HCl 7 80 3 h 22 50 44 269
9 [BMIm]Cl AuCl3·HCl 7 100 3 h 44 88 50 269
9 [BMIm]Cl AuCl3·HCl 7 120 3 h 48 98 49 269
9 [EMIm]Cl CrCl2 6 mol% 80 3 h 60 100 60 142
b 9 [EMIm]Cl CrCl2 6 80 3 h 65 92 71 134
c 9 [EMIm]Cl CrCl3 6 80 3 h 69 92 75 134
a 5 [BMIm]Cl CrCl3 20 100 1 min 78 - - 347
9 [BenzylMIm]Cl CrCl3•6 H2O 10 100 2 h 71 100 71 336
9 [BMIm]Cl CrCl3•6 H2O 10 100 2 h 75 100 75 336
9 [EMIm]Cl CrCl3•6H2O 6 mol% 80 30 min 80 100 80 142
9 [HMIm]Cl CrCl3•6 H2O 10 100 2 h 44 92 47 336
132| Chapter 2
9 [HMIm]Cl CrCl3•6 H2O 25 100 2 h 62 100 62 336
5 [BMIm]Cl CuCl2 18 80 10 min 80 94 85 346
90 N,N,N’,N'- FeCl3 10 100 1 h 89 - - 352 tetramethylurea
5 [BMIm]Cl GeCl4 10 50 20 min 84 96 88 268
5 [BMIm]Cl GeCl4 10 80 5 min 90 100 90 268
5 [BMIm]Cl GeCl4 10 100 5 min 92 100 92 268
5 [BMIm]Cl H2SO4 18 80 10 min 70 82 86 346
9 [BenzylMIm]Cl H2SO4 10 100 50 min 77 100 77 336
9 [BMIm]Cl H2SO4 10 100 50 min 83 100 83 336
9 [HMIm]Cl H2SO4 10 100 50 min 7 55 13 336
9 [HMIm]Cl H2SO4 25 100 50 min 82 100 82 336
5 [BMIm]Cl HfCl4 10 100 5 min 58 100 58 268
9 [BMIm]Cl IrCl3 7 80 3 h 63 93 68 269
9 [BMIm]Cl IrCl3 7 100 3 h 70 98 72 269
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|133
9 [BMIm]Cl IrCl3 7 120 3 h 65 100 65 269
9 [BMIm]Cl IrCl3 10 - 0.5 h 86 100 86 269
9 [BMIm]Cl NHC/CrCl2 9 100 6 h 96 - - 143
9 [BMIm]Cl NHC/CrCl2 9 100 6 h 89 - - 143
9 [BMIm]Cl NHC/CrCl3 9 100 6 h 96 - - 143
9 [BMIm]Cl NHC/CrCl3 9 100 6 h 90 - - 143
9 [EMIm]Cl PtCl2 6 80 3 h 83 99 84 134
8 [BMIm]BF4- PTSA 10 80 32 h 68 - - 331 DMSO (5:3 v/v)
40 ChCl pTsOH 10 100 0.5 h 67 - m/a 352
17 [EMIm]BF4 SnCl4 10 100 3 h 62 100 62 144
a: heating by microwave irradiation; b: 6% glucose yield; c: 3% glucose yield
134| Chapter 2
Table 29. Fructose dehydration to HMF in ionic liquids in the presence of heterogeneous catalysts
Fructose Solvent Catalyst Catalyst concentration Temperature Reaction Yield Conv. Sel. Ref. conc. (wt%) (wt%) (°C) Time (%) (%) (%)
5 [BMIm]Cl:MeOH (6.6:1 n/n) Amberlyst 15 100 25 6 h 82 93 88 174
8 [BMIm]BF4 Amberlyst 15 267 80 3 h 52 - - 331
8 [BMIm]BF4-DMSO (5:3 v/v) Amberlyst 15 100 80 32 h 75 - - 331
8 [BMIm]BF4-DMSO (5:3 v/v) Amberlyst 15 200 80 32 h 87 - - 331
5 [BMIm]Cl Amberlyst 15 100 80 10 min 83 99 85 346
5 [BMIm]Cl Dowex 50WX8 100 80 10 min 82 92 90 346
9 [EMIm]Cl PTA/MIL-101b 40 80 1 h 80 87 92 226
2- a 5 [BMIm]Cl SO4 /ZrO2 40 100 30 min 88 96 92 348
2- a 5 [HexylMIm]Cl SO4 /ZrO2 40 100 30 min 89 100 89 348
a 5 [BMIm]Cl ZrO2 40 100 30 min 55 60 91 348 a: heating by microwave irradiation; b: MIL-101 encapsulated polytungstic acid
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|135
Glucose dehydration in ionic liquids In 2007 a publication by Zhao et al. marked a breakthrough in sugar dehydration to HMF by producing HMF in approximately the same yield from fructose and glucose.134 From 9 wt% glucose in [EMIm]Cl with 1 mol% CrCl2 an HMF yield of 68% at 94% conversion was obtained at 100 °C after 3 h. From fructose under comparable conditions, though at 80 °C, the
HMF yield was 65% at 92% conversion in the presence of CrCl2. Cao et al. performed similar experiments in various ionic liquids with CrCl3•6H2O at 120 °C for 1 h, obtaining HMF yields around 60%.336 With the same catalyst in tetraethylammonium chloride 71% HMF yield was achieved after 10 min at 130 °C.356 Also work in tetrabutylammonium 357 chloride in combination with CrCl2 was performed with a reported HMF yield 56%. As discussed in section 2.4.3.2, Yong et al. used chromium chlorides in fructose and glucose dehydration in ionic liquids in combination with N-heterocyclic carbene (NHC) ligands, which are described in Scheme 32.143 A solution of 9 wt% glucose in 1-Butyl-3- methyl imidazolium chloride ([BMim]Cl) was reacted in the presence of 9 mol% of NHC/Cr catalyst at 100 ºC for 6 h. HMF yields of around 80% were obtained from fructose using a number of different NHC ligands in combination with either CrCl2 or CrCl3. A separate experiment under the exact same conditions, but under argon instead of air, gave only 65% HMF yield from glucose, leading to the suggestion that oxygen played a role in the catalytic process, though it is unclear how.
Lima et al. used 6 mol% of CrCl3 in 0.67 M glucose in a [BMIm]Cl and toluene mixture to obtain 91% HMF yield at 91% conversion after 4 h at 100 °C.337
Hu et al. applied SnCl4 as a catalyst for glucose dehydration in a number of ionic liquids 144 and DMSO. Reactions were performed in [EMIm]BF4 with glucose concentrations between 9 and 26 wt% and 10 mol% SnCl4 at 100 °C for 3 h. The HMF yield with 9 wt% glucose was 53% at 97% conversion. HMF yields around 60% were reported for glucose concentrations of 13-26 wt%. With fructose, sucrose and cellobiose they reported essentially the same yields. 268 Zhang et al. applied GeCl4 in [BMIm]Cl to glucose dehydration. The effect of temperature, catalyst loading and water content were investigated. When varying the temperature between 80 °C and 120 °C it was shown that both activity and selectivity improved with temperature, whereas increasing the catalyst loading from 5 wt% to 15 wt% only showed a slight effect in initial reaction rate, but not on the selectivity. From a 5 wt% reaction mixture an HMF yield of 48% was obtained at 99% conversion after 30 min at 120
136| Chapter 2
°C. The formation of humins was also observed. The addition of water to the reaction mixture had a negative effect on the HMF selectivity and conversely the addition of mol sieves led to an increase in HMF yield. Experiments in other alkyl imidazolium chloride ionic liquids showed a decrease in HMF selectivity with increasing alkyl chain length. This was later also reported by Shi et al. for fructose dehydration in imidazolium based ionic liquids.358 Replacement of the chloride counter ion by non-halides had a detrimental effect on the activity and selectivity of the system. Glucose conversion to HMF in an ionic liquid/CrCl3 system under microwave irradiation (MI) was investigated by Li et al.252 An HMF yield of 91% was reported after reacting 9 wt% glucose in [BMIm]Cl for 1 min under 400 W MI. When applying an oil bath at 100 ºC for 60 min only 17% HMF was obtained. Qi et al. published work on glucose dehydration in an ionic liquid system resembling that 134 347 of Zhao, using [BMIM]Cl in combination with CrCl3. HMF formation at temperatures between 90 and 140 °C was studied using 5 wt% glucose in ionic liquid containing around 18 wt% CrCl3. The highest yield of around 70% at 95% conversion was obtained at 120 °C after 10 min and at 140 °C after 30 s. HMF was found to be stable under the reaction conditions in the absence of glucose. Fructose, at a yield of less than 5%, was mentioned as one of the by- products. In recycling experiments a gradual decrease in activity was observed after the third cycle. The authors observed a colour change in the ionic liquid to deep brown that remained - upon recycling, which was attributed to soluble humin. Similar ionic liquids with HSO4 as counterion instead of Cl- performed significantly worse with yields below 10% at conversions - - over 80%. The authors provide an explanation for the difference between Cl and HSO4 based on the difference in acidity. According to other work the metal concentration in the ionic liquid determines if it behaves as a base or as an acid.359 Under the conditions that were used it behaves as a base, explaining its ability to isomerise glucose as described by 134,347 - Zhao. HSO4 , however, was mentioned to be a rather strong Brønsted acid, making the ionic liquid/CrCl3 mixture too acidic for the selective dehydration to HMF.
A bifunctional catalyst system was applied by Qi et al., using heterogeneous ZrO2 as a base to catalyse isomerisation to fructose, which subsequently converted to HMF in a mixture of [HexylMIm]Cl and water.360 The highest HMF yield obtained was 53% at 92% glucose conversion in a 1:1 (w/w) mixture of [HexylMIm]Cl and water with 40 wt% ZrO2 after 10 min at 200 °C. Recently a very high HMF yield of 68% was reported by Guo et al., applying a lignin- derived solid acid catalyst in DMSO-[BMIm]Cl mixture.361 This is a surprising result, since
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|137 the state of the art clearly indicates that a basic or Lewis acid catalyst is required in combination with an acid catalyst to obtain such high HMF yields from glucose. Chidambaram and Bell published work on glucose dehydration to HMF and a subsequent hydrogenation to 2,5-dimethylfuran (DMF) in two steps.362 For the dehydration to HMF a number of mineral acids, organic acids and heteropolyacids was tested in [BMIm]Cl at 1 mol% acid concentration for 3 h at 120 °C. With 12-tungstophosphoric acid (H3PW12O40) a yield of 66% at 82% conversion was obtained and with H2SO4 a yield of 61% at 93% conversion was claimed. The highest HMF yield of 61% at 93% glucose conversion was 362 obtained in [BMIm]Cl, with 1 mol% H2SO4 as catalyst at 120 ºC after 3 h. Zhang’s publication reported only 10% HMF yield at >90% glucose conversion under similar conditions with [EMIm]Cl as ionic liquid at 100 ºC.134 Chidambaram and Bell tested a 362 number of heteropolyacids, obtaining 66% HMF yield at 82% conversion with H3PW12O40. 362 An HMF yield of 63% at 71% conversion was obtained with H3PMo12O40. In [EMIm]Cl a 59% HMF yield at 66% conversion was reported. An improvement to 97% HMF yield at 99% conversion was claimed when the experiment was performed in the presence of acetonitrile.362 Differences in activity were found for different types of mineral acid with similar acidity in water. The role of the counterion in the catalysis and a difference in acidic behaviour in ionic liquids were mentioned as possible causes for this phenomenon. In a publication by Ståhlberg et al. ionic liquids were used as solvents in combination with lanthanide catalysts in order to convert glucose to HMF. A 24% HMF yield at 65% 363 134 conversion was obtained with Yb(OTf)3 in [BMIm]Cl. In contrast to findings by Zhang longer alkyl chains on the imidazolium ionic liquid were found to increase the activity of the system. Ståhlberg also applied boric acid as a catalyst at 120 °C, reporting an optimum boric acid/glucose ratio between 0.8 and 1.145 With 1 eq boric acid an HMF yield of 40% at 87% conversion was obtained after 3 h at 120 °C. With 0.8 eq boric acid 41% HMF yield at 95% conversion was reported. The HMF yield decreased with increasing boric acid concentrations, which is explained by the formation of more stable sugar-boric acid complexes that inhibit dehydration to HMF. It was also shown that the chloride counterion plays an important role in the process, as [EMIm] type ionic liquids with non-halide counterions, apart from 6% HMF yield with sulfonates, did not yield any HMF. Recently Kokhlova et al. performed an NMR monitoring of carbohydrate dehydration with boron derivatives in ionic liquids, presenting very high conversions.364
138| Chapter 2
Ilgen et al. obtained 45% HMF yield form glucose-choline chloride (4:6 w/w) melt with 10 352 mol% CrCl2 at 110 °C after 1 h. With CrCl3 in the yield was only 31%. The conversions were not reported.
Zhang and Zhao published work on glucose dehydration in [BMIm]Cl, catalysed by CrCl3 immobilised on hydroxyapatite.365 With 5 wt% glucose and 60 wt% catalyst an HMF yield of 40% was obtained at 78% conversion after 2.5 min at 400 W microwave irradiation. In the absence of catalyst 5% HMF was obtained at 81% conversion after 3 min and with chromium-free hydroxyapatite the yield was 8% at 81% conversion. Avantium patented the formation of HMF ethers and esters from glucose in ionic liquid by adding the appropriate alcohol or carboxylic acid to the ionic liquid.355 In [EMIm]Cl-acetic acid (1:4 w/w) 4 wt% of glucose was reacted in the presence of 4 wt% CrCl2 at 100 °C for 3 h, yielding 1.3% HMF and 5.1% AMF. Table 30 provides an overview of the research on glucose dehydration in ionic liquids. It is clear from multiple publications that chromium plays an important role in obtaining high
HMF yields from glucose. The HMF yields in the presence of CrCl2 or CrCl3 are in the same region as those obtained form fructose in ionic liquids. The differences in activity between
CrCl2 and CrCl3 are not significant. This phenomenon has been researched by Zhang et al., 142 who compared the activity of CrCl2, CrCl3 and CrCl3•6H2O. The initial rate of glucose conversion was highest for CrCl3•6H2O, followed by CrCl2 and CrCl3 respectively, with the latter showing an induction period. The CrCl3 was observed to dissolve much slower into the reaction mixture than the other CrCl2 and CrCl3•6H2O, explaining the lower initial conversion rate. The system with Cr3+ furthermore led to a significantly higher HMF yield of 72% at 97% conversion than 60% HMF yield at 93% conversion obtained from the Cr2+ system. What the publications by Zhang134 and Pidko136 and Hu144 have in common is the absence of an added Brønsted acid, so the actual dehydration could take place in a number of ways. As none of the three publications report the presence of fructose in the reaction mixture, starting from glucose, it is also possible that the dehydration takes place from an intermediate species in stead of fructose. Given the fact that regular Brønsted acid catalysed dehydration of fructose yields substantial amounts of HMF (depending on the solvent) in the absence of an isomerisation catalyst like CrCl2 or SnCl4 and glucose does not, it appears to be more likely that CrCl2 and SnCl4 function purely as an isomerisation catalyst. What catalyses the actual dehydration is not yet clear. It could be that free Cl- generates enough hydrochloric acid.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|139
Table 30. Glucose dehydration to HMF in ionic liquids
Glucose conc. Solvent Catalyst Catalyst T (°C) Reaction Yield Conv. Sel. Ref. (wt%) concentration Time (%) (%) (%)
8 [BMIm]Cl CF3COOH 1 mol% 120 3 h 44 58 75 362
8 [BMIm]Cl CF3SO3H 1 mol% 120 3 h 40 87 46 362
8 [BMIm]Cl CH3SO3H 1 mol% 120 3 h 42 73 58 362
9 [BMIm]Cl Cr(NO3)3 7 mol% 100 3 h 37 82 45 269
9 [EMIm]Cl CrCl2 6 mol% 100 3 h 62 93 67 142
9 [EMIm]Cl CrCl2 6 mol% 100 3 h 68 94 72 134
40 ChoCl CrCl2 10 mol% 110 1 h 45 - - 352
2.2 M NBu4Cl CrCl2 10 mol% 110 4 h 54 - - 357
a 5 [BMIm]Cl CrCl3 20 mol% 90 1 h 40 51 80 347
9 [BMIm]Cl CrCl3 3.6 wt% 100 1 h 17 - - 252
a 5 [BMIm]Cl CrCl3 20 mol% 100 30 min 56 77 73 347
140| Chapter 2
a 5 [BMIm]Cl CrCl3 20 mol% 100 1 h 67 85 78 347
9 [BMIm]Cl CrCl3 7 mol% 100 3 h 35 79 45 269
9 [EMIm]Cl CrCl3 6 mol% 100 3 h 44 72 62 134
40 ChoCl CrCl3 10 mol% 110 1 h 31 - - 352
a 5 [BMIm]Cl CrCl3 20 mol% 120 10 min 69 94 73 347
a 23 [BMIm]Cl CrCl3 20 mol% 120 10 min 55 97 56 347
a 5 [BMIm]Cl CrCl3 10 mol% 120 10 min 66 88 75 347
b 9 [BMIm]Cl CrCl3 6 mol% 120 4 h 91 91 100 337
a 5 [BMIm]HSO4 CrCl3 20 mol% 120 10 min 5 86 6 347
a 5 [EMIm]Cl CrCl3 20 mol% 120 10 min 72 97 74 347
a 5 [HexylMIm]Cl CrCl3 20 mol% 120 10 min 63 94 67 347
a 5 [BMIm]Cl CrCl3 20 mol% 140 0.5 min 71 96 74 347
a 9 [BMIm]Cl CrCl3 3.6 wt% Unknown 1 min 91 - - 252
9 [EMIm]Cl CrCl3•6H2O 6 mol% 100 3 h 72 97 74 142
9 [BenzylMIm]Cl CrCl3•6H2O 25 mol% 120 1 h 65 - - 336
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|141
9 [BMIm]Cl CrCl3•6H2O 25 mol% 120 1 h 67 - - 336
9 NEt4Cl CrCl3•6H2O 10 mol% 130 10 min 71 - - 356
5 [BMIm]Cl GeCl4 10 mol% 100 75 min 38 93 42 268
c 5 [BMIm]Cl GeCl4 10 mol% 100 75 min 48 93 52 268
5 [BMIm]Cl GeCl4 10 mol% 120 30 min 48 99 48 268
8 [BMIm]Cl H2SO4 1 mol% 120 3 h 61 93 66 362
a 9 [BMIm]Cl H2SO4 10 wt% Unknown 1 min 49 - - 252
9 [BMIm]Cl H3BO3 100 mol% 120 3 h 22 68 32 145
9 [BMIm]Cl H3BO3 80 mol% 120 3 h 14 47 30 145
9 [EMIm]Cl H3BO3 100 mol% 120 3 h 40 87 46 145
9 [EMIm]Cl H3BO3 80 mol% 120 3 h 41 95 43 145
9 [HexylMIm]Cl H3BO3 80 mol% 120 3 h 32 68 47 145
9 [HMIm]Cl H3BO3 80 mol% 120 3 h 19 95 20 145
9 [OctylMIm]Cl H3BO3 80 mol% 120 3 h 26 63 41 145
8 [BMIm]Cl H3PMo12O40 1 mol% 120 3 h 63 71 89 362
142| Chapter 2
8 [BMIm]Cl H3PW12O40 1 mol% 120 3 h 66 82 81 362
8 [BMIm]Cl HNO3 1 mol% 120 3 h 43 56 77 362
5 [BMIm]Cl Hydroxyapatite with 4.6 wt% Cr(III) 60 wt% Unknowna 2.5 min 40 78 52 365
9 [BMIm]Cl NHC/(CrCl2)2 9 mol% 100 6 h 81 - - 143
9 [BMIm]Cl NHC/CrCl2 9 mol% 100 6 h 81 - - 143
9 [BMIm]Cl NHC/CrCl2 9 mol% 100 6 h 80 - - 143
9 [BMIm]Cl NHC/CrCl3 9 mol% 100 6 h 78 - - 143
9 [BMIm]Cl NHC/CrCl3 9 mol% 100 6 h 78 - - 143
9 [EMIm]BF4 SnCl4 10 mol% 100 3 h 53 97 55 144
9 [BMIm]Cl Yb(OTf)3 10 mol% 140 6 h 24 65 37 363
9 [EMIm]Cl Yb(OTf)3 10 mol% 140 6 h 10 63 16 363
2 [HexylMIm]Cl/H2O ZrO2 40 wt% 200 10 min 53 92 56 360 1:1 w/w
10 DMSO/[BMIm]Cl Lignin-derived solid acid 50 wt% 160 50 68 99 69 350 a: heating by microwave irradiation; b: 2.3 volume equivalents of MIBK as extraction solvent; c: molsieves used for water adsorption
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|143
The dehydration of disaccharides and polysaccharides in ionic liquids Ionic liquids have also been used as solvents and catalysts for polymeric carbohydrates, like cellulose and inulin. Because this is a review on HMF, only work focussed on HMF synthesis from these carbohydrates will be discussed. Research on carbohydrate dissolution and hydrolysis are not covered by this review. Qi et al. performed inulin dehydration in [BMIm]Cl in which 5% inulin was dissolved along with ten equivalents of water relative to the amount of fructose units.366 Amberlyst 15 and a number of Brønsted acidic ionic liquids were tested as catalysts at 80 °C. A 67% HMF yield at full conversion was obtained with 100 wt% Amberlyst 15 after 3 h. With 33 mol%
[HMIm][HSO4], [EMIm][HSO4] and [BMIm][HSO4] HMF yields of around 55% at full conversion were obtained after 3 h. Fructose formation to a maximum of around 40% was observed in all reactions, but the rate of fructose formation and dehydration were different for all catalysts. An increase alkyl chain length of the ionic liquid corresponds with a decrease in acidity and therefore also a decrease in reaction rate. Experiments in which [BMIm]Cl was - replaced entirely by a Brønsted acidic ionic liquid, with [HSO4 ] as counterion, showed high 366 fructose yields in the initial stages of the reaction. For [HMIm][HSO4] a 55% fructose and
10% HMF yield were obtained within 3 min. In [EMIm][HSO4] and [BMIm][HSO4] the fructose yield exceeded 80% in the first 3-5 min of the reaction. The dehydration rate decreased with increasing alkyl chain length and in [HMIm][HSO4], [EMIm][HSO4] and
[BMIm][HSO4] the HMF yields were 60%, 52% and 37% respectively after 3 h. Significant amounts of fructose were still present after 3 h in [EMIm][HSO4] and [BMIm][HSO4]. In the presence of Amberlyst 15 as a catalyst, 60-65% HMF yields were obtained within 10 min at 100 °C. A positive influence of the chloride counterion in the dehydration rate of fructose to HMF was suggested as it was observed that the dehydration proceeded at a faster rate in the [BMIm]Cl based experiments. This is consistent with observations by Binder and Raines with regard to the role of halide ions in fructose dehydration to HMF.119 Qi et al. reported an HMF yield of 82% by applying a two-step procedure in which 5 wt% inulin was first hydrolysed with 20 molar equivalents of water in [BMim][HSO4] at 80 °C for 5min, followed by addition of an equal weight of [BMim]Cl and 100 wt% Amberlyst 15 for dehydration at 100 °C for 60 min.366 They also observed brown colouring of the reaction mixture due to humin formation. Use of inulin/ChoCl (1:1 w/w) melts by Ilgen et al. resulted in significant HMF yields around 55% with Amberlyst 15, FeCl3 and p-toluene sulfonic acid as catalysts at 90 °C for 1
144| Chapter 2
352 h. A sucrose/ChoCl (1:1) melt, catalysed by CrCl2 and CrCl3 yielded 62% and 43% HMF respectively after 1 h at 100 °C. Hu et al. applied catalyst systems that were found to be efficient in fructose and glucose dehydration on polysaccharides.144,367 Choline Chloride was used in combination with oxalic acid (ChoCl/OxAc) and citric acid (ChoCl/CitAc) in inulin dehydration.367 Both organic acids were used as hydrates, providing water for the hydrolysis of the polysaccharides. The ChoCl/OxAc system yielded 56% HMF at complete carbohydrate conversion after 2 h at 80 ºC. The ChoCl/CitAc system yielded 51% HMF after 2 h at about 90% carbohydrate conversion. The HMF yield in ChoCl/oxalic acid was improved to 64% at full conversion by applying a two-phase system with ethyl acetate as an extracting solvent. Inulin dehydration in
[EMIm]BF4 in the presence of 10 mol% SnCl4 yielded 40% HMF at 100% conversion with 17 wt% substrate at 100 °C for 3 h.144 Reactions with sucrose and cellobiose under the same conditions yielded 65% and 57% HMF respectively at full conversion. A 9 wt% reaction mixture with starch yielded 47% HMF after 24 h at 100 °C. Jiang et al. performed hydrolysis of cellulose in [BMim]Cl, catalysed by various Brønsted acidic ionic liquids.368 HMF yields around 15% were obtained after 1 h at 100 °C with sulfonic acid group containing imidazolium based ionic liquids as the catalysts. Zhang and co-workers used the system they had developed for te conversion of glucose for the direct conversion of cellulose into HMF.369,370 Thus, 6 mol% (based on glucose monomers) of a CuCl2-CrCl2 mixture of metal chlorides with a Cu-Cr ratio of 17:83 (n/n) in [EMIm]Cl was used. With a cellulose concentration of 10% at 120 ºC for 8 h an HMF yield of 55% was obtained.369 The catalyst in the ionic liquid could be recycled 12-fold without apparent loss of activity.
Li et al. investigated CrCl3-catalysed cellulose dehydration in [BMIm]Cl under 400 W microwave irradiation.252 From different types of cellulose HMF yields between 53% and
62% were obtained with 9 wt% substrate and 3.6 wt% CrCl3 after 2 min. The authors propose the same mechanism for glucose dehydration as Zhang,134 but additionally propose a role for 252 CrCl3 in cellulose hydrolysis. 2,2’-Bipyridine, a strong coordinating ligand, decreased the HMF yield to approximately 2%. The same system was applied in the direct dehydration of lignocellulosic biomass.371 A 5 wt% concentration of untreated biomass was reacted for 3 min resulting in HMF yields of around 50% with regard to the estimated amount of hexose present in the sample. In addition, furfural, in yields of around 30% with respect to the estimated amount of pentose were found, originating from the hemi-cellulose present in the biomass. A temperature of approximately 200 ºC was reached during 3 min microwave
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|145 irradiation. Model feedstock, consisting of cellulose and xylan, led to comparable yields of 53% HMF and 33% furfural after 2.5 min. Heating a reaction mixture of pine wood in ionic liquid with CrCl3 with an oil bath for 6 min at 200 ºC gave clearly lower yields of 35% HMF and 17% furfural.
Qi et al. published work on CrCl3-catalysed dehydration of di- and polysaccharides in 347 [BMIM]Cl. At a 5 wt% substrate concentration with 20 mol% CrCl3 sucrose yielded 76% HMF after 5 min at 100 °C.347 Reactions with cellobiose and cellulose yielded around 55% HMF at 140 °C after 5 min and 150 °C after 10 min respectively. Wang et al. reported HMF yields of around 60% from cellulose by reacting 2.5 wt% cellulose in [BMIm]Cl containing different ratio’s of CrCl3 and metal chloride (LaCl3, LiCl or LiBr), with a combined 100 mol% loading, at 140 °C for 40 min in the presence of 10 wt% of water (relative to cellulose).372 The best results were generally obtained with a 1:1 mixture of CrCl3 and metal chloride. The total reducing sugar (TRS) yield was typically around 30%. Reducing sugars are sugars that contain an aldehyde group or can form one via isomerisation. With wheat straw similar yields were obtained after 15 min at 160 °C. At 120 °C Hsu et al. reported 21% HMF yield from cellulose in [EMIm]Cl in the absence of catalyst.373 Chun et al. reported work on starch dehydration in 1-octyl-3-methylimidazolium chloride ([OMIm]Cl) in combination with aqueous HCl solution an ethyl acetate in a 4:5:0.8 ratio (w/w/w).374 When 0.5 M HCl was applied on 10 wt% starch in the described [OMIm]Cl-
HCl(aq)-EtOAc mixture, 30 wt% HMF was obtained after 1 h at 120 °C. When 20 wt% CrCl2 was added, the HMF yield increased to 60%. When a specific starch, namely tapioca starch, was used a 73 wt% yield was reported. With 10 wt% sucrose in a 1:1 (w/w) mixture of [OMIm]Cl and 0.5 M HCl a 79 wt% HMF yield was reported after 1 h at 120 °C, but no conversion was reported.375 No conversions and thus no selectivities are reported in this work. An HMF yield of 89% was reported from cellulose by Zhang et al. by reacting 17 wt% cellulose in [EMIM]Cl in the presence of 10 mol% CrCl2 at 120 °C after 6 h. The yield was determined by acetone extraction, followed by column purification and 1H-NMR analysis.376
A combined CrCl2/H-Y zeolite system was used by Tan et al. in which the zeolite catalysed 377 the hydrolysis of cellulose and the CrCl2 catalysed the conversion of glucose to HMF. With 6 wt% of cellulose in [BMIm]Cl in the presence of 100 wt% of zeolite and 393 mol%
CrCl2, 34% isolated HMF was obtained after 6 h at 100 °C. When an NHC, namely 1,3- bis(2,6-diisopropylphenyl)imidazolylidene (Ipr), was applied as ligand for CrCl2, a 37%
146| Chapter 2
HMF yield was reported. An increase in the reaction time to 12 h, with intermittent ether extraction, increased the yield to 48%. A significant number of recent publications also mention the combination of ionic liquid and chromium salts for producing HMF from cellulose378-379-381 and starch.382 Tao et al. reported a high HMF yield of 48% from cellulose in SO3H-functionalised ionic liquids in 383 combination with MnCl2. Tao et al. used manganese salts in 1-(4-sulfonic acid butyl)-3-methylimidazolium hydrogen sulfate ([SABMIm][HSO4]), for the hydrolysis and subsequent dehydration of microcrystalline cellulose in the presence of water and MIBK.384 An IL-water-MIBK mass ratio of 2:1:6.4 was applied, leading to HMF yields of 27-37% at 68-89% conversion depending on the counterion, with consistent selectivity around 40% after 5 h at 150 °C. The yields were defined as weight of product divided by weight of starting material. Significant amounts of furfural and levulinic acid were also observed, with furfural yields of 9-18% and levulinic acid yields of 3-7%. The blank experiment led to 15% HMF yield at 70% conversion. The same experiments performed with CoCl2 yielded 29% HMF at 81% conversion.385 Work on starch-rich acorn biomass by Lee et al. focussed on the application of chromium halide catalysts. Significant HMF formation was observed, but the definition of the yield was not clearly defined.386 Work on glucose-containing di- and polysaccharides by Ståhlberg in [EMIm]Cl, catalysed by 0.5 eq boric acid at 120 °C yielded 66% HMF from sucrose after 8 h, 33% HMF from maltose and 32% HMF from both starch and cellulose after 24 and 8 h respectively.145 These yields are in line with what was obtained form glucose under similar conditions, but the reaction rates are lower. It is surprising to find a higher reaction rate for cellulose than for starch, because hydrolysis of starch is generally considered to be easier than the hydrolysis of cellulose. With GeCl4 as catalyst in [BMIm]Cl, HMF yields of 55%, 41% and 35% were obtained from sucrose, cellobiose and cellulose respectively at almost full conversion by Zhang et al.268 The same conditions were applied as with glucose (See section 2.4.3.2). When considering a 48% HMF yield from glucose under the same conditions, the yields from cellobiose and cellulose are very good. However, a yield of around 75% should be expected from sucrose, because around 90% HMF was obtained from fructose and almost 50% was obtained from glucose. Tables 31-33 provide an overview of the research on ionic liquid-based di- and polysaccharide dehydration to HMF. At relatively low temperatures (80-120 °C) HMF yields
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|147 over 50% have been reported. For a polyfructan like inulin this is not surprising, though for cellulose and lignocellulosic biomass this is a significant breakthrough. Part of this is caused by the fact that some ionic liquids are able to efficiently dissolve cellulose. This is then combined with the known effectiveness of chromium chloride as a catalyst for HMF production from glucose.
2.4.3.3 Conclusion on HMF production in ionic liquids
In general the selectivities and yields of HMF synthesis in ionic liquids from different types of biomass are high compared to the other systems discussed, with the exception of aprotic polar organic solvents like DMSO. Due to their unique dissolution properties, some ionic liquids could dissolve over 10 wt% cellulose (see Table 32), which is notoriously difficult to solubilise. In general, hardly any levulinic acid formation has been reported in ionic liquid systems, even though water was present. Thus the ionic liquids apparently stabilise the formed HMF, preventing its rehydration. Brown colouration of the reaction mixtures was however generally observed, indicating formation of potentially polymeric byproducts. This apparent stabilising effect also has a disadvantage, as it is difficult to separate the HMF from the ionic liquid, requiring large amounts of extracting solvent. Since ionic liquids have essentially no vapour pressure and HMF is heat sensitive, solvent evaporation and HMF distillation are off the table, leaving extraction as the only method for HMF purification. Zhang et al. performed cellulose depolymerisation with [EMIm]Cl of different quality and from different suppliers.387,388 They reported clearly different results for different batches (purities) of the same ionic liquid. Because the ionic liquid is typically used as a solvent, it is used in large quantities compared to the amount of substrate; small impurities could therefore significantly influence its catalytic behaviour. This could, for instance, explain why Ståhlberg et al.363 and Zhao et al.134 published contradictory findings with regard to the effect of the length of the alkyl chains on the imidazolium ionic liquid. Furthermore imidazolium based ionic liquids were recently shown to react with HMF at temperatures over 200 °C.389 In order to apply ionic liquids as reaction media for HMF production form biomass, highly efficient recycling is required because of their high cost price. This will present challenges when using untreated biomass feedstock, because these contain many inorganic (ash) and organic impurities that will have to be removed form the ionic liquid at some point.
148| Chapter 2
Table 31. The dehydration of polysaccharides and disaccharides in ionic liquids, catalysed by ionic liquid
Substrate Substrate Solvent Catalyst Catalyst conc. T Reaction Yield Conv. Sel. Ref. conc.(wt%) (mol%) (°C) Time (h) (%) (%) (%)
Cellulose 8d [EMIm]Cl - - 120 3 21 - - 373
Cellulose 4 [BMIm]Cl [C4H8SO3HMIm]Cl 300 100 1 15 - - 368
Cellulose 4 [BMIm]Cl [C4H8SO3HMIm]HSO4 300 100 1 13 - - 368
b a Inulin 5 [BMIm]Cl [BMIm]HSO4 33 80 3 53 100 53 366
b a Inulin 5 [BMIm]Cl [EMIm]HSO4 33 80 3 55 100 55 366
b a Inulin 5 [BMIm]Cl [HMIm]HSO4 33 80 3 56 100 56 366
b a Inulin 5 [HMIm][HSO4] None - 80 3 59 97 61 366
b a Inulin 5 [EMIm][HSO4] None - 80 3 52 89 58 366
b a Inulin 5 [BMIm][HSO4] None - 80 3 37 71 52 366
Starch 10 [OMIm]Cl/0.5 M HCl/EtOAc None - 120 1 30 - - 374
Sucrose 20 [HMIm]Cl None - 90 0.5 52 100c 50 334 a: conversion was given as the amount of converted fructose units; b: 10 molar equivalents of H2O added relative to the amount of fructose units; c: 3% of obtained glucose was converted; d: 10 molar equivalents of H2O present
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|149
Table 32. The dehydration of polysaccharides in ionic liquids, catalysed by homogeneous catalysts
Substrate Substrate Solvent Catalyst Catalyst Temperature Reaction Yield Conversion Selectivity Ref. concentration concentration(°C) Time (%) (%) (%) (wt%)
a Cellobiose 5 [BMIm]Cl CrCl3 20 mol% 120 10 min 49 - - 347
b Cellobiose 9 [BMIm]Cl CrCl3 6 mol% 120 4 h 50 - - 337
a Cellobiose 5 [BMIm]Cl CrCl3 20 mol% 140 5 min 55 - - 347
Cellobiose 5 [BMIm]Cl GeCl4 10 mol% 120 30 min 41 99 41 268
Cellobiose 17 [EMIm]BF4 SnCl4 10 mol% 100 3 h 57 100 57 144
d d Cellulose 14 [C4H8SO3HMIm]HSO4 Co(NO3)2 6.5 mol% 150 5 h 19 51 37 385
d d Cellulose 14 [C4H8SO3HMIm]HSO4 CoCl2 6.5 mol% 150 5 h 29 81 36 385
d d Cellulose 14 [C4H8SO3HMIm]HSO4 CoSO4 6.5 mol% 150 5 h 24 85 28 385
e Cellulose 17 [EMIM]Cl CrCl2 10 mol% 120 3 h 82 - - 376
e Cellulose 17 [EMIM]Cl CrCl2 10 mol% 120 6 h 89 - - 376
a Cellulose 2.5 [BMIm]Cl (10 wt% CrCl3/LiCl 50 mol%/50 140 40 min 62 - - 372
150| Chapter 2
H2O) mol%
a Cellulose 5 [BMIm]Cl CrCl3 20 mol% 150 10 min 54 - - 347
h a Cellulose 9 [BMIm]Cl CrCl3 3.6 wt% Unknown 2 min 55 - - 252
a Cellulose 5 [BMIm]Cl CrCl3 3.6 wt% Unknown 3 min 62 - - 252
i a Cellulose 9 [BMIm]Cl CrCl3 3.6 wt% Unknown 2 min 55 - - 252
j a Cellulose 9 [BMIm]Cl CrCl3 3.6 wt% Unknown 2 min 53 - - 252
Cellulose 9 [EMIm]Cl CuCl2/PdCl2 18 mol% 120 1 h 18 - - 390 (1:10)
Cellulose 5 [BMIm]Cl GeCl4 10 mol% 120 30 min 35 97 36 268
Cellulose 9 [EMIm]Cl H3BO3 50 mol% 120 8 32 - - 145
Cellulose 6 [BMIm]Cl H-Y 100 wt% 120 6 h 34e - - 377
zeolite/CrCl2 zeolite/393 mol% Cr
Cellulose 6 [BMIm]Cl H-Y 100 wt% 120 6 h 37e - - 377 zeolite/Ipr- zeolite/393 k CrCl2 mol% Cr
Cellulose 6 [BMIm]Cl H-Y 100 wt% 120 12 h 48e - - 377
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|151
zeolite/Ipr- zeolite/393 k CrCl2 mol% Cr
d,e d Cellulose 14 [C4H8SO3HMIm]HSO4 Mn(CH3COO)2 6.5 mol% 150 5 h 32 84 39 384
d,e d Cellulose 14 [C4H8SO3HMIm]HSO4 Mn(H2PO4)2 6.5 mol% 150 5 h 31 79 39 384
d,e d Cellulose 14 [C4H8SO3HMIm]HSO4 Mn(NO3)2 6.5 mol% 150 5 h 27 68 40 384
d,e d Cellulose 14 [C4H8SO3HMIm]HSO4 MnCl2 6.5 mol% 150 5 h 37 89 42 384
d,e d Cellulose 14 [C4H8SO3HMIm]HSO4 MnSO4 6.5 mol% 150 5 h 35 83 42 384
d,e d Cellulose 14 [C4H8SO3HMIm]HSO4 None 6.5 mol% 150 5 h 15 70 21 384,385
a Cellulose 9 [BMIm]Cl CrCl3 3.6 wt% Unknown 2 min 61 - - 252
a Corn Stalk 5 [BMIm]Cl CrCl3 3.6 wt% Unknown 3 min 45 - - 252
Inulin 9 ChoCl Citric Acid 600 mol% 80 2 h 51 100g 56 367
Inulin 9 ChoCl Citric Acid 600 mol% 80 2 h 57 100g 65 367
f Inulin 50 ChoCl FeCl3 10 mol% 90 1 h 55 - - 352
Inulin 5 ChoCl Oxalic Acid 600 mol% 80 2 h 56 100 56 367
Inulin 5 ChoClc Oxalic Acid 600 mol% 80 2 h 64 100 64 367
152| Chapter 2
Inulin 50 ChoClf pTsOH 10 mol% 90 1 h 57 - - 352
Inulin 17 [EMIm]BF4 SnCl4 10 mol% 100 3 h 40 100 40 144
Maltose 9 [EMIm]Cl H3BO3 50 mol% 120 8 33 - - 145
Pine wood 5 [BMIm]Cl CrCl3 3.6 wt% 200 3 min 35 - - 376
a Pine wood 5 [BMIm]Cl CrCl3 3.6 wt% Unknown 3 min 52 - - 376
a Rice Straw 5 [BMIm]Cl CrCl3 3.6 wt% Unknown 3 min 47 - - 376
Starch 10 [OctylMIm]Cl/0.5 M CrCl2 20 wt% 120 1 h 60 - - 374 HCl/EtOAc
Starch 9 [EMIm]Cl H3BO3 50 mol% 120 24 32 - - 145
Starch 9 [EMIm]BF4 SnCl4 10 mol% 100 24 h 47 100 47 144
Sucrose 9 [BMIm]Cl AuCl3·HCl 7 mol% 100 3 h 36 - - 269
f Sucrose 50 ChoCl CrCl2 10 mol% 100 1 h 62 - - 352
Sucrose 10 [OctylMIm]Cl/0.5 M CrCl2 20 wt% 120 1 h 79 - - 375 HCl 1:1 w/w
a Sucrose 5 [BMIm]Cl CrCl3 20 mol% 100 5 min 76 - - 347
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|153
f Sucrose 50 ChoCl CrCl3 10 mol% 100 1 h 43 - - 352
b Sucrose 7 [BMIm]Cl CrCl3 6 mol% 120 4 h 100 100 100 337
Sucrose 5 [BMIm]Cl GeCl4 10 mol% 120 30 min 55 98 56 268
Sucrose 9 [EMIm]Cl H3BO3 50 mol% 120 8 66 - - 145
Sucrose 9 [BMIm]Cl IrCl3 7 mol% 100 3 h 37 - - 269
Sucrose 17 [EMIm]BF4 SnCl4 10 mol% 100 3 h 65 100 65 144
Tapioca 10 [OctylMIm]Cl/0.5 M CrCl2 20 wt% 120 1 h 73 - - 374 Starch HCl/EtOAc
a Wheat 2.5 [BMIm]Cl (10 wt% CrCl3/LiCl 50 mol%/50 160 15 min 61 - - 372 straw H2O) mol% a: heating by microwave irradiation; b: 2.3 volume equivalents of MIBK used as extracting solvent; c: ethyl acetate used as extracting solvent ; d: in wt%; e: significant furfural yield (9-18%); f: melt; g: around 10% fructose left; h: avicel; i: sigmacel; j: N. spruce; k: 1,3-bis(2,6- diisopropylphenyl)imidazolylidene as N-heterocyclic ligand)
154| Chapter 2
Table 33. The dehydration of inulin in ionic liquids, catalysed by Amberlyst 15
Inulin Solvent Amberlyst 15 Temperature Reaction Yield Conversion Selectivity Reference concentration loading (wt%) (°C) Time (%) (%) (%) (wt%)
5 [BMIm]Cla 100 80 3 h 67 100b 67 366
50 ChoClc 10 90 1 h 54 - - 352
5 [EMIm][HSO4]a 100 100 5 min 65 97b 67 366
5 [BMIm][HSO4]a 100 100 10 min 61 97b 63 366
2 [BMIm]Cl/[BMIm][HSO4] 100 100 1 h 82 99b 83 366 1:1 w/wd a: 10 molar equivalents of H2O relative to the amount of fructose units; b: the conversion was given as the amount of converted fructose units; c: melt; d: pretreatment in [BMIm]HSO4 for 5 min at 80 °C
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|155
2.5 Process technology
2.5.1 Introduction
This section focuses on HMF manufacturing methods and covers i) kinetic studies on HMF synthesis, ii) reactor configurations and designs, iii) separation and purification strategies, iv) pilot scale studies and v) an overview of HMF manufacturing costs. The most recent reviews covering this topic were published in 1990 by Kuster12 and a mini review limited to the process technology of HMF synthesis in 2009 by Boisen et al.391 Literature data until 29 July 2011 are included in this chapter. Details on the catalytic synthesis of HMF from lignocellulosic biomass using a wide range of types of catalysts and process conditions are provided in the Process Chemistry section. Some relevant features regarding solvents and catalyst selection, of prime importance for reactor- and process design, will be highlighted below. Here, a distinction based on reaction solvents is applied, i.e. i) aqueous systems, ii) mixed solvent systems and particularly biphasic systems involving water, and iii) non- aqueous systems including the application of ionic liquids. Solvent effects on HMF yields are pronounced, whereas solvent choice also affects product work-up and solvent/catalyst recycle streams. To the best of our knowledge, full scale commercial HMF plants are not operational yet. However, pilot scale studies have been reported and will be discussed.
2.5.1.1 Aqueous reaction systems
A well-known and established synthetic strategy for HMF synthesis is the acid catalysed dehydration of C6-sugars at elevated temperatures in aqueous solutions.7,392,393 Initially, sucrose was used as the feedstock and oxalic acid was employed as the catalyst. Later, a variety of catalytic systems were explored to obtain highly active and selective catalysts for the formation of HMF, mainly with fructose as a preferred sugar feed. Cottier et al.394 divided the catalysts into 5 categories: organic acids, mineral acids, salts, Lewis acids and solid catalysts. Table 34 presents an overview of catalysts used in the dehydration reaction of fructose to HMF in aqueous media based on Cottier’s classification394 with a literature update.
156| Chapter 2
Table 34. Catalysts used for the dehydration of fructose to HMF in aqueous solutions
Organic acids Mineral Salts Lewis acids Solid catalysts acids oxalic acid H3PO4 Ag3PW12O40 AlCl3 ion exchange resins levulinic acid H2SO4 ZnCl2 zeolites maleic acid HCl BF3 metal oxidesa: - formic acid HI B(OH)3 TiP, C-ZrP2O7, acetic acid HBr lanthanide salts ZrP, ZrO2, TiO2, eHTiNbO -MgO, p-toluenesulfonic 5 NbOPO , acid 4 trifluoracetic acid a TiP = titanium phosphate, C-ZrP2O7 = cubic zirconium pyrophosphate, ZrP = Zr(HPO4)2, eHTiNbO5 = exfoliated HTiNbO5, obtained by treating the H-compound with Bu4NOH.
Despite the advantage of using water as an environmentally benign solvent, the chemistry in aqueous systems suffers from relatively low HMF yields due to subsequent reactions to levulinic acid and insoluble polymeric substances (humins). Typical maximum yield of HMF for synthesis of HMF in water is around 50-60 mol% at fructose conversion level in the region of 50-95 % (Tables 7-9 in section 2.4.1.1). In addition, the formation of insoluble humins during the reaction also complicates work-up of the reaction mixture to obtain a high HMF purity.
2.5.1.2 Non-aqueous reaction systems
HMF synthesis in non-aqueous solvents can be categorised based on the boiling points of the solvents, viz. low boiling solvents (bp < 150 °C) and high boiling solvents with ionic liquids as the extreme. The use of low boiling solvents such as acetonitrile,212 ethyl acetate,212 butyl acetate,212 acetone,206 methanol,233,244 ethanol,244 1,2-dimethoxyethane (glyme)231 and acetic acid206,395 have been examined in detail. In acetone at sub- and supercritical conditions using sulfuric acid as the catalysts, HMF was obtained in high selectivities (up to 75%) at 95% fructose conversion. In the case of low chain alcohols and acetic acid, subsequent
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|157 reactions of the in situ formed HMF were observed leading to 5-methoxymethylfurfural (MMF), 5-ethoxymethylfurfural (EMF) and 5-acetomethylfurfural (AMF) as the main reaction product for methanol, ethanol and acetic acid as the solvents, respectively. At fructose conversions exceeding 98%, MMF can be obtained in selectivities up to 78%.233 Meanwhile, higher selectivities for HMF and EMF (up to 90%) can be attained at lower (25%) fructose conversions.244 These reactions are typical examples of in situ conversions of HMF and will be discussed in more detail in the last part of this section. Various high boiling solvents such as dimethyl sulfoxide (DMSO),117,209,212,213,216,3961- methyl-2-pyrrolidinone (NMP),230 N,N-dimethylacetamide (DMA),230,397 N,N- dimethylformamide (DMF)212,224,230,263,397 and sulfolane212,229 have been investigated for the synthesis of HMF using fructose as the starting material. High HMF yields (up to 92%)213 were observed in DMSO in combination with various catalytic systems ranging from Lewis acids213,216 to ion exchange resins209 and even in the absence of catalysts.117 Though advantageous regarding HMF yields, the use of high boiling solvents complicates HMF work-up (vide infra). Recently a number of studies have been carried out on the synthesis of HMF from cheaper feedstocks, such as glucose, cellulose and even lignocellulosic biomass in high boiling solvents, particularly polar aprotic solvents such as DMF, DMA and DMSO has been reported. Takagi et al. examined a combination of base and acid solid catalyst for the conversion of glucose to HMF in DMF.263 A combination of hydrotalcite and amberlyst-15 in DMF gave HMF from glucose with an HMF selectivity of 50% at 73% glucose conversion after 9 h at 80 °C. Binder and Raines reported that DMA containing 10 wt% LiCl is a good solvent to produce HMF in a single step with good yields from lignocellulosic biomass, cellulose, glucose and fructose.119 HMF yields of 48% can be obtained from untreated corn stover using 10 mol% CrCl3 and 10 mol% HCl as the catalyst in DMA/LiCl in the presence of 60 wt% [EMIM]Cl at 140 °C for 2 h reaction time. Using DMSO as the solvent, the conversion of glucose to HMF in yields up to 60 mol% were reported in the presence of chromium and aluminium salts as the catalysts.217,398 Recently, the use of ionic liquid as solvent/catalyst for the synthesis of HMF has been explored. A description of the various ionic liquids used for the synthesis of HMF is provided in section 2.4.3.2. In general, promising results have been obtained for the conversion of fructose or fructose based polymers (like inulin) to HMF. In addition, a breakthrough for the use of glucose and cellulose as a feedstock was reported recently.134,135,252,399 HMF yields of 62% were obtained when converting cellulose and typical examples of lignocellulosic
158| Chapter 2
372 biomass (e.g. straw) in [BMIm]Cl using CrCl3/LiCl as the catalysts. However, the development of an effective HMF isolation method combined with efficient ionic liquid recycle remains a challenge.
2.5.1.3 Mixed solvent reaction systems
When using mixed solvents for the synthesis of HMF, biphasic or single phase systems should be distinguished. For biphasic liquid-liquid systems, it is the intention to continuously extract the in situ formed HMF from the reaction phase to the other liquid phase during the reaction to prevent subsequent reactions of HMF, like hydration to levulinic acid and the formation of humins. In the 1950’s, Peniston introduced the use of 1-butanol in combination with water for sucrose/sulfuric acid systems.288 In this biphasic system, solid humin formation was not observed and the HMF was extracted to the 1-butanol phase. In the given example, a 0.58 M sucrose solution containing 0.1 N H2SO4 as the catalyst was contacted two times with fresh 1- butanol at 150 °C for 20 min for each stage. The reactions were carried out at a phase ratio of one for the aqueous and organic phase. After reaction, the yield of HMF was 72.5 % HMF at 86 % sucrose conversion. Most of the HMF was present in the 1-butanol phase (68.6 %) and minor amounts in (3.9 %) in the aqueous phase. This patent also disclosed that the presence of a sugar affects the partition coefficient (P) of HMF between the organic and aqueous phase (1-butanol/water: P= 1.57 and 1-butanol/20% sucrose solution: P= 1.79). In 1959, Cope was the first to apply methyl isobutyl ketone (MIBK) as an effective solvent for the in situ removal of HMF from an aqueous reaction phase.311 When using a continuous solvent extraction, the yield of HMF improved considerably from 20% to 63%. Two decades later, the application of biphasic systems consisting of water and MIBK as the organic solvent was revisited for the conversion of fructose to HMF using homogenous mineral acid199 and ion exchange resins.211,289,290,400 In the 1990’s, the use of heterogeneous catalysts (like zeolites) as catalysts for biphasic systems was explored.102 The reactions were performed either in batch211,289,290,400 or continuous mode,102,199 see the Process Chemistry section for more details. Because the partitioning of HMF is only slightly favourable at best 170 towards the organic layer, with [HMF]org/[HMF]aq between 0.9 and 1.9, it is advantageous to perform the reaction in a solvent mixture with a high MIBK to water ratio (5:1 or even 9:1) in order to obtain high HMF yields. In 2006, Dumesic and co-workers reported studies to enhance the partitioning of HMF between the reactive aqueous solution and the MIBK phase by applying 2-butanol as a co-
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|159 solvent.170,300 To suppress undesired side reactions in the aqueous phase, dimethyl sulfoxide (DMSO) and/or poly(1-vinyl-2-pyrrolidinone) were applied as co-solvents. This approach allowed processing at high fructose concentrations (10 to 50%) and yielded HMF with a selectivity of 80% at 90% fructose conversion. Further improvements for the biphasic system were achieved by adding NaCl to the aqueous reaction phase.297,401 Recently, Dignan and Sanborn reported on a biphasic aqueous-alcohol (higher alcohol such as fusel oil or pentanol) system in a flow reactor at 240 -270 °C (10–83 bar).402 At least 60% fructose conversion was observed to mainly HMF without substantial solids formation. Apart from biphasic systems, a number of single phase mixed solvents system for the synthesis of HMF from fructose have been reported. Among them are water/dioxane,403 water/triethylene glycol,403 water/PEG-600,156 water/acetone,206,207 acetone/DMSO221 and DMSO/MIBK292 systems. All studies highlight that further degradation of HMF can be suppressed/prevented by using mixed solvent systems, resulting in HMF yield improvements and less humin formation.
2.5.2 Kinetic studies on HMF formation
A number of kinetic studies have been performed on the kinetics of HMF formation from various feedstocks. Kinetic studies are not only of relevance to unravel the mechanism for HMF formation on a molecular level, but also for process development studies to identify optimum reactor configurations and process conditions for the highest HMF yields. In this paragraph an overview of kinetic studies will be provided. It starts with studies using fructose as the feedstock (2.5.2.1, Table 35), with a distinction on studies in aqueous -, non-aqueous- and two solvent systems. Subsequently, kinetic studies using glucose will be provided (2.5.2.2, Table 36), followed by studies using cellulose and more complex lignocellulosic biomass (2.5.2.3, Table 37). Finally, kinetic studies on the decomposition of HMF to levulinic acid and formic acid will be reported and discussed (2.5.2.4, Table 38). The rate of the latter reaction should be suppressed as much as possible to avoid a reduction in the HMF yields.
2.5.2.1 Kinetic studies on the formation of HMF from fructose
As noticed in many studies, fructose is the preferred C6-sugar for HMF formation in water. Compared to glucose, fructose is more reactive and also leads to higher HMF yields. An overview of kinetic studies reported in the literature using fructose as the starting material is given in Table 35.
160| Chapter 2
Single solvent, aqueous systems, homogenous catalysts Kinetic studies on the conversion of fructose to HMF in aqueous systems using HCl as the catalyst were reported by Kuster et al.155 and Asghari et al.404 Kuster and van der Baan155 investigated the reaction of fructose (0.2–1 M) to HMF at 95 °C in a batch reactor. The experimental data were best described using an order of one in fructose. A kinetic model including the formation of humins and the subsequent decomposition of HMF to levulinic acid (LA) and formic acid (FA) was proposed involving unknown intermediates X and Y (Table 35). Asghari et al. reported the conversion of fructose in subcritical water (210–270 °C, 40–150 bar) at fixed initial pH of 1.8.404 The reactions were carried out in a continuous flow system with residence times between 0.5–300 s. The initial concentration of fructose was 0.03 M. Main products were HMF, 2-furaldehyde, LA, FA and soluble polymers. A reaction network was proposed (Table 35), involving three parallel reactions for fructose: to 2-furfuraldehyde, to soluble polymers and to HMF. The reaction to HMF was modelled assuming a first order reaction in fructose and the activation energy was reported to be 160.6 kJ/mol. Li et al. reported a kinetic study on the reaction of fructose to HMF using formic or acetic acid as the catalysts in subcritical water (180–220 °C, 100 bar).123 The reactions were performed in a 500 ml batch reactor at a stirring rate of 300 rpm. The authors also examined the thermal decomposition of fructose under these conditions. Experimental data were modelled using a first order approach. At a temperature of 200 °C, the rate constants for fructose are 0.08, 0.12, and 0.32 min-1 for the thermal reaction and the catalytic reaction using acetic acid (10.8 mg/ml) and formic acid (10.8 mg/ml), respectively. Even though the addition of organic acids enhanced the conversion rate of fructose, the maximum yield of HMF (52 %) was similar for all cases. The apparent activation energies of fructose decomposition were 126.83.3 kJ/mol, 125.63.8 kJ/mol, and 112.013 kJ/mol for the thermal decomposition, and reactions catalysed by acetic- and formic acid, respectively. A kinetic study on the decomposition of fructose in aqueous solutions using sulfuric acid as the catalyst at a wide range of sulfuric acid concentrations (5 mM – 1 M) was reported by Abdilla et al.405 The reactions were carried out in a batch mode using glass ampoule reactors at temperatures ranging between 140–180 °C with fructose loadings between 0.1–1 M. A kinetic model using a power law approach was established for a reaction network involving formation of side products. The order of reaction for fructose was found to be close to 1 for the main reaction and 1.18 for the side reaction. The activation energy for the side reaction
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|161
(14712 kJ/mol) was higher than the activation energy of the main reaction (1235 kJ/mol). This indicated that when using sulfuric acid as the catalyst in aqueous media, lower temperatures are preferred to avoid humin formation and thus to increase the production of HMF. The microwave assisted non-catalysed reaction of fructose to HMF in water at subcritical conditions (180-250°C) has been reported.406 5-HMF was found to be the major product, with a maximum yield of 47%. The kinetic constant for fructose was 3.6 × 10-3 s-1 at 220 °C, assuming a first order reaction in fructose.
Single solvent, aqueous systems, heterogeneous catalysts
Carniti et al. reported the application of niobic acid (Nb2O5) and niobium phosphate 200 (NbOPO4) for the dehydration of fructose in water. The reactions were performed in a continuous packed bed reactor in a small temperature range (90–110 °C) and a fixed initial fructose concentration of 0.3 M. The results were modelled assuming i) first order reaction in fructose and ii) only kinetic limitations and no mass transfer issues. The activation energy was found to be 65.8 ± 8 kJ/mol when niobium phosphate was used as the catalyst.
Single solvent, non-aqueous systems, homogenous catalysts Bicker et al. reported kinetic studies on the dehydration of fructose in supercritical methanol and supercritical acetic acid.233 Reactions were carried out in continuous mode in a high pressure reactor using sulfuric acid (10 mM) as the catalyst and a fixed fructose inlet concentration of 0.06 M. A broad temperature range was applied for methanol (120–270 °C), whereas the reactions in supercritical acetic acid were performed at a fixed temperature of 180 °C. A kinetic network was proposed involving fructose conversion to the desired HMF as well as soluble polymeric by-products. The intermediate HMF is not inert under these conditions. In supercritical methanol a furfural ether (5-methoxymethylfurfural) is formed whereas an ester (5-acetoxymethylfurfural) was produced in supercritical acetic acid. All reactions were assumed to be irreversible and first order in substrate. The activation energy for the decomposition of fructose was 80 kJ/mol in supercritical methanol. Moreau et al. studied the conversion of fructose to HMF using an ionic liquid, 1-H-3- methylimidazoliumchloride ([HMIM]Cl).334 Here, the ionic liquid acts both as a solvent and catalyst. The reactions were carried out on small scale in magnetically stirred batch reactors within a temperature window of 90–120 °C. The initial fructose concentration ranged from 0.01–2.5 M. The data were modelled assuming a simple kinetic scheme involving the main
162| Chapter 2 reaction of fructose to HMF and the subsequent decomposition of HMF to products. A first order approach was applied; though the initial rate versus intake of fructose profiles indicate that this approach is valid only at low fructose intakes (< 0.5 mol/l). The activation energy for the reaction of fructose to HMF was 143 kJ/mol. Recently, Wei et al. reported the conversion of fructose to HMF in 1-butyl-3-methyl 269 imidazolium chloride ([BMIm]Cl) using IrCl3 as the catalyst. The reactions were performed in 25 ml stirred flasks in a temperature range of 80–120 °C and a fixed fructose concentration (200 mg fructose in 2 g [BMIM]Cl). A kinetic network was proposed involving fructose conversion to HMF and byproducts. The experimental data were modelled using first order reaction kinetics. The activation energies for fructose decomposition were estimated to be 165 kJ/mol and 124 kJ/mol for the formation of HMF and formation of byproducts, respectively. Recently, Caratzoulas et al. reported the use of hybrid quantum mechanics/molecular mechanics dynamics free energy calculations to investigate the reaction mechanism for the conversion of fructose to HMF in acidic water.407 Remarkably good agreement was observed between the calculated and experimentally determined concentration profiles from kinetic studies of Asghari et al.404
Single solvent, non-aqueous systems, heterogeneous catalysts Nakamura and Morikawa reported the conversion of fructose to HMF in DMSO using acidic ion-exchange resins.209 The reactions were carried out in glass stirred batch reactors at a constant temperature of 80 °C. Various ion exchange resins were examined as the catalysts, ranging from porous types such as Diaion PK-208, Diaion PK-216, and Diaion PK-228 to gel type catalysts such as Amberlite IR-118, Amberlite IR-120, and Lewatit SC-108. The formation of HMF was described by a first order reaction model. The rate constants for the reactions catalysed by the various resins were determined. Data analyses showed that the use of the porous resins resulted in higher reaction rates than the use of gel type resins.
Mixed solvents, homogeneous catalysts A kinetic study on the synthesis of HMF from fructose in a biphasic system consisting of water and MIBK using phosphoric acid as the catalyst was reported by Kuster et al.199,209 The experiments were performed in a continuous stirred tank reactor at temperatures between 170 and 220 oC. The data were modelled using a kinetic scheme involving the reaction of HMF to
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|163 an intermediate X, which either reacts to HMF or to humins. In a subsequent step HMF is converted to other products, with the major being LA and FA. The reaction of fructose to HMF was found to be first order in HMF and phosphoric acid concentration, with an activation energy of 93 ± 6 kJ/mol. Bicker et al. studied the use of acetone-water (90:10) mixtures for the conversion of fructose to HMF at subcritical conditions (180–300 °C, 200 bar) in a continuously operated tubular reactor with sulfuric acid as the catalyst (0–5 mM). 206 The reaction was modelled using a first order approach without taking any other reactions into account, indicating an activation energy of 98 kJ/mol in the presence of sulfuric acid. The thermal reaction was about twenty times slower than the catalysed reaction (300 °C) and an activation energy of 158 kJ/mol was calculated.
Mixed solvents, heterogeneous catalysts Moreau et al. studied the dehydration of fructose to HMF, catalysed by H-mordenite, in a biphasic water-MIBK batch system at three temperatures (150, 165 and 180 °C).102 Two reactions were considered in the kinetic analyses: the main reaction of fructose to HMF and the subsequent rehydration of HMF to LA and FA. When assuming first order kinetics for both reactions, the activation energy was 141 kJ/mol for the dehydration and 64 kJ/mol for the rehydration. Experiments with different rotor speeds and estimations of relevant mass transfer criteria revealed that the reactions were carried out in the kinetic regime and therefore were not biased by mass transfer effects. Qi et al. reported the conversion of fructose to HMF in water-acetone mixtures in the presence of a strongly acidic resin (DOWEX 50WX8-100).207 Reactions were performed in a batch reactor using microwave heating in the absence of stirring. The reactions were carried out in a temperature range of 100–180 °C with an initial fructose concentration of 2 wt%. The fructose concentration versus time profiles for 70:30 (w/w) acetone/water mixtures were modelled assuming a first order reaction in fructose. Without considering other reactions this led to an activation energy of 103.4 kJ/mol. The same group also performed comparable studies on the reaction of fructose to HMF in acetone-DMSO.221 The same catalyst was tested with similar temperature ranges and initial fructose concentrations as applied for the acetone-water mixtures. The activation energy for the reaction of fructose to HMF in a mixture of 70:30 (w/w) acetone/DMSO was considerably lower (60 kJ/mol) than in acetone/water (103.4 kJ/mol).
164| Chapter 2
Discussion The experimental data of the conversion of fructose to HMF and byproducts (humins, levulinic acid, formic acid, furfural) using various solvents and catalysts at a wide range of experimental conditions have been modelled using numerous kinetic schemes of which some involve up to more than 7 individual reactions. For the reaction of fructose to HMF, a first order approach in fructose is usually applied. Figure 2 presents an overview of the activation energies for the conversion of fructose to HMF and humins. The activation energy for the hydrothermal decomposition of fructose in water was determined in two separate studies and reported to be about 120 kJ/mol.123,408 In aqueous systems, the lowest activation energy (66 kJ/mol) was reported for the dehydration of fructose using a solid catalyst (niobium phosphate) at relatively low temperatures (90–120 °C).200 The highest activation energy (160 kJ/mol) was found for the conversion of fructose at high temperatures (210–270 °C) using HCl as the catalyst at a fixed pH of 1.8.404 The reported activation energies for the conversion of fructose to HMF in non-aqueous systems range from 60 to 165 kJ/mol. The lowest activation energy of 60.4 kJ/mol was reported for fructose conversion in acetone/DMSO (70/30) using an ion exchange resin as the catalyst.221 The authors suggest that a higher concentration of the furanoid forms of fructose in acetone/DMSO leads to a lower activation energy.221 As was discussed in 2.3.1.1, Amarasekara proposed a reaction mechanism in which DMSO acts as a catalyst, which could also explain the low activation energy (Scheme 12).117 In some of the kinetic studies both the activation energies for the main reaction of fructose to HMF, and the side reaction to humins were determined and compared in order to gain insights in the temperature effect on HMF selectivity. Different trends were observed. Asghari et al.reported a lower activation energy for the side reaction (102 kJ/mol) than for the dehydration reaction (160 kJ/mol) in an aqueous system, catalysed by HCl at a fixed pH of 1 using a flow reactor at 210–270 °C.404 In an ionic liquid system at relatively low temperatures of 80–100 °C Wei et al. reported an activation energy of 165 kJ/mol for the dehydration of fructose to HMF; 269 considerably higher than the activation energy for the side reactions (124 kJ/mol). This implies that for this system higher temperatures favour HMF selectivity. For the conversion of fructose in water with sulfuric acid as the catalyst, Abdilla et al. found the opposite, namely an activation energy of 123 kJ/mol for the dehydration and 147 kJ/mol for the side reaction,405 which indicates that operation at lower temperatures is favoured for high HMF selectivity.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|165
Apparently, different catalysts and/or solvents behave differently, making it impossible to draw general conclusions.
250
225 FRC HMF FRC HUM
200
175
150
125
100
75
activation energy (kJ/mol) energy activation 50
25
0
Li,2009 Li,2009Li,2009 Qi,2008b Qi,2008a Wei,2011 Wei,2011 Khajavi,2005Carniti,2006 Abdilla,2011Asghari,2007Kuster,1977Bicker,2003Moreau,2006Moreau,1996Bicker,2003Moreau,2001 Asghari,2007Abdilla,2011
Figure 2. Activation energies for the conversion of fructose (FRC: fructose; HUM: humins; white bars: thermal aqueous systems; light grey bars: catalytic aqueous systems; dark grey bars: non-catalytic aqueous systems including mixed solvents and biphasic systems)
2.5.2.2 Kinetic studies on the formation HMF from glucose
A number of kinetic studies on the dehydration of glucose to HMF have been reported, despite the known lower selectivity to HMF compared to fructose. This attention could be explained by the fact that HMF is an undesired side product when targeting for high glucose concentrations in biomass pre-treatment studies for subsequent fermentation processes. All studies reported so far have been conducted in water, except a recent study by Qi et al., which was performed in an ionic liquid.347 An overview of all kinetic data is provided in Table 36.
Conversion of glucose to HMF in aqueous systems In 1945 Saeman developed a kinetic model for acid-catalysed hydrolysis of glucose in water using sulfuric acid (0.4-1.6 wt%).409 This kinetic study was part of a larger study on the acid catalysed decomposition of cellulose and woody biomass (Douglas-fir). The experiments with glucose were carried out in batch setups (sealed glass bombs) at 170–190 °C and an initial glucose concentration of 0.278 M. The overall rate of glucose decomposition was determined, without considering the individual reaction rates to other products. The experimental data were successfully modelled with a first order approach in glucose. The
166| Chapter 2 order in acid was found to be slightly higher than 1 (1.02), with an activation energy of 137.5 kJ/mol. Heimlich et al. investigated the conversion of glucose in aqueous solutions in the presence of 0.35 M HCl at 100–150 °C.410 The reaction was modelled using a first order approach, leading to an activation energy of 133 kJ/mol. McKibbins et al. studied the kinetics of the reaction of glucose to HMF as well as the subsequent reaction of HMF to LA and FA in water using sulfuric acid as the catalyst (0.025– 0.8 N).411 The experiments were carried out in a batch reactor for a wide temperature range of 140–250 °C. The rates for glucose conversion to HMF were modelled using a first order approach in both glucose and sulfuric acid. An activation energy of 136.8 kJ/mol was reported for the dehydration of glucose to HMF. Smith et al. investigated the decomposition of glucose in water using sulfuric acid as the catalyst in a batch reactor (ampoules, 180-224 °C, 0.5–7 min reaction time) and a continuous flow reactor (178–241 °C, 0.06–0.27 min).412 In the batch reactor, the kinetic data from McKibbins et al.411 were reproduced, though the activation energy was slightly lower (128.6 kJ/mol). However, different results were obtained using the flow reactor. Higher initial reaction rates for the decomposition of glucose were obtained in the continuous setup compared to the batch reactor. Moreover, the experimentally determined activation energy was considerably lower (87.8 kJ/mol) for the continuous reactor than for batch (128.6 kJ/mol). It was concluded that the original batch models are not valid for short residence/reaction times (< 20 s). Isomerisation to fructose and reversible oligomerisation (reversion) reactions, particularly for higher loading of intake glucose, were assumed to be responsible for these differences and should be considered in kinetic models for glucose decomposition (Scheme 33).
Scheme 33.
Recently, Pilath et al. developed a kinetic model for glucose reversion in aqueous acidic solution.413 The experiments were conducted in a robotic microwave-heated batch reactor
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|167
o system in the presence of 1.2 wt% H2SO4 between 120 and 180 C. Reversion product yields of up to 12 wt% of the glucose feed were observed consisting mainly of di-saccharides. Levoglucosan, an anhydrosugar, was also formed, but larger oligosaccharides were not observed. These reversion products were in equilibrium with glucose, which was reached at t < 10 minutes at temperatures higher than 140 oC. These reversion reactions compete with the dehydration of glucose to form HMF and actually reduce the glucose concentration in solution. The kinetic parameters for the conversion of glucose to ten identified di-saccharides were best modelled using a second-order dependency in glucose.. The rate expression for the reaction of glucose to HMF was determined using a first order model and an activation energy of 133 kJ was obtained. Bienkowiski et al. studied the decomposition of glucose (4-12 wt%) in water, catalysed by 4-20 wt% sulfuric acid.414 The experiments were performed in sealed glass reactors at 100- 144 °C. The experimental data were analysed using five different kinetic models. The best results were obtained using a first order approach in glucose in combination with a more complex term, including the fugacity of H+, to describe the effect of acid concentration on the reaction rate. The activation energy was 130.4 kJ/mol, in agreement with the value from Saeman.409 Baugh et al. investigated the decomposition of glucose in water at initial glucose concentrations of 0.006–0.33 M.415 The reactions were carried out in batch, catalysed by sulfuric acid at 170–230 °C. The reactions were modelled using a pseudo-first order approach in glucose and a kinetic rate coefficient including three individual terms (no catalysis, acid and base catalysis). An activation energy of 121 kJ/mol was determined, which was slightly lower than determined by Saeman (137.5 kJ/mol). Xiang et al. investigated the kinetics of glucose decomposition in water at pH 1.5-2.2 (sulfuric acid) at 180–230 °C using glass ampoule reactors.416 The experimental data for glucose decomposition were modelled using a pseudo first order approach. Deviations of this model were observed at short reaction times and high glucose concentrations. Therefore a reaction network involving the rapid and reversible formation of reversion products was applied to improve the modelling results. The temperature dependence of the equilibrium constant for this reaction was also determined. The activation energy for the irreversible decomposition reaction of glucose was 139 kJ/mol, close to the value found by Saeman. Kabyemela et al. investigated the kinetics of glucose isomerisation and decomposition in flow in sub- and supercritical water at 300-400 °C, 25-40 MPa for 0.02-2 s.122 The experimental data were modelled using a reaction network involving isomerisation of glucose
168| Chapter 2 to fructose and the decomposition of both glucose and fructose to various products. All reactions were assumed to be irreversible and first order in substrate. The rate constants of both the isomerisation of glucose to fructose and the decomposition of glucose and fructose were independent of the pressure at subcritical conditions, whereas the rate decreased with increasing pressure at supercritical conditions. The activation energy for glucose decomposition was estimated to be 96 kJ/mol. Kabyemela et al. proposed reaction pathways for the decomposition of glucose and fructose in supercritical water in the absence of catalysts.121 A reaction network was proposed involving (i) isomerisation of glucose and fructose, (ii) dehydration to HMF and (iii) subsequent retro-aldol condensation to C4- and C3-sugars and their fragmented products (glyceraldehyde and pyruvaldehyde). Kinetic constants for reactions to C3-C4 sugars were determined, but the authors did not report the reaction rates for the formation and decomposition of HMF. Khajavi et al. performed kinetic studies on the thermal degradation of various monosaccharides (glucose, galactose, mannose, fructose, and sorbose) in subcritical water.408 The reactions were performed in a tubular reactor at 180–260 °C. Activation energies of 155 kJ/mol and 120 kJ/mol were estimated for the thermal decomposition of glucose and fructose, respectively. Fructose gave the highest HMF yields of all the monosaccharides that were tested (up to 50%). Matsumura reported hydrothermal decomposition studies of glucose in sub- and supercritical water (573-733 K, 25 MPa) in a continuous set-up.417 The product yields were determined and the highest HMF yield (25%) was obtained at 573 K and residence times of 50-70 s. The reactions were modelled using a very complex reaction scheme involving 12 individual reactions and assuming first order in substrates. The activation energy of the overall decomposition rate (consisting of 4 individual reactions) of glucose was 95.5 kJ/mol. The microwave assisted non-catalysed reaction of glucose to HMF in water at subcritical conditions (180-250°C) has been reported.406 5-HMF was found to be the major product, with a maximum yield of 30%. The kinetic constant for glucose was 5.2 × 10-4 s-1 at 220 °C, assuming a first order reaction in glucose. The reaction was about 8 times slower than observed for fructose. As part of a larger study on kinetic modelling of the acid catalysed decomposition of lignocellulosic biomass, Girisuta et al. performed a kinetic study on the acid-catalysed decomposition of glucose in water.165 In glass ampoules the effects of initial glucose concentration (0.1–1 M), sulfuric acid concentration (0.05–1M) and temperature (140–200
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|169
°C) were quantified. A reaction network was developed involving two parallel reactions, namely the dehydration of glucose to HMF and the decomposition of glucose to humins. The data were modelled using a power law approach, see Table 36 for details. The activation energy for the decomposition of glucose to HMF was determined at 1521 kJ/mol for the main reaction and 1651 kJ/mol for the side reaction. Chang et al. reported a kinetic study on the formation of levulinic acid from glucose using 1-3 wt% sulfuric acid as the catalyst at 170-190 °C.418 The reactions were carried out in a stainless steel batch reactor (125 ml). A reaction network comprising of two parallel reactions was proposed and the data were modelled using a first order approach. The activation energy was estimated at 86.3 kJ/mol for the dehydration to HMF and 56.9 kJ/mol for the reaction of glucose to solid humins. The activation energy for the lumped decomposition of HMF to levulinic acid and humins was 209 kJ/mol. In a subsequent paper, Chang and co-workers reported a kinetic study on the acid catalysed hydrolysis/dehydration of wheat straw to levulinic acid at 190-230 °C in batch using 1-5 wt% sulfuric acid as the catalyst.419 The kinetics of the reactions to the main intermediates glucose and HMF were established. An activation energy of 54.5 kJ/mol was reported for glucose dehydration to HMF . Though a side reaction of glucose to humins was proposed in their reaction network, the activation energy of this side reaction was not reported. Meanwhile, an activation energy of 56.5 kJ/mol was reported for the decomposition of HMF to LA , which was considerably lower than the value in their first paper (209 kJ/mol). Kupiainen et al. reported a kinetic study on the decomposition of glucose in 5-20 wt% aqueous formic acid solutions at 180–220°C.420 The experiments were carried out in zirconium batch reactors heated in a fluidised sand bath. The initial concentration of glucose was either 56 mM or 112 mM. The data were modelled using first order reactions. A reaction scheme involving an unknown intermediate in the conversion of glucose to HMF was used to model the kinetic profiles. The activation energy for glucose decomposition to the intermediate was found to be 1532 kJ/mol. Activation energies for the decomposition of the intermediate to HMF and humins were found to be 1174 kJ/mol and 1105 kJ/mol, respectively. Duru et al. recently reported the conversion of glucose to HMF in water using metal chlorides (CrCl2, CoCl2, CuCl2, SnCl2 and FeCl3) at 373 K.421 Best yields were obtained using CrCl2 and CuCl2, though yield data were not reported. Remarkably, the yields for
170| Chapter 2
CrCl2 were highest in alkali media (pH> 9). The reaction was modelled using a first order approach, giving a k value of 6.5 × 10-6 s-1 for CrCl2 at 373 K. Wu et al. performed an in depth kinetic study on the decomposition of dilute aqueous glucose solutions (5.6 × 10-8 – 5.6 × 10-3 mol/l) in the absence of catalysts in hot compressed water (175-275°C).422 The initial glucose concentration has a major effect on the selectivity of the reaction, ascribed to a change in reaction mechanism. At initial glucose concentrations exceeding 5.6 × 10-5 mol/l the selectivity to HMF increases significantly. The apparent activation energy for glucose decomposition is a function of the initial glucose concentration and decreases with decreasing glucose concentration (109 +/- 5 kJ/mol at high to 90+/- 4 at kJ/mol at the lowest initial glucose concentration).
Conversion of glucose to HMF in non-aqueous solvents Qi et al. reported kinetic research on the conversion of glucose to HMF in the ionic liquid
1-butyl-3-methyl imidazolium chloride ([BMIm]Cl) using CrCl3 as catalyst and applying microwave irradiation as a heating source.347 A first order reaction approach was used to model the experimental data. An activation energy of 114.6 kJ/mol was reported for the dehydration of glucose to HMF. Recently, thermodynamic insights in the reactions of glucose to HMF and anhydroglucose in DMSO/water mixtures were reported using a G4 ab initio method in combination with the COSMO-SAC solvation model.423 The predicted ΔG values were in quantitative agreement with experimental values. The calculations showed that the conversion of glucose to HMF is irreversible whereas glucose dehydarion to anhydroglucose is reversible, with equilibrium constants depending on reaction conditions and solvent choice.
Discussion Kinetic studies on the conversion of glucose in aqueous systems are mostly available as part of a reaction scheme on the conversion of cellulosic materials in order to obtain either fermentable sugar or levulinic acid. In the latter reaction, HMF is proposed as the intermediate reaction product, and the formation rate of HMF from glucose is quantified. For the development of kinetic models for producing fermentable sugars, the rate of decomposition of glucose was evaluated as a single, lumped, reaction without considering the individual reactions to various products. Parallel reactions were proposed to incorporate the formation of HMF and humins for the most recent studies on the decomposition of glucose. Both reactions in the presence and absence of catalysts have been studied. The non-catalytic,
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|171 hydrothermal decomposition of glucose was studied at elevated temperatures and pressures in sub-and supercritical water (e.g. 180–260 oC, 100 bar).122,249,408 The experimental data for glucose decomposition using mineral acids as the catalysts were successfully modelled using either a first order reaction in glucose or a power law approach. The acid concentration is explicitly added to the rate laws using various approaches, e.g. a power law factor, activity factor, fugacity, pH or concentration. Figure 3 presents an overview of the reported activation energies for the decomposition of glucose, including the activation energies obtained from studies using cellulose or cellobiose as starting material and modelled with glucose as an intermediate. For the conversion of glucose to decomposition products, most of studies report activation energies ranging from 130 to 140 kJ/mol; close to the activation energy reported in the first study on this topic by Saeman (137 kJ/mol).409 Smith et al. found a considerably lower activation energy of 82 kJ/mol for acid catalysed glucose decomposition in a flow system compared to experiments at similar conditions in a batch reactor (128 kJ/mol).412 Isomerisation and reversion reactions have been suggested to be responsible for these differences. Only a limited number of studies is available for the conversion of glucose in ionic liquid systems. An activation energy of 114 kJ/mol was reported for the conversion of glucose to 347 HMF in [BMIm]Cl , catalysed by CrCl3. As part of the acid-catalysed hydrolysis of cellobiose in [EMIm]Cl the activation energy for the decomposition of glucose was reported to be 90 kJ/mol.424 A number of studies have been reported where the decomposition reaction of glucose is not lumped into a single reaction. Here the rates of the reactions to individual products have been considered. For the dehydration of glucose to HMF the reported values for the activation energy show a large spread and range between 54 and 152.2 kJ/mol. For the reaction of glucose to humins, three datasets are available, which also show a large spread in activation energy of 56–135.7 kJ/mol. The lowest activation energy for the decomposition of glucose of around 50 kJ/mol, far below the value reported in other studies, was reported by Chang and co-workers on the decomposition of glucose or wheat straw to levulinic acid in aqueous sulfuric acid. The authors did not provide an explanation for this anomalous value.418,419 With the kinetic data available for the desired reaction to HMF and the undesired reaction to humins, the effect of temperature on HMF selectivity may be estimated. Girisuta et al.165 (sulfuric acid as the catalyst) and Jing et al.249 (no catalyst) found that the activation energy for undesired decomposition reaction of glucose to humins is higher than for the desired
172| Chapter 2 reaction of glucose to HMF; suggesting that high HMF selectivity is preferred at lower temperature for these systems.
200 GLC DP GLC HMF GLC HUM 175
150
125
100
75
50 activationenergy (kJ/mol)
25
0
Jing,2008 Jing,2008 Qi,2010 Jing,2008 Kabyamela Chang,2009Smith,1982Baugh,1988Smith,1982 Fagan,1971Xiang,2004 Chang,2009Chang,2006Pillath,2010 Chang,2005 Khajavi,2005 Saeman,1945Malester,1992Vanoye,2009 Heimlich,1960Girisuta,2006 Girisuta,2006 Thompson,1979 Kupiainen,2011 McKibbins,1962 Bienkowskwi,1987
Figure 3. Activation energies for the conversion of glucose (light grey bars: aqueous systems; dark grey bars: non-aqueous systems including mixed solvents and biphasic systems; white bars: thermal aqueous systems), DP = decomposition products; HUM = humins
2.5.2.3 Kinetic studies on the formation of HMF from cellulose, lignocellulosic biomass or fructan based biomass
As the major C6-sugars are available in biopolymers such as cellulose, starch, or inulin, it is relevant to highlight the conversions of these polymers to C6-sugars and derivatives (Table 37). Though the focus of most of these studies is not on HMF formation, HMF is involved as a decomposition or intermediate product in processes for obtaining sugar monomers or subsequent derivates such as levulinic acid. Surprisingly, kinetic studies on the formation HMF directly from polyfructans like inulin are not available.
Cellulose and lignocellulosic biomass, aqueous systems Saeman was the first to report kinetic studies on the hydrolysis of cellulose in dilute sulfuric acid at elevated temperatures.409 Woody biomass (Douglas-fir, 5-20 wt%) was used as feedstock and treated with 0.4-1.6 wt% sulfuric acid at 170–190 °C. The reactions were carried out in an isothermal batch setup (glass bomb). A homogenous pseudo first order reaction was assumed to describe the kinetic decomposition of cellulose to glucose. The
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|173 reaction order for sulfuric acid was found to be 1.34. The activation energy was estimated at 179.5 kJ/mol, which was higher than the activation energy of the decomposition of glucose to HMF (137.5 kJ/mol). Fagan et al. performed experiments on the hydrolyses of kraft paper slurry using 0.2–1 wt% sulfuric acid as the catalyst at 180 – 230 °C.425 The experiments were carried out in a non-isothermal batch setup. A kinetic pseudo-first order model developed for the hydrolysis of wood chips by Saeman was used to fit the data.409 The activation energies were found to be 188.7 kJ/mol and 137.2 kJ/mol for the decomposition of cellulose and glucose, respectively. Using an isothermal plug-flow reactor, Thompson and Grethlein performed a kinetic study on the acid-catalysed hydrolysis of purified Solka-Floc® purified cellulose (5–13.5 wt%) using 0.5–2 wt% sulfuric acid at 180-240 °C and short reaction times.426 The experimental data were best described using a the cellulose decomposition model proposed by Saeman,409 resulting in estimated activation energies of 177.8 kJ/mol and 136.8 kJ/mol for cellulose and glucose decomposition, respectively . Malester et al. reported the kinetics of dilute acid hydrolysis of cellulose originating from municipal solid waste (MSW) with sulfuric acid as the catalyst.427 The experiments were carried out in a batch reactor at 200–240 °C. Experimental data were modelled using a homogeneous pseudo first order reaction network viz the reaction of cellulose to glucose and subsequent conversion of glucose to its decomposition products. The activation energies were estimated at 171.5 kJ/mol for the conversion of cellulose to glucose and 142.3 kJ/mol for the decomposition of glucose. Antal and co-workers examined the acid catalysed hydrolysis of cellulose in a percolation reactor using 5 mM dilute sulfuric acid as the catalyst at 190–250 °C.272 They reported that the classical model of two sequential reactions for the decomposition of cellulose to glucose does not adequately describe cellulose hydrolysis at temperatures below 220 °C. The authors proposed a side reaction, i.e. the decomposition of cellulose to non-hydrolysable oligomers, to compensate for the imperfect glucose yields. Activation energies of 140 kJ/mol for the decomposition of cellulose to glucose and 100 kJ/mol for the decomposition of cellulose to non-hydrolysable sugars were reported. Sasaki et al. 428 investigated the non-catalysed thermal degradation of cellobiose in subcritical and supercritical water. The experiments were carried out in a continuous microreactor at 325–400 °C, 25–40 MPa and residence times of 0.01–0.54 s. The decomposition of cellobiose proceeds through two parallel reactions: hydrolysis to glucose
174| Chapter 2 and retro-aldol condensation to lycosil-erythrose. Both pathways resulted in glucose, which subsequently converted to fragmented products (HMF, furfural, erythrose, glyceraldehyde, dihydroxyacetone). The degradation of cellobiose was modelled using a first-order reaction. An activation energy of 111.2 kJ/mol was obtained. The authors also reported the conversion of microcrystalline cellulose in subcritical and supercritical water.429 The reactions were performed at 290–400 °C, at a fixed pressure of 25 MPa and residence times of 0.02–13.1 s using a continuous microreactor. The rate of cellulose conversion was modelled using a shrinking-core (grain) model, in which two regimes were discriminated. Below 370 °C, the activation energy was 145.94.6 kJ/mol. Above 370 °C, the conversion rate of cellulose was much faster due to the contribution of swelling or dissolution of cellulose and possibly pyrolytic depolymerisation of cellulose, with an activation energy of 547.9 27.8 kJ/mol. Girisuta et al. reported the kinetics of the acid-catalysed decomposition of cellulose to levulinic acid using 0.05–1 M sulfuric acid as the catalysts.189 The reactions were carried out in glass batch reactors with 1.7–14 wt% microcrystalline cellulose as the starting material at 180-230 oC. Experimental data were modelled using a power law model for 6 individual reactions, including that of the formation and degradation of HMF. The activation energy for the degradation of cellulose to humins (174.7 kJ/mol) was higher than the activation energy for the main reaction, i.e. the hydrolysis of cellulose to glucose (151 kJ/mol). This kinetic model was applied on the acid catalysed conversion of a real biomass source, namely water hyacinth.167 Though these studies were performed to optimise LA yields, HMF was included as an intermediate and the results may also be used to select optimal reaction conditions for high HMF yields. Recently Shen and Wyman reported a kinetic study on the acid-catalysed decomposition of microcrystalline cellulose using HCl (0.309 – 0.927 M) as the catalyst.430 The reactions were carried out in an isothermal batch reactor at 160–200 °C for up to 50 min. Experimental data were modelled using a homogenous pseudo first order reaction for six individual reactions, involving the formation and degradation of HMF. Compared to the scheme proposed by Girisuta et al.,189 the author did not include the decomposition of cellulose to insoluble products. The activation energy for cellulose decomposition to glucose was estimated at 95.6 kJ/mol, which was considerably lower than the activation energies reported for sulfuric acid catalysed cellulose decomposition (151–184 kJ/mol), indicating that HCl is a more effective catalyst than sulfuric acid for this reaction.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|175
Recently, Vazquez et al. reported a kinetic study on the acid catalysed hydrolysis of wheat straw in batch.431 The reactions were carried out at a fixed temperature of 130°C using sulfuric acid with the main objective to study optimum conditions for xylose production. Kinetic models were determined, not only for the target xylose but also for HMF and other intermediates, using a pseudo homogeneous irreversible first order approach.
Cellulose and lignocellulosic biomass, non-aqueous systems Vanoye et al. performed a kinetic study on the acid-catalysed hydrolysis of cellobiose in [EMIm]Cl using 3.5 mM methanesulfonic acid as the catalysts and small amounts of water (3.5 mM) as the co-solvent.424 The reactions were carried at 80–110 °C in a magnetically stirred (700 rpm) glass micro-reactor. The hydrolysis of 10 wt% cellobiose was modelled using two successive first order reactions involving the decomposition of cellobiose to glucose and the subsequent decomposition of glucose. Activation energies of 111 12 kJ/mol and 102 9 kJ/mol were reported for cellobiose hydrolysis and glucose degradation. Bell et al. carried out kinetic studies in a batch set up on the hydrolysis of cellulose in typical ionic liquids ([EMIm]Cl and [BMIm]Cl) using mineral acid catalysts with variable amounts of water present.432 Glucose, cellobiose and HMF were observed as primary reaction products. The rate of glucose formation was determined to be first order in the concentration of dissolved glucans and zero order in the concentration of water with an activation energy of 96 kJ/mol.
Polyfructans and Oligofructans Though several studies revealed that polyfructans such as inulin, are excellent starting materials for HMF production in water183,184,186 and ionic liquid systems,367,433 attempts to quantify and model HMF yields have not been published to date. Heyraud et al. examined the acid-catalysed hydrolysis of oligofructans obtained from the Jerusalem artichoke. The purified oligofructans with degrees of polymerisation (d.p.) between 2 and 7 were hydrolysed individually at a fixed temperature of 70 °C using aqueous sulfuric acid at pH 2.434 The experimental data were modelled using pseudo first order kinetics. Cleavage of the glucose-fructose linkage was identified as the main limiting step for the acid catalysed hydrolysis of polyfructan from the juice of Jerusalem artichokes. Christian et al. studied the kinetics of the formation of di-D-fructose dianhydrides, a family of isomeric cyclic difructans, during thermolysis of pure inulin in the presence of citric
176| Chapter 2 acid.154 The reactions were performed at 160-180 °C in a teflon-lined batch reactor immersed in an oil bath. Blecker et al. reported a kinetic study on the acid catalysed hydrolysis of five commercial oligofructans.435 The reactions were carried out at 7 – 130 °C using HCl as the catalyst with pH values between 2.0 and 4.2. The decomposition of the oligofructans was modelled using first order kinetics with fructose considered as the main reaction product. The formation of HMF was not reported. The activation energy for the conversions of oligofructans to fructose was estimated at 10910 kJ/mol. L’Homme et al. performed experiments on the acid catalysed hydrolysis of oligofructans such as 1-kestose (GF2), nystose (GF3), and fructofuranosylnystose (GF4).436 The reactions were performed in mineral-acid buffered aqueous solutions at pH 4, 7, and 9 at 80–120 °C. Experimental data were modelled using first order kinetics. The decomposition rate of oligofructan is faster in acidic conditions than in neutral or basic conditions. Activation energies of 80.9, 76.2, and 84.6 kJ/mol were reported for 1-kestose, nystose, and fructofuranosylnystose respectively. The formation rates of the decomposition products of these reactions, either fructose or HMF, were not reported.
Discussion The decomposition of cellulose in aqueous systems is often modelled using homogeneous models based on a pseudo first order approach. Figure 4 shows the reported activation energies for the decomposition of cellulose, cellobiose and oligofructan. The activation energies for the acid catalysed decomposition of cellulose in water are between 79 and 189 kJ/mol. However, when excluding the low values obtained by Chang et al.418 and Shen and Wyman.430 (75 kJ/mol and 96 kJ/mol, respectively), the average value is about 168 kJ/mol. Two studies have been performed on the decomposition of cellobiose, one in water and one in an ionic liquid. The activation energy for cellobiose decomposition in the ionic liquid using methanesulfonic acid as a catalyst was found to be 100 kJ/mol, which is close to the value found for the non-catalytic thermal decomposition of cellobiose in subcritical conditions.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|177
250
225 CELL GLC CELL HUM FOS DP 200
175 CELO GLC 150
125
100
75
activationenergy (kJ/mol) 50
25
0
Mok,1992 Mok,1992 Chang,2009Shen,2011 Fagan,1971 Girisuta,2007Saeman,1945Malester,1992 Girisuta,2007 Sasaki,2002Vanoye,2009 Blecker,2002 Thompson,1979 L’Homme,2003
Figure 4. Activation energies for the conversion of cellulose (CELL: cellulose, GLC: glucose, HUM: humin, CELO: cellobiose, FOS: oligofructan, DP: decomposition product; light grey bars: aqueous systems; dark grey bars: non-aqueous systems including mixed solvents and biphasic systems)
2.5.2.4 Kinetic studies on the decomposition of HMF
HMF is an intermediate in the proposed reaction network for the conversion of C6-sugars to levulinic acid. For HMF synthesis, the undesired reactions of HMF to LA, FA and humins should be suppressed as much as possible. To gain insight in the reactivity of HMF, kinetic studies using HMF as the starting material have been reported. An alternative approach involves the investigation of the dehydration of C6-sugars where both the formation and the subsequent conversion of HMF are modelled in an integrated manner (see Table 35). In this paragraph, only the studies with HMF as the starting material are discussed and the results are compiled in Table 38.
Thermal decomposition of HMF Luijkx et al. investigated the non-catalytic decomposition of HMF in subcritical water.437 The experiments were carried out with 0.05 M HMF in a tubular flow reactor at 290-400 °C, a fixed pressure of 27.5 MPa, and residence times ranging from 1 to 15 minutes. 1,2,4- benzenetriol was identified as the main reaction product with selectivities up to 46% at 50% HMF conversion. The rate of HMF decomposition was modelled as a pseudo-first order reaction. A linear Arrhenius plot could only be obtained in the temperature range of 290–350 °C, from which an activation energy of 47.7 kJ/mol was obtained.
178| Chapter 2
Recently Chuntanapum et al. studied the thermal decomposition of HMF in both subcritical and supercritical water.438 The experiments were performed in a continuous flow reactor at 175–450 °C and a fixed pressure of 25 MPa for 80–400 s. A series of first-order reactions was proposed to model the decomposition of HMF to intermediate liquid products and the subsequent decomposition to gaseous products, mainly CO2, CO, and H2. The intermediate liquid products were not identified and thus the kinetic data were analysed using total organic carbon TOC yield of liquid products and carbon content of the gaseous products. A characteristic Arrhenius plot could be obtained for the temperature window of 175 – 450 °C. An activation energy of 75.8 kJ/mol was obtained, which was considerably higher than reported by Luijkx et al.437 In addition, contrary to Luijkx’s work, the authors found a lower rate of decomposition of HMF and tar, char and other high-molecular-weight compounds were not detected. The authors suggested that these differences were caused by different heating trajectories and the use of more dilute HMF solutions (0.02 M). As part of a kinetic study on the thermal decomposition of glucose in water, Jing et al. investigated the kinetics of the thermal decomposition of HMF in subcritical water in batch (180–260 °C, 100 bar, 300 rpm).249 Experimental data were modelled using a first order approach using a kinetic scheme involving two parallel reactions to incorporate the formation of LA and insoluble humins. Activation energies of 89.28 kJ/mol and 108.5 kJ/mol were reported for the reaction of HMF to LA and the side reaction to humins, respectively.
Acid-catalysed hydrolysis of HMF, aqueous systems The first kinetic study on the reaction of HMF to LA and FA was performed in water under reflux with a protective atmosphere by Teunissen in 1930.393 An initial HMF concentration around 0.08 M was applied with a range of acids at 0.1-0.5 N . The reaction was modelled using a first order approach for both HMF and the catalyst. The rate of the reaction was shown to be a function of the acid catalyst, with HI giving the highest reaction rates and oxalic acid the lowest. McKibbins et al. studied the kinetics of the rehydration of HMF (0.06–0.14 M) to levulinic acid in 0.025–0.4 N aqueous sulfuric acid at 160-220 °C in glass ampoules.411 The experimental data were modelled using a first order approach, leading to an activation energy of 96.8 kJ/mol. Baugh et al. performed kinetic studies on the decomposition of HMF in water using a mixture of butyric acid and phosphoric acid as the catalyst at pH values between 1 and 4.415 The reactions were performed at 170-230 °C in batch. The concentration of HMF versus time
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|179 was determined and the data were modelled using a pseudo first order approach in HMF leading to an activation energy of 55.9 kJ/mol. Girisuta et al. investigated the kinetics of HMF decomposition to LA/FA and humins in glass ampoules, with an initial HMF concentration of 0.1-1 M in 0.05-1 M aqueous sulfuric acid at 98-181 °C.164 An extended kinetic framework was proposed, including the decomposition of HMF to humins besides the reaction to LA and FA, using a power law model. This resulted in an activation energy of about 111 ± 2 kJ/mol for both the main reaction to LA/FA and the side reaction to humins. The order in HMF for the main reaction (0.88) was lower than for the reaction to humins (1.23). As part of a kinetic study on the decomposition of fructose, Asghari et al. examined the decomposition of HMF, catalysed by HCl, in subcritical water.404 The reactions were performed at 210-270 °C, 4 MPa and an initial pH of 1.8 in a continuous tubular reactor with residence times of 0.5–300 s. The experimental data were modelled assuming pseudo first order kinetics in HMF for both the reaction to LA/FA and the formation of humins, resulting in activation energies of 94.1 kJ/mol for the main reaction and 121.5 kJ/mol for the side reaction to humins.
Acid-catalysed hydrolysis of HMF, multi-solvent systems In 1931 Teunissen investigated the kinetics of the acid catalysed decomposition of HMF in mixtures of water with up to 40 wt% methanol and 75 wt% ethanol.439 The experiments were either carried out at the boiling point of the mixture in a reflux system or at a fixed temperature of 100 °C in a closed bottle, with 0.5 N HCl as the catalyst for all experiments. In the presence of either methanol or ethanol, the decomposition rate of HMF to levulinic acid and formic acid was considerably lower, with the reaction constants decreasing considerably with increasing alcohol concentration. The hydrolysis of HMF was not observed when heated in 40 wt% methanol or 75 wt% ethanol. The amount of humins produced after 8 h reaction time decreased at higher alcohol intakes. Even though there were indications of alcoholysis reactions, the products of these reactions were not determined at that time. More recently, the formation of HMF-ethers and levulinate esters were reported when reacting fructose and glucose in acidic alcohol mixtures.233,440
Discussion Figure 5 provides the activation energies for the decomposition of HMF, both from studies using HMF as the starting materials and studies in which data were derived from sugar
180| Chapter 2 dehydration data. The activation energies range between 47 and 210 kJ/mol for the decomposition of HMF to levulinic acid. For the decomposition reaction of HMF to humins, the activation energies vary between 100 and 125 kJ/mol. Asghari et al.,404 Jing et al.249 and Abdilla et al.405 reported that the formation of humins from HMF has a higher activation energy than the rehydration of HMF to levulinic acid. Contrary to these observations, Girisuta et al. found similar activation energies for both reactions.164
250
225 HMF DP HMF LA + FA HMF HUM
200
175
150
125
100
75
activationenergy (kJ/mol) 50
25
0
Luijcx,1993 Jing,2008 Shen,2011 Jing,2008 Shen,2011 Chang,2009Baugh,1988Chang,2009Abdilla,2011Asghari,2007 Girisuta,2006Chang,2006Moreau,1996Kuster,1977 Girisuta,2006Abdilla,2011Asghari,2007 McKibbins,1962Kupiainen,2011 Kupiainen,2011 Chuntanapun,2008
Figure 5. Activation energies for the conversion of HMF (light grey bars: aqueous systems; dark grey bars: non-aqueous systems including mixed solvents and biphasic systems; white bars: thermal aqueous systems)
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|181
Table 35. Kinetic expressions for the conversion of fructose
Scheme Solvent [FRC],0 Catalyst T kinetic expression (M/min) activation Ref. energy
(kJ/mol)
Aqueous Systems
No catalyst
H2O 0.25 wt% - 300 – 350 °C at T = 300 °C P = 25 MPa - 121 (25- 40 MPa) RF 28.8FRC 400 °C (40 MPa)
Homogenous Catalysts
H2O 0.25–1 [HCl] = 95 °C 2 - 155 R1F 0.4010 FRC M 0.2–1.0 M
2 R1H 0.4210 HMF at [HCl] = 0.5 M
182| Chapter 2
H2O 0.03 M [HCl], 210-270 °C, at T = 210 °C, 40 bar ER1F 160.6 404 pH = 1.8 40 – 150 bar R1F 0.272FRC ER2F 132.2
R2F 0.031FRC
ER3F 101.9 R3F 6.12FRC R 0.181HMF 1H ER1H 95.6
R2H 0.003HMF
ER2H 141.5
H2O - 180–220 °C, at T= 200 °C E 123 RF 100 bar 126.8 3.3 RF 0.08FRC
10.8 formic acid= 180–220 °C, at 10.8 mg/ml formic acid and T = 200 ERF 123 mg/ml 100 bar °C 112.0 13 2.8–13.9
g/ml RF 0.36FRC
10.8 acetic acid= 180–220 °C, at 10.8 mg/ml acetic acid and T = 200 ERF 123 mg/ml 100 bar °C 125.60 13 10.8 mg/ml
RF 0.12 FRC
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|183
0.96 H2O 0.1–1M [H2SO4]= 140–180 °C 1.006 ER1F 123 5 405 R1F k1G H FRC 5 mM–1M ER1H 1 1 k1F 1.093exp ER2F 147 12 R T 413
1.06 1.18 R2H k2H H FRC
ER2H 1 1 k2H 0.547 exp R T 413
Heterogeneous Catalysts
H2O 0.3 M NbOPO4 90 – 110 °C T=100 °C R 0.0229 FRC E 65.88 200 F RF
Non Aqueous Systems
Homogenous Catalysts
MeOH 0.06 M [H2SO4]= 120–270 °C, at T=180 °C, 200 bar ERF 80 233 10 mM 150–350 bar (R = R 8.7FRC F 1F R1F+ R2F) R2F 2.88FRC
184| Chapter 2
R1H 14.6HMF
R2H 0.18HMF
CH3CO2H 0.06 M [H2SO4]= 180 °C, R1F 3.18FRC - 233 R 7.38 FRC 10 mM 200 bar 2F
R1H 9.12HMF
[HMIm]Cl 0.01-2.5 90–120 °C RF 0.1681FRC ERF 143 334
ERH 69
[BMIM]Cl 200 mg 7 mol% 80–100 °C at T = 80 °C ER1F 165 269
in 2 g IrCl3 R1F 0.0019 FRC E 124 ionic R2F liquid R2F 0.0011FRC
Two solvents/Biphasic Systems
Homogeneous Catalysts
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|185
H O 0.2-1 M [H PO ]= 170–220 °C, E 93 6 199 2 3 4 ER1F 1 1 R1F R1F k1F exp 0.1-0.5 M R T 453 ERx 130 42 50 atm E 100 42 k1F 0.0481 35 KW [H 3 PO4 ] R1H
ER1F 1 1 R1X R2X kX exp R T 453
kX 13KW [H3PO4]
ER1F 1 1 R1H k1H exp R T 453
k1H 0.33KW [H 3 PO4 ]
Acetone/ 0.05 M [H SO ]= 180–300 °C, at T= 240 °C 233 2 4 H2O 0-5 mM 200 bar 0 mM H2SO4 R 0.492FRC E 158 (90:10) F RF
3 mM H2SO4 RF 34.14FRC ERF 99
5 mM H2SO4 RF 50.99FRC ERF 98
Heterogeneous Catalysts
H2O: 0.093 H- 160 – 178 °C - ERF 141 102
MIBK
186| Chapter 2
(1:5) mordenites ERH 64
Acetone/ 2 wt% Dowex 100-180 °C, R 0.15 FRC at 150 °C E 103.4 207 F RF H2O 50wx8-100 MW (70/30 w/w)
Acetone/ 2 wt% Dowex 100-180 °C, R 0.1681 FRC at 150 °C E 60.4 221 F RF DMSO 50wx8-100 MW (70:30 w/w)
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|187
Table 36. Kinetic expressions for the conversion of glucose
Scheme Solv [GLC], Catalyst T kinetic expression (M/min) activation Ref.
ent 0 energy (kJ/mol)
H O 0.278 M [H SO ]= 170–190 °C 409 2 2 4 14 1.02 ERG ERG 137.5 RG 2.38 10 H 2SO4 exp [GLC] 0.4, 0.8, RT 1.6 wt%
3 H2O 0.056 M HCl= 100–150 °C at 150 °C R 2.810 GLC ERG 133 410 G 0.35 N
H O 0.56– [H SO ]= 140–250 °C 411 2 2 4 11 ERH ERH 96.8 RH 2.410 H H2SO4 exp HMF 1.11 M 0.025 – RT 0.8 N
14 ERG RG 9.27 10 G H 2SO4 exp GLC E 136.8 RT RG
H2O 0.4–6 [H2SO4]: 180–224 °C ampoule (0.5 – 7 min): ERG 128.6 412 wt% 0.5–4.0 12 0.8955 ERG wt% RG 2.5510 H2SO4 exp GLC RT
188| Chapter 2
H2O 0.4–6 [H2SO4]= flow reactor (0.06 – 0.27 min): ERG 87.8 412 wt% 0.5–3.8 E wt% 9 0.5687 RG RG 3.84 10 H 2SO4 exp GLC RT
H O 0.006- a mixture 170-230 °C 415 2 ERG ERG 121 RG kG exp 0.33 M of butyric RT 11 13 pH acid and kG [4.910 1.510 10 22 pH14 H3PO4 4.710 10 ] adjusted
to
pH = 1-4
H O 0.125 M H SO , 190–210 °C Glu 13 15 416 2 2 4 k 2.13210 2.14810 H ERG 139 pH 1.5– ERG exp 2.2 RT
10640 ke 1.253135.37 H exp RT
H2O 4 – 12 [H2SO4]= 100 – 144 °C 16 ERG ERG 130.4 414 R1G 1.9110 H fH exp( )GLC wt% 4–20 wt% RT
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|189
H2O 0.007M - 300 – 350 °C Rgf 14.4GLC ERg Rgf 96
122 Rg 12.6GLC
R f 49.2GLC
H2O 0.5 % - 180–260 °C C E 155 408 exp k n RG w/v C0
0.427 H2O [H2SO4]: 170–190 °C ER1G 86.33 418 R1G k1G H GLC 1%, 3%, E 56.95 E R2G k 4.597109 exp R1G 5% 1G ERH 209.5 RT
0.973 R2G k2G H GLC
7 ER2G k1G 4.31810 exp RT
0.119 RH k H H GLC
23 ERH k H 1.2010 exp RT
1.13 H2O 0.1–1 M [H2SO4]: 140–200 °C 1.08 ER1G 165 R1G k1G H GLC 152.2 0.7 0.05–1 M ER1G 1 1 k1G 0.013exp R T 413
190| Chapter 2
1.12 1.13 ER2G R2G k2G H GLC 164.7 0.6 ER2G 1 1 k2G 0.013exp R T 413
H2O 0.06 M - 180–260 °C E 249 R 5.06 109 exp( R1G )GLC ER1G 108.03 1G RT ER2G 135.71 E 89.28 R1H 12 ER2G E 108.91 R2G 4.29 10 exp( )GLC R2H RT
E R 2.62 106 exp( R1H )HMF 1H RT
E R 3.57 109 exp( R2H )HMF 2H RT
H2O 10 – [H2SO4]: 120–180 °C E ER1G 133 413 R 1.881014 exp( R1G )GLC 200 1.2 wt% 1G RT mg/ml
`6 H2O CrCl2 100 °C 421 RG 6.510 GLC
H2O 56 mM Formic 180–220°C R k GLC E 420 G G RG or 112 acid: ERG 1 1 153 2 kG (0.018 2.6H ) exp mM 5-20 wt% R T 473
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|191
R k INT E 1I I R1I 110 5 ER1I 1 1 k1I (0.109 8.6H ) exp R T 473 E R2I 117 4 R2I k2I INT
ER2I 1 1 E k2I (0.058 2.9H ) exp R1H R T 473 107 5
R k HMF E 1H 1H R2H ER1H 1 1 127 9 k1H (0 5.5H ) exp R T 473
R k HMF 2H 2H ER2H 1 1 k2H (0.031 2.5H ) exp R T 473
192| Chapter 2
Table 37. Kinetic expression for the acid-catalysed conversion of Cellulose and Lignocellulosic biomass
Scheme Sol- [Cell],0 Catalys T kinetic expression (M/min) activation Ref vent t energy
(kJ/mol)
Cellulose/Lignocellulosic Biomass
H O Douglass- [H SO ]: 170–190 °C 409 2 2 4 19 1.34 ERC ERC 179.5 RC 1.7310 H2SO4 exp [CEL] fir, Liquid 0.4, 0.8, RT
to solid 1.6 wt% ERG 137.5 E ratio: 5 – R 2.381014H SO 1.02 exp RG [GLC] G 2 4 RT 20
H O kraft paper, [H SO ]: 180–230 °C 425 2 2 4 19 1.78 ERC ERC 189.4 RC 28 10 H 2SO4 exp [CEL] 0.5 g in 20 0.2 – 1 RT ml wt% ERG 137.8
14 0.55 ERG RG 4.19 10 H 2SO4 exp [GLC] RT
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|193
H O Solka-floc, [H SO ]: 180–240 °C 426 2 2 4 19 1.16 ERC ERC 177.8 RC 1.22 10 H 2SO4 exp [CEL] purified 0.5 – 2.0 RT cellulose: wt% ERG 136.8 5.0 – 14 0.69 ERG 13.5% RG 3.79 10 H 2SO4 exp [GLC] RT
H O cellulose H SO at 200–240 °C 427 2 2 4 19 pH ERB ERC 171.5 RC 1.16 10 10 exp [CEL] from pH 0.34 RT
Municipal – 0.85 ERG 142.3 E Solid R 1.16 101910pH exp RG [CEL] C RT Waste
H2O Cellobiose - 325–400 °C RC k[Cellobiose] ERB 112 428 (25–40 MPa)
H2O micro - 290–400 °C at T< 370 °C 429 crystalline (250 bar) E cellulose, RC 145.9 4.6 10 wt%
194| Chapter 2
at T> 370 °C
E RC 547.9 27.8
H2O micro [H2SO4]: 150- 200 °C 0.98 E 151.5 189 R1C k1c CEL R1C crystalline 0.05 – 1 ER1C 1 1 0.96 k1c 0.410 exp H cellulose M R T 448 1.7 – 14 R k CEL 1.01 wt% 2C 2c ER2C 174.7
ER2C 1 1 0.94 k2c 0.065exp H R T 448
[EMI 0.5 g 3.5 mM 80 – 110 °C RC k[Cellobiose] ERC 11112 424 M]Cl Cellobiose methane
in 5 g sulfonic RG kG[GLC] ERG 102 9 [EMIm]Cl acid
H2O [H2SO4] 190, 210, 1.434 E 78.66 419 RC kC H CELL RC =1,2, 230 °C 10 ERC kC 1.48810 exp 3wt% RT
ERG 61.06
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|195
RG R1G R2G 0.679 R k H GLC G G 7 ERG k1G 1.430 10 exp RT
0.268 E 54.51 R1G k1G H GLC R1G
5 ER1G k1G 1.502 10 exp RT
RH R1H R2H ERH 51.37 0.620 RH kH H HMF
6 ERH k1G 2.42510 exp RT
ER1H 56.47 0.804 R1H k1H H HMF
7 ER1H k1H 1.42510 exp RT
H2O Micro- [HCl] : 160-200 °C at 99.6 M cellulose, T=200 °C, HCl: 0.309 M ERC 95.6 430 crystalline 0.309 – RC 1.35CEL R1G 1.80 GLC cellulose8. 0.927 M ER1G 137.4
09 – 24.1 R2G 0 g/L ER1H 144 R1H 3.17 HMF R2H 3.08 HMF
196| Chapter 2
Oligofructan
H2O Commercia 7- 130 °C E 109 10 435
l FOS
H2O 1-kestose aqueous 80 -120 °C at pH: 4, T:120 °C at pH = 4 436 (GF2) solutions 4 E 80.9 RF 580 10 GF 2 GF 2 nystose buffered
4 EGF 3 76.2 (GF3) at pH RF 472 10 GF3 values of fructofuran 4 EGF 4 84.6 4.0, 7.0, RF 315 10 GF 4 osylnystose and 9.0 (GF4)
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|197
Table 38. Kinetic expression for the decomposition of HMF
Scheme Solvent [HMF],0 Catalyst T kinetic expression (M/min) activation Ref energy (M) (°C) (kJ/mol)
Thermal decomposition
H2O 0.05 - 290-400 at T =290 °C RH 0.107 HMF ERH 47.7 437 (27.5
MPa) at T= 350 °C RH 0.301HMF
H O 0.02 - 175-450 438 2 5 ERH ERH 75.76 RH 2.05 10 exp HMF (25 MPa) RT
ERL 109.58
H2O 0.06 - 180 – at T = 200 °C ER1H 89.28 249 260 R 3.2 104 HMF P=100 1H ER2H 108.81
bar 4 R2H 3.58 10 HMF
Acid-catalysed hydrolysis
198| Chapter 2
H2O 0.08-0.09 [H2SO4], 100 2 - 393 0.2 N oxalic acid R1H 1.24 10 H HMF [HCl], [HBr] or [HI] = 2 0.2 N H2SO4 R1H 1.57 10 H HMF 0.1-0.5 N 2 0.2 N HCl R1H 3.82 10 H HMF oxalic acid =
0.2 – 0.5 N 2 0.2 N HBr R1H 5.6710 H HMF
2 0.2 N HI R1H 5.8910 H HMF
H2O 0.061- [H2SO4] = 160-220 11 ER1H ER1H 96.8 411 R1H 2.410 αH H2SO4 exp HMF 0.139 0.025-0.4 N RT
H2O 0.25-1 [HCl] = 95 1.2 - 155 R1H 0.001H HMF 0.5-2.0 N
H O 0.024 a mixture of 170-230 415 2 6 pH ER1H ER1H 55.9 R1H 1300 4.110 10 exp HMF butyric acid RT
and H3PO4 adjusted to
pH = 1-4
H O 0.1-1 [H SO ] = 98 -181 E 164 2 2 4 ER1H 1 1 1.38 0.88R1H R1H 0.340exp H HMF 0.05-1 M R T 413 110.5 0.7
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|199
ER2H 1 1 1.07 1.23 R2H 0.117 exp H HMF R T 413
E R2H 111.0 2.0
H2O 0.04 [HCl], 210 – at T = 210 °C ER1H 95.6 404
270, 4 R1H 0.199HMF pH = 1.8 MPa ER2H 114.8
R2H 0.003HMF
H2O/ 0.08-0.09 0.5 N HCl Reflux at T=100 °C - 439 methanol and methanol or
H2O/ T = 100 2 10%v RH 6.66 10 HMF ethanol °C 2 30%v RH 1.210 HMF
ethanol
200| Chapter 2
2 10%v RH 8.34 10 HMF
2 35%v RH 3.310 HMF
2 50%v RH 1.310 HMF
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|201
2.5.3 Reactor Concepts
Most of the reported studies on HMF preparation from monomeric sugars like fructose as well as from more complex feeds are exploratory in nature and have been performed in lab scale equipment. As such, most studies are typically performed in batch setups (autoclaves, glass ampoules). This also appears to be the case for the few reported studies on kg scale HMF production. For instance, Rapp conducted the conversion of 11 kg fructose in water (33 wt%) using oxalic acid as the catalyst at 140 °C for 3 h in a stirred batch autoclave.195 Using DMSO as the solvents, M’bazoa performed close to a kg-scale operation for the conversion of fructose to HMF in a stirred autoclave.396 In this chapter, reactor concepts other than typical laboratory batch set-ups will be discussed. Further details on pilot scale production are given in section 2.5.5.
2.5.3.1 Reactions in water
Some studies, particularly those at elevated temperature and pressure like in sub- and supercritical water, have been performed in continuous setups to allow for short contact times.203,204,404,441 Asghari and Yoshida utilised a home built continuous tubular reactor of SS 316 steel with an inside diameter of 25 mm to develop a kinetic model for the acid catalysed production of HMF from fructose at temperatures of 210 – 270 °C, pressures of 1-15 MPa, and residence times of 0.5 – 300 s. To allow accurate and fast heating, water was preheated separately to the target temperature, then mixed to the feed solution and subsequently passed through the reactor. Tarabanko et al. reported the formation of HMF from fructose and sucrose at temperatures >200 °C in a flow reactor using phosphoric acid as the catalyst.203 HMF yields up to 40 % were attained when processing 0.25 M fructose at 240–250 °C using 0.01 M phosphoric acid. The corrosion rate of the reactor material (stainless steel 12Kh18N10T) was found to be 1-2 mm/year at the prevailing reaction conditions. Continuous microreactors have also been explored for the synthesis of HMF in the absence of any catalyst. At extreme severity of 400 °C, 10 MPa and 10 s residence time, Japanese researchers at Advance Industrial Science and Technology (AIST) reported the synthesis of 80% pure HMF at 70% yield from aqueous fructose.441 The estimated continuous production capacity was 500 kg/year for the setup. Tuercke et al. examined the conversion of fructose to HMF in aqueous solutions using a continuous microreactor setup (Figure 6) at elevated temperatures.442 Aqueous fructose and
202| Chapter 2
HCl were introduced to the microreactor via two fluid inlets. A wide range of operating conditions (10-50 wt% fructose, 80-200 °C,1–20 bar, 0.1–0.5 M HCl, 1–3 min residence times, 0.5–0.6 ml/min) was systematically explored to determine optimum process conditions. At optimum conditions of 10 wt% fructose and 0,1 M HCl in equal flow at 185 °C and 17 bar with 1 min residence time), an HMF selectivity of 75% at 71% fructose conversion was achieved when using a feed consisting of 10 wt% fructose. Compared to batch data at comparable conditions, the selectivity and conversion were increased considerably by 24% and 21%, respectively. Further improvement of the system by using DMSO as a co-solvent and MIBK/2-butanol as an extraction agent resulted in 85 % HMF selectivity at 97% fructose conversion.
Figure 6. A continuous microreacor setup as used by Tuercke et. Al. Reproduced from Ref 442 with permission from John Wiley and Sons
2.5.3.2 Reactions in non-aqueous and mixed solvent systems
A continuous setup has also been applied for the synthesis of HMF and derivatives, such as MMF, in low boiling solvents like acetone and methanol at supercritical conditions.206,233 The reactions were carried in a continuous high pressure reactor (pipe-in-pipe system) which can be heated up to 350 °C. Bicker et al. also proposed a continuous setup, as described in Figure 7, for the technical production of HMF in acetone and MMF in methanol.233 The flow system was operated at 180 °C for HMF production and at 240 °C for MMF production at a pressure of 10 MPa. Under milder reaction conditions than mentioned above (180 °C), Gruter and Dautzenberg demonstrated the production of HMF and EMF from sucrose, fructose or glucose in a continuous setup using either sulfuric acid or zeolites as the catalysts.244 The combined selectivity of HMF and EMF was over 90% at 25% sucrose conversion when dilute
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|203 sucrose solutions in ethanol/water (10 mM) was used at 195 °C in a continuous reactor with residence times of 6–60 s at a flow rate of 10 ml/min. When zeolite beta was used as the catalyst, the reactor was operated in a fixed bed mode. In this configuration, various sugar solutions, containing 55 mM sugar in 90% ethanol, were reacted at 180 °C. At high sugar conversion (98%), the yields of EMF were 37%, 22% and 28% for fructose, glucose, and sucrose respectively. The HMF yields were 1–2% for all the sugars.
Figure 7. Simplified process flow sheet for HMF and MMF production in supercritical acetone or methanol. Reprinted from Ref 233 with permission from Elsevier For mixed solvent systems, both batch stirred reactors298 and continuous reactors (continuous stirred tank reactor (CSTR)199 and tubular flow reactors102,294) have been explored for the preparation of HMF in mixed solvent systems. For biphasic systems, several reactor configurations were employed to promote intense contact between the reaction and extraction phases. Cope et al. used a tubular flow reactor, operated in a semi-batch mode, to study the reaction of fructose in a biphasic system consisting of MIBK and water (7:1 v/v) for 9 h at 175 °C.311 The 60% HMF yield under prevailing conditions was considerably higher
204| Chapter 2 compared to HMF yield in the absence of the organic extraction solvents, which was only 20–25 %. Kuster reported the application of a CSTR for the production of HMF from fructose using
MIBK as an in situ extracting agent at 180-200 °C, catalysed by H3PO4. The highest HMF selectivity was 75% at 65–85% conversion, obtained at 200 °C with 0.1 M H3PO4 and an
MIBK/H2O feed ratio of 4. Rivalier et al. developed a continuous heterogeneous pulsed column reactor to promote the yield of HMF from fructose in a biphasic water/MIBK system using a solid zeolite as the catalyst.294 Initially, the reactor was operated as a counter current fixed bed column reactor. Pulsing was applied at the base of the column to improve liquid-liquid mass transfer rates. Major disadvantages of this configuration were caused by limitations of hydrodynamic origin, namely flooding and emulsion formation, and also of kinetic origin, due to catalyst inactivation by the formation of brown soluble polymeric material. A considerable improvement was obtained when the reaction was performed in a slurry configuration. With this concept, Moreau et al. demonstrated an improvement in HMF selectivity of about 10% in subsequent research on the dehydration of fructose to HMF using H-mordenite as the catalyst.102 Recently McNeff et al. reported the continuous production of HMF from various feeds such as glucose, fructose, starch and cellulose in a biphasic system consisting of water and MIBK using a fixed bed reactor with a porous metal oxide-catalyst.303 The two liquids were mixed before entering a pre-heater. In the case of cellulose an additional solubilisation chamber was added after the pre-heater, as described in Figure 8. An HMF yield of 29% was obtained in a biphasic system consisting of a 23 wt% aqueous glucose solution and MIBK
(1:3 v/v), catalysed by TiO2 with a residence time of 3 min at 180 °C. Higher yields up to 34% were obtained when cellulose was reacted at a temperature of 270 °C using a MIBK to
H2O volume ratio of 5, a residence time of 2 min and TiO2 as the catalyst.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|205
Figure 8. Schematic representation of the continuous flow system for the production of HMF in biphasic systems. Reprinted from Ref.303 with permission from Elsevier Mihovilovic et al. reported studies on the conversion of glucose and fructose in two continuous microreactor devices (cartridge based and stop flow microwave reactor) in DMSO using HCl as the catalyst.443 Highests yields were obtained for fructose, viz 90.3% using the cartridge based reactor (180°C) and 85.6% undermicrowave heating (150°C). For glucose, the yields were considereably lower and 33% at maximum for the cartridge reactor system (200°C) and 29% for the microwave reactor (200°C).
206| Chapter 2
The main challenge regarding reactor design and operation is the improvement of the HMF yield, which is difficult considering the instability of HMF under reaction conditions. This means that follow-up reactions should be suppressed in order to prevent the formation of rehydration and polymerisation products. The solid polymeric by-products (humins) create an additional problem, complicating reactor design and operation because of plugging and deactivation of heterogeneous catalysts. Some studies mention that the solids are easily filtered from the reaction mixtures.195 However, other reports indicate that the humins deposit on the reactor wall, affecting heat transfer during the reaction and complicating operation and maintenance. To avoid accumulation of the deposition of humins on the reactor wall, the use of a PTFE layer in the reactor was suggested.444 Reactor design is further complicated when starting with cheaper biopolymer feeds like cellulose and lignocellulosic biomass. In these cases, introduction of a solid feed stream to a pressurised reactor poses an additional challenge.
2.5.4 Separation and purification strategies
Separation strategies to obtain pure HMF by removal of by-products, catalysts and solvents depend among others on the solvent system used for the reaction. This section is divided in two main parts, viz. separation and purification strategies for: i) aqueous systems and ii) non- aqueous systems.
2.5.4.1 Separation and purification strategies for aqueous systems
Typically HMF is isolated from aqueous reaction products by the following sequence: i) filtration of solids (humins), ii) neutralisation, iii) HMF isolation (solvent extraction or others) and finally iv) HMF purification (vacuum distillation or others).392,393,445 Each step will be discussed individually in the following sections.
Filtration Rapp reported the use of a pressure filter and stated that separation of the insoluble products is an unexpectedly simple process.195 Haworth suggested the possibility of centrifugation, though further details are lacking in their paper.445
Neutralisation When using homogeneous acid catalysts, the acid is neutralised before further processing to 195 195,392,393,445 a pH of about 5-6, for which CaCO3 is typically used. The use of other bases
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|207
311 195 392 445 such as K2CO3 and Ca(OH)2 was also reported. Middendorp and Haworth et al. proposed the use of lead acetate after a neutralisation step with CaCO3, though the exact purpose and benefits are unclear.
Isolation: Extractions Solvent extraction is a traditional approach to isolate HMF from the aqueous phase,392,393,445 often with ethyl acetate as the solvent of choice. Generally, the ethyl acetate layer is dried after extraction using typical drying agents, for instance MgSO4 or Na2SO4. Subsequently, the extraction phase is concentrated by a solvent evaporation step to obtain crude HMF, but both systematic solvent screening studies and detailed engineering studies on thermodynamic properties, like partitioning as a function of process conditions, are lacking to date. In reactive aqueous-organic systems, 1-butanol288 and MIBK311 were shown to be very useful for the in situ extraction of HMF from the aqueous phase. Recently, Román-Leshkov and Dumesic297 examined a number of organic solvents for in situ solvent extraction from aqueous systems saturated by NaCl at 180 °C. A broad range of solvents was examined, ranging from primary and secondary alcohols to ketones and cyclic ethers. The partitioning coefficient R of HMF, which was defined as the concentration of HMF in the organic phase divided by the concentration of HMF in the aqueous phase, was determined for all combinations and attempts were made to correlate the R value to the HMF selectivity. The results are described in Figure 9, which evidently shows that a strong relation between HMF selectivity and the partitioning is absent. The effect of the temperature on the partitioning and the mixing of the solvents was not studied.
208| Chapter 2
Figure 9. Correlation of HMF selectivity and partition coefficient (R), and solvent type for the dehydration of fructose in saturated NaCl (pH=1 HCl, 180 °C). Legend for solvent types: ketones (Δ), ethers (○), secondary alcohols (◊), and primary alcohols (□). Open symbols correspond to solvents that are partially miscible with water and closed symbols correspond to completely miscible in water without presence of salts at room temperature. Reprinted from Ref.297 with kind permission from Springer Science+Business Media.
Recently, Menkhaus et al. reported the recovery of HMF from aqueous solutions using a polyethyleneimine (PEI), a soluble cationic polyelectrolyte.446 After separation of the PEI with adsorbed HMF from the aqueous phase using a membrane filtration (MWCO, 10000 cutof), followed by a centrifugation step and acidification with dilute sulfuric acid, up to 59% of HMF was recovered.
Isolation: Chromatography and adsorption A second approach to separate and isolate HMF from aqueous reaction mixtures involves chromatographic separations as disclosed in a patent assigned to Sueddeutsche Zucker AG.195 The HMF was separated from unreacted sugars and other decomposition products by the use of calcium loaded ion exchange columns such as Lewatit TSW 40. The chromatographic process was performed after the removal of solids and subsequent neutralisation. A large scale chromatographic separation trial was performed with water as the eluent, using three columns equipped with 13 m3 of Ca2+-resin at preferred temperatures of 55-85 °C. HMF- enriched fractions were separated and crystallised at temperatures below 20 °C to obtain HMF crystals with a purity exceeding 90%. Fructose and disaccharides were identified as
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|209 major impurities. In case the reaction mixture is not neutralised, a columns loaded with an ion exchange resins in acidic form can be used. Recently, a method for the isolation of HMF from aqueous reaction mixtures was described in a patent assigned to Archer Daniels Midland Company.447 It involves the use of non- functional polymeric resins to adsorb HMF. The authors demonstrated the concept with a 15.3 wt% HMF mixture containing fructose and furfurals using Lewatit S7768 resin. The adsorbed HMF was subsequently desorbed by elution with a volatile organic solvent such as acetone. After removal of acetone, the HMF-enriched fraction (82.9%) was recovered. Further purification was performed with Finex CS13GC293, a cation exchanger in acidic form, yielding HMF with a purity of 94.1%. In a another example furfural was first adsorbed from a 1 kg mixture containing 3.1 wt% of HMF, 18.5 wt% of fructose and 0.37 wt% of furfural, using Lewatit VP OC 1064. After furfural adsorption, fructose and HMF were subjected to a second separation step with the HMF selective Lewatit S7768 resin. After washing of the resin with acetone, 99.1% pure HMF was obtained. An attempt to adsorb in situ formed HMF in aqueous solutions using activated carbon was reported by Vinke et al.194 Isoalting HMF from a mixture obtained by acid-catalysed dehydration of fructose, an HMF yield improvement of up to 40% was obtained. About 90% of the adsorbed HMF was recovered by washing the loaded activated carbon with an organic solvent such as methanol or ethanol. Subsequent studies on the use of activated biochars as adsorbents for HMF from aqueous solutions were reported by Boihem et al.448 HMF is readily adsorbed on phosphoric acid and steam activated biochars from agricultural residues. Best results were obtained with steam activated biochars and these were shown to to be able remove 99% of the HMF (1 g/L) at a biochar intake of 2.5wt%.
Purification: Distillation After solvent removal, distillation under reduced pressure is typically used to obtain pure HMF. Table 39 provides an overview of the distillation properties of HMF. Teunissen performed the distillation at 114 – 117 °C at less than 1 mm Hg to obtain pure HMF by following a method described earlier by Middendorp.392,393 Cope applied distillation at 118 °C and 0.4 mm Hg to purify the product.311 The distillation of crude HMF is recognised as troublesome due to thermal degradation of HMF associated with the formation of tarry carbonaceous materials.449,450 Jones reported that the isolated yield of HMF from the crude HMF was only 45 – 60 % by distillation.450 Two
210| Chapter 2 patents have been published disclosing procedures to prevent further degradation of HMF during the distillation.449,450 Jones proposed to protect the crude HMF from moisture by degassing at a high vacuum and continuously distilling the HMF at higher temperature under pressure, claiming a pure HMF yield over 90%.450 To reduce contact time at elevated temperatures during distillation of crude HMF, Hunter proposed the use of a falling film evaporator.449 Moreover, the addition of a non-volatile flowing agent like PEG-600 was proposed to avoid the formation of tarry materials and deposits in the still during the distillation.
Table 39. Overview of HMF purification by distillation
Temperatures Pressures HMF Reference (°C) (mmHg) yield
114 – 117 < 1 - 392, 393
118 0.4 - 311
175 0.19–3.34 93%a 450
130 – 190 0.05–0.5 -b 449 a: moisture removal by degassing; b: falling film evaporator, additional flowing agent (PEG-600)
Purification: Crystallisation Crystallisation of HMF from concentrated HMF fractions in aqueous solution and DCM was reported in a Patent assigned to Südzucker AG.195,396 After obtaining a crude HMF fraction by a chromatographic process, this was concentrated by heating to 80 °C under vacuum until the remaining water-content was about 7%.195 The HMF was then crystallised by cooling the concentrated HMF fraction in a multistage stirred crystalliser. The mixture was cooled from room temperature to 10 °C with a cooling rate of 5 °C/h. When the mixture reached 10 °C, it was seeded with HMF-crystals, followed by further cooling to 4 °C. HMF crystals with a purity of 97% were obtained, and further recrystallisation yielded 99.4% of pure crystalline HMF.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|211
2.5.4.2 Separation and purification strategies for non-aqueous systems
When high boiling solvents such as DMSO or even non-volatile solvent such as ionic liquids are applied in the preparation of HMF, the non-isolated HMF yields are generally high. However, only limited attention has been paid to the separation of the HMF from the reaction mixture. M’bazoa et al. isolated HMF from a DMSO solution by means of solvent extraction with dichloromethane (DCM).396 In order to allow phase separation, a certain amount of water was added to the mixture. In this way 96% of the HMF was extracted to the DCM phase, along with a small fraction of DMSO. After solvent removal, HMF was purified by crystallisation. As part of a process to obtain high purity HMF (>99%) from the in situ formed intermediate AMF, Reichert and co-workers isolated and purified HMF from a reaction mixture obtained by hydrolysis of AMF to HMF in methanol.395 After removal of impurities and volatiles, HMF was crystallised from a methyl tert-butyl ether (MTBE) solution by cooling it slowly to 5 °C. Reports on isolation and purification of HMF from ionic liquid systems are hardly available. A wash step with an organic solvent such as ethyl acetate has been successfully tested by Qi et al.347 Hu et al. reported HMF synthesis from fructose in a biphasic system of ethyl acetate and choline chloride (ChoCl) with citric acid.335 Chan et al. studied the continuous extraction of HMF by THF from tungsten salt catalysed reaction systems in [BMIm]Cl.451 The extract phase was concentrated under vacuum at room temperature, yielding 80% HMF the purity of which was not reported. A novel entrainer-intensified vacuum reactive distillation process for the separation of HMF from ionic liquids and particularly [OMIm]Cl was proposed by Ren et al.452 It involves heating (batch) reaction mixtures under vacuum (ca 300 Pa) to 150 -180 °C under a flow of an entrainer (nitrogen). After 10 min at 180°C, 95% of the HMF was recovered from the reaction mixture. The system was also tested for the integration of the dehydration reaction and susbsequent HMF separation. Recently, Liu and co-workers disclosed a patent application on a process for the recovery of HMF from an ionic liquid system by column adsorption.453 A schematic overview of the adsorption step is presented in Figure 10. Various adsorbents have been examined, ranging from activated carbon and carbon molecular sieves to zeolites, silica based materials and ion exchange resins. The column is fed with the reaction mixture (including an ionic liquid) and de-ionised water in order to adsorb the HMF. The loaded column is subsequently eluted with
212| Chapter 2 water and/or an alcohol in order to wash of the HMF. The ionic liquid and the unconverted feed are recycled after water removal.
Figure 10. A schematic overview of the adsorption step used by Liu et al. Taken from Liu et al.453
2.5.5 Pilot Scale Production of HMF
Several studies on pilot scale production of HMF have been reported and will be discussed in this paragraph. It details the challenges for further HMF development activities at larger (commercial) scale. Aqueous phase, non-aqueous phase and mixed solvent reaction systems are discussed separately.
2.5.5.1 Pilot scale studies of aqueous HMF processes
Teunissen prepared HMF by the acid catalysed dehydration of sugar cane on 1 kg scale.393 Yields of 40 – 50 g of pure HMF were reported. The reaction and subsequent purification of HMF were performed following earlier methods described by Düll,6 Kiermayer,7 Blanksma454 and Middendorp.392 Teunissen heated 30 wt% aqueous sugar cane in the presence of 0.3 wt% oxalic acid in an autoclave at 134 °C for 3 h . The reaction mixture was then neutralised with CaCO3, followed by HMF extraction with ethyl acetate. The solvent was evaporated and HMF was subsequently distilled under reduced pressure (<1 mm Hg) at 114 – 117 °C.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|213
Pilot scale production of HMF in an aqueous reaction system was disclosed in a patent issued by Süddeutsche Zucker-aktiengesellschaft in 1988.195 The patent reports the multi-kg scale manufacture of HMF involving chromatographic separation and crystallisation to obtain 99% pure HMF in crystalline form (Figure 11). In one of the examples in the patent, a mixture of 11 kg fructose and 33 L of water was heated in a stirred batch autoclave at 135 – 142 °C for 3 h in the presence of oxalic acid. After the reaction mixture was cooled, the solid material was filtered in an air pressure filter. About 1.3 kg of solid residue was obtained from this filtration step. The filtrate was neutralised to a pH of 5 using CaCO3. Analysis of the filtrate revealed 55 % fructose conversion and 33 % (2.58 kg) of HMF yield. The isolation of HMF from the filtrate was performed by column chromatography using a calcium loaded ion-exchange resin. Four fractions were obtained: salts, a mixed fraction, unreacted fructose and HMF. The mixed fraction was again separated, the unreacted fructose was re- used, and the HMF fraction was concentrated. Two large HMF batches (143 kg and 176 kg) were concentrated at temperatures below 80 °C under vacuum until the remaining water content was less than 7%. Subsequently, the concentrated HMF solution was crystallised in a crystalliser. The mixture was cooled down from 20 °C with a cooling rate of 5 °C/h. When the mixture temperature reached 10 °C, the mass was seeded with HMF-crystals and further cooled to 4 °C. 124 kg of crystalline HMF (97%) was obtained from processing the two batches. Further re-crystallisation yielded 82.7 kg of crystalline HMF with a purity of 99.4%. The patent also discloses an example for the use of a polymeric fructose rich feedstock in the form of chicory roots. About 20 kg of chicory roots were mixed with 20 kg of water and then acidified with sulfuric acid to a pH of 1.8. The mixture was heated in an autoclave at 140 °C for 2 h. The reaction mixture was cooled to 70 °C and filtered in a pressure filter. This process yielded 13% HMF and 30% fructose, based on the available amount of inulin in the chicory roots. Two different chicory roots pretreatment were conducted before hydrolysis. Inulin was first extracted from chicory roots in a counter current extraction apparatus (DDS- extractor) at 70 °C, yielding an extract with 16% inulin. The second pretreatment involved liquefaction/pre-hydrolysis of the chicory-roots with 20% sulfuric acid at 80 °C for 2.5 h and subsequent heating by introducing 6 bar of steam for 1 h. After that, the pressure was released and the mixture was filtered before further hydrolysis. Unfortunately, the yield of HMF using these pretreatment methods was not disclosed.
214| Chapter 2
Figure 11. HMF production by Suedzucker195
2.5.5.2 Pilot scale production using non-aqueous solvents
A pilot scale batch production process for HMF in DMSO was patented by Furchim, France.396 In the absence of catalyst, 0.9 kg of fructose was reacted in 2 kg DMSO at 160 °C for 4 h, yielding 85% HMF at full conversion. Subsequently 2.7 kg water was added to the reaction mixture before extraction with 6.5 kg of DCM in a counter-current column at 20 °C. Around 96% of the HMF was extracted together with 0.05 kg of DMSO. The extraction phase was concentrated by evaporation of the DCM, followed by crystallisation (rapid cooling from ambient temperature to 5 °C, and then to -5 °C at rate of 4-5 °C/h). In total, 0.5 kg of 98% pure crystalline HMF was obtained from processing 0.9 kg of fructose, adding up to a molar yield of over 75%.
2.5.5.3 Pilot scale production using mixed solvent systems
Cope reported the production of HMF from a 50 wt% aqueous sucrose solution (3 kg) without catalyst using MIBK as an in situ extraction solvent.311 The reaction was performed at 160 °C for 9 h with about 60 L of MIBK to yield around 63% HMF. After neutralisation of the MIBK phase with K2CO3, MIBK was evaporated using a Turba-film rotavapor at a pressure of 180 mmHg, with a steam pressure of 3 psig. The HMF was isolated after a vacuum distillation at 118 °C and 0.4 mmHg in a batch laboratory setup. The isolated yield and HMF purity were not reported. Using a similar approach with glucose, the yield of HMF in the MIBK phase was 21 to 25% after 9 h reaction time.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|215
A patent assigned to Roquette Frères, France discloses the kg-scale production of HMF from fructose in a water-MIBK biphasic system in the presence of cationic resins at relatively low temperatures of 70-95 °C.455 The reaction was performed with 1 kg of fructose in 4 L of water using 36 L of MIBK in a 50 L stainless steel reactor with an overhead stirrer. After dissolution the reaction was performed at 89 °C for 5 h in the presence of 0.6 kg of cationic pre-dried Lewatit SPC 108. Subsequently, the reaction mixture was filtered and the aqueous and organic phases were separated. The aqueous phase was analysed and contained residual fructose and small amounts of HMF and levulinic acid. The organic phase was rich in HMF and also contained some levulinic acid, adding up to an HMF yield of 38% at 51% fructose conversion. Further isolation of HMF from the organic phase and subsequent HMF purification were not described.
2.5.5.4 Pilot scale production of HMF involving an HMF derivative
In this section, studies will be reported involving the (in situ) formation and subsequent purification of HMF derivatives followed by a final conversion to HMF. The main idea is that the derivative is obtained in higher isolated yields than HMF itself due to both higher reaction yields as well as improved isolation/purification yields. As such, it resembles the classical protective group approach in synthetic organic chemistry. Evonik Degussa Gmbh. performed a kg-scale production of high purity HMF (>99%) using 5-acetoxymethylfurfural (AMF) as the HMF derivative.395 First about 39.4 kg of fructose in 90 L NMP was reacted in the presence of 5.9 kg Dowex 50WX8-200, an acidic ion exchange resin, at 110 °C for 6 h to form HMF. After cooling, the reaction mixture was filtered and washed with 8 L NMP. Subsequently, the filtrate was mixed with 4-(N,N- dimethylamino)pyridine and acetic acid anhydride at 25 °C . After 1 h reaction, the solvents were removed from the reaction mixture at 90-110 °C at 50-100 mbar. The residue was subsequently mixed with 160 L MTBE, 60 L water and 4 kg activated carbon. The suspension was filtered and phase separated. The solvent was removed from the filtrate by distillation at 50 °C under reduced pressure (20 mbar). The residue was further purified by fractional distillation at 106-110 °C and 5 mbar. About 15.5 kg of AMF (42%) was obtained after solvent removal (distilled at 50 °C under pressure) and subsequent fractional distillation at 106-110 °C at 5 mbar. About 7 kg high purity HMF (<99 %) was obtained by the hydrolysis of 10.9 kg AMF in 60 L methanol in the presence of 1.1 kg Amberlyst A26 OH, a strongly basic ion exchange resin, at 25 °C for 1 h. After the reaction, the mixture was contacted with activated carbon for
216| Chapter 2
1 h, and subsequently filtered and washed with 10 L methanol. The filtrate was concentrated at 40 °C under reduced pressure. The residue was mixed with 8 L MTBE and subsequently cooled slowly to 5 °C to precipitate the HMF. The crude HMF was washed with 1.5 L ice cold MTBE and dried at 20 °C (under pressure). About 85% of HMF was obtained from AMF using this method.
2.5.6 Techno-economic evaluations of different modes of HMF production
Evaluation of the pilot scale production processes for HMF indicates 2 examples using homogeneous catalysts (oxalic acid), one with a heterogeneous catalyst (a cationic resin) and 2 without the use of a catalyst. The use of a catalyst is considered advantageous regarding activity and selectivity, though the exact effect is a function of the solvent composition. The typical homogeneous catalysts (oxalic acid, mineral acids) have the advantage that they are relatively cheap, though recycling is often cumbersome. In most pilot studies, a neutralization step is incorporated using a base and the resulting salts are discarded, which is evidently not a good example of green chemistry and technology principles. In this respect, the use of heterogeneous catalysts has certain advantages. However, catalyst lifetime needs to be established and a major concern is the deposit of humic substances on the heterogeneous catalysts leading to irreversible catalyst deactivation. Detailed, long duration experiments will be required to address this issue and assess the true potential of heterogeneous catalysts. To the best of our knowledge, these are not available in the open literature yet. Scattered data are available in the open literature on the manufacturing costs of HMF. An overview of reported data is shown in Table 40. For HMF production from fructose in an aqueous reaction system patented by Südzucker AG, an HMF manufacturing price of 6 DM/kg was reported when fructose is available at a price of 0.5 DM/kg (1988). Bicker et al. investigated the production of HMF from fructose in supercritical acetone.233 The process consisted of a flow-reaction system, solvent recovery, and further HMF purification by chromatographic methods. Based on this process, an HMF cost price of 2 €/kg was estimated, using the assumption that fructose is available at 0.5 €/kg.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|217
Table 40. Estimated HMF manufacturing costs
Process Design Capacity Fructose HMF Reference Price Price aqueous process - 0.5 DM/kg 6 DM/kg 195, referenced by Kuster12
Low boiling - 0.5 $/kg 2 €/kg 206 solvent: acetone
Two solvents: Production of 0.55 $/kg 2.16 $/kg 456 MIBK and 2- HMF: 7000 butanol Ton/year two solvents: 0.55 $/kg 1.97 $/kg 456 water and THF two solvents: 0.55 $/kg 2.43 $/kg 456 water and MIBK/2-butanol two solvents: Feed rate of 0.3 $/kg 1.33 $/l 457 water and 1- fructose: 300 (1.08 butanol Ton/day $/kg) (49.2 mega l/year)
Recently, two techno-economic evaluations on the production of HMF using a biphasic reaction system were published.456,457 Torres et al. performed a design case on the continuous production of HMF employing MIBK and 2-butanol as the extraction solvents.456 The process design involves a liquid-liquid reactor coupled with an extractor and evaporator (Figure 12). The authors developed a mathematical model of the process consisting of mass balances, equilibrium relations and design constraints based on published work by Roman- Leskhov et al.170 A first order kinetic model proposed by Kuster and Temmink was used as input in the reactor model.157 The manufacturing costs of HMF production were estimated at 2.16 $/kg based on a fructose price of 0.55 $/kg. The fructose price was shown to be the
218| Chapter 2 major factor determining in the HMF cost price at an 80% contribution. The authors suggested also possible process improvements such as the use of extraction solvents with a higher partitioning coefficient for HMF (e.g. THF). In a follow up paper,458 an alternative flow sheet was proposed involving combination of the fructose dehydration reaction and HMF separation in a single unit. The minimum HMF costs were estimated and found to be between 0.21 and 0.24 $/mol.
Figure 12. Process diagram for the production of HMF in biphasic systems (Legend: vi denotes molar flow rate in the stream i; FJi stand for the molar flow of component J in stream i, A: fructose, BPA: byproducts from fructose, B: HMF, C: levulinic acid, D: formic acid, BPB: other decomposition products from HMF, W: water, S: solvent) Reproduced from Ref. 456 with permission of the Royal Society of Chemistry.
As part of a techno-economical evaluation of the production of dimethylfuran, a HMF derivative with potential to be used as a biofuel (component), Kazi et al. reported a techno- economic analysis of the production of HMF.457 A conceptual plant design, combined with the production of 2,5-dimethylfuran, was proposed based on a biphasic reaction system of water saturated with NaCl and 1-butanol (Figure 13).
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|219
Figure 13. Conceptual design for the production HMF and DMF. Adapted from Ref.457 with permission from Elsevier.
The design incorporates a reaction section, HMF purification, reactant and solvent recycle and isolation of valuable by-products such as levulinic acid. For a 300 metric ton/day of fructose input, the minimum cost price of HMF in this highly selective and efficient process was estimated at 1.08 $/kg. The HMF cost price is very sensitive to the feedstock cost, but also dependent on the HMF yield, by-product price, catalyst cost and total purchased equipment cost (in that order). A price of around $ 1.00/Kg of HMF would seem to be a good basis for its use in bulk- scale applications as this is the same order of magnitude as current fossil fuel based raw materials. In order to reach such a price level the plant size should be at least 100 kTon/y. Unfortunately, most bulk products that can be made from HMF, such as FDCA are not drop- ins, i.e. they cannot replace identical existing bulk chemicals, but rather need to find their own place in the market, which may require substantial time and normally goes through a phase were the new products are produced on sub-optimal scale at too high costs. Nevertheless, this process could be accelerated by the finding of favourable properties of the final (polymeric) material. In addition, a strong market pull, caused by a desire for the use of
220| Chapter 2 products based on renewable raw materials and a desire to bring down the carbon footprint of the current product can help to overcome this small volume high price dilemma. And indeed, Avantium has shown that the polyester PEF, based on biobased FDCA and biobased ethylene glycol has both superior barrier properties and a higher Tg than the petrochemical analogue PET. The company recently announced the opening of pilot-scale production of FDCA. Avantium has also announced partnerships with The Coca-Cola Company and Danone. Here, the prime objective is to replace PET for bottle applications, which would allow for immediate large-scale production. Production of HMF for fuel applications is not an easily obtainable target. Here the price should be substantially lower. This would require the direct production of HMF in good yields from cellulose or even lignocellulose. Although proof-of- principle has been obtained for such conversions, the reported methods are still far removed from an industrial process.
2.6 The relevance of 5-hydroxymethylfurfural as a platform chemical
HMF is a very important building block for a wide range of applications. In section 2.5.2.4 the in situ formed derivatives were discussed of which levulinic acid is the most important one. In this Chapter other applications in the areas of polymers (section 2.6.1), fine chemicals (section 2.6.2) and fuels (section 2.6.3) are discussed.
2.6.1 Conversion of HMF to monomers for polymers
When HMF is produced at high efficiency follow-up products will become an attractive option to replace petrochemical analogs.459,460 In addition to FDCA, as mentioned in section 2.1, other platform chemicals can be produced as well. 5-Hydroxymethylfuroic acid, 2,5- diformyl furan, the 2,5-di-aminomethylfuran and 2,5-bishydroxymethylfuran are most versatile intermediate chemicals of high industrial potential because they are six-carbon monomers that could replace e.g. adipic acid, or alkyldiols, or hexamethylenediamine in the production of polymers.13,459-462
2.6.1.1 HMF-based diols
Utne described the hydrogenation of HMF at 350 bar hydrogen pressure and 150 °C to 2,5- bishydroxymethyl-furan (5) using copper chromite as a catalyst.463 The product was obtained
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|221 in 100% yield. Surprisingly, if the pressure and the temperature were increased somewhat, a mixture of products was obtained containing 2,5-bishydroxymethyltetrahydrofuran (50), 1,2,6-hexanetriol (51), 1,2-hexanediol (52) and 1,6-hexanediol (10, Scheme 34). The relative amounts of these could be varied by subtle changes in the solvent and the temperature. The formation of polymers based on HMF-derived monomers was recently reviewed.464
Scheme 34. Reduction of HMF
For synthetic purposes a NaBH4 reduction of HMF to 5 can be quite convenient. This reduction was reported by Cram and co-workers to proceed in 92% yield.465 Descotes performed a systematic study of the use of several different metal catalysts for the hydrogenation of HMF.466 He was able to convert HMF to 5 in 100% selectivity using either
10 mol% of 2CuO.Cr2O3 at 60 °C, 1 mol% Pt/C at 30 °C, or 1 mol% PtO2 at 60 °C. Battelle researchers reported the use of Raney-Co (97% selectivity), Co/SiO2 (96% selectivity) and
Pt/Al2O3 (98% selectivity) as catalysts for the hydrogenation of HMF to 5 at 35 bar H2, 60 °C; water was the best performing solvent.467 Use of Ra-Ni led to unselective reactions due to over-reduction to 50, whereas use of Ra-Cu gave poor selectivities. Heeres, de Vries and co- workers reported the use of bimetallic Ni-Cu catalysts on zirconia and Ru on alumina as catalysts for the hydrogenation of HMF to 5 with good selectivity.468 Enzymatic methods have also been reported for the conversion of HMF to 5. Whole bacteria where used by Boopathy and co-workers and reported to give full conversion to 5.469 Ras has done an extensive screening study on the reduction of EMF, the ethyl ether of HMF, into the corresponding alcohol. It was found that Pt and Rh supported on silica are efficient catalysts for the selective hydrogenation of EMF to the unsaturated alcohol, obtaining 100% selectivity for conversions up to 66%.470
222| Chapter 2
HMF can be quite toxic for microorganisms. This is particularly problematic in industrial yeast fermentations towards ethanol where the presence of HMF can slow down the process. Liu reported the development of a strain of Saccharomyces cerevisiae capable of reducing HMF to 5, which is not an inhibitor.471 The diol 5 has been converted into polyethers that are used as components in polyurethane foams.472 Hales reported the use of a supported nickel catalyst that also contained small amounts of copper and iron. Using this catalyst in water as solvent at 105-140 bar of H2 at 70-100 °C, he obtained 50 in 97% yield. Researchers from ADM reported the hydrogenation of HMF to 50 in its cis-form using a heterogeneous nickel catalyst doped with zirconium at 103 bar and 200 °C in ethanol. However, as much as 10 mol% of catalyst was necessary.473 Probably the best catalyst for the hydrogenation of HMF to 50 is Raney-nickel. Virtually quantitative yields have been reported using this catalyst by Descotes,466 Connolly474 and de Vries, Heeres and co-workers 468 Dumesic and co-workers have tested supported Ru, Pd, and Pt catalysts in monophasic and biphasic reactor systems to determine the effects of the metal, support, solution phase acidity, and the solvent to elucidate the factors that determine the selectivity for hydrogenation of HMF to 50. Highest yields (88–91%) were achieved using Ru supported on materials with high isoelectric points, such as ceria, magnesia–zirconia, and γ-alumina.475 Both 5 and 50 have been used as monomers for polyesters.14,476 Reaction with glycerol (Scheme 35) is another method for converting HMF into a diol (53) that can be used as a monomer. Use of 53 for the production of polyesters and polyurethanes may lead to materials 477 with a high tg by virtue of the presence of the two rigid rings. Similarly, it is possible to convert HMF into an aldehyde acid by its reaction with a cyclic anhydride such as phthalic anhydride or maleic anhydride.478
Scheme 35. Acetalisation of HMF with glycerol
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|223
2.6.1.2 2,5-Diformylfuran
2,5-Diformylfuran (DFF, 54) is an attractive building block that can be converted into a host of new products. DFF has been polymerised to polypinacols and polyvinyls; it has been used as a starting material for the synthesis of anti-fungal agents, pharmaceuticals and ligands.392,410 It is also possible to decarbonylate DFF to furfural or to furan.479 Many methods have been reported for its synthesis from HMF via oxidation (Scheme 36). The stoichiometric oxidation of HMF using potassium dichromate in DMSO under ultrasonic 480 irradiation at room temperature gave DFF in 75% yield. Use of MnO2 in refluxing CH2Cl2 led to a very clean conversion of HMF to DFF although only 80% conversion was reached 481 after 8h. HMF was efficiently oxidised to 54 in 63–89% yield using a Mn(III)–salen catalysts and sodium hypochlorite as oxidant in a pH 11.3 phosphate buffer–CH2Cl2 biphasic system at room temperature.482 Lee and co-workers claimed a yield of 100% of DFF upon stoichiometric oxidation of HMF with immobilised IBX (2-iodoxy-benzoic acid).483
Scheme 36. Manganese-catalysed oxidation of HMF to DFF
A number of catalytic oxidations have been reported using various metal catalysts and oxygen or air as oxidants (Scheme 37). Use of catalytic amounts of TEMPO and CuCl2 in 484 DMF with O2 led to the formation of 54 in 55% yield. Use of the acetamino variant of
TEMPO was also reported in conjunction with HNO3/O2 as oxidant; the authors claim a 100% yield to DFF.485 The use of vanadium catalysts has been reported a number of times. 486 Use of V2O5 on TiO2 in toluene at 170 °C led to the formation of 54 in 64% yield. With the same catalyst at higher air pressure (16 bar) and lower temperature (90 °C) a much higher selectivity of 97% was obtained, however, relatively large amounts of catalyst were used.487
224| Chapter 2
The effect of the carrier material on the structure of the VOx domains has been investigated.488 The results can serve as a guidance for the development of better catalysts. Grushin and co-workers tested a range of different vanadium salts and complexes in DMSO at 150 °C. They were able to combine two steps by first converting fructose into HMF using a Dowex type acidic ion exchange resin at 110 °C for 5h (85% selectivity to HMF) followed by removal of the acidic catalyst and oxidising in the presence of the vanadium salt. Use of 5 489 mol% of VOHPO4•0.5H2O led to the formation of 54 in a combined 41% yield. They also showed that the combination of PdCl2/CuCl2 is an excellent catalyst for the conversion of HMF to DFF in near 100% selectivity. However, the catalysis stalls after about 70% conversion, which the authors ascribe to the presence of water.
Intererstingly, Xu and co-workers showed that upon use of VO(acac)2 as catalyst at 90 °C oxidation of HMF at 10 bar gave up to 52% of maleic anhydride in addition to smaller amounts of DFF.33 Very high yields of DFF where obtained upon use of a mixed catalyst 33,490 consisting of Cu(NO3)2 and VOSO4. Chinese researchers were capable of producing DFF from glucose in a 2-step sequence.
First they reacted glucose in DMA at 100 °C in the presence of catalytic CrCl3•6H2O and NaBr for 6 h to give a crude mixture containing 74% of HMF. This mixture was oxidised with air in the presence of catalytic NaVO3•2H2O at 110 °C for 10 h to give DFF in 51% yield.491
Scheme 37. Metal catalysed oxidation of HMF to DFF in air
Carlini and co-workers also attempted the direct conversion of fructose to DFF by using the same vanadium catalyst both for the dehydration step of fructose to HMF and as oxidation 492 catalyst. The catalyst VOPO4∙2H2O was quite effective in the dehydration of fructose in water. After 2h at 80 °C 50% conversion was reached with a selectivity to HMF of 82%. The use of biphasic MIBK/H2O mixtures reduced the selectivity to HMF somewhat to 70-74%.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|225
Oxidation of HMF in water using the same catalyst at room temperature and 1 bar O2 was very slow with only 3% conversion, albeit with 100% selectivity to DFF. Use of biphasic
MIBK/H2O mixtures led to somewhat higher conversions of HMF, but the selectivity to DFF deteriorated strongly with 5-acetoxymethyl-2-furfural (AMF) and a dimeric ether (OBMF) as main side products. Oxidation of HMF in DMSO at 150 °C led to 84% conversion after 6 h and formation of DFF with 97% selectivity. However, as the product is not easily separated from DMSO the authors also investigated the use of other solvents. Whereas use of MIBK, toluene or benzene led to poor selectivity to DFF, use of DMF at 100 °C gave a 56% conversion of HMF with 93% selectivity to DFF after 6 h. At higher temperatures the selectivity started to deteriorate at higher conversion. Partial substitution of the vanadium in
VOPO4∙2H2O by other metals led to lower selectivities. Use of the vanadium complex (8- hydroxyquinoline)2V(O)OiPr in the oxidation of HMF with air gave DFF in 94% yield.493 Corma and co-workers tested a series of metal catalysts supported on either polyvinylpyridine, cross-linked with polystyrene or on aminopropylated zeolite (SBA-15) that had been treated with 4-pyridinecarboxaldehyde. In the initial tests using the soluble metal salts they obtained some very interesting results (Table 41).
Table 41. Oxidation of HMF to dialdehydes
Catalyst T (°C) t(h) Conv of Selectivity (%) HMF (%) 54 7 48
CuCl/PdCl2 160 10 85 97 2.5 0.5
115 24 20 >99 0 0
CuCl 160 24 26 97 2.5 0.5
CuCl2 160 8 65 98.5 1.5 0
a Reaction conditions: air flow rate 0.5 ml/s, DMSO (7ml), 0.3g of HMF, HMF/Metal = 10, P = 0.1 MPa
226| Chapter 2
After 10 h up to 85% conversion of HMF was achieved using CuCl/PdCl2 as catalyst and DFF was obtained with 97% selectivity. However, the authors note that DMSO is oxidised to dimethylsulfone in substantial amounts. Use of the immobilised salts led to comparable activities and selectivities. However, addition of extra pyridine was necessary to retain the activity of the catalyst upon reuse. The immobilised catalysts also functioned reasonably well in toluene and very well in trifluorotoluene. Partenheimer and Grushin have applied the catalyst system that the former had developed for the commercial scale oxidation of xylene to phthalic acid. Thus, use of mixtures of cobalt, manganese and zirconium acetates with bromide allowed the oxidation of HMF to DFF in 57% isolated yield at 1bar oxygen pressure.494 Increase of the pressure to 70 bar gave DFF in 61% yield. More recently Sanborn from Archer Daniels Midland found that oxidation of HMF in acetic acid using cobalt and manganese acetate as catalyst gives up to 86% selectivity to DFF when the reaction is performed without bromide, but in the presence of an equivalent of MIBK.495 Indian researchers claimed 100% selectivity to DFF using a catalyst made from calcined manganese mesoporous material substituted with silver in isopropanol at 145 °C.496 Direct formation of DFF from fructose and glucose was alos reported using a mixed catalyst system of Amberlyst-15 and ruthenium supported on hydrotalcite. Yields of DFF were low to moderate.497 Electrochemical oxidation of HMF allowed the isolation of DFF in <68% yield.498
HMF has also been oxidised using the enzyme chloroperoxidase and H2O2 as oxidant. A selectivity of 60-74% to DFF was reached with FDCA and 5-formyl-2-furancarboxylic acid as main side products. The reaction is relatively fast with 87% conversion obtained after 21 min.499 The use of DFF was reviewed by Lewkowski.13
2.6.1.3 2,5-Furandicarboxylic acid (FDCA), adipic acid and 5- hydroxymethyl-2-furan carboxylic acid
Another interesting molecule that can be derived from HMF is 2,5-furandicarboxylic acid (FDCA, 3). It can be obtained via the oxidation of HMF. FDCA was identified by the Department of Energy, USA, to be a key bio-derived platform chemical which can serve as starting point for several interesting molecules including, succinic acid, 2,5-furandicarboxylic acid dichloride and 2,5-furandicarboxylic acid dimethyl ester.1,2 2,5-furandicarboxaldehyde and 2,5-hydroxymethylfuroic acid can be considered intermediates to FDCA in the oxidation
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|227 of HMF and were discussed before. Recently, several new applications of FDCA and its diester have emerged, which will be discussed below. Several oxidation methods have been described in the literature (Table 42).
Table 42. Oxidation of HMF to FDCA
Reaction conditions HMF conversion (%) FDCA Yield (%) Ref.
N2O4 in DMSO 100 24 196
Ag2O and HNO3 100 47 196
KMnO4 100 70 196
Pt/Al2O3 (basic conditions) 100 500
Co/Mn/Br (Zn), air 91 61 494
Co/Mn/Br with Mol Sieves in HOAc 70 501
Co(acetylacetonate)-SiO2 (from Fructose), air 72 99 502
Pb-Pt/C (NaOH), air 100 81 503,504
Pt/C, air 100 95 505
Pt/C, air 100 91 506
Pt-ZrO2; Pt/Al2O3, air 100 98 507,508
PtBi/C (from fructose, solid acid, H2O/MIBK), air 50 25 509
Au/TiO2 (basic conditions/O2) 100 71 510
Au (hydrotalcite), air 100 100 511
Pt/C, Pd/C, Au/C, Au/TiO2, air 100 79 512
NiO2/OH anode (electrochemical oxidation) 71 513
Ru on several supports - 20-100 514,515
Bimetallic Cu-Au nanoparticles on TiO2 99 516
228| Chapter 2
HMF oxidation into FDCA (3, Scheme 38) was achieved with several different 517 stoichiometric oxidants like N2O4, HNO3 and KMnO4. The oxidation of HMF using air or oxygen has been described using several different catalysts. The catalysts which is currently used for terephthalic acid production (Co/Mn/Br) was used for oxidation under high pressure (70 bar air).494,501 Similar processes have also been patented.495,518,519 Heterogeneous catalysts also resulted in FDCA via HMF oxidation with molecular oxygen. Supported platinum catalysts were first used in the presence of base, resulting in near quantitative FDCA yields. The base is used in stoichiometric amounts to keep the FDCA formed in aqueous solution as the di-alkaline salt.500,520 Direct synthesis routes of FDCA from fructose by combining dehydration and oxidation have also been reported using Pt-Bi/C in combination with a solid acid in water/MIBK. An FDCA yield of only 25% was obtained, constituting a 50% selectivity.509 Ribeiro reported the direct conversion of fructose to FDCA with high conversion and excellent selectivity (99%) using Co(acac)–SiO2 as bifunctional catalyst at 160 °C and 20 bar air pressure.502 Recently, two examples were reported using supported 510 gold catalysts for aqueous HMF oxidation. Gorbanev et al. demonstrated that Au/TiO2 could oxidise HMF into FDCA in 71% yield at near room temperature.510 Casanova et al. 521 showed that Au/CeO2 was more active and selective. Similar to the Pt systems reported by Vinke500 the addition of homogeneous base (1–20 equiv. NaOH) and high oxygen pressure (10–20 bar) are required. This was corroborated in recent research from Davis and co- workers.522 Surprisingly, Saha and co-workers recently reported that addition of trifluoroacetic acid to the oxidation of HMF had a positive effect on yield and selectivity, both with Co(OAc)2/Zn(OAc)2/NaBr and with gold nanoparticles immobilised on CeO2 or 523 TiO2 as catalyst. Davis and co-workers compared a number of different catalysts and found that Pt/C and Pd/C oxidised HMF to FDCA an order of magnitude faster than Au/C or 524 Au/TiO2. Recently, Gupta et al. reported the base free oxidation over gold catalysts supported on hydrotalcites.511 Since the FDCA formed is an acid, it can be expected that a reaction between the basic hydroxyl groups of the hydrotalcite and the product formed will take place. Riisager and co-workers used Ruthenium on several different supports as catalysts in the oxidation of HMF (O2, H2O). Best results were obtained using a basic support, such as MgO; however, the authors found that the metals were leached from the basic support.514,515
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|229
Scheme 38. Oxidation of HMF into FDCA
Avantium reported the oxidation of ethers of HMF using a Co/Mn/Br catalyst. The FDCA monoester thus obtained was esterified further to obtain the diesters such as 55.525 Furthermore Avantium Chemicals have reported the use of FDCA-ethylene glycol polyesters starting from the FDCA dimethyl ester 55. Though the polymer was already described by Drewitt and Lincoln in 1946526 real progress was only achieved recently. Gruter and co- workers527 and Gandini476 have recently published work on FDCA-based polyesters. Gandini has prepared extensive reviews on the synthesis and initial characterisation of a wide range of FDCA based polymers.461,462 Avantium has been able to produce multi-kilogram quantities. No details have been reported for the production of this diester, but it highlights the growing interest in these kinds of compounds. An interesting application of this polyester would be the replacement of PET in soft drink bottles. In view of the large size of the market, this 528 could have an enormous impact on energy use and CO2 emissions. The interest in green polymers is also exemplified by a patent of Evonik529 that claims the use of iso-decanol ethers of FDCA as good polymer plasticisers. Esterification of FDCA to obtain esters has been described by Gaset.530 Taarning et al. reported production of the diester 55 (dimethyl
2,5-furandicarboxylate) from HMF through oxidative esterification using Au/TiO2 as catalyst in a sodium methoxide-containing methanol solution under oxygen pressure.531,532 Casanova et al. demonstrated base-free oxidative esterification of HMF into the diester using Au/CeO2 as catalyst in methanol under 10 bar oxygen in an autoclave reactor (Scheme 39).521
Scheme 39. Formation of FDCA di-esters from HMF
From FDCA several other compounds can be made that find application in polymer applications. A useful monomer that can be obtained from FDCA is adipic acid (9, Scheme 40), one of the two monomers used in the nylon 6,6. Boussie described the hydrogenation of FDCA to adipic acid in two steps.533 In the first step 50 was produced in 88% yield by
230| Chapter 2 hydrogenating FDCA at 140 °C for 3 h in acetic acid, catalysed by Pd on silica. Yields up to 99% of adipic acid were claimed by reacting 50 under hydrogen at 160 °C for 3 h in acetic acid in the presence 0.2 M of and 5% Pd on silica catalyst.
Scheme 40. Conversion of FDCA into adipic acid
Sato described the preparation of diisocyanates.534 Benecke described the reaction with aromatic amines and their uses as cross-linking agents for polyureas, hybrid epoxy-urethanes, hybrid urea-urethanes, chain extenders for polyurethane and polyurea elastomers.535 Improved hydrogenation of the furan ring to tetrahydrofuran cis-2,5-dicarboxylic acid has been described by Moore.536 Amides and their subsequent hydrogenation to amines have been described by Mndzhoyan 537 A highly selective oxidation of HMF to 5-hydroxymethyl-2-furan carboxylic acid 4 using whole cells of Acetobacter rancens or Serratia liquefaciens was reported by Nagasawa et al. (Scheme 41).538 Thus 2.6 mmol of HMF was oxidised in 26 h using 182 mg of whole cells of S. liquefaciens with 97% conversion.
Scheme 41. Enzymatic oxidation of HMF
Koopman has described a novel HMF/furfural oxidoreductase from Cupriavidus basilensis that converts HMF into FDCA. The oxidoreductase gene was introduced into Pseudomonas putida S12. In fed-batch experiments using glycerol as the carbon source, 30 g/l of FDCA was produced from HMF at a yield of 97%.539 This approach has also been patented.540,541 The same authors recently published a review regarding microbial metabolism of HMF and other furans.83 Riisager and co-workers have reported the oxidation of HMF in methanol using supported gold nanoparticles as catalyst. The resulting methyl 5-hydroxymethyl-2-furancarboxylate was reacted in situ with hexyalamine to form the n-hexylamide.542
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|231
The oxidation of HMF to FDCA was recently reviewed.543
2.6.1.4 Conversion of HMF into other monomers
HMF has been converted into caprolactam, the monomer for nylon-6, by de Vries, Heeres and co-workers (Scheme 42).468 HMF was hydrogenated in >99% yield to 50. They were able to hydrogenate 50 to 1,6-hexanediol in 86% yield, by using a Rh-Re/SiO2catalyst in the presence of nafion. This is in fact a tandem three step process proceeding through formation of 1,2,6-hexanetriol, which is cyclised under the influence of the acid to tetrahydropyran-2- methanol (56), which in turn is hydrogenated to 1,6-hexanediol using the same catalyst. The diol was converted into caprolactone using a ruthenium-catalysed Oppenauer oxidation in virtually quantitative yield. Overall selectivity from HMF to caprolactone was 86%. Conversion of caprolactone into caprolactam using ammonia has been practiced on production scale in the past, so this constitutes an interesting route from HMF to Caprolactam in only 4 steps, whereas the current industrial process, which is based on benzene, contains seven steps. Scheme 42 provides an overview of the conversion of 50 into caprolactam.
Scheme 42. Conversion of 50 into caprolactam
Another outlet for HMF is the synthesis of novel biomass-based vinyl polymers. HMF or its methylated derivative MMF was efficiently converted to its vinyl derivative (57 and 58)
232| Chapter 2 by the Wittig reaction in a solid-liquid phase transfer process (Scheme 43), followed by free radical polymerisation in a bulk.544 Especially the MMF derived polymers 59 showed good thermal stability.
Scheme 43. Vinyl furans from HMF and their polymers
Polymeric building blocks based on dimers from HMF have also been described. Casanova has investigated the conversion of HMF into 5,5’(oxy(bismethylene))-2-furaldehyde (7, OBMF), which is an interesting prepolymer and a precursor for an anti-viral compound (Scheme 44).545
Scheme 44. Conversion of HMF into5,5'-(oxybis(methylene))bis(furan-2-carbaldehyde)
Toste has published a synthesis of xylene, which starts from HMF derived 2,5-dimethyl- furan which is reacted with glycerol-derived acroleine in a Diels-Alder reaction (Scheme 45). The Diels-Alder adduct is oxidised to the carboxylic acid using a stoichiometric oxidant, followed by acid catalysed rearrangement to 2,5-dimethyl-benzoic acid. This is
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|233 decarboxylated in a copper-catalsyed reaction to p-xylene, which is the raw material for terephthalic acid, an important monomer for polyesters such as PET.546
Scheme 45. Xylene synthesis, starting from 2,5-dimethylfuran
2.6.2 Fine chemicals
2.6.2.1 Pharmaceuticals
In view of the rigid furan structure and the two substituents that can be easily modified, HMF has been used in quite a number of drug studies. Lukevics and co-workers investigated a series of adducts between Meldrum’s acid and substituted furans (Scheme 46).547 The derivative 60 made from HMF had some neurotropic activity.
Scheme 46. Meldrum’s acid derivative of HMF
Lewis and co-workers synthesised a range of furan containing compounds they designed to be active both as lipoxygenase inhibitors (N-hydroxy-urea moiety) and as Histamine H1 receptor antagonists (benzhydrylpiperazine moiety).548 Compound 62 showed poor lipoxygenase inhibition properties (IC50 = 1700 nm) however, it was an excellent antagonist of the Histamine H1 receptor (Ki= 4 nM) (Scheme 47).
234| Chapter 2
Scheme 47. Dual action lipoxygenase inhibitor and Histamine receptor antagonist based on HMF
Condensation of HMF with 2,4-pentanedione in the presence of B2O3, tri-butylborate and butylamine led to the double Knoevenagel product 63 (Scheme 48) which was tested for cytotoxicity against two tumour cell lines. It did not show activity in these tests.549
Scheme 48. Furan-based Curcumin analogue as anti-cancer agent
A series of phenyl-furanyl-rhodanines was synthesised and tested as antibacterial inhibitors of RNA polymerase.550 These molecules have been evaluated for their ability to inhibit transcription and to affect the growth of bacteria living in suspension or in a biofilm. An HMF-based compound was prepared via the bromomethyl derivative 64 (Scheme 49), which was arylated via a Suzuki reaction and aldol condensation of this product with N-allyl- rhodanine. The product 65 showed only low to moderate activity in the antiobacterial tests.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|235
Scheme 49. HMF-based reverse transcriptase inhibitor
Schinzer and co-workers reported the total synthesis of an epothilone d analogue 66 (Scheme 50) that was built from HMF. This compound had moderate cytotoxic activity, which was less then paclitaxel, which was used as reference.551
Scheme 50. Epothilone analogue based on HMF
2.6.2.2 Agrochemicals
5-Amino-levulinic acid (67) and its derivatives are herbicides. A synthesis from HMF, decribed in Scheme 51, was published by Descotes in collaboration with Südzucker.552,553 HMF was subjected to a Ritter reaction with acetonitrile, which delivered the desired amide 68 in 47% yield and 97% selectivity due to incomplete conversion. This amide was subjected 1 to a reaction with singlet oxygen ( O2) which was prepared by irradiation of oxygen in the presence of immobilised Rose Bengal. The crude product contained 64% of the furanone 69,
236| Chapter 2 which was reduced with zinc dust using ultrasonic irradiation to give N-acetyl 5-amino- levulinic acid (70) in 55% yield. Hydrolysis to 67 was effected by reflux in hydrochloric acid.
Scheme 51. Conversion of HMF into 5-amino-levulinic acid
Li, Xu and co-workers prepared a series of analogues of Imidacloprid, a much-used insecticide, by reaction of 5-ring aromatic aldehydes with the nitromethylene neonicotinoid 6-Cl-PMNI (71), which also has insecticidal activity (Scheme 52). Reaction with HMF in acetonitrile was catalysed by HCl and proceeded in 72% yield.554 The product 72 was active against a certain strain of aphids and against armyworm.
Scheme 52. Insecticide based on HMF
The propionic acid ester of HMF is a fungicide that can be obtained via acylation of HMF with propionic anhydride.13
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|237
2.6.2.3 Flavours & Fragrances
The Maillard reaction between reducing carbohydrates and amino acids is undoubtedly one of the most important reactions in the F&F world, leading to the development of the unique aroma and taste as well as the typical browning, which contribute to the sensory quality of thermally processed foods, such as cooked or roasted meat, roasted coffee or cocoa. Although numerous studies have addressed the structures and sensory attributes of the volatile odour- active compounds, the information available on non-volatile, sensory-active components generated during thermal food processing is scarce. Researchers in Germany examined the reaction between alanine and glucose in boiling water at pH 5.555 From the Maillard reaction product they were able to isolate a compound, which they named alapyridaine. This compound, although tasteless by itself, has the property that it enhances sweet taste. Thus the threshold detection level for tasting glucose was decreased by a factor of 16 in a 1:1 mixture of glucose and alapyridaine. They were able to synthesise alapyridaine 73 in 51-65% yield by refluxing HMF with alanine in EtOH. Although L-alanine was used as starting material the product was isolated as a racemic mixture. A two-step synthetic method was developed that allowed the isolation of the enantiopure compounds in poor yield (Scheme 53).556 It was shown that only the (S)-alapyridaine 73a was active as sweetness enhancer, whereas the (R)- enantiomer had no effect.
Scheme 53. Sweet taste enhancer from HMF and alanine
238| Chapter 2
Similar pyridinium compounds have been made by the reaction between HMF and methylamine,557 1-propylamine,557,558 glycine,559 -alanine,559 -aminobutyric acid559 and N- acetyl-lysine. 557 Terada and co-workers reported the preparation of a trimeric acetal of HMF in only 2.3% yield by treating HMF with an acidic ion exchange resin for a prolonged period of time.560 The compound was used for the preparation of flavour enhancing compounds.
2.6.2.4 Natural Products
The naturally occurring furan derivatives rehmanone A, rehmanone B, and rehmanone C have been recently isolated from the dried roots of Rehmannia glutinosa,1 and 3 from Salvia miltiorrhiza Bunge (Labiatae). Rehmanone A (74a), B (74b) and C (74c) display significant biological activity, since Rehmanone A inhibited blood platelet aggregation and Rehmanone A and C promoted immune activity. Moreover, the latter activity is inhibited when Rehmanone B and C are tested at higher concentrations. Their synthesis is straightforward as outlined in Scheme 54 below.561
Scheme 54. Synthesis of Rhemanones from HMF
Treatment of HMF with 2 eq. of TBDMS-protected p-hydroxybenzylalcohol catalysed by sulfuric acid gave a good yield of 5-{[(4-hydroxybenzyl)oxy]methyl}-2-furfural (75) (Scheme 55), a compound that was recently isolated from the rhizome of Gastrodia elata Blume (Orchidaceae), and exhibited weak cytotoxicity against the HT-29 cell line. Similarly, treatment of HMF with 2-phenylethanol catalysed by Yb(OTf)3 gave the ether, which was immediately reduced to Pichiafuran C (76), a compound that was recently isolated from the yeast Pichia membranifaciens, derived from the marine sponge Petrosia sp.562
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|239
Scheme 55. Etherification of HMF with aromatic derivatives to form natural products 75 and 76
2.6.2.5 Macrocycles
In addition to the epothilone d analogue reported by Schinzer many more macrocyclic compounds have been made using HMF as a building block. Cram reported a series of crown ethers (82-85) that were prepared from HMF and 2- chloroethanol as building blocks (Scheme 56).465
Scheme 56. Crown ethers based on HMF
240| Chapter 2
Cottier and co-workers prepared furan-containing macrocycles from 5 by etherification with bromoalkenes. The resulting bis-olefins were cyclised at high dilution using the Grubbs 1st generation catalyst. High yields of monomeric (86) or dimeric (87) macrocycles were obtained (Scheme 57).
Scheme 57. Macrocycles from HMF
Goswami and co-workers synthesised a ditopic receptor (88) for the selective binding of -dicarboxylic acids in which 2,5-bishydroxymethylfuran was used as a spacer (Scheme 58).563 Binding could be measured via NMR or using fluorescence quenching.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|241
Scheme 58. Ditopic macrocyclic receptors from HMF
2.6.2.6 Heterocycles
Furan-substituted chromones 89 were prepared by Krafft and co-workers in a two-step sequence (Scheme 59).564 These compounds were designed to be photoactivatable fluorophores. Irradiation at 350 nm for 4 min. converts them into an E/Z mixture of the highly fluorescent l-arylfuro[3,4b]chromones 90 in low yield. Prolonged irradiation (2.5h) resulted in 50% conversion and an E/Z ratio of 97:3.
242| Chapter 2
Scheme 59. Synthesis of furan-substituted chromones
A range of substituted pyrazoles was prepared via a Wittig-Horner reaction between a series of substituted aldehydes with a novel tosylhydrazone phosphonate synthon 91. The resulting sodium salt 92 of the -unsaturated tosylhydrazone was refluxed in THF to give the pyrazole 93 with concomitant elimination of sodium tosylate (Scheme 60). Thus HMF was converted to the 3-furyl-pyrazole in 60% overall yield.565
Scheme 60. Furanylpyrazole from HMF
The synthesis of 3-hydroxy-pyridinium compounds from HMF and amino acids or alkylamines was already described above. It is also possible to prepare the parent compound 94 in a two-step sequence from HMF (Scheme 61).566 The same authors also prepared
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|243 compound 95, which is a 6-hydroxymethyl analogue of the parasympathomimetic pyridostigmine.
Scheme 61. Pyridinols from HMF
Lichtenthaler and co-workers converted HMF into a number of different heterocycles by first effecting a ring-opening of the furan-ring using oxidative methods. The resulting E- and Z-1,6-dioxy-3-hexene-2,5-diones 96 were first selectively double-bond reduced and then converted into heterocycles by treatment with either alkyl- or arylamines to afford pyrroles (97), or by Lawesson’s reagent to afford the thiophenes (98) or by hydrazine to afford the pyridazines (99) in good yields (Scheme 62).567
Scheme 62. Conversion of HMF into heterocycles
244| Chapter 2
A similar series of reactions was performed on -glucosylated HMF (100, Scheme 63). Oxidation of the HMF moiety in this adduct, using m-chloro-perbenzoic acid or singlet oxygen, provided the glucosylated hydroxybutenolide 101 in yields of 70% and 95% respectively. The reaction of this product with o-phenylenediamine led to formation of the glucosylated benzodiazepinone 102 (65%). This could be dehydrogenated using 2,3-dichloro- 5,6-dicyano-benzoquinone (DDQ) to give the dehydrogenated glucosylated benzodiazepinone 103 in 89% yield.
Scheme 63. glucosylated benzodiazepinone from HMF
Connolly and co-workers described the preparation of 8-oxa-3-aza-bicyclo[3.2.1]octane as its HCl-salt 104 in a straightforward 3-step sequence from 50 (Scheme 64).474 Thus bistosylation of 50 (83% yield) followed by reaction with benzylamine (93.5% yield) gave the bicyclic N-benzylated compound 105, which was debenzylated by hydrogenolysis using
Pearlman’s catalyst (Pd(OH)2). The product was isolated as the HCl-salt in 68% yield.
Scheme 64. Synthesis of 8-oxa-3-azabicyclo[3.2.1]octane from 50
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|245
Amarasekara and co-workers converted the aldehyde moiety of HMF into the nitrileoxide, using established technology, forming 106 (Scheme 65). Cycloaddition of the nitrileoxide with a number of alkenes gave the 4- or 4,5-disubstituted 4,5-dihydroisoxazoles substituted in the 3-position (107-110) with the furfuryl unit in yields from 71 to 84%. Reaction with dimethyl acetylenedicarboxylate gave the 4,5-disubstituted isoxazole 111 in 96% yield.568
Scheme 65. Conversion of HMF into 3-furyl-isoxazole or -4,5-dihydroisoxazoles
Reaction of HMF with 1,2 aminothiols leads to the formation of thiazolidine- (112) or thiazoline-substituted hydroxymethylfuran 113 in good yields (Scheme 66).569,570
Scheme 66. Thiazolidine and thiazoline from HMF
Hanefeld and co-workers synthesised a series of rhodanine derivatives via aldol condensation of a range of aldehydes with the parent compound 114 (Scheme 67).571 In this
246| Chapter 2 fashion the HMF derivative 115 was prepared in 73% yield. These products are have potential as aldose reductase inhibitors.
Scheme 67. The synthesis of rhodanine derivates via aldol condensation
2.6.2.7 Sugar derivatives
HMF has been used as oxygen nucleophile in the Ferrier rearrangement of glycals (116) to form 117. Depending on the Lewis acid that was used as the catalyst, yields of 45-93% were achieved (See Scheme 68; only the best results are shown).572
Scheme 68. HMF as a nucleophile in the Ferrier rearrangement
Glucosylated HMF was already mentioned above in the section on heterocycles. Cottier and co-workers explored different methods of preparation of the alpha and beta-anomers. Direct coupling of HMF with protected activated glucose equivalents generally gave the glucosylated HMF in low yields. Better yields were obtained by coupling 2-furfurylalcohol and formylating the adduct afterwards. They also performed some further conversions selectively on the HMF moiety. Thus, treatment with singlet oxygen gave the corresponding glucosylated 5 hydroxy-2,5-dihydro-2-furanone 118, which could be further reduced to 119. Palladium-catalysed formate reduction thereof led to the formation of the corresponding glucosylated 5-hydroxylevulinic acid, which was esterified to 120 using dimethyl sulfate and bicarbonate (Scheme 69).573
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|247
Scheme 69. Transformations of glucosylated HMF
Hungarian researchers were able to convert HMF into dideoxysugars (Scheme 70).574 HMF was first benzoylated and next the aldehyde functionality was protected as the acetal with pinacol. This compound (121) was oxidised with bromine to the diketone (not shown), which was immediately reduced with NaBH4 to give a mixture of diastereomeric diols (122 and 123) which could be separated by chromatography. Removal of the benzoyl group with methanolic sodium methoxide was followed by ring-closure with methanolic HCl, resulting in the formation of the - and -dideoxysugars (124 and 125).
Scheme 70. Dideoxysugars from HMF
248| Chapter 2
2.6.2.8 Spiroketals
The trioxadispiroketal functionality is present in a number of biologically active marine natural products as well as in antibiotic polyether ionophores. Stockman and coworker reported a short synthesis of such a unit based on furandialdehyde, prepared in situ from
HMF with excess MnO2, which was immediately converted to the bis-acrylate ester 126 using a stabilised Wittig reagent.575 Consecutive reduction of the ester units and the double bonds gave the diol 127, which was cyclised to the trioxadispiroketal 128 using NBS (Scheme 71). They also prepared the 6,5,6-ring analogue using a Wittig C-3 synthon containing a silylated hydroxygroup. Other steps were similar.
Scheme 71. The synthesis of a trioxadispiroketal, starting from HMF
2.6.2.9 Other synthetic conversions
Acetalisation The dimethylacetal of HMF 129 has been prepared by two different methods. The classical acetalisation with MeOH, catalysed by a zeolite, led to formation of the acetal in 96% yield,521 whereas use of trimethyl orthoformate in the presence of catalytic ytterbium sulfate immobilised on amberlite led to the product in 80% yield (Scheme 72).573
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|249
Scheme 72. The synthesis of the dimethyl acetal of HMF
Acycloine condensation Chinese researchers reacted HMF with pyruvic acid, catalysed by pyruvate dehydrogenase,576 resulting in enantiopure hydroxyketone 130 which was reacted with an alcohol dehydrogenase to give the (1S,2R)-diol 131 (Scheme 73).
Scheme 73. Enzymatic conversion of HMF
The Aldol condensation The aldol reaction on the aldehyde moiety of HMF (Scheme 74) has been reported many times and usually proceeds in high yields. Dumesic investigated the condensation between acetone and HMF in biphasic THF/aqueous salt solutions using NaOH as basic catalyst. Using exactly 2 equivalents of HMF to one equivalent of acetone good yields of the double aldol condensation product 132 were obtained.577 This product can be seen as a precursor to biodiesel as full hydrogenolysis will give rise to a C-13 hydrocarbon.578 Using an excess of acetone, good yields of the single aldol adduct 133 could be obtained. Shantz and co-workers also looked at the aldol between HMF and acetone in acetone solution, using a solid secondary amine type basic catalyst.579 The reaction was very slow at 40 °C. No data were given regarding selectivity of the process.
250| Chapter 2
Scheme 74. Aldol condensation products between HMF and acetone
The reaction of HMF with 1,2,3,4-tetrahydropyridine gave the aldol condensation product described in Scheme 75.580
Scheme 75. Reaction of 1,2,3,4-tetrahydropyridine with HMF
Aldehyde to amide Sugai and co-workers described a one-pot procedure to convert aldehydes into amides 581 (Scheme 76). Thus HMF was first converted into its nitrile 135 by reaction with NH3 and
I2 in DMSO. After reduction of excess I2 with Na2S2O3, harvested cells of Rhodococcus rhodochrous IFO 15564 containing nitrile hydratase enzyme in phosphate buffer (pH 6.0, 0.1 M) were added and the reaction was stirred at 30 °C for 1–24 h to yield 136.
Scheme 76. Conversion of HMF into 5-hydroxymethyl-2-furylamide
Arylation
Beller and co-workers reported a remarkable FeCl3-catalysed alkylation reaction of activated arenes using benzyl alcohols and benzyl esters(Scheme 77).582 They were able to
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|251
react HMF with o-xylene, catalysed by 10 mol% of FeCl3, which resulted in 37% yield of a mixture of 4- and 3-substituted xylenes 137 and 138 (62:38).
Scheme 77. FeCl3-catalysed alkylation of xylene with HMF
Baylis-Hillman reaction Hu and co-workers reported the Baylis-Hillman reaction between HMF and methyl acrylate in a mixture of dioxane and water, in the presence of DABCO (100 mol%).583 The product 139 was isolated in 63% yield (Scheme 78). It was also possible to effect the Baylis-Hillman reaction between HMF and acrylamide in 61% yield.584
Scheme 78. Baylis-Hillman reaction of HMF and acrylate ester and amide.
Carbonylation Sheldon and co-workers subjected HMF to palladium-catalysed carbonylation in water using TPPTS as water-soluble ligand, described in Scheme 79.585 They used 4 mol% of palladium and 24 mol% of ligand; in addition, H2SO4 was used as a promoter. After 20 h a conversion of 90% was achieved. Unfortunately the product was not the pure carboxylic acid
140, but rather a 3:1 mixture of this compound with the hydrogenolysis product 141.
252| Chapter 2
Scheme 79. Palladium-catalysed aqueous carbonylation of HMF
Chinese researchers reported the enzymatic conversion of HMF into its cyanohydrin 142. Using an oxynitrilase enzyme they achieved only 8% conversion and 31% ee (Scheme 80).586
Scheme 80. Enzymatic cyanohydrin formation from HMF
Decarbonylation Klankermayer, Leitner and co-workers studied the decarbonylation of HMF to 2- furfurylalcohol (143).587 Using an iridium precursor in combination with triaryl- or tri- alkylphosphine ligands they were able to decarbonylate HMF with high selectivity but never exceeding 50% conversion after 48h in refluxing dioxane (110 °C). Increasing the reaction temperature to 220 °C led to full conversions, but the selectivity decreased dramatically and only decomposition products were observed. Surprisingly, use of CO2-expanded dioxane (50 bar of CO2) led to 99% conversion at 220 °C with near perfect selectivity to furfurylalcohol (Scheme 81).
Scheme 81. Decarbonylation of HMF
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|253
Dimerisation and ether formation A number of papers describe the acid-catalysed formation of dimeric ether 7 from HMF.213,465,588 The highest reported isolated yield is 44% (pTsOH, toluene, azeotropic reflux).465 Better results were obtained by carrying out the reaction in DMSO as solvent without any acid. In that case the yield reaches a maximum of 55% at 84% selectivity (Scheme 82).589
Scheme 82. Dimeric ether from HMF
The HMF dimer 7 was converted into the diisocyanate by Klein and co-workers via the diacid, which was reacted with ethyl chloroformate and sodium azide to from the bis- acylazide. Treatment of this compound at 100 °C in toluene gave the bis-isocyanate which was used to make polyurethanes.590 Ethers from HMF have also been formed by acid-catalysed reaction of HMF with alcohols, such as methanol (50% yield) and ethyleneglycol (24% yield).591 The reaction between HMF and dihydropyran was catalysed by pyridinium p-toluenesulfonate. The resulting tetrahydrofurfuryl ether was isolated in 72% yield.592 Bell reported the formation of a mixture of HMF ethyl ether (81%) and ethyl levulinate (16%) upon treatment of HMF in EtOH with sulfuric acid as catalyst. He also reported the reductive etherification to the bis ether of furan dimethanol.593 The use of solid acid catalysts, such as mesoporous aluminasilicates, zirconia and sulphated zirconia has also be reported.594 It is also possible to make ethers using the Williamson ether synthesis by reacting HMF with an alkyl halide in the presence of a base or a halogen scavenging agent, usually in the form of a silver salt.553,561,573
Esterification Esters of HMF can be formed quite easily either via acylation with aroyl chlorides574,595 or treatment with alkanoyl anhydride.553 Esters of HMF were also prepared on solid-phase using a Sieber amide resin that was functionalized by reaction with succinic anhydride. In this case the reaction was accomplished using di-isopropylcarbodiimide with hydroxyl-benzotriazole en 4-dimethylaminopyridine as co-catalysts.596 It was found that in mice the enzyme
254| Chapter 2 sulfotransferase converts HMF into the highly carcinogenic sulfate ester.93 These reactions sequences were discussed in more detail in section 2.2.3.
Formation of 5-halomethylfurfuraldehyde In section 2.4.2 the in situ formation of halomethylfurfurals from carbohydrates has been described. Also substantial work has performed on the halogenation of HMF. Gaset and co- workers carefully screened all the possible reagents that can be used to convert the hydroxymethyl moiety of HMF into a chloromethyl (77) or bromomethyl (64) group (Scheme 597 83). They reported the use of HCl and HBr in various solvents: SOCl2, with and without pyridine, SOBr2, with and without pyridine, SO2Cl2 with pyridine, PCl5 + CaCO3, PBr5 +
CaCO3, PCl3 + Et3N, PBr3 + Et3N, PCl5 or PBr3 immobilised on an anionic ion exchange resin and POCl3 + pyridine, Me3SiCl and Me3SiBr. The authors observed that the acid that was formed as a side product of the halogenation reaction induced tar formation, particularly in the presence of water, hence the need for the addition of base. Nevertheless, best results were obtained with HCl gas in Et2O (87% yield, 92% selectivity), POCl3 with pyridine (82% yield, 91% selectivity), Me3SiCl in CHCl3 (92% yield, 92% selectivity) and Me3SiBr in 1,1,2-trichloroethane (99% yield, 99% selectivity). The same authors also converted fructose in a DMF/MIBK mixture using solid acid catalysts in a yield of around 88%. After filtering off the acidic catalyst the solvent was dried by partial azeotropic distillation. Addition of POCl3 led to in situ formation of the Vielsmaier reagent, which they had previously shown to be a highly effective reagent for the chlorination of HMF.598 Thus the chloromethyl compound was obtained in 95% yield based on HMF and 85% based on fructose.599
Scheme 83. Formation of halomethyl derivatives of HMF
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|255
Ketone formation Treatment of HMF with diazomethane for 14 days gave a 40% yield of ketone 144 (Scheme 84).600
Scheme 84. Ketonisation of HMF
The Molish reaction The reaction between hexoses or pentoses with 1-naphthol in sulfuric acid leads to the formation of an intensely reddish-violet colour. This colour test for sugars is known as the Molish reaction. Unfortunately, so far it has been impossible to isolate a single compound from the Molish reaction that would allow the identification of the highly coloured product. This is possibly due to random sulfonation of the naphthyl nucleus. Thus, Ueda and co- workers screened a number of 2-substituted 1-naphthols in the Molish reaction hoping to reduce the number of sulfonated products in this way.601 Surprisingly, when using 2-methyl- 1-naphthol in the Molish reaction with glucose (Scheme 85) they obtained, after column chromatography, low yields of crystals of a reddish brown compound that gave a reddish- violet colour in concentrated sulfuric acid. The UV-spectrum of this protonated compound was identical to that of an authentic Molish reaction of 1-naphthol. The compound was identified as the substituted furan 145. Reaction between HMF and 2-methyl-1-naphthol resulted in formation of the same product.
256| Chapter 2
Scheme 85. Identification of the product of the Molish reaction between glucose/HMF and 2-methyl-1-naphthol
Phenols Van Bekkum and co-workers subjected HMF to hydrothermal treatment.437 The major product observed was 1,2,4-trihydroxybenzene (146, Scheme 86). At 350 °C a maximum selectivity to 1,2,4-trihydroxybenzene of 46% was obtained at 50% conversion of HMF.
Scheme 86. Hydrothermal conversion of HMF into 1,2,4-trihydroxybenzene
Attempts by Hashmi and co-workers to convert HMF propargyl ether to isobenzofuran-4- 602 ol, catalysed by AuCl3, failed.
Reactions with N-nucleophiles The aldehyde functionality of HMF reacts in good yield with nitrogen nucleophiles, such as hydroxylamine, to form the oxime (95% yield)568 and substituted hydrazines, mainly for identification purposes.603-606 It is possible to convert HMF into the imines (147) by reaction with aliphatic primary amines in excellent yields by simply stirring in water at room temperature. Anilines reacted more sluggishly and needed methanol or ethanol as solvent. Addition of molecular sieves and heating, either conventional or by microwave, also allowed
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|257
the isolation of these imines in good yields. Addition of 1.5 eq. of NaBH4 to the crude solution of imine gave the amines (148) in excellent yields. Scheme 87 describes the reductive amination of HMF through an imine.
Scheme 87. Reductive amination of HMF
The addition of dialkylphosphite to imines of HMF led to formation of the aminophosphonic acids. A diastereoselective version was also reported.607
Silylation In a study of Diels-Alder activity of substituted furans Sternbach and co-workers silylated the hydroxyl group of HMF with tert-butyl-dimethylsilylchloride and imidazole in DMF to form 149 (Scheme 88).608 Descotes and co-workers reported both the trimethylsilyl derivative and the tert-butyl-dimethylsilyl derivative.573,609
Scheme 88. Silylation of the hydroxyl group of HMF
258| Chapter 2
Diels-Alder reaction
Scheme 89. Internal Diels-Alder reaction of HMF derivative
The silylated HMF 149 was further functionalised as shown in Scheme 89 to give compound 152. During the acidic work up of the alkylation reaction the silyl protection group was hydrolysed off. It was replaced with a pivaloyl group. Compound 152 was treated with 525 ZnI2 and TMSCN to give 61% of a mixture of isomeric Diels Alder products 153 a,b. Another Diels-Alder reaction of an HMF derived furan was reported by Toste (Scheme 50).468
Wittig reaction Wittig and Horner-Wittig type reactions have been performed many times on HMF, a number of which were already discussed earlier. Gaset and co-workers performed a Wittig-
Horner reaction on HMF with methyl 2-(diethoxyphosphoryl)acetate using K2CO3 as base in an alcoholic solvent.610 Under these conditions a simultaneous transesterification took place with the alcoholic solvent to form 154 (Scheme 90).
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|259
Scheme 90. Simultaneous Wittig-Horner and transesterification reactions on HMF
Jacobsen and co-worker prepared -unsaturated imides (155) via a Wittig-Horner reaction. The reaction with HMF proceeded in 81% yield (Scheme 91). The products are interesting substrates for asymmetric azide addition.611
Scheme 91. Synthesis of -unsaturated imide from HMF
Orsini and co-workers performed a Wittig reaction using an in situ prepared Wittig reagent from ethyl -bromoacetate and PBu3 in the presence of tetrabutylammonium fluoride as base (Scheme 92).612 The reaction with HMF proceeded in 81% yield (156, E:Z = 85:15).
Scheme 92. Reaction of HMF with in situ formed Wittig reagent
2.6.3 HMF as precursor of fuel components
HMF is a solid at room temperature with very poor fuel blend properties, therefore HMF cannot be used and has not been considered as a fuel or a fuel additive. The SME company Avantium is developing chemical, catalytic routes to produce furan derivatives “Furanics” for a range of biofuel applications.613,614 Avantium targets biofuels with advantageous qualities, both over existing biofuels such as bioethanol and biodiesel as well as over traditional transportation fuels. Another major goal is minimising the H2 demand for their production. These compounds have a relatively high energy density, and good chemical and physical
260| Chapter 2 characteristics (a.o. boiling point, cloud point, flash point, viscosity).613 Using a regular Paccar PR 9.2 litre 183 kW diesel engine Avantium has been testing a range of blends of its novel biofuels (both C5 derived monoethers and C6 derived diethers) with regular diesel, with different concentrations (up to 30%) of Avantium’s novel biofuel.614 At all conditions tested (different blends and different steps in the ESC cycle) no difference in the engine operation was observed. Smoke and particulates as well as sulfur content decreased significantly with increasing furanics blending concentrations. Fuel consumption increases with increasing furanics amount, but is completely in line with the calculated lower energy content of furanics. The CO, CO2, NO2 exhaust percentages and THC content appeared to be independent of furanics concentrations. NOX only shows a slight increase at higher blending percentages (> 10%). The use of furans, such as HMF and furfural, as precursors of liquid hydrocarbon fuels is also an option for the production of linear alkanes in the molecular weight range appropriate for Diesel or jet fuel. The group of Dumesic has researched and evaluated the different strategies possible for upgrading HMF to liquid fuels.615 HMF can be transformed by hydrogenolysis to 2,5-dimethyl furan (2,5-DMF) with 76–79% yields over a Cu–Ru/C 401 catalyst or over CuCrO4 with 61% yield. 2,5-DMF is not soluble in water and can be used as blender in transportation fuels. A techno-economic evaluation of this process was recently published.616 Binder and Raines also reported the preparation of 2,5-DMF from fructose with a two-step method.119 Luijkx et al reported the production of 2,5-DMF by hydrogenolysis of HMF over a palladium catalyst in 1-propanol.617 To form larger hydrocarbons, HMF and other furfural products can be upgraded by aldol condensation with ketones, such as acetone, over a basic catalyst (NaOH) already at room temperatures.577 Single condensation of MF and acetone produces a C9 intermediate, which can react further with a second molecule of HMF to produce a C15 intermediate. Condensation products can then undergo hydrogenation/dehydration over bifunctional catalysts with metal and acid sites (Pd/gAl2O3 at
373–413 K and 25–52 bar; Pt/NbPO5 at 528–568 K and 60 bar) to produce linear C9 or C15 alkanes that are hydrophobic and separate spontaneously from water, reducing the cost of purification.618 Silks and co-workers claimed the use of zinc and Ytterbium salts of proline as catalyst for the condensation between HMF and acetone.619 Huber and co-workers tested a range of solid acid catalysts (MgO–ZrO2, NaY and nitrogen substituted NaY) in the aldol condensation of HMF with acetone or propanal at 120 °C. With acetone, mixtures of mono- and bis-adduct were obtained. The authors propose that with propanal a double aldol condensation occurs in 100% yield. However, the product was not isolated and
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|261 characterised.620 Aldol condensation can be coupled with hydrogenation steps using a bifunctional catalyst like Pd/MgO–ZrO2, leading to high yields of condensation products (>80%) at 326–353 K.578 Although aldol condensation is not itself a hydrogen consuming reaction, it is typically equilibrium limited and products are generally hydrogenated to achieve high yields. Thus, aldol condensations utilise high pressures of hydrogen and bi- functional (metal/base) catalysts. Additionally, biomass derived condensation products, particularly those derived from furfural or HMF, are extensively oxygenated and require a large input of hydrogen to produce alkane fuels. It is estimated that the production of C9 alkanes through condensation of HMF and acetone, for example requires 8 moles of hydrogen per mole of alkane.618 A major benefit accompanying this large consumption of hydrogen, however, is that strategies based on aldol condensation allow for selective production of jet fuel and Diesel range linear alkanes with minimal carbon branching, which is not possible with hydrogen-neutral strategies such as oligomerisation. Also several LA derivatives have been proposed for fuel applications, for instance ethyl levulinate (EL), -valerolactone (gVL), and methyl-tetrahydrofuran (MTHF).621-623 However, these components do not always exhibit satisfactory properties when blended in current fuels. Recently, scientists from Shell presented a new platform of LA derivatives, the “valeric biofuels”, which can deliver both gasoline and diesel components that are fully compatible with current transportation fuels.624 The manufacture of these valeric biofuels consists of the acid hydrolysis of lignocellulosic materials to LA, the hydrogenation of the acid to gVL and valeric acid (VA), and finally esterification to alkyl (mono/di)valerate esters. The potential of LA as an intermediate for biofuel manufacture is further confirmed by the conversion of gamma-valerolactone to kerosene- and diesel-range hydrocarbons through decarboxylation to butenes and subsequent butene oligomerisation.625 Many of the different furan and furan derived components are also covered by patent applications.245,298,626-628 629-637 The conversion of HMF to fuels has recently been reviewed.638
2.7 Conclusions
The world’s desire to be less dependent on fossil resources and to lower the carbon foot print of production processes has led to a significant increase in research and development in the field of biomass conversion to building blocks for fuels and chemicals. HMF has been known as a product from hexose dehydration for over 100 years and is an omnipresent
262| Chapter 2 component of especially heat-treated foods. Extensive toxicological assessments have not revealed major concerns. HMF is considered to be one of the most promising platform molecules that can be converted to a wealth of interesting chemicals. This has led to a large number of publications, especially in recent years, illustrated by the size of this review. A general observation here is that the vast majority of the work is exclusively focussed on optimising the yield of HMF by testing a whole range of acidic catalysts without efforts to intrinsically understand the chemistry. This is illustrated by work reporting unrealistically high yields in acidic aqueous reaction systems in which HMF is known to decompose. Nevertheless, very high yields have been obtained in non-protic solvents, in particular in DMSO. However, it seems unlikely that such a reaction can be scaled up to an industrial process. It has always been a challenge to obtain HMF in an efficient way. Although a lot of improvements have been made in understanding the mechanism and kinetics of the dehydration process, significant challenges still remain in transferring it to an industrial scale. The vast majority of the research summarised in this review has been performed on lab-scale in batch reactions without a real focus on finding an efficient and economically viable process. Improvements have been made in recent years by applying different solvent types and extraction methods, and by applying bifunctional catalyst systems. The use of biphasic systems appears to be straight-forward; good yields of HMF have been obtained, and such processes would appear to be scalable although they require relatively large amounts of extracting solvents. There has been little focus on the purification of HMF and the efficient recycling of reaction- and extraction solvents or catalysts. Application of ionic liquid solvent systems and microwave heating are showing great promise, especially in the conversion of glucose to HMF, but are still mostly uncharted territory. The focus appears to be mainly on testing new ionic liquids in stead of finding efficient ways to tackle the main draw-backs of using ionic liquids, namely efficient separation of HMF from the ionic liquid and recycling it by removing impurities. The biggest challenge still to overcome is the use of glucose and glucose based polymers, like starch and cellulose, as the carbohydrate feedstock due to the unstable nature of the product HMF and the severe reaction conditions required. Key to the success of HMF will be an economical production process. So far two pilot plant efforts have not been successful in the direct production of HMF. Recently, Avantium Chemicals started operating a pilot plant for production of furan-based chemicals and plastics in which HMF is in situ converted to its ether to help improve yield and purification. A
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|263 similar strategy is followed by Mascal in a process that produces 5-chloromethyl-furfural, another more stable derivative of HMF in order to obtain high yields in combination with efficient separation. It is expected that once production has been scaled up to 100-200 kTon/year a cost price of HMF of around $ 1.00/Kg should be possible. In order to reach the lower cost price that would be necessary for fuel applications, it will be necessary to produce HMF or a derivative directly from cellulose or preferably lignocellulose. This review underlines the progress that has been made towards an economic production of furan derivatives from biomass; a development which will open the way to a wealth of interesting materials all the way from fuels and bulk chemicals to fine chemicals and pharmaceuticals applications.
264| Chapter 2
2.8 References
(1) Bozell, J. J.; Petersen, G. R. Green Chem. 2010, 12, 539-554. (2) Werpy, T.; Petersen, G. Top Value Added Chemicals from Biomass Vol. I-Results of Screening for Potential Candidates from Sugars and Synthesis Gas, NREL/TP-510-35523; National Renewable Energy Laboratory: Golden, CO, 2004. (3) Mulder, G. J. J. Prakt. Chem. 1840, 21, 203-240. (4) von Grote, A.; Tollens, B. Ber. 1874, 7, 1375. (5) Newth, F. H. Adv. Carbohydr. Chem. 1951, 6, 83-106. (6) Düll, G. Chem. Ztg. 1895, 19, 216. (7) Kiermayer, J. Chem. Ztg. 1895, 19, 1003. (8) Data was taken from the following: Sigma-Aldrich Catalog; Sigma-Aldrich: St. Louis, MO, 2010. (9) The Merck Index; 11 ed.; Budavari, S., Ed.; Merck & Co., Inc.: Rahway, NJ, 1989, p767. (10) Moye, C. J. Rev. Pure Appl. Chem. 1964, 14, 161-170. (11) Feather, M. S.; Harris, J. F. Adv. Carbohydr. Chem. 1973, 28, 161-224. (12) Kuster, B. F. M. Starch/Stärke 1990, 42, 314-321. (13) Lewkowski, J. ARKIVOC 2001, i, 17-54. (14) Moreau, C.; Belgacem, M. N.; Gandini, A. Top. Catal. 2004, 27, 11-30. (15) Tong, X.; Ma, Y.; Li, Y. Appl. Catal. A: Gen. 2010, 385, 1-13. (16) Rosatella, A. A.; Simeonov, S. P.; Frade, R. F. M.; Afonso, C. A. M. Green Chem. 2011, 13, 754-793. (17) Stark, A.; Ondruschka, B. In Handbook of Green Chemistry, Volume 6: Ionic Liquids; Wasserscheid, P., Stark, A., Eds.; WILEY-VCH Verlag GmbH & Co. KGaA: Weinheim, 2010. (18) Ståhlberg, T.; Fu, W.; Woodley, J. M.; Riisager, A. ChemSusChem 2011, 4, 451-458. (19) Lima, S.; Antunes, M. M.; Pillinger, M.; Valente, A. A. ChemCatChem 2011, 3, 1686-1706. (20) Zakrzewska, M. E.; Bogel-Łukasik, E.; Bogel-Łukasik, R. Chem. Rev. 2010, 111, 397-417. (21) Corma, A.; Iborra, S.; Velty, A. Chem. Rev. 2007, 107, 2411. (22) Patel, A. D.; Serrano-Ruiz, J. C.; Dumesic, J. A.; Anex, R. P. Chem. Eng. J. 2010, 160, 311- 321. (23) Xu, Q.; Li, Y.-H.; Lü, X.-Y. J. Zheijang Univ. Sci. B 2007, 8, 439-445. (24) Li, Y.-X.; Li, Y.; Qian, Z.-J.; Kim, M.-M.; Kim, S.-K. J. Microbiol. Biotechnol. 2009, 19, 1319-1327. (25) Ding, X.; Wang, M.-Y.; Yu, Z. L.; Hu, W.; Cai, B.-C. Zhonqquo Zhong Yao Za Zhi 2008, 33, 392-396. (26) Podesta, F.; Fajardo, V.; Freyer, A. J.; Shamma, M. Arch. Pharm. 1988, 321, 949. (27) Kroh, L. W. Food Chem. 1994, 51, 373-379. (28) Martins, S. I. F. S.; Van Boekel, M. A. J. S. Food Chem. 2005, 92, 437-448. (29) Van Boekel, M. A. J. S. Nahrung (Food) 2001, 45, 150–159. (30) Kus, S.; Gogus, F.; Eren, S. Int. J. Food Prop. 2005, 8, 367-375. (31) Rufián-Henares, J. A.; De la Cueva, S. P. Food Add. Contamin. 2008, 25, 1306-1312. (32) EFSA J. 2005, 215, 1-73. (33) Du, Z.; Ma, J.; Wang, F.; Liu, J.; Xu, J. Green Chem. 2011, 13, 554-557. (34) Capuano, E.; Fogliano, V. LWT - Food Sci. Technol. 2011, 44, 793-810. (35) Abraham, K.; Gürtler, R.; Berg, K.; Heinemeyer, G.; Lampen, A.; Appel, K. E. Mol. Nutr. Food Res. 2011, 55, 667-678. (36) Florian, S.; Bauer-Marinovic, M.; Taugner, F.; Dobbernack, G.; Monien, B. H.; Meinl, W.; Glatt, H. Mol. Nutr. Food Res. 2012, 56, 593-600. (37) Arribas-Lorenzo, G.; Morales, F. J. Food Chem. Toxicol. 2010, 48, 644-649. (38) Ulbricht, R. J.; Northup, S. J.; Thomas, J. A. Fundam. Appl. Toxicol. 1984, 4, 843-853.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|265
(39) Delgado-Andrade, C.; Seiquer, I.; Navarro, M. P.; Morales, F. J. Molecular Nutrition Food Research 2007, 51, 341-351. (40) Capuano, E.; Ferrigno, A.; Acampa, I.; Serpen, A.; Açar, Ö. Ç.; Gökmen, V.; Fogliano, V. Food Res. Int. 2009, 42, 1295-1302. (41) Gökmen, V.; Açar, O. C.; Serpen, A.; Morales, F. J. Eur. Food Res. Technol. 2008, 226, 1031-1037. (42) Arena, E.; Ballistreri, G.; Tomaselli, F.; Fallico, B. J. Food Sci. 2011, 76, C1203-1210. (43) Gensberger, S.; Mittelmaier, S.; Glomb, M. A.; Pischetsrieder, M. Anal. Bioanal. Chem. 2012. (44) Teixidó, E.; Santos, F. J.; Galceran, M. T. J. Chromatogr. A 2006, 1135, 85-90. (45) Teixidó, E.; Núñez, O.; Santos, F. J.; Galceran, M. T. Food Chem. 2011, 126, 1902–1908. (46) Ait Ameur, L.; Mathieu, O.; Lalanne, V.; Trystram, G.; Birlouez-Aragon, I. Food Chem. 2007, 101, 1407-1416. (47) Kirit, A. B.; Erdogdu, F.; Ozdemir, Y. J. Food Process Eng. 2012. (48) Husøy, T.; Haugen, M.; Murkovic, M.; Jöbstl, D.; Stølen, L. H.; Bjellaas, T.; Rønningborg, C.; Glatt, H.; Alexander, J. Food Chem. Toxicol. 2008, 46, 3697-3702. (49) Murkovic, M.; Pichler, N. Mol. Nutr. Food Res. 2006, 50, 842-846. (50) Bachmann, S.; Meier, M.; Kaènzig, A. Lebensmittelchemie 1997, 51, 49-50. (51) Rada-Mendoza, M.; Olano, A.; Villamiel, M. Food Chem. 2002, 79, 513-516. (52) Ruiz-Matute, A. I.; Soria, A. C.; Sanz, M. L.; Martínez-Castro, I. J. Food Compos. Anal. 2010, 23, 260-263. (53) Corradini, D.; Corradini, C. J. Chromatogr. 1992, 624, 503-509. (54) Zhao, B.; Hall III, C. A. Food Chem. 2008, 108, 511-518. (55) Zappalà, M.; Fallico, B.; Arena, E.; Verzera, A. Food Control 2005, 16, 273-277. (56) Khalil, M. I.; Sulaiman, S. A.; Gan, S. H. Food Chem. Toxicol. 2010, 48, 2388. (57) Delgado-Andrade, C.; Morales, F. J.; Seiquer, I.; Pilar Navarro, M. Food Res. Int. 2010, 43, 1304-1311. (58) Del Campo, G.; Berregi, I.; Caracena, R.; Zuriarran, J. Talanta 2010, 81, 367-371. (59) Masino, F.; Chinnici, F.; Franchini, G. C.; Ulrici, A.; Antonelli, A. Food Chem. 2005, 92, 673-679. (60) Theobald, A.; Müller, A.; Anklam, E. J. Agric. Food Chem. 1998, 46, 1850-1854. (61) Bononi, M.; Tateo, F. J. Food Compos. Anal. 2009, 22, 79-82. (62) Giordano, L.; Calabrese, R.; Davoli, E.; Rotilio, D. J. Chromatogr. A 2003, 1017, 141-149. (63) Câmara, J. S.; Alves, M. A.; Marques, J. C. Anal. Chim. Acta 2006, 563, 188-197. (64) Pérez-Coello, M. S.; González-Viñas, M. A.; García-Romero, E.; Díaz-Maroto, M. C.; Cabezudo, M. D. Food Control 2003, 14, 301-306. (65) Cutzach, I.; Chatonnet, P.; Dubourdieu, D. J. Agric. Food Chem. 1999, 47, 2837-2846. (66) Lo Coco, F.; Valentini, C.; Novelli, V.; Ceccon, L. Anal. Chim. Acta 1995, 306, 57-64. (67) Li, M.; Yang, Z.; Yang, M.; Shan, L.; Dong, J. J. Inst. Brew. 2009, 115, 226-231. (68) Quesada Granados, J.; Villalón Mir, M.; García-Serrana, L.; López Martínez, M. C. J. Agric. Food Chem. 1996, 44, 1378-1381. (69) Caligiani, A.; Acquottti, D.; Palla, G.; Bocchi, V. Anal. Chim. Acta 2007, 585, 110-119. (70) Cocchi, M.; Ferrari, G.; Manzini, D.; Marchetti, A.; Sighinolfi, S. J. Food Eng. 2007, 79, 1438-1444. (71) Cerezo, A. B.; Tesfaye, W.; Soria-Díaz, M. E.; Torija, M. J.; Mateo, E.; Garcia-Parilla, M. C.; Troncoso, A. M. J. Food Compos. Anal. 2010, 23, 175-184. (72) Masino, F.; Chinnici, F.; Bendini, A.; Montevecchi, G.; Antonelli, A. Food Chem. 2008, 16, 90-95. (73) Chinnici, F.; Masino, F.; Antonelli, A. J. Chromatogr. Sci. 2003, 26, 2161-2171. (74) Crump, D. R.; Gardner, D. Environ. Int. 1989, 15, 455-462. (75) Janzowski, C.; Glaab, V.; Samimi, E.; Schlatter, J.; Eisenbrand, G. Food Chem. Toxicol. 2000, 38, 801-809. (76) Jöbstl, D.; Husøy, T.; Alexander, J.; Bjellaas, T.; Leitner, E.; Murkovic, M. Food Chem. 2010, 123, 814-818. (77) Jellum, E.; Børresen, H. C.; Eldjarn, L. Clin. Chim. Acta 1973, 47, 191-201.
266| Chapter 2
(78) Godfrey, V. B.; Chen, L.-J.; Griffin, R. J.; Lebetkin, E. H.; Burka, L. T. J. Toxicol. Environ. Health, Part A 1999, 57, 199-210. (79) Germond, J. E.; Philippossian, G.; Richli, U.; Bracco, I.; Arnaud, M. J. J. Toxicol. Environ. Health 1987, 22, 79-89. (80) Prior, R. L.; Wu, X.; Gu, L. J. Agric. Food Chem. 2006, 54, 3744-3749. (81) Surh, Y. J.; Liem, A.; Miller, J. A.; Tannenbaum, S. R. Carcinogenesis 1994, 15, 2375-2377. (82) Glatt, H. R.; Sommer, Y. In Acrylamide and other health hazardous compounds in heat- treated foods; Skog, K., Alexander, J., Eds.; Woodhead Publishing: Cambridge, 2006. (83) Wierckx, N.; Koopman, F.; Ruijssenaars, H. J.; Winde, J. H. Appl. Microbiol. Biotechnol. 2011, 92, 1095-1105. (84) Y.-X., L.; Li, Y.; Qian, Z.-J.; Kim, M.-M.; Kim, S.-K. J. Microbiol. Biotechnol. 2009, 19, 1319-1327. (85) Geng, F.; Wang, X. J. World Sci. Technol. 2005, 7, 52-55. (86) Fu, Z. Q.; Wang, M.-Y.; Cai, B.-C. Chin. Arch. Tradit. Chin. Med. 2008, 26, 508-510. (87) Abdulmalik, O.; Safo, M. K.; Chen, Q.; Yang, J.; Brugnara, C.; Ohene-Frempong, K.; Abraham, D. J.; Asakura, T. J. Haematol. 2005, 128, 552-561. (88) Ding, X.; Wang, M.-Y.; Yao, Y.-X.; Li, G.-Y.; Cai, B.-C. J. Ethnopharmacol. 2010, 128, 373-376. (89) Wang, M.-Y.; Zhao, F.-M.; Peng, H.-Y.; Lou, C.-H.; Li, Y.; Ding, X.; Yu, X.-Y.; Yang, G.- M.; Xu, D.-G.; Jiang, L.-H.; Zhang, X.; Ye, L.-H.; Cai, B.-C. J. Ethnopharmacol. 2010, 130, 424–428. (90) Sriwilaijaroen, N.; Kadowaki, A.; Onishi, Y.; Gato, N.; Ujike, M.; Odagiri, T.; Tashiro, M.; Suzuki, Y. Food Chem. 2011, 127, 1-9. (91) Severin, I.; Dumont, C.; Jondeau-Cabaton, A.; Graillot, V.; Chagnon, M.-C. Toxicol. Lett. 2010, 192, 189–194. (92) Glatt, H.; Schneider, H.; Murkovic, M.; Monien, B. H.; Meinl, W. Mutagenesis 2012, 27, 41- 48. (93) Monien, B. H.; Frank, H.; Seidel, A.; Glatt, H. Chem. Res. Toxicol. 2009, 22, 1123. (94) Bakhiya, N.; Monien, B.; Frank, H.; Seidel, A.; Glatt, H. Biochem. Pharmacol.2009, 78, 414- 419. (95) Svendsen, C.; Husoey, T.; Glatt, H.; Paulsen, J. E.; Alexander, J. Anticancer Res. 2009, 29, 1921-1926. (96) Seiquer, I.; Díaz-Alguacil, J.; Delgado-Andrade, C.; López-Frías, M.; Muñoz Hoyos, A.; Galdó, G.; Pilar Navarro, M. Am. J. Clin. Nutr. 2006, 83, 1082-1088. (97) Tomo, T. Peritoneal Dial. Int. 2008, 28, S123-S127. (98) Lee, D.-H.; Choi, S.-Y.; Ryu, H.-M.; Kim, C.-D.; Park, S.-H.; Chung, H.-Y.; Kim, I.-S.; Kim, Y.-L. Peritoneal Dial. Int. 2009, 29, 44-51. (99) Hoydonckx, H. E.; Van Rhijn, W. M.; Van Rhijn, W.; De Vos, D. E.; Jacobs, P. A. In Ullmann’s Encyclopedia of Industrial Chemistry; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, 2007. (100) Zeitsch, K. J. The Chemistry and Technology of Furfural and its Many By-products; Elsevier: The Netherlands, 2000. (101) Dias, A. S.; Lima, S.; Pillinger, M.; Valente, A. A.; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, 2010. (102) Moreau, C.; Durand, R.; Razigade, S.; Duhamet, J.; Faugeras, P.; Rivalier, P.; Ros, P.; Avignon, G. Appl. Catal. A: Gen. 1996, 145, 211-24. (103) Anet, E. F. L. J. Adv. Carbohydr. Chem. 1964, 19, 181-218. (104) Antal, M. J.; Mok, W. S. L.; Richards, G. N. Carbohydr. Res. 1990, 199, 91-109. (105) Speck, J. C., Jr. Adv. Carbohydr. Chem. 1958, 13, 63-103. (106) Anet, E. F. L. J. J. Am. Chem. Soc. 1960, 82, 1502. (107) Anet, E. F. L. J. Chem. Ind. 1962, 262. (108) Wolfrom, M. L.; Wallace, E. G.; Metcalf, E. A. J. Am. Chem. Soc. 1942, 64, 265-269. (109) Jadhav, H.; Pedersen, C. M.; Sølling, T.; Bols, M. ChemSusChem 2011, 4, 1049-1051. (110) Akien, G. R.; Qi, L.; Horvath, I. T. Chem. Commun. 2012, 48, 5850-5852. (111) Alberda van Ekenstein, W.; Blanksma, J. J. Ber. 1910, 43, 2355.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|267
(112) Dekker, C. A.; Hashizume, T. Arch. Biochem. Biophys. 78, 1958, 78, 348-358. (113) Tiemann, F.; Haarmann, R. Ber. 1886, 19, 1257-1281. (114) Zhang, J.; Weitz, E. ACS Catal. 2012, 2, 1211-1218. (115) Schneider, B.; Lichtenthaler, F. W.; Steinle, G.; Schiweck, H. Liebigs Ann. Chem. 1985, 2443-2453. (116) Kimura, H.; Nakahara, M.; Matubayasi, N. J. Phys. Chem. A 2011, 115, 14013-14021. (117) Amarasekara, A. S.; Williams, L. D.; Ebede, C. C. Carbohydr. Res. 2008, 343, 3021-3024. (118) Nicole, D. J.; Gillet, B.; Eppiger, E. N.; Delpuech, J. J. Tetrahedron Lett. 1982, 23, 1669- 1672. (119) Binder, J. B.; Raines, R. T. J. Am. Chem. Soc. 2009, 131, 1979-1985. (120) Yoshida, T.; Yanachi, S.; Matsumara, Y. J. Jpn. Inst. Energy 2007, 87, 700-706. (121) Kabyemela, B. M.; Adschiri, T.; Malaluan, R. M.; Arai, K. Ind. Eng. Chem. Res. 1999, 38, 2888-2895. (122) Kabyemela, B. M.; Adschiri, T.; Malaluan, R. M.; Arai, K. Ind. Eng. Chem. Res. 1997, 36, 1552-1558. (123) Li, Y.; Lu, X.; Yuan, L.; Liu, X. Biomass Bioenergy 2009, 33, 1182-1187. (124) Lima, S.; Antunes, M. M.; Fernandes, A.; Pillinger, M.; Ribeiro, M. F.; Valente, A. A. Molecules 2010, 15, 3863-3877. (125) Bicker, M., PhD Thesis, Technische Universität Darmstadt, Darmstadt, Germany 2005. (126) Watanabe, M.; Aizawa, Y.; Iida, T.; Aida, T. M.; Levy, C.; Sue, K.; Inomata, H. Carbohydr. Res. 2005, 340, 1925-1930. (127) Watanabe, M.; Aizawa, Y.; Iida, T.; Nishimura, R.; Inomata, H. Appl. Catal. A: Gen. 2005, 295, 150-156. (128) Aida, T. M.; Sato, Y.; Watanabe, M.; Tajima, K.; Nonaka, T.; Hattori, H.; Arai, K. J. Supercrit. Fluids 2007, 40, 381-388. (129) Qi, X.; Watanabe, M.; Aida, T. M.; Smith Jr., R. L. Catal. Commun. 2008, 9, 2244-2249. (130) Yan, H.; Yang, Y.; Tong, D.; Xiang, X.; Hu, C. Catal. Commun. 2009, 10, 1558-1563. (131) Ishida, H.; Seri, K. J. Mol. Catal. A: Chem. 1996, 112, L163-165. (132) Erixon, M.; Wieslander, A.; Lindén, T.; Carlsson, O.; Forsbäck, G.; Svensson, E.; Jönsson, J. Å.; Kjellstrand, P. Peritoneal Dial. Int. 2006, 26, 490-497. (133) Erixon, M.; Lindén, T.; Kjellstrand, P.; Carlsson, O.; Ernebrant, M.; Forsbäck, G.; Wieslander, A.; Jönsson, J. Å. Peritoneal Dial. Int. 2004, 24, 392-398. (134) Zhao, H.; Holladay, J. E.; Brown, H.; Zhang, Z. C. Science 2007, 316, 1597. (135) Zhao, H.; Holladay, J. E.; Zhang, Z. C.; (Battelle Memorial Institute) WO2008/019219, 2008. (136) Pidko, E. A.; Degirmenci, V.; van Santen, R. A.; Hensen, E. J. M. Angew. Chem. Int. Ed. 2010, 49, 2530-2534. (137) Pidko, E. A.; Degirmenci, V.; van Santen, R. A.; Hensen, E. J. M. Inorg. Chem. 2010, 49, 10081-10091. (138) Fenn, T. D.; Ringe, D.; Petsko, G. A. Biochemistry 2004, 43, 6464. (139) Meilleur, F.; Snell, E. H.; van der Woerd, M. J.; Judge, R. A.; Myles, D. A. A. Eur. Biophys. J. 2006, 35, 601. (140) He, X.; Agnihotri, G.; Liu, H. Chem. Rev. 2000, 100, 4615. (141) Ferenczy, G. C.; Nára-Szabó, G.; Várnai, P. Int. J. Quantum Chem. 1999, 75, 215. (142) Zhang, Y.; Pidko, E. A.; Hensen, E. J. M. Chem. Eur. J. 2011, 17, 5281-5288. (143) Yong, G.; Zhang, Y.; Ying, J. Y. Angew. Chem. Int. Ed. 2008, 47, 9345-9348. (144) Hu, S.; Zhang, Z.; Song, J.; Zhou, Y.; Han, B. Green Chem. 2009, 11, 1746-1749. (145) Ståhlberg, T.; Rodriguez-Rodriguez, S.; Fristrup, P.; Riisager, A. Chemistry – A European Journal 2011, 17, 1456-1464. (146) Hicks, K. B.; Symanski, E. V.; Pfeffer, P. E. Carbohydr. Res. 1983, 112, 37-50. (147) Mendicino, J. F. J. Am. Chem. Soc. 1960, 82, 4975-4979. (148) Martins, S. I. F. S., PhD Thesis, Wageningen University, the Netherlands, 2003. (149) Nikolov, P. Y.; Yaylayan, V. A. J. Agric. Food Chem. 2011, 59, 6099-6107. (150) Cämmerer, B.; Wedzicha, B. L.; Kroh, L. W. Eur. Food Res. Technol. 1999, 2099, 261-265. (151) Murkovic, M.; Bornik, M.-A. Mol. Nutr. Food Res. 2007, 51, 390-394. (152) Sasaki, M.; Goto, K.; Tajima, K.; Adschiri, T.; Arai, K. Green Chem. 2002, 4, 285-287.
268| Chapter 2
(153) Weenen, H.; Tjan, S. B. In Flavor Precursors; Teranishi, R., Takeoka, G. R., Guntert, M., Eds.; American Chemical Society: Washington, DC, 1992. (154) Christian, T. J.; Manley-Harris, M.; J., F. R.; Parker, B. A. J. Agric. Food Chem. 2000, 48, 1823-1837. (155) Kuster, B. F. M.; van der Baan, H. S. Carbohydr. Res. 1977, 54, 165-176. (156) Kuster, B. F. M. Carbohydr. Res. 1977, 54, 177-183. (157) Kuster, B. F. M.; Temmink, H. M. G. Carbohydr. Res. 1977, 54, 185-191. (158) Van Dam, H. E.; Kieboom, A. P. G.; Van Bekkum, H. Starch/Stärke 1986, 38, 95-101. (159) Hansen, T. S.; Woodley, J. M.; Riisager, A. Carbohydr. Res. 2009, 344, 2568-2572. (160) Hensen, E. J. M. Lecture at CATCHBIOo Symposium, Amsterdam, 2011. (161) Yang, G.; Pidko, E. A.; Hensen, E. J. M.; Eindhoven University of Technology, 2011. (162) Wang, Z.; Lin, W.; Song, W. BioResources. 2011, 6, 4045-4054. (163) Seri, K.; Inoue, Y.; Ishida, H. Bull. Chem. Soc. Jpn. 2001, 74, 1145-1150. (164) Girisuta, B.; Janssen, L. P. B. M.; Heeres, H. J. Green Chem. 2006, 8, 701-709. (165) Girisuta, B.; Janssen, L. P. B. M.; Heeres, H. J. Chem. Eng. Res. Des. 2006, 84, 339-349. (166) Heeres, H.; Handana, R.; Chunai, D.; Borromeus Rasrendra, C.; Girisuta, B.; Jan Heeres, H. Green Chem. 2009, 11, 1247-1255. (167) Girisuta, B.; Danon, B.; Manurung, R.; Janssen, L. P. B. M.; Heeres, H. J. Bioresour. Technol. 2008, 99, 8367-8375. (168) Kallury, R. K. M. R.; Ambidge, C.; Tidwell, T. T.; Boocock, D. G. B.; Agblevor, F. A.; Stewart, D. J. Carbohydr. Res. 1986, 158, 253-261. (169) Silberman, H. C. J. Org. Chem 1961, 26, 1967-1969. (170) Román-Leshkov, Y.; Chheda, J. N.; Dumesic, J. A. Science 2006, 312, 1933. (171) Kabyemela, B. M.; Adschiri, T.; Malaluan, R. M.; Arai, K. Ind. Eng. Chem. Res. 1997, 36, 2025-2030. (172) Aida, T. M.; Tajima, K.; Watanabe, M.; Saito, Y.; Kuroda, K.; Nonaka, T.; Hattori, H.; Smith Jr., R. L.; Arai, K. J. Supercrit. Fluids 2007, 42, 110-119. (173) Witowski, J.; Jörres, A. Peritoneal Dial. Int. 2000, 20, S31-S36. (174) Qi, X.; Watanabe, M.; Aida, T. M.; Smith Jr., R. L. ChemSusChem 2009, 2, 944-946. (175) Patil, S. K. R.; Lund, C. R. F. Energy Fuels 2011, 25, 4745-4755. (176) Zhang, M.; Yang, H.; Liu, Y.; Sun, X.; Zhang, D.; Xue, D. Nanoscale Res. Lett. 2012, 7, 38. (177) Assary, R. S.; Redfern, P. C.; Hammond, J. R.; Greeley, J.; Curtiss, L. A. J. Phys. Chem. B 2010, ?, ? (178) Caratzoulas, S.; Vlachos, D. G. Carbohydr. Res. 2011, 346, 664-672. (179) Mushrif, S. H.; Caratzoulas, S.; Vlachos, D. G. Phys. Chem. Chem. Phys. 2012, 14, 2637- 2644. (180) Qian, X. J. Phys. Chem. A 2011, 115, 11740-11748. (181) Guan, J.; Cao, Q.; Guo, X.; Mu, X. Comput. Theor. Chem. 2011, 963, 453-462. (182) Perez Locas, C.; Yaylayan, V. A. J. Agric. Food Chem. 2008, 56, 6717-6723. (183) Carlini, C.; Giuttari, M.; Raspolli Galletti, A. M.; Sbrana, G.; Armaroli, T.; Busca, G. Appl. Catal. A: Gen. 1999, 183, 295-302. (184) Benvenuti, F.; Carlini, C.; Patrono, P.; Raspolli Galletti, A. M.; Sbrana, G.; Massucci, M. A.; Galli, P. Appl. Catal. A: Gen. 2000, 193, 147-153. (185) Haworth, W. N.; Jones, W. G. M. J. Chem. Soc. 1944, 667. (186) Rigal, L.; Gaset, A. Biomass 1983, 3, 151-163. (187) Mascal, M.; Nikitin, E. B. Angew. Chem. 2008, 120, 1-4. (188) Seri, K.; Sasaki, T.; Shibata, M.; Inoue, Y.; Ishida, H. Bioresour. Technol. 2002, 81, 257-260. (189) Girisuta, B.; Janssen, L. P. B. M.; Heeres, H. J. Ind. Eng. Chem. Res. 2007, 46, 1696-1708. (190) Peng, L.; Lin, L.; Zhang, J.; Zhuang, J.; Zhang, B.; Gong, Y. Molecules 2010, 15, 5258-5272. (191) Popoff, T.; Theander, O. Carbohydr. Res. 1972, 22, 135-149. (192) Usuki, C.; Kimura, Y.; Adachi, S. Chem. Eng. Technol. 2008, 31, 133-137. (193) Kuster, B. F. M.; Tebbens, L. M. Carbohydr. Res. 1977, 54, 159-164. (194) Vinke, P.; Van Bekkum, H. Starch/Stärke 1992, 44, 90-96. (195) Rapp, K. M. (Süddeutsche Zucker-Aktiengesellschaft) US Patent 4740605, 1987. (196) El Hajj, T.; Masroua, A.; Martin, J. C.; Descotes, G. Bull. Soc. Chim. Fr. 1987, 5, 855-860.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|269
(197) Armaroli, T.; Busca, G.; Carlini, C.; Giuttari, M.; Raspolli Galletti, A. M.; Sbrana, G. J. Mol. Catal. A: Chem. 2000, 151, 233-243. (198) Carlini, C.; Patrono, P.; Raspolli Galletti, A. M.; Sbrana, G. Appl. Catal. A: Gen. 2004, 275, 111-118. (199) Kuster, B. F. M.; van der Steen, H. J. C. Starch/Stärke 1977, 29, 99-103. (200) Carniti, P.; Gervasini, A.; Biella, S.; Auroux, A. Catal. Today 2006, 118, 373-378. (201) Carniti, P.; Gervasini, A.; Marzo, M. Catal. Commun. 2011, 12, 1122-1126. (202) By Deng, T.; Cui, X.; Qi, Y.; Wang, Y.; Hou, X.; Zhu, Y. Chem. Commun. 2012, 48, 5494- 5496. (203) Tarabanko, V. E.; Chernyak, M. Y.; Nepomnyashchiy, I. V.; Smirnova, M. A. Chem. Sus. Dev. 2006, 14, 49-53. (204) Asghari, F. S.; Yoshida, H. Carbohydr. Res. 2006, 341, 2379-2387. (205) De, S.; Dutta, S.; Saha, B. Green Chem. 2011, 13, 2859-2868. (206) Bicker, M.; Hirth, J.; Vogel, H. Green Chem. 2003, 5, 280-284. (207) Qi, X.; Watanabe, M.; Aida, T. M.; Smith Jr., R. L. Green Chem. 2008, 10, 799-805. (208) Qi, X.; Watanabe, M.; Aida, T. M.; Smith Jr., R. L. Catal. Commun. 2009, 10, 1771-1775. (209) Nakamura, Y.; Morikawa, S. Bull. Chem. Soc. Jpn. 1980, 53, 3705-3706. (210) Szmant, H. H.; Chundury, D. D. J. Chem. Tech. Biotechnol. 1981, 31, 135-145. (211) Mercadier, D.; Rigal, L.; Gaset, A.; Gorrichon, J. P. J. Chem. Tech. Biotechnol. 1981, 31, 489-496. (212) Brown, D. W.; Floyd, A. J.; Kinsman, R. G.; Roshan-Ali, Y. J. Chem. Tech. Biotechnol. 1982, 32, 920-924. (213) Musau, R. M.; Munavu, R. M. Biomass 1987, 13, 67-75. (214) Traynelis, V. J.; Hergenrother, W. L.; Livingstone, J. R.; Valicenti, J. A. J. Org. Chem 1962, 27, 2377. (215) Gillis, B. T.; Beck, P. E. J. Org. Chem 1963, 28, 1388. (216) Seri, K.; Inoue, Y.; Ishida, H. Chem. Lett. 2000, 22-23. (217) Wang, F.; Shi, A.-W.; Qin, X.-X.; Liu, C.-L.; Dong, W.-S. Carbohydr. Res. 2011, 346, 982- 985. (218) Shimizu, K.-i.; Uozumi, R.; Satsuma, A. Catal. Commun. 2009, 10, 1849-1853. (219) De, S.; Dutta, S.; Patra, A. K.; Bhaumik, A.; Saha, B. J. Mater. Chem. 2011, 21, 17505- 17510. (220) Dutta, S.; De, S.; Patra, A. K.; Sasidharan, M.; Bhaumik, A.; Saha, B. Appl. Catal. A: General 2011, 409-410, 133-139. (221) Qi, X.; Watanabe, M.; Aida, T. M.; Smith, R. L., Jr. Ind. Eng. Chem. Res. 2008, 47, 9234- 9239. (222) Tong, X.; Li, Y. ChemSusChem 2010, 3, 350-355. (223) Qu, Y.; Song, Y.; Huang, C.; Zhang, J.; Chen, B. Adv. Mat. Res., 287-290, 1585-1590. (224) Ohara, M.; Takagaki, A.; Nishimura, S.; Ebitani, K. Appl. Catal. A: Gen. 2010, 383, 149-155. (225) Bao, Q.; Qiao, K.; Tomida, D.; Yokoyama, C. Catal. Commun. 2008, 9, 1383-1388. (226) Zhang, Y.; Degirmenci, V.; Li, C.; Hensen, E. J. M. ChemSusChem 2011, 4, 59-64. (227) Sidhpuria, K. B.; Daniel-da-Silva, A. L.; Trindade, T.; Coutinho, J. A. P. Green Chem. 2011, 13, 340-349. (228) Binder, J. B.; Raines, R. T.; (Wisconsin Alumni Research Foundation) Int. Patent WO 2009/155297, 2009. (229) Caes, B. R.; Raines, R. T. ChemSusChem 2011, 4, 353-356. (230) Sanborn, A. J. (Archer Daniels Midland Company) US Patent 7,317,116, 2008. (231) Chen, J. D.; Kuster, B. F. M.; Van der Wiele, K. Biomass Bioenergy 1991, 1, 217-223. (232) Haworth, W. N.; Hirst, E. L.; Nicholson, V. S. J. Chem. Soc. 1927, 1513. (233) Bicker, M.; Kaiser, D.; Ott, L.; Vogel, H. J. Supercrit. Fluids 2005, 36, 118-126. (234) Tarabanko, V. E.; Smirnova, M. A.; Chernyak, M. Y. Chem. Sus. Dev. 2005, 13, 551-558. (235) Liu, J.; Tang, Y.; Wu, K.; Bi, C.; Cui, Q. Carbohydr. Res. 2012, 350, 20-24. (236) Lai, L.; Zhang, Y. ChemSusChem 2011, 4, 1745-1748. (237) Qu, Y.; Huang, C.; Zhang, J.; Chen, B. Bioresour. Technol. 2012, 106, 170-172. (238) Zhu, H.; Cao, Q.; Li, C.; Mu, X. Carbohydr. Res. 2011, 346, 2016-2018.
270| Chapter 2
(239) Gruter, G. J. M.; Dautzenberg, F.(Avantium International BV) Int. Patent WO 2007104514, 2007. (240) Gruter, G. J. M.; Manzer, L. E.(Furanix Technologies, B. V.) US Patent Appl. 2010/0058650, 2010. (241) Gruter, G. J. M.(Furanix Technologies, B. V.) 2010; US 2010/0218416 A1. (242) Gruter, G. J. M.; Manzer, L. E.(Furanix Technologies, B. V.) US Patent Appl. 2010/0218415, 2010 (243) Gruter, G. J. M.(Furanix Technologies, B. V.) US Patent Appl. 2010/0083565, 2010. (244) Gruter, G. J. M.; Dautzenberg, F. (Furanix Technologies B. V.) US Patent Appl. 2009/0131690, 2009. (245) Gruter, G. J. M.; Dautzenberg, F.; Purmova, J. (Avantium International B. V.) Int. Patent WO 2007104515, 2007. (246) Gruter, G. J. M. (Furanix Technologies B. V.) US Patent Appl. 2010/0212217, 2010. (247) Sanborn, A. J.; Howard, S. J. US Patent Appl 20090156841, 2008. (248) Tong, X.; Li, M.; Yan, N.; Ma, Y.; Dyson, P. J.; Li, Y. Catal. Today 2011, 175, 524-527. (249) Jing, Q.; Lü, X. Chin. J. Chem. Eng. 2008, 16, 890-894. (250) Mednick, M. L. J. Org. Chem 1962, 27, 398-403. (251) Smith, N. H. (Rayonier Inc.) US Patent 3,118,912, 1960. (252) Li, C.; Zhang, Z.; Zhao, Z. K. Tetrahedron Lett. 2009, 50, 5403-5405. (253) Tyrlik, S. K.; Szerszen, D.; Kurzak, B.; Bal, K. Starch/Stärke 1995, 47, 171-174. (254) Tyrlik, S. K.; Szerszen, D.; Olejnik, M.; Danikiewicz, W. J. Mol. Catal. A: Chem. 1996, 106, 223-233. (255) Tyrlik, S. K.; Szerszen, D.; Olejnik, M.; Danikiewicz, W. Carbohydr. Res. 1999, 315, 268- 272. (256) Lourvanij, K.; Rorrer, G. L. Ind. Eng. Chem. Res. 1993, 32, 11-19. (257) Chareonlimkun, A.; Champreda, V.; Shotipruk, A.; Laosiripojana, N. Fuel 2010, 89, 2873- 2880. (258) Chareonlimkun, A.; Champreda, V.; Shotipruk, A.; Laosiripojana, N. Bioresour. Technol. 2010, 101, 4179-4186. (259) Grande, P. M.; Bergs, C.; Dominguez de Maria, P. ChemSusChem 2012. (260) Nakajima, K.; Baba, Y.; Noma, R.; Kitano, M.; Kondo, J. N.; Hayashi, S.; Hara, M. J. Am. Chem. Soc. 2011, 133, 4224-4227. (261) Beckerle, K.; Okuda, J. J. Mol. Catal. A: Chemical 2012, 356, 158-164. (262) Tuteja, J.; Nishimura, S.; Ebitani, K. Bull. Chem. Soc. Japan 2012, 85, 275-281. (263) Takagaki, A.; Ohara, M.; Nishimura, S.; Ebitani, K. Chem. Commun. 2009, 6276-6278. (264) Chen, T.; Lin, L. Chin. J. Chem. 2010, 28, 1773-1776. (265) Yasuda, M.; Nakamura, Y.; Matsumoto, J.; Yokoi, H.; Shiragami, T. Bull. Chem. Soc. Japan 2011, 84, 416-418. (266) Yang, Y.; Hu, C.; Abu-Omar, M. M. Bioresour. Technol. 2012, 116, 190-194. (267) Lew, C. M.; Rajabbeigi, N.; Tsapatsis, M. Ind. Eng. Chem. Res. 2012, 51, 5364-5366. (268) Zhang, Z.; Wang, Q.; Xie, H.; Liu, W.; Zhao, Z. ChemSusChem 2011, 4, 131-138. (269) Wei, Z.; Li, Y.; Thushara, D.; Liu, Y.; Ren, Q. J. Taiwan Inst. Chem. Eng. 2011, 42, 363-370. (270) Yang, Y.; Xiang, X.; Tong, D.; Hu, C.; Abu-Omar, M. M. Bioresour. Technol. 2012, 116, 302-306. (271) Wu, S.; Fan, H.; Xie, Y.; Cheng, Y.; Wang, Q.; Zhang, Z.; Han, B. Green Chem. 2010, 12, 1215-1219. (272) Mok, W. S. L.; Antal, M. J.; Varhegyi, G. Ind. Eng. Chem. Res. 1992, 31, 94-100. (273) Ehara, K.; Saka, S. Cellulose 2002, 9, 301-311. (274) Ehara, K.; Saka, S. J. Wood Sci. 2005, 51, 148-153. (275) Sasaki, M.; Fang, Z.; Fukushima, Y.; Adschiri, T.; Arai, K. Ind. Eng. Chem. Res. 2000, 39, 2883-2890. (276) Yin, S.; Pan, Y.; Tan, Z. Int. J. Green Energy 2011, 8, 234-247. (277) Dutta, S.; De, S.; Alam, M. I.; Abu-Omar, M. M.; Saha, B. J. Catal. 2012, 288, 8-15. (278) Snyder, F. H. (Dendrol Inc.) US Patent 2,851,468, 1958. (279) Amarasekara, A. S.; Ebede, C. C. Bioresour. Technol. 2009, 100, 5301-5304.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|271
(280) Lightner, G. E. US Patent Appl. 2003/0032819, 2003. (281) Lightner, G. E. US Patent Appl. 2002/0123636, 2002. (282) Asghari, F. S.; Yoshida, H. Carbohydr. Res. 2010, 345, 124-131. (283) Dedsuksophon, W.; Faungnawakij, K.; Champreda, V.; Laosiripojana, N. Bioresour. Technol. 2011, 102, 2040-2046. (284) Zhao, S.; Cheng, M.; Li, J.; Tian, J.; Wang, X. Chem. Commun. 2011, 47, 2176-2178. (285) Omari, K. W.; Besaw, J. E.; Kerton, F. M. Green Chem. 2012, 14, 1480-1487. (286) Daengprasert, W.; Boonnoun, P.; Laosiripojana, N.; Goto, M.; Shotipruk, A. Ind. Eng. Chem. Res. 2011, 50, 7903-7910. (287) Wang, C.; Fu, L.; Tong, X.; Yang, Q.; Zhang, W. Carbohydr. Res. 2012, 347, 182-185. (288) Peniston, Q. P. (Food, Chemical and Research Laboratories, Inc.) US Patent Appl. US2750394, 1956. (289) Mercadier, D.; Rigal, L.; Gaset, A.; Gorrichon, J. P. J. Chem. Tech. Biotechnol. 1981, 31, 497-502. (290) Mercadier, D.; Rigal, L.; Gaset, A.; Gorrichon, J. P. J. Chem. Tech. Biotechnol. 1981, 31, 503-508. (291) Rigal, L.; Gorrichon, J. P.; Gaset, A.; Heughebaert, J.-C. Biomass 1985, 7, 27-45. (292) Gaset, A.; Rigal, L.; Paillassa, G.; J.-P., S.; Flèche, G. (Roquette Freres) 1986 US Patent 4,590,283, 1986. (293) Moreau, C.; Durand, R.; Pourcheron, C.; Razigade, S. Ind. Crops Prod. 1994, 3, 85-90. (294) Rivalier, P.; Duhamet, J.; Moreau, C.; Durand, R. Catal. Today 1995, 24, 165-171. (295) Moreau, C.; Durand, R.; Duhamet, J.; Rivalier, P. J. Carbohydr. Chem. 1997, 16, 709-714. (296) Ordomsky, V. V.; van der Schaaf, J.; Schouten, J. C.; Nijhuis, T. A. J. Catal. 2012, 287, 68- 75. (297) Román-Leshkov, Y.; Dumesic, J. A. Top. Catal. 2009, 52, 297-303. (298) Dumesic, J. A.; Román-Leshkov, Y.; Chheda, J. N. (Wisconsin Alumni Research Foundation) Int. Patent WO 2007/146636, 2007. (299) Crisci, A. J.; Tucker, M. H.; Dumesic, J. A.; Scott, S. L. Top. Catal. 2010, 53, 1185-1192. (300) Chheda, J. N.; Román-Leshkov, Y.; Dumesic, J. A. Green Chem. 2007, 9, 342-350. (301) Chheda, J. N.; Dumesic, J. A. Catal. Today 2007, 123, 59-70. (302) Crisci, A. J.; Tucker, M. H.; Lee, M.-Y.; Jang, S. G.; Dumesic, J. A.; Scott, S. L. ACS Catal. 2011, 1, 719-728. (303) McNeff, C. V.; Nowlan, D. T.; McNeff, L. C.; Yan, B.; Fedie, R. L. Appl. Catal. A: Gen. 2010, 384, 65. (304) McNeff, C. V.; Nowlan, D. T. (Sartec Corporation) WO2010/075437, 2008. (305) Hansen, T. S.; Mielby, J.; Riisager, A. Green Chem. 2011, 13, 109-114. (306) Yang, F.; Liu, Q.; Bai, X.; Du, Y. Bioresour. Technol. 2011, 102, 3424-3429. (307) Yang, F.; Liu, Q.; Yue, M.; Bai, X.; Du, Y. Chem. Commun. 2011, 47, 4469-4471. (308) Fan, C.; Guan, H.; Zhang, H.; Wang, J.; Wang, S.; Wang, X. Biomass Bioenergy 2011, 35, 2659-2665. (309) Zhao, Q.; Wang, L.; Zhao, S.; Wang, X.; Wang, S. Fuel 2011, 90, 2289-2293. (310) Brasholz, M.; von Kanel, K.; Hornung, C. H.; Saubern, S.; Tsanaktsidis, J. Green Chem. 2011. (311) Cope, A. C. US2917520, 1959. (312) Pagan-Torres, Y. J.; Wang, T.; Gallo, J. M. R.; Shanks, B. H.; Dumesic, J. A. ACS Catal. 2012, 2, 930-934. (313) Yang, Y.; Hu, C.-w.; Abu-Omar, M. M. Green Chem. 2012, 14, 509-513. (314) Azadi, P.; Carrasquillo-Flores, R.; Pagan-Torres, Y. J.; Guerbuez, E. I.; Farnood, R.; Dumesic, J. A. Green Chem. 2012, 14, 1573-1576. (315) Nikolla, E.; RomaÌ•n-Leshkov, Y.; Moliner, M.; Davis, M. E. ACS Catal. 2011, 1, 408-410. (316) Degirmenci, V.; Pidko, E. A.; Magusin, P. C. M. M.; Hensen, E. J. M. ChemCatChem 2011, 3, 969-972. (317) Yang, L.; Liu, Y.; Ruan, R. Adv. Mat. Res. 2012, 335-336, 1448-1453.
272| Chapter 2
(318) Pedersen, S.; Christensen, T. B.; Boisen, A.; Juergensen, V. W.; Hansen, T. S.; Kegnaes, S.; Riisager, A.; Woodley, J. M.; Jensen, J. S.; Fu, W. (Novozymes AS) Int. Patent WO 2011124639, 2011. (319) Zhuang, J.; Lin, L.; Pang, C.; Liu, Y. Adv. Mat. Res. 2011, 236-238, 134-137. (320) Mascal, M.; Nikitin, E. B. Green Chem. 2010, 12, 370-373. (321) Mascal, M.; Nikitin, E. B. ChemSusChem 2009, 2, 859-861. (322) Mascal, M.; Nikitin, E. B. ChemSusChem 2009, 2, 423-426. (323) Seddon, K. R. J. Chem. Tech. Biotechnol. 1997, 68, 351-356. (324) Walden, P. Bull. Acad. Imper. Sci. (St. Petersburg) 1914, 1800. (325) Holbrey, J. D.; Seddon, K. R. Clean Prod. Proc. 1999, 1, 223-236. (326) Welton, T. Chem. Rev. 1999, 99, 2071-2083. (327) Gordon, C. M. Appl. Catal. A: Gen. 2001, 222, 101-117. (328) Wasserscheid, P.; Keim, W. Angew. Chem., Int. Ed. 2000, 39, 3772-3789. (329) Wilkes, J. S.; Zaworotko, M. J. J. Chem. Soc., Chem. Commun. 1992, 965. (330) Marsh, K.; Deev, A.; Wu, A.; Tran, E.; Klamt, A. Korean J. Chem. Eng. 2002, 19, 357-362. (331) Lansalot-Matras, C.; Moreau, C. Catal. Commun. 2003, 4, 517-520. (332) Fayet, C.; Gelas, J. Carbohydr. Res. 1983, 122, 59-68. (333) Zakrzewska, M. E.; Bogel-Łukasik, E.; Bogel-Łukasik, R. Energy Fuels 2010, 24, 737-745. (334) Moreau, C.; Finiels, A.; Vanoye, L. J. Mol. Catal. A: Chem. 2006, 253, 165-169. (335) Hu, S.; Zhang, Z.; Zhou, Y.; Han, B.; Fan, H.; Li, W.; Song, J.; Xie, Y. Green Chem. 2008, 10, 1280-1283. (336) Cao, Q.; Guo, X.; Yao, S.; Guan, J.; Wang, X.; Mu, X.; Zhang, D. Carbohydr. Res. 2011, 346, 956-959. (337) Lima, S.; Neves, P.; Antunes, M. M.; Pillinger, M.; Ignatyev, N.; Valente, A. A. Appl. Catal. A: Gen. 2009, 363, 93-99. (338) Mittal, N.; Nisola, G. M.; Chung, W.-J. Tetrahedron Lett. 2012, 53, 3149-3155. (339) Jadhav, A. H.; Kim, H.; Hwang, I. T. Catal. Commun. 2012, 21, 96-103. (340) Ma, H.; Zhou, B.; Li, Y.; Argyropoulos, D. S. BioResources 2012, 7, 533-544. (341) Ray, D.; Mittal, N.; Chung, W.-J. Carbohydr. Res. 2011, 346, 2145-2148. (342) Li, C.; Zhao, Z. K.; Cai, H.; Wang, A.; Zhang, T. Biomass Bioenergy 2011, 35, 2013-2017. (343) Cao, Q.; Guo, X.; Guan, J.; Mu, X.; Zhang, D. Appl. Catal. A: Gen. 2011, 403, 98-103. (344) Vigier, K. D. O.; Benguerba, A.; Barrault, J.; Jerome, F. Green Chem. 2012, 14, 285-289. (345) Zhang, Z.; Liu, B.; Zhao, Z. Carbohydr. Polym. 2012, 88, 891-895. (346) Qi, X.; Watanabe, M.; Aida, T. M.; Smith, R. L. Green Chem. 2009, 11, 1327-1331. (347) Qi, X.; Watanabe, M.; Aida, T. M.; Smith Jr., R. L. ChemSusChem 2010, 3, 1071-1077. (348) Qi, X.; Guo, H.; Li, L. Ind. Eng. Chem. Res. 2011, 50, 7985-7989. (349) By Guo, X.; Cao, Q.; Jiang, Y.; Guan, J.; Wang, X.; Mu, X. Carbohydr. Res. 2012, 351, 35- 41. (350) By Guo, F.; Fang, Z.; Zhou, T.-J. Bioresour. Technol. 2012, 112, 313-318. (351) Xie, H.; Zhao, Z. K.; Wang, Q. ChemSusChem 2012, 5, 901-905. (352) Ilgen, F.; Ott, D.; Kralisch, D.; Reil, C.; Palmberger, A.; König, B. Green Chem. 2009, 11, 1948-1954. (353) Liu, F.; Barrault, J.; Vigier, K. D. O.; Jerome, F. ChemSusChem 2011. (354) Tao, F.; Song, H.; Chou, L. RSC Adv. 2011, 1, 672-676. (355) Gruter, G. J. M.; Manzer, L. E.; De Sousa Dias, A. S. V.; Dautzenberg, F.; Purmova, J. (Furanix Technologies B. V.) Eur. Patent 2008/007429, 2008. (356) Hu, L.; Sun, Y.; Lin, L. Ind. Eng. Chem. Res. 2012, 51, 1099-1104. (357) Jadhav, H.; Taarning, E.; Pedersen, C. M.; Bols, M. Tetrahedron Lett., 53, 983-985. (358) Shi, C.; Zhao, Y.; Xin, J.; Wang, J.; Lu, X.; Zhang, X.; Zhang, S. Chem. Commun. 2012, 48, 4103-4105. (359) Wilkes, J. S.; Levisky, J. A.; Wilson, R. A.; Hussey, C. L. Inorg. Chem. 1982, 21, 1263-1264. (360) Qi, X.; Watanabe, M.; Aida, T. M.; Smith, R. L. Bioresour. Technol. 2012, 109, 224-228. (361) Guo, F.; Fang, Z.; Zhou, T.-J. Bioresour. Technol. 2012, 112, 313-318. (362) Chidambaram, M.; Bell, A. T. Green Chem. 2010, 12, 1253-1262. (363) Ståhlberg, T.; Grau Sørensen, M.; Riisager, A. Green Chem. 2010, 12, 321-325.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|273
(364) Khokhlova, E. A.; Kachala, V. V.; Ananikov, V. P. ChemSusChem 2012, 5, 783-789. (365) Zhang, Z.; Zhao, Z. Bioresour. Technol. 2011, 102, 3970-3972. (366) Qi, X.; Watanabe, M.; Aida, T. M.; Smith Jr., R. L. Green Chem. 2010, 12, 1855-1860. (367) Hu, S.; Zhang, Z.; Zhou, Y.; Song, J.; Fan, H.; Han, B. Green Chem. 2009, 11, 873-877. (368) Jiang, F.; Zhu, Q.; Ma, D.; Liu, X.; Han, X. J. Mol. Catal. A: Chem. 2011, 334, 8-12. (369) Su, Y.; Brown, H. M.; Huang, X.; Zhou, X.; Amonette, J. E.; Zang, Z. C. Appl. Catal. A: Gen. 2009, 361, 117-122. (370) Zhang, Z. C.; Brown, H. M.; Su, Y. (Battelle Memorial Institute) Int. Patent WO 2009134631, 2009. (371) Zhang, Z.; Zhao, Z. K. Bioresour. Technol. 2010, 101, 1111-1114. (372) Wang, P.; Yu, H.; Zhan, S.; Wang, S. Bioresour. Technol. 2011, 102, 4179-4183. (373) Hsu, W.-H.; Lee, Y.-Y.; Peng, W.-H.; Wu, K. C.-W. Catal. Today 2011, 174, 65-69. (374) Chun, J.-A.; Lee, J.-W.; Yi, Y.-B.; Hong, S.-S.; Chung, C.-H. Starch/Stärke 2010, 62, 326- 330. (375) Chun, J.-A.; Lee, J.-W.; Yi, Y.-B.; Hong, S.-S.; Chung, C.-H. Korean J. Chem. Eng. 2010, 27, 930-935. (376) Zhang, Y.; Du, H.; Qian, X.; Chen, E. Y. X. Energy Fuels 2010, 24, 2410-2417. (377) Tan, M.; Zhao, L.; Zhang, Y. Biomass Bioenergy 2011, 35, 1367-1370. (378) Bali, S.; Tofanelli, M. A.; Ernst, R. D.; Eyring, E. M. Biomass Bioenergy 2012, 42, 224-227. (379) Yi, Y.-B.; Lee, J.-L.; Choi, Y.-H.; Park, S.-M.; Chung, C.-H. Environ. Chem. Lett. 2012, 10, 13-19. (380) Qi, X.; Watanabe, M.; Aida, T. M.; Smith, R. L., Jr. Cellulose 2011, 18, 1327-1333. (381) Kim, B.; Jeong, J.; Lee, D.; Kim, S.; Yoon, H.-J.; Lee, Y.-S.; Cho, J. K. Green Chem. 2011, 13, 1503-1506. (382) Yi, Y.-B.; Ha, M.-G.; Lee, J.-W.; Chung, C.-H. Chem. Eng. J. 2012, 180, 370-375. (383) Tao, F.; Song, H.; Chou, L. Bioresour. Technol. 2011, 102, 9000-9006. (384) Tao, F.; Song, H.; Yang, J.; Chou, L. Carbohydr. Polym. 2011, 85, 363-368. (385) Tao, F.; Song, H.; Chou, L. Carbohydr. Res. 2011, 346, 58-63. (386) Lee, J.-W.; Ha, M.-G.; Yi, Y.-B.; Chung, C.-H. Carbohydr. Res. 2011, 346, 177-182. (387) Zhao, H.; Holladay, J. E.; Brown, H.; Zhang, Z. C. Topsoe Catalysis Forum, 2008. http://www.topsoe.com/research/~/media/PDF%20files/Topsoe_Catalysis_Forum/Topsoe_Ca talysis_Forum_2008.ashx (accessed on January 20, 2013). (388) Zhao, H.; Brown, H. M.; Holladay, J. E.; Zhang, Z. C. Top. Catal. 2012, 55, 33-37. (389) Zhang, Z.; Liu, W.; Xie, H.; Zhao, Z. K. Molecules 2011, 16, 8463-8474. (390) Su, Y.; Brown, H. M.; Li, G.; Zhou, X.-d.; Amonette, J. E.; Fulton, J. L.; Camaioni, D. M.; Zhang, Z. C. Appl. Catal. A: Gen. 2011, 391, 436-442. (391) Boisen, A.; Christensen, T. B.; Fu, W.; Gorbanev, Y. Y.; Hansen, T. S.; Jensen, J. S.; Klitgaard, S. K.; Pedersen, S.; Riisager, A.; Ståhlberg, T.; Woodley, J. M. Chem. Eng. Res. Des. 2009, 87, 1318-1327. (392) Middendorp, J. A. Rec. trav. chim. 1919, 38, 1. (393) Teunissen, H. Rec. Trav. Chim. Pays-Bas 1930, 49, 784. (394) Cottier, L.; Descotes, G. Trends Heterocycl. Chem. 1992, 2, 223. (395) Reichert, D.; Sarich, M.; Merz, F. (Evonik Degussa GmbH) Eur. Patent EP1958944, 2008. (396) M'Bazoa, C.; Raymond, F.; Rigal, L.; Gaset, A. (Furchim SaRL) Fr. Patent FP2669635, 1990. (397) Bonner, T. G.; Bourne, E. J.; Ruszkiewicz, M. J. Chem. Soc. 1960, 787-791. (398) Yuan, Z.; Xu, C.; Cheng, S.; Leitch, M. Carbohydr. Res. 2011, 346, 2019-2023. (399) Zhao, H.; Holladay, J. E.; Zhang, Z. C. (Battelle Memorial Institute) U.S. Patent Appl. 2008/0033187, 2008. (400) Rigal, L.; Gaset, A.; Gorrichon, J. P. Ind. Eng. Chem. Res. 1981, 20, 719-721. (401) Roman-Leshkov, Y.; Barrett, C. J.; Liu, Z. Y.; Dumesic, J. A. Nature 2007, 447, 982-985. (402) Dignan, C.; Sanborn, A. J. (Archer Daniels Midland Co.) U.S. Patent Appl. 2009030215, 2009. (403) Hales, R. A.; Maistre, J. L. W.; Orth Jr., G. O. (Atlas Chemical Industries Inc.) U.S. Patent Appl. 3071599, 1963. (404) Asghari, F. S.; Yoshida, H. Ind. Eng. Chem. Res. 2007, 46, 7703-7710.
274| Chapter 2
(405) Abdilla, R. M.; Rasrendra, C. B.; Fachri, B. A.; Heeres, H. J. 2011. Unpublished results. (406) Moeller, M.; Harnisch, F.; Schroeder, U. Biomass Bioenergy 2012, 39, 389-398. (407) Nikbin, N.; Caratzoulas, S.; Vlachos, D. G. ChemCatChem 2012, 4, 504-511. (408) Khajavi, S. H.; Kimura, Y.; Oomori, R.; Matsuno, R.; Adachi, S. J. Food Eng. 2005, 68, 309. (409) Saeman, J. F. Ind. Eng. Chem. 1945, 37, 43-52. (410) Heimlich, K. R.; Martin, A. N. J. Am. Pharm. Assoc., Sci. Ed. 1960, 49, 592. (411) McKibbins, S. W.; Harris, J. F.; Saeman, J. F.; Neill, W. K. For. Prod. J. 1962, 12, 17-23. (412) Smith, P. C.; Grethlein, H. E.; Converse, A. O. Sol. Energy 1982, 28, 41-48. (413) Pilath, H. M.; Nimlos, M. R.; Mittal, A.; Himmel, M. E.; Johnson, D. K. J. Agric. Food Chem. 2010, 58, 6131-6140. (414) Bienkowski, P. R.; Ladisch, M. R.; Narayan, R.; Tsao, G. T.; Eckert, R. Chem. Eng. Commun. 1987, 51, 179-192. (415) Baugh, K. D.; McCarty, P. L. Biotechnol. Bioeng. 1988, 31, 50-61. (416) Xiang, Q.; Lee, Y. Y.; Torget, R. W. Appl. Biochem. Biotechnol. 2004, 113-116, 1127. (417) Promdej, C.; Matsumura, Y. Ind. Eng. Chem. Res. 2011, 50, 8492-8497. (418) Chang, C.; Ma, X.; Cen, P. Chin. J. Chem. Eng. 2006, 14, 708. (419) Chang, C.; Ma, X.; Cen, P. Chin. J. Chem. Eng. 2009, 17, 835. (420) Kupiainen, L.; Ahola, J.; Tanskanen, J. Chem. Eng. Res. Des. 2011. (421) Duru, C. E.; Duru, I. A.; Nwagbara, N. K. Macromol.: An Indian J. 2012, 8, 24-27. (422) Yu, Y.; Wu, H. Ind. Eng. Chem. Res. 2011, 50, 10500-10508. (423) Choudhary, V.; Burnett, R. I.; Vlachos, D. G.; Sandler, S. I. J. Phys. Chem. C 2012, 116, 5116-5120. (424) Vanoye, L.; Fanselow, M.; Holbrey, J. D.; Atkins, M. P.; Seddon, K. R. Green Chem. 2009, 11, 390-396. (425) Fagan, R. D.; Grethlein, H. E.; Converse, A. O.; Porteous, A. Environ. Sci. Technol. 1971, 5, 545-547. (426) Thompson, D. R.; Grethlein, H. E. Ind. Eng. Chem. Prod. Res. Dev. 1979, 18, 166-169. (427) Malester, I. A.; Green, M.; Shelef, G. Ind. Eng. Chem. Res. 1992, 31, 1998-2003. (428) Sasaki, M.; Furukawa, M.; Minami, K.; Adschiri, T.; Arai, K. Ind. Eng. Chem. Res. 2002, 41, 6642-6649. (429) Sasaki, M.; Adschiri, T.; Arai, K. AIChE J. 2004, 50, 192-202. (430) Shen, J.; Wyman, C. E. AIChE J. 2011, 85, 236. (431) Guerra-Rodriguez, E.; Portilla-Rivera, O. M.; Jarquin-Enriquez, L.; Ramirez, J. A.; Vazquez, M. Biomass Bioenergy 2012, 36, 346-355. (432) Dee, S. J.; Bell, A. T. ChemSusChem 2011, 4, 1166-1173. (433) Yi, Y.-B.; Lee, J.-W.; Hong, S.-S.; Choi, Y.-H.; Chung, C.-H. J. Ind. Eng. Chem. 2011, 17, 6- 9. (434) Heyraud, A.; Rinaudo, M.; Taravel, F. R. Carbohydr. Res. 1984, 128, 311. (435) Blecker, C.; Fougnies, C.; Van Herck, J. C.; Chevalier, J. P.; Paquot, M. J. Agric. Food Chem. 2002, 50, 1602-1607. (436) L'Homme, C.; Arbelot, M.; Puigserver, A.; Biagini, A. J. Agric. Food Chem. 2003, 51, 224- 228. (437) Luijkx, C. A.; Van Rantwijk, F.; Van Bekkum, H. Carbohydr. Res. 1993, 242, 131-139. (438) Chuntanapum, A.; Yong, T. L. K.; Miyake, S.; Matsumara, Y. Ind. Eng. Chem. Res. 2008, 47, 2956. (439) Teunissen, H. Rec. Trav. Chim. Pays-Bas 1931, 50, 1. (440) Gruter, G. J. M.(Furanix Technologies, B. V.) U.S. Patent Appl. 2010/0083565, 2010. (441) Ondrey, G. Chem. Eng. 2009, June, 14. (442) Tuercke, T.; Panic, S.; Loebbecke, S. Chem. Eng. Technol. 2009, 32, 1815-1822. (443) Schoen, M.; Schnuerch, M.; Mihovilovic, M. D. Mol. Divers. 2011, 15, 639-643. (444) Laurent, F.; Seguinaud, A.-l.; Faugeras, P. (Sucreries, Raffineries D'Erstein) Int. Patent WO2005018799, 2005. (445) Haworth, W. N.; Wiggins , L. F. U.S. Patent 2498918, 1950. (446) Carter, B.; Gilcrease, P. C.; Menkhaus, T. J. Biotechnol. Bioeng. 2011, 108, 2046-2052. (447) Geier, D. F.; Soper, J. G. (Archer Daniels Midland Company) WO2008157617, 2008.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|275
(448) Klasson, K. T.; Uchimiya, M.; Lima, I. M.; Boihem, L. L., Jr. BioResources 2011, 6, 3242- 3251. (449) Hunter, R. H.; Atlas, C. I. (Atlas Chemical Industries Inc.) U.S. Patent 3201331, 1965. (450) Jones, R. E.; Lange, H. B. (Merch & Co., Inc.) U.S. Patent 2994645, 1961. (451) Chan, J. Y. G.; Zhang, Y. ChemSusChem 2009, 2, 731-734. (452) Wei, Z.; Liu, Y.; Thushara, D.; Ren, Q. Green Chem. 2012, 14, 1220-1226. (453) Liu, W.; Holladay, J. E.; Zheng, F.; Brown, H. M.; Cooper, A. R. (Batelle Memorial Institute) U.S. Patent Appl. 2011105770, 2011. (454) Blanksma, J. J. Ber. 1910, 43, 2355. (455) Fleche, G.; Gaset, A.; Gorrichon, J.-P.; Truchot, E.; Sicard, P. (Roquette Freres) U.S. Patent 4339387, 1982. (456) Torres, A. I.; Daoutidis, P.; Tsapatsis, M. Energy Environ. Sci. 2010, 3, 1560. (457) Kazi, F. K.; Patel, A. D.; Serrano-Ruiz, J. C.; Dumesic, J. A.; Anex, R. P. Chem. Eng. J. 2011, 169, 329. (458) Torres, A. I.; Tsapatsis, M.; Daoutidis, P. Comput. Chem. Eng. 2012, 42, 130-137. (459) Lichtenthaler, F. W.; Peters, S. C. R. Chimie 2004, 7, 65–90. (460) Lichtenthaler, F. W. Industrial Chemicals from Carbohydrate Feedstocks: Current Status and Challenges Ahead. In Production of Fuels, Specialty Chemicals and Biobased Products from Agro-Industrial Waste; Fava, F., Canepa, P., Eds.; INCA: Venezia, Italy, 2008; p 230. (461) Gandini, A.; Belgacem, M. N. Prog. Polym. Sci. 1997, 22, 1203-1379. (462) Gandini, A. Green Chem. 2011. (463) Utne, T.; Garber, J. D.; Jones, R. E. (Merck & Co, Inc,) U.S. Patent 3083236, 1963. (464) Amarasekara, A. S. In Renewable Polymers; Mittal, V., Ed.; Scrivener Publishing LLC, 2012. (465) Timko, J. M.; Cram, D. J. J. Am. Chem. Soc. 1974, 96, 7159-7160. (466) Schiavo, V.; Descotes, G.; Mentech, J. Bull. Soc. Chim. France 1991, 128, 704-711. (467) Lilga, M. A.; Hallen, R. T.; Werpy, T. A.; White, J. F.; Holladay, J. E.; Frye, J. R.; Zacher, A. H. (Battelle Memorial Institute) U.S. Patent Appl. 2007/0287845, 2007. (468) Buntara, T.; Noel, S.; Phua, P. H.; Melián-Cabrera, I.; de Vries, J. G.; Heeres, H. J. Angew. Chem. Int. Ed. 2011, 50, 7083-7087. (469) Boopathy, R.; Bokang, H.; Daniels, L. J. Ind. Microbiol. 1993, 11, 147-150. (470) Ras, E.-J.; McKay, B.; Rothenberg, G. Top. Catal. 2010, 53, 1202–1208. (471) Liu, Z. L. (The Unites States of America as represented by the Secretary of Agriculture) Int. Patent WO2009006135, 2009. (472) Pentz, W. J. (The Quaker Oats Company) Gr. Brit. Patent 2131014, 1984. (473) Sanborn, A. J.; Bloom, P. D. (Archer Daniels Midland Company) U.S. Patent 7393963, 2008. (474) Connolly, T. J.; Considine, J. L.; Ding, Z.; Forsatz, B.; Jennings, M. N.; MacEwan, M. F.; McCoy, K. M.; Place, D. W.; Sharma, A.; Sutherland, K. Org. Proc. Res. Dev. 2010, 14, 459- 465. (475) Alamillo, R.; Tucker, M.; Chia, M.; Pagan-Torres, Y.; Dumesic, J. Green Chem. 2012, 14, 1413-1419. (476) Gomes, M.; Gandini, A.; Silvestre, A. J. D.; Reis, B. J. Polym. Sci. Part A: Polym. Chem. 2011. (477) Bonsignore, B. V. (Stepan Company) Int. Patent WO2008022287, 2008. (478) Benecke, H. P.; King, J. L., II; Kawczak, A. W.; Zehnder, D. W.; Hirschl, E. E. (Batelle Memorial Institute) Int. Patent WO2007/092569, 2007. (479) Grushin, V.; Herron, N.; Halliday, G. A. (E. I. du Pont de Nemours & Co.) Int. Patent WO2003024947, 2003. (480) Cottier, L.; Descotes, G.; Lewkowski, J.; Skowronski, R. Org. Prep. Proc. Int. 1995, 27, 564- 566. (481) Lilga, M. A.; Hallen, R. T.; Hu, J.; White, J. F.; Gary, M. J. (Batelle Memorial Institute) Int. Patent WO2008/054804, 2008. (482) Amarasekara, A. S.; Green, D.; McMillan, E. Catal. Commun. 2008, 9, 286-288. (483) Yoon, H.-J.; Choi, J.-W.; Jang, H.-S.; Cho, J. K.; Byun, J.-W.; Chung, W.-J.; Lee, S.-M.; Lee, Y.-S. Synlett 2011, 165-168. (484) Takada, H. (Kao Corporation) Jpn. Patent 03101672, 1991.
276| Chapter 2
(485) Haas, T.; Tacke, T.; Pfeffer, J. C.; Klasovsky, F.; Rimbach, M.; Volland, M.; Ortelt, M. (Evonik Degussa GmbH) Int. Patent WO 2012004069, 2012. (486) Durand, G.; Faugeras, P.; Laporte, F.; Moreau, C.; Neau, M.-C.; Roux, G.; Tichit, D.; Toutremepuich, C. (Agrichimie) Int. Patent WO9617836, 1996. (487) Martin, T. (Episucres, S. A.) Ger. Patent DE 19615878, 1997. (488) Nie, J.; Liu, H. Pure Appl. Chem. 2012, 84, 765-777. (489) Halliday, G. A.; Young, R. J.; Grushin, V. V. Org. Lett. 2003, 5, 2003-2005. (490) Ma, J.; Du, Z.; Xu, J.; Chu, Q.; Pang, Y. ChemSusChem 2011, 4, 51-54. (491) Hu, C.; Xiang, X.; Yang, Y.; Guo, B.; Tong, D.; Zhu, L. (Sichuan University) Chin. Patent CN 101768142, 2010. (492) Carlini, C.; Patrono, P.; Raspolli Galletti, A. M.; Sbrana, G.; Zima, V. Appl. Catal. A: General 2005, 289, 197-204. (493) Hanson, S. K.; Wu, R.; Silks, L. A. Org. Lett. 2011, 13, 1908-1911. (494) Partenheimer, W.; Grushin, V. V. Adv. Synth. Catal. 2001, 343, 102-111. (495) Sanborn, A. (Archer Daniels Midland Company) Int. Patent WO2010132740, 2009. (496) Yadav, G. D.; Sharma, R. V. 2012; Int. Patent WO 2012073251, 2012. (497) Takagaki, A.; Takahashi, M.; Nishimura, S.; Ebitani, K. ACS Catal. 2011, 1, 1562-1565. (498) Skowronski, R.; Cottier, L.; Descotes, G.; Lewkowski, J. Synthesis 1996, 1291-1292. (499) van Deurzen, M. P. J.; van Rantwijk, F.; Sheldon, R. A. J. Carbohydr. Chem. 1997, 16, 299- 309. (500) Vinke, P.; Poel, W. V.; Van Bekkum, H. Stud. Surf. Sci. Catal. 1991, 59, 385. (501) Yutaka, K.; Miura, T.; Eritate, S.; Komuro, T. U.S. Patent Appl. 20110092720, 2011. (502) Ribeiro, M. L.; Schuchardt, U. Catal. Commun. 2003, 4, 83. (503) Merat, N.; Verdeguer, P.; Rigal, L.; Gaset, A.; Delmas, M. (Furchim) Fr. Patent FR 2669634, 1992. (504) Verdeguer, P.; Merat, N.; Gaset, A. J. Mol. Catal. A: Chem. 1993, 85, 327–344. (505) Lew, B. W. U.S. Patent 3,326,944, 1967. (506) Leupold, E. I.; Wiesner, M.; Schlingmann, M.; Kapp, K. (Hoechst AG) Eur. Patent EP 0356703, 1993. (507) Lilga, M. A.; Hallen, R. T.; White, J. F.; Frye, G. J. (Battelle Memorial Institute) U.S. Patent Appl. 20080103318, 2008. (508) Lilga, M. A.; Hallen, R. T.; Gray, M. Top. Catal. 2010, 53, 1264–1269. (509) Kröger, M.; Prüsse, U.; Vorlog, K.-E. Top. Catal. 2000, 13, 237–242. (510) Gorbanev, Y. Y.; Klitgaard, S. K.; Woodley, J. M.; Christensen, C. H.; Riisager, A. ChemSusChem 2009, 2, 672-675. (511) Gupta, N. K.; Nishimura, S.; Takagaki, A.; Ebitani, K. Green Chem. 2011, 13, 824. (512) Davis, S. E.; Houk, L. R.; Tamargo, E. C.; Datye, A. K.; Davis, R. J. Catal. Today 2011, 160, 55–60. (513) Grabowski, G.; Lewkowski, J.; Skowroński, R. Electrochim. Acta 1991, 36, 1995. (514) Gorbanev, Y. Y.; Kegnaes, S.; Riisager, A. Top. Catal. 2011, 54, 1318-1324. (515) Gorbanev, Y. Y.; Kegnaes, S.; Riisager, A. Catal. Lett. 2011, 141, 1752-1760. (516) Pasini, T. P., Marco; Blosi, Magda; Bonelli, Rosa; Albonetti, Stefania; Dimitratos, Nikolaos; Lopez-Sanchez, Jose A.; Sankar, Meenakshisundaram; He, Qian; Kiely, Christopher J.; et al Green Chem. 2011, 13, 2091-2099. (517) Miura, T.; Kakinuma, H.; Kawano, T.; Matsuhisa, H. (Canon Kabushiki Kaisha) U.S. Patent 7411078, 2008. (518) Muñoz de Diego, C.; Schammel, W. P.; Dam, M. A.; Gruter, G. J. M. (Furanix Technologies B.V.) Int. Patent WO2011/043660, 2011. (519) Muñoz de Diego, C.; Schammel, W. P.; Dam, M. A.; Gruter, G. J. M. (Furanix Technologies B.V.) Int. Patent WO2011/043661, 2011. (520) Borsotti, G.; Digioia, F. Int. Patent WO 2012017052, 2012. (521) Casanova, O.; Iborra, S.; Corma, A. J. Catal. 2009, 265, 109-116. (522) Zope, B. N.; Davis, S. E.; Davis, R. J. Top. Catal. 2012, 55, 24-32. (523) Saha, B.; Dutta, S.; Abu-Omar, M. M. Catal. Sci. Technol. 2012, 2, 79-81.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|277
(524) Davis, S. E.; Houk, L. R.; Tamargo, E. C.; Datye, A. K.; Davis, R. J. Catal. Today 2011, 160, 55-60. (525) Munoz de Diego, C.; Schammel, W. P.; Dam, M. A.; Gruter, G. J. M. (Furanix Technologies B.V.) Int. Patent WO 2011043660, 2011. (526) Drewitt, J. G. N.; Lincoln, J., ; Gr. Brit. Patent GB621,971, 1946. (527) Gruter, G. J. M.; Sipos, L.; Dam, M. A. Comb. Chem. High Throughput Screening 2011. (528) Eerhart, A. J. J. E.; Faaij, A. P. C.; Patel, M. K. Energy Environ. Sci. 2012, 5, 6407-6422. (529) Grass, M.; Becker, H. G. (Evonik Oxeno GmbH) Int. Patent WO 2011023491, 2011. (530) Gaset, A.; Rigal, L.; Sene, B.; Ralainirina, R.; Demande, F. (Agro Industrie Recherches et Developpements (ARD)) Fr. Patent FR2723946, 1996. (531) Taarning, E.; Nielsen, I. S.; Egeblad, K.; Madsen, R.; Christensen, C. ChemSusChem 2008, 1, 75–78. (532) Taarning, E.; Christensen, C. H. Chem. Today 2007, 25, 70–73. (533) Boussie, T. R.; Dias, E. L.; Fresco, Z. M.; Murphy, V. J. (Rennovia, Inc.) Int. Patent WO 2010144873, 2010. (534) Sato, Y.; Hagitani, K.; Koho, J. K. T., (Toyo Kasei Kogyo Co., Ltd) Jpn. Patent JP2006257004, 2006. (535) Benecke, H. P.; Kawczak, A. W.; Garbark, D. B.; (Battelle Memorial Institute) U.S. Patent Appl. 20080207847, 2008. (536) Moore, J. A.; Kelly, J. E. Org.Prep. Proc. Int. 1972, 4, 289-292. (537) Mndzhoyan, A. L.; Adzhibekyan, A. S.; Sanasaryan, A. A. Armyanskii Khimicheskii Zhurnal 1969, 22, 933-935. (538) Mitsukura, K.; Sato, Y.; Yoshida, T.; Nagasawa, T. Biotech. Lett. 2004, 26, 1643-1648. (539) Koopman, F.; Wierckx, N.; De Winde, J. H.; Ruijssenaars, H. J. Bioresour. Technol. 2010, 101, 6291-6296. (540) Hanke, P. D., 2009 (Archer Daniels Midland Co.) Int. Patent WO2009/023174, 2009. (541) Koopman, F.; Wierckx, N.; De Winde, J. H.; Ruijssenaars, H. J. (DSM IP Assets BV) Int. Patent WO2011/026913, 2011. (542) Kegnaes, S.; Mielby, J.; Mentzel, U. V.; Jensen, T.; Fristrup, P.; Riisager, A. Chem. Commun. 2012, 48, 2427-2429. (543) Dutta, S.; De, S.; Saha, B. ChemPlusChem 2012, 77, 259-272. (544) Yoshida, N.; Kasuya, N.; Haga, N.; Fukuda, K. Pol. J. 2008, 40, 1164–1169. (545) Casanova, O.; Iborra, S.; Corma, A. J. Catal. 2010, 275, 236-242. (546) Shiramizu, M.; Toste, F. D. Chem. Eur. J. 2011, 17, 12452-12457. (547) Lukevics, E.; Ignatovich, L.; Shestakova, I. Appl. Organometal. Chem. 2003, 17, 898-905. (548) Lewis, T. A.; Bayless, L.; Eckman, J. B.; Ellis, J. L.; Grewal, G.; Libertine, L.; Nicolas, J. M.; Scannell, R. T.; Wels, B. F.; Wenberg, K.; Wypij, D. M. Bioorg. Med. Chem. Lett. 2004, 14, 2265-2268. (549) Lin, L.; Shi, Q.; Nyarko, A. K.; Bastow, K. F.; Wu, C.-C.; Su, C.-Y.; Shih, C. C.-Y.; Lee, K.- H. J. Med. Chem. 2006, 49, 3963-3972. (550) Villain-Guillot, P.; Gualtieri, M.; Bastide, L.; Roquet, F.; Martinez, J.; Amblard, M.; Pugniere, M.; Leonetti, J.-P. J. Med. Chem. 2007, 50, 4195-4204. (551) Schinzer, D.; Bourguet, E.; Ducki, S. Chem. Eur. J. 2004, 10, 3217-3224. (552) Descotes, G.; Cottier, L.; Eymard, L.; Rapp, K. M. (Südzucker Aktien Gesellschaft) Eur. Patent EP584585, 1994. (553) Descotes, G.; Cottier, L.; Eymard, L.; Rapp, K. M. Synthesis 1995, 303-306. (554) Shao, X.; Li, Z.; Qian, X.; Xu, X. J. Agric. Food Chem. 2009, 57, 951-957. (555) Ottinger, H.; Soldo, T.; Hofmann, T. J. Agric. Food Chem. 2003, 51, 1035-1041. (556) Villard, R.; Robert, F.; Blank, I.; Bernardinelli, G.; Soldo, T.; Hofmann, T. J. Agric. Food Chem. 2003, 51, 4040-4045. (557) Koch, J.; Pischetsrieder, M.; K., P.; Severin, T. Carbohydr. Res. 1998, 313, 117-123. (558) Frank, O.; Ottinger, H.; Hofmann, T. J. Agric. Food Chem. 2001, 49, 321-238. (559) T., S.; Hofmann, T. J. Agric. Food Chem. 2005, 53, 9165-9171. (560) Terada, I.; Takeda, T.; Kobayashi, T.; Hiramoto, T.; Tsuyoshi, K. (Takasago International Corporation) Int. Patent WO2008044784, 2008.
278| Chapter 2
(561) Quiroz-Florentino, H.; Aguilar, R.; Santoyo, B. M.; Diaz, F.; Tamariz, J. Synthesis 2008, 1023-1028. (562) Quiroz-Florentino, H.; Hernandez-Benitez, R. I.; Avina, J. A.; Burgueno-Tapia, E.; Tamariz, J. Synthesis 2011, 1106-1112. (563) Goswami, S.; Dey, S.; Jana, S. Tetrahedron 2008, 64, 6358. (564) Cummings, R. T.; DiZio, J. P.; Krafft, G. A. Tetrahedron Lett. 1988, 29, 69-72. (565) Almirante, N.; Cerri, A.; Fedrizzi, G.; Marazzi, G.; Santagostino, M. Tetrahedron Lett. 1988, 39, 3287-3290. (566) Muller, C.; Diehl, V.; Lichtenthaler, F. W. Tetrahedron 1998, 54, 10703-10712. (567) Lichtenthaler, F. W.; Brust, A.; Cuny, E. Green Chem. 2001, 3, 201-209. (568) Amarasekara, A. S.; Edigin, O.; Hernandez, W. Lett. Org. Chem. 2007, 4, 306-308. (569) Undheim, K.; J., R.; Greibrok, T. Acta Chem. Scand. 1969, 23, 2501-2504. (570) Sattler, L.; Zerban, F. W.; Clark, G. L.; Chu, C. C. J. Am. Chem. Soc. 1951, 73, 5908-5910. (571) Hanefeld, W.; Schlitzer, M.; Debski, N.; Euler, H., J. Heterocycl. Chem. 1996, 33, 1143- 1146. (572) De Freitas Filho, J. R.; Srivastava, R. M.; Soro, Y.; Cottier, L.; Descotes, G. J. Carbohydr.Chem. 2001, 20, 561-568. (573) Cottier, L.; Descotes, G.; Soro, Y. J. Carbohydr. Chem. 2005, 24, 55-71. (574) Bognar, R.; Herczegh, P.; Zsely, M.; Batta, G. Carbohydr. Res. 1987, 164, 465-469. (575) McDermott, P. J.; Stockman, R. A. Org. Lett. 2005, 7, 27-29. (576) Tang, L.; Wang, Q.; Liu, X.; Wang, R.; Deng, C.; Chen, W. (University of Jiangnan) Chin. Patent CN101293884, 2008. (577) West, R. M.; Liu, Z. Y.; Peter, M.; Gartner, C. A.; Dumesic, J. A. J. Mol. Catal. A: Chemical 2008, 296, 18-27. (578) Barrett, C. J.; Chheda, J. N.; Huber, G. W.; Dumesic, J. A. Appl. Catal. B: Environ. 2006, 66, 111-118. (579) Wang, Q.; Guerrero, V. V.; Ghosh, A.; Yeu, S.; Lunn, J. D.; Shantz, D. F. J. Catal. 2010, 269, 15-25. (580) Miller, R. Acta Chem. Scand. Ser. B 1987, 41, 208-209. (581) Kashiwagi, M.; Fuhshuku, K.-I.; Sugai, T. J. Mol. Catal. B: Enzym. 2004, 29, 249-258. (582) Iovel, I.; Mertins, K.; Kischel, J.; Zapf, A.; Beller, M. Angew. Chem. Int. Ed. 2005, 44, 3913- 3917. (583) Yu, C.; Liu, B.; Hu, L. J. Org. Chem. 2001, 66, 5413-5418. (584) Yu, C.; Hu, L.; 2002, 219-223. J. Org. Chem. 2002, 67, 219-223. (585) Papadogianakis, G.; Maat, L.; Sheldon, R. A. J. Chem. Soc. Chem. Commun. 1994, 2659- 2660. (586) Lin, G.; Chen, P.; Han, S.; Li, Z.; Huang, H. (Shanghai Institute of Organic Chemistry) Chin. Patent CN1308074, 2001. (587) Geilen, F. M. A.; Vom Stein, T.; Engendahgl, B.; Winterle, S.; Liauw, M. A.; Klankermayer, J.; Leitner, W. Angew. Chem. Int. Ed. 2011, 50, 6831-6834. (588) Chundury, D.; Szmant, H. H. Ind. Eng. Chem. Prod. Res. Dev. 1981, 158-163. (589) Larousse, C.; Rigal, L.; Gaset, A. J. Chem. Tech. Biotech. 1992, 53, 111-116. (590) Neumann, C. N. D.; Bulach, W. D.; Rehahn, M.; Klein, R. Macromol. Rapid. Commun. 2011, 32, 1373-1378. (591) Cottier, L.; Descotes, G.; Nigay, H.; Parron, J. C.; Gregoire, V. Bull. Soc. Chim. Fr. 1986, 844–850. (592) El Hajj, T.; Martin, J. C.; Descotes, G. J. Heterocycl. Chem. 1983, 20, 233-235. (593) Balakrishnan, M.; Sacia, E. R.; Bell, A. T. Green Chem. 2012, 14, 1626-1634. (594) Lanzafame, P.; Temi, D. M.; Perathoner, S.; Centi, G.; Macario, A.; Aloise, A.; Giordano, G. Catal. Today 2011, 175, 435-441. (595) Jogia, M. K.; Vakamoce, V.; T., W. R. Aust. J. Chem. 1985, 38, 1009-1016. (596) Gupta, P.; Singh, S. K.; Pathak, A.; Kundu, B. Tetrahedron 2002, 58, 10469-10474. (597) Sanda, K.; Rigal, L.; Gaset, A. Carbohydr. Res. 1989, 187, 15-23. (598) Sanda, K.; Rigal, L.; Delmas, M.; Gaset, A. Synthesis 1992, 541-542. (599) Sanda, K.; Rigal, L.; Gaset, A. J. Chem. Tech. Biotechnol. 1992, 55, 139-145.
Hydroxymethylfurfural, a versatile platform chemical made from renewable resources|279
(600) Ramonczai, J.; Vargha, L. J. Am. Chem. Soc. 1950, 72, 2737. (601) Ohta, M.; Iwasaki, M.; Kouno, K.; Ueda, Y. Chem.Pharm. Bull. 1985, 33, 2862-2865. (602) Hashmi, A. S. K.; Weyrauch, J. P.; Kurpejovic, E.; Frost, T. M.; Michlich, B.; Frey, W.; Bats, J. W. Chem. Eur. J. 2006, 12, 5806-5814. (603) Wahhab, A. J. Am. Chem. Soc. 1948, 70, 3580-35892. (604) Erdmann, E. Ber. 1910, 43, 2391-2398. (605) Erdmann, E.; Schaefer, C. Ber. 1910, 43, 2398-2406. (606) Gardner, T. S.; Smith, F. A.; Wenis, E.; Lee, J. J. Org. Chem. 1951, 16, 1121-1125. (607) Cottier, L.; Descotes, G.; Lewkowski, J.; Skowronski, R. Phosphorus, Sulfur, Silicon Relat. Elem. 1996, 116, 93-100. (608) McNelis, B. J.; Sternbach, D. D.; MacPhail, A. T. Tetrahedron 1994, 50, 6767-6782. (609) Cottier, L.; Descotes, G.; Lewkowski, J. Synth. Commun. 1994, 24, 939-944. (610) Mouloungui, Z.; Delmas, M.; Gaset, A. Synth. Commun. 1985, 15, 491-494. (611) Goodman, S. N.; Jacobsen, E. N. Adv. Synth. Catal. 2002, 344, 953-956. (612) Fumagalli, T.; Sello, G.; Orsini, F. Synth. Commun. 2009, 39, 2178-2195. (613) Gruter, G. J.; De Jong, E. Biofuels Technol. 2009, 1, 11-17. (614) De Jong, E.; Vijlbrief, T.; Hijkoop, R.; Gruter, G. J.; Van der Waal, J. C. Biomass Bioenergy 2012, 36, 151. (615) Alonso, M. D.; Bond, J. Q.; Dumesic, J. A. Green Chem. 2010, 12, 1493-1513. (616) Kazi, F. K.; Patel, A. D.; Serrano-Ruiz, J. C.; Dumesic, J. A.; Anex, R. P. Chem. Eng. J. 2011, 169, 329. (617) Luijkx, G. C. A.; Huck, N. P. M.; Van Rantwijk, F.; Maat, L.; Van Bekkum, H. Heterocycles 2009, 77, 1037-1044. (618) Huber, G. W.; Chheda, J. N.; Barrett, C. J.; Dumesic, J. A. Science 2005, 308, 1446-1450. (619) Silks, L. A.; Gordon, J. C.; Wu, R.; Hanson, S. K. (Los Alamos National Security, LLC) Int. Patent WO 2011022042, 2011. (620) Shen, W.; Tompsett, G. A.; Hammond, K. D.; Xing, R.; Dogan, F.; Grey, C. P.; Conner, W. C., Jr.; Auerbach, S. M.; Huber, G. W. Appl. Catal. A: Gen. 2011, 392, 57-68. (621) Geilen, F. M. A.; Engendahl, B.; Harwardt, A.; Marquardt, W.; Klankermayer, J.; Leitner, W. Angew. Chem. 2010, 122, 5642-5646. (622) Hayes, D. J.; Fitzpatrick, S.; Hayes, M. H. B.; Ross, J. R. H. In Biorefineries: Industrial Processes and Products, Vol. 1; Kamm, B., Gruber, P. R., Kamm, M., Eds.; Wiley-VCH: Weinheim, 2006. (623) Horváth, I. T.; Mehdi, H.; Fábos, V.; Boda, L.; Mika, L. T. Green Chem. 2008, 10, 238-242. (624) Lange, J. P.; Price, R.; Ayoub, P. M.; Louis, J.; Petrus, L.; Clarke, L.; H., G. Angew. Chem. Int. Ed. 2010, 49, 1-6. (625) Bond, J. Q.; Martin Alonso, D.; Wang, D.; West, R. M.; Dumesic, J. A. Science 2010, 327, 1110-1114. (626) Dumesic, J. A.; Roman-Leshkov, Y. (Wisconsin Alumni research Foundation) Int. Patent WO2008/151178, 2008. (627) Dumesic, J. A.; Ruiz, J. C. S.; West, R. M. (Wisconsin Alumni research Foundation) Int. Patent WO2010/151343, 2010. (628) Gruter, G. J. M.; Dautzenberg, F. (Furanix Technologies BV) Eur. Patent EP1,834,950, 2007. (629) Gruter, G. J. M.; De Jong, E. (Furanix Technologies BV) Int. Patent WO2009/141166, 2009. (630) Gruter, G. J. M.; De Jong, E. (Furanix Technologies BV) Eur. Patent EP2128227, 2009. (631) Gruter, G. J. M. (Furanix Technologies BV) Eur. Patent EP2034005, 2009. (632) Gruter, G. J. M. (Furanix Technologies BV) U.S. Patent Appl. US2010212218, 2010. (633) Van de Graaf, W. D.; Lange, J.-P. (Shell Internationale Research Maatschappij BV) Int. Patent WO2007/023173, 2007. (634) Haan, R. J.; Lange, J.-P. (Shell Internationale Research Maatschappij BV) Int. Patent WO2009/077606, 2009. (635) Lange, J.-P. (Shell Internationale Research Maatschappij BV) Int. Patent WO2009/156439, 2009. (636) Felix-Moore, A.; Lange, J.-P.; Smith, J. (Shell Internationale Research Maatschappij BV) Int. Patent WO2010/136436, 2010.
280| Chapter 2
(637) Felix-Moore, A.; Lange, J.-P.; Smith, J.; Price, R. J. (Shell Internationale Research Maatschappij BV) Int. Patent WO2010/136437, 2010. (638) Maki-Arvela, P.; Salminen, E.; Riittonen, T.; Virtanen, P.; Kumar, N.; Mikkola, J.-P. Int. J. Chem. Eng. 2012, 674761.
|281
3 The dehydration of different ketoses and aldoses to 5-hydroxymethylfurfural
Abstract
5-Hydroxymethylfurfural is considered an important biobased building block for future biobased chemicals. Here we present an experimental study using different ketoses (fructose, sorbose, tagatose) and aldoses (glucose, mannose, galactose) under aqueous acidic conditions -1 (65 gL substrate, 100-160 °C, 33-300 mM H2SO4) to gain insights in reaction pathways for hexose dehydration to HMF. Both reaction rates and HMF selectivities were significantly higher for ketoses than for aldoses, which is in line with literature. Screening and kinetic experiments showed that the reactivity of the different ketoses is a function of the hydroxyl group orientation at the C3 and C4 positions. These results, in combination with DFT calculations, point to a dehydration mechanism involving cyclic intermediates. For aldoses, no influence of the hydroxyl group orientation was observed, indicating a different rate- determining step. The combination of the knowledge from the literature and the findings in this work indicates that aldoses require an isomerisation to ketose prior to dehydration to obtain high HMF yields.
* Published as ‘The dehydration of different ketoses and aldoses to 5-hydroxymethylfurfural,’ Van Putten, R.-J.; Soetedjo, J. N. M.; Pidko, E. A.; Van der Waal, J. C.; Hensen, E. J. M.; De Jong, E.; Heeres, H. J. ChemSusChem 2013, 6, 1681-1687.
282| Chapter 3
3.1 Introduction
The replacement of fossil feedstocks with sustainable resources for energy generation, transportation fuels, bulk and fine chemicals and materials is currently considered as a pivotal challenge, receiving increasing social and scientific interest. To achieve this, alternative sources of organic carbon need to be identified. Biomass is an attractive option as it is the most abundant non-fossil source of organic carbon. Biomass mainly comprises carbohydrates, lignin, fatty acids, lipids and proteins. Carbohydrates represent the largest fraction of biomass, predominantly present in polymeric form (cellulose, hemi-cellulose, starch, inulin) and are built up from hexoses (glucose, fructose, mannose, galactose) and pentoses (arabinose, xylose). The acid-catalysed dehydration of pentoses1-3 and hexoses4 leads to the formation of furfural and 5-hydroxymethylfurfural (HMF), respectively, along with many by-products. Both molecules, and derivatives thereof, are in Bozell’s ‘Top 10 + 4’ list of biobased chemicals and are considered to be key components in the development of a biobased economy.5 This has led to an enormous increase in research published on acid- catalysed dehydration of sugars over the last decade.3,4
Both furfural and HMF can be used in different application areas. Furfural has high potential in fuel and solvent applications. HMF is considered a promising platform chemical due to its high derivatisation potential. It can be converted to a wide range of interesting bulk and fine chemicals, for instance as a monomer for novel biobased polymers. Avantium is currently developing a process for the production of polyethylenefurandicarboxylate (PEF) from C6 sugars as a next generation replacement material for PET, having improved barrier properties.6-9
The attention for the development of highly efficient routes to HMF has strongly increased in recent years,4 Glucose or glucose-based oligomers and polymers, especially those derived from lignocellulosic sources, are favoured feedstocks due to their availability and presence in agricultural side streams and other waste.10,11 The vast majority of experimental studies, however, shows that fructose, a ketose, is much more efficiently dehydrated to HMF than glucose, an aldose. Under aqueous acidic conditions fructose yields a maximum of around 50% HMF at best, due to the formation of polymeric material, known as humins, and hydration of HMF to levulinic and formic acids (Scheme 1).4,12,13 For glucose the maximum HMF yield is only around 5%. Higher HMF yields from fructose (>80%) have been reported in other solvent systems, especially in ionic liquids and aprotic polar solvents The dehydration of different ketoses and aldoses to 5-hydroxymethylfurfural |283 such as DMSO.4 Work on glucose dehydration with heterogeneous base/acid bi-catalytic systems,14-17 chromium catalysed glucose dehydration in ionic liquids18,19 and organic solvents,20 and other catalysts,21,22 indicate that apart from an acid, an additional catalyst is required to efficiently dehydrate glucose to HMF. This additional catalyst is generally believed to facilitate the isomerisation of glucose to fructose prior to dehydration to HMF.4
Scheme 1. The hydration of HMF to levulinic acid and formic acid
To improve the yields, significant steps must be made in the development of catalysts, preferably heterogeneous in nature. In order to do so detailed knowledge of the reaction mechanism of the main and side reactions is required. A number of reaction mechanisms have been proposed for the dehydration reaction in water, though no definitive evidence has yet been found to confirm these.4 The postulated mechanisms can be divided in mechanisms with cyclic (Scheme 2) or acyclic intermediates (Scheme 3).
Scheme 2. Proposed dehydration mechanism with cyclic intermediates.
Scheme 3. Proposed dehydration mechanism with acyclic intermediates.
284| Chapter 3
Knowledge about the effect of the relative orientation of the hydroxyl groups of different ketoses and aldoses on the dehydration rate and selectivity to HMF is scarcely available in the literature. Seri et al. reported a lower HMF yield from sorbose than from fructose in DMSO, though no conversion data were provided.23 Determination of the influence of the orientation of hydroxyl groups in hexoses could very well provide new insights in the mechanisms of ketose and aldose dehydration. For this reason the acid-catalysed dehydration of fructose, sorbose and tagatose (all ketoses, Scheme 4), as well as glucose, mannose and galactose (all aldoses, Scheme 5) in an aqueous environment was studied using sulphuric acid as the catalyst. A complementary, integrated research strategy was applied. First a high- throughput screening study was performed at a range of temperatures, acid concentrations and reaction times to identify trends in reactivity for various ketoses and aldoses. Subsequently, a detailed kinetic study at relevant process conditions, established in the screening, was performed to determine the kinetic constants and activation energies for the rate of reaction of the various hexose feeds. Finally, DFT calculations were performed, the results of which were compared with the experimental data.
Scheme 4. Ketoses used in this study.
Scheme 5. Aldoses used in this study.
The dehydration of different ketoses and aldoses to 5-hydroxymethylfurfural |285
3.2 Experimental section
3.2.1 High-throughput screening
Glucose (99%), fructose (99%), sulphuric acid (96%), 1,4-dioxane (99.8%) and saccharine (98%) were purchased from Sigma-Aldrich, galactose (99%) and mannose (99%) were purchased from Acros and sorbose (99.6%) and tagatose (99.8%) were purchased from Carbosynth. Milli-Q quality water was used for all experiments and sample preparations. All experiments were performed at a 1 mL scale under 20 bar N2 in batch on an Avantium Quick Catalyst Screening system with a substrate concentration of 65 gL-1 (0.36 M) in water. Sulphuric acid concentrations of 33, 100 and 300 mM were tested. The ketoses were reacted at 100, 120 and 140 °C and the aldoses were reacted at 120, 140 and 160 °C. After the appropriate reaction time the reactor blocks were cooled in an ice bath. The analysis of sugars and furans was performed on a Waters Acquity UPLC with an Acquity ULPC BEH C18 2.1x 5.0 mm, 1.7 µm column (supporting information, Table S1). The sugars were detected using an ELS detector and HMF was detected on a PDA at 230 nm. Saccharine was used as external standard. The analysis of levulinic acid was performed on an Interscience TraceGC with an Agilent J&W FactorFour VF-WAXms, 30 m x 0.25 mm, 0.25 µm column with an FID and 1,4-dioxane as the external standard. The standard mixture was added to the reaction mixture upon opening of the reactor. This was followed by the appropriate dilutions for UPLC and GC analyses.
3.2.2 Kinetic experiments
Fructose (99%), sorbose (98%), tagatose (98%) and sulphuric acid (96%) were purchased from Sigma-Aldrich. Milli-Q quality water was used for all experiments and sample preparations. All experiments were performed at 0.5 mL scale in sealed glass ampoules which were heated in an oven at 137 °C. A substrate concentration of 65 gL-1 (0.36 M) in water was used. Sulphuric acid concentrations of 33, 100 and 300 mM were used. The ampoules were cooled in cold water after the appropriate reaction time. The reaction mixture was then filtered over a 0.45 µm PTFE syringe filter and diluted 7-9 times in water for analysis on an Agilent 1200 HPLC with a Bio-rad Aminex HPX-87H column. 5 mM sulphuric acid was used as the eluent with a flow rate of 0.55 mLmin-1. Refractive index and UV (210 nm) detection were used.
286| Chapter 3
3.2.3 Determination of the kinetic parameters
The kinetic parameters were determined using a maximum likelihood approach, which is based on minimization of errors between the experimental data and the kinetic model. For each hexose, the datasets obtained at the three different acid concentrations were solved simultaneously. Error minimisation was performed using the MATLAB toolbox lsqnonlin.
3.2.4 DFT calculations
Reaction Gibbs free energies (ΔGº298K) of the dehydration of alpha- and beta-anomers of fructo-, sorbo-, psico- and tagatofuranose in water resulting from protonation of the anomeric hydroxyl group were calculated at the B3LYP/6-311+G(d,p) level of theory using the Gaussian 09 program.24 Bulk solvent effects were approximated using the polarised continuum solvent (PCM) model within the conductor reaction field (COSMO) approach.25,26 + 27 Brønsted acid was modelled as a H5O2 cation.
3.3 Results and discussion
3.3.1 High-throughput screening
In order to determine trends in the reactivity of ketoses, (fructose, tagatose and sorbose), a high-throughput screening in a batch mode was performed. Experiments were done at 100, 120 and 140 °C in water containing 33, 100 or 300 mM sulphuric acid at 30, 45 and 60 min reaction time and a fixed initial hexose concentration of 65 gL-1 (0.36 M). The results for the experiments at 120 °C with 100 mM sulphuric acid are given in Figure 1 and Figure 2. In Figure 1 the conversion of the three ketoses is plotted against the reaction time. After 30 min over 60% of the tagatose is converted, compared to around 40% conversion for fructose and sorbose. After 60 min over 90% of the tagatose is converted, whereas around 70% fructose and less than 60% sorbose is converted. This shows a clear difference in reaction rates between the ketoses, with tagatose being by far the most reactive. This trend was observed for the entire dataset (Figures 3-5).
The dehydration of different ketoses and aldoses to 5-hydroxymethylfurfural |287
Figure 1. Conversion of fructose (▲), tagatose (■) and sorbose (●) (65 gL-1) against time at 120 °C with 100 mM aqueous sulphuric acid. Duplicate experiments are shown.
The HMF yield over time at 120 °C with 100 mM sulphuric acid is given in Figure 2. The graph shows that under these conditions the HMF yield is consistently the highest for tagatose. The HMF yield increases in time till about 40% at a conversion of around 85% (Figure 1) and then remains constant, due to the subsequent decomposition of HMF to levulinic acid and formic acid. The HMF yields from fructose and sorbose are equal after 30 min (around 22%), and increase to 35% and 28% after 60 min reaction time, respectively.
Figure 2. HMF yield from fructose (▲), tagatose (■) and sorbose (●) (65 gL-1) against time at 120 °C with 100 mM aqueous sulphuric acid. Duplicate experiments are shown.
The experimental results for the high-throughput screening of the ketoses, exemplified by Figure 1 and Figure 2, indicate that tagatose possesses the highest reactivity for the acid-
288| Chapter 3 catalysed dehydration in the entire process window. The difference in reaction rate between fructose and sorbose is very small and within the error of experimentation. The key difference between the ketoses tested is the orientation of hydroxyl groups on the C3 and C4 position. This is a strong indication that the stereochemistry at C3 and C4 has a significant influence on the reactivity.
Figure 3. Conversion of fructose (▲), tagatose (■) and sorbose (●) (65 gL-1) against time at 100 °C with 33 mM aqueous sulphuric acid. Duplicate experiments are shown.
Figure 4. Conversion of fructose (▲), tagatose (■) and sorbose (●) (65 gL-1) against time at 100 °C with 300 mM aqueous sulphuric acid. Duplicate experiments are shown.
The dehydration of different ketoses and aldoses to 5-hydroxymethylfurfural |289
Figure 5. Conversion of fructose (▲), tagatose (■) and sorbose (●) (65 gL-1) against time at 140 °C with 33 mM aqueous sulphuric acid. Duplicate experiments are shown.
Under reaction conditions where hexose conversion is complete, the HMF yield from tagatose is consistently lower than the HMF yield from fructose and sorbose under the same conditions (Figure 6). This can be explained by the higher reaction rate of tagatose. As a result, the subsequent HMF decomposition to, among others, levulinic and formic acid, also occurs to a larger extent, leading to lower HMF yields.12,13
Figure 6. HMF yield from fructose (▲), tagatose (■) and sorbose (●) (65 gL-1) against time at 140 °C with 100 mM aqueous sulphuric acid. Duplicate experiments are shown. Figure 7 provides an overview of all experimental data obtained in the high-throughput screening. The experimental error in some cases is rather large, due to the fact that this was a screening with the purpose of finding general trends for detailed kinetic experiments. The
290| Chapter 3 maximum attainable yield of HMF is about the same for all three ketoses within the margin of error, and at around 40-45% at 90% conversion. This is in the lower range of the general trends reported in the literature for acid-catalysed fructose dehydration in water.4
Figure 7. HMF yield against conversion for fructose (▲), tagatose (■) and sorbose (●). Experiments were performed at 100, 120 and 140 °C with 33, 100 or 300 mM H2SO4 at 30, 45 and 60 min reaction time. Duplicate experiments are shown.
Figure 8 shows a strong increase in levulinic acid yield at high conversions. Together with the decrease in HMF yield at high conversions as (Figure 7) this confirms that at high conversion, the HMF yield is reduced by a consecutive reaction to levulinic acid and formic acid, as described in the literature.4,12,13
Apart from the ketoses discussed above, three aldoses were also screened in batch mode at similar conditions as for the ketoses. Glucose, mannose and galactose were reacted at 120, 140 and 160 °C with 33, 100 and 300 mM sulphuric acid for 30-60 min. A higher temperature range was selected because glucose is known to be less reactive than fructose.4 The observed maximum HMF yield was around 5%, at conversions of 30-70%, which is in line with the majority of the literature.4 Furthermore no clear differences in aldose conversion rates or selectivities to HMF were observed for the three aldoses. This indicates that the orientation of the hydroxyl groups does not affect the dehydration rate of the aldoses, which suggests that the acid-catalysed aldose dehydration possesses a different rate-determining step than the acid-catalysed ketose dehydration.
The dehydration of different ketoses and aldoses to 5-hydroxymethylfurfural |291
Figure 8. Levulinic acid yield against conversion for fructose (▲), tagatose (■) and sorbose (●).
Experiments were performed at 100, 120 and 140 °C with 33, 100 or 300 mM H2SO4 at 30, 45 and 60 min reaction time. Duplicate experiments are shown.
3.3.2 Kinetic study
The screening results indicated clear differences in the reactivity of the three ketoses. In the next step, kinetic experiments were performed to quantify the differences in reactivity of the ketoses at 137 °C with 33, 100 and 300 mM sulphuric acid in a time frame of 0-90 minutes. The experiments were performed in a batch mode using glass ampoules. At t = 0 min, the ampoules were placed in an oven at 137 °C and allowed to react for a pre- determined time. The experimental data were modelled using a simple kinetic expression assuming first order in sugar concentration and acid concentration (equations 1 and 2). Here, kref describes the reaction rate constant at 137 °C (Tref) and Ti is the initial temperature (25 °C). At the initial stage of the reaction, the temperature is not constant as it takes typically 5- 10 minutes to reach 137°C. This effect was compensated for by extending the model with an energy balance. After integration, equation 3 is obtained, which, combined with the mass balances in batch for the hexoses, allows calculation of the concentration and the temperature as a function of the batch time. The value for the parameter h (0.2135 min-1, equation 3) was determined independently.12
In Figure 9 the experimental data points (137 °C, 33 mM sulphuric acid) and the model line for the three ketoses are provided. Agreement between model and experimental data is 2 very satisfactory (c.f. R values and the error in k1,ref as seen in Table 1). The results from the kinetic study are in line with the screening experiments, with tagatose being the most reactive
292| Chapter 3
ketose. This is clearly illustrated by the significantly higher k1,ref value and significantly lower Ea value for acid-catalysed dehydration of tagatose than for the acid-catalysed dehydration of fructose and sorbose. The k1,ref and Ea values for sorbose and fructose dehydration are essentially equal within the experimental error.
Figure 9. The concentration of sugar in time at 137 °C with 33 mM sulphuric acid, with the points describing the experimental results (fructose ▲, tagatose ■ and sorbose ●) and the lines describing the model.
Table 1. Kinetic data for the acid-catalysed dehydration of different ketoses - -1 2 Ketose kref (Lmol Ea (kJ mol ) R 1min-1) Sorbose 0.9 ± 0.2 138 ± 41 0.976 Fructose 0.9 ± 0.1 124 ± 22 0.989 Tagatose 2.0 ± 0.6 89 ± 15 0.990
r - u ar 4 (1) 1 1 - a - T T refe real ref (2) -ht Treal Tref- Tref-Ti e (3)
Figure 10 shows the yield against the conversion for all kinetic experiments. For fructose and tagatose the general trend is the same as for the experiments in the high-throughput screening, with approximately 55% selectivity until around 60% conversion and a maximum HMF yield of just over 40% at around 85% conversion. The trend for sorbose, however, is different. At conversions higher than 30%, the HMF yield line for sorbose is below that of the The dehydration of different ketoses and aldoses to 5-hydroxymethylfurfural |293 other two ketoses. This indicates that additional reactions are involved in sorbose dehydration. This effect was not observed in the screening, presumably due to analytical issues (e.g. overlapping product peaks). Further research is in progress to identify additional reaction products.
Figure 10. The HMF yield plotted against conversion for the kinetic studies using the three ketoses at 137 °C with 33, 100 and 300 mM acid at 0-90 minutes (fructose ▲, tagatose ■ and sorbose ●).
3.3.3 DFT calculations
To gain insights in the origin of the experimentally observed differences in reactivity between the three ketoses, model DFT calculations were performed. It has previously been proposed that the reactivity and selectivity of the acid-catalysed hexose conversion is determined by the regeoselectivity of the initial protonation and the accompanying dehydration step.27 A characteristic aspect of fructofuranose dehydration by Brønsted acids is the preferred protonation of the OH group at the anomeric C2 carbon (Figure 11a)) that initiates a sequence of fast reactions towards HMF. In the case of glucose conversion, the initial protonation is much more difficult.27 Here the preferred protonation site is the anomeric hydroxyl at C1. However, its activation does not lead to the desired products but rather opens pathways towards humin precursors. In both cases, the reactivity is determined by the stability of cationic intermediates resulting from the dehydration of the initial protonated complex.
294| Chapter 3
The optimised structures of the cationic sugar dehydration products and the respective
ΔGº298K values are summarised in Figure 11 (b). The most favourable reaction is predicted for tagatofuranose. The free energies of fructofuranose and sorbofuranose dehydration are very similar and are substantially higher than that computed for psicofuranose, which was not included in the experimental study. Interestingly, it appears that the computed reaction free energies correlate with the direction of the dipole moment in the dehydration products. The most stable cationic species is formed when the dipole moment is oriented parallel to the furanose ring (tagatofuranose). Deviations from such an orientation lead to the destabilization of the dehydrated cation (Figure 11 (b)).
Figure 11. (a) Furanose dehydration initiated by the preferred protonation of the anomeric hydroxyl group and (b) the optimised structures of the cationic dehydration products originating from psico-, sorbo-, fructo- and tagatofuranose and the respective DFT-computed average Gibbs free energies of dehydration of alpha- and beta-anomers of the sugars (kJ mol-1). The arrows depict the direction of the dipole moment in the cations.
The dehydration of different ketoses and aldoses to 5-hydroxymethylfurfural |295
3.3.4 Mechanistic aspect
The experimental data from both the screening and kinetic experiments clearly show differences in reactivity for the different ketoses. The kinetic models obtained for the three ketoses at different acid concentrations are consistent and fit well with the experimental results. A possible explanation for the differences in reactivity of the ketoses is the relative orientation of their hydroxyl groups at the C3 and C4 position.
DFT calculations on acid–catalysed furanose dehydration indicate that the carbocation formed by dehydration at C2 is most stable for tagatose, followed by fructose and sorbose. This is also the order of decreasing reactivity as found in the kinetic study, suggesting a relationship between the determined activation energies and the calculated ΔG from the DFT calculations (Figure 12).
Figure 12. The observed Ea plotted a ainst calculated ΔGº298K of the single dehydrated ketose carbocations.
The studies provided above imply that the differences in reaction rate of the different ketoses are due to the orientation of the hydroxyl groups on C3 and C4. This suggests that the reaction follows a mechanism with cyclic intermediates (Scheme2). For a dehydration mechanism with acyclic intermediates (Scheme 3), it is more difficult to explain the significant differences in reactivity of the different ketoses. The mechanisms are based on a series of ß-dehydrations. According to literature, intermediates are generally not observed in water-based acid-catalysed fructose dehydration, which implies that the formation of the 1,2- enediol would be the rate-determining step in the case of a mechanism with acyclic intermediates as shown in Scheme 6. This shows that the orientation of the hydroxyl group on
296| Chapter 3 the chiral C3 of the 1,2-enediol relative to the double bond between C1 and C2 is not relevant, as C2 en C1 are both achiral. The same can be said for C4 in 3-deoxyhexosulose. This means that any differences in reactivity between the ketoses can only be explained on basis of the interactions between groups on C3 with those on C4 and C5, which are connected with free-rotating carbon-carbon bonds, which in turn makes large differences in reaction rates less likely.
Since a cyclic reaction mechanism requires the ketose to react through its furanose form, it is relevant to take the tautomeric distribution into account. The different tautomeric forms of ketohexoses are shown in Scheme 7. The tautomeric distribution of the ketohexoses in water was studied by Que and Gray (Table 2),28 showing that of the ketoses addressed in this research, fructose has the largest fraction present in the furanose forms (28%) at equilibrium at 30 °C, followed by tagatose (14%) and sorbose (5%). There is no relationship between these values and the differences in reaction rate observed in our study. A study on galactopyranose tautomerisation in water showed that at slightly acidic conditions (pH 4.3) equilibrium can be reached after about an hour at room temperature.29 At increasing temperatures and acid concentrations this rate will increase. Therefore it can be safely assumed that the tautomerisation rate is much higher than the dehydration rate under the conditions studied in this research. This, in combination with the data in Table 2 excludes the tautomeric distribution as an important factor in the dehydration rate.
Scheme 6. The formation of HMF from hexose through acyclic intermediates
The difference in HMF yield between sorbose and the other two ketoses is difficult to explain. The DFT calculations did indicate that the proposed carbocation intermediate for sorbose is less stable than for fructose and especially tagatose. This could be favourable for a parallel reaction pathway that forms other products. This phenomenon is currently under investigation. The dehydration of different ketoses and aldoses to 5-hydroxymethylfurfural |297
Another important observation in this study is that the orientation of the hydroxyl groups on the aldoses has no influence on their dehydration rate, indicating that this orientation has no effect on the rate-determining step. It suggests that aldose dehydration to HMF involves a different rate-determining step than ketose dehydration to HMF, which likely demands a different type of catalyst. Since a number of researchers have found evidence that suggests that an isomerisation co-catalyst is required in the glucose dehydration to HMF,4,14-18 this could mean that the rate-determining step in aldose dehydration under acidic conditions is the isomerisation to a ketose. The so-called Lobry de Bruyn-Alberda Van Eekensteijn transformation for aldose-ketose isomerisation is base-catalysed.30 It is thus not surprising that the HMF yield in dehydrations from aldoses under acidic conditions is so low. This information furthermore suggests a cyclic dehydration mechanism starting from the ketofuranose form, since 1,2-enediol formation appears to be favoured in the presence of basic catalysts.
Scheme 7. The possible tautomers of ketohexoses.
Table 2. Proportions of pyranose and furanose forms of ketoses at equilibrium in aqueous solutions at 30 °C28
Ketose α- ß- α- ß- Pyranose Pyranose Furanose Furanose
Fructose 0 72 5 23
Tagatose 71 15 5 9 Sorbose 95 0 5 0
Psicose 26 21 38 15
298| Chapter 3
3.4 Conclusions
In this research the combination of high throughput screening and detailed kinetic studies was successfully applied to determine the differences in reactivity between hexoses in the acid-catalysed dehydration to HMF. The screening showed that different ketoses have different reactivity in acid-catalysed dehydration to HMF, which must be caused by the different orientation of the hydroxyl groups on C3 and C4. Tagatose showed significantly higher conversion rates than fructose and sorbose. Differences between fructose and sorbose were much smaller. For these ketoses (sorbose, fructose and tagatose) k and Ea values were successfully determined from kinetic experiments.
Between the aldoses tested (glucose, mannose and galactose) no differences in reactivity were observed. This indicates that HMF formation from ketoses and aldoses possesses different rate-determining steps.
Initial screening experiments using UPLC analyses indicated a maximum HMF yield around 40% at around 85% conversion. The kinetic experiments in combination with HPLC analyses, however, showed that the maximum HMF yield from sorbose was only 30%, which leads to the conclusion that a different parallel pathway is available for sorbose. This aspect is still under investigation.
DFT calculations on the ketoses indicate that tagatose forms the most stable furanose carbocation upon single dehydration at the anomeric carbon (C2), followed by fructose and sorbose. As this is also observed to be the order of decreasing reactivity, the calculations provide a plausible explanation for the obtained results by a dehydration mechanism with cyclic intermediates.
The results of the experimental data in combination with DFT calculations are best explained by the formation of HMF via a ketose dehydration mechanism with cyclic intermediates. This means that for aldose conversion to HMF, an isomerisation to a ketose would be required prior to dehydration. The fact that this is typically base-catalysed explains why the acid-catalysed dehydration of aldoses is much less efficient than the acid-catalysed dehydration of ketoses.
The dehydration of different ketoses and aldoses to 5-hydroxymethylfurfural |299
3.5 References
(1) Dias, A. S.; Lima, S.; Pillinger, M.; Valente, A. A. In Ideas in Chemistry and Molecular Sciences: Advances in Synthetic Chemistry; Pignataro, B., Ed.; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, Germany, 2010. (2) Lange, J.-P.; van der Heide, E.; van Buijtenen, J.; Price, R. ChemSusChem 2012, 5, 150-166. (3) Van Putten, R.-J.; Dias, A. S.; De Jong, E. In Catalytic Process Development for Renewable Materials; Imhof, P., van der Waal, J. C., Eds.; Wiley-VCH Verlag GmbH & Co., 2013. (4) Van Putten, R.-J.; Van der Waal, J. C.; De Jong, E.; Rasrendra, C. B.; Heeres, H. J.; De Vries, J. G. Chem. Rev. 2013, 113, 1499-1597. (5) Bozell, J. J.; Petersen, G. R. Green Chem. 2010, 12, 539-554. (6) Gruter, G. J. M.; Dautzenberg, F. (Avantium International BV) WO 2007104514, 2007. (7) Munoz de Diego, C.; Schammel, W. P.; Dam, M. A.; Gruter, G. J. M. (Furanix Technologies BV) Int. Patent WO2011043660, 2011. (8) Sipos, L. (Furanix Technologies BV) Int. Patent WO2010077133, 2010. (9) De Jong, E.; Dam, M. A.; Sipos, L.; Gruter, G. J. M. In ACS Symp. Ser.; Smith, P. B., Gross, R., Eds., 2012. (10) Bauen, A.; Berndes, G.; Junginger, M.; Londo, M.; Vuille, F.; Ball, R.; Bole, T.; Chudziak, C.; Faaij, A.; Mozaffarian, H. “Bioenergy: a sustainable and reliable energy source. A review of status and prospects,” 2009. (11) Serrano-Ruiz, J. C.; Dumesic, J. A. Energy Environ. Sci. 2011, 4, 83-99. (12) Girisuta, B.; Janssen, L. P. B. M.; Heeres, H. J. Green Chem. 2006, 8, 701-709. (13) Kuster, B. F. M. Starch/Stärke 1990, 42, 314-321. (14) Ohara, M.; Takagaki, A.; Nishimura, S.; Ebitani, K. Appl. Catal. A: Gen. 2010, 383, 149-155. (15) Nikolla, E.; RomaÌ•n-Leshkov, Y.; Moliner, M.; Davis, M. E. ACS Catal. 2011, 1, 408-410. (16) Grande, P. M.; Bergs, C.; Dominguez de Maria, P. ChemSusChem 2012, 5, 1203-1206. (17) Chareonlimkun, A.; Champreda, V.; Shotipruk, A.; Laosiripojana, N. Fuel 2010, 89, 2873-2880. (18) Zhao, H.; Holladay, J. E.; Brown, H.; Zhang, Z. C. Science 2007, 316, 1597. (19) Pidko, E. A.; Degirmenci, V.; van Santen, R. A.; Hensen, E. J. M. Angew. Chem. Int. Ed. 2010, 49, 2530-2534. (20) Binder, J. B.; Raines, R. T. J. Am. Chem. Soc. 2009, 131, 1979-1985. (21) Ståhlberg, T.; Rodriguez-Rodriguez, S.; Fristrup, P.; Riisager, A. Chem. Eur. J. 2011, 17, 1456- 1464. (22) Hu, S.; Zhang, Z.; Song, J.; Zhou, Y.; Han, B. Green Chem. 2009, 11, 1746-1749. (23) Seri, K.; Inoue, Y.; Ishida, H. Chem. Lett. 2000, 22-23. (24) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; J. A. Montgomery, J.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J.; Revision A.01 ed.; Gaussian, Inc.: Wallingford CT, 2009. (25) Barone, V.; Cossi, M. J. Phys. Chem. A 1998, 102, 1995. (26) Eckert, F.; Klamt, A. AIChE J. 2002, 48, 369. (27) Yang, G.; Pidko, E. A.; Hensen, E. J. M. J. Catal. 2012, 295, 122. (28) Que, L., Jr.; Gray, G. R. Biochemistry 1974, 13, 146-153.
300| Chapter 3
(29) Wertz, P. W.; Garver, J. C.; Anderson, L. J. Am. Chem. Soc. 1981, 103, 3916-3922. (30) Speck, J. C., Jr. Adv. Carbohydr. Chem. 1958, 13, 63-103.
|301
4 A comparative study on the reactivity of various ketohexoses to furanics in methanol
Abstract
In this chapter, the acid-catalysed dehydration of the four 2-ketohexoses (fructose, sorbose, tagatose and psicose) to furanics was studied in methanol (65 gL-1 substrate concentration, 17 and 34 mM sulphuric acid, 100 °C) using Avantium high-throughput technology. Significant differences in the reactivity of the hexoses and yields of 5-hydroxymethylfurfural (HMF) and its methyl ether MMF were observed. Psicose and tagatose were shown to be the most reactive, with psicose also giving the highest combined yield of MMF and HMF of around 55% at 96% sugar conversion. Hydroxyacetylfuran and its corresponding methylether were formed as a by-product for particularly sorbose and tagatose, with a maximum combined yield of 8% for sorbose. The formation of hydroxyacetylfuran was studied using 13C-NMR with labelled sorbose, providing new insights into the mechanism of its formation.
302| Chapter 4
4.1 Introduction
Diminishing fossil resources and the debate on climate change require the development of sustainable processes for the production of materials and energy. Production of materials and liquid fuels present an especially interesting case, as it is clear that non-fossil sources of fixed carbon are required. Biomass is the obvious alternative to fossil resources for these applications, since it is the largest available source of fixed carbon on earth.1 Carbohydrates represent 60-80% of all biomass.2 They occur predominantly in polymeric form (cellulose, hemi-cellulose, starch) and are built up from hexoses (a.o. glucose, fructose, mannose, galactose) and pentoses (a.o. arabinose, xylose). The development of efficient catalytic processes for producing fuels and materials is the most important challenge in the conversion of biomass, especially in the case of carbohydrates. This challenge is mainly caused by the over-functionalised nature of these compounds, which have a high oxygen content. This results in both chemical and physical issues, such as reduced selectivity of potential processes and solvent incompatibilities. For most applications compounds with high carbon and low oxygen content are required, so oxygen has to be removed. The main chemical strategies to do so are dehydration,3 decarboxylation/decarbonylation4 and hydrogenation/hydrogenolysis.4 Decarboxylation and decarbonylations result in significant carbon loss in the form of COx. Hydrogenation/hydrogenolysis requires additional hydrogen, which is expensive and is typically produced from non-renewable sources.5 As such, dehydration is an interesting approach, as it conserves all the carbon atoms in the molecule and the remaining product is characterised by having unsaturated, reactive groups (either olefinic or carbonyl) for further reactions.
The dehydration of hexoses is well known and initially leads to the formation of 5- hydroxymethylfurfural (HMF).3 This molecule is considered as a high potential platform chemical and it is mentioned in the DoE top-12 biobased chemicals.6 HMF can be converted into a number of interesting building blocks, such as 2,5-furandicarboxylic acid (FDCA, DoE top-12),3,6 and levulinic acid (LA, DoE top-12).6-9 Another interesting derivative is dimethylfuran, with applications as fuel additive, solvent, and as a precursor for green aromatics (p-xylene).10
FDCA has been identified as a polymer precursor for polyethylenefuranoate (PEF), which is an example of a high-potential biobased polyethyleneterephtalate (PET) replacement, in which the terephtalic acid moiety is replaced by FDCA. Compared to PET, A comparative study on the reactivity of various ketohexoses to furanics in methanol |303
PEF production reduces non-renewable energy use and greenhouse emissions significantly.11 Additionally PEF has superior physical properties compared to PET.12,13
Besides HMF, another well known platform chemical from hexoses is levulinic acid (LA). LA may be converted to both bulk- and fine chemicals.14 Examples are the multistep conversion to δ-aminolevulinic acid, an active ingredient in a biodegradable herbicide;14 the oxidation to succinic acid, a DoE top-12 compound applicable as a monomer in polyesters;6,14 and reduction to γ-valerolactone, a precursor for methyltetrahydrofuran, which can be used as a solvent and a fuel additive.14
Over the last decade, global research activities on the development of techno- economically viable HMF processes have increased dramatically. The main focus has been on increasing yields via the development of selective catalysts, and the use of solvents other than water.3 The state of the art has shown that the maximum obtainable yield of HMF is limited by the instability of the molecule under reaction conditions.3 Under acidic conditions HMF reacts with water to form levulinic and formic acid (Scheme 1).7-9 In addition, insoluble black-brown by-products are formed, referred to as humins, that reduce the yield and cause processing issues such as reactor fouling.15 The conversion of fructose to HMF under aqueous conditions has a maximum yield of around 50%.3 In ionic liquids and polar aprotic solvents like DMSO yields over 90% have been reported.3 For glucose, an aldose, the reported yields are generally much lower (<10%), unless specific co-catalysts and advanced solvent systems are used.3 Avantium is currently developing a process in which hexoses are dehydrated in alcohols. This leads to the formation of HMF ethers as the predominant products (Scheme 1). Dehydration in alcohols has a number of advantages compared to HMF production in water, aprotic polar solvents and ionic liquids. The use of organic solvent increases the attainable yield of furanics by suppressing the formation of levulinic acid. Additionally, the HMF will be mainly converted to its alkyl ether. In the case of lower alcohols these ethers have a lower boiling point than HMF, which simplifies work-up by distillation.16,17 It is generally very challenging to separate HMF from aprotic polar solvents or ionic liquids, since these solvents generally have a very high affinity for HMF.3 Lower alcohols, on the contrary, can be readily removed by evaporation.
304| Chapter 4
Scheme 1: The dehydration of hexose, in the absence and presence of alcohol, to form RMF, followed by hydration to levulinates
Despite the large volume of research on HMF formation from hexoses, the reaction mechanism is still under debate.3 A better understanding is of pivotal importance for the development of efficient catalytic systems for the production of HMF and derivatives. We have recently shown that aldose dehydrations follow a different reaction pathway than ketose dehydrations.18 This led to the conclusion that different catalyst systems are required for glucose- and fructose dehydration to HMF. Furthermore we reported that the relative position of the hydroxyl groups in the different ketoses (Scheme 2) resulted in different reactivity for the dehydration to HMF in water.18 Tagatose was shown to react significantly faster than sorbose and fructose. These results essentially disprove a reaction mechanism with acyclic intermediates, making a reaction mechanism with cyclic intermediates more likely.
Apart from differences in reaction rates for the ketoses, it was also shown that the maximum obtainable HMF yield in water was significantly lower for sorbose than for fructose and tagatose.18 It is not yet clear whether this is related to the lower rate of HMF formation relative to the rate of HMF decomposition or whether it also indicates the existence of a competing reaction pathway for sorbose that is not as accessible for the other ketoses. A comparative study on the reactivity of various ketohexoses to furanics in methanol |305
The reactivity of four ketohexoses (fructose, sorbose, tagatose and psicose) in methanol using sulphuric acid as the catalyst was studied in detail in order to gain insights in sugar- furan yield relationships. For this purpose, an experimental study was performed using the Avantium high-throughput platform. In addition, a 13C-NMR experiment with labelled sorbose was performed to gain insights into the pathways leading to the formation of HAF, a by-product for sorbose.
Scheme 2: The α-furanose structures of the D-ketoses
4.2 Experimental section
4.2.1 Chemicals
D-Fructose, L-sorbose, D-tagatose (>99%), sulphuric acid (96%), 1,4-dioxane (99.8%),
CDCl3 (99.8%) and saccharin (98%) were purchased from Sigma-Aldrich, D-psicose and L- [6-13C]sorbose (99.9%) were purchased from Carbosynth, 2-hydroxyacetyfuran (HAF, 95%) was purchased from Otava Ltd., acetonitrile (HPLC-S grade) from Biosolve and methanol (HPLC grade) from Fisher. Milli-Q quality water was used for all sample preparations.
4.2.2 High-throughput experimentation
All experiments were performed with a substrate concentration of 65 gL-1 (0.36 M) and with two sulphuric acid concentrations (17 and 34 mM) at 1 mL scale in methanol under 20 bar N2 at 100 °C. An Avantium Quick Catalyst Screening system was used. After the appropriate reaction time, the reactor blocks were cooled in an ice bath. After opening the reactors, 4 mL of an external standard solution containing 5.63 gL-1 saccharin (UPLC) and 3.75 gL-1 1,4- dioxane (GC) in 1:1 acetonitrile/water was added to each reactor. Upon complete dissolution, the resulting mixtures were transferred to vials for storage at -20 °C. These mixtures were measured on GC and diluted in water (25 times) for UPLC analysis. For GC analysis the mixtures were neutralised with 10-30 µl aqueous NaOH solution (1 M), depending on the acid content of the sample. The level of neutralisation was checked with pH paper.
306| Chapter 4
4.2.3 Synthesis of 2-methoxyacetylfuran
2-Methoxyacetylfuran (MAF) was synthesised by reacting 2-hydroxyacetylfuran (HAF) using the reaction conditions as described for the high-throughput experimentation, except that H2SO4 concentrations of 17, 34 and 67 mM were applied at a fixed reaction time of 60 min. The reaction mixtures were analysed using GC-MS to confirm MAF formation. These mixtures were consequently used to confirm the retention time of MAF on both the UPLC and the GC.
4.2.4 Chromatographic analysis
The analysis of sugars and furanics was performed on a Waters Acquity UPLC (Table 1) equipped with an Acquity ULPC BEH C18 2.1x 5.0 mm, 1.7 µm column. A mobile phase (0.6 mL/min) with acetonitrile/methanol (1:1, A) and water (0.2 wt% TFA, B) was used, with a gradient from 98% A and 2% B up to 2% A and 98% B. The sugars were detected using an ELS detector (ELSD) and HMF was detected on a PDA at 230 nm. Saccharine was used as external standard, measured at 250 nm for UV and in the ELSD for sugar analysis. As the reactions were performed in alcohols, remaining sugars were mostly present in methylated form. These methylated sugars were categorised as unconverted sugars, since they are known to convert to furanics.17
Levulinic acid and methyl levulinate were analysed on an Interscience TraceGC with an Agilent J&W FactorFour VF-WAXms, 30 m x 0.25 mm, 0.25 µm column with an FID and 1,4-dioxane as the external standard. The GC as operated at a temperature of 250 °C at a constant flow of 2 mL/min. GC-MS (Thermo Scientific traceGC, DSQ II mass detector) was used to confirm the formation of 2-hydroxyacetylfuran (HAF) and its methyl ether (MAF). Ion source and transfer line temperatures of 250 °C were used.
Calibration of the equipment was done using pure compounds. The UPLC response factor for MAF was estimated using the known response factors for HAF, HMF and MMF. The response factors of the latter two were shown to be related to the molecular weights, with the response factor ratio of the two being equal to their molecular weight ratio. The response factor of MAF was therefore calculated from that of HAF using a correction for the difference in molecular weight. A comparative study on the reactivity of various ketohexoses to furanics in methanol |307
4.2.5 Experiments with L-[6-13C]sorbose
Three experiments with L-[6-13C]sorbose were performed at 1 mL scale with 65 gL-1 substrate and 17 mM H2SO4 in methanol for 150 min at 100 °C and 20 bar N2. After reaction the samples were combined and neutralised with NaOH(aq) (8.1 mg, 50 wt%, 0.10 mmol) in 1.5 mL methanol. All volatiles were then removed under reduced pressure (5 mbar, 35 °C), leaving a brown coloured residue. 1.0 mL CDCl3 was added to the residue and the resulting suspension was filtered over a syringe filter (Millex FH 0.45 µm) before measurement.
1 13 H and C (Inverse Gated Decoupled) NMR spectra were recorded in CDCl3 on a Bruker Avance 500 (125.78 MHz for 13C, 2048 scans, 2.0 s delay time) digital NMR with a
Cryo Platform. The residual hydrogens of CDCl3 (δ 7.22 ppm) were used to reference the spectra.
4.3 Results and discussion
The acid-catalysed dehydration of the ketoses fructose, tagatose, sorbose and psicose to furanics was studied in methanol using a high-throughput batch setup with the objective to compare the reactivity of the four ketohexoses systematically under similar reaction conditions. Duplicates and triplicates of selected experiments were performed for statistical analysis regarding product yields and hexose conversions. The reactions were followed in time at a fixed temperature (100 °C) with two H2SO4 concentrations (17 and 34 mM) at a constant substrate concentration (65 gL-1). Detailed product analysis was performed using UPLC and GC, and the main soluble reaction products were HMF and its corresponding methyl ether (MMF), levulinic acid (LA) and its methyl ester (ML, see Scheme 10). In general insoluble humins were not formed, though the reaction mixtures did show brown colouring, which increased with the severity of the reaction conditions. In addition, for some of the hexoses, significant amounts of hydroxacetylfuran (HAF) and its corresponding ether (MAF) were formed. Furfural, which has been observed for the Brønsted acid catalysed conversions of hexoses to HMF and LA,3 was observed in very low quantities and yields on hexoses were always below 0.5%. In the following, the concentration versus time profiles and yield-selectivity data will be provided and discussed in detail for the four ketohexoses used in this study.
308| Chapter 4
4.3.1 High-throughput experimentation
Figure 1 shows the conversion versus time profiles for experiments with all four ketohexoses. The spread in the duplicates and triplicates does not exceed 3%, showing that the experiments are highly reproducible. When comparing the conversion-time trends for the different hexoses, it is clear that tagatose and psicose show very similar reactivity. The trends for fructose and sorbose show a similar pattern, though both are less reactive than tagatose and psicose. A similar trend was observed previously in water, though a full comparison is challenging as psicose was not tested in this study.18
Figure 1: The conversion vs. reaction time for fructose (▲), sorbose (●), tagatose (■) and -1 psicose (★) at 65 gL substrate in methanol at 100 °C with (a) 17 mM H2SO4 and (b) 34 mM H2SO4
The yields of HMF in time are shown in Figure 2. Again the reproducibility was shown to be very high. The HMF yield shows a clear optimum in all cases; for both acid concentrations, the HMF yield is the highest for psicose, followed by tagatose, fructose and sorbose, respectively. Contrary to sugar dehydration in water,18 the HMF yield decreases in time at low conversions. This is mainly caused by the acid-catalysed methylation of HMF to methoxymethylfurfural (MMF)19 and at this early stage of the reaction not to the subsequent conversion to LA and ML. The MMF yield (Figure 3) is also consistently the highest for A comparative study on the reactivity of various ketohexoses to furanics in methanol |309 psicose, followed by tagatose and fructose with approximately equal values, and sorbose. The highest MMF yield was around 55%, obtained from psicose. For fructose and tagatose the maximum MMF yield was around 43% and for sorbose this was only around 25%.
Figure 2: The yield of HMF vs. reaction time for fructose (▲), sorbose (●), tagatose (■) and -1 psicose (★) at 65 gL substrate in methanol at 100 °C with (a) 17 mM H2SO4 and (b) 34 mM H2SO4
310| Chapter 4
Figure 3: The yield of MMF vs. reaction time for fructose (▲), sorbose (●), tagatose (■) and -1 psicose (★) at 65 gL substrate in methanol at 100 °C with (a) 17 mM H2SO4 and (b) 34 mM H2SO4
The combined furan yield (HMF and MMF, abbreviated as RMF) is of particular importance for this study and the results are given in Figure 4. The RMF yield versus time data show an optimum when working at the highest acid concentration and reach up to 55% for psicose at around 80 min reaction time. The subsequent decrease in yield is due to the successive reaction to LA and ML (vide infra).
When the sugar conversion (Figure 1) and RMF yield versus time profiles (Figure 4) are compared, it is clear that psicose and tagatose have comparable conversion rates, but psicose forms more RMF. This implies that psicose is converted to RMF with a higher selectivity than tagatose. A comparable trend is observed for fructose and sorbose, where fructose converts more selectively to RMF.
Additional information is obtained by plotting the yields of RMF for all hexoses versus the conversion for both acid concentrations (Figure 5). A clear trend is visible for each hexose, irrespective of the acid concentration. This is an important observation, as it reveals that the acid concentration does not have a measurable influence on the selectivity to RMF. When comparing the results for the four hexoses, it is clear that the selectivity to RMF is A comparative study on the reactivity of various ketohexoses to furanics in methanol |311 consistently higher for fructose and psicose than for sorbose and tagatose. Due to their higher reactivity, the amount of data points below 70% conversion for tagatose and psicose is limited. As such, it is difficult to distinguish between psicose and fructose or between sorbose and tagatose at lower conversions. The data available for psicose below 80% conversion appear to follow the same trend as the fructose data, which suggests comparable selectivity for fructose and psicose in this range. The tagatose data at conversions below 65% appear to fit with the trend of the sorbose data, which points to comparable selectivity at low conversions for tagatose and sorbose. At higher conversions, though, a clear separation is visible in both cases. At conversions over 80%, the RMF yield from psicose is consistently higher than for fructose. This results in a maximum RMF yield around 55% from psicose, whereas the maximum yield from fructose is around 45%. These results clearly show that from a yield and selectivity point of view, psicose is the favoured substrate for the formation of HMF and MMF. Sorbose is clearly the worst substrate, with a maximum yield that is only about half of the yield obtained from psicose. The differences between fructose and tagatose are smaller, though significant, as fructose converts slower but more selectively to HMF and MMF than tagatose.
The observed differences in selectivity between the four ketoses could be caused by the differences in the rate of RMF formation and competing side-reactions. A known side- reaction for the dehydration of sugars is the formation of 2-hydroxyacetylfuran (HAF, Scheme 1).20-23 As such, attempts were made to identify and quantify HAF in the product samples. Improved analytics showed an additional peak in the UPLC chromatogram next to HMF with a similar UV spectrum. Spiking experiments with pure HAF on the UPLC and GC-MS confirmed the presence of HAF. Besides HAF, its methyl ether (2- methoxacetylfuran, MAF), was also observed on the UPLC. Etherification experiments with pure HAF, in combination with GC-MS analysis, confirmed MAF was formed under the reaction conditions. Subsequent spiking experiments on the UPLC provided the retention time of MAF. The formation of MAF had not been previously reported in sugar dehydration reactions, though it is a known compound.24
312| Chapter 4
Figure 4: The yield of RMF vs. reaction time for fructose (▲), sorbose (●), tagatose (■) and -1 psicose (★) at 65 gL substrate in methanol at 100 °C with (a) 17 mM H2SO4 and (b) 34 mM H2SO4.
Both HAF and MAF were only formed in significant amounts from sorbose and tagatose (Figure 6). The yields of both are consistently higher for sorbose than for tagatose. HAF is by far the major component, thus the methylation of HAF to MAF appears relatively slow on the timescale of the experiments. The sum of the HAF and MAF yields is at maximum around
8% for sorbose at 150 min with 17 mM H2SO4 (Figure 6a) or at 75 min with 34 mM H2SO4 (Figure 6b).
A comparative study on the reactivity of various ketohexoses to furanics in methanol |313
Figure 5: The yield of RMF vs. sugar conversion for fructose (▲), sorbose (●), tagatose (■) and psicose (★) at 65 gL-1 substrate in methanol at 100 °C for all reaction times with 17 or 34 mM H2SO4.
In Figure 7 the combined yield of all identified furanics (HMF, MMF, HAF and MAF) vs. the conversion is shown for the four ketohexoses. The total furanics yield is consistently higher for psicose and fructose than for tagatose and sorbose. Especially at conversions over 85% large differences in furanics yield are observed, with psicose yielding around 55% furanics and sorbose yielding only around 32% furanics at best. This means that the formation of HAF and MAF in the case of tagatose and especially sorbose does not compensate entirely for the lower RMF yields observed in Figure 5.
314| Chapter 4
Figure 6: The yield of HAF (grey) and MAF (black) vs. reaction time for sorbose (●) and -1 tagatose (■) at 65 gL substrate in methanol at 100 °C with (a) 17 mM H2SO4 and (b) 34 mM H2SO4.
Figure 7: The combined yield of HMF, MMF, HAF and MAF vs. sugar conversion for fructose (▲), sorbose (●), tagatose (■) and psicose (★) at 65 gL-1 substrate in methanol at
100 °C for all reaction times with 17 or 34 mM H2SO4. A comparative study on the reactivity of various ketohexoses to furanics in methanol |315
In sugar dehydration to furanics, the yield is known to be limited by two important side- reactions: the hydration of HMF to levulinic acid (LA) and the formation of humins. Brown solid materials were not present after reaction, so any humins formed should be soluble, which hampers quantification. As the reactions were performed in methanol under acidic conditions, LA was mainly present as its methyl ester (ML). In Figure 8 the yield of ML versus time is shown. The ML yield is a clear function of the time and, as expected, is formed particularly at longer reaction times (Figure 8). The yields of LA were typically below 1%, showing that almost all LA is esterified to ML under the prevailing reaction conditions. In Figure 9 the yield of ML is plotted against the RMF yield, which shows clearly different trends for the different ketoses. At the same RMF yield, psicose shows clearly the lowest yield of methyl levulinate (ML), followed by fructose and tagatose with comparable values, and sorbose with by far the highest yield in ML relative to the RMF yield. This relates the maximum obtainable RMF yield to the formation of levulinates. From Figure 4 it was observed that the RMF formation rate is the highest for psicose, followed by fructose and tagatose with approximately equal values, and sorbose with clearly the lowest formation rate of RMF. When this information is combined with the observations from Figure 9, it shows that a high rate of RMF formation is favoured for high RMF yields. This can be explained by a simplified reaction pathway, considering a two-step consecutive reaction (Scheme 3) involving the dehydration of the sugar to MMF (step 1) and the formation of ML (step 2).
The experimental data reveal that k1 is a function of the sugar substrate. Based on this scheme, k2 should not be a function of the sugar substrate. As such, a higher reactivity of the sugar (higher k1) is expected to lead to higher RMF yields, which was also shown experimentally in this research.
Scheme 3: Simplified reaction network for C6 sugar conversion to MMF and ML
316| Chapter 4
Figure 8: The yield of methyl levulinate vs. time for fructose (▲), sorbose (●), tagatose (■) -1 and psicose (★) at 65 gL substrate in methanol at 100 °C with (a) 17 mM H2SO4 and (b) 34 mM H2SO4
A comparative study on the reactivity of various ketohexoses to furanics in methanol |317
Figure 9: The yield of methyl levulinate vs. the combined yield of HMF and MMF for fructose (▲), sorbose (●), tagatose (■) and psicose (★) at 65 gL-1 substrate in methanol at 100 °C for all reaction times with 17 or 34 mM H2SO4
The sum of all measured products (HMF, MMF, HAF, MAF, furfural, ML and LA) is plotted against the hexose conversion in Figure 10. In this figure the maximum attainable yield line is given (100% mass balance). Interestingly, below 80% conversion the trend for the sum of measured products vs. conversion for each sugar is approximately parallel with the 100% mass balance line. However, the difference between the measured yields and the 100% mass balance closure line is significant, meaning that other, unidentified, products are formed as well. For instance, the fructose trend is consistently 20% below the 100% mass balance line. However, at higher conversions the difference between the actual total identified yield line and the 100% mass balance closure line is much lower. This is an important observation as it suggests the presence of unidentified intermediates in the early stage of the reaction that in a later stage are converted to RMF and ML, rather than to humin by-products. A possible explanation is the formation of sugar dimers or oligomers in the initial stage of the reaction when the sugar concentration is relatively high. It is possible that these compounds are not detectable by UPLC or that they overlap with the known sugars, but have different
318| Chapter 4 response factors, thereby underestimating the total amount of sugars present. Fructose is known to form dimers, so-called difructose dianhydrides which makes it likely that the other ketoses are also prone to dimerisation.25 As the reactions are performed in methanol, the initial water concentration is very low, favouring dimerisation reactions. These oligomers are known to be in equilibrium with the monomeric sugars and as such the equilibrium will shift to the monomers at higher monomer conversions and accompanying water formation from the dehydration, in line with the experimental data. It also suggest that humin formation, a major by-product in water, is suppressed considerably in methanol and that the amounts of humins are at maximum 20-30 % at high sugar conversion (assuming that all unknowns are humins and not unidentified methanol solubles). For example for psicose, the most reactive hexose with the highest attainable RMF yield, the amount of unidentified products is about 10-15% at psicose conversions higher than 90%, indicating that humin formation is less than 10%.
Additionally, the results in Figure 10 show that the trend in the total product yield for psicose and fructose is comparable. For psicose and fructose the combined yields of furanics and levulinates follow approximately the same trend vs. the conversion. This shows that the lower furanics yield at higher conversion from fructose (Figure 7) is completely compensated by the additional amount of ML formed. As levulinates are generally considered to be formed from HMF/MMF species, they should be considered as a product in the RMF reaction pathway.3 The fact that only the ratio of RMF and ML differs, and not their combined yield, indicates that the selectivity of psicose and fructose towards the RMF pathway is approximately identical. Therefore the conversion rate of these two sugars to RMF has no measurable influence on the selectivity towards the RMF pathway (Scheme 1) relative to other available sugar conversion routes, such as humin formation.
The total product yields for tagatose and sorbose follow a similar trend, though the total yield of identified products is consistently lower than for psicose and fructose. Both tagatose and sorbose form significant amounts of HAF and MAF, whereas for psicose and fructose only traces of HAF/MAF were observed. This indicates the existence of a parallel reaction pathway to HAF/MAF for tagatose and sorbose. As such, the lower amount of total identified products for tagatose and sorbose compared to tagatose and fructose could be (partly) due to additional side-reactions in the HAF formation pathway. A comparative study on the reactivity of various ketohexoses to furanics in methanol |319
Figure 10: The sum of the yields of HMF, MMF, HAF, MAF, furfural, LA and ML vs. the sugar conversion for fructose (▲), sorbose (●), tagatose (■) and psicose (★) at 65 gL-1 substrate in methanol at 100 °C for all reaction times with 17 or 34 mM H2SO4. The dashed line represents 100% mass balance
It is of interest to compare the results from the current study to those published recently by our group for the dehydration of three of the four ketoses in water catalysed by sulphuric acid.18 The reactivity of the ketoses is clearly higher in methanol than in water. In methanol significantly less severe conditions were applied (100 °C, 17, 34 mM H2SO4) to obtain comparable conversion levels to those in the previous study in water (120°C, 100 mM
H2SO4). In water, the experiments were carried out with tagatose, fructose and sorbose. Tagatose was by far more reactive than fructose and sorbose, which is in line with the data reported here in methanol. Thus the reactivity patterns for the three ketohexoses in methanol are in line with the results obtained earlier in water. The study in water was combined with DFT calculations relating the structure of the ketoses to their reactivity. The DFT calculations were in line with the experimental reactivity data for tagatose, fructose and sorbose, and predicted psicose to be the least reactive ketose. These conclusions were based on the differences in calculated stability of proposed intermediate carbocations. The current experimental study in methanol shows that psicose is not the least reactive but as reactive as
320| Chapter 4 tagatose. Thus, additional experiments in water with psicose will be reported to experimentally verify the DFT calculations. An important difference between the reactions in water and methanol can be found in the rate of RMF formation from fructose relative to tagatose. In water, the HMF yields at the same reaction time were higher for tagatose than for fructose whereas in methanol the values for RMF yield in time are comparable. This could be caused by the difference of the orientation of the hydroxyl group on C4, which is the key difference (Scheme 2). The formation of HMF is slower in water than in methanol.18 This suggests that the interaction of the sugar with water has a stronger negative effect on the HMF formation rate if the hydroxyl group on C4 is pointed down than when it is pointed up in Scheme 2. The same relative drop in rate would be expected for psicose relative to tagatose in the case of water. Water has a higher capacity to form hydrogen bonds than methanol and is therefore more likely to disrupt intramolecular hydrogen bonds within the sugars. This makes it plausible that for the formation of HMF an intramolecular hydrogen bond with the C4 hydroxyl group is of importance.
4.3.2 Mechanistic considerations and 13C labelling experiments with sorbose
The observed differences in reactivity of the four ketoses to furanics in methanol are remarkable. The ketoses differ only in the orientation of the hydroxyl group orientations on C3 and C4 (Scheme 2). When the hydroxyl groups are cis-oriented in the Fischer projection, and therefore also in the furanose structures, the conversion rate is higher than when these hydroxyl groups are trans-oriented. This implies that the relative orientation of the hydroxyl groups on C3 and C4 is a key factor with regard to the conversion rate of the sugars and indicates that it plays a major role in the rate-determining step of the reaction sequence. This strongly suggests that the C3-C4 lacks free rotation in the rate-determining step. As such, reaction mechanisms with cyclic intermediates (Scheme 4) are more likely than reaction mechanisms with acyclic intermediates, as the relative orientation of the latter is of less importance due to free rotation of the carbon-carbon bonds. This cyclic mechanism was also considered the most appropriate to explain the differences in the reactivity of ketohexoses and aldohexoses in water with sulphuric acid as the catalyst.18 A comparative study on the reactivity of various ketohexoses to furanics in methanol |321
The formation of HAF and MAF was only observed in significant amounts for tagatose and sorbose, both having the same orientation of the hydroxyl group at C4 (Scheme 2). This suggests that the hydroxyl group orientation at C4 plays a major role in the formation of HAF and MAF.
Scheme 4: The dehydration of fructose through cyclic intermediates as proposed in literature.3
An experiment with L-[6-13C]sorbose was performed to gain insights in the reaction pathways from sorbose to HAF and HMF/MMF. Sorbose was selected as the amount of HAF/MAF is the highest for all ketohexoses tested. The experiments were performed with 17 mM H2SO4 for 150 min at 100 °C, as these conditions gave the highest HAF yields (7%, MAF 2% and MMF around 25%), which was favourable for the analysis by NMR. A 13C- NMR measurement of the reaction mixture after work-up is shown in Figure 11. The largest peak, at δ 66.3 ppm, is attributed to MMF, the major product of hexose dehydration in methanol under these conditions. As expected from the proposed dehydration mechanisms in 3 literature, the label at C6 in sorbose ends up in the CH2 unit of the methoxymethyl group of MMF. The significant peak at δ 29.8 ppm corresponds to the methyl group next to the carbonyl moiety of LA. Thus, the carbonyl group of HMF/MMF is converted into formic acid. Both observations are in agreement with NMR studies by Zhang and Weitz for fructose in water and DMSO and suggests that the formation of HMF from sorbose in methanol and fructose in water follows a similar reaction pathway.26 The smaller peak at δ 66.6 ppm is assigned to the hydroxymethyl group of HAF (see also Figure 12a), which shows that the label at C6 on sorbose ends up in the hydroxymethyl group of HAF. Besides HAF, MAF is also formed in the reaction. The C6 peak of a representative sample of MAF is present at δ 76 ppm (see Figure 12b). Indeed, it is also clearly present in the NMR spectrum of the reaction mixture and indicates that the label is also present at the etherified hydroxymethyl group of MAF (Figure 11).
322| Chapter 4
At around δ 147 ppm HAF resonances are expected in case the labelling is present in the furan ring. Here no significant peak is observed, only a cluster of small peaks is present, excluding the possibility that these originate from HAF. It is more likely that the peaks at around δ 147 ppm originate from minor furan-like products, like furfural,27 furan dimers and trimers, and/or soluble humin by-products.28
The group of very small peaks in the δ 60-64 ppm area are expected to arise from sorbose and methyl sorbosides.29,30
Figure 11: 13C-NMR of the reaction mixture of the acid catalysed dehydration of L-[6-13C]sorbose in methanol
A comparative study on the reactivity of various ketohexoses to furanics in methanol |323
Figure 12: 13C-NMR spectrum of (a) pure 2-hydroxyacetylfuran (HAF) and (b) 2- methoxyacetylfuran (MAF) in CDCl3
The interpretation of the NMR results regarding HAF formation pathways is not straightforward. The reaction mechanism for HAF formation proposed in the literature involves an isomerisation to a 2,3-enediol, which would have placed the 13C label in the furan ring (Scheme 5).21,31
Scheme 5: Reaction mechanism for HAF formation from fructose with acyclic intermediates as proposed in literature21
The labelling experiments presented here do not agree with this mechanism, as the label ends up in the CH2OR side chain. Based on the observation that the orientation of the hydroxyl group on C4 of the ketose plays a key role in the selectivity towards HAF formation, it is hypothesised that the pathway to HAF proceeds through a 1,4-anhydroketose (Scheme 6).
324| Chapter 4
Scheme 6: Proposed formation of HAF with the involvement of a 1,4-anhydrosorbose
The hydroxyl groups on C1 and C4 can be oriented in such a way that they are in close proximity to allow the formation of a second five-membered ring through etherification, particularly in the pyranose (six-membered ring) form, but also in the α-furanose form. The hemiacetal bond in the pyranose ring could then potentially open up due to the ring strain to form 1,4-anhydrosorbose. This would place the labelled carbon (C6) in the proper position to end up in the CH2OR side chain of HAF. A mechanism that explains the conversion of 1,4- anhydrosorbose to HAF is not straightforward. In Scheme 7 four potential pathways are described in order to form HAF from sorbose: two through monocyclic and one through bicyclic intermediates. A comparative study on the reactivity of various ketohexoses to furanics in methanol |325
Scheme 7: Possible mechanisms to form HAF through a bicyclic anhydrosorbose
Both proposed pathways through monocyclic 1,4-anhydrosorbose require the migration of double bonds through the ring to C5 in order to form a carbonyl at the proper position. The main difference is that in one an enediol between C2 and C3 is proposed, whereas in the other an enol is formed between C1 and C2. The first pathway would require migration of the double bond, followed by dehydration at C2. The second pathway would lead to an initial dehydration at C3. The challenge in both cases is the required migration of a double bond to C5, followed by an additional dehydration on the ring to form HAF, since both intermediates are already conjugated hydroxyfuran rings.
326| Chapter 4
The alternative bicyclic pathways suggests dehydration at the C2 position upon formation of the 1,4-anhydrosorbose to form a bicyclic carbocation. This step is comparable to the initial dehydration suggested in the conversion of fructose to HMF (Scheme 4). Steric constraints make it less likely that a carbocation is formed at the expected C2 position, leading to a hydride shift from C1 to C2 to form a carbocation on C1. Even if a carbocation would be formed at C2, then a hydride shift from C1 to C2 is required in order to break the C2-O bond in the following step to form a C1-C2 double bond. For the furanose ring this is then followed by deprotonation on C5, opening the five-membered ring and forming a ketone on C5 and a double bond between C1 and C2. A regular ß-dehydration on C3 then leads to HAF. For the pyranose form the steps are comparable, but an additional isomerisation of the C6 aldehyde to the C5 ketone is required to form HAF. The isomerisation of the aldehyde to the ketone should take place through an enediol and should be favoured because of the formation of a conjugated furfural system.
4.4 Conclusions
In this research, the Brønsted acid catalysed conversion of four different 2-ketohexoses (fructose, sorbose, tagatose and psicose) was investigated in methanol with the objective to identify the most suitable precursor for HMF/MMF formation, and to establish hexose structure-yield relations. Significant differences in the reactivity and selectivity were observed. Psicose and tagatose were the most reactive and psicose gave the highest selectivity towards HMF and MMF (appr. 55% combined yield at around 96% conversion). As such, psicose would be the ideal monosaccharide for the production of HMF and its derivatives. However, psicose is a rare sugar, present in small amounts in wheat, Itea plants, processed sugar cane and beet molasses.32-35 For a techno-economically viable process, either a psicose producing organism or an isomerisation process to obtain psicose from readily accessible sugars would need to be developed.36
The significant differences in reactivity between the four ketohexoses were related to their molecular structure. The hexoses differ only in the orientation of the vicinal hydroxyl groups on C3 and C4, which showed that these structural differences are essential in the dehydration to both HMF/MMF and HAF. It was shown that the reactivity of the sugar is determined by the relative orientation of these hydroxyl groups, with cis-orientation leading A comparative study on the reactivity of various ketohexoses to furanics in methanol |327 to faster conversion than trans-orientation. This is best explained by a dehydration mechanism with cyclic intermediates.
The formation of HAF was shown to depend on the orientation of the hydroxyl group on C4 of the ketohexoses, which consequently also affected the selectivity for the HMF/MMF formation. The reaction mechanism of HAF formation was proposed to involve a 1,4- anhydroketose as an intermediate. This would have all the carbons in the right position to form HAF, as evidenced by performing dehydration experiments with 13C-labelled sorbose.
328| Chapter 4
4.5 References
(1) Clark, J. H.; Deswarte, F. E. I.; Farmer, T. J. Biofuels, Bioprod.Biorefin. 2009, 3, 72-90. (2) Mosier, N.; Wyman, C.; Dale, B.; Elander, R.; Lee, Y. Y.; Holtzapple, M.; Ladisch, M. Bioresour. Technol. 2005, 96, 673-686. (3) Van Putten, R.-J.; Van der Waal, J. C.; De Jong, E.; Rasrendra, C. B.; Heeres, H. J.; De Vries, J. G. Chem. Rev. 2013, 113, 1499-1597. (4) Kamm, B.; Gruber, P. R.; Kamm, M. Biorefineries - Industrial Processes and Products; Wiley-VCH: Weinheim, 2010. (5) Van der Waal, J. C.; De Jong, E. In Producing Fuels and Fine Chemicals from Biomass Using Nanomaterials; Luque, R., Balu, A. M., Eds.; CRC Press: Boca Raton, 2013. (6) Bozell, J. J.; Petersen, G. R. Green Chem. 2010, 12, 539-554. (7) Girisuta, B.; Janssen, L. P. B. M.; Heeres, H. J. Green Chem. 2006, 8, 701-709. (8) Girisuta, B.; Janssen, L. P. B. M.; Heeres, H. J. Chem. Eng. Res. Des. 2006, 84, 339-349. (9) Girisuta, B.; Janssen, L. P. B. M.; Heeres, H. J. Ind. Eng. Chem. Res. 2007, 46, 1696-1708. (10) Roman-Leshkov, Y.; Barrett, C. J.; Liu, Z. Y.; Dumesic, J. A. Nature 2007, 447, 982-985. (11) Eerhart, A. J. J. E.; Faaij, A. P. C.; Patel, M. K. Energy Environ. Sci. 2012, 5, 6407-6422. (12) De Jong, E.; Dam, M. A.; Sipos, L.; Gruter, G. J. M. In ACS Symp. Ser.; Smith, P. B., Gross, R., Eds., 2012. (13) Burgess, S. K.; Leisen, J. E.; Kraftschik, B. E.; Mubarak, C. R.; Kriegel, R. M.; Koros, W. J. Macromolecules 2014, 47, 1383-1391. (14) Girisuta, B. PhD thesis, University of Groningen, 2007. (15) Van Zandvoort, I.; Wang, Y.; Rasrendra, C. B.; van Eck, E. R. H.; Bruijnincx, P. C. A.; Heeres, H. J.; Weckhuysen, B. M. ChemSusChem 2013, 6, 1745-1758. (16) Gruter, G. J. M.; Dautzenberg, F.; Avantium International BV, 2007; WO 2007104514. (17) Dias, A. S.; Van Putten, R.-J.; Gruter, G. J.; Furanix Technologies, 2012; WO2012091570. (18) Van Putten, R.-J.; Soetedjo, J. N. M.; Pidko, E. A.; Van der Waal, J. C.; Hensen, E. J. M.; De Jong, E.; Heeres, H. J. ChemSusChem 2013, 6, 1681-1687. (19) Bicker, M.; Kaiser, D.; Ott, L.; Vogel, H. J. Supercrit. Fluids 2005, 36, 118-126. (20) Miller, R. E.; Cantor, S. M. J. Am. Chem. Soc. 1952, 74, 5236-5237. (21) Harris, D. W.; Feather, M. S. Tetrahedron Lett. 1972, 13, 4813-4816. (22) Heyns, K.; Hauber, R. Liebigs Ann. Chem. 1970, 733, 159-169. (23) Rigal, L.; Gaset, A. Biomass 1983, 3, 151-163. (24) Moriarty, R. M.; Prakash, O.; Duncan, M. P.; Vaid, R. K.; Musallam, H. A. J. Org. Chem 1987, 52, 150-153. (25) Christian, T. J.; Manley-Harris, M.; Field, R. J.; Parker, B. A. J. Agric. Food Chem. 2000, 48, 1823-1837. (26) Zhang, J.; Weitz, E. ACS Catal. 2012, 2, 1211-1218. (27) Kallury, R. K. M. R.; Ambidge, C.; Tidwell, T. T.; Boocock, D. G. B.; Agblevor, F. A.; Stewart, D. J. Carbohydr. Res. 1986, 158, 253-261. (28) Aida, T. M.; Tajima, K.; Watanabe, M.; Saito, Y.; Kuroda, K.; Nonaka, T.; Hattori, H.; Smith Jr., R. L.; Arai, K. J. Supercrit. Fluids 2007, 42, 110-119. (29) Angyal, S. J.; Bethell, G. S. Aust. J. Chem. 1976, 29, 1249-1265. (30) Bock, K.; Pedersen, C. Adv. Carbohydr. Chem. Biochem. 1983, 41, 27-66. (31) Kuster, B. F. M. Starch/Stärke 1990, 42, 314-321. (32) Miller, B. S.; Swain, T. J. Sci. Food Agric. 1960, 11, 344-348. (33) Binkley, W. W. Int. Sugar. J. 1963, 65, 105-106. (34) Hossain, A.; Yamaguchi, F.; Matsunaga, T.; Hirata, Y.; Kamitori, K.; Dong, Y.; Sui, L.; Tsukamoto, I.; Ueno, M.; Tokuda, M. Biochem. Biophys. Res. Commun. 2012, 425, 717-723. (35) Baek, S. H.; Park, S. J.; Lee, H. G. J. Food Sci. 2010, 75, 49–53. (36) Izumori, K.; Yamakita, M.; Tsumura, T.; Kobayashi, H. J. Ferment. Bioeng. 1990, 70, 26-29.
|329
5 Reactivity studies on the acid- catalysed dehydration of ketohexoses to 5-hydroxymethylfurfural in water
Abstract
The use of the four possible ketohexoses (fructose, tagatose, sorbose and psicose) for HMF synthesis was explored in water using sulphuric acid as the catalyst (33 mM H2SO4, 120 °C). Significant differences in reactivity were observed and tagatose (48% conversion after 75 min) and psicose (35% conversion after 75 min) were clearly more reactive than fructose and sorbose (around 20% conversion after 75 min). The results indicate that the relative orientation of the hydroxyl groups on C3 and C4 has a major effect on the reactivity. This suggests that the dehydration towards HMF takes place via a mechanism with cyclic intermediates in which the C3-C4 bond is fixed in a ring structure. The selectivity to HMF was found to be higher for fructose and psicose than for tagatose and sorbose. Hydroxyacetylfuran (HAF) was shown to be a by-product for mainly sorbose and tagatose (as high as 2% yield). A reaction mechanism scheme involving a bicyclic structure is proposed to explain the formation of HAF. The reactivity of the sugars was significantly lower in water than previously observed in methanol.
330| Chapter 5
5.1 Introduction
Hydroxymethylfurfural (HMF) is considered a high-potential building block for biobased materials and fuels.1 It can, for instance, be oxidised to 2,5-furandicarboxylic acid (FDCA), which can be used to produce polyethylenefuranoate (PEF), a high-potential biobased polyethyleneterephtalate (PET) replacement, in which the terephtalic acid moiety is replaced by FDCA. PEF production was calculated to reduce non-renewable energy use and greenhouse emissions significantly when compared to PET production.2 Moreover materials research on PEF has revealed superior product properties compared to PET.3 HMF is obtained by the acid-catalysed dehydration of C6 sugars (hexoses, Scheme 1).4 HMF is an intermediate product and reacts further with water in the presence of acid to form levulinic acid5-7 (Scheme 1) and with sugars and intermediates to form polymeric material (humin).8 The maximum obtainable HMF yield is highly dependent on the type of hexose used and the solvent system applied.4 For fructose, a ketose, typical yields are around 50% HMF in aqueous systems, whereas yields over 90% are reported in DMSO and certain ionic liquids. Glucose, an aldose, gives much lower HMF yields than fructose. For instance, the Brønsted acid-catalysed dehydration of glucose in water typically results in an HMF yield below 10%. HMF yields over 50% from glucose are possible though, when using aprotic polar solvents and ionic liquids in which a catalyst is added that facilitates glucose-fructose isomerisation.4
Scheme 1. The dehydration of hexoses to form HMF, hydroxyacetylfuran (HAF), and levulinic and formic acid
Besides differences in reactivity between ketoses and aldoses, considerable differences in reactivity between the various ketoses (Scheme 2) have been observed.9-11 These differences in reactivity are relevant for the development of an economically viable process for biobased Reactivity studies on the acid-catalysed dehydration of ketohexoses to HMF in water |331
HMF. High yields of HMF are required as downstream processing is cumbersome. Particularly HMF separation from unreacted sugars and HMF purification are challenging. As such, sugar feeds which produce HMF at high yields, i.e. minimum by-product formation at high sugar conversions, are strongly favoured. In addition, the reactivity differences can provide additional understanding of the important factors determining the selectivity in HMF formation, which allows rational design of improved, highly active and selective catalysts.
In an earlier study (see chapter 3 of this thesis), we have performed experimental studies on the reactivity of fructose, tagatose and sorbose in water, which showed that tagatose was the most reactive. Density functional theory (DFT) calculations were in agreement with the experimental reactivity pattern and, in addition, also predicted that psicose would be the least reactive ketose.10 A follow-up study in methanol was performed using all four possible ketohexoses and revealed that tagatose and psicose were equally reactive, with psicose exhibiting a higher selectivity towards HMF and its methyl ether methoxymethylfurfural (MMF).11 The latter was unexpected based on the DFT calculations in water and as a result, a subsequent study for psicose in water seems justified. Due to improved analytical methods, the study in methanol also revealed the formation of hydroxyacetylfuran (HAF) from tagatose and sorbose, which was not observed previously in water,10 even though it is a known by- product in HMF formation from hexoses.12-14 The first report on the formation of HAF from C6-sugars dates from 1952 (Miller).14 HAF has earlier been reported as a degradation component of fructose-containing solutions for intravenous nutrition.15 Rapp and co- workers16 identified HAF in wine and concluded that its presence is circumstantial evidence that inverted sucrose solutions have been added to the grape must or wine.
The current study involves experimental research on the acid-catalysed dehydration of fructose, tagatose, sorbose and psicose in water, using the improved analytics not available in the previous study in water10. The objectives are i) to experimentally determine the reactivity of psicose in water and ii) to compare the results with the other three ketoses, iii) to quantify the formation of by-products like HAF and, as such, to gain insights in the reactivity differences of the four ketoses in water. Finally, the results in water will be compared with those obtained earlier in methanol.
332| Chapter 5
Scheme 2. The α-D-furanose structures of the four ketohexoses
5.2 Experimental section
The experimental procedures were similar to those described in a previous study using methanol by our group,11 the main differences being that the experiments were performed in water (Milli-Q quality) instead of methanol and at 120 °C instead of 100 °C. Sulphuric acid (33 mM) was used as the catalyst and the sugar intake was set at 65 gL-1. All experiments were carried out in triplicate. The average value of the substrate conversion and product yield, including the range of the experimental results, are given in the relevant figures.
5.3 Results and discussion
The experiments were performed in a batch mode using a high throughput screening device. The experimental conditions for this study were taken from our previous study10 and involve a sulphuric acid concentration of 33 mM, a reaction temperature of 120 °C and 65 gL-1 aqueous sugar solutions as the initial intake.
5.3.1 Sugar reactivity
The conversions of all ketohexoses versus the batch time are given in Figure 1. Clear differences in reactivity are visible. Tagatose is the most reactive, with a conversion of 48% after 75 min batch time, followed by the slightly less reactive psicose (35% conversion after 75 min). Fructose and sorbose are equally reactive, though much less reactive than tagatose and psicose. For instance, the conversion after 75 minutes for both fructose and sorbose was about 20%, which is a factor of 2.5 lower than for tagatose. The experimental conversion versus time trends for fructose, sorbose and tagatose are in line with those reported previously by us in water.10 The lower reactivity of psicose compared to tagatose is surprising based on their equal reactivity observed earlier in methanol. Reactivity studies on the acid-catalysed dehydration of ketohexoses to HMF in water |333
Figure 1. The conversion vs. reaction time for fructose (▲), sorbose (●), tagatose (■) and -1 psicose (◆) at 65 gL substrate concentration with 33 mM H2SO4 at 120 °C.
In our previous research on ketose dehydration in water, in which psicose was not tested experimentally, DFT calculations were performed to support the experimental findings.10 Based on the calculated stability of proposed carbocation intermediates, these calculations provided an explanation for the differences in reactivity between tagatose, fructose and sorbose, and predicted psicose to be the least reactive ketose. The present results, however, show that psicose is clearly not the least reactive ketohexose, which is not in agreement with the DFT results and thus the model requires further refinement.
Of interest is the lower reactivity of the ketoses in water compared to methanol,11 indicating that the sugars are apparently more stable in water than in methanol. For instance, 48% tagatose conversion was observed after 75 min batch time in water (120 °C, 33 mM sulphuric acid), whereas in methanol the same conversion was obtained after 15 min under much less severe conditions (100 °C, 17 mM sulphuric acid). This shows that the solvent has a significant effect on the rate of the dehydration reaction. This can be explained by differences in stabilisation of the sugar or potential reactive intermediates by hydrogen bonding (intramolecular vs. intermolecular) in water and methanol.
334| Chapter 5
Other factors that could play a role are differences in the tautomeric distribution in water and methanol or the formation of methyl ketosides in the case of methanol. The time scale of tautomeric rearrangements17 of sugars is in general much faster than that of the dehydration reaction of hexoses to HMF.10 Therefore the equilibrium tautomeric distribution is not expected to play an important role. The formation of methyl ketosides in acidic methanol requires relatively mild conditions compared to the dehydration of ketoses and as such are expected to be formed relatively fast and play a major role.11 The rate of the dehydration reaction to HMF could be positively affected by the formation of the methyl acetal at C2 (Scheme 3), as it affects possible hydrogen bonding modes of the sugar molecule, and therefore also its reactivity.
Scheme 3. The formation of methyl ketosides in methanol
5.3.2 HMF yield
The HMF yields in time for the 4 substrates are shown in Figure 2. HMF yields of up to 30% at 75 min batch time were observed for the most reactive sugar, tagatose. The HMF yields for the less reactive sugars, fructose and sorbose, are at least a factor of two lower than for tagatose (16% for fructose and 12% for sorbose after 75 min batch time). However, a proper comparison of yield/selectivity data is only possible at similar conversion levels and therefore the HMF yield vs. the conversion for all four substrates is shown in Figure 3. These findings suggest that fructose and psicose follow the same trend, like sorbose and tagatose, though the latter two at a lower yield level.
The reactivity pattern for the various sugars in water is in line with the results at low sugar conversion in methanol.11 However, different trends are observed when comparing the combined yields of HMF and MMF vs. conversion with those reported in methanol.11 At lower conversions, the measured selectivity for HMF/MMF in methanol was significantly lower than HMF selectivity in water. This may be explained by the likely equilibrium formation of ketose dianhydrides and oligomers in methanol.11 Unfortunately, the spread in the conversion values (see also Figure 1) is large for some of the data points (triplicate Reactivity studies on the acid-catalysed dehydration of ketohexoses to HMF in water |335 measurements), making it difficult to draw sound conclusions. This is somewhat surprising, as all experiments were performed in the same high-throughput device, using similar methodology and in the same time frame as the previous work in methanol, for which the reproducibility of the conversion data was much better.11 The spread in the yield data is very small (see also Figure 2) and of comparable quality to that in methanol. The key difference between the experiments is the use of water instead of methanol. In water, solid humins were formed in the course of the reaction, whereas this was not the case in methanol. These solid humins are known to physically entrap products (e.g. levulinic acid) and reactants (e.g. sugars) to a variable extent,8 which could be an explanation for the large variance in the conversion data.
Figure 2. The yield of HMF vs. reaction time for fructose (▲), sorbose (●), tagatose (■) and -1 psicose (◆) at 65 gL substrate concentration with 33 mM H2SO4 at 120 °C.
336| Chapter 5
Figure 3. The yield of HMF vs. conversion for fructose (▲), sorbose (●), tagatose (■) and -1 psicose (◆) at 65 gL substrate concentration with 33 mM H2SO4 at 120 °C.
The yield vs. conversion data (Figure 3) in this research are in line with our previous findings in water.10 The average HMF yield values (Figure 3) suggest that fructose and psicose are converted to HMF at a comparable selectivity and that the selectivity is higher than for tagatose and sorbose. This trend was also observed in the previous study in methanol with regard to the combined selectivity for MMF and HMF.11 Based on these observations, the proposed reaction mechanism involving cyclic intermediates (Scheme 4) is most likely also valid for water.
Scheme 4. HMF formation from fructose through cyclic intermediates
Reactivity studies on the acid-catalysed dehydration of ketohexoses to HMF in water |337
5.3.3 Product selectivity
For sugar dehydrations the selectivity towards HMF is generally limited by humin and levulinic acid formation. Within the sugar conversion window (< 50%), the levulinic acid yield was always below 2% for all ketoses. The majority of the unknown products are expected to be humins, though some sugar dimers and oligomers could also be present. The lower selectivity for HMF formation from tagatose and sorbose compared to psicose and fructose in methanol was previously shown to be caused in part by the formation of significant amounts of hydroxyacetylfuran (HAF), a product hardly formed from fructose and psicose.11 In this study, the formation of HAF was indeed observed from tagatose and sorbose (Figure 4), reaching yields over 2%, whereas only traces of HAF formation were observed from psicose and fructose. These results are in line with the findings in methanol11 and therefore support the conclusion that the orientation of the hydroxyl group on C4, which is the same for tagatose and sorbose (Scheme 2), determines the rate of formation of HAF. This is consistent with a reaction mechanism through a 1,4-anhydroketose as was previously proposed (Scheme 5).11
Figure 4. The HAF yield vs. reaction time for sorbose (●) and tagatose (■) at 65 gL-1 substrate concentration with 33 mM H2SO4 at 120 °C. The error bars indicate the range of the obtained results.
338| Chapter 5
Scheme 5. HAF formation via 1,4-anhydrosorbose11
5.4 Conclusions
Significant differences in reactivity for the four ketohexoses for the acid-catalysed dehydration to furans in water were observed. Psicose and tagatose were shown to be more reactive and were converted at a higher rate than fructose and sorbose. Furthermore, the selectivity for HMF formation was higher for psicose and fructose than for sorbose and tagatose. Thus, psicose appears to be the best substrate for HMF production in water, as it has a higher selectivity for HMF formation than sorbose and tagatose, and a higher reactivity than fructose. These results support a dehydration mechanism with cyclic intermediates where the relative orientation of the hydroxyl groups on C3 and C4 determines the reaction rate to HMF. The results regarding the relative reactivity of the four ketohexoses in water are consistent with those obtained in a previous study from our group in water for three ketohexoses (excluding psicose). However, the findings, particularly those for psicose, are not in line with earlier DFT calculations with proposed single dehydrated carbocation intermediates. Further activities are in progress to get a better understanding of this anomaly between experiments and calculations. For tagatose and sorbose, small though significant amounts of the by-product hydroxyacetylfuran (HAF, as high as 2%) were formed. This indicates that the orientation of the hydroxyl group on C4 of the sugar determines the rate of HAF formation. This finding is in line with previous studies in methanol.
Reactivity studies on the acid-catalysed dehydration of ketohexoses to HMF in water |339
5.5 References (1) Bozell, J. J.; Petersen, G. R. Green Chem. 2010, 12, 539-554. (2) Eerhart, A. J. J. E.; Faaij, A. P. C.; Patel, M. K. Energy Environ. Sci. 2012, 5, 6407-6422. (3) Burgess, S. K.; Leisen, J. E.; Kraftschik, B. E.; Mubarak, C. R.; Kriegel, R. M.; Koros, W. J. Macromolecules 2014, 47, 1383-1391. (4) Van Putten, R.-J.; Van der Waal, J. C.; De Jong, E.; Rasrendra, C. B.; Heeres, H. J.; De Vries, J. G. Chem. Rev. 2013, 113, 1499-1597. (5) Girisuta, B.; Janssen, L. P. B. M.; Heeres, H. J. Green Chem. 2006, 8, 701-709. (6) Girisuta, B.; Janssen, L. P. B. M.; Heeres, H. J. Chem. Eng. Res. Des. 2006, 84, 339-349. (7) Girisuta, B.; Janssen, L. P. B. M.; Heeres, H. J. Ind. Eng. Chem. Res. 2007, 46, 1696-1708. (8) van Zandvoort, I.; Wang, Y.; Rasrendra, C. B.; van Eck, E. R. H.; Bruijnincx, P. C. A.; Heeres, H. J.; Weckhuysen, B. M. ChemSusChem 2013, 6, 1745-1758. (9) Binder, J. B.; Cefali, A. V.; Blank, J. J.; Raines, R. T. Energy Environ. Sci. 2010, 3, 765-771. (10) Van Putten, R.-J.; Soetedjo, J. N. M.; Pidko, E. A.; Van der Waal, J. C.; Hensen, E. J. M.; De Jong, E.; Heeres, H. J. ChemSusChem 2013, 6, 1681-1687. (11) Van Putten, R.-J.; Van de Bovenkamp, H. H.; Van der Waal, J. C.; De Jong, E.; Heeres, H. J. 2014, Manuscript in preparation, see Chapter 4 of this thesis. (12) Harris, D. W.; Feather, M. S. Tetrahedron Lett. 1972, 13, 4813-4816. (13) Heyns, K.; Hauber, R. Liebigs Ann. Chem. 1970, 733, 159-169. (14) Miller, R. E.; Cantor, S. M. J. Am. Chem. Soc. 1952, 74, 5236-5237. (15) Jellum, E.; Børresen, H. C.; Eldjarn, L. Clin. Chim. Acta 1973, 47, 191-201. (16) Rapp, A.; Mandery, H.; Heimann, W. Vitis 1983, 22, 387-394. (17) Wertz, P. W.; Garver, J. C.; Anderson, L. J. Am. Chem. Soc. 1981, 103, 3916-3922.
340| Chapter 5
|341
6 Experimental and modelling studies on the solubility of D-arabinose, D- fructose, D-glucose, D-mannose, sucrose and D-xylose in methanol and methanol-water mixtures*
Abstract
The solubilities of D-glucose, D-arabinose, D-xylose, D-fructose, D-mannose and sucrose in methanol and methanol-water mixtures (<25 wt% water) were determined at temperatures between 295 and 353 K using a unique high-throughput screening technique. The data were modelled with a UNIQUAC framework with an average error between calculated and experimental data of 3.7%. The results provide input for the design of efficient chemical processes for the conversion of these sugars into valuable biobased building blocks in methanol-water mixtures.
* Published as ‘Experimental and modelling studies on the solubility of D-arabinose, D-fructose, D- glucose, D-mannose, sucrose and D-xylose in methanol and methanol-water mixtures,’ Van Putten, R.-J.; Winkelman, J. G. M.; Keihan, F.; Van der Waal, J. C.; De Jong, E.; Heeres, H. J., Ind. Eng. Chem. Res., 2014, 53, 8285–8290
342| Chapter 6
6.1 Introduction
Lignocellulosic biomass is considered an attractive renewable source for the production of biofuels and biobased chemicals and intense research and development activities have been performed the last decade to identify techno-economically viable routes. Well-known examples are the conversion of lignocellulosic biomass to bioethanol as a second generation biofuel, and the conversion of C6-sugars to platform chemicals like 5-hydroxymethylfurfural (HMF) and levulinic acid.1-4
For most of these conversions, water was the initial solvent of choice as it is considered environmentally benign and a good solvent for monomeric sugars. Especially aqueous fermentation of sugars to biofuels and biobased chemicals will play an important role.5 There are, however, significant disadvantages for using water as solvent for chemical processes. General disadvantages of using water as the solvent for any chemical process are the high cost of its removal due to its high heat of vaporisation (2.4 kJg-1)6 and its corrosiveness at elevated temperatures. The importance of a more classical chemocatalytic approach to the conversions of sugars is expected to gain momentum.7 For some high-potential chemo- catalytic processes the carbon yields in water cannot meet techno-economic targets due to excessive by-product formation. Alternative solvents have been explored , especially in the case of the production of furans from sugars.1
Unfortunately, the solubility of monomeric sugars in common organic solvents is rather limited and sugars typically only dissolve well in high-boiling polar aprotic solvents such as DMSO and NMP. However, the relatively high boiling points of these solvents result in serious issues in the product work-up and therefore lower boiling solvents are favoured. As such, the use of lower alcohols is currently extensively explored in the development of carbohydrate (pre-)treatment and conversion processes, such as the organosolv processes using ethanol8 (boiling point 351 K, heat of vaporisation 0.84 kJg-1)6 and sugar dehydration to 5-hydroxymethylfurfural derivatives.1, 9 Saka and co-workers have looked extensively into cellulose and lignocellulosic biomass pre-treatment in methanol10, 11 (boiling point 338 K, heat of vaporisation 1.11 kJg-1).6 Research by Bicker et al.9 and Avantium12 has shown that methanol is the solvent of choice for the production of 5-hydroxymethylfurfural (HMF) and its methyl ether (MMF). Experimental and modelling studies on the solubility of D-arabinose, D-fructose, D-glucose, |343 D-mannose, sucrose and D-xylose in methanol and methanol-water mixtures
For the development of efficient green processes solvent usage and recycle should be minimised and as such there is a large incentive for performing reactions in highly concentrated solutions. For this purpose, detailed solubility data of the sugars in solvents other than water are required. Limited research has been performed on the solubility of sugars in alcohols, see Table 1 for an overview. The majority of the studies focus on glucose and fructose solubilities using methanol and ethanol as the solvent. The emphasis is mainly on alcohol-water mixtures and less on the pure alcohols. Furthermore the temperature ranges for individual studies are often very narrow. Limited attention is given to higher alcohols and to other sugars.
Table 1. Overview of research studies on sugar solubility in alcohols
Temperature range Sugar Solvent (K) Model type Reference Galactose Ethanol/water 273, 288 and 303 UNIFAC 13
Xylose and mannose Ethanol and ethanol/water 278, 288 and 298 UNIFAC 14
Glucose, fructose, xylose, Methanol, ethanol, n- 298-398 UNIFAC 15 galactose, mannose, propanol, n-butanol, t- sucrose pentanol
Glucose, fructose, sucrose Methanol, methanol/ethanol 298, 313, 333 UNIFAC 16 and methanol/water and Ethanol/water
Fructose Methanol, 298, 313, 333 UNIFAC 17 methanol/ethanol, and methanol/water and UNIQUAC Ethanol/water
Sucrose Ethanol and ethanol/water 310 UNIQUAC 18
Fructose and glucose Ethanol/water 303 UNIQUAC 19
Glucose Methanol, ethanol, 313 and 333 UNIQUAC 20 methanol/water, methanol/ethanol
Glucose, fructose, sucrose Methanol, methanol/ethanol 298, 313, 333 UNIQUAC 21 and methanol/water and ethanol/water
Glucose, fructose, Ethanol and ethanol/water 295, 303, 313 UNIQUAC 22 galactose
344| Chapter 6
In this paper, a systematic study is presented on the solubility of a range of C6 and C5 sugars (D-arabinose, D-fructose, D-glucose, D-mannose and D-xylose) and one disaccharide (sucrose) in methanol and methanol-water mixtures containing up to 25 wt% water. The sugars are shown in Scheme 1 in which the monosaccharides are presented in their pyranose forms as it is likely that these will be the most abundant tautomers in water and in methanol.23, 24
To the best of our knowledge, this is the most extensive study reported in this field so far, allowing a proper comparison of the solubilities of the sugars under study at a wide range of conditions. The experiments were performed using a unique Avantium high-throughput device (Crystal16™). A wide range of temperatures (295-353 K) and methanol-water ratios (0-25 wt%) were applied. The equipment allows sixteen parallel experiments at temperatures higher than the boiling point by measuring within a closed vial. The solubility data were modelled using a UNIQUAC model.
Scheme 1: The sugars tested in this study, with all monosaccharides in their pyranose forms
Experimental and modelling studies on the solubility of D-arabinose, D-fructose, D-glucose, |345 D-mannose, sucrose and D-xylose in methanol and methanol-water mixtures
6.2 Experimental section
6.2.1 Chemicals
Sucrose (99.6%) was obtained from Fluka. All other sugars (pure D-form) were obtained from Sigma-Aldrich at >99% purity. Methanol was obtained from Fisher and was HPLC gradient grade. Milli-Q grade water was used for all experiments.
6.2.2 Solubility measurements
All experiments were performed using the Avantium Pharmatech Crystal16™ equipment. This equipment has been specifically designed for determining clear points and cloud points for crystallisation studies. It consists of four heating blocks with magnetic stirring, with four slots each for standard 1.6 mL HPLC vials. Validation of the equipment showed that the temperature of the liquid in the vial was always within 1 K of the set point temperature for the complete temperature range of the experiments. Each reactor has a light source and a sensor. The equipment determines the percentage of light passing through the vials allowing for automatic turbidity measurements and thus for the determination of the clear points and cloud points of the solutions under investigation.
The software provided with the Crystal16™ equipment allows for automatic and regulated temperature increase and cooling sequences. Here, one cycle consisted of heating up at a rate of 3 K/h to 353 K, followed by cooling at a rate of 1 or 2 K/min to 268 K. The experiments generally consisted of 3 cycles. Solutions were prepared from weighed amounts of all components, making the error in sample preparation significantly less than 1%. The experiments were performed at approximately 1 mL scale. After the experiments the vials were weighed to check for loss of the solvent by evaporation due to leakage. Significant loss of solvent was not observed.
346| Chapter 6
6.2.3 UNIQUAC modelling
From thermodynamics the well-known equation for the solubility of a solid in a liquid is obtained as:25
H f ,i 1 1 Cpi Tf ,i Tf ,i ln(ixi ) ( ) (1 ln ) (1) R Tf ,i T R T T where i and xi denote the activity coefficient and the molar fraction of the solute in the liquid phase. Cpi denotes the difference of the solute's specific heat in the solid and liquid phases, T is the actual temperature and T f ,i the fusion point temperature. Equation (1) is obtained with the assumptions that the solid phase that is in equilibrium with the solution consists of the pure solute and that is independent of the temperature in the range of to .
Equation (1) can be used to evaluate the experimental results if a model for the activity of the solute is implemented. Here the UNIQUAC model formulation26 was chosen to model the activity coefficients because of its suitability for the systems under investigation, i.e. a highly polar, non electrolyte, low pressure mixture that contains both large and small 27 molecules. From Equation (2) ln i is calculated:
i i x j ji, j lni 1i lni 5qi (1 ln ) qi (1 ln Si ) S i i j j (2) where and depend on the volume and surface area parameters ri and qi , and S also on the binary parameters ij (see list of symbols).The values of the parameters and were taken from Poling et al.27 (Table 2).
The adjustable binary parameters are defined in terms of binary energy interaction parameters Aij :
exp(A / T) ij ij (3) where in general Aii A jj 0 and Aij A ji . The values of the parameters were obtained by fitting the calculated solubilities of equations (1)-(2) to the measured ones using a standard Newton routine for non-linear optimisation. Experimental and modelling studies on the solubility of D-arabinose, D-fructose, D-glucose, |347 D-mannose, sucrose and D-xylose in methanol and methanol-water mixtures
The physical properties of the carbohydrates mentioned in equation (1), i.e. the heat of fusion, the temperature of fusion and the solid-liquid specific heat difference, were taken from literature when available, or estimated following literature (Table 3).
27 Table 2. UNIQUAC parameters ri and qi
M [g/mol] methanol 32.04 1.9011 2.048 water 18.01528 0.92 1.40 arabinose 150.13 6.7089 6.492 fructose 180.16 8.1589 8.004 glucose 180.16 8.1558 7.92 mannose 180.16 8.1558 7.92 sucrose 342.30 14.5586 13.764 xylose 150.13 6.7089 6.492
Table 3. Physical properties of the sugars used in this study sugar H f T f [K] C p [kJ/mol] [J/mol.K] arabinose 35.7828 43529 120b fructose 33.030 37830, 31 120a glucose 3228, 30, 32 42029-32 120a mannose 24.728 40428, 29 120a sucrose 46.228 45229, 31 25433 xylose 31.728 42028, 29, 31 120a a estimated by Ferreira et al.15 b estimated following Ferreira et al.15
348| Chapter 6
6.3 Results and discussion
6.3.1 Experimental studies
The solubility of the six sugars in methanol and methanol/water mixtures (0-25 wt% water) was studied at temperatures in the range of 298-353 K. Determination of the solubility of fructose at a water content higher than 10 wt% proved not possible due to stirring issues as a results of the high solid loading required by the high solubility under those conditions. The experimental results, combined with the modelled solubilities (vide infra) are shown in Figures 1-6. As expected the solubility of the sugars increased with temperature and water content. Of the sugars tested, sucrose shows by far the lowest solubility in methanol, followed by glucose and arabinose with almost equal solubility, xylose, mannose and fructose. The order was the same in methanol containing 10 wt% water, with the exception that glucose now possessed significantly better solubility than arabinose.
Figure 1. The solubility of glucose in pure methanol and methanol/water mixtures containing up to 25 wt% water
Experimental and modelling studies on the solubility of D-arabinose, D-fructose, D-glucose, |349 D-mannose, sucrose and D-xylose in methanol and methanol-water mixtures
Figure 2. The solubility of arabinose in pure methanol and methanol/water mixtures containing up to 25 wt% water
Figure 3. The solubility of xylose in pure methanol and methanol/water mixtures containing up to 25 wt% water
350| Chapter 6
Figure 4. The solubility of fructose in pure methanol and methanol/water mixtures containing up to 10 wt% water
Figure 5. The solubility of mannose in pure methanol and methanol/water mixtures containing up to 25 wt% water
Experimental and modelling studies on the solubility of D-arabinose, D-fructose, D-glucose, |351 D-mannose, sucrose and D-xylose in methanol and methanol-water mixtures
Figure 6. The solubility of sucrose in pure methanol and methanol/water mixtures containing up to 25 wt% water
6.3.2 Modelling studies
The UNIQUAC binary energy interaction parameters were obtained by matching the calculated to the measured solubilities and the results are given in Table 4. The data were modelled on a mass basis to avoid possible complications due to density changes.
Table 4. Modelled UNIQUAC binary energy interaction parameters.
Aij Methanol water arabinose fructose glucose mannose sucrose xylose methanol 0 -13.88 79.48 -6.64 92.31 13.72 43.73 37.67 water 167.19 0 57.73 65.91 -124.81 6558.79 -127.48 306.31 arabinose 112.39 -1.34 0 - - - - - fructose 171.59 48.58 - 0 - - - - glucose 122.22 380.87 - - 0 - - - mannose 226.14 -225.16 - - - 0 - - sucrose 187.87 410.38 - - - - 0 - xylose 152.50 -164.80 - - - - - 0
352| Chapter 6
A parity plot of experimental versus calculated values for 323 datapoints is shown in Figure 7. This figure clearly shows that the model fits the experimental results very well.
Figure 7. Parity plot of the solubilities of six carbohydrates (323 points)
The modelling results for all sugars are given in Figures 1-6. The measured solubilities show consistent trends and the data points are all very close to the model lines, again indicating that the model fits the data very well. Deviations between the model and experimental data in the figures are mainly due to small differences in the actual water concentration in the solutions and not due to modelling inaccuracies. For instance, in the specific case of Figure 1 at 15 wt% water, the experimental points above the line were actually obtained at 16.1 wt% water in methanol and those below that at 15.6 wt% water, whereas the line represents the model prediction at a 15 wt% water content. The actual water contents were used in the model.
The average absolute error between the calculated model and the experimental data was 3.7%, which shows that the obtained models are accurate representations of the solubilities of these sugars in methanolic mixtures. As such, the high throughput technology is very well suited for this type of solubility research.
6.3.3 Literature comparison
The solubility of fructose, glucose and sucrose in methanol and methanol-water mixtures has been reported (Table 1). Figures 8-10 show a comparison of the experimental solubility data Experimental and modelling studies on the solubility of D-arabinose, D-fructose, D-glucose, |353 D-mannose, sucrose and D-xylose in methanol and methanol-water mixtures obtained in this study with those previously reported. In addition, the model lines are provided. For fructose (Figure 8) agreement between our and the literature data is in general good in the range 310-325K. In addition, the combined data are predicted well using the model provided in this paper. However, at both the high and low end of the temperature range, some deviations between our model and the experimental literature data are observed. We were not able to obtain datapoints for fructose in this temperature regime (max 10 wt% fructose loading) making comparison cumbersome.
0.3 fructose 0.25
0.2 10% ] - 0.15 x [ x 0% 0.1
0.05
0 290 300 310 320 330 T [K]
Figure 8. Comparison of fructose solubility vs. temperature in pure methanol (0%) and methanol with 10% by weight water (10%). Symbols: closed symbols: 0% water; open symbols: 10% water; ▲: Montañes et al.22; , : Macedo and Peres17; □, ■: this work. Lines: calculated according to the model of this work.
In Figure 9 the experimental glucose data and the model developed in this work are compared with the results from literature. In pure methanol, our experimental data and those obtained by Montañes et al.22 and Macedo and Peres20 are in agreement, though the literature results are consistently slightly higher. The glucose solubilities in methanol with 10 wt% water at 313 and 333 K reported by Macedo and Peres20 are only slightly above the model line obtained in this work. At 20 wt% water content the results reported by Macedo and Peres20 are significantly higher than the model line obtained in this work.
354| Chapter 6
0,1 glucose 20% 0,08
10%
0,06
]
- x x [ 0,04
0,02 0%
0 300 310 320 330 340 350 360 T [K]
Figure 9. Comparison of glucose solubility vs. temperature in pure methanol (0%) and methanol with 10% and 20% by weight water (10%, 20%). Symbols: closed symbols: 0% and 20% water; open symbols: 10% water; ▲: Montañes et al.22 , : Macedo and Peres20; □, ■: this work. Lines: calculated according to the model of this work.
In Figure 10 the sucrose solubility data and model from this work are compared to results from literature. At all three water concentrations given, the data reported by Macedo and Peres20 are exactly in line with the model and data obtained in this work.
0.05 sucrose 20% 0.04
0.03
]
- x [ x 0.02 10%
0.01 0%
0 290 300 310 320 330 340 350 360 T [K]
Figure 10. Comparison of sucrose solubility vs. temperature in pure methanol (0%) and methanol with 10% and 20% by weight water (10%, 20%). Symbols: closed symbols: 0% and 20% water; open symbols: 10% water; , : Macedo and Peres20; □, ■: this work. Lines: calculated according to the model of this work. Experimental and modelling studies on the solubility of D-arabinose, D-fructose, D-glucose, |355 D-mannose, sucrose and D-xylose in methanol and methanol-water mixtures
6.4 Conclusions
The solubilities of six sugars in methanol and methanol-water mixtures (< 25 wt% water) were successfully determined using a unique method employing Avantium high-throughput technology. This allowed the measurement of many datapoints with high accuracy and reproducibility at a wide temperature range and even above the boiling point of the solvent. The data were successfully modelled using the UNIQUAC model. The results provide valuable input for the development of efficient processes for biomass conversions in methanol, and methanol water mixtures.
6.5 Symbols
Ai, j UNIQUAC binary energy interaction parameter, K
-1 -1 C p solid-liquid difference of specific heat, J mol K
-1 H f heat of fusion, J mol qi UNIQUAC surface area parameter ri UNIQUAC volume parameter
Si UNIQUAC parameter, Si xkk k,i k
T f temperature of fusion, K x molar fraction, -
i UNIQUAC parameter, i ri / xkrk k
i UNIQUAC parameter, i qi / xkqk k
i, j UNIQUAC binary interaction parameter
356| Chapter 6
6.6 References (1) Van Putten, R.-J.; Van der Waal, J. C.; De Jong, E.; Rasrendra, C. B.; Heeres, H. J.; De Vries, J. G., Hydroxymethylfurfural, A Versatile Platform Chemical Made from Renewable Resources. Chem. Rev. 2013, 113, 1499-1597. (2) Van Putten, R.-J.; Dias, A. S.; De Jong, E., Furan based building blocks from carbohydrates. In Catalytic Process Development for Renewable Materials, Imhof, P.; van der Waal, J. C., Eds. Wiley-VCH Verlag GmbH & Co.: 2013; pp 81-117. (3) Bozell, J. J.; Petersen, G. R., Technology development for the production of biobased products from biorefinery carbohydrates - The US Department of Energy's "top 10" revisited. Green Chem. 2010, 12, 539-554. (4) Corma, A.; Iborra, S.; Velty, A., Chemical Routes for the Conversion of Biomass into Chemicals. Chem. Rev. 2007, 107, 2411-2502. (5) Van der Waal, J. C.; Imhof, P., Catalytic Process Development for Renewable Materials. Wiley-VCH: Weinheim, 2013. (6) Verkerk, G.; Broens, J. B.; Kranendonk, W.; Van der Puijl, F. J.; Sikkema, J. L.; Stam, C. W., BINAS. 2nd ed.; Wolters-Noordhoff: Groningen, 1986. (7) Van der Waal, J. C.; De Jong, E., Chemocatalytic Processes for the Production of Bio-Based Chemicals from Carbohydrates. In Producing Fuels and Fine Chemicals from Biomass Using Nanomaterials, Luque, R.; Balu, A. M., Eds. CRC Press: Boca Raton, 2013; pp 223-265. (8) Pye, E. K.; Rushton, M., Organosolv Biorefining: Creating Higher Value from Biomass. In Catalytic Process Development for Renewable Materials, Imhof, P.; Van der Waal, J. C., Eds. Wiley-VCH: Weinheim, 2013; pp 239-263. (9) Bicker, M.; Kaiser, D.; Ott, L.; Vogel, H., Dehydration of D-fructose to hydroxymethylfurfural in sub- and supercritical fluids. J. Supercrit. Fluids 2005, 36, 118-126. (10) Ishikawa, Y.; Saka, S., Chemical conversion of cellulose as treated in supercritical methanol. Cellulose 2001, 8, 189-195. (11) Minami, E.; Saka, S., Comparison of the decomposition behaviors of hardwood and softwood in supercritical methanol. J. Wood Sci. 2003, 49, 73-78. (12) Gruter, G. J. M.; Dautzenberg, F. Method for synthesis of 5-alkoxymethyl furfural ethers and their use. WO 2007104514, 2007. (13) Zhang, L.; Gong, X.; Wang, Y.; Qu, H., Solubilities of Protocatechuic Aldehyde, Caffeic Acid, d-Galactose, and d-Raffinose Pentahydrate in Ethanol-Water Solutions. J. Chem. Eng. Data 2012, 57, 2018-2022. (14) Gong, X.; Wang, C.; Zhang, L.; Qu, H., Solubility of Xylose, Mannose, Maltose Monohydrate, and Trehalose Dihydrate in Ethanol-Water Solutions. J. Chem. Eng. Data 2012, 57, 3264-3269. (15) Ferreira, O.; Brignole, E. A.; Macedo, E. A., Phase Equilibria in Sugar Solutions Using the A- UNIFAC Model. Ind. Eng. Chem. Res. 2003, 42, 6212-6222. (16) Peres, A. M.; Macedo, E. A., A modified UNIFAC model for the calculation of thermodynamic properties of aqueous and non-aqueous solutions containing sugars. Fluid Phase Equilib. 1997, 139, 47-74. (17) Macedo, E. A.; Peres, A. M., Thermodynamics of Ternary Mixtures Containing Sugars. SLE of d-Fructose in Pure and Mixed Solvents. Comparison between Modified UNIQUAC and Modified UNIFAC. Ind. Eng. Chem. Res. 2001, 40, 4633-4640. (18) Bouchard, A.; Hofland, G. W.; Witkamp, G.-J., Properties of Sugar, Polyol, and Polysaccharide Water-Ethanol Solutions. J. Chem. Eng. Data 2007, 52, 1838-1842. (19) Flood, A. E.; Puagsa, S., Refractive Index, Viscosity, and Solubility at 30 °C, and Density at 25 °C for the System Fructose + Glucose + Ethanol + Water. J. Chem. Eng. Data 2000, 45, 902-907. Experimental and modelling studies on the solubility of D-arabinose, D-fructose, D-glucose, |357 D-mannose, sucrose and D-xylose in methanol and methanol-water mixtures
(20) Peres, A. M.; Macedo, E. A., Measurement and Modeling of Solubilities of d-Glucose in Water/Alcohol and Alcohol/Alcohol Systems. Ind. Eng. Chem. Res. 1997, 36, 2816-2820. (21) Peres, A. M.; Macedo, E. A., Phase equilibria of d-glucose and sucrose in mixed solvent mixtures: Comparison of UNIQUAC 1-based models. Carbohydr. Res. 1997, 303, 135-151. (22) Montañés, F.; Olano, A.; Ibáñez, E.; Fornari, T., Modeling solubilities of sugars in alcohols based on original experimental data. AIChE J. 2007, 53, 2411-2418. (23) Allavudeen, S. S.; Kuberan, B.; Loganathan, D., A method for obtaining equilibrium tautomeric mixtures of reducing sugars via glycosylamines using nonaqueous media. Carbohydr. Res. 2002, 337, 965-968. (24) De Wit, G. Gedrag van Glucose, Fructose en Verwante Suikers in Alkalisch Milieu. Ph.D. thesis, Faculty of Applied Sciences, Delft University of Technology, Delft, the Netherlands 1979. (25) Prausnitz, J. M.; Lichtenthaler, R. N.; Gomes de Azevedo, E., Molecular Thermodynamics of Fluid-Phase Equilibria. 2 ed.; Prentice Hall Inc.: Englewood Cliffs, N.J., 1986. (26) Abrams, D. S.; Prausnitz, J. M., Statistical thermodynamics of liquid mixtures: a new expression of the excess Gibbs energy of partly or completely miscible systems. AIChE J. 1975, 21, 116-128. (27) Poling, B. E.; Prausnitz, J. M.; O’Connell, J. P., The Properties of Gases and Liquids. 5 ed.; McGraw-Hill: 2004. (28) Roos, Y., Melting and Glass Transitions of Low Molecular Weight Carbohydrates. Carbohydr. Res. 1993, 238, 39-48. (29) Dean, J. A., Langes Handbook of Chemistry. 13 ed.; McGraw-Hill: New York, 1985. (30) Flood, A. E. In Measurement and modeling of solid - liquid equilibrum in the system fructose + glucose + ethanol + water, Regional Symposium on Chemical Engineering (RSCE2000), Singapore, December 11-13, 2000; Singapore, 2000. (31) Raemy, A.; Schweizer, T. F., Thermal Behavior of Carbohydrates Studied by Flow Calorimetry. J. Therm. Anal. 1983, 28, 95. (32) Parks, G. S.; Thomas, S. B., The heat capacities of crystalline, glassy and undercooled liquid glucose. J. Am. Chem. Soc. 1934, 56, 1423. (33) Miller, D. P.; Pablo, J. J., Calorimetric Solution Properties of Simple Saccharides and Their Significance for the Stabilization of Biological Structure and Function. J. Phys. Chem. B 2000, 104, 8876.
358| Chapter 6
|359
7 Concluding remarks and recommendations
The objectives of the research presented in this thesis were to obtain new insights in the reaction mechanism of the acid-catalysed sugar dehydration to HMF and to identify the most suitable hexose from an HMF selectivity and activity point of view. The research showed that structural differences and particularly the stereochemistry of the C3 and C4 carbons of the ketoses play an intricate role in the rate of sugar conversion, as well as the selectivity and the rate of HMF formation (Chapters 3-5). It was shown that the relative orientation of the vicinal hydroxyl groups on C3 and C4 determines the conversion rate of the sugars, whereas the orientation of the hydroxyl group on C4 determines the extent of formation of a by- product, 2-hydroxyacetylfuran (HAF). Psicose and tagatose have both hydroxyl groups on C3 and C4 oriented on the same side (cis configuration) of the furanose ring (Scheme 1). Apparently this is a favourable configuration as it leads to higher sugar conversion rates than for fructose and sorbose, for which the two hydroxyl groups are on the opposite side (trans configuration) of the furanose ring (Scheme 1). These results are in agreement with a rate- determining step with cyclic intermediates, which does not allow free rotation of the C3-C4 bond and thus fixes the relative hydroxyl group orientation on these carbons. The involvement of acyclic intermediates in the rate-determining step is less likely as this would allow free rotation of the C3-C4 bond, which in turn would reduce the effect of the hydroxyl group orientation at these positions.
Scheme 1. The four 2-ketohexoses in their α-D-furanose form
The results presented in Chapter 3 indicate that the reactivity and selectivity towards HMF formation is much lower for the aldoses than for the ketoses. In addition, it was shown 360| Chapter 7 that the orientation of the hydroxyl groups in the aldoses (glucose, mannose and galactose) does not lead to significant differences in reactivity and selectivity to HMF. These findings are in contrast with those for the ketoses and imply that ketoses and aldoses have different rate-determining steps in the acid-catalysed dehydration of sugars to HMF. A possibility is an initial isomerisation of the aldose to the corresponding ketose prior to the dehydration reaction to HMF. In Chapter 4 and Chapter 5 the orientation of the hydroxyl group on C4 was proven to determine the extent of formation of HAF from ketoses. HAF was observed in significant amounts for sorbose and tagatose and not for fructose and psicose. Dehydration experiments with 13C-labelled sorbose disproved a mechanism via acyclic intermediates as proposed in the literature, presumably involving a 1,4-anhydroketose intermediate. This led to the proposition of four potential reaction mechanisms to form HAF. Further research is required in order to draw more detailed conclusions on the reaction pathway for HAF. The acid-catalysed dehydration reactions of ketoses were performed both in water and methanol. The sugars are by far more reactive in methanol than in water. These results indicate that methanol is a more suitable solvent for the production of HMF, in the form of its methyl ether, methoxymethylfurfural (MMF). This is best illustrated by the results obtained with fructose, which yielded at best around 45% of HMF and MMF combined in methanol, and around 40% HMF in water. In general, comparable trends regarding the relative reactivity of the four ketoses were observed in water and methanol. Different acid concentrations were tested for the ketose dehydration reactions in both solvents. Each ketose followed a consistent yield vs. conversion trend for all products, indicating that the acid concentration has no significant effect on the selectivity. In water also different temperatures were tested (Chapter 3), which likewise has no significant effect on the selectivity. These observations indicate that the maximum obtainable HMF/MMF yield for each ketose is solvent dependent and essentially independent of the catalyst concentration and temperature within the experimental window. The maximum obtainable combined yield of HMF and MMF in methanol is around 55%, using psicose as the substrate (Chapter 4). Thus, both from a selectivity and activity point of view, psicose is the preferred feedstock for the production of HMF and its alkyl ethers. As shown in Chapters 3-5, it is beneficial to perform catalytic sugar reactions in alcohols instead of water. Detailed knowledge regarding the solubility of the sugars in alcohols is lacking, though of high importance for the design of an efficient process in methanol. As such, a detailed solubility study of six sugars (glucose, fructose, mannose, Concluding remarks and recommendations |361 xylose, arabinose and sucrose) was performed in methanol and methanol-water mixtures (up to 25% water). The results were successfully modelled with a UNIQUAC model. The solubilities of the sugars in methanol differ considerably. Fructose is clearly the most soluble sugar and sucrose is clearly the least soluble. In patent PCT/NL2011/050907 it was shown that much higher sugar concentrations in methanol (>25%) are possible by converting the sugars into methyl glycosides at relatively low temperatures (<80 °C). Future work should elaborate on the solubility of mixtures of fructose and alkyl fructosides in pure methanol and methanol-based solvent mixtures. The results presented in this thesis illustrate that the use of methanol as the solvent, instead of water, is favoured for a number of reasons. The selectivity for the formation of the desired product HMF/MMF is higher than in water. In addition, the amount of insoluble humins formed is much less in methanol than in water. This has a positive effect on the variable costs of the process while simultaneously reducing potential process issues such as reactor plugging. Methoxymethylfurfural (MMF), the major product in methanol, is thermally more stable than HMF and has a much lower boiling point, making it much easier to distil off without further loss of product by thermal decomposition. The latter is especially important, as a prominent issue in the development of HMF processes in the past 30 years has been the efficient separation of HMF from the reaction mixture. This also touches on another important point, namely that an economically viable process will likely have to be performed at high substrate concentrations as well as high conversion levels, because MMF has a higher boiling point than methanol and water, requiring extensive solvent recycling when operating the process at low substrate concentration and/or at low conversion. On the basis of this study, yields of 50-55% MMF and 25% ML at >95% sugar conversion is possible using psicose as the substrate, not taking into account losses in the work-up section. However, psicose is a very rare and expensive sugar, making fructose the preferred substrate for the imminent future. In conclusion, novel information on hexose structure-reactivity relations regarding HMF and its MMF synthesis was obtained. Strong evidence has been generated that the dehydration of ketoses proceeds via cyclic intermediates. It was also found that psicose is the most suitable substrate for HMF/MMF synthesis from a selectivity and activity point of view. Finally, methanol appears to be a more suitable solvent for HMF/MMF formation than water.
361
362| Chapter 7
|363
Summary
Fossil resources play a key role in energy generation, the transportation sector and for the production of materials, food and pharmaceuticals and this has been the case for more or less the past 100 years. The environmental footprint of an economy based totally on fossil resources is high and the development of alternative, sustainable sources and processes for energy and materials production is of prime importance. Alternative energy sources are already used (solar, wind and water power), although significant improvements are still needed. The production of materials generally requires sources of organic carbon. The only source of sustainable organic carbon is biomass and it is for this reason that biobased chemicals need to be developed, especially for bulk applications. When sugars are heated in the presence of Brønsted acids, 5-hydroxymethylfurfural (HMF) is formed. HMF is considered a ‘sleeping giant’, as it may be converted into a number of high-potential building blocks, especially for polymers. The synthesis of HMF presents challenges in terms of yield and work-up/purification. In order to overcome these challenges, more knowledge is required about the actual dehydration chemistry of sugars to HMF. The research discussed in the present thesis entitled ‘Experimental and modelling studies on the synthesis of 5-hydroxymethylfurfural from sugars’ was performed with two main objectives in mind: (i) to gain new insights in the reaction pathways of the dehydration of ketohexoses to HMF and (ii) to identify the most suitable hexose for HMF production from a selectivity and activity point of view. The state of the art on the dehydration of sugars to HMF and the conversion of HMF into building blocks for chemicals, materials and fuels are reviewed In Chapter 2. It shows that the vast majority of the research performed on the synthesis of HMF so far is done on small scale in a research laboratory setting. Initially mainly fructose and glucose were used as substrates, but recent years have seen an increase in the use of polysaccharides. The state of the art clearly shows the effects of solvents on HMF yields and underpins the importance of solvent selection. For the acid-catalysed fructose dehydration in water, the maximum yield of HMF is around 50%, a value that can be increased to around 90% by using ionic liquids or polar aprotic solvents. For glucose the yields are typically below 10%. Yields from glucose
364| can be improved by using bifunctional catalysts, which presumably involves an initial isomerisation of glucose to fructose. The vast majority of the research so far has focused on yield optimisation by (systematic) catalyst screening studies. Mechanistic studies to intrinsically understand the chemistry and to postulate reaction mechanisms are virtually absent. In general, the highest HMF yields and selectivities have been obtained in exotic reaction systems that are difficult to translate and scale up to economically viable and industrially relevant processes. In Chapter 3 an experimental and modelling study is reported on the reactivity of different hexoses in water using sulphuric acid as the catalyst. Three aldoses (glucose, mannose and galactose) and three ketoses (fructose, sorbose and tagatose) were tested using high-throughput experimentation and supported by DFT calculations. The aldoses are less active and less selective for HMF formation than the ketoses. Major differences in reactivity were not observed for the three aldoses. On the contrary, the ketoses show significant differences in reactivity and selectivity to HMF, explained by considering the hydroxyl group orientations on the C3 and C4 positions for the various ketoses. These findings point to a reaction mechanism involving cyclic intermediates. Experimental studies on the use of methanol instead of water as the reaction medium for the conversions of ketoses to HMF are described in Chapter 4. The ketoses are much more reactive in methanol than in water. In addition, HMF was formed in minor amounts as it is etherified to its methyl ether, 5-methoxymethylfurfural (MMF). Fructose, sorbose, tagatose and psicose again show significant differences in reactivity and selectivity to furanics. Psicose and tagatose show the highest conversion rates, whereas fructose and psicose show the highest selectivities for HMF and MMF formation. The highest maximum MMF yield was found for psicose, at around 55%. The formation of significant amounts of an HMF isomer, 2-hydroxyacetylfuran (HAF), was observed for sorbose and tagatose. This study supports the claim (Chapter 3) that the stereochemistry of the OH groups at C3 and C4 plays a key role in the reaction network, supporting a reaction mechanism with cyclic intermediates in the rate-determining step. HAF formation was studied in more detail by performing experiments with 13C-labelled sorbose and subsequent analyses of the reaction mixtures using 13C-NMR. The results imply that HAF formation proceeds via a 1,4-anhydroketose. Experimental studies on the reactivity of psicose in acidic water are discussed in Chapter 5 and the results are compared to fructose, sorbose and tagatose. The results are consistent with those in Chapters 3 and 4, pointing to the importance of the orientation of the hydroxyl groups on C3 and C4 regarding the reaction rate and selectivity to HMF. The |365 formation of HAF was only observed for tagatose and sorbose, indicating a key role of the stereochemistry of the hydroxyl group on C4 in the reaction pathway to HAF. As shown in Chapter 4, it is beneficial to perform catalytic sugar conversions in alcohols instead of water. Since the solubility of sugars in alcohols may differ considerably from that in water, the solubility of six sugars (glucose, fructose, mannose, xylose, arabinose and sucrose) in methanol and methanol-water mixtures has been determined experimentally and the results are described in Chapter 6. The results were successfully modelled with a UNIQUAC model and may be used for further scale-up studies.
366|
|367
Samenvatting
Fossiele grondstoffen spelen al ongeveer 100 jaar een sleutelrol in onze maatschappij als de voornaamste bron voor de productie van energie en materialen. Een economie die afhankelijk is van fossiele bronnen heeft een zeer negatief effect op het milieu en om die reden is het van het grootste belang om duurzame alternatieven te vinden voor de productie van energie en materialen. Duurzame alternatieven voor energieopwekking zijn al in gebruik (waterkracht, zonne- en windenergie), maar vereisen nog significante verbeteringen. De productie van materialen vereist vaak organische koolstof. De enige duurzame bron van organische koolstof is biomassa en daarom is het noodzakelijk om grondstoffen voor met name de bulkchemie te produceren van biomassa. Bij het verwarmen van suikers in de aanwezigheid van Brønsted-zuren wordt 5- hydroxymethylfurfural (HMF) gevormd. HMF wordt gezien als een ‘slapende reus’, aangezien het kan worden omgezet in een aantal potentiëel zeer belangrijke bouwstenen, met name voor polymeren. De synthese van HMF is uitdagend in termen van opbrengst en opwerking. Om deze uitdagingen te overwinnen is meer kennis vereist van de dehydratatiereactie van suikers naar HMF. De twee voornaamste doelstellingen van het onderzoek beschreven in dit proefschrift, genaamd ‘Experimental and modelling studies on the synthesis of 5-hydroxymethylfurfural from sugars’, zijn: (i) het verwerven van meer inzicht in de reactiepaden van de dehydratatie van monomere C6 suikers (hexoses) naar HMF, in het bijzonder vanuit fructose en andere ketohexoses en (ii) het identificeren van de meest geschikte suiker voor HMF productie met het oog op selectiviteit en activiteit. Hoofstuk 2 geeft een overzicht van de huidige stand van zaken omtrent de dehydratatie van suikers naar HMF en de conversie van HMF naar bouwstenen voor chemicaliën, materialen en brandstoffen. Dit laat zien dat de overgrote meerderheid van het onderzoek naar HMF-synthese plaatsvindt op kleine schaal. Tot een paar jaar geleden werden voornamelijk fructose en glucose toegepast als substraat, maar tegenwoordig wordt er in toenemende mate onderzoek gedaan naar polysaccharides. De stand van zaken laat duidelijk de invloed van het oplosmiddel op de dehydratatiereactie zien en het benadrukt het belang van de oplosmiddelkeuze. In water is de maximale opbrengst van HMF uit fructose ongeveer
368|
50%, maar de opbrengst kan worden verhoogd tot ongeveer 90% door ionogene of aprotische polaire oplosmiddelen te gebruiken. De HMF-opbrengst uit glucose ligt normaal gesproken onder de 10% en kan worden verbeterd door bifunctionele katalysatorsystemen toe te passen, waarbij wordt aangenomen dat glucose initiëel wordt geïsomeriseerd naar fructose. Tot dusver is de overgrote meerderheid van het onderzoek aan HMF-synthese gefocust op het verhogen van de opbrengst door (systematisch) katalysatoren te testen. Fundamenteel onderzoek naar de dehydratatiereactie van suikers om informatie over het reactiemechanisme te verkrijgen wordt nauwelijks gedaan. In het algemeen worden de hoogste opbrengsten en selectiviteiten voor HMF behaald met exotische reactiesystemen die heel moeilijk te vertalen zijn naar economisch levensvatbare en industriëel relevante processen. Een studie naar de reactiviteit van verschillende hexoses in water in de aanwezigheid van zwavelzuur is beschreven in Hoofdstuk 3. Hierbij is zowel experimenteel werk als modellering (DFT) gedaan. Drie aldoses (glucose, mannose en galactose) en drie ketoses (fructose, sorbose en tagatose) werden getest volgens een ‘high-throughput’ methode en de resultaten in overeenstemming waren met de DFT-berekeningen. De aldoses zijn duidelijk minder reactief en minder selectief naar de vorming van HMF dan de ketoses. Tussen de aldoses onderling werden geen significante verschillen in reactiviteit waargenomen. Daarentegen blijken er wel duidelijke verschillen in reactiviteit en selectiviteit tussen de ketoses te bestaan. Dit kan worden verklaard door de verschillen in hydroxylgroep-oriëntatie op de C3 en C4 posities voor de verschillende ketoses. Deze bevindingen wijzen op een reactiemechanisme met cyclische intermediairen. In Hoofdstuk 4 zijn experimentele studies naar het gebruik van methanol in plaats van water als reactiemedium voor ketosedehydratatie naar HMF beschreven. De ketoses zijn aanzienlijk reactiever in methanol dan in water. Verder werd de opbrengst van HMF beperkt door de vorming van de veel stabielere en vluchtigere methyl ether, 5-hydroxymethylfurfural (MMF). Fructose, sorbose, tagatose en psicose laten wederom significante verschillen zien in reactiviteit en selectiviteit naar furanen. Psicose en tagatose worden het snelst omgezet, terwijl psicose en fructose de hoogste selectiviteit hebben voor de vorming van HMF en MMF. De hoogste maximale MMF opbrengst werd behaald met psicose, ongeveer 55%. In het geval van sorbose en tagatose werd de vorming van significante hoeveelheden van een isomeer van HMF, namelijk 2-hydroxyacetylfuran (HAF) geobserveerd. Deze studie bevestigt de stelling dat de stereochemie van de hydroxylgroepen op C3 en C4 van de ketoses een doorslaggevende rol speelt in het reactienetwerk, wat een reactiemechanisme met cyclische intermediairen in de snelheidsbepalende stap ondersteunt. De vorming van HAF |369 werd in meer detail bestudeerd door experimenten uit te voeren met 13C-gelabeld sorbose en deze vervolgens te bestuderen met 13C-NMR. De resultaten impliceren dat HAF-vorming plaatsvindt via een 1,4-anhydroketose. De reactiviteit van psicose in zuur water wordt besproken in Hoofdstuk 5, waarbij de resultaten worden vergeleken met die van fructose, sorbose en tagatose. De resultaten zijn in overeenstemming met de bevindingen in Hoofdstuk 4 en 5, wijzend op het belang van de oriëntatie van de hydroxylgroepen op C3 en C4 aangaande de reactiviteit en selectiviteit naar HMF. De vorming van HAF werd uitsluitend geobserveerd bij sorbose en tagatose, wijzend op een sleutelrol voor de stereochemie van de hydroxylgroep op C4 in het reactiemechanisme voor HAF-vorming. Hoofdstuk 4 laat zien dat het voordelig kan zijn om katalytische suikerconversies uit te voeren in alcohol in plaats van water. Aangezien de oplosbaarheid van suikers in alcoholen behoorlijk kan afwijken van die in water, werd de oplosbaarheid van zes suikers (glucose, fructose, mannose, xylose, arabinose en sucrose) in methanol en methanol-watermengsels experimenteel bepaald. Dit wordt beschreven in Hoofdstuk 6. De resultaten werden succesvol gemodelleerd met een UNIQUAC-model en kunnen worden gebruikt voor verdere opschalingsstudies.
370|
|371
Acknowledgements
I have now arrived at the end of my PhD thesis and it is time to thank all the people that supported me on this journey. It is clear that all this would not have been possible without their assistance/support/guidance.
When I started my first real job at Avantium in March of 2008, my scientific curiosity was triggered and I quickly realised that I wanted to pursue a career in research. For the first time I was absolutely certain that I wanted to pursue a PhD degree. Ed, you were my supervisor at the time, and I cannot thank you enough for your positive response and support to start this project. Without it, it would probably not have happened. I realise that this was a great opportunity for me to combine my PhD research with obtaining working experience in a company. For this reason I am very thankful to Ed and the Avantium management team; Tom van Aken, Gert-Jan Gruter and Frank Roerink.
I am grateful to my promoter, Erik Heeres, for taking me on as an external PhD student. Erik, for me your positivity and enthusiasm were contagious. The cooperation between the Department of Chemical Engineering at the University of Groningen and Avantium was very fruitful. Jos Winkelman, C.B. Rasrendra, Jenny Soetedjo, prof. Kamerling and Henk van de Bovenkamp provided key contributions in discussions and publications. Henk, I additionally thank you for helping me out with lots of little things.
Apart from support in the way of resources I also had great scientific support. Jan Kees, Ed and Erik have challenged me and triggered my scientific curiosity. This was effective, great fun and kept me motivated during difficult periods in my research, especially at the beginning. Jan Kees, you have taught me to look at data in a different way, a key skill required to fill this thesis.
Avantium is a great company to work for. Therefore I would like to thank my colleagues for providing a positive and challenging workplace and a great atmosphere. I especially learned a lot from Jan Kees, Gert-Jan, Ed, Etienne, Ben and Erik-Jan. Farhad helped me out 372| on a number of occasions, consistently producing excellent quality work. The pubquiz team provided welcome distraction to compensate for the busy times.
I also want to thank the direct supervisors I have had during the years, especially Ana and Gert-Jan, for making sure that I had time to work on my PhD research and for providing the essential support and resources (analytics, equipment, etc.).
The large amount of experiments presented in this thesis clearly required a lot of chromatography. I was fortunate to have very strong analytical support from the Avantium analytical team, especially Jurriaan, Suzanne, Martin and Angela.
The publications/chapters I have written would not have been possible without the input of the co-authors. The Chemical Reviews publication (Chapter 2) would have been impossible without C.B., Erik, Ed, Jan Kees and especially Hans de Vries. Chapters 3 and 6 likewise required key contributions of the co-authors. Evgeny, thank you for your scientific input and for the entertaining discussions about sugar and HMF chemistry.
I would also like to thank the members of the reading committee: Prof. Centi, Prof. Palkovits and Prof. Broekhuis, for taking the time to read and judge this thesis.
In the second half of my PhD, the majority of the work was carried out with financial support from the European Community under the seventh framework programme (Project SPLASH, contractnr. 311956). This was essential in completing this work and it proved to be an excellent tool for companies to venture into more fundamental research and use this to train their people.
Most importantly of course my thanks go out to family and friends. Unfortunately I cannot share this achievement with my father, since he passed away shortly after I got the green light to start my PhD research. He taught me to form my own opinion and his support helped me overcome a number of hurdles earlier in my life. We received warm support from great friends and family. Felix, Sven, Ria, Berna and Hans, I cannot thank you enough for supporting us during these difficult times.
Obviously there were also plenty of good times to compensate for the long evenings and weekends with deadline stress, thanks to Sven, Christine, Mick and Aziz.
Finally I want to thank my mother and my sister Elise for their loving support; ik hou van jullie! |373
Publications
Papers
‘Hydroxymethylfurfural, a versatile platform chemical made from renewable resources’ (Chapter 2) Van Putten, R.-J.; Van der Waal, J. C.; De Jong, E.; Rasrendra, C. B.; Heeres, H. J.; De Vries, J. G. Chem. Rev. 2013, 113, 1499-1597.
‘The dehydration of different ketoses and aldoses to 5-hydroxymethylfurfural’ (Chapter 3) Van Putten, R.-J.; Soetedjo, J. N. M.; Pidko, E. A.; Van der Waal, J. C.; Hensen, E. J. M.; De Jong, E.; Heeres, H. J. ChemSusChem 2013, 6, 1681-1687.
‘A comparative study on the reactivity of various ketohexoses to furanics in methanol’ (Chapter 4) Van Putten, R.-J.; Van de Bovenkamp, H. H.; Van der Waal, J. C.; De Jong, E.; Heeres, H. J. 2014. Manuscript in preparation
‘Reactivity studies on the acid-catalysed dehydration of 2-ketohexoses to 5- hydroxymethylfurfural in water’ (Chapter 5) Van Putten, R.-J.; Van der Waal, J. C.; De Jong, E.; Heeres, H. J. 2014. Manuscript in preparation
‘Experimental and modelling studies on the solubility of D-arabinose, D-fructose, D-glucose, D- mannose, sucrose and D-xylose in methanol and methanol-water mixtures’ (Chapter 6) Van Putten, R.-J.; Winkelman, J. G. M.; Keihan, F.; Van der Waal, J. C.; De Jong, E.; Heeres, H. J. Ind. Eng. Chem. Res. 2014, 53, 8285-8290.
Presentation
‘Different ketoses in HMF synthesis’ Van Putten, R.-J.; Van der Waal, J. C.; De Jong, E.; Pidko, E.; Heeres, H. J. Netherlands Catalysis and Chemistry Conference (NCCC), 2013, Noordwijkerhout, the Netherlands (Oral presentation)
|374