University of Groningen. Levulinic acid from lignocellulosic biomass Girisuta, Buana

Similar documents
University of Groningen. Levulinic acid from lignocellulosic biomass Girisuta, Buana

Production of 5-Hydroxymethylfurfural from Glucose Using a Combination of Lewis and

Study on the Production of Gamma valerolactone from Hybrid Poplar

Carbon dioxide removal processes by alkanolamines in aqueous organic solvents Hamborg, Espen Steinseth

CATALYSTS FOR SELECTIVE CONVERSION OF PLANT CELL WALL POLYSACCHARIDES

Plug Flow Reactor Model to Analyse Operating Conditions on the Dilute H 2 SO 4 -acid Hydrolysis of Sugarcane Bagasse at High-Solids Loading

Green Chemicals from d-glucose Rasrendra, C. B.; Makertihartha, I. G. B. N.; Adisasmito, S.; Heeres, Hero

Chapter 6 Experimental and Kinetic Modelling Studies on the Acid-Catalysed Hydrolysis of the Water Hyacinth Plant to Levulinic Acid

Tetrahydrofuran (THF) is a versatile chemical that has commercial application as a solvent for

University of Groningen. Catalytic inulin conversions to biobased chemicals Boy Arief Fachri, B

Citation for published version (APA): Nouri-Nigjeh, E. (2011). Electrochemistry in the mimicry of oxidative drug metabolism Groningen: s.n.

Estimation of Temperature Transients for Biomass Pretreatment in Tubular Batch Reactors and Impact on Xylan Hydrolysis Kinetics

Kinetics Studies of Xylan and Acetyl- Group Hydrolysis

University of Groningen

University of Groningen

University of Groningen. Enantioselective liquid-liquid extraction in microreactors Susanti, Susanti

Production of Levulinic Acid from Pennisetum alopecuroides in the Presence of an Acid Catalyst

Carbohydrates - Introduction

Carbon dioxide removal processes by alkanolamines in aqueous organic solvents Hamborg, Espen Steinseth

Kinetic Analysis of Partial Decomposition of Starch in Hot-Compressed Water

Theoretical simulation of nonlinear spectroscopy in the liquid phase La Cour Jansen, Thomas

Personalised Learning Checklists AQA Chemistry Paper 2

Published in: ELECTRONIC PROPERTIES OF NOVEL MATERIALS - PROGRESS IN MOLECULAR NANOSTRUCTURES

LAP-013. Procedure Title: Date: 6/28/96. Author: Raymond Ruiz and Tina Ehrman ISSUE DATE: 8/26/96 SUPERSEDES: 11/01/94.

Chemical Reactions and Chemical Reactors

Bimetallic ruthenium-copper nanoparticles embedded in. mesoporous carbon as an effective hydrogenation catalyst

Transformation of lignocellulose into aromatic building blocks

Citation for published version (APA): Sok, R. M. (1994). Permeation of small molecules across a polymer membrane: a computer simulation study s.n.

University of Groningen. Levulinic acid from lignocellulosic biomass Girisuta, Buana

University of Groningen

Alcohols. Ethanol Production. 182 minutes. 181 marks. Page 1 of 25

Investigation of the Mechanism of Acidic Hydrolysis of Cellulose

Biofuels Series. Application Note 301LS RAPID MEASUREMENT OF XYLOSE AND GLUCOSE MONITORING CORN STOVER FERMENTATION IN BIOETHANOL PRODUCTION

Effect of phenol concentrations on the condensation reaction during the liquefaction of waste woody materials with phenol

Experiment 2 Solvent-free Aldol Condensation between 3,4-dimethoxybenzaldehyde and 1-indanone

Catalytic Conversion of Cellulose to Levulinic Acid by Metal Chlorides

Hydrogen Sulfide Removal from Waste Air by Oxidation reaction with Sodium Hypochlorite in CSTR

2005 Elsevier : Reproduced in accordance with the copyright policy of the publisher : This journal is available online - use hypertext links.

SUPPLEMENTARY INFORMATION

PART 2 Dr. Stephen Glover

Valorisation of Lignocellulosic Waste and the Potential for Supercritical Water and Wet Air Oxidation Part 1

University of Groningen. Laser Spectroscopy of Trapped Ra+ Ion Versolato, Oscar Oreste

GCSE Chemistry. Module C7 Further Chemistry: What you should know. Name: Science Group: Teacher:

Le Lycee Mauricien. Proposed Syllabus Chemistry (5070) - Form 5

Synthesis of jet fuel range cycloalkanes with diacetone alcohol. from lignocellulose

University of Groningen. Statistical Auditing and the AOQL-method Talens, Erik

SUPPLEMENTARY TOPIC 3 ENERGY AND CHEMICAL REACTIONS

2/18/2013 CHEMISTRY OF CELLS. Carbon Structural Formations. 4 Classes of Organic Compounds (biomolecules)

UvA-DARE (Digital Academic Repository) Converting lignin to aromatics: step by step Strassberger, Z.I. Link to publication

Oxidation of Phenolic Wastewater by Fenton's Reagent

GCSE CHEMISTRY REVISION LIST

The Chemistry and Energy of Life

A New Oil Spill Adsorbent from Sulfuric Acid Modified Wheat Straw

Importance of Carbohydrates

Extraction of spruce hemicelluloses. J. V. Rissanen, H. Grénman, D. Yu. Murzin, T. Salmi

ENZYME SCIENCE AND ENGINEERING PROF. SUBHASH CHAND DEPARTMENT OF BIOCHEMICAL ENGINEERING AND BIOTECHNOLOGY IIT DELHI LECTURE 3

Paper Atomic structure and the periodic table

Contents. Efficient synthesis of 5-(chloromethyl)furfural (CMF) from high fructose corn syrup (HFCS) using continuous flow processing

University of Groningen. Extraction and transport of ion beams from an ECR ion source Saminathan, Suresh

Lecture (9) Reactor Sizing. Figure (1). Information needed to predict what a reactor can do.

Methods of purification

KINETICS OF ESTERIFICATION OF CITRIC ACID WITH BUTANOL USING SULPHURIC ACID

Catalytic One-Pot Production of Small Organics from Polysaccharides

Decomposition of lignin alkaline and chemicals recovery in sub- and supercritical water

Review Topic 8: Phases of Matter and Mixtures

Characterization of liquefied products from model woody components in the presence of mineral acid catalysts

3.2.2 Kinetics. Maxwell Boltzmann distribution. 128 minutes. 128 marks. Page 1 of 16

SAMPLE CHEMISTRY QUESTIONS MIXTURE OF UNIT 3 & 4 MATERIALS

FLUDEOXYGLUCOSE ( 18 F) INJECTION: Final text for addition to The International Pharmacopoeia (January 2009)

Kinetic investigations on the esterification of phthalic anhydride with n-heptyl, n-nonyl or n-undecyl alcohol over sulfuric acid catalyst

Chemical Reactions and Kinetics of the Carbon Monoxide Coupling in the Presence of Hydrogen

BIOLOGY LTF DIAGNOSTIC TEST CHEMISTRY OF LIFE

Apparent Melting: A New Approach to Detecting Drug-Excipient Incompatibility

PAPER CHEMISTRY. APPLETON. WISCONSIN IPC TECHNICAL PAPER SERIES NUMBER 42

Chemical Reaction Engineering

CHEMICAL KINETICS (RATES OF REACTION)

University of Groningen

Xylose Monomer and Oligomer Yields for Uncatalyzed Hydrolysis of Sugarcane Bagasse Hemicellulose at Varying Solids Concentration

3.5 Alcohols H H C C. N Goalby chemrevise.org 1 H H. Bond angles in Alcohols. Boiling points. Different types of alcohols H 2 C CH 2 CH 2

Determination of Reducing Sugar Content: Clinitest, Benedict s Solution and the Rebelein Titration

THE CHEMISTRY OF AUTOCATALYTIC PROCESSES IN THE CONTEXT OF THE SHROUD OF TURIN

Gravity is a force which keeps us stuck to the earth. The Electrostatic force attracts electrons to protons in an atom.

Cherry Hill Tuition A Level Chemistry OCR (A) Paper 9 THIS IS A NEW SPECIFICATION

Supplemental Information

, can be completely combusted to give carbon dioxide and water. (s) + 6O 2

AQA Chemistry (Combined Science) Specification Checklists. Name: Teacher:

University of Groningen. Morphological design of Discrete-Time Cellular Neural Networks Brugge, Mark Harm ter

Conversion of Pretreated Biomass into Levulinic Acid via Continuous Extraction at Atmosphere Pressure

Studies on the Kinetics of Heavy Oil Catalytic Pyrolysis

Chapter 9 The Chemical Reaction Equation and Stoichiometry 9.1 Stoichiometry

KINETIC MODELING OF HARDWOOD PREHYDROLYSIS. PART II. XYLAN REMOVAL BY DILUTE HYDROCHLORIC ACID PREHYDROLYSIS 1

Expt RO 1. Determination of Reaction Order (RO)

Substrate and Cation Binding Mechanism of Glutamate Transporter Homologs Jensen, Sonja

Methanolysis of Poly(ethylene terephthalate) in Supercritical Phase

2017 VCE Chemistry (NHT) examination report

1. Introduction. Olatunde Jogunola 1,*, Tapio Salmi 1, Johan Wärnå 1, Sébastien Leveneur 1,2, Jyri-Pekka Mikkola 1,3

Development of a continuous process for EPDM devulcanization in an extruder Sutanto, Poppy

CHEMICAL ENGINEERING LABORATORY CHEG 4137W/4139W. Reaction Kinetics Saponification of Isopropyl Acetate with Sodium Hydroxide

Chapter 11 Reactions of Alcohols. Types of Alcohol Reactions

Tuning Porosity and Activity of Microporous Polymer Network Organocatalysts by Co-Polymerisation

Transcription:

University of Groningen Levulinic acid from lignocellulosic biomass Girisuta, Buana IMPRTANT NTE: 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: 2007 Link to publication in University of Groningen/UMG research database itation for published version (APA): Girisuta, B. (2007). Levulinic acid from lignocellulosic biomass s.n. opyright ther 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 reative ommons). 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/UMG 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: 16-11-2018

hapter 4 A Kinetic Study on the onversion of Glucose to Levulinic Acid * Abstract Levulinic acid has been identified as a promising biomass-derived platform chemical. A kinetic study on one of the key steps in the conversion of biomass to levulinic acid, i.e., the acid-catalysed decomposition of glucose to levulinic acid has been performed. The experiments were carried out in a broad temperature window (140 200 ), using sulphuric acid as the catalyst (0.05 1 M) and an initial glucose concentration between 0.1 and 1 M. A kinetic model of the reaction sequence was developed including the kinetics for the intermediate 5- hydroxymethyl-2-furaldehyde (MF) and humins by-products using a power-law approach. The yield of levulinic acid is favoured in dilute glucose solution at high acid concentration. n the basis of the kinetic results, continuous reactor configurations with a high extent of back-mixing are preferred to achieve high levulinic acid yields. Keywords: biomass; green chemistry; levulinic acid; kinetic studies; reactor configurations. * Based on: Girisuta, B.; Janssen, L. P. B. M.; eeres,. J., A kinetic study on the conversion of glucose to levulinic acid. hem. Eng. Res. Des. 2006, 84, 339 349.

hapter 4 4.1 Introduction A substantial amount of research activities is currently undertaken worldwide to identify attractive chemical transformations to convert biomass into organic (bulk) chemicals and to develop economically feasible processes for these transformations on a commercial scale. ur research activities involve the acidcatalysed decomposition of lignocellulosic biomass into valuable chemicals. An attractive option is the conversion of lignocellulosic biomass into levulinic acid (4- oxopentanoic acid) by acid treatment at relatively mild conditions. Levulinic acid contains a ketone group and a carboxylic acid group. These two functional groups make levulinic acid a potentially versatile building block for the synthesis of various organic (bulk) chemicals as shown in Figure 4.1 [1-5]. For instance, 2- methyltetrahydrofuran and various levulinate esters may be used as gasoline and biodiesel additives, respectively. δ-aminolevulinate is a known herbicide, and the bisphenol derivative may be an interesting substitute for bisphenol A [6,7]. 3 3 2 2-Methyltetrahydrofuran γ-valerolactone Acrylic acid 3 α-angelica lactone 3 Levulinic acid 3 1,4-Pentanediol 3 R 3 Levulinate esters β-acetylacrylic acid 2 N δ-amino levulinic acid 3 4,4-Bis-(4-hydroxyphenyl) valeric acid Figure 4.1 Potentially interesting derivatives of levulinic acid. n a molecular level, the conversion of lignocellulosic biomass to levulinic acid is known to follow a complicated reaction scheme involving several intermediates and by-products (Figure 4.2) [8,9]. emicellulose and cellulose, two of the three main constituents of biomass, are carbohydrate-based polymers that can be broken down to low molecular weight sugars by hydrolysis using an acid catalyst. The acid-catalysed decomposition of the 6-sugar fragments (e.g., glucose) leads to 5- hydroxymethyl-2-furaldehyde as the intermediate product, which is subsequently rehydrated to give levulinic and formic acids as the final products. ydrolysis of the 5-sugars of hemicellulose may lead to furfural. In addition, other constituents in the hemicellulose matrix may produce side products like acetic acid and galacturonic acid [9]. Lignin, the third main constituent of lignocellulosic biomass, 64

A Kinetic Study on the onversion of Glucose to Levulinic Acid is a resin-like polymer matrix with various substituted phenolics present. During the acid hydrolysis, various acid soluble lignin-derived components may be formed, increasing the product complexity. The simplified reaction scheme given in Figure 4.2 does not explicitly show the reactions leading to the undesired insoluble-polymeric materials known as humins. BIMASS emicellulose ellulose Lignin 3 Acetic Acid Galacturonic Acid 2 Xylose 2 Glucose Acid-soluble compounds Glycolic Acid Furfural 2 5-ydroxymethyl-2-furaldehyde Formic Acid 3 LEVULINI AID Figure 4.2 Possible pathways and products of the acid-catalysed hydrolysis of a typical lignocellulosic material. As part of a larger project to develop efficient reactor configurations for the conversion of biomass to levulinic acid, we have initiated a study to determine the kinetics of all steps involved in the process. A stepwise approach was followed, starting with the conversion of 5-hydroxymethyl-2-furaldehyde (MF) to levulinic acid [10]. We here report our results of the kinetic study of the acid-catalysed decomposition of glucose in a broad range of process conditions, including the kinetics of the reactions leading to humins. The acid-catalysed decomposition of glucose has been studied by a number of authors [11-20]. owever, in all studies, only the rate of decomposition of glucose has been taken into account, often represented by a simple first-order reaction (Table 4.1). The development of a kinetic scheme for the conversion of glucose to levulinic acid, including the kinetics of by-product formation and incorporation of 65

hapter 4 MF as an intermediate, has not been reported to date. In addition, the rate equations provided in the literature are often valid for small temperature, substrate and/or catalyst concentration windows, whereas our study was performed in a large window for all variables. Table 4.1 verview of kinetic studies of glucose decomposition. T ( ) GL,0 acid rder in substrate and acid Reference R w [11] 1.02 2 4 170 190 5 wt % 2 S 4 (0.4 1.6 wt %) ( S ) GL 100 150 0.056 M l (0.35 M) R GL [12] 160 240 0.278 1.112 M 2 S 4 (0.025 0.8 N) R acid GL [13] R w [15] 0.8955 2 4 180 244 0.4 6 wt % 2 S 4 (0.5 4 wt %) ( S ) GL 1.33 R + 100 144 4 12 wt % 2 S 4 (4 20 wt %) ( ) GL [16] 170 230 0.006 0.33 M p 1 4 R + GL [17] 190 210 0.125 M p 1.5 2.2 R + GL [20] The results of this kinetic study will be used as input to obtain a full kinetic model for the acid-catalysed hydrolysis of the lignocellulosic biomass to levulinic acid. In addition, it allows the selection and the development of an efficient continuous reactor technology, in which the yield of levulinic acid is optimised and the amount of undesired by-products is reduced. 4.2 Experimental 4.2.1 Experimental procedure All chemicals (analytical grade) were purchased from Merck (Darmstadt, Germany) and used without further purification. The reactions were carried out in glass ampoules with an internal diameter of 3 mm, a wall thickness of 1.5 mm, and a length of 15 cm. An ampoule was filled at room temperature with a solution of glucose and sulphuric acid in the predetermined amounts (V liquid 0.5 cm 3 ). The ampoule was sealed with a torch. A series of ampoules was placed in a special rack and subsequently positioned in a constant-temperature oven (± 0.1 ), which was pre-set at the desired reaction temperature. At different reaction times, an ampoule was taken from the oven and directly quenched into an ice-water bath (4 ) to stop the reaction. The ampoules were opened, the reaction mixture was taken out and subsequently diluted with water to 10 ml. Insoluble humins, formed during the decomposition reaction, were separated from the solution by filtration over a 0.2 µm cellulose acetate filter (Schleicher & Schuell MicroScience, Dassel, Germany). The particle-free solution was then analysed using high performance liquid chromatography (PL). 66

4.2.2 Analytical methods A Kinetic Study on the onversion of Glucose to Levulinic Acid The composition of the liquid phase was determined using an PL system consisting of a ewlett Packard 1050 pump, a Bio-Rad rganic Acid column Aminex PX-87 and a Waters 410 differential refractive index detector. The mobile phase consisted of aqueous sulphuric acid (5 mm), which was set at a flow rate of 0.55 cm 3 min -1. The column was operated at 60. The analysis for a sample was complete in 40 minutes. A typical chromatogram is shown in Figure 4.3. The concentrations of each compound in the product mixture were determined using calibration curves obtained by analysing standard solutions of known concentrations. Glucose Levulinic Acid Formic Acid MF 0 5 10 15 20 25 30 35 40 Retention Time / min Figure 4.3 Typical chromatogram of a product mixture obtained from the acidcatalysed decomposition of glucose. Identification of side-products of glucose decomposition (i.e., the reversion products) was performed by connecting the PL system to an API3000 triple quadrupole L/MS/MS mass spectrometer (Perkin-Elmer Sciex Instruments, Boston, MA). The mass spectrometer was supplied with an atmospheric pressure ionisation source at a temperature of 400. 4.2.3 eat transfer experiments At the start-up of the reaction, the reaction takes place nonisothermally due to heating-up of the contents of the ampoule from the room temperature to the oven temperature. To gain insight into the time required for heating-up the reaction mixture and to compensate for this effect in the kinetic modelling studies, the temperature inside the ampoules as a function of the time during the heating-up process was determined experimentally. For this purpose, an ampoule equipped with a thermocouple was filled with a representative reaction mixture. The ampoule was then closed tightly using a special bolt-and-screw system to prevent evaporation of the liquid. The ampoule was subsequently placed in the oven at a 67

hapter 4 specified temperature and the temperature of reaction mixture was followed in time. Before and after an experiment, the amount of liquid inside the ampoule was measured to ensure that evaporation of the liquid did not occur. The experimental profiles at different temperatures were modelled using a heat balance for the contents in an ampoule: d ( M T ) dt p UA ( T T ) (4.1) t oven When assuming that the heat capacity of the reaction mixture is constant and not a function of temperature, rearrangement of equation (4.1) will give: dt dt UA ( T T ) h( T T t oven oven M p ) (4.2) Solving the ordinary differential equation (4.2) with the initial value T T i at t 0 leads to: ht ( T T ) T Toven oven i exp (4.3) The value of h was determined by fitting all experimental data at different oven temperatures (100 160 ) using a non-linear regression method and was found to be 0.596 min -1. Figure 4.4 shows an experimental and modelled temperature profile performed at an oven temperature of 100. Equation (4.3) was incorporated in the kinetic model to describe the non-isothermal behaviour of the system at the start-up of the reaction. 120 100 Experimental data Model according to equation (4.3) 80 T / o 60 40 20 0 2 4 6 8 10 12 14 16 t / min Figure 4.4 eating profile of the reaction mixture at T oven 100. 68

4.2.4 Determination of the kinetic parameters A Kinetic Study on the onversion of Glucose to Levulinic Acid The concentrations of all compounds involved in the decomposition reaction of glucose were obtained from PL analysis. All concentrations were normalized with respect to the initial concentration of glucose as follows: X Y Y ( ) GL,0 GL GL (4.4) GL,0 ( ) MF MF,0 MF (4.5) GL,0 ( ) LA LA,0 LA (4.6) GL,0 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. Details about this procedure can be found in the literature [21,22]. Error minimization to determine the best estimate of the kinetic parameters was performed using the MATLAB toolbox fminsearch, which is based on the Nelder-Mead optimisation method. 4.3 Results and discussions 4.3.1 Effects of process variables on the decomposition reaction of glucose A total of 22 experiments were performed in a temperature window of 140 200, 2S4 ranging between 0.05 and 1 M and GL,0 between 0.1 and 1 M. A typical concentration profile is given in Figure 4.5. MF was observed as an intermediate product in all experiments. The MF showed a maximum with respect to reaction time, although its maximum value is generally very low and less than 5% of the GL,0. This observation indicates that the conversion of MF to levulinic acid is much faster than the conversion of glucose to MF. 0.12 0.10 GL LA MF oncentration / M 0.08 0.06 0.04 0.02 0.00 0 20 40 60 80 100 120 t / min Figure 4.5 Typical concentration profile ( GL,0 0.1 M, 2S4 1 M, T 140 ). 69

hapter 4 The rate of glucose decomposition is a strong function of the temperature, and the time to reach 99 mol % of glucose conversion ranged between 12 h at 140 ( GL,0 0.1 M, 2S4 0.1 M) and 6 min at 200 ( GL,0 1 M, 2S4 0.5 M). The reaction rate is also considerably higher at higher acid concentrations. At 200, only dilute solutions of sulphuric acid (0.05 0.1 M) could be used as catalyst. Due to the very fast reaction rates at these conditions, representative sampling and analysis proved not possible. In all reactions, the formation of substantial amounts of black insolubleproducts, also known as humins, was observed. The composition and the formation pathways of these polymeric sugar-derived compounds are poorly understood [23]. The yield of levulinic acid is a function of the reaction time, temperature, GL,0 and 2S4. The highest yield was about 60 mol % at GL,0 0.1 M, 2S4 1 M and T 140. The yield of levulinic acid as a function of the reaction time and GL,0 (0.1 1M) is given in Figure 4.6. It is evident that more dilute solutions of glucose results in higher yields of levulinic acid. The effect of temperature on the yield is given in Figure 4.7. The maximum yield decreases when operating at the high end of the temperature window. The concentration of sulphuric acid only has a small effect on the yield of levulinic acid (Figure 4.8). 70 60 50 GL,0 0.1 M GL,0 0.5 M GL,0 1.0 M Y LA / mol % 40 30 20 10 0 0 20 40 60 80 100 120 t / min Figure 4.6 Yield of levulinic acid versus time for different GL,0 (T 140, 2S4 1 M). 70

A Kinetic Study on the onversion of Glucose to Levulinic Acid 70 60 50 T 140 o T 170 o T 200 o Y LA / mol % 40 30 20 10 0 1 10 100 t / min Figure 4.7 Yield of levulinic acid versus time at different temperatures ( GL,0 0.1 M, 2S4 0.5 M). 70 60 50 2 S 4 0.1 M 2 S 4 0.5 M 2 S 4 1.0 M Y LA / mol % 40 30 20 10 0 1 10 100 t / min Figure 4.8 Yield of levulinic acid versus time at different 2S4 ( GL,0 0.5 M, T 170 ). 4.3.2 Development of a kinetic model for glucose decomposition to levulinic acid The acid-catalysed decomposition of glucose (1) to levulinic acid (LA, 3) and formic acid (FA, 4) is schematically given in Scheme 4.1. In line with literature data and our experimental findings, MF (2) is considered as an intermediate product. 71

hapter 4 1-3 2 + + 2 2 + 3 + 2 3 4 Scheme 4.1 Acid-catalysed decomposition of glucose to LA. This simplified scheme does not take into account the formation of humins and other possible by-products. Substantial amounts of insoluble humins are formed in the course of the reaction. There are strong indications that the humins may be formed from both glucose and MF [13,17]. LA is not a source for humins. This was checked independently by reacting LA with 1 M sulphuric acid at 150 for 6 hours. It was found out that the concentration of LA was constant during the reaction. The rate of formation of humins from glucose and MF was included in the kinetic model. Fructose (5) is a known intermediate in the acid-catalysed decomposition of glucose [24-26]. It is likely formed from glucose (1) according to a reaction mechanism given in Scheme 4.2 [27,28]. ere, 1,2-enediol (6) is proposed as the common intermediate. owever, fructose could not be detected in our reaction mixtures. This is not surprising, as previous studies [29,30] have already shown that the dehydration of fructose to MF is much faster than that of glucose. Therefore, any fructose formed from glucose is expected to be converted to MF rapidly. 2 2 2 2 1 6 5-3 2 2 + 2 2 + 3 4 Scheme 4.2 Reaction mechanisms for the acid-catalysed decomposition of glucose to LA. 72

A Kinetic Study on the onversion of Glucose to Levulinic Acid ther possible by-products are so-called reversion products, like levoglucosan or 1,6-anhydro-β-D-glucopyranose (9), 1,6-anhydro-β-D-glucofuranose (10), isomaltose (11) and gentiobiose (12), is shown in Scheme 4.3. In acidic solutions, the acyclic form of D-glucose (1) exists in equilibrium with its anomeric forms, i.e., α- D-glucopyranose (7) and β-d-glucopyranose (8). The anomeric forms may be involved in a number of reactions leading to reversion products [27,31]. Intramolecular condensation reactions produce anhydro sugars, mainly levoglucosan and 1,6-anhydro-β-D-glucofuranose. Inter-molecular condensation reactions between two glucose units will give disaccharides such as isomaltose and gentiobiose. Several investigators [32,33] have also found and isolated other type of disaccharides, i.e., (1 2)-linked and (1 3)-linked dimers. Most studies [33,31] revealed that the yields of anhdyro sugars were higher than the yields of disaccharides, although other investigator [32] found opposite results. 10-2 - 2 7 11-2 1-2 9 12 8 Scheme 4.3 Reversion reactions of glucose in acid solutions. Some of the reversion products were detected in our experiments (Figure 4.9). Gentiobiose and levoglucosan were identified in the product mixture from the retention times of their pure compound, i.e., 8.4 min (gentiobiose) and 13.6 min 73

hapter 4 (levoglucosan). Isomaltose (7.7 min) and 1,6-anhydro-β-D-glucofuranose (12.6 min) were identified using L-MS. Gentiobiose Glucose 1,6-Anhydro-β-D-glucofuranose Isomaltose Levoglucosan 0 4 8 12 16 20 Retention time / min Figure 4.9 Identification of reversion products ( GL,0 1 M, 2S4 0.1 M, T 170, t 10 min). The reversion products were observed at the initial stage of the reactions. At full glucose conversion, reversion products were absent. The maximum concentrations of the reversion products in the course of the reaction were very low which made it very difficult to determine the concentrations of every component accurately. Therefore, we have not incorporated the reversion products in the kinetic model. n the basis of these considerations, the following kinetic model (Scheme 4.4) was applied to model the acid-catalysed decomposition of glucose. R 1G R 1 3 + R 2G R 2 umins umins Scheme 4.4 Reaction network for the acid-catalysed decomposition of MF to LA. 74 The reaction rates were defined using a power-law approach: 1G 1G a ( ) G R k (4.7) 2G 2G GL b ( ) G R k (4.8) GL

A Kinetic Study on the onversion of Glucose to Levulinic Acid 1 1 a ( ) R k (4.9) 2 2 MF b ( ) R k (4.10) MF The temperature dependence of the kinetic constants was defined in terms of modified Arrhenius equations: k k k 1G 2G 1 E T T ( ) 1G R α R T T G R + k exp 1RG E T T ( ) 2G R β R T T G R + k exp 2RG E T T ( ) 1 R α R T T R + k exp 1R (4.11) (4.12) (4.13) k 2 E T T ( ) 2 R β R T T R + k exp 2R (4.14) where T is a function of time defined in equation (4.3), and T R is the reference temperature (140 ). The catalytic effect of sulphuric acid is included in reaction rates in term of +, which can be calculated as follow: 2 ( K + K S K ) 1 + + + 2S 4 (4.15) 4 a,s 4 a,s 4 2 4 a, S 4 2 The term K a,s 4 in equation (4.15) represents the dissociation constant of (S 4 ), which ranges between 10-4.5 10-3.6 in the temperature window of 140 200 [34]. The kinetic constants and the reaction orders for the decomposition to MF to LA and FA have been determined earlier in our previous study [10]. The results are given in Table 4.2. These values were used as input for the kinetic model for glucose decomposition. Table 4.2 Kinetic parameter estimates for the MF decomposition to LA. Parameter Estimate k 1R (M 1-a -α min -1 ) a 0.340 ± 0.010 E 1 (kj mol -1 ) 110.5 ± 0.7 k 2R (M 1-b -β min -1 ) a 0.117 ± 0.008 E 2 (kj mol -1 ) 111 ± 2.0 a ( ) 0.88 ± 0.01 b ( ) 1.23 ± 0.03 α ( ) 1.38 ± 0.02 β ( ) 1.07 ± 0.04 a T R 140 75

hapter 4 In a batch system, the concentrations of the compound involved in decomposition reaction can be represented as follow: d dt GL d dt d dt MF ( 1G + R2G R ) (4.16) 1G ( R1 + R2 R ) (4.17) LA R (4.18) 1 4.3.3 Modelling results The best estimates of the kinetic parameters, as determined by minimization of the errors between all experimental data and the kinetic model, are shown in Table 4.3. The experimental data consisted of 660 data points (22 experiments, 10 samples per experiment, concentrations of LA, MF and glucose for each sample). omparisons of the experimental data and the output of the kinetic model demonstrate a good fit for a broad range of reaction condition (Figure 4.10). A parity chart (Figure 4.11) shows the goodness-of-fit between the experimental and model data. Table 4.3 Kinetic parameter estimates for the glucose decomposition to LA. Parameter Estimate k 1RG (M 1-a G -α G min -1 ) a 0.013 ± 0.001 E 1G (kj mol -1 ) 152.2 ± 0.7 k 2RG (M 1-b G -β G min -1 ) a 0.013 ± 0.001 E 2G (kj mol -1 ) 164.7 ± 0.6 a G ( ) 1.09 ± 0.01 b G ( ) 1.30 ± 0.02 α G ( ) 1.13 ± 0.01 β G ( ) 1.12 ± 0.02 a T R 140 76

A Kinetic Study on the onversion of Glucose to Levulinic Acid 0.12 0.6 0.6 0.10 0.5 0.5 0.08 0.4 0.4 / M 0.06 / M 0.3 / M 0.3 0.04 0.2 0.2 0.02 0.1 0.1 0.00 0 60 120 180 240 300 360 420 0.0 0 20 40 60 80 100 120 0.0 0 2 4 6 8 10 12 t / min t / min t / min (a) T 140 o, 2 S 4 0.1 M, GL,0 0.1 M (b) T 140 o, 2 S 4 1.0 M, GL,0 0.5 M (c) T 171 o, 2 S 4 1.0 M, GL,0 0.5 M 1.2 0.12 1.2 1.0 0.10 1.0 0.8 0.08 0.8 / M 0.6 / M 0.06 / M 0.6 0.4 0.04 0.4 0.2 0.02 0.2 0.0 0 3 6 9 12 15 0.00 0 3 6 9 12 15 0.0 0 1 2 3 4 5 6 7 t / min t / min t / min (d) T 171 o, 2 S 4 0.5 M, GL,0 1.0 M (e) T 200 o, 2 S 4 0.05 M, GL,0 0.1 M (f) T 200 o, 2 S 4 0.5 M, GL,0 1.0 M Figure 4.10 omparison of experimental data ( : GL, : MF, : LA ) and kinetic model (solid lines). 100 X GL 80 Y LA Predicted value (mol %) 60 40 20 0 0 20 40 60 80 100 Experimental value (mol %) Figure 4.11 Parity plot of all experimental data and model prediction. 77

hapter 4 4.4 Application of the kinetic model 4.4.1 Batch simulation and optimisation With the model available, it is possible to gain insight into the conversion, selectivity and yield of LA as a function of the process conditions. A typical batch time for 90 mol % glucose conversion as a function of the temperature is given in Figure 4.12. 10 5 10 4 batch time / min 10 3 10 2 10 1 10 0 10-1 100 120 140 160 180 200 T / o Figure 4.12 Reaction time to achieve 90 mol % of glucose conversion in isothermal batch reactor as a function of temperatures ( GL,0 0.1 M, 2S4 1 M). The kinetic model also allows the determination of the optimum reaction conditions to achieve the highest selectivity of LA. For this purpose equation (4.4) is differentiated to give: dx d GL GL (4.19) GL,0 ombination of equations (4.16) (4.19) leads to the following expressions: d dx d dx d dx GL GL (4.20) R GL,0 R R MF 1G 1 2 GL,0 (4.21) GL R1G + R2G R LA 1 GL,0 (4.22) GL R1G + R2G Equations (4.20) (4.22) were solved using the numerical integration toolbox ode45 in the MATLAB software package from 0 to 90 mol % glucose conversion. The selectivity of LA (σ LA) is defined as the ratio of the amount of the desired product (LA) formed and the key reactant (glucose) converted. 78

A Kinetic Study on the onversion of Glucose to Levulinic Acid Y LA LA,0 LA σ LA (4.23) GL,0 GL XGL Figure 4.13 shows the predicted σ LA as a function of the T and GL,0 at 90 mol % glucose conversion and a 2S4 of 0.5 M. The experimental data points are also given, demonstrating the good fit between experiments and model. 100 80 GL,0 0.1 M GL,0 0.5 M GL,0 1.0 M σ LA / mol % 60 40 20 0 100 120 140 160 180 200 T / Figure 4.13. Temperature effect on σ LA at X GL 90 mol % and 2S4 0.5 M. Symbols ( ), ( ) and ( ) represent experimental values of σ LA at GL,0 0.1, 0.5 and 1.0 M respectively. The σ LA is strongly temperature depending, with high temperatures leading to lower selectivity. This is in line with the observed activation energies for the main and side reactions. The activation energy for humins formation from glucose (Table 4.3) is significantly higher (164.7 kj mol 1 ) than all other activation energies, implying that the kinetics of this reaction is the most sensitive to temperature. To reduce humins formation, reactions at low temperature are favoured. It is also evident that higher GL,0 will lead to lower σ LA. This may be rationalised when looking at the orders in substrate for the various reactions involved. The order of glucose for the desired reaction to MF (1.09) is lower than that of the side reaction to humins (1.30), hence a higher GL,0 will lead to reduced σ LA. 4.4.2 ptimisation of continuous reactor systems The yield of LA in continuous reactors will be a function of typical process parameters (T, GL,0 and 2S4) and the extent of mixing in the reactor. In Figure 4.14, the yield of LA as a function of the glucose conversion at different temperatures (140 and 200 ) is provided for the two extremes with respect to mixing, i.e., a plug-flow reactor (PFR) and a continuous ideal stirred tank reactor (ISTR). 79

hapter 4 100 80 PFR ISTR T 140 o 100 80 PFR ISTR T 200 o Ψ LA / mol % 60 40 Ψ LA / mol % 60 40 20 20 0 0 20 40 60 80 100 0 0 20 40 60 80 100 X GL / mol % X GL / mol % Figure 4.14 omparison of LA yields in two ideal continuous reactors at different temperatures ( GL,0 0.1 M, + 0.5 M). ere the yield of LA, Ψ LA, is defined as the ratio between the amounts of LA formed during the reaction and of glucose fed into the reactor. Ψ LA (4.24) out in LA LA in GL The graphs were constructed from the mass balance design equations for the two model reactors in combination with the rate equations for the reactions. The reactor design equations of the PFR are similar to the design equations for the batch reactor (equations (4.19) (4.22)). The general reactor design equation for a ISTR reads: out in i i τ ISTR (4.25) R i The relationship between glucose conversion (X GL ) and τ ISTR is given by the following equation: X in GL GL τ ISTR (4.26) R1G + R2G Substitution of equation (4.26) into equation (4.25) and executing some rearrangement gives: out MF R R R 1G 1 2 in GLXGL R1G + R (4.27) 2G out LA R 1 in GLXGL R1G + R (4.28) 2G 80

A Kinetic Study on the onversion of Glucose to Levulinic Acid Based on the results shown in Figure 4.14 it is clear that the LA yields increases with the glucose conversion and that the yields in a ISTR are higher than in a PFR, particularly at high conversion levels. The yields at low temperature are higher than the yields at high temperature for both reactor configurations. To select the optimum operating conditions for the reactor, it is also necessary to consider the full process configuration. If a high glucose conversion is desired, e.g., when the separation of the LA from the glucose/humins/sulphuric acid mixture is difficult, it might be advantageous to apply a reactor with a high extent of backmixing (Figure 4.14). A number of options are available like a stirred tank reactor equipped with an impeller or a recycle reactor with a high recycle ratio. An important feature will be the scaling properties of the insoluble, humins byproducts. owever, information on this topic is lacking and further research will be required. In case separation of the product mixture is simple and cheap, it might be advantageous to operate at relatively low conversions of glucose to reduce reactor volume and associated costs. At low conversions, the yield is not a strong function of the extent of back-mixing (Figure 4.14) and other reactor configurations may be applied as well. 4.5 onclusions A kinetic model for the acid-catalysed decomposition of glucose in a broad operating window ( 2S4 0.05 1 M, GL,0 0.1 1 M, T 140 200 ) has been developed. Glucose decomposes in a consecutive reaction mode to give LA as the final product through MF as the intermediate. Glucose as well as MF decomposes in parallel reaction modes to give insoluble humins as the byproduct. The model implies that the highest yield of LA in continuous reactor configurations may be achieved by applying dilute solution of glucose, a high concentration of sulphuric acid as the catalyst and using a reactor configuration with a high extent of back-mixing. 4.6 Nomenclature a G : Reaction order of GL in the decomposition of glucose to MF ( ) α G : Reaction order of + in the decomposition of glucose to MF ( ) a : Reaction order of MF in the decomposition of MF to LA and FA ( ) α : Reaction order of + in the decomposition of MF to LA and FA ( ) A t : eat transfer area (m 2 ) b G : Reaction order of GL in the decomposition of glucose to humins ( ) β G : Reaction order of + in the decomposition of glucose to humins ( ) b : Reaction order of MF in the decomposition of MF to humins ( ) β : Reaction order of + in the decomposition of MF to humins ( ) GL : oncentration of glucose (M) 81

hapter 4 GL,0 : Initial concentration of glucose (M) + : oncentration of + (M) 2 S 4 : oncentration of sulphuric acid (M) MF : oncentration of MF (M) MF,0 : Initial concentration of MF (M) in : oncentration of the i th compound at the inflow (M) i out : oncentration of the i th compound at the outflow (M) i LA : oncentration of LA (M) LA,0 : Initial concentration of LA (M) p : eat capacity of reaction mixture (J g -1 K -1 ) E 1G : Activation energy of k 1G (kj mol -1 ) E 1 : Activation energy of k 1 (kj mol -1 ) E 2G : Activation energy of k 2G (kj mol -1 ) E 2 : Activation energy of k 2 (kj mol -1 ) h : eat transfer coefficient from the oven to the reaction mixture (min -1 ) k 1G : Reaction rate constant of glucose decomposition to MF (M 1-a G min -1 ) k 1RG : Reaction rate constant k 1G at reference temperature (M 1-a G -α G min -1 ) k 1 : Reaction rate constant of MF for the main reaction (M 1-a min -1 ) k 1R : Reaction rate constant k 1 at reference temperature (M 1-a -α min -1 ) k 2G : Reaction rate constant of glucose decomposition to humins (M 1-b G min -1 ) k 2RG : Reaction rate constant k 2G at reference temperature (M 1-b G -β G min -1 ) k 2 : Reaction rate constant of MF for the side reaction to humins (M 1-b min -1 ) k 2R : Reaction rate constant k 2 at reference temperature (M 1-b -β min -1 ) K a,s4 : Dissociation constant of (S 4 ) - ( ) M : Mass of the reaction mixture (g) R : Universal gas constant, 8.3144 J mol -1 K -1 R 1G : Reaction rate of glucose decomposition to MF (mol L -1 min -1 ) R 1 : Reaction rate of MF decomposition to LA and FA (mol L -1 min -1 ) R 2G : Reaction rate of glucose decomposition to humins (mol L -1 min -1 ) R 2 : Reaction rate of MF decomposition to humins (mol L -1 min -1 ) t : Time (min) T : Reaction temperature ( ) T i : Temperature of reaction mixture at t 0 ( ) T oven : Temperature of oven ( ) T R : Reference temperature ( ) U : verall heat transfer coefficient (W m -2 K -1 ) w 2 S 4 : Weight percentage of sulphuric acid (%) X GL : onversion of glucose (mol %) Y MF : Yield of MF (mol %) Y LA : Yield of LA (mol %) 82

A Kinetic Study on the onversion of Glucose to Levulinic Acid Greek symbols σ LA : Selectivity of LA (mol %) τ ISTR : Residence time of ISTR (min) Ψ LA : Yield of LA in continuous reactors (mol %) 4.7 References [1] Leonard, R.., Levulinic Acid as a Basic hemical Raw Material. Ind. Eng. hem. 1956, 48 (8), 1330-1341. [2] Kitano, M.; Tanimoto, F.; kabayashi, M., Levulinic acid, a new chemical raw material; its chemistry and use. hem. Econ. Eng. Rev. 1975, 7, 25-29. [3] Thomas, J. J.; Barile, G. R. In onversion of cellulose hydrolysis products to fuels and chemical feedstocks., 8th Symposium on Energy from Biomass and Waste, Lake Buena Fista, FL, 1984; pp 1461-1494. [4] Ghorpade, V. M.; anna, M. A., Industrial applications for levulinic acid. In ereal Novel Uses and Processes., ampbell, G. M.; Webb,.; McKee, S. L., Eds. Plenum Press: New York, 1997; pp 49-55. [5] Timokhin, B. V.; Baransky, V. A.; Eliseeva, G. D., Levulinic Acid in rganic Synthesis. Russ. hem. Rev. 1999, 68 (1), 73-84. [6] Bozell, J. J.; Moens, L.; Elliott, D..; Wang, Y.; Neuenscwander, G. G.; Fitzpatrick, S. W.; Bilski, R. J.; Jarnefeld, J. L., Production of levulinic acid and use as a platform chemical for derived products. Resour. onserv. Recycl. 2000, 28, 227-239. [7] Werpy, T.; Petersen, G. Top Value Added hemicals from Biomass Volume I-Results of Screening for Potential andidates from Sugars and Synthesis Gas.; NREL/TP-510-35523; National Renewable Energy Laboratory (NREL): 2004. [8] Grethlein,. E., hemical Breakdown of ellulosic Materials. J. Appl. hem. Biotechnol. 1978, 28, 296-308. [9] Danon, B.; Girisuta, B.; eeres,. J., A Kinetic Study of the Acidic ydrolysis of Water yacinth to Levulinic Acid. In University of Groningen: Groningen (Netherlands), 2005. [10] Girisuta, B.; Janssen, L. P. B. M.; eeres,. J., A Kinetic Study on the Decomposition of 5- ydroxymethylfurfural into Levulinic Acid. Green hem. 2006, 8, 701-709. [11] Saeman, J. F., Kinetics of Wood Saccharification - ydrolysis of ellulose and Decomposition of Sugars in Dilute Acid at igh Temperature. Ind. Eng. hem. 1945, 37, 43-52. [12] eimlich, K. R.; Martin, A. N., A Kinetic Study of Glucose Degradation in Acid Solution. J. Am. Pharm. Assoc. 1960, 49, 592-597. [13] McKibbins, S.; arris, J. F.; Saeman, J. F.; Neill, W. K., Kinetics of the Acid atalyzed onversion of Glucose to 5-ydroxymethyl-2-furaldehyde and Levulinic Acid. Forest Prod. J. 1962, 12, 17-23. [14] Fagan, R. D.; Grethlein,. E.; onverse, A..; Porteus, A., Kinetics od the Acid ydrolysis of ellulose Found in Paper Refuse. Environ. Sci. Technol. 1971, 5, 545-547. [15] Smith, P..; Grethlein,. E.; onverse, A.., Glucose Decomposition at igh-temperature, Mild Acid, and Short Residence Times. Solar Energy 1982, 28, 41-48. 83

hapter 4 [16] Bienkowski, P. R.; Ladisch, M. R.; Narayan, R.; Tsao, G. T.; Eckert, R., orrelation of Glucose (Dextrose) Degradation at 90 to 190-Degrees- in 0.4 to 20-Percent Acid. hem. Eng. omm. 1987, 51, 179-192. [17] Baugh, K. D.; Mccarty, P. L., Thermochemical Pretreatment of Lignocellulose to Enhance Methane Fermentation.1. Monosaccharide and Furfurals ydrothermal Decomposition and Product Formation Rates. Biotechnol. Bioeng. 1988, 31, 50-61. [18] Bergeron, P.; Benham,.; Werdene, P., Dilute Sulfuric-Acid ydrolysis of Biomass for Ethanol-Production. Appl. Biochem. Biotechnol. 1989, 20-1, 119-134. [19] Mosier, N. S.; Ladisch,. M.; Ladisch, M. R., haracterization of acid catalytic domains for cellulose hydrolysis and glucose degradation. Biotechnol. Bioeng. 2002, 79, 610-618. [20] Xiang, Q.; Lee, Y. Y.; Torget, R. W., Kinetics of glucose decomposition during dilute-acid hydrolysis of lignocellulosic biomass. Appl. Biochem. Biotechnol.. 2004, 113-16, 1127-1138. [21] Bard, Y., Nonlinear Parameter Estimation. Academic Press: New York, 1974; p 61-71. [22] Knightes,. D.; Peters,. A., Statistical analysis of nonlinear parameter estimation for Monod biodegradation kinetics using bivariate data. Biotechnol. Bioeng. 2000, 69, 160-170. [23] orvat, J.; Klaic, B.; Metelko, B.; Sunjic, V., Mechanism of levulinic acid formation. Tetrahedron Lett. 1985, 26 (17), 2111-2114. [24] arris, D. W.; Feather, M. S., Evidence for A -2-->-1 Intramolecular ydrogen Transfer During Acid-atalyzed Isomerization of D-Glucose to D-Fructose. arbohydr. Res. 1973, 30, 359-365. [25] arris, D. W.; Feather, M. S., Intramolecular -2-->-1 ydrogen Transfer-Reactions During onversion of Aldoses to 2-Furaldehydes. J. rg. hem. 1974, 39, 724-725. [26] arris, D. W.; Feather, M. S., Studies on Mechanism of Interconversion of D-Glucose, D- Mannose, and D-Fructose in Acid Solution. J. Am. hem. Soc. 1975, 97, 178-182. [27] van Dam,. E.; Kieboom, A. P. G.; van Bekkum,., The onversion of Fructose and Glucose in Acidic Media - Formation of ydroxymethylfurfural. Starch-Starke 1986, 38, 95-101. [28] Moreau,.; Durand, R.; Razigade, S.; Duhamet, J.; Faugeras, P.; Rivalier, P.; Ros, P.; Avignon, G., Dehydration of fructose to 5-hydroxymethylfurfural over -mordenites. Appl. atal. A: General 1996, 145, 211-224. [29] Kuster, B. F. M.; van der Baan,. S., Dehydration of D-Fructose (Formation of 5- ydroxymethyl-2-furaldehyde and Levulinic Acid).2. Influence of Initial and atalyst oncentrations on Dehydration of D-Fructose. arbohydr. Res. 1977, 54, 165-176. [30] Kuster, B. F. M., 5-ydroxymethylfurfural (MF) - A Review Focusing on Its Manufacture. Starch-Starke 1990, 42, 314-321. [31] elm, R. F.; Young, R. A.; onner, A.., The Reversion Reactions of D-Glucose During the ydrolysis of ellulose with Dilute Sulfuric-Acid. arbohydr. Res. 1989, 185, 249-260. [32] Thompson, A.; Anno, K.; Wolfrom, M. L.; Inatome, M., Acid Reversion Products from D- Glucose. J. Am. hem. Soc. 1954, 76, 1309-1311. [33] Peat, S.; Whelan, W. J.; Edwards, T. E.; wen,., Quantitative Aspects of the Acid Reversion of Glucose. J. hem. Soc. 1958, 586-592. [34] Dickson, A. G.; Wesolowski, D. J.; Palmer, D. A.; Mesmer, R. E., Dissociation-onstant of Bisulfate Ion in Aqueous Sodium-hloride Solutions to 250-Degrees-. J. Phys. hem. 1990, 94, 7978-7985. 84